The Search for MH370 Begins Again!

What many of us have been encouraging has finally transpired–the seabed search for the wreckage of MH370 has been re-started. The search vessel Seabed Constructor has just arrived in the new search area, outfitted with a team of eight autonomous underwater vehicles (AUVs). Ocean Infinity, the company under contract with Malaysia to conduct the search, has agreed to start by searching the 25,000 square kilometers identified by the ATSB and CSIRO as most likely. Included in that area are three locations that CSIRO has identified as high priority, as determined from satellite images of floating objects and complex drift models. Last August, the highest priority location was described by CSIRO’s David Griffin in these words: We think it is possible to identify a most likely location of the aircraft, with unprecedented precision and certainty. This location is 35.6[degrees south], 92.8 [degrees east]. 

At Ocean Infinity’s touted scan rate of 1,200 square kilometers per day, the entire 25,000 square kilometers would be completed in 21 days of searching, and the highest priority area of 5,000 square kilometers would be completed in less than a week.

The nominal location of the 7th arc that is shown in the figure above is a based on the assumption that the last transmission from the aircraft occurred at 20,000 ft, and our best estimate of the final BTO value is 18390 μs. The final two BTO values that were used for the best estimate occurred when the SATCOM of MH370 initiated a log-on to Inmarsat’s satellite network at 00:19 UTC on March 8, 2014, minutes after the engines stopped due to fuel exhaustion. (The re-boot of the SATCOM likely occurred after the APU automatically started and briefly supplied electrical power.)

I performed a statistical analysis of previous log-on events that occurred on March 7, 2014, including those that occurred on flight MH371 from Beijing to Kuala Lumpur. Using the results of this analysis, the final two BTO values from MH370 were first corrected and then appropriately weighted based on their respective uncertainties in order to arrive at the best estimate of 18390 μs.  The procedure was briefly described in a previous comment of mine.

Also shown in the figure are two other arcs that are positioned at +/- 25 NM from the nominal location of the 7th arc. These might serve as limits for some parts of the search. The figure shows that the +/-25 NM limits do not correspond to the boundaries of the 25,000 square kilometer area that was previously identified. In fact, the highest priority location identified by CSIRO (labeled CSIRO Priority 1) falls slightly outside of the 25-NM outer limit.

If not found in the initial 25,000 square kilometer area, the contract with Ocean Infinity indicates that the search will continue further northeast along the 7th arc. Likely, the search will continue along the 7th arc as far northeast as time and weather permit.

I often get asked whether I believe this search will succeed in finding the wreckage of MH370. I long ago arrived at the conclusion that based on the evidence we have, it is impossible to determine any one location with a high level of certainty, and I stopped trying. The satellite data and the drift models allow a broad range of possible impact sites. Within that range, there are at best some “warm spots” that are based on assumptions about navigation inputs. So, it becomes a numbers game–the more area searched, the higher probability of finding the wreckage. I subjectively believe there is a 33% chance of finding the wreckage in the first 25,000 square kilometers. If there is time and money to search at +/- 25 NM from the 7th arc all the way to a latitude of 26S, I subjectively put the chances of success at around 67%. That might seem like bad odds, but realistically, that’s higher than they’ve ever been.

The highest priority location identified by CSIRO is about 66 NM from Seabed Constructor’s present location, and might be reached within the next day. We’ll all be watching.

[Don Thompson reminds me that the data from an AUV mission is available only after the AUV is recovered after the completion of a dive, which could last 2+ days, based on the endurance of the batteries. It might take another 18 hours to analyze the data. That means that although the AUVs could reach “CSIRO Priority 1” by tomorrow, we would not know until Wednesday or Thursday whether or not the debris field was found.]

1,025 Responses to “The Search for MH370 Begins Again!”

  1. Ulric says:

    I’m still gravely concerned that the proposed search area does not extend far enough north. I guess that with the rate Seabed Constructor can scan, we will find out relatively quickly whether they are looking in the right place.

    Do we know whether the search area will be extended in the event of the wreck not being discovered?

  2. Victor Iannello says:

    @Ulric: I think it depends on time, weather, and if the equipment performs as anticipated. I would be surprised if the search doesn’t proceed to the north past 30S.

  3. airlandseaman says:

    Victor: I can’t put a number on it, but I am more optimistic. Josh, Andy and crew have the right stuff. It is hard to be patient on such an exciting day, but patient we must be. It may take the next 90 days to locate the debris. But they have a good chance by April…whatever number you want to associate with good!

  4. Don Thompson says:

    Considering the information described in the SUT/Perth conference presentation it will be Wednesday (UTC) before premininary images of the seabed are available.

    The AUV missions likely to be mid-50 hours, recover data from AUV disk packs and post-process for preliminary results adds some further time.

    This is consistent with speculation that Maersk Mariner is ferrying the two TLDM (MY navy) observers for a rendezvous with Constructor.

  5. Ulric says:

    @Victor

    Thanks for that reassurance. I’d hate to end up in that terrible limbo which would exist following two negative searches.

  6. Victor Iannello says:

    @Don Thompson: Thank you for reminding us that there is a distinction between when an area of the seabed is scanned and when the data is analyzed. There could be a 2+ day lag.

  7. Rob says:

    @Victor

    To DennisW “No. Some of us voluntarily help when we can.”

    Same here. Between us, we will make sure OI find the plane, whatever our differences of opinion and interpretation, that is the common goal.

  8. IR1907 says:

    @victor,

    66% ? Why ?

  9. Victor Iannello says:

    IR907: It might be further north, further out from the arc, or scanned and missed (past or future efforts).

  10. Richard Godfrey says:

    @Victor

    I am glad to see that Ocean Infinity have restarted the search for MH370.

    This is a major milestone!

    I put the chances of Ocean Infinity finding MH370 in the next 90 days at around 80%.

  11. ErikN says:

    @all Many thanks to @Victor for driving this amazing dialogue here. Herding so much brainpower and its concomitant ego into one discussion is a monumental task. I am probably less optimistic than Victor or Mike about the probabilities. But I am as hopeful as anyone that some way, some how, somewhere 9M-MRO is found. We have the data and evidence that we have. However good it is, and however good we can interpret it–let’s pray she’s found. I think this is the last great hope for a long time into the future. But a great hope it is. #FTW Crew and Scientists of Ocean Infinity and M/V Seabed Constructor!

  12. Victor Iannello says:

    @ErikN: Thank you for your comment. Egos here sometimes clash (mine included), but the tone here is for the most part collegial, factual, and helpful. I think that collectively contributors here have helped advance our understanding of the data, and have also helped keep the search alive.

  13. DennisW says:

    @Ulric

    Thanks for that reassurance. I’d hate to end up in that terrible limbo which would exist following two negative searches.

    It would be “dissappointing” to say the least. I would probably take a breather, but not give up. MH370 is, indeed, a fascinating puzzle. So far most of the heavy lifting has been done by physics, ee, cs, and math types (nice to have professionals like Andrew and David on board as well). I am bit surprised by the relative lack of input from the pysch, ME, and criminology types.

    As far as what to do if the plane is not found in the current search, I am mulling that over as well. AI, in particular Google’s recent Open Sourcing of Tensorflow, might be worth a long look. I can’t click on an Amazon email without buying something. Their AI engine has me completely calibrated. Sorry, just babbling during the “wait”.

  14. Indeed ! very exiting moments !
    I guess we all have our fingers crossed that something will come out, no matter our views.
    Relief of the families first !
    Thanks for the living exchanges 🙂

  15. DennisW says:

    @Victor

    BTW, great graphic in your recent post above.

  16. Victor Iannello says:

    @DennisW: Thanks, Dennis.

    Here are some other interesting facts. If we use the 20,000 ft, 18390 μs arc as the reference, then CSIRO-1 is 25 NM to the outside of the arc, and CSIRO-2 and -3 are 22 NM to the inside of the arc. At those latitudes, the search area already extended to about 22 NM outside the arc, and 17 inside the arc, so if CSIRO is right, the debris field would be located just miles beyond the 120,000 sq km previously searched. It would also imply the plane glided as much as 20 NM after the last BFO values. That’s not consistent with an aircraft that was already descending at 15,000 fpm and in a 0.65g downward acceleration. Those observations dampen my enthusiasm for the CSIRO locations, despite the floating objects that were observed by the French satellite, but who knows?

  17. George Tilton says:

    There was a very interesting drift study by the University of Washington on the ScienceDaily website.

    Summary: An experiment using hundreds of plastic drifters in the Gulf of Mexico shows that rather than simply spread out, as current calculations would predict, many of them clumped together in a tight cluster.

    I wonder what impact on future drift models will result from this?

    https://www.sciencedaily.com/releases/2018/01/180118175318.htm

  18. Don Thompson says:

    @ George Tilton,

    Concerning particles spreading, or not.

    The Global Drifter Program drops single drifters into the ocean, spread temporally and geographically.

    I can’t recall where I heard/read it, but David Griffin made a similar comment about ‘clumping’. My interpretation may not be wholly accurate but oceanography studies do now include sea surface height, from radar altimetry, so as to consider effect of identified troughs and eddies on the ocean surface in the mesoscale processes. Particles do congregate in the troughs and eddies.

  19. M Pat says:

    Hi Victor

    With the renewed search getting under way, I just wanted to post a quick note to express appreciation for the tremendous job you have done in hosting this blog. It has turned out to be everything I hoped it would be. I read it every day and I believe I have not missed a single comment. It is a testament to the knowledge and ingenuity of key contributors that almost 4 years after the loss of the aircraft there is almost every day a new nugget of information to absorb.

    I remember when the end of the first 60K km2 search was in sight, and again at the end of the final 120K km2, some of us dared to speculate that a private search effort might pick up where the official attempt left off. It is very exciting that this is actually occurring. Never say never.

    In the early days of the search, we all hoped that the Venn diagram of intersecting constraints (BTO/BFO/fuel consumption/navigation modes/performance envelope/drift analysis, etc.) could meaningfully restrict the area of seabed to be scanned to a fraction of the original ATSB ‘wide search area’ (that famous grey box). We have come a long way since. But while analysis of terminal BFO data makes a compelling case for restricting the search width around the 7th arc as has frequently been suggested, I was very happy indeed to read your penultimate paragraph on the true consequences of the fundamental uncertainty in the data that might allow a refinement in the range of possible positions along the arc.

    The apparently high degree of confidence in the first 25K km2 worries me. It is reassuring that regardless of outcome, OI are prepared for a systematic search towards the north as far as time and funds allow.

    Good luck to OI in the coming days.

  20. Victor Iannello says:

    @M Pat: Thank you for contributing your wonderful comment. Although we should remain realistic about the prospects of finding the wreckage, this is all very exciting.

  21. lkr says:

    The Other Site pananoia/cynicism surges: “Obviously if they find anything right of the bat (in Gibsonian fashion) it will be without any observers on board…” –JW

  22. Victor Iannello says:

    @GlobusMax: That’s an interesting analysis that will take some time to fully digest. Obviously, your thinking is evolving.

  23. Victor Iannello says:

    @lkr: The antics there are predictable and best ignored.

  24. rogan josh says:

    Would the initial rise in altitude have likely killed everyone on board, including the captain before continuing on in to the SIO?

  25. lkr says:

    @VI: I know, but can’t resist. The sanity over here is remarkable — as has been said today, there’s a basic respect both for data and for other contributors.

    As with others, crossing my fingers…

  26. Victor Iannello says:

    @rogan josh: Despite the press reports at the time, there is no evidence for a steep climb.

  27. TBill says:

    @Victor
    I like your choice of words “subjectively feel” 33% to 66% chance…remind me to use that word next time. I had estimated 25% so I accept 33% for the first part.

    Not much energy for new path projections right now, since we have the real world search. But if they find the debris or not, that will spur a new set of activity to project path. It would be a lot less speculative though if we actually knew the end point.

    @sk999’s estimate of 32.4 South based on best fit of the sat data sounds good to me right now, in part subjectively.

  28. GlobusMax says:

    @Victor

    My thinking is more catching up. Yours and I believe Guarded Don’s subjective probability estimates are pretty accurate if my 16 hours or so analyses are any good.

    May I used your great circle route graphic in a future post assessing performance limit at 40S? (with attribution and link to post).

  29. Victor Iannello says:

    @GlobusMax: Sure.

  30. airlandseaman says:

    Here’s a link to Blaine’s interview on ABC-OZ. It works for some, and not others. Geo blocking? SharePoint issue? Not sure. Give it a try. https://goo.gl/nXtco5

  31. ventus45 says:

    @ALSM It goes to a microsoft sign in page for me, and I am in Sydney.

  32. airlandseaman says:

    OK. Now on Google Drive. Try this: https://goo.gl/WZvkQH

  33. sk999 says:

    ALSM,

    I am half a world away from Sydney, but get the same response as ventus45. Once again, thank you Microsoft.

  34. ventus45 says:

    @ALSM – Google drive ref gives a video – downloading.

  35. ventus45 says:

    @ALSM – got it – a reasonable media report for once.

  36. airlandseaman says:

    The google drive link seems to work, but it is only available as a download so far (409mb). It should finish processing the video for live play soon.

  37. Brock McEwen says:

    @IG members who monitor this forum: I am looking for probabilities from you. Each of these probabilities…

    …MUST be conditioned on hypothetical impact taking place within the current search zone on which Victor places a 67% chance of success, but

    …must NOT be conditioned on the ISAT data (i.e. pretend it has never existed)

    Here are the probabilities I wish to crowd-source:

    Pr(Indian radar would not detect the path)
    Pr(Indonesian radar would not detect the path)
    Pr(Australian radar would not detect the path)
    Pr(US detection capabilities would not detect the path)
    Pr(Chinese detection capabilities would not detect the path)
    Pr(Russian detection capabilities would not detect the path)
    Pr(intensive Australian shoreline debris searches in Q4 2014 would all come up empty)
    Pr(the “Roy” piece would be discovered where it was, when it (first) was)

    Please feel free to specific any path you like; all I ask is that you use the SAME path to assess each probability.

    For crystal clarity: I am NOT looking for “100%, because the ISAT data says it must be so”, or “if we find wreckage, 100%, otherwise zero”. These are not helpful to the scientific exercise I am undertaking. The exercise requires that the ISAT data is NOT known, but the search zone IS. I am asking you to think like an outsider who has no access to search zone decision-making, an is trying merely to square the search zone to the radar and drift data.

    Many thanks in advance for any time and consideration you are able to give to these questions.

  38. Michael John says:

    In regards to Maersk Mariner it would appear that the ship has a contract with Quadrant Energy starting in March & lasting approximately 150-200 days. It will be working on the North Western shelf just off the coast of Western Australia.

    Whilst it’s reasons for a rendezvous with SeaBed Constructor are seemingly unclear (there is nothing in the public realm about the meet up) I fail to see that it is something to be suspicious of (As suggested in other Blogs).

  39. Mick Gilbert says:

    @Michael John

    The contract for Maersk Mariner (and Maersk Master) with Quadrant doesn’t commence until March.

    Since August last year Mariner has been operating under contract to Woodside Energy as an offshore support vessel for their North West Shelf Project. My understanding is that Mariner is ferrying two Tentera Laut Diraja Malaysia (TLDM) – Royal Malaysian Navy – observers to the Seabed Constructor and then it will proceed to Dampier on Western Australia’s north-west coast.

  40. Ge Rijn says:

    @MPat

    Nice to read your post and to hear from you again. That has been some years back in my memory..
    Like to mention your contribution few years ago with your drift-analysis has been quite important IMO. To me it was the first study which gave clear indications the crash area had to be north of 36S.
    To me your analysis marked a turning point in this regard.
    Just wanted to mention this to your credit.

  41. Richard Godfrey says:

    Here is graphic showing the Ocean Infinity Search Area Progress for 21st January 2018:

    https://www.dropbox.com/s/qv16vylwlsigrq6/Ocean%20Infinity%20Search%20Area%20Progress%20Map%2021%20Jan%202018.pdf?dl=0

  42. Ge Rijn says:

    @ALSM @others

    In the video the commentator states; “..there were more clues; 35S fits with a different A/P-setting Malaysian pilots often use..”

    Can you or someone else explain what kind of different A/P setting is suggested with this remark?

  43. Ge Rijn says:

    ..the comment starts after 3:21..

  44. @Brian Anderson
    @TBill,

    I realise I did not answer your post on January 10, 2018 at 8:27 pm. Sorry for this.
    So please find here the answer keeping in mind that our trajectory should be consirered in 5D: 4D + variable speed.

    Because the aircraft was descending from FL270 to FL150 the distance Aircraft-Statellite in the proposed piloted trajectory varied fast. The minimum LOS is about 36,733km (BTO=11,415) and occured at about 19:22:50 UTC at altitude 25,100 ft. (at that time the ground speed was about 350kn).

    You can recompute the data based on our report available at http://www.mh370-captio.net where the descent data is provided.

    To answer Brian’s objection “flight direction from outside to inside at Arc2 crossing”: in our trajectory the aircraft was descending and thus the radius of the evolving in time spheres Satellite-aircraft increased in the vicinity of Arc2 at 15,000ft, this implies that it flew from the inside to the outside of the successive spheres as a function of time.

    Our assumptions:
    1) flight vertical projection on tangent plan is in straight line
    2) decreasing altitude
    3) Variable speed

    I hope this data convinces you that a piloted trajectory cannot be rejected from Brian’s computation which is only valid for a quasi planar flight at constant altitude and constant speed.

  45. Victor Iannello says:

    @Ge Rijn: I believe the comment refers to the DSTG studies in which those paths using LNAV and CTT modes produced a hot spot near 38S, and CMH, CTH, and CMT modes terminate further north. If you eliminate the LNAV and CTT possibilities, then the posterior distribution shifts north. CMH would be a typical setting. It also requires a change in speed setting, which the DSTG models can accommodate. (The DSTG model does not include the possibility of an ongoing descent at 18:40, nor a “structured” hold manoeuver.)

  46. Victor Iannello says:

    @Brock McEwen said: @IG members who monitor this forum: I am looking for probabilities from you. Each of these probabilities…
    …MUST be conditioned on hypothetical impact taking place within the current search zone on which Victor places a 67% chance of success, but
    …must NOT be conditioned on the ISAT data (i.e. pretend it has never existed)

    This is a ridiculous question, predicated on a hypothetical that is extremely unlikely. The ISAT data is the best, i.e., most precise, data we have. Without that data, there is no basis for conducting the subsea search. You refuse to accept the data, yet you offer no solid reason to reject it.

  47. Victor Iannello says:

    Michael John: Whilst it’s reasons for a rendezvous with SeaBed Constructor are seemingly unclear (there is nothing in the public realm about the meet up) I fail to see that it is something to be suspicious of (As suggested in other Blogs).

    If you haven’t yet discovered the pattern, I’ll reveal it. You’ll also be able to use the pattern to predict future suggestions:

    Whenever there is uncertainty, attribute that uncertainty to malicious Russian actions. Uncertainty about financing of OI? Look for Russian connections. Uncertainty about a rendezvous with a support ship? Must be Russians handing off debris to be planted. Satellite data disproves path was a Russian diversion to Kazakhstan? Russia must have spoofed the data. Recovered debris confirms crash in the SIO? Russian agents must be planting debris. Independent investigators confirm data consistent with a crash in the SIO? Investigators are colluding with Russians.

    Of course, this comment just proves I’m working with the Russians.

  48. Victor Iannello says:

    @ALSM: The news report you provided includes the statement from David Griffin that the lack of debris recovered along the shores of Western Australia (WA) is an important discriminator of the latitude of the impact. There seems to be some disagreement about this statement among prominent oceanographers. Chari Pattiaratchi believes that offshore currents prevent debris in the vicinity of WA from actually coming ashore, and therefore the lack of recovered debris provides no indication of the latitude of impact. That’s quite a difference of opinion on how to interpret the evidence.

  49. Ge Rijn says:

    @VictorI

    To shime in on this. @MPat’s study I referred to in my comment to him also includes a link to a CSIRO-study.
    Both studies prove historical buoys from the Global Drifter Program that past the search area between 39S and 32S beeched on Australian shores.
    In @MPat’s study this were 7 (undrogued) buoys and the CSIRO/Griffin study 12.
    It turned out though that only buoys that crossed the search area south of 36S reached Australian beeches.
    North of this latitude none of the 177 buoys that past the projected search area did reach Australia:

    Once more @Mpat’s article:

    http://jeffwise.net/2016/07/07/guest-post-where-mh370-search-area-debris-has-historically-gone/

  50. TBill says:

    @Victor @Ge Rijn
    “I believe the comment refers to the DSTG studies in which those paths using LNAV and CTT modes produced a hot spot near 38S, and CMH, CTH, and CMT modes…”

    OK I had the same question about what the news report was saying…I like CTH, possibly CMH. For some reason (either pilot action or unexplained aircraft flying performance after Arc5) the final end point may have thrown off the Bayesian analysis.

    I suggest if the Bayesian analysis had set Arc5 as the final “predictable” Arc, then 180S may emerge as the “hot spot”. That why I think @sk999’s 32.4 South estimate might be a valid technique, as it presumably allows for deviation after Arc5.

  51. TBill says:

    @Jean-Luc
    I am OK with your path in theory. In practice I have adopted the hypothesis of a straight flight South, because (from many FS9 flights) I believe Arc3 to Arc5 is almost exact fit to 180S CTH from about ISBIX. My 1941 crossing is quite similar to yours…I do think the 1941 BFO suggests change is still happening.

    Ultimately it gets into judgement (picking a scenario). The Xmas island scenario needs a little more evidence to sway the search, even though TimR has long said there may have been strong rumors of a clandestine plan like that. The “plan” could still be true but the flight did not seem to follow the plan very well, for unexplained reasons.

  52. Victor Iannello says:

    This French article from Le Figaro reveals that Ocean Infinity has engaged the French-British company Deep Ocean Sea (DOS) to assist in the search, under the direction of Nicolas Vincent. From the article:

    According to our information, Ocean Infinity is associated with Deep Ocean Search, a Franco-British company, which recently found the black boxes of the plane Egyptair that had crashed at sea on May 19 with 66 people on board. The group has 20 oceanographers, all French, composed of veterans of the French Navy and former graduates of Intechmer, the French school of science and techniques of the sea. DOS was founded by the Britain John Kingsford in 2010, and the Frenchman Nicolas Vincent directs the operations. In autumn 2013, the team of French researchers working for DOS had found the wreck of the City of Cairo, more than 5,150 meters under the sea. Never had any cargo been recovered to such depth. Thus, the company brings its expertise to Ocean Infinity to carry out this mission.

  53. Ge Rijn says:

    @TBill

    What VictorI also said is the DSTG did not include the possibility of an ongoing descent after 18:40 or a holding pattern.
    I understand they (DSTG) took 18:40 as the latest possible FMT. Then LNAV or CTT modes lead to the 38S hot spot.
    It seems to me that when an ongoing descent after 18:40 or a holding pattern with a FMT (or MT) farther (much farther) north-west is considered, still a LNAV or CTT mode could be possible leading to latitudes north of 38S.

    CMH/CTH routes seem to me only possible if there was a route discontinuety at/after FMT without any further pilot inputs after.
    Those modes suggest IMO the pilot had no specific end-coördinates/location/track in mind. This sure could have been the case but we can’t know this for sure.

    I still believe the PIC had a specific end-location in mind. Best way to fly to such a location would be LNAV or CTT in my (limited) understanding regarding the expected great magnetic deviations on those far south latitudes and the wind-changes along the route.

    ~32.4S is also my favorite latitude as I suggested often but mainly for different reasons (drift-analysis, PIC motives).
    My greatest concern is OI will limit their search width to +/- 25Nm when they are going to start searching from ~33S till ~32.2S.
    IMO they by then should at least widen the search width till the 97th longitude and better a bit further.

    We’ll see. This is at least three steps away thinking.
    I hope they find the plane tomorrow..

  54. Ge Rijn says:

    @VictorI

    This is great news again. Awesome!

  55. DennisW says:

    @TBill

    The Xmas island scenario needs a little more evidence to sway the search,

    Where do you think this additional information/evidence is likely to come from? I know of no activity that is in the process of generating any additional information. While there is information that has not been publicly disclosed, the people engaged in the search presumably have it.

  56. TBill says:

    @DennisW
    Right now there is probably not enough justification for moving part of the Arc7 search up there. It would take a whistle-blower or corroboration of some kind. If the current search is unsuccessful, then the concept probably gains some strength.

  57. Ge Rijn says:

    @TBill

    Even a 26S latitude would need almost impossible extraordinairy drift-patterns to fit the arrival data of the flaperon. Let alone C.I. or other latitudes more north of 26S. Not even mentioning no debris was found on Indonesian Islands or other shores in the NIO.

    Even Richard Godfrey’s model from ~30S could only fit the flaperon arrival time when including a ~140 day hold-up in an eddy close to the 7th arc.
    All possible perhaps but not likely IMO.

  58. DennisW says:

    @TBill,

    Yes, I suppose the current activity is generating information regardless of the outcome. I have pretty much lost hope relative to traditional investigative methodologies producing anything useful.

  59. Victor Iannello says:

    @Ge Rijn: As you proceed north from 30S, the timing of the flaperon discovery requires a delay between arrival near the shore and the discovery. However, proceeding north also makes the timing of the discovery of Roy easier to explain. 26S latitude probably is a bit further north than where the BFO and drift data suggest, but the limit is not well-defined.

  60. DennisW says:

    @Ge Rijn

    Even Richard Godfrey’s model from ~30S could only fit the flaperon arrival time when including a ~140 day hold-up in an eddy close to the 7th arc.

    I tend to put the “late” arrival of debris in the questionable value category. There are many factors that could be related to that. It is simply not worth putting much weight on “late”.

    You might be interested in the article below in which Mr. Gibson says some things about your glide theory.

    http://www.news.com.au/travel/travel-updates/incidents/wreckhunter-david-blaine-gibson-says-mh370-crash-was-not-result-of-pilot-murdersuicide/news-story/1579b3d728b0454dabb1ced33e3681e0

  61. TBill says:

    @Ge Rijn
    You and I always shared the hypothesis that Broken Ridge end point might have been a strategy. Still a hypothesis, but willing to find it anywhere as soon as possible.

  62. Victor Iannello says:

    Recent AIS data indicates Seabed Constructor is at -36.2608, 91.9645, at 13.9 kn on a course of 46°T, and already about 9 NM into the interior of the first 5,000 sq km area. That seems consistent with prioritizing the search near CSIRO-1 at -35.6,92.8. HT @LabratSR for the AIS data.

  63. Nederland says:

    To me, the state of debris, as described by Blaine Gibson and elsewhere, strongly suggests uncontrolled spiral dive (11 or so pieces have pull rather than crush damage).

    If the BFOs are to be explained by pilot input and thus a linear vertical motion, it is unlikely flaps and other debris would have come off the aircraft as apparently they did before impact.

    This is, for example, explained in the Silk Air 185 accident report (Boeing 737). The investigators concluded that only parts of the empennage came off and this only as a result of supersonic speed caused by engine power. They otherwise ruled out flutter damage is possible in the vertical dive scenario. (p. 21)

    https://reports.aviation-safety.net/1997/19971219-0_B733_9V-TRF.pdf

  64. Ge Rijn says:

    @VictorI

    Yes, the limits are not well defined and the Roy-piece is better explained by latitudes north of ~35S till even 26S. But this constraints also count for the flaperon IMO. They have to be in balance.
    It’s about arrival times and then those two pieces constrain a certain time frame and latitudes by their fresh barnacles when they were first found.
    From 26S the flaperon would have arrived a year too soon at Reunion if not caught up in an eddy for a year. Which seems a bit unlikely to me.

  65. TimR says:

    @TBill “The Xmas island scenario needs a little more evidence to sway the search, even though TimR has long said there may have been strong rumors of a clandestine plan like that. The “plan” could still be true but the flight did not seem to follow the plan very well, for unexplained reasons.”

    There is a possibility MH370, under pilot control, ditched some 100nm off the South coast of Java near the 7th arc.
    This was alluded to within days of MH370 going missing by someone in the know.
    Flying at a steady speed between waypoints the BTO’s fit.
    The BFO’s broadly fit.
    The drift studies do not disprove the possibility of a ditching in this area, in fact independent studies point to a high probability this was a likely area.
    There were a number of people who knew sufficient detail of what was intended to happen and that the plane could have ended up in this area.
    It is disappointing that not one of them has had the compassion to come forward in some way after almost 4 years.
    This area needs to be searched.

  66. Richard Godfrey says:

    @Ge Rijn

    There you go again, misquoting me again …

    When will you ever stop?

  67. Nederland says:

    @TimR

    What information do you have on this?

    There was the one Geomar study, not taking into account stoke drift, which was later revised, but it did not take into accont any debris other than the flaperon.

  68. Victor Iannello says:

    @TimR: It is disappointing that not one of them has had the compassion to come forward in some way after almost 4 years.

    Perhaps the reason nobody has come forward is that the scenario did not occur.

    If somebody would come forward with evidence of a plan to hijack and negotiate, that would blow this case wide open. Without that, there’s not much to consider.

  69. DennisW says:

    @Richard,

    I feel your pain… but I am now distracted.

    Amazon just delivered my new book. It is written for physicists, so it should be a quick read. The two previous books I got were oriented toward mathematicians, and I struggled with them (I am old and slow now). I’ll be gone for a couple of weeks.

    https://photos.app.goo.gl/xKoLwSQqM8cUf0pB2

  70. M Pat says:

    Hartelijk bedankt for the very kind words Ge Rijn!

    You do me too much credit on the ‘north of 36’ story however. I think both Brock and yourself were making that observation around that time if I remember correctly. I remain a bit agnostic on the significance of the apparently null result on Ozzie shores. It seems to me to be one of those things that over time becomes accepted wisdom but needs a bit more work to be given true weight.

    A shame for example that there isn’t an easily measurable statistic of the form “average duration of exposure of each square metre of coastline to appropriately knowledgeable human vision per day”. I’ve never been to Madagascar, but I have spent quite a bit of time on the coast of Tanzania and Zanzibar. In the third world the shoreline is a major resource. Pretty much all of it. The footfall is incredible. Flotsam will be picked over, even if its significance is not understood. Conversely I have little difficulty in imagining a few pieces of composite wedged between boulders at various locations along the empty stretches of the western or southern Ozzie coast. And we know from the drifter work that the likelihood of beaching there seems much lower in the first place.

  71. Michael John says:

    Just to clarify…. Are the drift timings of the debris based on when the objects were found or when they actually washed ashore? I’m a little confused because I don’t see how anyone could know when the latter could have been.

  72. Brian Anderson says:

    @Jean-Luc

    It seems to me that it is unnecessary to complicate the argument over the time [and point] of closest approach to the satellite by introducing the concept of spheres evolving in time, and 4D and speed. All that is necessary is to determine the time of the minimum LOS distance to the satellite. Yes, the BTO points lay on the surface of a sphere, but altitude is already included in the calculation of the LOS distance.

    In this argument the aircraft altitude becomes virtually insignificant, because we are talking about distances of around 36,700 km.

    I noticed that Jean-Luc has estimated BTOs that differ from the original Inmarsat BTOs by between 40 and 50 u secs around this time. Those differences alone contribute to differences in the LOS of more than 6km, or more than 20,000ft in altitude.

    So I tried to be consistent. Use the Inmarsat BTO numbers [recognizing that there are some known error bounds there too], and estimate by whatever means the time where the minimum occurs. In my view the minimum occurs after 19:41 and not before.

    It is then a matter of reasonably simple geometry to determine the straight-line distance between 19:41 and 20:41, and hence determine the average ground speed of the aircraft between these two points.

    Of course, I recognize that the two points at 19:41 and 20:41 plus the point of the minimum LOS distance will not necessarily sit exactly on a straight line, and hence it could be argued that I use quasi-planar flight. However, I do not assume constant altitude, nor constant speed. Both altitude and actual speed are irrelevant.

    I’ll stick to my hypothesis that the minimum LOS distance occurred at about 19:52, and hence the average ground speed between 19:41 and 20:41 was about 485 knots.

  73. DennisW says:

    @Brian

    Yes, I have looked at your analytics and they hold together.

    What is difficult is reconciling your average speed of 485kts with the measured BFO at 19:41 (see Bayesian Methods figure 5.7). For a track of 180 degrees the BFO reported is closer to compatibility with a ground speed of 410kts. It takes a significant deviation from a 180 degree track to achieve compatibility with 485kts ground speed at that time. Not to say the speed or the track is static. Just and observation. FWIW, I get the same numbers as shown in Figure 5.7.

  74. Michael John says:

    The theory behind Mh370 reaching the SIO was based on a General assumption of Ghost flight, is that correct? If so then it is generally assumed that once fuel exhaustion is reached the aircraft would effectively plummet from the sky? Thus all the calculations about how the debris found to date would be based on these assumptions. So If the components show pull as opposed to crush damage then the damage must have been caused by the aircraft being pushed beyond forces it is esigned to withstand… Then the spiral theory makes sense.

    A ditching scenario wouldn’t fit the scenario so is unlikely to be looked at in great detail. From my understanding aircraft ditchings are not an exact science & how the aircraft reacts on ditching is dependant on many varying factors. I am trying to research how an aircraft ditching would generally affect certain components. In particular all those that have been found to date. The majority of pieces found to date (aside from the interior cabin piece) have belonged to the nachelle or wings. I would like the experts here to look at previous aircraft ditchings & compare those results to the debris found from Mh370.

    I found a research paper & in it is this:

    “The nacelle would be torn away after the airplane pitched down to a low attitude angle. The slotted flap strength is usually designed such that failure occurs almost as soon as the flaps strike the water surface and do not cause any nose-down pitching moment”

    Which basically to me explains the flap being “torn off” & the “Roy” piece. Whether Mh370 suffered any crush damage is dependant on the angle the aircraft hit the water surely? I am not an expert so it would be good to learn from the people that are.

  75. sk999 says:

    Michael John asks, “The theory behind Mh370 reaching the SIO was based on a General assumption of Ghost flight, is that correct?”

    No. 100% wrong. Originally Inmarsat and the JIT only had the BTO data. Remember the “North and South Corridors”? The plane could have gone either to Kazakhstan or the SIO, and the data couldn’t distingush which. However, it was thought unlikely that MH370 could have evaded all the air defense radars in India and surrounding countries, so the SIO was deemed more likely on those grounds. Once the BFO data were understood, the SIO conjecture was confirmed. At no point was “ghost flight” invoked.

    Whether that impacts the rest of your post I don’t know.

  76. airlandseaman says:

    MJ: “The theory behind Mh370 reaching the SIO was based on a General assumption of Ghost flight, is that correct?”

    NO. 100% wrong. You should know that by now. If your whole theory is based on that, toss it all in the circular file. BTW…is has been emphasized a 1000 times here. Are you intentionally trying to distract everyone down another of your rabbit holes?

  77. Don Thompson says:

    @MJ meet @Ge Rijn, @Ge Rijn meet @MJ

    Other, quiet, corners of the internet are available.

  78. sk999 says:

    Michael John, you ask: “… then it is generally assumed that once fuel exhaustion is reached the aircraft would effectively plummet from the sky?”

    Independent of your previous assumptions and statements, CNN was able to determine conclusively the following:

    https://pbs.twimg.com/media/BkGANyyIYAAXK4B.jpg

    “Boeing 777 will struggle to maintain altitude once the fuel tanks are empty”. Mind you – it was a developing story.

  79. Michael John says:

    Come on Gents. Get off your pompous high horses. We may not hold the qualifications you do But I would argue that doesn’t make us any less of a man. We all know that it is highly likely that Mh370 is where you think it is. What I’m doing is talking hypothetically about subject none of us could know about. That is how Mh370 came to it’s final demise. Personally I can’t see why the “Poor man” or at least that is what I detect the intonation is can not share ideas in the mystery. The Ghost Flight was a popular theory. Now on the grounds of logic are you now suggesting Mh370 could have ended under pilot control? I wouldn’t think that likely. Most humans have a self preservation trigger & those that are hell bent on murder-suicide tend to do things sooner than later. So as we have decided that Mh370 must be on the 7th ARC in the location that is currently being searched. No issues with that from me. But from a logical point of view Ghost flight (at least at some stage must have played a part) surely. Thus my argument. You don’t like the ditch idea primarily IMO because it doesn’t fit with the pattern. So even if the debris evidence did support the argument you would dismiss it without a 2nd thought. We aren’t lawyers or barristers in a court room. So what is this pompodom all about? The difference between us gents is that I am able to keep my ego in check.

  80. Don Thompson says:

    @MJ

    I’m averse to terms such as ‘ghost flight’ and ‘mysteries’: populist junk. Stock-in-trade elsewhere. Not my thing. The objective has been to determine where the aircraft is. The specifics of who & the why come later when the recorders are located. My position is that who & why are not significant concerns at this time.

    Try to work with tangibles, it’s less risk to one’s ego.

  81. airlandseaman says:

    MJ: I have no problem with ill informed or even stupid questions. Happy to help people understand as those questions come up the first time. But you insist on revisiting the same questions previously answered many times, just to create a stir. That is annoying.

  82. Brian Anderson says:

    @MJ,

    I think the reaction to your posts is because all your comments have been debated at length, ad infinitum, over the last 3 years. It is tiresome going back over all the old ground. Almost everyone here has moved on and we are waiting with much anticipation to see what develops with the OI search.

    It isn’t about egos. The serious contributors here have all learned to park their egos. The bristling is because it is frustrating and tiresome having to continually educate newcomers, especially those who have not kept up with the debate over the last [nearly] four years.

  83. Brian Anderson says:

    @Dennis,

    Compare the calculated speed at 19:41 from other models:

    Yap, 475 or 481 knots
    Richard, 490 knots

    I readily accept a deal of tolerance on my number because of limitations of the very simple methodology, say -20 + 15 knots, but I think 410 is out of the ballpark.

  84. DennisW says:

    @Brian

    ~410kts is what you get using the 19:41 BFO and 180 track. Someone else here can check me on that, but I have been through it a few times.

    @MJ

    Don’t get wadded up. We all experience the same thing from time to time, and usually for a good reason. I am not going to debate telecom with ALSM or how to fly a 777 with Andrew. Make an effort to understand the “boundary conditions” on your posts.

  85. sk999 says:

    Michael John. you complain, “Come on Gents. Get off your pompous high horses. We may not hold the qualifications you do But I would argue that doesn’t make us any less of a man.”

    Sorry, but that is a typical response from someone who couldn’t be bothered to conduct basic research on the internet.

    A Google search for the terms “mh370 corridor kazakhstan southern indian ocean” instantly turns up online articles about the early days of the search for MH370, with headlines such as “MH370: Focus on southern Indian Ocean intensifies”. No mention of a ghost flight.

    The qualifications needed to perform the above search are minimal. However, laziness is not one of them.

    Disclaimer – I am not a member of the IG, nor do I speak for it.

  86. lkr says:

    MJ insists that he is “not a Troll”. Also claims that he’ll defer to the quants here. Elsewhere he declares that he “does not believe MH 370 is in the SIO”.

    Troll or not is a matter of definition, but responses to the babble above are not helpful — he will blink and later come back with the same whatabouts.

    There was a conversation going on — let’s get back to it.

  87. TBill says:

    @Brian @DennisW
    The thing I noticed (but did not try to reconcile) is for the case when the satellite close approach is 19:41, the speed calc is 400. Then if it is just a little past 19:41 at 19:51 the speed calc goes to 485ish. I just assume the severity of the curvature of the Arcs makes that a sensitive calc.

  88. Victor Iannello says:

    @Michael John: I have often advised people here to not use the terms “ghost flight” and “death dive”. They are inexact phrases that reporters have used to sensationalize stories. It is more exact to talk about periods of time with automated flight and no pilot input, for instance. Or after fuel exhaustion, an uncontrolled steep descent, or a nose-down pilot input, or a controlled glide.

  89. Victor Iannello says:

    @TBill: What is relevant in using the BTO to qualify a path is whether or not the calculated BTO values match the measured values within a limit dictated by what we believe the uncertainty of the measurement to be. There was a time when we didn’t know what the uncertainty was. That’s not the case now.

  90. DennisW says:

    @ikr

    I am reminded of my medical training. The instructor would ask “what were you thinking when you did such and such a thing?” (trick question)

    My response “blah, blah, blah, blah…”

    Instructor – “We are not training you to think. We are training you to follow the f’ing script. If you have to think, you will most likely f’ something up”.

    How can it get more humbling (but correct) than that? Likewise in many disciplines. A script (prepared by people way smarter than you) is a very good thing. You can be relaxed and execute competently, and not waste time thinking about what you should or should not do.

  91. Mick Gilbert says:

    @Brian Anderson

    Brian, did you ever create a graphic or kmz file for the arc of closest approach? I’ve been looking back at some of your earlier work (eg Deducing the mid-flight speed …, etc) but haven’t been able to find anything.

  92. DennisW says:

    @ikr

    more

    training, training, training,… is where it is at. There is simply no substitute. I am a believer.

    Having said that, I think the lack of CE for advanced degree people is a mistake.

  93. DennisW says:

    @sk999

    Disclaimer – I am not a member of the IG, nor do I speak for it.

    Well, that is an interesting disclaimer. The IG has done a number of very good things, but their terminal predictions have been embarrassing to science in the whole. Funny shit, actually.

    But hey, I can live with it despite my colleagues disparaging remarks. I am not a member either.

  94. Brock McEwen says:

    @Victor: the probabilities above – in whose crowdsourcing I invite you to participate – will in fact scientifically construct some of the evidence (or counter-evidence – it depends on the responses) you challenge me to present. They are only ridiculous questions if the answers are all “very high probability”. Is that your position? Or is your position that it doesn’t matter how improbable the above concatenation of events is – it must have happened, because the ISAT data’s reliability is so guaranteed that scientific scrutiny is absurd? If so, then please rest assured: I will not take any of your a priori probability estimates – no matter how low – as evidence that you yourself question the ISAT data. I will simply present the crowd sourced probabilities – aggregate them into an overall a priori probability (something in which I am trained, and offer freely and happily for the good of our collective search for truth), and let readers draw their own conclusions.

    If you wish not to participate in this crowdsourcing exercise, that is of course your right. All you have to do is say: “I refuse to participate in your effort to assess these probabilities”, and I will happily record that response in my report.

    I will also eagerly accept probability estimates for each of the above from non-IG members – I should have been more inclusive in the original request. I was just focusing on the IG members because, if memory serves (and it frequently doesn’t!), Victor wrote when he set up this blog that its purpose was for members of the general public to ask the IG questions.

    (This is not an academic exercise. If deep sea wreckage is ever found – as predicted by authentic and faked ISAT data alike – we’ll need to move quickly to the more important question of assessing the likely authenticity of any such find. It is much easier to fake seabed debris than ISAT data. And those who’ve been arguing we should search for wreckage BEFORE assessing authenticity will no longer be able to stand in our way.)

  95. ventus45 says:

    We can only watch and wait at this point – and wish calm seas for OI.

  96. Brian Anderson says:

    @TBill,

    “I just assume the severity of the curvature of the Arcs makes that a sensitive calc.”

    No. The calculation depends only on the time of the closest approach relative to 19:41 and 20:41, and then it is a matter of geometry. The curvature of the arcs can be completely ignored. I agree that if the time of the closest approach was 19:41, then the speed works out at close to 400 knots. But the calculation is certainly quite sensitive to the time that is estimated.

    @Mick,

    The speed calculation is totally independent of possible paths, and hence there can be no specific graphic [relating to the arcs] that can be depicted.

    On the other-hand many [maybe even most . . ] of the possible tracks produced by Richard, Victor, Barry, Yap, and others, illustrate the minimum point inside the 2nd arc quite well, even although the latitude of the minimum may vary considerably.

  97. Richard Cole says:

    Based on an analysis of their release and mission start points, where the AUVs are located at 09GMT today, Tuesday. Probable rendezvous point with Maersk Mariner also shown.

    http://www.dropbox.com/s/xu5w6uivpk6bcip/23-1-18c.jpg?dl=0

  98. Victor Iannello says:

    @Richard Cole: Thank you for that figure.

    For anybody that is interested in learning more about how the AUVs are probably deployed for this mission, read the excellent work of Richard at:

    http://recole.org.uk

  99. Niels says:

    @DennisW, Brian Anderson
    A while ago I found the same discrepancy in 19:41 groundspeed as Dennis did. My tool gives 417 knots at 186 degrees track (based on D = df_up + df_comp = -38.5 Hz and a latitude of 0.8 degrees N). After re-reading Brian’s paper I concluded that a factor in the discrepancy could be the straight path assumption. By the BFO/BTO interpolation method I find indeed a slightly curved path between 19:41 and 20:41, see fig. 3 in the linked document:

    https://www.dropbox.com/s/p477ng11mrezhc7/Straight%20paths_V0_8.pdf?dl=0

    Another explanation could be that the 19:41 BFO value has an error of around 6 Hz.

  100. Richard says:

    @Michael John

    Please take your pompous attitude elsewhere.

    You and Ge Rijn are trying to ruin an otherwise excellent website by repeatedly repeating what you have already repeated.

    You are boring!

  101. Victor Iannello says:

    @Niels: If MH370 followed a great circle path ending near 30S latitude on the 7th arc, then the BFO errors have a mean of about 5.3 Hz, and the RMS error is 5.7 Hz. If the bias drifted up by 5.3 Hz, the remaining RMS error is only about 2.0 Hz. Did that occur? The DSTG believes it’s possible.

  102. Victor Iannello says:

    In a new article, Geoff Thomas makes the case that the path of MH370 indicates human input, and suggests that MH370 was a case of suicide/murder by the captain. He’s getting a lot of flak.

  103. Ge Rijn says:

    @Richard

    I just connected YOUR drift-model from ~30S with YOUR buoy that landed on Reunion just like YOUR flaperon in a similar time-frame from ~30S.
    Your buoy got stuck in an eddy for about ~140 days.
    This is all YOUR information, not mine.
    I only made an interpretation and did not (mis)quote you.

    And I’m not here to please you or anyone. I support the common goal; finding the plane.
    It seems to me you don’t take different interpretations on your drift-modeling very well the way you react to me.
    You just serve me off with degrading comments without reflecting on content. Did I touched a weak point in your drift-model??
    I think so. And as long as you cann’t confince me with arguments I’ll keep thinking this way.
    Your asking about banning me or silence me with other comments won’t help you change my views. Only arguments.

    Anyway, different point of views don’t matter that much anymore.
    The new search has started and OI has got his search-plan ready. Unless major discoveries take place during the search this plan won’t change much IMO.

    I opt for the area between ~32S/33S and 95/~97E. You put your cards on ~30S close to the 7th arc. VictorI goes as far as ~26S. Which all could be possible IMO. Others go even further north or south. Which is impossible IMO regarding all the known data.

    We’ll see.

  104. Richard Godfrey says:

    @Ge Rijn

    I do not allow you to interpret my work.

    Please go and interpret your own work.

    You are not set up as God to judge over everyone else!

  105. DennisW says:

    @Victor

    @Niels: If MH370 followed a great circle path ending near 30S latitude on the 7th arc, then the BFO errors have a mean of about 5.3 Hz, and the RMS error is 5.7 Hz. If the bias drifted up by 5.3 Hz, the remaining RMS error is only about 2.0 Hz. Did that occur? The DSTG believes it’s possible.

    I truly hate being a “bad guy”. I am not trying to look smart or impress anyone, honest. However, it is not proper analytics to compute and assign meaning to the “mean” and “variance” of a non-stationary and non-ergodic process. I had the same argument with the DSTG, and they acknowedged that (sort of). The reality is that if MH370 flew whatever path it flew to wherever it terminated a hundred times, the mean and variance of the BFO profile would not be consistent. Simply not useful metrics to describe that process.

  106. ArthurC says:

    Interesting thread, especially where it seems to be going…

    How does this help find the plane?

    I can see why some are very fond of their ideas, but the reality is that only experts should be taken into consideration.
    Yes, everyone’s entitled to an opinion, but if it’s not backed by facts…

    And yes, I have to say this, the comment above about crowdsourcing reminds me of a Simpsons episode where Homer clones himself repeatedly and they all end up falling off a cliff while following a giant doughnut.
    “It was a good crowd”

  107. Ge Rijn says:

    @Richard

    I expressed my respect to you many times. I tried to be reasonable (in my flawed way I know) but it has no use I see.
    Go search the definition of ‘Projection’. God does not like these imperfect humans to interpret his truth. I completely understand.
    Leave it here. I will.
    It has no use anyway anymore. The search will show who was right or wrong.
    To me it’s just the same as long as the plane gets found.

  108. Richard Godfrey says:

    @Ge Rijn

    Er, No! It is not the same for you, as long as the plane gets found.

    You cannot leave it here.

    You have a psychological problem, first you accuse Don of being God:

    @Don Thompson “It’s not my confirmation-bias but yours you are projecting on me IMO in your God-like know it all perceptions.”

    Now you accuse me of being God:

    @Richard “Go search the definition of ‘Projection’. God does not like these imperfect humans to interpret his truth. I completely understand.”

    I know you are a former Psychologist, but I think you desperately need to visit a member of your former profession.

    I can promise you, neither Don nor myself are God. You are suffering from delusions.

  109. Victor Iannello says:

    @Brock McEwen: I already responded to you. Without the ISAT data that you refuse to accept, there is no basis to conduct a search.

    Consider this. Since the probability that a commercial flight takes off and disappears without reaching its destination is near zero, perhaps we should be looking in Kuala Lumpur and Beijing for the plane.

  110. Rob says:

    @ArthurC

    Welcome to the discussion.

    You said “Yes, everyone’s entitled to an opinion, but if it’s not backed by facts…”

    It’s all very well to say that, but there is no independent arbiter of the “facts” here. It’s more a case of “yes we are all striving for the facts, but my facts are more factual than your facts”.

  111. Victor Iannello says:

    @Ge Rijn et al: This is not a place to discuss theology. We’ll know soon enough whether or not the ATSB area is correct, and soon after whether or not 32S-33S area is correct. It’s also ok not to get in the last word in an exchange.

  112. Ge Rijn says:

    @Richard

    I know neither Don or you, or I are God. I have the utmost respect for both of you for your efforts and contributions. I’ve been not the one starting to devalue contributors (like me) the way Don and you did.
    If you do this to me you get my response whether you like it or not.
    I’m probably too long seriously involved in this drama to take any ‘bullshit’ for granted from anyone regardless their expert-status.
    If you don’t like it then just skipp it. Or come with arguments.

    I’ve come a long way and thanks to all the information I myself studied, VictorI, Don, you and many others over the years provided, I’ve been able to make my own interpretations for what it’s worth. Nothing perhaps. But I’ve tried with the best intend.

    We’ll see how it turns out. That’s what only matters in the end.

  113. @Brian Anderson

    Brian, please could you take another look at the data below ? I put down precise data centered on the arcs as if they were of no dimension and at the corresponding altitude, to demonstrate my point.
    In your paper you assume that the plane made a Direct flight just after the turn at about 18h40 (as did assume Inmarsat) with no manoeuver. Thus in that path the evolution of the LoS is probably what you describe. Now if the a/c made a small detour via around MEMAK, this is a different story and you curve fitting does not apply anymore as the a/c was flying towards the satellite and thus making the shortest LOS happening earlier in time. Then flying from MEMAK in a straight line heading 174°, descending in steps from FL270 to FL140 (at Arc3) and flying at a constant KIAS=240kn (thus varying ground speed) leads to the data below (I used Yap’s model for calculating the BTOs after a new flight simulation passing exactly on the center of the arcs as said above).

    It shows that the minimum LoS occurs much earlier than the one when flying in a straight line from what you call the FMT.

    I would be grateful if you could cross check and let me know if your findings are the same?
    Thank You.

    KIAS=240kn UTC Distance Sat BTO Lat Long Altitude (100ft)
    MEMAK (25NM before) 19:09:01 36738 11442 6.6092 93.2667 270
    Top of Desce 19:17:00 36732 11405 5.2592 93.3406 270
    19:22:48 36733 11415 4.8852 93.3876 251
    Middle descent 19:29:48 36735 11428 4.1854 93.458 216
    Arc2-centered 19:41:03 36746 11502 2.8330 93.6767 150
    middle Arc2-arc3 20:09:35 0.54 93.91 150
    Arc3-centered 20:41:05 36786 11740 -1.8455 94.127 140

    In reality, one should not forget that each arc is actually a band of up to 16Nm wide with a radius (from the Foot of the satellite on Earth) decreasing as the altitude decreases. In short, it shifts West as its altitute decreases.

  114. @Brian
    Sorry for the table whicg did not go throught properly. If you read from right to left then the tabulation can be recovered :-).

  115. Victor Iannello says:

    @Jean-Luc Marchand: Yes, your BTO values are correct at 19:41 and 20:41 for the positions and altitudes you reference.

    The point (and time) of nearest approach provides insight, but it is not a rigorous test of an acceptable path, as it uses a curve fit of BTO measurements, and it assumes those measurements have no uncertainty, among other assumptions, such as constant groundspeed. As long as the calculated BTO values match the measured BTO values within the uncertainty band, there is really no basis to use the BTO curve fit to reject the path.

  116. Brian Anderson says:

    @Jean-Luc,

    I understand how you arrived at an earlier time for the closest approach to the satellite, and I am not disagreeing with your approach to solving the issue.

    Please understand that I first proposed my hypothesis in May 2014, and at that time only very limited and sketchy information was available. In fact I began thinking about it before the Inmarsat data became available, before the correct interpretation of the BTO was determined and before any analysis of the BFO information was carried out, by anyone. I started when the so-called “fuzzy chart of elevation angles” became available. One might interpret that chart as showing that the maximum elevation occurred shortly after 19:41.

    Yes, my calculation is based on a [more-or-less] straight flight track, although a slightly curving track would not change the result significantly. However, if there is a significant discontinuity in the track [an intentional deviation . . as in the flight track you propose] and no corresponding BTO numbers, then the methodology may fail.

    My approach is very simple, and within the limitations of the assumptions I make, it is a useful method to test or corroborate paths generated using the full BTO and BFO data in sophisticated models. It has stood the test of time over the past 3 years.

    If we knew the track that the aircraft took, then finding it would be a lot easier. Your approach, developing a track using a logical argument, and adjusting the logic to ensure that it still fits the BTO and BFO numbers is unique. I applaud you for the effort. I’m not convinced though that this is the right answer. Hopefully we will know soon.

    Cordialement.

  117. TBill says:

    @Victor
    “In a new article, Geoff Thomas makes the case that the path of MH370 indicates human input, and suggests that MH370 was a case of suicide/murder by the captain. He’s getting a lot of flak.”

    Is the flak available to read, or is it behind the scenes? I now realize I’ve been confusing Geoff Thomas and Geoff Taylor (author of MH370 book). Geoff Thomas I know from MH370 TV coverage.

  118. AB says:

    With regards to the comment about auto pilot settings – the ATSB first principles report stated the 5 possible headings, and on p15 here says that Malaysian crew say they often use the CMH heading

    https://www.atsb.gov.au/media/5772107/ae2014054_final-first-principles-report.pdf

  119. David says:

    @Victor. Below (3 pages), if a relight did not bank the aircraft steeply, to account for the final log-on, I think it more likely that there was a pilot. A simulation would help with this.

    However even if that proved so, to me the intent indicated by the 7th arc log-on BFOs would suggest the maximum distance he was likely to glide would be 13 miles or less, tending to reduce the risk of a glide beyond the currently intended 25 Nm search width.

    https://www.dropbox.com/s/69717uzs7dfihum/Piloted%20final%20descent.docx?dl=0

  120. Mick Gilbert says:

    @Victor
    @TBill

    Re: “In a new article, Geoff Thomas makes the case that the path of MH370 indicates human input, and suggests that MH370 was a case of suicide/murder by the captain. He’s getting a lot of flak.

    Despite some stiff competition from Mr Ghost Flight-Death Dive at The Oz, Geoff Thomas is probably the worst aviation reporter in Australia. But a broken clock is right twice a day, so, as the Zen Master said, “We’ll see.”

    Victor, you’d hardly call that loose collection of ill-informed and misleading thought bubbles, “making a case”, would you?

  121. Barry Carlson says:

    All,

    At 24-0000UTC / 24-0600LMT the Seabed Constructor and Maersk Mariner were holding DP positions about 120 meters apart centered on -34.8481, 093.6424 and both heading +/- 230T into a light breeze from that direction.

    I expect the inter-ship transfer(s) will be taking place now.

  122. Victor Iannello says:

    @Mick Gilbert: I think there is consensus here that there was human input after 17:21, which I think was Geoff’s main point. Other statements about the guilt of the captain were proposed as possibilities. He also included the possibility of a stow-away in the MEC.

    Those that believe the captain is innocent are complaining that this article is unfair to the captain. Others likely see it as old news. Because there are no new facts, I don’t see a reason to debate the article here. It’s just a distraction, and doesn’t help OI find the plane.

  123. Victor Iannello says:

    @David said: if a relight did not bank the aircraft steeply, to account for the final log-on, I think it more likely that there was a pilot. A simulation would help with this.

    That simulation exists. It’s the case of double engine flameout witnessed by Mike Exner in a Level D simulator. About 1m20s after flameout, the re-started left engine has spooled up fast enough to roll the plane to a bank of 50. I posted the time sequence previously.

    In any event, I agree with your conclusion that if the BFOs are correct, the wreckage should be found well within 25 NM from the 7th arc.

    Thank you for your persistence to nail this down.

  124. Victor Iannello says:

    @AB: Yes, thank you for that. The two most common A/P modes would be LNAV, which is used to follow the waypoints of a route, and constant magnetic heading (CMH).

  125. Mick Gilbert says:

    @Victor

    Re: GT’s article – “It’s just a distraction

    It most assuredly is just that.

  126. sk999 says:

    The following is the results of an analysis I had been intending to do for several months. Please feel free to ignore – it will have no impact on OI’s current search strategy.

    As many have noted previously, the BTO values after the 18:25:27 logon varied smoothly with time. This gives rise to the following questions:

    1. How smooth is smooth?
    2. The first BTOs came before the FMT (assuming that the 18:40 phone call means that the plane turned South.) Wouldn’t a major turn mess up the smooth fit?
    3. What are the implications, if any, regarding possible flight paths?

    To address these questions, I set up two calculations. The first simply makes a polynomial fit to the actual data. The second starts with a simulation, samples the predicted BTO at the same UT times as the actual BTOs, and make a polynomial fit to those values. The polynomial by itself is not intended to represent any actual flight path – instead, it is used simply as a way to measure smoothness.

    First, to quantify smoothness, I fit a polynomial to the reliabe BTO values after 18:25. I use 12 values, separated into 6 distinct UT intervals. (The 24:11 and 24:19 values count as one interval). I find that a cubic fit (4 free parameters) fit the BTO values well, with an rms of 17 microseconds. That is already less than the known scatter intrinsic to the BTOs, so I am starting to fit the noise.

    Second, to assess the impact of the first BTO value being before the FMT, I made use of the calculation using simulations. The simulation models start with a track WNW at the 18:01:49 DSTG initial point, adjust the speed and/or heading to cross the 1st arc at the appropriate time, and then make an FMT at IGOGU. The exact details and autopilot mode don’t really matter. In all cases a cubic fit matches the simulated BTOs about the same as the actual data. It turns out that the fact that the 1st point is before the FMT DOES impact the smoothness of the fit. If I vary the timing of the 1st logon, it causes the rms of the fit to vary in the expected way – earlier times cause the rms to increase, later times cause it to decrease. If the 1st logon had been at the time of the FMT (around 18:39), a good fit could be had using just a quadratic (3 parameters.)

    Third, what are the implications for possible flight paths? As an initial matter, I ran this calculation using the actual flight path that MH371 used while traveling down from Beijing. That flight path went via multiple waypoints. Using an offset of 16.5 hours to align the two flights in time, I found that a cubic fit to the predicted BTOs has an rms resdiual of 269 microseconds – a terrible fit. I think we can safely say that MH370 did not fly via waypoints in a manner similar to that of MH371.

    What about a pilot actively maneuvering the aircraft? At this point one is fighting an uphill battle. Major maneuvers cause the BTO to bounce around – thus, one would have to structure them that, by chance, the timing of the pings was sufficiently fortuitous that the simple dependence of the BTO on time was preserved.

    Experience suggests that complex scenarios that magically align with simple patterns in the data are just wrong.

  127. Brian Anderson says:

    @SK999,

    “How smooth is smooth”

    Some interesting observations. Ha! Maybe it is because your observations seem to align with my own views. When developing the speed hypothesis and I was testing the “smoothness” of the fit [I used a cubic polynomial too], it was pretty clear that the BTOs around 18:25 did not fit very well. It was for that reason that I reverse engineered a new BTO for 19:01 in order to get a series that arguably might fit on a reasonably straight path, as an aid to determining the minimum.

    The fact that it results in a “smooth” curve with very low “R”is, I think, telling us something.

  128. Andrew says:

    @David

    First, I agree with the conclusion that the aircraft should be found close to the 7th arc, assuming the final BFOs are correct.

    Regarding your paper, you wrote “I expect the APU would deliver hydraulic pressure early in its auto-start cycle…”

    I don’t believe that to be the case. Hydraulic pressure will not be available until the APU generator starts delivering AC power to the ACMPs. That will not occur until two seconds after the APU reaches 95 percent RPM, ie two seconds after APU RUNNING, as stated in the following extract from the training manual:
    APU Starting

    You also stated a relight “could last up to 34 secs”, based on a residual fuel quantity of 30 lb at the APU DC fuel pump output pressure of 3150 lb/hr. That calculation seems to assume the 30 lb of residual fuel is all in the tank. If so, how do you account for the ATSB’s statement: “…the difference in location between the left engine fuel inlet and the APU fuel inlet would result in approximately 30 lb of fuel being available to the APU after a left engine fuel exhaustion.”? Wouldn’t some of that 30 lb have to include fuel in the APU fuel line, which would not be available to the engine?

    Regarding your previous comments about the hydraulic EDP, the pump will continue to deliver fluid at low RPM, but it can only regulate the pressure at 3,000 psi above a certain threshold RPM. I don’t have the details for the B777, but consider what happens after an engine failure in an Airbus A330, with RR Trent engines and a similar type of hydraulic pump. On that aircraft, which does not have demand pumps to assist the EDPs, the windmilling engine causes intermittent hydraulic system low pressure cautions, which are generated when the system pressure drops to 1450 psi.

    In the B777 case, I still believe the hydraulic pressure generated by a windmilling engine would be insufficient to prevent the ACEs recognising a hydraulic failure, at which point the PCUs would revert to bypass mode.

  129. David says:

    @Victor. Yes I was conscious of that simulation but have supposed it not to be representative. I would have expected that had it been, Boeing would not have said they could not replicate fuel movement, for example.

    The APU start-up time at 47 secs maximum is short and the engine relight runs too long, particularly since the engine should have been supplied fuel simultaneously with the APU as the engine coasted down past idle. That would have lengthened engine fuel supply time more, ie before relight. If fuel supply to both was an appreciable interval after fuel exhaustion, since both APU and the left engine would not have been supplied fuel before that, the apparent APU start time would have been delayed further.

    I do not know what that simulation’s engine fuel source was but it should have been limited to 3020 lbs/hr, the APU DC pump maximum with the APU also drawing. There are other concerns too about fidelity in this scenario and indeed whether total fuel exhaustion simulation would have been specified as a requirement for the United simulator software, if it was.

    As you say there was a relight all the same and an accompanying bank, though what part RAT deployment played in that is unknown. Indeed since bank started with the relight there may have been none, though the ATSB has spoken of RAT deployment inducing left roll I believe. Again, if supplied in the simulation with more fuel than the APU DC pump could have, its apparent effect would have been exaggerated.

    In any event that relight does not realise the bank needed in my opinion for the log-on descent, despite it running for those 48 secs. If that is to be the guide then I conclude that a piloted end was likely, though I do not think it has that fidelity.

    I would have preferred a simulated 34 seconds relight, without A/P and a fuel supply of 3020 lbs/hr, the maximum, where the relight was the only variable.

    However all this is academic unless there remains a strong possibility that a piloted maximum glide would be greater than 13 NM or if search width were decreased beneath that.

  130. David says:

    Last line amendment please, quick. “….search half width……”

  131. Richard Godfrey says:

    CLARIFICATION

    For the record, my drift simulation software takes the average historic data from ALL the Global Drifter Program transoceanic buoys in the Southern Indian Ocean (over 1,000,000 km of track) and calculates the average drift speed and direction for each time and location.

    THE DRIFT SIMULATION SOFTWARE HAS NO CAPABILITY TO INTRODUCE EDDIES OR DELAYS.

    THE DRIFT SIMULATION SOFTWARE IS NOT BASED ON A SINGLE BUOY, BUT ALL TRANSOCEANIC GDP BUOYS OVER THE LAST 20 YEARS.

  132. David says:

    @Andrew.
    – “Hydraulic pressure will not be available until the APU generator starts delivering AC power to the ACMPs….”

    Yes, thank you Andrew. Brain fade.

    – ““…the difference in location between the left engine fuel inlet and the APU fuel inlet would result in approximately 30 lb of fuel being available to the APU after a left engine fuel exhaustion.”? Wouldn’t some of that 30 lb have to include fuel in the APU fuel line..?”

    I do not think that was their intended meaning since their wording implies to me it is all, “between the left engine fuel inlet and the APU fuel inlet…”

    – “the windmilling engine causes intermittent hydraulic system low pressure cautions, which are generated when the system pressure drops to 1450 psi.”

    Yes I understand. The question remains though as to what prompts reversion to by-pass etc in the PCUs. There was an earlier post to the effect that ACEs will place systems in a hydraulically failed state when their position feed-back gets out of kilter with the requirement. That sounds pretty right to me and I will track down that post. I would expect that to be at pretty low pressures when the aircraft is not manoeuvering or fast.

    In any case I see the low pressure cautions while windmilling, to be the result of demand, in other words, little demand (no pilot) and the pumps will deliver their normal pressure.

  133. David says:

    @Andrew.

    This was it.

    Pilatus says:
    January 14, 2018 at 11:45 pm
    @Andrew,

    RE: “The pumps would keep turning and deliver some fluid as the engines windmill, but at a much-reduced pressure. I doubt the pressure would be sufficient to prevent the ACEs eventually recognising a hydraulic failure, at which point they would de-energise the flaperon PCU bypass solenoid valves and the PCUs would revert to bypass mode.”

    At a much-reduced pressure; should be at a much reduced flow rate. Hydraulic demand (as in flow) at the point of fuel exhaustion would not be particularly high IMO.

    The flaperon PCU is in the bypass mode when there is an electrical or hydraulic failure with the PCU or its ACE.
    To do this:
    The ACE has a failure monitor to do a check of the correct operation of the PCU. The monitor uses the ACE command signal and the feedback signals from the EHSV LVDT and the actuator LVDT. When the failure monitor finds a fault with the PCU or the ACE, it de-energizes the bypass solenoid of the PCU.

    A hydraulic failure of the PCU is a discrepancy between command signal and either or both LVDT feedback signals; not a failure of the hydraulic system.

  134. @Brian Anderson
    Thanks, Brian. Your demonstration for the “direct” path was inspired.
    As you say let’s see the search …

    @Victor
    Please note that we don’t assume a constant ground speed. We followed the aircraft and we assume a constant Indicated Air Speed at 240kn as it is the most likely speed that the FMS would have selected since 18h40 because of the descent in steps. This means that the FMS adapts and constantly decreases the ground speed as the aircraft is descending when the people in command ordered to do so. That’s why a “computational” demonstration is much more complex than assuming a pseudo constant ground speed and that’s why we would welcome some help.
    My previous post was mainly an academic case for the discussion with Brian.

    In conclusion, a piloted trajectory cannot be ruled out 🙂

  135. Andrew says:

    @David

    RE: “There was an earlier post to the effect that ACEs will place systems in a hydraulically failed state when their position feed-back gets out of kilter with the requirement.”

    Thank you, I missed that earlier post from @Pilatus. Given that information I agree, the PCUs would most likely remain in normal mode, assuming electrical power remained available.

    @Pilatus

    Thank you for your post on January 14; it answers a question that was raised a while back. Unfortunately I did not read it earlier.

  136. Mick Gilbert says:

    @Don Thompson
    @David
    @Andrew

    Don, back on 5 June last year you posted the following:

    “The AMM (Ch 27, Flight Controls) states:

    “The flaperon PCU is in the bypass mode when there is an electrical or hydraulic failure with the PCU or its ACE the ACE then de-energizes the bypass solenoid. If both PCUs on a flaperon are in the bypass mode, the flaperon can move freely. In flight, the aerodynamic lift then causes the flaperon to move 10 degrees up from the faired position.

    Training Information Point

    When the RAT is extended and hydraulics off, the airplane rolls left. Two to three units of right control wheel rotation are necessary to hold the wings level.”

    David and Andrew, can we agree that that is an accurate quotation?

    If so, why would the Training Information Point refer to the specific condition of “the RAT is extended and hydraulics off”, a condition that should give rise to the lack of pressurisation of the left and right hydraulic systems, as being associated with the behaviour that “the airplane rolls left”, a behaviour that is associated with both PCUs on the left flaperon (but not the right) being in bypass mode thereby allowing the flaperon to move up out of the faired position.

    It would be beyond odd, would it not, for Boeing to think it necessary to call out a specific behaviour as being associated with a specific condition if that condition did not cause the behaviour?

  137. Rob says:

    @sk999

    Interesting and useful analysis on your part.

    In particular, when you compared MH370 with MH371: “That flight path went via multiple waypoints. Using an offset of 16.5 hours to align the two flights in time, I found that a cubic fit to the predicted BTOs has an rms resdiual of 269 microseconds – a terrible fit. I think we can safely say that MH370 did not fly via waypoints in a manner similar to that of MH371”

    I have come to the same conclusion as you, regarding waypoints. It took me a very long time, though. Mine was reached by trial and error, plus and a bit of help from the Captain’s simulator and DR Bobby’s BFO work.

  138. Andrew says:

    @Mick Gilbert

    Yes, that is an accurate quote. The two conclusions are not necessarily inconsistent; it depends on your definition of ‘hydraulics off’. If there is no pressure in the system, then I assume the ACE will de-energise the PCU bypass solenoid when the ACE failure monitor detects that the PCU does not respond to ACE commands. That’s different to the windmilling case, where there is some hydraulic pressure in the system and the PCU does respond, albeit at a slower rate. There may be a point before zero system pressure is reached, at which the failure monitor decides the PCU isn’t responding fast enough and de-energises the solenoid; I don’t know.

    In the case mentioned in the training point, the RAT will power the centre hydraulic system primary flight controls. Consequently the right flaperon outboard PCU will be powered. If the left and right hydraulic systems are unpowered (ie ‘off’), then both PCUs in the left flaperon will be in bypass mode and the flaperon will float up.

  139. Victor Iannello says:

    @sk999 said Experience suggests that complex scenarios that magically align with simple patterns in the data are just wrong.

    The DSTG Bayesian analysis accounts for this by making paths that require more manoeuvers less probable. The end result is a hot spot at around 38S latitude, which seems to be contraindicated by the timing and location of the debris, as well as the null result of the previous subsea search. In your opinion, how is this reconciled?

  140. TBill says:

    @Jean-Luc @Victor
    Thinking ahead, of course I am hoping OI finds the crash site soon.
    If not, and if the weather gets bad, and if Jean-Luc’s path starts to look like a option, would the weather be better up there to allow some searching? Lot’s of “ifs” but just throwing it out there for planning purposes.

  141. Don Thompson says:

    @Mick, @Andrew, others interested…

    Hydraulic power without Engine Driven Pumps: one must assume that the ‘DEMAND’ selectors on the P5 HYDRAULICS panel are set to AUTO. If the ‘DEMAND’ selectors were set to OFF then no hydraulic pressure would be restored to the L and R systems subsequent to fuel exhaustion of the respective engine. AUTO would be the normal selection, but I think it’s necessary to consider the OFF condition.

    The primary C system hydraulic supply is the ACMPs, these would be supplied with power during the APU run.

    The Hydaulic system status is sensed through the HYDIM cards. The HYDIM cards take engine status as an input – I’d expect the logic to use ‘engine above idle’ in an AND function to signal hydraulic system status to the ACE units.

  142. DennisW says:

    @Victor

    Bayes works great if you have reliable statistics relative to the variables used i.e. how many false positives occur for a breast cancer test. The successful use of Bayes for AF447 was luck.

    1> what is the probability the pingers did not work
    2> what is the probability of an underwater scan missing the debris
    3> what is the probability of the drift data accuracy

    Blah, blah, blah,… If anyone thinks assigning values to any of the above is more than a guess, they are out to lunch.

    The reality is:

    1> The BFO data tells us the plane went South.
    2> The BFO data tells us the plane came down rapidly at the end.
    3> The BTO data (7th arc) limits the terminal locations.
    4> Drift analytics indicate a range of latitudes on the 7th arc.

    Beyond the above we have nothing. The whole notion of flight path reconstruction has been a flawed concept from the get-go. If I was running the ATSB, the attention half life of an IG or DSTG person in my office would be measured in minutes.

  143. Victor Iannello says:

    @DennisW: If you were running the ATSB, you’d be fired before you had a chance to interact with any staff.

  144. DennisW says:

    @Victor,

    Yes, you are probably right about that. Maybe I would have enough time to point out (as you did) that the CSIRO areas are too far from the 7th arc to merit a priority status.

  145. Ge Rijn says:

    @DennisW

    Some thoughts on your comment:

    1> what is the probability the pingers did not work:

    If there has been a ditch-like impact G-forces could have been too low to activate the pingers.

    2> what is the probability of an underwater scan missing the debris:

    This probability has- and can be directly made by the results and expected performance of the scanners. About 99% probability?

    3> what is the probability of the drift data accuracy:

    The best drift-data come from Godfrey, Pattiarachi and Griffin (IMO).
    Those data limit the search area between ~30S and 36S.
    Probability? I would say at least 80%. But I agree this is a subjective estimate. North of ~30S till ~26S could also be possible if debris has been caught-up in eddies for a (very) long time (IMO).

    1> The BFO data tells us the plane went South:

    In my mind there is no doubt about that whatever the Inmarsat-data would have told us. The debris finds confirm the Inmarsat-data.

    2> The BFO data tells us the plane came down rapidly at the end:

    The final BFO’s only tell us the plane made a rapid descent during 8 seconds from an unknown altitude, although most probably from an altitude between ~30.000ft and ~20.000ft.
    IMO we cann’t be sure what happened after this 8 second snap-shot.
    The plane could have pulled-out and level-out within ~1500ft (the PIC of ChinaAir 006 did it with a 747 after all engines had flamed-out during a spiral high speed dive)).
    Somewhere after this snap-shot a glide to a destination with the intent to ditch the plane is still possible IMO if it was under pilot control.

    IMO there is no evidence yet this is impossible.
    On the contrary I think but I won’t bore you with repeating my further arguments again..

  146. DennisW says:

    @Ge Rijn

    I was referring to the black box pingers. Both failed.

    The controlled ditch is very unlikely.

  147. Nederland says:

    ChinaAir 006 had lost only one out of four engines.

  148. Mick Gilbert says:

    @Andrew
    @Don Thompson

    Thank you for those clarifications.

    A quick question on post-second engine fuel exhaustion hydraulic loads; when the second engine flames out engine thrust data is lost and the thrust asymmetry compensation will disengage. My understanding is that when the TAC disengages the rudder is returned to its “normal” (pre-asymmetric) flight position. What sort of load would that rudder movement place on the three hydraulic systems?

  149. David says:

    @Don. PCU general isolation on engine failure would need logic exceptions for RAT operation, APU supply (through ACMPs or air driven) and ground operation.

    I see no benefit in general PCU shut down when sufficient pressure remains available, eg up until RAT deployment (centre system) and even possibly APU start (all), even though normally I expect there would be pilot/autopilot demand consuming supply rapidly. Any control overpowering/reversal would be detected by the position monitoring system, leaving flutter prevention with some PCUs(not the flaperons) to block. I think that would be normal hydraulic lock when not in movement but maybe not, in which case a minimum pressure would be required.

    Can you say how hydraulics would come to be selected off?

  150. Paul Smithson says:

    @sk999. Thank you for your post on how “smooth” the BTO curve is. I share your view. As I have previously pointed out, you can even (with a third-order polynomial fit):-

    1) “predict” 1828 BTO within about 100 microseconds by backward propagation of the fit 1941-0011
    2) see virtually no change in quality of fit if 1828 points are included (Rsq value and RMS of residuals)
    3) predict a very similar BTO_min value and timing (aka closest point of approach) using points after 1941 only, or indeed points from 1828ff but omitting 1941

    All of which suggests to me that if the 1825-1828 cluster of BTOs don’t “belong” to the same path that produced the rest, then they must be very close indeed in space and time. In fact if we didn’t have the radar and the BFO to tell us otherwise, I think we’d expect 1825-1828 to be a part of a single smooth trajectory. These observations are only reinforced by the observation that 1840 BFO also appears to “belong” to the rest of the series and we can only argue otherwise by invoking a change in altitude happening to be underway during the phonecall.

    @Victor raised the question, if smooth trajectory goes so far back (ie FMT so early), then the plane ought to have been found in the first search area. How could we account for the null result. It is a good question and one that continues to vex me.

    If we accept last radar somewhere in the 1822 vicinity and BFO 1825-1828 as indicative of WNW trajectory, then it seems to me our only ripostes could be:

    – plane was there but missed
    – the plane was further away from 7th arc than we have assumed
    – a steadily-curving path was followed that is not CMT or CMH (both of which bend too little early on and too much at the end)

    I am beginning to wonder whether the latter could be possible. An unpiloted aircraft that happens to fly ever-so-slighly “left” from ~1830 through until final fuel exhaustion. Could such a scenario be possible – perhaps with autopilot disengaged? @Andrew, would you care to comment?

  151. David says:

    James Jeffrey, Strewth Editor in The Australian enjoying himself today:

    “Holding pattern

    In our latest instalment in Adventures in Bureaucracy, the renewed hunt for missing Malaysia Airlines flight MH370 has afforded our illustrious colleague Ean Higgins the opportunity to quiz Australian Transport Safety Bureau spokesman Paul Sadler. Asked if the ATSB still held to its “ghost flight” and “death dive” theory, Sadler said the ATSB did not answer questions about what it thought about MH370 any more. That, he said, was the job of the federal government’s interdepartmental Joint Agency Co-ordination Centre run by Department of Infrastructure and Regional Development deputy secretary Judith Zielke. Early last year, Zielke addressed the Institute of Public Administration Australia on how essential it had been for the JACC to be transparent with the MH370 families and the media. But Zielke declined to answer those questions Sadler had steered towards the JACC, and her spokesman told Higgins any questions “should be directed to Malaysian officials”. Meanwhile, Sadler said it would be “misleading” to identify him as not answering the questions about MH370 because, he said, “I have not provided you a response”. This constitutes a form of perfection. “

  152. Andrew says:

    @Mick Gilbert
    @Don Thompson

    The ACE failure monitor controls the bypass solenoid on the PCU. It monitors the position of the EHSV (electro-hydro servo valve) that controls the hydraulic power to the extend and retract sides of the PCU actuator. It also monitors the differential pressure across the actuator piston. As @Pilatus mentioned, if there is a discrepancy between either or both of those feedback signals and the ACE command (ie a ‘hydraulic failure on the PCU’), the failure monitor de-energises the bypass solenoid, which puts the PCU in bypass mode. The ACE failure monitor might also have an input from the HYDIM cards, as @Don commented, but there is no mention of that in the manuals.

    RE: ”What sort of load would that rudder movement place on the three hydraulic systems?”

    I’d say little or none. The TAC disengages when the PFCS switches to secondary mode. When that happens, the rudder input from the TAC is removed and the air load moves the rudder back to its previously trimmed position.

  153. David says:

    Bodies could be preserved.

    “http://www.abc.net.au/news/2018-01-25/what-air-france-447-could-tell-us-about-mh370/9356854”

  154. TBill says:

    @Paul Smithson said
    “I am beginning to wonder whether the latter could be possible. An unpiloted aircraft that happens to fly ever-so-slighly “left” from ~1830 through until final fuel exhaustion. Could such a scenario be possible…”

    I speculated the same approx idea further above, that the Arc5-Arc6 spacing being too close caused the Bayesian analysis to miss the 180 south path “hot spots”…this could be due to pilot intervention (my preference) but maybe some other type of flight path upset in a passive flight (high winds is one thing, let’s say the flaperon fell off earlier than we though, etc)

  155. Brock McEwen says:

    @Victor wrote: “Consider this. Since the probability that a commercial flight takes off and disappears without reaching its destination is near zero, perhaps we should be looking in Kuala Lumpur and Beijing for the plane.”

    With respect: we are now, my fine friend, on my turf.

    It is silly, as you imply, to marvel over the “original coincidence” of the plane not making it to its destination. Of the hundreds of thousands that did, headlines rivet our attention on the one or two that did not. Accordingly, it would be absurd to search where the “Bayesian” analysis puts the plane by going “a priori” all the way back to take-off.

    In stark contrast: GIVEN that the first improbable event (the disappearance) has occurred, we should be highly skeptical of FURTHER coincidences. Probabilistically speaking, we are correct to reset the “coincidence meter” back to zero after the first odd event, but run that meter for every odd event thereafter.

    I will use the same logic in my report when computing, for example, the probability that MH370 and MH17 are utterly unrelated. The a priori probability of two consecutive major aviation disasters both being MAS 777-200ERs by sheer fluke is something on the order of “only”

    P = (# of MAS Boeing 777-200ERs in the world) / (# of large commercial aircraft in the world)

    …not P^2. Squaring would fail to realize that the first event must be treated as a given in the analysis.

    The only thing sillier than using the extremely remote a priori probability of an aircraft disappearance to deem the disappearance itself as “unlikely to have occurred” would be using it to deem every single coincidence thereafter as “LIKELY to have occurred”. BOTH are patently absurd. So I am quite sure you are NOT making that argument. Which brings us back to the need to construct the above probabilities scientifically, rather than setting each to 0 or 1 on the basis of blind trust or mistrust. Something I will do whether IG members assist me or not. I continue to hope it is the former.

    But, in jest or not, you were indeed on the right track when you started considering outside the box explanations of MH370’s true fate (something highly embarrassing to one or more major world powers, for example, ticks a lot more boxes than do any of the various theories endlessly circulating here). This is definitely a useful exercise.

    Whether seabed wreckage is discovered soon, or not.

    * * *

    While on the topic of math:

    One thing that often bubbles up in this forum is the misconception that the ISAT data – by virtue of its PRECISION – should earn our trust. I could publish infinitely inane analysis whose inputs and outputs are carried to 34 decimal places. Would that attract many IG members to its hypothetical proposed search area?

    Of course not. Because precision has nothing whatsoever to do with reliability. A dataset’s reliability derives from our ability to

    1) replicate it by experiment (repeat the event, & reproduce the data),
    2) validate it by comparison to the predictions of other evidence,
    3) validate it by comparison to the predictions of common sense, and/or
    4) trust it by reference to reputation/track record of its source.

    The ISAT data, in my opinion (and not yours, as well-noted) currently stands at 0 for 4. But we don’t need to agree on the score to mutually agree that, when assessing how “good” its data is, its PRECISION is neither here nor there.

  156. Andrew says:

    @Paul Smithson

    RE: “I am beginning to wonder whether the latter could be possible. An unpiloted aircraft that happens to fly ever-so-slighly “left” from ~1830 through until final fuel exhaustion. Could such a scenario be possible – perhaps with autopilot disengaged?”

    I think it’s extremely unlikely the aircraft would continue flying for almost six hours with the autopilot disengaged and no manual input.

  157. sk999 says:

    Victor,

    My comment r.e. complex scenarios with magical outcomes was meant to be general, not limited to MH370. Jean-Luc’s Captio scenario is one such scenario. JW’s Northern route is a second. Locally where I live, the police and prosecutors invented a complex scenario to get someone convicted of murder; it took over 3 years to show that the scenario was BS and get the guy released (his alibi was just as he said all along), and the mess is still not cleaned up.

    Paul Smithson,

    Your thoughts r.e. smoothness and possible ways for an unpiloted aircraft to fly are similar to mine – nothing conclusive yet. A CMT route is the closest approximation to a data-driven route (ending at -33.6 to -34.0), but the fit to the BTOs is rather unsatisfactory (while the fit to the BFOs is fine). Has anyone flown a B-777 for 5 hrs in CMT mode and verified that it follows the route that it is supposed to? One can things of ways to perturb it.

  158. Rob says:

    @Paul Smithson

    Apologies, I appreciate your post wasn’t directed at me, but I will bet the reason turns out to be that they didn’t search far enough downrange of the DSTG hotspot.

    The ATSB used Boeing’s recommended procedure for ditching at sea, as a justification for excluding the piloted glide scenario. The recommended procedure s to ditch under engine power, with flaps down. When they found that the flaps were retracted at the time of impact, they used it to exclude the pilot in control scenario. But why pilot would want to
    soft ditch in preference to a hard impact that got it over quickly as possiblevand sent the plane to the bottom with the minimum of floating debris as a reminder. was something they conveniently overlooked. The debris actually suggests a hard, but non the less, controlled impact, imho.

    I had a short email correspondence with David Griffin, shortly after the ATSB/OI meeting in London. I told him that I had identified a flight path aligned on manual waypoint S41 E88, that had almost exactly an endurance/range match with Dr Bobby’s M0.82 FL400 fuel model result, that used the navigation method that the pilot had practiced on his simulator. I was surprised and concerned at his reaction. Instead of defending the drift study outcome, he just wanted me to know that he went to the ATSB presentation about their flap examination result, and he said the evidence the flap was retracted was very convincing.

    I have a feeling the current search will end exactly the same way as the first

  159. David says:

    @Victor. Further, on the Boeing vs United Airline’s simulators, FYI I commented to the ATSB in part on 28th March, 2017 as follows,”…No other level four simulations I have seen results of replicate that 10,800 fpm rate increase in 8 secs.”

    In response the ATSB responded, on 3rd April, 2017, “…..any discrepancies between the simulations you have seen and those which the MH370 Working Group used may be the result of access to different simulators. The Working Group’s simulations were performed in the Boeing Engineering Simulator which has a significant fidelity to the actual systems on the aircraft, and which was therefore considered more accurate than many other available simulators.”

    On another matter of interest still, they added, “The required increase in rate of descent (consistent with the BFO analysis) was recorded in several simulations with different configurations of the aircraft…”

    I notice (as have you) that the issue of whether these descents were coincident with the log-on, as distinct from the 8 secs acceleration, was and remains unaddressed. The unpiloted simulations illustrated in the Search and Debris Analysis of November 2017 suggests no coincidence, IMO and I think Gysbreght has long held this view.

    However this coincidence is crucial, or else explanation is needed as to how the conclusion implied can be sustained without it.

    Hammering on a locked door though I think. I hope it is not confirmation bias but IMO a piloted possibility is more consistent with the evidence as released/described than the piloted.

    It remains faith/confidence in the SSWG that the evidence is there; it is just that it has not been disclosed.

    I promise to beat this drum no longer.

  160. David says:

    One last beat. Third last para, “…..than the unpiloted.”

  161. DrB says:

    @Dennis,

    You said: “I was referring to the black box pingers. Both failed.”

    I assume you mean the acoustic locators on the CVR and FDR (not the Emergency Locator Transmitters, which certainly did fail). Why do you think they failed? I don’t think there is reliable information on whether they operated correctly or not, based on the locations of the limited areas where acoustic searches were conducted.

  162. Brock McEwen says:

    @David: thanks for sharing. I suggest you follow up with the ATSB, and ask whether they have any hard data demonstrating a bias toward rapid acceleration sufficient to explain the gap you’ve observed. As a hardened veteran of the war to extract anything of value from the ATSB, I can see immediately from their response that

    1) they want you to BELIEVE their extra model fidelity explains the acceleration gap you highlight
    2) they did not STATE any such thing. At all.

  163. TBill says:

    Some FS9 calcs tonite:

    Assuming 180 south from ISBIX, comparing CTH vs. CTT
    with the opposing winds, both CTH and CTT end up at same spot (93.67E) by 27.5S and at the same approx. time and fuel consumed at constant Mach.

    So I see almost no difference (CTT vs. CTH) to 27S. The CTH will then start bending to the East and will appear to hit Arc6 a little too soon unless slow down is imposed. But CTH probably matches rings better for Arcs 3,4,5.

    Close approach to satellite appears to happen before 19:41 in this scenario.

  164. lkr says:

    @DrB

    Pardon me for jumping in, but @Dennis’s original context was to compare the use of Bayesian prob for AF447 [where pingers did in fact fail] with what we know in the present case [which he consistently points out is inadequate for a search if he were in charge].

    EG: [DW, original quote]
    “The successful use of Bayes for AF447 was luck.

    1> what is the probability the pingers did not work
    2> what is the probability of an underwater scan missing the debris
    3> what is the probability of the drift data accuracy”

    This dragged in GR previously and Dennis’ clarification didn’t help. But [I could be wrong] I don’t see a rabbit to chase here..

  165. HB says:

    @Rob, RE: “The debris actually suggests a hard, but non the less, controlled impact, imho”

    This may be true from the point of view that very few debris were discovered to date and very few pieces were observed to fail due to compression damage. However, from a structural failure analysis point of view, not all debris suggest that.

    Take the reunion flaperon for instance, it separated following hinge failure (apparently from the photos – torsional failure). This is possible for a ditch case as experienced by Asiana 214 however the failure pattern on Mh370 are very dissimilar and the MH part do not indicate, unexpectedly, any aluminium frame deformation. The lack of apparent structural deformation of the aluminium frame and the lack of composite shattering/ fastener failure suggests that mid air separation yielding a lower impact energy cannot be ruled out at this stage(though i disagree with flutter as a cause as the natural frequency of these components are relatively high regardless of the boundary conditions free or rigid considered). IE it is well possible that the current search (based on reverse drift analysis) may only find the place where the debris fell and not where the main body is.
    My understanding of the assumptions in the CISRO study is that all the debris fell at roughly the same location near the 7th arc. If this is the case, the distance apart could be very significant.
    Anyway, dice are rolled and we will know soon enough.

  166. Ge Rijn says:

    @Nederland

    On ChinaAir 006 I still had this ‘Mayday-video’ in mind.
    Which indeed is not accurate on the loss of all engines considering the NTSB conclusions. Starting at 16:10:

    http://www.dailymotion.com/video/x3evpw2

    Thanks for the correction.

  167. Ge Rijn says:

    @Andrew

    On your comment:

    “The TAC disengages when the PFCS switches to secondary mode. When that happens, the rudder input from the TAC is removed and the air load moves the rudder back to its previously trimmed position.”

    I have a question you maybe can answer.
    I assume that after first (right) engine flame-out the TAC will cause the rudder to deflect to the left to compensate for the trust and drag difference between left and right side of the plane.

    Now when second engine flames out and TAC disengages the rudder will still be in this left deflected attitude for some (short)time before finding it’s neutral position again. Initiating a left turn initially after second engine flame out I assume.

    Do you have a guess how long it would take the rudder to find it’s neutral position again when TAC disengages after second engine flame-out?
    Could this be of any effect on the attitude (turn) the plane took right after second engine flame-out and without pilot-inputs (ofcourse)?

  168. Andrew says:

    @Ge Rijn

    Re: ”Do you have a guess how long it would take the rudder to find it’s neutral position again when TAC disengages after second engine flame-out?
    Could this be of any effect on the attitude (turn) the plane took right after second engine flame-out and without pilot-inputs (ofcourse)?”

    It would only take a second or two. It is unlikely to have any effect on the aircraft’s flight path.

  169. Paul Smithson says:

    @Andrew. Thanks for your earlier response about flight without A/P engaged. Am I right in understanding, then, that you can’t see any mode of flight (other than an AP controlled magnetic heading/track, or by CTH with crosswind) where the path might curve gently and progressively to the left? Can you conceive of any atypical/impaired situation that could produce the same result?

    The reason I persist in asking is that I think there is a number of reasons to prefer a “curved” rather than straight path solution to the ISAT data.

  170. Rob says:

    @HB

    “Anyway, dice are rolled and we will know soon enough.”

    Yes, very true.

    But there is a lot at stake. A second search failure could effectively put an end to any further attempts. Would the ATSB and Malaysia be particularly concerned about that? I really think not.

    For the relatives It will be a further tragic turn of events, however. And OI will end up as the fall guys.

  171. Andrew says:

    @Ge Rijn

    Bear in mind that it takes about 5-6 seconds for the electrical power to fail after the second flame-out, because the generator continues to produce power until the engine RPM winds down. The TAC still operates during that time and senses the decreasing thrust differential between the engines. Hydraulic power is also available for a short time as the engine winds down, so the rudder will already be close to its trimmed position before the power fails and the TAC disengages.

    @Paul Smithson

    Re: ”Am I right in understanding, then, that you can’t see any mode of flight (other than an AP controlled magnetic heading/track, or by CTH with crosswind) where the path might curve gently and progressively to the left? Can you conceive of any atypical/impaired situation that could produce the same result?”

    No, I can’t see any other modes that would produce such a result, nor do I know of any ‘impaired’ modes that would do the same.

  172. Ge Rijn says:

    @Andrew

    Thanks. I understand now it’s a gradual, controlled proces for about 5/6 seconds in which the rudder turns back towards it’s trimmed position till the lack of electrical power shuts down the TAC.

    I assume hydraulic pressure is maintained fully during this time and only starts to fade when the electrical power supply has stopped after this 5/6 seconds. Then I assume it will take some extra time for the hydraulic pressure to drop to a certain level that will cause the left flaperon to go in bypass mode and then lift the left flaperon upwards.
    I wonder how much additional time this will take before the APU kicks in and restores hydraulic pressure completely (and the left flaperon neutral position).

    Just thinking out loud trying to understand better what possibly could have happened after second engine flame-out (without pilot inputs).

  173. @sk999
    Magic has nothing to do in our trajectory. It is all logical. The 24 events are logically deduced from the aircraft configuration and capacity, the airspace structure, the safety rules, the operational procedures, the fuel consumption, the timing…
    And it participates to the understanding: for example, the recent adoption by Bobby Ulrich of the Contingency Procedure after Ping1 comes from our deductions and from the B777 TRI/TRE pilot we have consulted.

    Now that you raise the question of magic, one return the question if the magic is not in assuming that after intensive human interactions in the cockpit from 18h22 till after 18h40 suddenly the aircraft is left to herself for the rest of the flight ?

    I leave to the statistical experts to determine what is the most probable between “after human activity there is still human activity” or “after there is no more human activity in the cockpit”…

    As we are all trying to help, let’s continue to progress in the “understanding” in all possible logical directions until the wreck is found.

  174. Paul Smithson says:

    @Andrew – thanks again.

  175. TBill says:

    @Jean-Luc
    …agreed. Another way to say it, looking at FR24, it would be an unusual flight path for a commercial jet to hang a sharp left turn at IGOGU and go down Indonesian flight space and radar reach. The only flights that seem to go down that way are from Thailand and China to Madagascar.

    On the other hand, it is fairly common to see commercial flight deviate from N571 similar to your flight path. Separately I shared with Victor an FR24 screen grab showing EK343 basically taking your path deviation to MEKAR over Great Nicobar…I don’t think they descended in that case, but other cases I have seen the descent to FL280-FL300 for flight path rules. I have suggested previous to your paper that the descent may have been related to flight path rules deviating from N571.

  176. TBill says:

    CORRECTION-
    I have not actually seen a descent, rather I have seen flights like MAS3/4 stay at FL300 in preparation from deviating off N571 to DOTEN. From that I infer that a rouge aircraft might make the descent if one were higher altitude.

  177. TBill says:

    CORRECTION #2: MEMAK not MEKAR (I keep doing that typo)

  178. @sk999
    pls read “aircraft configuration and capabilities”

    @TBill
    That’s interesting, thanks 🙂 (I guess you meant MEMAK)

  179. Ge Rijn says:

    @Jean Luc M.

    Rather than trying to pull focus on your scenario now from here, it seems to me more appropriate that you and your colleagues use your knowledge to help making the current search a succes or else argument/prove the assumptions and data it’s based on make no sence.

    Your scenario is only a distraction now IMO and I honestly wonder why it shows up at this particular time at the beginning of the new search.

    Your scenario maybe becomes more actual when this search has ended without finding the plane. I doubt it but who knows.
    Till that time I don’t consider it helpfull in finding the plane.
    All you suggest in piloted versus unpiloted has been discussed in lenght here and else before. Nothing new added by you.

    I choose to concentrate on the ongoing search and not on trying to ‘help’ you with your scenario at this time for it has no use now IMO.

  180. Victor Iannello says:

    David said: I hope it is not confirmation bias but IMO a piloted possibility is more consistent with the evidence as released/described than the piloted.

    Noted. That assumes the BFOs are correct, and we have no reason to believe they are not. If correct, the wreckage should be close to the 7th arc, no matter what the cause of the descent, unless there was an extraordinary and successful effort to recover and glide.

  181. sk999 says:

    All,

    Need a little help, as I am a bit lost. The location of the 7th arc depends on four numbers: the bias offset K, the BTO of the final logon, the UT of that logon, and the altitude of the aircraft. Given the same set of such numbers, the comparisons that I and others have done suggeest that all here get basically the same location.

    My own current values are K=495678 (1 off from the DSTG value), BTO=18390, UT=00:19:33 (both an average of the last 2 values), and geometric altitude of 6 km (approx 20,000 feet). At longitude 93 deg, I find a latitude of -34.855

    Where is the actual 7th arc being used by OI, presumably supplied by the ATSB? The only ATSB figure that seems to show it accurately is Figure 9 of the Dec 3, 2015 report, which was derived by the DSTG; for longitude 93, I estimate the latitude is -34.977. Richard Cole’s charts show a latitude of -34.935. The maps posted to the MH370 official website are not well calibrated but seem to indicate a value of -34.99. This range corresponds to a spread of about 5 nm in the cross-arc direction, which seems rather large. OI’s first round of scanning is all just outside of where I would draw the 25 nm limit but within the limit using the ATSB arc.

    Has anyone here made a comparable comparison?

  182. Paul Smithson says:

    @sk999. You might have added one seldom-mentioned criterion, the position of the satellite.

  183. Paul Smithson says:

    …but thanks for pointing out that discrepancy. I hope that others will be able to comment because if OI’s search position is as you say then I’d agree it’s a little odd.

  184. Victor Iannello says:

    @sk999: OI is aware of the discrepancy you noted. Some of us have advised they use BTO = 18,390 μs, A = 20,000 ft, K = 495679 μs, and have supplied them with coordinates based on those values. What they decide to actually use is their decision.

    Regarding their initial search area, I believe they need to include at least all the area included in the 25,000 sq km prioritized by the ATSB, independent of the relation to the centerline they will follow outside of that area.

  185. DrB says:

    @lkr,

    Thanks for the clarification. I thought DennisW was referring to the MH370 pingers.

  186. sk999 says:

    Victor,

    Great, thanks for the clarification.

  187. Ge Rijn says:

    @sk999

    I only hope OI does not restrict themselves to this +/-25Nm search width as it were the Bible after this 5000km2 search has ended (unsuccesfully, hope not..).

    The chance of an extraordinary and successful effort to recover and glide is still not proven to be impossible.
    This could have ended the plane up to ~45NM from the 7th arc after a steep descent IMO.
    IMO OI has to consider this possibility at some time if necessary.
    I hope they take the opportunity as they start searching north of 33S and take the width between 95E and ~97E from there till ~32S starting with the east part of this sector.

    It’s the most difficult part to search by its Broken Ridge underwater mountains and trenches but maybe better do it then than later if little time is left and the weather starts to deteriorate.

  188. sk999 says:

    Jean Luc,

    I am sure your route is wonderfully logical, but you ignored what I wrote. The measured BTOs after the 18:25 logon are well fit by a cubic polynomial. Your piloted route would not generically give that result. Take the extra BTOs that you provided at the 18:40 and 23:14 phone calls – they do not follow the cubic fit at all – errors are over 200 and 100 microsec respectively. Even if you eliminate the two nearest Inmarsat points and recalculate the cubic fit, it is terrible compared to the original fit (60 microsec rms error v. 17). That is why your route is “magical”.

  189. Andrew says:

    Correction:

    I previously said: The ACE failure monitor controls the bypass solenoid on the PCU. It monitors the position of the EHSV (electro-hydro servo valve) that controls the hydraulic power to the extend and retract sides of the PCU actuator. It also monitors the differential pressure across the actuator piston.

    The ACE uses the EHSV position and actuator position signals to control the bypass solenoid. The differential pressure signal is used in the servo loop that produces the command signals for the PCUs. It is also used by the PFCs.

  190. airlandseaman says:

    sk999: Until last month, ATSB had been using for the “reference arc” the 40,000 ft arc. Following a discussion with them about the implications of the BFO data, they agreed to switch to 20,000 ft for the reference arc. (It is obvious the plane was much lower than 40,000 feet at 00:19.) That causes a ~3nm shift toward the sub satellite point. In addition, we recently learned that the ATSB (Boeing) assumed a corrected 00:19:29 BTO value was 18440 usec. They did not attempt to calibrate and use the 00:19:37 value.

    As victor noted, we believe a (much) better estimate can be derived by combining the 00:19:29 and 00:19:37 observations, appropriately weighted according to the estimated noise on each observation, resulting in an estimate of 18390 usec. That also moves the ATSB reference arc in ~3nm, for a total change of 6nm from has been used in prior publications.

    As Victor also noted, what OI is actually using is not known, but they are fully briefed on all of the above.

  191. TBill says:

    @Ge Rijn
    Jean-Luc’s work is very interesting even if you do not accept the Xmas Island diversion. Their account of how the flight might have gone before 19:41 is a contribution of an alternative which opens some new ideas. Also the objective of the paper is to try to impact the search area, which even if not appearing relevant at the moment, helps to focus on what the future objectives should or could be.

    As Victor has said, nobody seems to be changing their personal theories 4-yrs out here, but I accept that my theory is just one of multiple possible solutions to the problem. Having one’s own path theory worked out mathematically helps one to understand the discussions here, even if you do not agree with a certain discussion.

  192. Ge Rijn says:

    @TBill

    I agree on that. It might be helpfull in this regard to again re-evaluate all that has been investigated and discussed before for some.
    To me it’s only a distraction at this point in time.
    At first glance I see too many big discrepancies in @JLM’s scenario to consider now while the search has started and more important things have to be considered IMO.
    Christmas Island is not on the list of OI.
    It’s just a non-issue now. Maybe later.

  193. Andrew says:

    @Ge Rijn

    Re: ”I assume hydraulic pressure is maintained fully during this time and only starts to fade when the electrical power supply has stopped after this 5/6 seconds. Then I assume it will take some extra time for the hydraulic pressure to drop to a certain level that will cause the left flaperon to go in bypass mode and then lift the left flaperon upwards.
    I wonder how much additional time this will take before the APU kicks in and restores hydraulic pressure completely (and the left flaperon neutral position).”

    The pressure should remain constant initially, but the flow rate will decrease as the engine winds down. The pressure will drop once the engine RPM drops below idle.

    There’s an ongoing debate about whether the flaperon PCU will go to bypass mode. According to the manuals, the ACE failure monitor compares the ACE command signal with feedback signals from the PCU servo valve and the actuator to determine if there is a failure in the PCU. In the absence of control inputs, the PCU might remain in normal mode.

  194. GlobusMax says:

    Does anyone know if any non-flaperon items considered “high windage” were found that were very likely or confirmed to be MH370? CSIRO places a bar of first detect on their drift model in Africa on the high windage charts, but I am not clear as to whether they are saying that high windage items were found there. None of the debris I have seen in reports seems to be classic “high windage.”

  195. Don Thompson says:

    Rob wrote “he went to the ATSB presentation about their flap examination result, and he said the evidence the flap was retracted was very convincing

    Yes, the evidence for the flaps remaining retracted is very convincing. It’s not just what ATSB posit in their debris analysis. You quote Dr Ulich’s fuel endurance analysis: when the fuel runs out, the flaps cannot be extended. Others have discussed that the average speed required to span the distance, 6th to 7th arc, must be kept high. You might try calculating how that could be realised with the extension of flaps that’s necessary for a significant downward deflection of the outer flap.

  196. Don Thompson says:

    Andrew wrotethere’s an ongoing debate about whether the flaperon PCU will go into bypass mode

    The PFCS design delivers hydraulic supply to the left flaperon’s PCUs from the L and R hydraulic systems. The hydraulic system HYDIM functions will sense system and pump pressures for each system (L, R and C) plus engine speed below idle. The HYDIM functions also control the ACMPs for the L & R systems, if the DEMAND selection is set to AUTO. The HYDIM statuses for each hydraulic system is distributed to the PFCS and ACE units. Accordingly, the ACE will set the appropriate PCU mode if hydraulic supply is lost. In the case that the associated PSA cannot supply power, the PCU’s solenoids will ‘relax’.

    PCUs for other control surfaces act differently, e.g. aileron PCUs lock to prevent upward deflection in slipstream.

  197. Don Thompson says:

    Ge Rijn, Ge Rijn, Ge Rijn, Ge Rijn

    (channelling Mrs Doyle)

    You are Alex Siew reincarnate: in every way, identical MO.

    No matter what the topic of discussion, the response is classic ‘siewage‘. His thing was lightning, secret backup batteries everywhere, and a rear upper fuselage crown floating on the Sth China Sea for 8 hours: utter nonsense. Yours is “ the chance of an extraordinary and successful effort to recover and glide is still not proven to be impossible“.

    I’ll suggest that a chance may be evaluated, but an outcome is required as proof. I will evaluate the chance for you: 0.0001%

    Advice unchanged, as before.

  198. Barry Carlson says:

    @ALSM,

    I note that Richard Cole has a 1900 UTC update to his site, and has mentioned that it looks like the SC is heading back to the original start point.

    Closer examination of his graphic has me to believe that the manner in which the AUV’s have now been deployed is different to that used initially. However, there is evidence that they have all been launched and are now heading SW “outside” the nominal search area.

  199. Victor Iannello says:

    @Barry Carlson: Based on this graphic from Kevin Rupp, it looks as though for the most recent set of launches, they gained operational efficiency in the timing and location of each launch. The spacing between the AUV trajectories looks a tad over 2 km.

  200. Barry Carlson says:

    @Victor Iannello: On a closer look at Richard Cole’s graphic, it now appears that there may have been a problem with AUV#8 on the northward trip, and that there are now only 7 AUV’s going south, i.e. the track originally taken by #8 is now being rescanned on this southward pass.

  201. DennisW says:

    @ikr@DrB

    “Thanks for the clarification. I thought DennisW was referring to the MH370 pingers.”

    Yes, thanks also, ikr. I did not realize there was some ambiguity.

  202. sk999 says:

    ALSM, you pointed out that “In addition, we recently learned that the ATSB (Boeing) assumed a corrected 00:19:29 BTO value was 18440 usec.”

    That is worth noting. Up to now we have used the Inmarsat-derived correction of -4600 usec for the R600 BTOs, but never verified it. Presumably Boeing was using -4560 instead. I went back to the MH371 data to ferret out the correction, and found a value of -4580, but with 20 usec rms error in that value (due to R600 BTOs being so noisy), so there is no good reason to select Boeing over Inmarsat; one just needs a larger dataset.

  203. Victor Iannello says:

    @sk999: I get an offset of 4578 μs with a standard deviation of the mean of 19 μs (N=25). That’s pretty close.

  204. Andrew says:

    @Don Thompson

    RE: ”The HYDIM statuses for each hydraulic system is distributed to the PFCS and ACE units. Accordingly, the ACE will set the appropriate PCU mode if hydraulic supply is lost.”

    I can see where the HYDIM cards interface with the PFCS via AIMS and the flight control buses, but I can’t find any references that describe how the HYDIM card data is used, apart from auto speedbrake control. The AMM does not mention hydraulic system status as an input to the ACE failure monitor that controls the PCU bypass solenoids. The only inputs seem to be the ACE command, EHSV position and actuator position. Do you have a reference?

  205. PaxLambda says:

    Hi,

    A document on MH370 official site (http://mh370.gov.my/index.php/en/) is a PDF where the 4th page show what could be the search area(s) if the three first (5000 + 10000 + 10000 km²) will come empty :
    http://mh370.gov.my/phocadownload/News/MH370%20Search%20Operation%20by%20OI%20Map.pdf
    The created date is 10th january, so perhaps it has been “published” there before: in this case, I am sorry and this post can be deleted.

    PL

  206. Rob says:

    @Don Thompson said to me iro the retracted flap subject “You quote Dr Ulich’s fuel endurance analysis: when the fuel runs out, the flaps cannot be extended. Others have discussed that the average speed required to span the distance, 6th to 7th arc, must be kept high. You might try calculating how that could be realised with the extension of flaps that’s necessary for a significant downward deflection of the outer flap.”

    Don, I’m not sure what you’re trying to get across here. Perhaps the Ge Rijn issue is beginning to mess with your psychic composure. I sense a rising panic encroaching on your ability to remain rational. Take it from me, the pilot was in control right up to the end.

    No calculation is necessary because there was no flap reflection involved. On my flight path, the aircraft maintains average M0.81 between 0011 and 0019.50, quite in keeping with an aircraft performing an unpowered glide from an altitude of about 38,000ft from 0017.50 onwards, maintaining a bearing of 187 deg.

    I have no problem with flaps retracted. As I said earlier, the ATSB’s presumption that a piloted ditching would necessarily be carried out according to Boeing’s recommendations under engine power with flaps extended, while refusing to consider what the debris is evidencing, which is a piloted impact intended to damage the fuselage sufficiently to cause rapid sinking of the airframe without creating a large floating debris field, is simply a cynical ploy designed to limit the search area, and at the same time, avoid causing Malaysia any inconvenient embarrassment.

    No pilot, either sane or insane, would entertain a controlled, flaps down ditching after kidnapping 238 innocent people and flying them to a watery grave in the SIO. Some of the passengers may have survived the journey.

    The ATSB stance on this issue is patently untenable to anyone with enough intellectual capacity to properly grasp the situation.

    So, it’s flaps up all the way. It does away with sweating about how long the APU will keep going.

  207. Brock McEwen says:

    @ALSM wrote: “Until last month, ATSB had been using for the “reference arc” the 40,000 ft arc. Following a discussion with them about the implications of the BFO data, they agreed to switch to 20,000 ft for the reference arc. (It is obvious the plane was much lower than 40,000 feet at 00:19.)”

    Are you saying the ATSB spent the entirety of its budget searching its box without ever making this 3nm adjustment – an adjustment that was self-evident to all from cursory inspection of the BFO values, which were published well before the search box was ever constructed in the first place?

    How about in 2016, when the primary strategic decision they made was to WIDEN the box further?

    If that simple, common sense altitude adjustment is made – as every analyst of consequence would have been screaming at them to do – then the hypothetical aircraft position at Arc 7 is 3nm NW, but also a couple of nm LOWER in the sky. These two shifts effectively cancel out when considering the INSIDE boundary of the search box, but work together to cause the OUTSIDE of the search box to move dramatically NW (i.e. a much narrower box). A narrower box could have been searched to a much greater length. Heck, they may even have had time to search where their own chosen drift analysis told them to search.

    What a disgustingly dangerous dereliction of duty. We should audit them, shouldn’t we?

    Second question: on what date did you yourself first realize that the BFO data (if reliable) meant that the plane, by Arc 7, had to be well below cruising altitude, losing altitude sharply, and accelerating rapidly toward the ground?

    Third question: on what date did you first publicly call for the narrowing of the search box to reflect this realization?

    Last question: if the gap between those two dates was due to uncertainty in the BFO data’s reliability, what changed to increase
    its reliability in your eyes?

    Thanks and regards for any time and attention you can give these questions.

  208. Don Thompson says:

    @Rob,

    Fair enough, you’ve caught me out for giving scant attention to the detail of your hypothesis.

    But please do, ever so briefly, explain the rationale for “Take it from me, the pilot was in control right up to the end.

    I’ll freely admit frustration at the absence of a cogent argument supporting the notion of control through the end of flight. I can’t help but read “Take it from me” as simply prefacing another GRinion.

  209. David says:

    @Brett McEwen. “I suggest you follow up with the ATSB”.

    They had added, ‘Given the commercial origins of the information, and the confidential nature, we do not have any additional details of those simulations for public release.”

    Gysbreght also has tried to access more information on those simulations, with limited success.

    I do not draw a conclusion that there is anything deliberately defensive or deceptive about this.

    I had raised with them the likelihood of flaperon imbalance in RAT sole operation as early as mid-February 2017 having also raised it on the JW web. I do not know what happened to that.

  210. Donald says:

    @Don Thompson

    You say “But please do, ever so briefly, explain the rationale for “Take it from me, the pilot was in control right up to the end.”

    If you accept that this event occurred because of a deliberate and purposeful action of the PIC (which seems bloody obvious to most, though notably not to all), and you accept the simulator data AND that the aircraft is still missing is evidence of intent (to fly into the SIO) and ALSO evidence of a successfully executed plan…then you are left with only only one scenario in which the PIC is not in control until the EOF.

    That one and only scenario is the pilot choosing to take his own life prior to the EOF, but AFTER knowing good and well that his actions will have resulted in the murder and disappearance of 238 human beings.

    The other scenario, of course, is a PIC knowing he has murdered 238 human beings but has yet to complete the task at hand. That task being the disappearance of the aircraft.

    Premeditated, complex, calculating and cunning acts of such barbarity and malevolence don’t relinquish control until either they are forced to or have satisfactorily completed their mission and attained the objective. They control everything that’s within their power and capabilities. The way and manner MH370 impacts the water and finds its final resting place would have been at the very top of Z’s concerns and considerations…it was everything, really. Sorry this is so difficult for so many to grasp.

    I’m as sure as sure can be that MH370 was put into the steep dive by Capt. Zaharie.

    So what do you think happened Don Thompson?

  211. HB says:

    @Rob,
    I cannot agree more. For me, a detailed debris forensic should have been a pre-requisite to any new search. We all know what happens when we cut corners. Now there is still likely that some debris which are the basis of the search location did not fall at the same location as the main body. If this is what actually happened, the scanned area should be quite large to get any chance to find the main body. I don’t know how long a plane can remain airborne without a flaperon.

  212. airlandseaman says:

    Brock:

    You are blowing the 7th arc issue up bigger than it needs to be. First, there is nothing right or wrong about using a reference arc defined as the 40K arc, or 20K ft arc or the 0ft arc. Any reference arc works, as long as the altitude is noted and everyone takes that into consideration when planning the detailed search area. We have often published papers using both the 0ft and 35K ft arcs to emphasize the difference associated with the altitude assumption.

    Switching to 20K ft just makes the search areas slightly easier to lay out since it is closer to the likely true altitude, and thus can be used as the “centerline” for the arc, rather than the outer limit.

    The Boeing calculations date back to April 2014. At that time, the calibration on the R600 offset was not known as well as it is today. I do not know how they came up with 4560 usec back then. But considering the chaos at that early time, they did an excellent job in my opinion.It’s actually pretty close. Their April 2014 estimate differs by less than 3.0 nm from the best 40k arc we can estimate today, based on nearly 4 more years of additional study and calibration. At the time, a 3 nm error was considered a small error compared to the other uncertainty, especially the post fuel exhaustion path uncertainty at that time (think Hardy, et. al.). Therefore, it was decided not to peruse further perfection of the arc at that time, and instead, they turned their attention to the larger uncertainties. Good choice.

    Could the 7th arc BTO calibration have been improved earlier? Yes it could. Would that have changed things materially? Probably not. Should OI take advantage of all the latest information? Of course. Let’s leave it there.

  213. Pilatus says:

    @Andrew,

    The maintenance manual tells us 3 things:

    1. The flaperon PCU is in the normal mode when there is no failure in the PCU or its ACE.

    2. The flaperon PCU is in the bypass mode when there is an electrical or hydraulic failure with the PCU or its ACE.

    3. With the aircraft on ground, airspeed < 85 knots and an engine near take-off thrust, the flaperon bypass logic of the PFC normal mode operates which puts the PCUs in bypass.

    I would suggest that an aircraft electrically powered at the gate with hydraulic power off would have the PCUs in normal mode. I would also suggest that the HYDIM inputs to the ACEs and PFCs would be used by various fault and failure monitors to prevent erroneous faults and messages from occurring eg. flaperons drooping at the gate without hydraulic power.

  214. Brock McEwen says:

    @ALSM: I’m afraid you mixed me up with someone else. My questions (a few comments above) were specifically about 00:19 BFO reliability, and its implications for end-flight modeling driving search zone width. I also sought two specific dates. Thanks again.

    @all: somewhere out there is a commenter who sought ALSM’s opinions on logon offset minutiae and/or model risk motherhoods. He accidentally replied to me (couple of comments above). It would be a shame if you missed it.

  215. David says:

    @Brock McEwen. Sorry to have renamed you Brett above, knowing your name full well; that is supposing that Brock to be your preference.

  216. Don Thompson says:

    @Pilatus

    I agree with yout statements 1, 2, and 3. 3 is a consequence of the PFCs blending the HLCS status into the flaperon commands, similar occurs on landing.

    Your point about conditions at the gate. With no electrical power, the various PFCS PCUs are in a state where all solenoid valves are de-energised. Each control surface has a particular design of PCU, and in the de-energised state all but the elevator allow some movement of the control surface: flaperons droop, ailerons droop (over time), rudder may deflect (slowly, with PCU damping). With electrical power only, the ACE will detect the control surface position (feedback from the PCU LVDTs) but without hydraulic power, the control surface cannot be restored to the normal position.

    Actuation of a control surface is normally effected by two PCUs, an inboard and an outboard unit. Each PCU is supplied commands & electrical interface by an ACE-PSA lane (L1, L2, C, or R) and hydraulic power from the corresponding system (L, C, or R).

    The aircraft maintenance manual (or training manual) does state that when the aircraft operates under RAT power alone, the left flaperon (controlled by L1 and R ACE, supplied by L & R hydraulic systems) will fair upwards into the slipstream and, if uncorrected, invoke a left wing down roll. Considering electrical and hydraulic supply to the L1 ACE-PCU and R ACE-PCU, during RAT only generation, the logic of their operation suggests that both ACEs must de-energise their flaperon PCU bypass valve for the flaperon to fair. The common ‘non-normal’ to both, with only RAT generation, is a loss of all pump pressure in the L and R hydraulic systems.

  217. David says:

    @Andrew. About minimum hydraulic pressure you said to me, “I agree, the PCUs would most likely remain in normal mode, assuming electrical power remained available.” The flaperon actuator LVDT was the operative sensor I had in mind as providing feedback of PCU failure under low hydraulic delivery.

    You commented later to @Mick Gilbert,@Don Thompson, to the effect that change to by-pass mode would a result of a discrepancy in the EHSV position or in differential pressure across the actuator via their LVDTs. Did you mean the actuator LVDT?

  218. Don Thompson says:

    @Donald,

    I’m more interested in an exploration of the aircraft behaviour under the conditions experienced after second engine fuel exhaustion. I’m not satisfied that RAT-APU-RAT sequence after main engine fuel exhaustion has been wrung dry. Knowledge is more useful than blind faith.

    Thanks

  219. Andrew says:

    @David

    RE: ”You commented later to @Mick Gilbert,@Don Thompson, to the effect that change to by-pass mode would a result of a discrepancy in the EHSV position or in differential pressure across the actuator via their LVDTs. Did you mean the actuator LVDT?”

    Yes. I posted a correction.

  220. HB says:

    @Don, if not already read, this document decribes the logic and set points for all the interlock and permissive logic. http://www.flight-mechanic.com/large-aircraft-hydraulic-systems-boeing-777-hydraulic-system/

    My understanding from this document is that the PCU control logic is totally independent from the hydraulic power / pressure logic. The only relationship is the hydraulic user shedding in case of leakage. The only common element in case of both engine failure is the RAT driving both power and hydraulic pressure.

    I am not sure what you are trying to prove but I personnaly don’t think hydraulic pressure will reduce to the point that the PCUs will switch to bypass. In my experience of hydraulic systems, if there is no major hydraulic power consumption from the users, it will take some time for pressure to drop (minutes). Also, there is effectively a hydraulic user shedding corresponding to the electrical load shedding.

    The interlock logic, combined with check valves, will maintain the pressure to the critical users after both engine failures. There should be enough residual pressure to maintain the required hydraulic pressure until RAT deployment as per design intent. Ie the discrepancy in hyraulic pressure mentioned in the above posts should not be experienced for this case.

    In case of power failure, before the RAT deployment, the system should revert to default designed condition ie bypass or locked depending on the PCU configuration (like mentioned above). What is not clear is whether the PCUs reset to original control set point conditions or not after RAT power restablishment. It should autorest as per any control system (ie is the ailerons droops, it should go back to the required position as per the control system requirement if auto pilot on and original position if autopilot disengage, and not remaining in the “droop position” as a natural control response.
    If it is the case, in the RAT-APU-RAT sequence the controlled default position of the PCUs should be observed at all time ie PCUs under RAT power remain under control and PCUs not under RAT are set to default position (after electical power/ hyraulic power shedding).

  221. Pilatus says:

    @Don Thompson,

    I think you’ve made some assumptions based on your own beliefs on how hydraulic flight control systems actually operate under engine failure.

    The report below, although relatively old, describes the operation of a 747 following an engine failure (refer page 19/20). We can take from that how a system operates on a windmilling engine (for the 747, 32 GPM and 777, 48 GPM).
    https://www.atsb.gov.au/media/1994828/aair199403038_001.pdf

    Interesting to note is that the system involved also operated the landing gear and flap system. If you are aware of the 747 flap system, it has a load alleviation function that contributed to the incident due to the non-stabilised approach although not specifically highlighted in the report (the flap bit that is); the ATSB tend to focus on blame rather than cause but that’s my opinion.

    The quote you referenced indicates hydraulics off, not depressurised. Remember these are simplified English manuals.

  222. airlandseaman says:

    Brock McEwen says:
    January 26, 2018 at 1:43 am
    @ALSM: I’m afraid you mixed me up with someone else.

    No. I was answering your post.

  223. Don. Thompson says:

    @HB, @pilatus

    I’m attempting to understand the behaviour described in the 777 tech docs where it’s clearly stated that the left flaperon fairs when the RAT is the only source of hydraulic pressure (on C system).

    We’ve been ‘around the houses’ with flaperon behaviour during the take-off & landing phases. You alluded to that above. The flaperon PCUs are unique among the ailerons, elevators & rudder, in that the bypass behaviour is unrestricted.

    If the only possible sensing for hydraulic power is ACE-PCU using the LVDTs to detect PCU response to ACE commmand signalling, that’s fine & I’d be very happy to get that emphatically confirmed.

    However, I’m concerned that such a scenario presents an unpredictability. I’d be more confident that safety would be best realised if the pump pressures for a given system contributed to the ACE putting specific PCUs into bypass mode. The manual describes a predictable condition.

  224. airlandseaman says:

    Brock asked: “Are you saying the ATSB spent the entirety of its budget searching its box without ever making this 3nm adjustment – an adjustment that was self-evident to all from cursory inspection of the BFO values, which were published well before the search box was ever constructed in the first place?”

    No. I did not say that.

  225. Andrew says:

    @Pilatus

    Thanks for your earlier comments. I agree with the three points you raised, although I don’t see how #3 is relevant to the in-flight case.

    Regarding the ‘parked at the gate’ scenario, I agree, the PCUs should remain in normal mode unless electrical power is removed. This photo shows a parked B777 with its right flaperon ‘undrooped’.

  226. Brock McEwen says:

    @ALSM: thanks for clarifying. I just wanted to give you the benefit of the doubt, since the response had nothing whatsoever to do with the question.

    I’ll just put you down as “evasive”, and we’ll leave it at that.

  227. Victor Iannello says:

    @Brock McEwen: You’ve worn out my patience. I’ve for too long let this blog be used by you to obtain information from contributors that you solicit, record, and distort for your own agenda. You’ve done it to me many times. You do it to others. You add nothing to the discussion here except to distract contributors that are seriously trying to help solve the mystery. You are banned.

  228. Ge Rijn says:

    @Andrew

    On your photo of the right flaperon in neutral when parked (if engines off, APU off, only electrics) I see also the left flaperon is in neutral in this photo:

    https://www.flickr.com/photos/raihanshahzad/6056814009

    I would think if both solenoids of the PCU’s on the left flaperon would be activated after stopping the engines (and APU) that left flaperon should droop when in bypass and free-floating if this depended on hydraulic pressure faiding.
    It seems it does not. Then it seems to me this only depends/happens when electrical power is lost/cut off.

    You mentioned before that batteries would take over electric supply for about a minute after second engine flame-out till the APU takes over supply of electrics and hydraulics.
    This would mean there would be no time in between where the left flaperon could lift to its free-floating upwards position.

    Does this make any sence to you?

  229. TBill says:

    @Don Thompson
    ” Others have discussed that the average speed required to span the distance, 6th to 7th arc, must be kept high. You might try calculating how that could be realised…”

    Don you make a strong argument for the passive flight theory. Let me just say what I perceive the generic problem is. For some reason on MH370, most of the theoretical physicists/scientists are advocating for the passive flight scenario. We lack theoretical support for the active pilot case, which could mean it is not a realistic case, but without exploring that case I am not sure. The fact that Victor does not squelch that discussion would seem to imply it is not disproven theoretically. I would say the pilot could have bridged the gap between Arc6 and Arc7 by several options, one being descent and slow down so the fuel lasted longer. But I got other ways to do it too if that idea fails.

  230. airlandseaman says:

    TBill et. al.

    Why is it that so many people keep conflating (a) the evidence for a rapid descent with (b) the question of whether the plane was under human control at the time. They are independent. We know with reasonable certainty that a rapid descent was underway at 00:19. We do not know if anyone was alive at that time, and if someone was, whether there was any human control involved.

  231. IR1907 says:

    @ALSM,

    Just because there was a rapid descent doesn’t mean Zaharie is innocent. People automatically connect

  232. IR1907 says:

    Continuing from above…

    People automatically connect a rapid descent with a ghost flight. A rapid descent was part of the plan. It was carefully executed.

  233. Ge Rijn says:

    @TBill

    On your comment:

    “..For some reason on MH370, most of the theoretical physicists/scientists are advocating for the passive flight scenario. We lack theoretical support for the active pilot case, which could mean it is not a realistic case..”

    To me your definition is the other way around. We lack any logical, evidence based, theoretical (and with it statistical) support this active pilot flight turned into a passive flight before, at- or after 18:40.
    The odds it was an all active pilot case is far more logical and should have been the prime assumption in the first place.

    IMO a lot of time and effort has been wasted by people trying to prove a theoretical, practical and statistical near to zero impossible scenario happened. While ignoring/denying the most obvious.

    Implications for the current search can be great. If OI takes the same ‘logic’ they are at great risk off failing like the ATSB did based on this irrational assumption.

    Like @ALSM says; “We do not know if anyone was alive at that time, and if someone was, whether there was any human control involved.”

    This statement at least leaves room to the IMO more obvious possibility there was someone active at the controlles till the end.
    I hope OI has taken this possibility into account seriously.

  234. Nederland says:

    @Ge Rijn

    Having worked on possible flight routes and proposed my own, I don’t think it is possible to reconcile the northern area currently searched with complete automated flight past 18:40.

  235. paul smithson says:

    @Victor. I applaud you long suffering patience with our self-appointed auditor as well as your decision to ban him.

  236. Victor Iannello says:

    @paul smithson: Thanks.

  237. Mick Gilbert says:

    @Victor

    I echo Paul’s remarks, your decision should take a couple of points off everyone’s blood pressure.

    Regarding the OI search, I note that over on Richard Cole’s site he has observed that Seabed Constructor appears to have redeployed the AUVs to search a swathe to the south and outside of the CSIRO’s priority search area.

  238. Pilatus says:

    @Don Thompson
    @Andrew,

    The flaperon PCU bypass solenoid provides pilot pressure to operate the mode selector valve to the normal mode. If the solenoid is in normal and there is a bleed down in pilot pressure the mode selector valve is spring biased to the bypass mode.

    On other PCUs the blocking damping mode is the biased condition.

  239. lkr says:

    @Mick: The blood pressure does seem to be rising among long-time contributors — I think it’s sinking in that “this is it, folks” and those who have long nursed pins in the map are feeling a bit desperate. But please, at least a little ‘IMO’ is helpful. And a willingness to see that time spent in one’s favorite hole is lost opportunity elsewhere.

    On your substantive point, is it correct that we all use the convention “outside the arc” to mean further from the satellite and further from the sixth arc? So to the south and east? [I could imagine a contrary Ozzie orientation..] If that’s where OI is looking, GR must be a little pleased, but are there other, non-glide reasons to do so?

  240. Don Thompson says:

    @Pilatus

    The pressurised system flow into the PCU is directed through a ‘check valve’, through the energised bypass valve, causing the mode selector valve to be set to the ‘normal’ position.

    Given that the circuit includes the check valve, can you explain the ‘bleed down’ that permits the mode selector valve to relax to the bypass position?

  241. Mick Gilbert says:

    @Ikr

    Regarding OI latest AUV sortie, yes, per Richard’s plot of the Seabed Constructor, they appear to be searching to the south of the southern boundary of the CSIRO’s priority search area (ie further away from the 7th arc). The southern boundary of the latest search zone appears to be about 50 km from the 7th arc (I’m not sure which variant of the 7th arc Richard uses in his illustrations).

    The extra swathe to the south might just represent an abundance of caution on OI’s behalf with regards to a rapid descent scenario. If they continue to repeat the approach with their next AUV sortie(s) you might well infer that they are entertaining the controlled glide scenario.

  242. Don Thompson says:

    @HB

    Your referred document doesn’t once mention the PCUs or ACE, I don’t see how it’s useful to the discussion.

  243. Andrew says:

    @Pilatus

    Thanks – I was thinking about that point earlier this morning and I agree; the loss of pilot pressure would cause the mode selector valve in the flaperon PCUs to move to the bypass position, assuming sufficient leakage past the check valve while the hydraulic system is depressurised.

    @Ge Rijn

    RE: ”Does this make any sence [sic] to you?”

    The ACE will de-energise (not ‘activate’) the bypass solenoid if there is an electrical or hydraulic failure in the PCU, or an electrical failure in the ACE itself. In the dual engine flame-out case, there should be continuity of electrical supply from the main battery or FCDC batteries (worst case) and I don’t think the ACE will detect a hydraulic failure. If that’s the case, the flaperon PCUs should remain in normal mode following the dual engine flame-out.

  244. David says:

    @Don Thompson.
    About automatic shutdown, what would be the criterion even were there one?
    A clue as to minimum pressure desirable for manned operation might come from RAT deployment, which can be by HYDIM when the centre system or pump pressures are ‘low’ when both engines are below idle. However the manuals do not say how low is ‘low’.
    Also, RAT hydraulic pump output is by-passed until it has spun up so appears to be speed dependent, not hydraulic pressure, maybe for electrical output. So that does not help either.

    Logically there will be no minimum pressure required for flutter suppression, which I speculated about earlier, because the aircraft design would foresee an eventual drop by leakage to whatever pressure is left in the reservoirs, flutter suppression therefore being left to the selection of ‘blocked’ in the PCUs of susceptible flight controls.

    I think the same logic would see little purpose in an automatic shut-down at all.

    However there is a necessary case for an automatic HYDIM initiated reawakening of ACEs on restoration of hydraulic pressure by RAT or APU, though that is unaddressed in the manuals.

    Incidentally, the 2 seconds I have been describing as RAT spin-up time post-deployment is in fact deployment time, after its initiation 15 secs after transfer-bus loss of electrical power; or alternatively when centre system/pump pressures are low, by HYDIM as above. Supposing that centre system pressure remains sufficient though, to those 17 secs of the alternative should be added spin-up time (unstated) and @Andrew’s 5-6 secs engine run down to idle, to total the delay in the RAT coming on line after fuel exhaustion.

    About system pressures more generally after fuel exhaustion, a possible complication. While the centre ACMPs will shut down at engine speed decay below idle I do not know whether there would be sufficient engine air bleed below idle to power the centre air driven hydraulic pumps, which might be selected on automatically at that point, that joining pressure sustenance by engine driven pumps in L & R systems, a possibility @Pilatus has increased via the 747 example. The effect could be to delay or stop RAT deployment if so. All the same I suppose this would have been replicated in simulations.

    Finally, on the glide Ge Rijn said, “the chance of an extraordinary and successful effort to recover and glide is still not proven to be impossible“. I think that is fair, just as the final BFOs might prove fallible – though we would agree that the prospects are not such that they should influence the current search.

  245. Richard Godfrey says:

    @Mick

    You stated: “The extra swathe to the south might just represent an abundance of caution on OI’s behalf”

    I agree, OI are being very thorough.

    They appear to be using a separation between AUVs of around 1,600 m, so with 8 AUVs operational, this allows them a swathe width of 12.8 km (6.9 NM). This was just too small to cover the 8 NM from 17 NM out to 25 NM required for this part of the search.

    In addition the CSIRO hot spot #1 was just outside 25 NM from the 7th Arc referenced at 20,000 feet.

    So if you have to do a second swathe, you might as well use all 8 AUVs and cover from 17 NM out to 30.8 NM.
    This gives a good margin of error around the CSIRO hot spot #1 as well.

    The original ATSB search area used the 7th Arc at 40,000 feet as a reference. The residual search area from 17 NM out to 25 NM is with reference to the 7th Arc at 40,000 feet. The new reference is the 7th Arc at 20,000 feet. This moves the reference line between 3 NM and 4 NM to the northwest. Another reason for OI to be cautious.

    Victor and Mike have made significant improvements in how the 7th Arc is calculated, as explained in the post above, but there is still a small margin of error. This margin of error will in the worst case add another 1 NM, to the 4 NM noted above gives you the 5 NM discrepancies noted by @sk999 above where he states “This range corresponds to a spread of about 5 nm in the cross-arc direction, which seems rather large.”

    Richard Cole has previously depicted the 7th Arc at 10,000 m (32,808 feet).

  246. Richard Cole says:

    @Mick Gilbert

    Definition of the 7th arc – The 7th arc definition I have used in the maps was from analysis of the BTO data I did some while ago. Checking with the DSTG book and particularly to the CSIRO drift study reports it is 5km inside the 7th arc definitions in those sources.

    However, re-checking Fig 14 of the First Principles report (the only figure with useable scales), the 7th arc definition is massively different from the other sources, particularly at the north end. At 93E it is plotted 50km north of the CSIRO definition (and hence the plotted search areas are also 50km north since they are shown with respect to the 7th arc). Since Fig 14 is used in page 1 of the maps issued by the Malaysian authorities it has caused confusion. It’s the reason why the Malaysian map indicates the CSIRO priority point to be someway outside the ‘First Principles’ area, while David Griffin set it to outside the searched area, but inside the ‘First Principles’ area.

    The unscaled coloured map of the three priority search areas on page 3 of the Malaysian maps can only be positioned geographically (and hence with respect to the 7th arc) by using the grey area that corresponds to the already searched area, which I have measured from the actual ship tracks. The positioning is about right in the maps I have produced, but is only indicative.

    Conclusion: There is no published source (that I have seen) that gives an authoritative and believable definition, with numbers, of where is the First Principles search area and how it matches to the priority areas in the Malaysian Authorities’ maps. Hence, statements, including those by me, that OI is operating outside the agreed areas are premature.

  247. Mick Gilbert says:

    @Richard Godfrey
    @Richard Cole

    Gentlemen, thank you both for those clarifications.

    Richard C, thank you for the detailed progress reports on your blog, they’ve been fascinating. Following OI carve through in two days what it would have taken Furgo more than two weeks to accomplish is really something (apple v. 8 apples comparison notwithstanding).

    Might I ask what is the source of such accurate position, speed and direction data for the Seabed Constructor?

  248. Richard says:

    @Mick

    I use the Big Ocean Data website. You can get a free trial for 7 days, thereafter it costs. I am only tracking SC, so the cost is not prohibitive.

  249. Mick Gilbert says:

    @Richard C

    Thanks for that Richard, I shall go and have a look at Big Ocean Data.

  250. Richard Godfrey says:

    Apologies Mick! That was Richard G. not Richard C.

    My commenter name defaults to Richard and I must get used to using Richard Godfrey.

  251. Richard Cole says:

    @Mick

    For the record I have used MarineTraffic and currently Big Ocean Data also. The latter is a superset of the former, but has a lot more data points per day in total, but a large fraction of the difference is multiple data points over short periods of time, which don’t particularly help long term tracking.

    The most useful data in both sources (IMO) is from the ORBCOMM constellation which seems to give one (sometimes two) datapoints per satellite pass, passes are roughly every 15minutes at the current latitude. The multiple datapoints/short period are from a different satellite source, I think.

  252. airlandseaman says:

    @Mick G. Contct me by email or Twitter DM.

  253. Don Thompson says:

    @David,

    Short response to a long comment: the HYDIMs sense pump pressure, not system pressure alone: two parameters. If the pumps are no longer delivering pressure, the system is unreliable. I am concerned about L & R systems, not the C which is supplied by the RAT.

  254. Victor Iannello says:

    @Richard Cole: In the First Principles Report, the width of the 25,000 sq km, from the inner limit to outer limit, is 52.5 NM, representing 27.5 NM to the inside of the ATSB’s 40,000 ft arc, and 25 NM to the outside of that arc. The bias towards the inside was to account for a lower altitude than 40,000 ft at the time of the last transmission. Whether or not they correctly achieved this in their graphics, I believe this was the intention.

    Independent of that, the CSIRO priority locations were derived from precise locations from satellite images and from drift models, and independent of the precise location of the 7th arc. It appears that these locations will be searched out to a generous radius of each. That should leave little doubt about validity of these locations. Since the data from the first two passes is under review, we might be close to ruling out CSIRO-1 (which most of us already assigned a low probability because it is 25+ NM from our best estimate of the 7th arc).

  255. sk999 says:

    All,

    Virtually all the maps from the ATSB are made using Google Earth, and in no case that I am aware of, where longitude and latitude lines are drawn, is the map rectilinear in those coordinates. The source for Fig. 14 of First Principles is not credited, but presumably it is also Google Earth. Longitude and latitude scales are marked on the periphery, but one should not assume that the map is rectilinear either. In fact, the scales aren’t even correctly drawn for the map edge either.

    As Richard Cole noted, the first figure of the Malaysian maps is a replica of Fig 14. The second figure has the same underlying map, but the coordinate scales on the periphery (although poorly drawn) appear to be drawn more correctly this time. The same caveats about rectilinear assumptions apply.

    For now, I’m using Richard Cole’s maps as the best representation of OI’s actual search zone, recognizing that there could be of order a 2 nm shift required to align it with the map that OI is actually using.

  256. Richard Cole says:

    @Victor

    >CSIRO priority locations were derived from precise locations from satellite images and from drift models, and independent of the precise location of the 7th arc.

    David Griffin’s work was always constrained by the 7th arc position. His and Oke’s report EP174155 dated 26 June 2017 states:

    “We now know that 9M-MRO is not within the area searched by SONAR (ATSB 2016). … If impact was between 36°S and 32°S, as concluded by the First Principles Review, the aircraft must (obviously) be farther from the 7th arc than the region that has been searched, but still within a distance that it could have glided after commencement of descent. This is how the boundary of the new search area proposed by ATSB in Dec 2016 was defined.

    Our confidence in our ability to associate the locations of the Pleiades images with points within the proposed search area is not high enough to attempt to comment on the optimal width of the two narrow (10-30km) strips east and west of the completed search area. We are quite sure, however, that we can (at least) rank the four quarters of the proposed search area in terms of likelihood of connection with the Pleiades images. We will refer here to these 4 quarters as west1 (southern half of western strip), west2 (northern half), east1 and east2.”

    So he placed his candidate areas just outside the searched area, but inside the first principles review area. The paper then prioritises the areas using the drift modelling and satellite images. His first priority point is 20nm outside the 7th arc definition he was using in the paper, though the optimal 7th arc position estimate may have changed subsequently.

    In any case the search has now covered to 29nm outside (his) arc position, so his first prediction has been well covered.

  257. Victor Iannello says:

    @Richard Cole. I agree.

  258. ventus45 says:

    It would be nice if the ATSB simply published their definition of the arc altitude and it’s coordinates in detail, but alas, what has been noticeable from the beginning, is that they never have, and presumably have no intention of doing so. Since the arc is so central to the whole exercise, it makes you wonder why they haven’t.

    Anyway, to another matter. The data gathered is slowly being released in stages to the geospacial community. Some may find these interesting.
    http://proceedings.esri.com/library/userconf/oceans17/papers/ocean_16.pdf
    https://www.youtube.com/watch?v=dOcac4S_osQ

  259. Ge Rijn says:

    @Andrew

    On your comment to me;

    “.. In the dual engine flame-out case, there should be continuity of electrical supply from the main battery or FCDC batteries (worst case) and I don’t think the ACE will detect a hydraulic failure. If that’s the case, the flaperon PCUs should remain in normal mode following the dual engine flame-out.”

    Do I understand it well then when I conclude the left flaperon will stay in normal mode after second engine flame-out till the APU takes over and then also stays in normal mode till the APU runs out of fuel?

  260. Victor Iannello says:

    @Richard Cole has tweeted that Seabed Constructor is now moving towards the inside sector of the 5,000 sq km primary area. That suggests that nothing was found at CSIRO-1. The next sector will cover CSIRO-2 and CSIRO-3.

  261. Ge Rijn says:

    @Andrew

    And forgive me my non-senSe Dutch English at times. I lack the finesses of the English language which I understand also cause some misinterpretation the way I intended to state things.

  262. DennisW says:

    @Ventus

    It would be nice if the ATSB simply published their definition of the arc altitude and it’s coordinates in detail, but alas, what has been noticeable from the beginning, is that they never have, and presumably have no intention of doing so. Since the arc is so central to the whole exercise, it makes you wonder why they haven’t.

    It would need to be a table of values (derived by iteration) since a closed form solution is not possible, IMO. I say IMO, since I could not generate one after recent discussions on the subject. The reality is the earth radius at the equator is 6378km and at 40S is 6369km. By using the correct subsatellite point, and an approximate average radius of 6373km one ends up with errors smaller than the errors in the altitude assumptions for the arc location when using a great circle radius approximation.

    Bottom line is that it should not be an issue.

  263. Ge Rijn says:

    @VictorI

    This is going to be a blood-pressure raising endavour for all us so closely involved. But the most for the NoK.
    I wish them strenght and some comfort in realising; again there are people on location who will do their utmost to find the plane. And with the support of most of the involved (Internet) community.

    The speed and thoroughness OI is showing is realy amazing. Going some extra Nm outside the box gives me confidence in their flexibility.
    That’s the way to go IMO.

    CSIRO-1 seems deleted. Not defeat but success. Next move.
    It seems they are doing great so far.

  264. Tim says:

    All, I’m curious to know.

    If the Penang turn never happened(I’m still to be convinced the military radar trace is correct) and MRO continued over Sumatra and out into the IO before turning south. Can you still make all the BFO/BTOs fit?

    Might help if you assume a wobbly route- as it was seen to be doing as it flew across the Malay Peninsula.

  265. HB says:

    @Don
    I was also speaking from my experience on hydraulic system failure. All HPUs follow the same principle in terms of response to power failure having analysed many.
    There is no low hydraulic oil pressure interlock acting on the PCUs for this configuration. PCUs are users and they will only default to their designated fail safe conditions in case the control system detects an incorrect response. The C hydraulic system and its users should be unaffected by the transition to RAT ie no momentary loss of hydraulic power (still within operating range). The L and R system will be shed as per the logic ie the users will default to their designated fail safe conditions. All in all the system should meet its design intent in case of both engine failures. I am not sure what is your concern.
    The only thing which is not documented is what is the schedule of default fail safe condition for each user.

  266. Paul Smithson says:

    @Tim. I was wondering the same. You can certainly obtain intersection with the 1825-1828 ping rings around the right time (I think it was Ventus who pointed this out) in vicinity of 3.1N 96.2E. However, the same (~242T) course is not going to fit the BFO of ~143 at 1828. You would also still require a turn to the south to fit subsequent BTOs. I haven’t tried path fitting from that assumed starting location. Given the rotational symmetry of the BTOs, I would expect that a BTO solution of sorts could be found but I’d expect that the BFOs would be way out of whack because it would be a long way further south and west.

  267. DennisW says:

    @Paul/Tim

    The suggested path would also have a very early effective FMT, and the plane would wind up in Rob territory and beyond. The “semi-concensus” late FMT is needed for the more Northern terminal locations. I am not at all a fan of a continuation over Summatra.

  268. Victor Iannello says:

    @Ge Rijn said: It seems they are doing great so far.

    Much to the dismay of the naysayers and mudslingers.

  269. Rob says:

    @DennisW

    “The suggested path would also have a very early effective FMT, and the plane would wind up in Rob territory and beyond. The “semi-concensus” late FMT is needed for the more Northern terminal locations. I am not at all a fan of a continuation over Summatra.”

    I’m not a fan of a very early FMT or a very late one, come to that. UTC1835.50 will do just fine. It’s in what I call the Goldilocks zone.

    The sooner they are done with the 25,000sq km zone, the better. OI can then move on to more promising “territory”

  270. Don Thompson says:

    @HB

    You appear to be describing generalities, not 777 systems specific knowledge.

  271. DennisW says:

    @Rob

    It’s in what I call the Goldilocks zone.

    My life coach instructor always advised us to worry less about Goldilocks than why momma bear and papa bear don’t sleep in the same bed anymore.

  272. airlandseaman says:

    BTW… An OI person mentioned to me today that they are very please with the SSS image and data quality. So, if 370 was at CSIRO #1, they should have detected it. I doubt if they have finished processing all the data collected in the last 24 hrs, so it is still possible that it will be found there. But by tomorrow, it should be known.

  273. Ge Rijn says:

    @Rob

    “The sooner they are done with the 25,000sq km zone, the better. OI can then move on to more promising “territory”.

    I then have to disappoint you it won’t be to the south but to the north.

  274. Rob says:

    @Ge Rijn

    “I then have to disappoint you it won’t be to the south but to the north.”

    I’ve got news for you then – You had better be wrong, otherwise the search is doomed to fail, and I’m certain that’s the last anybody wants.

    The further north you go, the less you are likely to find the aircraft.

  275. DennisW says:

    @ALSM

    BTW… An OI person mentioned to me today that they are very please with the SSS image and data quality.

    Excellent. That is very good news.

  276. DennisW says:

    @Ventus

    Sorry, I refuse to use dropbox. It is frustrating and obsolete.

  277. ventus45 says:

    @DennisW
    Yes, dropbox is frustrating.
    Just copied it here:
    https://drive.google.com/file/d/1SDUzYy1oFALceu-Ev5tcZsXZOAnWWJw2/view?usp=sharing

  278. falken says:

    @Richard

    “Please tell your Aunt with Schadenfreude, that according to Heisenberg’s Uncertainty Principle, MH370 is more likely where it is, than elsewhere.”

    yeah, including uncertainty that he knew howto…

  279. DennisW says:

    @ventus

    thx!

  280. sk999 says:

    Ventus,

    Those links to the talk and slides by Megan McCabe were very useful. Slide 5 is made by Geoscience, presumably with its in-house software (not Google Earth) and is replicated as slide 4 in the Malaysian OI this with a lat/long rectangular grid (unlabled, but one can figure it out.) Digitizing the 7th arc shows that is is, indeed, an arc of constant BTO.

    Likewise, Fig 1 of the ATSB Dec 3, 2015 report shows an arc that is also one of constant BTO (I find a difference of 9 microsec between the two).

    The average of the 2 is abot 24 microsec greater than the BTO used by Richard Cole, corresponding to an outward displacement of 3 nm.

  281. David says:

    Ventus45. About your video some points:
    – Budget prevailed in the original 60,000 sq km, pruned from 1m sq km, and some later area adjustments.
    – Political pressure to convert the surface search to bottom.
    – At 18:50, data gap of 100m X 100m seens as acceptable. Given the reflective targets and clean flat seabed used for calibration and the small number of wreckage items which might have that reflectivity (size, material, incident angle, part burial) and also the possibility they would escape detection if solitary, I hope there were few such gaps. If not, were they treated as of less than 100% confidence in search success probability assessment I wonder?
    – overlap of 500m @1100m range quoted.
    – Another tranche of ATSB/GeoScience data to come.

  282. Andrew says:

    @Don Thompson

    RE: ”If the pumps are no longer delivering pressure, the system is unreliable.”

    That’s possibly the key misunderstanding we have here; hydraulic pumps don’t generate pressure, they generate flow. The pressure is a function of resistance to the flow of fluid within the system. In this case, the pump and system pressures will be maintained at windmilling RPM, because there is little or no flow, assuming no control inputs or other demands on the system.

    More later, I’ll be on the road for the next few hours…

  283. DennisW says:

    @sk999

    The average of the 2 is abot 24 microsec greater than the BTO used by Richard Cole, corresponding to an outward displacement of 3 nm.

    Not sure how 24 usec BTO difference translates to ~3nm in the arc position.

  284. Mick Gilbert says:

    @Andrew
    @Don Thompson

    Re: “… assuming no control inputs or other demands on the system.
    Aye, there’s the rub.

    My knowledge of hydraulic systems is scant so I might be well off piste here but wouldn’t there be some demands on the system even in straight and level flight? We know that the flaperons’ “natural” position is to float up out of the faired position. That means that holding them in the faired position must be generating a hydraulic load. I suspect (but don’t know) that the spoilers probably exhibit the same characteristics as the flaperons in that aerodynamic forces will tend to pull them up and out of the faired position. If so, that’s another 14 control surfaces to add to the two flaperons that are loading the hydraulics.

    By my count there are 31 control surface related PCUs and 18 of them are used to control the flaperons and spoilers. Only 7 of the 18 flaperon/spoiler PCUs are on the centre hydraulic system. This may be the wrong way to approach the problem but between second engine flame-out and RAT deployment 7 PCUs will be loading the centre system, 5 will be loading the left system and 6 will loading the right system. Allowing for wind-down to IDG disengagement and 15 seconds to RAT deployment, some 20-odd seconds later the centre hydraulic system will be start to be powered by the RAT. APU start is still 40-60 seconds away which means 11 PCUs will continue to load the left and right systems.

    Would I be correct in assuming that if there is insufficient flow in the left and right systems to hold the left flaperon in the faired position that as soon as it starts to move this will be read as a failure (viz mismatch betwene the commanded and actual position) and the PCUs will fail over to by-passed mode?

  285. DennisW says:

    @Rob

    The further north you go, the less you are likely to find the aircraft.

    The terminus is near 30S on the 7th arc. I am quite confident about that, and I have not stuck a pin in the map since Christmas Island (a very long time ago).

    I have weak analytics to support 30S, and Richard’s drift analytics are very good (IMO).

    Trying to reinforce my analytics with a couple of new geometric algebra books “written for physicists” has not worked out as I expected. They are not a quick read. A sign of age I suppose.

  286. sk999 says:

    DennisW,

    I did something dangerous – applied reasoning. See if I got this right.

    24 microsec – 7.2 km light-travel distance, or 3.9 nm. If the satellite were straight overhead, all that distance would be vertical travel, so no change in long/lat. If the satellite were on the horizon, all that distance would be linear distance on the surface of the Earth – all goes into long/lat change. I multiple by cos(elevation), where elevation of the satellite is about 39 degrees at the 7th arc. That gives 3 nm.

    The absolute BTO depends on the altitude you assume; the differential doesn’t matter as long as you use the same altitude for all.

    Does that make sense?

    Regardless, in a future post I will make a prediction (sure to go wrong) of where the plane will be found. Unfortunately is only in the $30 million zone – was hoping it would leak into the $50 million zone, but no.

  287. DennisW says:

    @Andrew

    hydraulic pumps don’t generate pressure, they generate flow.

    Just a word of advice to someone I respect greatly. Be very careful going forward.

  288. HB says:

    @Don, Andrew. I am talking about the B777 configuration. You need to understand the relatonship between flow, pressure and time.
    The pumps are delivering flow and pressure following back pressure control like Andrew mentioned. If pumps stop the pressure gradually decrease due to ongoing hydraulic power consumption. The hydraulic reservoir combined with the interlocking logic will ensure that there is sufficient pressure to the users to ensure a smooth transition to RAT operation. Andrew is correct that if no or little consumption the system will be mantaned at low rpm. All this is completely independent to the Control of the PCUs.

  289. DennisW says:

    @sk999

    Thx. But the disconnect I am having is that BTO is the two way time of flight distance, and an error is halved. Right?

  290. sk999 says:

    DennisW

    Duh, yeah. Like I said, reasoning is dangerous.

    Once you get tired of geometric algebra, I can loan you my book on “Catastrophe Theory for Scientists and Engineers”. Wonderful read.

  291. DennisW says:

    @sk999

    I have had a very humbling experience with my new books. I cannot take anymore. 🙂

  292. DennisW says:

    @IG@DSTG

    After a second glass of wine I reflect. I cannot hang with you guys. I do not know how to transform scalers into vectors.

  293. DrB says:

    @Andrew,

    Do you know how accurately a True Track can be held using the MCP?

    What I really want to know is, if one set a true bearing and flew for an hour, what would the typical lateral offset be, after an hour, from a perfect track line?

    I would think the error would depend primarily on wind gusts. There are at least two effects. First, variable cross-track winds can translate the aircraft off the desired line. Also, variable winds can apply moments to the aircraft and produce rotational (yaw) disturbances. The AFDS will dampen those out, but there won’t be any compensation to get back to a fixed path like there is in LNAV. So small angular and linear disturbances will occur, and these may integrate into a noticeable cross-track error. I would think the right order of magnitude is about 1% of the distance or 0.6 degree. Do you have any documents or experience addressing this question?

  294. Andrew says:

    @Mick Gilbert

    RE: “…wouldn’t there be some demands on the system even in straight and level flight? We know that the flaperons’ “natural” position is to float up out of the faired position. That means that holding them in the faired position must be generating a hydraulic load.”

    That would be true if the PCUs put a constant demand on the system to hold the control surface(s) in position, but they don’t. The PCU actuators are controlled by a servo valve that directs hydraulic fluid to one side of the actuator or the other, depending on which way the actuator needs to move. The actuator position is monitored by the ACE as part of a ‘servo loop’. When the actuator reaches the commanded position, the ACE commands the servo valve to move back to the neutral position. The actuator and the associated control surface are then effectively ‘locked’ in position until the next command input. The only time there is a demand on the system is when the actuator is moving to reposition the control surface.

    RE: “Would I be correct in assuming that if there is insufficient flow in the left and right systems to hold the left flaperon in the faired position that as soon as it starts to move this will be read as a failure (viz mismatch betwene the commanded and actual position) and the PCUs will fail over to by-passed mode?”

    No! As described above, the control surfaces are held in position by the servo valve in each PCU, not by a demand on the system.

  295. Mick Gilbert says:

    @Andrew

    Thanks for those amswers. (That was a quick trip.)

  296. Andrew says:

    @Mick Gilbert

    Yes; we were ‘down south’ for the long weekend. It only takes about three hours to get home these days and fortunately there wasn’t too much traffic.

  297. David says:

    @ Andrew. “The pressure is a function of resistance to the flow of fluid within the system.”

    A comment.

    The system pressure is controlled by pressure compensators within the pumps to maintain the nominal system pressure of approximately 3000 psi. That alters the displacement of the pumps’ pistons to meet the flow requirement at that pressure, increasing displacement as demand quantity rises. When system flow rises beyond a primary pump’s supply at full piston displacement, a standby demand pump cuts in. Air driven demand pumps for the centre system cut in at 2400 psi, out at 2700 (ACMPs for L & R, 2400/2800).

    In other words in normal operation the supply will match the demand at about constant pressure rather than demand setting pressure.

    In windmilling there will be no demand pumps (unless bleed air is still sufficient to power the air driven some). Hence primary pumps will need to meet normal system internal leakage past seals etc plus their own case oil supply, used for their cooling. @Pilatus referred to 747 engine windmilling delivery (11-12 galls/min@ 3000 psi), of which a little over half was consumed by leakage.

    In the general if the demand exceeds the remainder, system pressure will fall and it will act as you say. However, as system pressure falls so does leakage so there can be some added useful flow recovered from that.

  298. David says:

    @Andrew. To be finickity, a little LESS than half the 747’s was consumed by leakage.

  299. David says:

    @HB. You say, “The hydraulic reservoir combined with the interlocking logic will ensure that there is sufficient pressure to the users to ensure a smooth transition to RAT operation.”

    Perhaps you mean the accumulators, though they have very little capacity. The reservoirs are pressurised by bleed air supply. As to maximum pressure a relief valve operates at 85-90 psi. The ‘LO’ warning pressure there comes on at 21 psia and less. Hence they are at very low pressure compared to system hydraulic pressure

    When system hydraulic pumps are off, the system pressure indication is reservoir pressure.

  300. Andrew says:

    @David

    Thank you. I was trying to keep it simple without getting into complicated explanations that only confuse the issue.

    RE: “In other words in normal operation the supply will match the demand at about constant pressure rather than demand setting pressure.”

    I didn’t say that ‘demand sets pressure’, if that’s what you’re implying. I said ‘pressure is a function of resistance to the flow of fluid…’. For example, the pressure at the pump outlet is zero if the pump is not connected to a load.

    RE: “In the general if the demand exceeds the remainder, system pressure will fall and it will act as you say.”

    I said: In this case, the pump and system pressures will be maintained at windmilling RPM, because there is little or no flow, assuming no control inputs or other demands on the system.

    There will be some leakage, but with little demand on the system in uncontrolled flight, the primary pumps should have no problem maintaining pressure at windmilling RPM, just like the 747 case you quoted. The piston displacement shouldn’t be anywhere near the maximum.

  301. David says:

    @Andrew. Thanks but you say, “For example, the pressure at the pump outlet is zero if the pump is not connected to a load.”

    I guess we differ on that, though there may be a misunderstanding. If the load is zero, the pressure at the outlet will be the normal 3000 psi. The pump pressure compensator maintains the pump delivery at 3000 psi, under no load, by essentially zeroing piston displacement, as it does when its outlet is blocked when turned off.

  302. Andrew says:

    @David

    This somewhat amateurish diagram shows what I mean by ‘not connected to a load’. The pump is simply discharging fluid ‘overboard’. I respectfully suggest the outlet pressure in this case is zero. The pressure would rise when a load is added to the system and the fluid needs to do work to overcome the resistance of the load.

  303. Andrew says:

    @DrB

    RE: “Do you know how accurately a True Track can be held using the MCP?”

    Sorry, I don’t have any information that might help.

  304. David says:

    @Andrew. I agree the outlet pressure would be zero in that case where the flow greatly exceeds the pump capacity. I would assert however that it is under no load. It is the same as a burst pipe, no load.

    Never mind. Matters of greater import to resolve.

    We are agreed that windmilling could hold the pressure at the PCUs I think.

  305. Andrew says:

    @David

    RE: “We are agreed that windmilling could hold the pressure at the PCUs I think.”

    Yes!!

  306. Ge Rijn says:

    @David @Andrew

    From your info I then conclude there is not a time after second engine flame-out the left flaperon will go to its upward position.
    Not even when the APU runs out of fuel.
    As long as the engines keep windmilling the left flaperon stays in its neutral position.
    Is this a correct conclusion? Or will the PCU’s on left go in bypass as electrical power cuts off when the APU stops running?

  307. Andrew says:

    @Ge Rijn

    I believe the left flaperon would stay in the neutral position until both PCUs lost electrical power. The left flaperon’s inboard PCU would be powered by an FCDC battery, which would only last about one minute. The outboard PCU would be powered by the aircraft main battery. That would last quite a long time if the RAT were powering the standby power system.

  308. Andrew says:

    @Ge Rijn

    CORRECTION:

    The outboard PCU would receive electrical power from the RAT via the battery bus. It would not lose power unless the RAT stopped operating and the main battery went flat. In short, the left flaperon should remain in the neutral position.

  309. Ge Rijn says:

    @Andrew

    Thank you. Then I conclude the left flaperon could not have been a cause of a roll/banking to the left. Is this correct in your view?

  310. Mick Gilbert says:

    @Andrew

    Remember this graphic from this article by Victor in June last year? I know we’re talking a desktop sim but it seems to be illustrating an outcome that is broadly consistent with the training note in the AMM. Are we saying that the synoptic is definitely not an accurate representation of system and control status on RAT power only?

  311. Don Thompson says:

    @David, @Andrew, @Pilatus, and @HB

    The discussion seems to have clarified the PCU functions, and the behaviour to be expected from the L & R hydraulic systems: that’s progress.

    However, the condition described by the Training Information Point in AMM Chapter 29, 29-22-00, pg 28:

    The minimum airspeed of 115 knots is necessary for the RAT to supply rated capacity.
    When the RAT is extended and hydraulics off, the airplane rolls left. Two to three units of right control wheel rotation are necessary to hold the wings level.

    Previous discussions concerning the above discounted that the deployed RAT itself could invoke the left wing down roll, whereas this excerpt from the Training Manual, 27-11-00, Flaperon PCU, Functional Description:

    If both PCUs of the same flaperon are in bypass mode, the flaperon can move freely in both directions. In flight, the airloads then cause the flaperon to move up a maximum of 10 degrees.

    suggests that the LEFT flaperon is the source of the described roll.

    The two PCUs of the LEFT flaperon are supplied by the L hydraulic system (outbd PCU) and R hydraulic system (inbd PCU). The outbd PCU is associated with the L1 ACE, the inbd PCU with the R ACE.

    The fallback conditions of PCUs driving other PFCS control surfaces do not appear to initiate a roll.

    Left Flaperion: when the aircraft is operating with only RAT supplied electrical and hydraulic power the R ACE and, consequently, the inbd PCU loses electrical power as the FCDC battery quickly discharges: the inbd PCU reverts to bypass mode.

    The L1 ACE and outbd PCU should remain powered via the Hot Battery bus (or C1 TRU when RAT is generating). However the outbd PCU is supplied by the R hydraulic system which has no functioning pump but residual pressure can be expected to persist for some unpredictable period. It is possible for the L1 ACE to disable the outbd PCU by de-energising the bypass valve, however, no condition is specifically described to result in that outcome.

    I hope the above summarises the ‘dilemma’ presented by the two document excerpts. Do note I have described the conditions for the situation when only RAT power is available, so prior or subsequent to the APU start.

  312. Don Thompson says:

    @Ge Rijn: any conclusion you might make now would be premature. However, its obvious that has never stopped you.

    Surprisingly, I applaud the approach that has been adopted elsewhere in response to your online behaviour.

  313. Don Thompson says:

    To add to above where I wrote “suggests that the LEFT flaperon is the source of the described roll.

    The left flaperon is one of only two PFCS* surfaces, the other is the right elevator, where both PCUs rely on the L and R hydraulic systems.

    (* excluding spoilers, each of which are powered by only one hyd sys)

  314. Andrew says:

    @Mick Gilbert

    RE: ”I know we’re talking a desktop sim but it seems to be illustrating an outcome that is broadly consistent with the training note in the AMM. Are we saying that the synoptic is definitely not an accurate representation of system and control status on RAT power only?”

    Yes, the synoptic display is consistent with the Training Information Point. However, bear in mind that Victor selected the fuel cutoff switches to OFF to force the hydraulic systems to fail during the simulation. I don’t know if the PMDG model accurately replicates the system behaviour in that event. I hope to have some more information late this week. Stay tuned.

  315. Ge Rijn says:

    @Don Thompson

    The right elevator PCU’s would be in blocked/damped mode as concluded long time ago (not by me).
    Or is this now also free-floating IYO?

  316. Andrew says:

    @Don Thompson

    I think that paints an accurate picture of the ‘dilemma’, although I assume you meant to say “the outbd PCU is supplied by the L hydraulic system” in the second last paragraph.

    As I said to @Mick Gilbert, I hope to pin down some additional information late this week. I think a lot of it rests on what they actually mean by ‘hydraulics off’.

  317. Richard Godfrey says:

    A quote from elsewhere:

    “@Ge Rijn, It’s pretty clear that your motive is to stir shit, rile people up, and add nothing substantive to the conversation. I’m going to put your comments on manual approval from now on.”

    It is extremely rare for me to agree with this particular source.

  318. Victor Iannello says:

    @Andrew said: However, bear in mind that Victor selected the fuel cutoff switches to OFF to force the hydraulic systems to fail during the simulation.

    It also prevented the IDGs and Backup Generators from intermittently powering the main AC and transfer busses due to windmilling.

    So here’s a thought. Is it possible that a backup generator continued to supply power to the transfer busses until just before 00:19:29, and the A/P did not disengage until that time? For instance, after fuel exhaustion, if altitude was maintained by the A/P, the airspeed might have decreased until the backup generator dropped offline. With the A/P disengaged, a banked descent might have started at that time due to rudder out-of-trim.

  319. Tim says:

    @victor,
    I don’t believe a windmilling engine would power the IDGs or backup Gen. Supplying enough hydraulic pressure is one thing, but to keep a generator online needs a running engine.

  320. airlandseaman says:

    FTR…I’ve been told by multiple knowledgeable sources that a windmilling engine would not power the IDGs or backup Gen. RPM too low.

    I find the discussion about hydraulics etc. interesting. But there are so many unknowns, I don’t see how it gets us any closer to knowing exactly what caused the high rate of descent. What I do know is that every plane I have flown over the last 53 years will not fly straight ahead hands (and feet) off the controls. It does not work that way, no matter what the dihedral and “natural tendencies” are. A spiral descent is very, very common. We observed it in the simulator. So did Boeing. Sometimes it takes a few minutes to wrap up. But if there is any trim imbalance, or CG issue, it can start immediately.

  321. DennisW says:

    @ALSM

    I feel the same way you do, Mike. While the discussions are interesting, I don’t see them changing anything. One of my standard questions to an employee struggling with some problem was “what would you do differently if you knew the answer?”. Much of the time the response was “nothing”. I think that is the case here.

    The only facts we have are the BFO’s. You either trust them or you don’t. At this moment I am OK with the work being done. It is not what I would do, but it is reasonable.

  322. IR1907 says:

    @all

    Seems the first priority area designed by CSIRO turned empty.

    Very disappointing

  323. Victor Iannello says:

    @ALSM: I’ve been told by multiple knowledgeable sources that a windmilling engine would not power the IDGs or backup Gen. RPM too low.

    Yes, considering that the rotational speed of all spools at idle is much higher than the windmilling speed, that’s probably true. I am retracing steps along a path I abandoned long ago.

    I have to admit that pilot input does seem to be where we land when we try to explain the timing of the high downward acceleration.

  324. Ge Rijn says:

    @Richard Godfrey

    So you also still keep playing on the person instead of the content mentioned by me. You don’t grok me at all that’s clear. How hard I try to confince you I respect and acknowledge your work does not matter.
    I only challenged your drift-model which basically is based on drogued buoys which you stated yourself.
    Then you came up with the buoy that landed on Reunion in time but was caught up in an eddy for ~140 days. I had my questions about this.

    All I got was a request for getting me banned here from you.
    That would be the easy way out won’t it.

    The matter on JW’s blog has quite different reasons.
    Nothing to do with the argument we had.
    Kind of childish you referre to this IMO.
    I’m doing all in my own way to contribute in solving this drama.
    That’s all.

  325. Richard Godfrey says:

    @ Ge Rijn

    You have upset 20 commentators on this website!

    That you also do this on other websites is simply further proof of your goals.

  326. Victor Iannello says:

    @Ge Rijn: Please stop making unsubstantiated claims, thinly veiled with your “IMO” label.

    @Ge Rijn, @Don Thompson, @Richard Godfrey: We’re all tired of the bickering. Please stop it.

  327. sk999 says:

    DennisW,

    To follow up from yesterday, let me try again.

    24 microsec – 7.2 km light-travel distance, or 3.9 nm. If I move by a certain distance on Earth, only a portion is in the direction of the satellite. I divide distance moved by cos(elevation), where elevation of the satellite is about 39 degrees at the 7th arc. BTO gets a factor of 2 from the round trip, so I divide by 2. 2.5 nm. That now matches a direct calculation.

  328. airlandseaman says:

    Victor wrote: “I have to admit that pilot input does seem to be where we land when we try to explain the timing of the high downward acceleration.”

    I do not agree. I don’t think we can infer pilot control from that. A few years ago, we had a fatal crash here in Boulder involving a glider. The pilot was in the wave (high altitude) and apparently lost contentiousness due to an O2 equipment problem. The glider traveled uncontrolled about 10 miles east before the spiral started. Once started, it wound up very tight and went in almost vertically. I think this is the most common ending for any uncontrolled flight.

  329. Victor Iannello says:

    @ALSM: Nobody doubts that uncontrolled flight may end in a steep descent. What we are trying to do is explain the 0.65g acceleration that occurred at the time of the log-on of the AES. Other than the possibility of a roll that is induced by an engine re-start, I’ve yet to hear an explanation. If you can explain it, please do.

  330. Don Thompson says:

    @Andrew,

    Thank you for pointing out that error, yes

    The L1 ACE and outbd PCU should remain powered via the Hot Battery bus (or C1 TRU when RAT is generating). However the outbd PCU is supplied by the R L hydraulic system which has no functioning pump but residual pressure can be expected to persist for some unpredictable period.

    You wrote “However, bear in mind that Victor selected the fuel cutoff switches to OFF to force the hydraulic systems to fail during the simulation.” Why would moving the fuel cut-off switches to off provoke a different outcome for hydraulics than fuel exhaustion? Certainly, when the P5 Hydraulic/RAT Panel Primary L (or R) switch is selected to OFF the EDP Depressurize solenoid valve will be energised and the pump no longer provides any supply to the system. If fuel supply is shutoff in flight the engines will windmill and continue drive to the EDPs.

  331. DennisW says:

    @sk999

    Seems right to me. Short reply – iPhone. Headed to beach house.

  332. DrB says:

    @sk999,
    (and to @DennisW),

    You said: “If I move by a certain distance on Earth, only a portion is in the direction of the satellite. I divide distance moved by cos(elevation), where elevation of the satellite is about 39 degrees at the 7th arc. BTO gets a factor of 2 from the round trip, so I divide by 2. 2.5 nm. That now matches a direct calculation.”

    If one moves horizontally on the Earth’s surface, the range to the satellite changes by that distance TIMES the COSINE of the elevation angle. You don’t divide by the cosine. You can’t change the range to the satellite by more than the distance moved.

    The change in range for a 24 microsecond round-trip change in light travel time is 24 microseconds (round-trip) / 2 = 12 microseconds one way /3E+8 m/s = 3,600 m change in range at 39 degrees elevation. The horizontal shift is 3,600 m / COS(39 deg) = 4.62 km = 2.5 NM.

  333. David says:

    @Don Thompson. I agree with you that the connection between flaperon asymmmetry and the AMM statement is implied only. Having deduced there would be a flaperon asymmetry, independent of that statement, I contacted the ATSB early last year about that as I have mentioned, though without response. In my 24th January paper I raise other possibilities, which generally were discounted earlier as you say. I agree that they are tenous but at the same time point out that the ATSB has never confirmed the connection between the AMM result with the flaperon being its cause, to my knowledge.

    That flaperon influence is what Victor simulated by other means in June last year.

    Now we have the possibility that windmilling would scrub flaperon asymmetry entirely as a cause of bank; and indeed I raised this on 24th January as part of a study into whether the final descent was piloted or not. That scrubbing would leave the cause of the descent rates as a relight as Victor says, yet that was not included in the Boeing simulations….I will not go into that again. Apparently it was the cause of bank in the ALSM/United Airlines simulation though I think that, like the Boeing simulations, neither would have been representative.

    That said, while doubting simulation realism in replicating the final descent accurately in some respects I would have expected the effect of windmilling producing hydraulic power to be apparent all the same, raising some uncertainty still about whether our conclusion about that is valid.

    I did ask the ATSB about windmilling engines producing electrical and hydraulic power. They responded that the manufacturer had advised that there would be no electrical power generation, but made no comment about the hydraulics.

    I think Victor and ALSM would agree that the search width is unlikely to be affected by all this and so temporarily it should go into the too-hard basket. I have drawn a deduction from it as to the likelihood of the descent being piloted but whether more can be drawn I have my doubts.

    @Ge Rijn. Since in my estimation the APU would run for the entire short glide, the effect of its fuel exhaustion was not pertinent to my thesis. However with windmilling speed likely to increase in the descent and electrical power available for the left flaperon PCU which could use the hydraulics, there would be no flaperon asymmetry then either, as it stands.

  334. David says:

    @Andrew. The above said I think we can await your further explorations with interest.

  335. @Victor
    @ALSM who said “I’ve been told by multiple knowledgeable sources that a windmilling engine would not power the IDGs or backup Gen. RPM too low”.

    Yes, I agree. Considering the numerous stages of compressors (HPCs and IPCs) building a pressure ratio of ~700, the interia of the mechanical parts, the pumps and the IDG, if some electrical load is put on the IDG and/or Backup generator, the windmil of the bypass turbine will not suffice to sustain the sum of all these loads at this level of speed.

  336. Andrew says:

    @Victor

    RE: ”So here’s a thought. Is it possible that a backup generator continued to supply power to the transfer busses until just before 00:19:29, and the A/P did not disengage until that time?”

    I agree with the earlier comments on the matter; the windmilling RPM is too low for the IDGs or backup generators to operate.

  337. Andrew says:

    @Don Thompson

    RE: “Why would moving the fuel cut-off switches to off provoke a different outcome for hydraulics than fuel exhaustion?”

    It shouldn’t do in the aircraft; I was referring to the PMDG modelling. Victor mentioned his concerns that it does not model some aspects of the system behaviour correctly. I only meant that perhaps this was something else that is not properly modelled.

  338. Victor Iannello says:

    @Andrew: Yes, we’re all in agreement. What I meant to say above but inadvertently left out a couple of words was:

    “Yes, considering that the rotational speed of all spools at idle is much higher than the windmilling speed, that’s probably true. I am retracing steps along a path I abandoned long ago.”

  339. Andrew says:

    @Victor

    Thanks – That makes more sense!

  340. sk999 says:

    Bobby Ulich,

    The question of the constancy of a true track (I presume you mean the MCP mode) route is interesting, not discussed by the DSTG, and I have found no useful information regarding it on pilot forums, presumably because it is not a mode used for significant periods of time. Presumably the MCP uses the ADIRU, not FMC, to derive track angle. The LaserRef VI Micro IRS documentation gives the error in True Track angle as 2 deg 2-sigma. That is big enough to be interesting. What is not specified is the time-dependent behavior of this error.

    Suppose the track angle error varied slowly, say as 2 (deg) * cos(wt), where w = 2*pi/period, and the period were 5 hours. The 1-sigma error is 1 deg and the cross-track distance would vary with amplitude V*(2-deg)/w = 13 nm. The change in BTO depends a bit on latitude (assuming a path 180 deg South) but could be of order 75 microsec. That is large enough to be interesting. But it involves a lot of unverified assumptions.

  341. Rob says:

    @DennisW

    Re “The terminus is near 30S on the 7th arc. I am quite confident about that, and I have not stuck a pin in the map since Christmas Island (a very long time ago).
    I have weak analytics to support 30S, and Richard’s drift analytics are very good (IMO).
    Trying to reinforce my analytics with a couple of new geometric algebra books “written for physicists” has not worked out as I expected. They are not a quick read. A sign of age I suppose.”

    Dennis, you are the one person here I really enjoy conversing with. Something to do with wavelength perhaps. And possibly it’s your forgiving nature that endears. However, when it comes to the 7th arc and flight paths pertaining thereof, I see we are still miles apart. You are comfortable with S30, me much, much less so. You have bought into Richard’s drift studies – well, you have my sympathy at least.

    I’ve been giving the whole thing a lot of renewed thought recently. The search in full swing again has a powerfully concentrating effect on the mind. I don’t think I’m alone. I will let you know the results.

    If only things had started off differently, and Malaysia had given MIT the job of finding the plane, instead of those bungling but lovable cowboys in Canberra, the plane would have been found long since. Fantasy, of course.

  342. DrB says:

    @sk999,

    I think that the largest contributor to cross-track drift with time when using the MCP for lateral navigation might well be the inability of the AFDS to recover its track from disturbances (because it lacks a position sensor), even if the ADIRU were perfect. Certainly the ADIRU can also drift, and that feeds directly into angular error and therefore into cross-track error. In very quiet air, with no turbulence, the ADIRU errors might dominate, but I suspect that, in most cases, wind-caused variations will be larger. Your 1-degree 1-sigma number is larger than I expected (but this could be because it is for time scales longer than 1 hour).

    It is hard to see how the 1-sigma MCP navigation error over an hour, for example, could be as small as 0.1 degrees. I would guess at best it would be more like several times that.

  343. DennisW says:

    @Rob

    You are comfortable with S30, me much, much less so. You have bought into Richard’s drift studies – well, you have my sympathy at least.

    “Comfortable” might be a stretch. Richard’s analytics are very good, and I have a couple of other weak supporting arguments. If I were asked to pick the best spot to look, that would be it. Of course, no one has asked.

    Traveled to the beach house today to sit in the hot tub, stare at the ocean, and pretend I am waiting for a eureka experience.

    As far as MIT is concerned, I doubt they would rise to the challenge. It would take them about 10 minutes to conclude that the problem is under-constrained and intractable.

  344. David says:

    @DennisW.”As far as MIT is concerned, I doubt they would rise to the challenge. It would take them about 10 minutes to conclude that the problem is under-constrained and intractable.”

    Yet OI with its shareholders are putting their money on it.

  345. DennisW says:

    @David

    Yet OI with its shareholders are putting their money on it.

    There are a lot of shareholders in the world (me included). The SEC does not require an intelligence test to buy stock. Even with decades worth of analytical tools I occasionally pick a loser.

    Before OI the ATSB spent a ton of taxpayer money on what I have consistently (before the the underwater search was even started) said was a flawed exercise (in fact, I even said it was the best example of flawed analytics since anti-lock brakes). There is absolutely no way to know where the terminus is with the data we have.

    http://www.thetruthaboutcars.com/2006/08/killer-abs-abs-braking-increases-rollover-risk-by-51/

  346. David says:

    @DennisW. Thanks. Lets hope they luck it out.

  347. DrB says:

    @DennisW,

    OI is controlled by one person, a hedge fund manager in Britain, worth roughly $700M. Hedge fund managers take risks. Usually with other people’s money. This guy wants to be a billionaire, and he might make it.

  348. Rob says:

    @DennisW

    “As far as MIT is concerned, I doubt they would rise to the challenge. It would take them about 10 minutes to conclude that the problem is under-constrained and intractable.”

    Yes but don’t forget, MIT would also have me as an unpaid, unsolicited independent technical advisor as a given, and MIT are a clever bunch of cookies; They would recognize a free leg up when they see it. 😋

  349. Michael John says:

    @DrB

    That guy works for Marshall Wace. Is it proven that he is actually behind OI? Personally I think Oliver Plunkett got the job through his connection to Paul Marshall Via Prosperity UK LTD. Both are listed on Company House records & OIs London HQ shares the same address as Prosperity. Bottom line is that if Marshall Wace are underpinning OI we can be confident that OI won’t run out of cash mid search.

  350. Michael John says:

    OI has plenty of experts from Jake Klara downwards. Above him you have Melanie Smith & Oliver Plunkett. Now it seems to me that whoever is investing in OI has put MS & OI into the company. These are the only 2 British people in the company. Why British? My view is that the Investor themselves is most likely British & would seem that this investor has put OP & MS at the head as part of the deal. Jake Klara is from Lafayette & that is around 220 miles from Houston. Makes sense if he is the real guy who is behind OI. What wobbles my nerves about OI isn’t their expertise or credibility but the reliability of their investor. Without a contract this investor could simply withdraw it’s investment at any time. Some might say that is unlikely & maybe it is. But how much money is this investor likely to put at risk? Are they dedicated to completing the full promised search or could they pull out as soon as the higher probability areas have been searched? What are the investors standing to gain from the search for Mh370? What is their moral standing?

  351. Victor Iannello says:

    @Michael John, et al.: Recently, there has been a lot written recently about OI, and I know that most of it is wrong. The operational team and the financial backers of OI are contributing to solve this mystery, and as private citizens, they have the same rights to privacy as the contributors on this blog. A smear campaign has already begun in an attempt to discredit the capabilities and motivations of OI by those not interested in finding the MH370 wreckage in the SIO.

    I will not let this blog be used without cause to invade the privacy and level false accusations against private citizens.

  352. Michael John says:

    @VI

    Fair enough. Point taken. In my defence I was just expanding on a comment by Dr B. Although on reading my comment again I can see why it would be viewed as being negative. I have no qualms with you removing any comments I make if you feel that they are not viable.

  353. Victor Iannello says:

    @Michael John: Don’t be surprised if the smear campaign will attempt to link the political leanings of the acknowledged and alleged financial backers of OI to Russia, and by extrapolation, the MH370 search in the SIO also to Russia. That’s been the consistent pattern. I won’t allow such silliness here, especially since it involves besmirching the reputations of private citizens without cause.

  354. Michael John says:

    @VI

    I can assure you that my views do not involve Russia or even conspiracies.

  355. Ge Rijn says:

    @VictorI

    I won’t react anymore the way I did when obviously provoked. I’ll hold my breath to some people from now on. I also don’t want(ed) the level of discussion fall to this kind of devaluing level.

    The issue discussed though about the sequence of hydraulic/electric failure after second engine flame-out specifically regarding the mode of the flaperon after this time could be significant also to the current search.

    The search is in its first stages. When important new information shows up (like this maybe) there is still time enough to adjust the search accordingly.

    Maybe the search-width has to be taken wider at some point.
    Your june 2017 uncontrolled ~100Nm glide phugoid scenario would probably be an unrealistic option.
    But the pilot controlled steep descent and recovery afterwards would still be possible (I’ll avoid ‘IMO’ from now on for it’s all IMO I’m saying, so ask all to read it this way from now on).

    In the latter case I firmly believe a long glide would not have been an objective. In this case the objective would have been to descent as fast as possible at the right time, recover and take the shortest glide necessary to ditch the plane.

  356. Nederland says:

    @Ge Rijn

    Is there any reason to say any pilot (suicidal/non-suicidal) would want to a) descend as fast as possible only in order to b) recover and glide (and how, without engine power) and then c) ditch so poorly that even the vertical stabiliser gets fractured so badly it couldn’t get worse in a high speed impact? And if so, why?

    Would it not make more sense to search as far as possible towards the northwest when there are good fits both with the satellite data and the drift studies?

  357. Ge Rijn says:

    @Nederland

    Firts the best drift studies; Godfrey, Pattiarachi, Griffin point to an area between ~30S and ~35S. Anything north of 30S or south ~35S becomes increasingly problematic considering arrival times and locations of found debris.

    I argue a pilot who has the objective to ditch a plane wants to have the least amount of fuel onboard possible. This is the first thing any emergency ditching/landing pilot does; get rid of the fuel if you can to prevent fire or explosion on impact as much as possible.
    The reasons this PIC could have had could have been different; to avoid detection and a large debris field after impact as much as possible.

    After fuel exhaust at altitude (~35.000ft) the need to descent fast would be obvious. Any all engine failure would require a steep descent towards altitudes where normal breathing is possible again (~20.000ft).

    A PIC would not opt for a long glide in these circumstances. This makes no sense. Without engine power it’s completely possible to perform a ‘succesfull’ landing or ditch. AirTransit flight 236, the Gimly glider and the Flight 1549 are defining examples among others.

    The tail section would take the brunt of the impact in a ditch or tail landing like shown in several occasions.
    Asiana 214 showed quite similar damage to the vertical stabiliser as the MH370 piece. At exactly the same area. The rest of the plane stayed fairly intact and together though after cartwheeling across the runway with ~150M/h.
    This B777 planes can take a lot of impact without falling apart it seems.

    The ‘why’ is impossible to answer.

  358. Nederland says:

    @Ge Rijn

    There is at least one peer reviewed drift study, published in a scientific journal and taking account of a number of debris, including that from South Africa, suggesting the impact point may be as far north as 28S, possibly even further north. A 90 day search won’t go beyond 28S, in fact, searching a greater width will mean a more southern terminal point of the search.

    https://www.nat-hazards-earth-syst-sci.net/16/1623/2016/

    A quick comparison with other accident flights in which a failure of air conditioning was an issue clearly shows that it takes quite a long time until all oxygen is gone (unless there is rapid depressurisation). So, there is no need for a very steep descent because of that. At any rate, this would severely reduce chances of succesful ditching, if there is indeed any realistic chance of recovering from such a descent as observed, without engine power. If the intent was to avoid a debris field, then it would make no sense at all to first descend steeply and then to try and ditch (and there are also quite a few pieces from the cabin which are badly damaged).

    None of the cases you mention is an example for ditching after fuel exhaustion. But all those examples you mention clearly show that the correct procedure is to glide, not to descend after fuel exhaustion, whether the intention is to land or to ditch.

    Asiana 214 was not a ditching attempt either. As you correctly observe, the aircraft’s tail section hit a wall, and there is no surprise the vertical stabiliser suffered damage from this. This is not very good case of comparison.

  359. Nederland says:

    @Ge Rijn

    Also, none of the cases you mention is an example for successful recovery after steep descent and no engine power.

  360. David says:

    @Andrew. On 26th january, 07:16 AM HB wrote, “What is not clear is whether the PCUs reset to original control set point conditions or not after RAT power restablishment.”

    As a general point, do you think the ACEs have memory after electrical power loss? I am thinking specifically of right elevator PCU restoration by the APU following AC loss after MEFE, by L2 and R ACEs. Then also there could be force fight likewise between aileron and flaperon PCUs for the same reason if L2 and R have no memory.

    It looks reasonable to assume that they do. Do you know?

  361. Don Thompson says:

    @David,

    Hope you don’t mind me butting in!

    Power for each PFC, ACE* and associated PCUs is ‘held up’ by the Flight Controls DC battery supplying each Power Supply Assembly (L, C & R PSA), long enough for the RAT or APU to come up.

    [* the L2 ACE & associated PCUs are the exception, these are powered directly off the L 28V DC Bus without PSA or FCDC battery.]

    PCUs connected to ACE L2 will lose electrical power immediately the L 28V DC supply fails. If the APU does not come online, or does not remain online, the R ACE and associated PCUs lose electrical power.

    The right elevator ‘no DC power’ failsafe is blocking mode. Interesting thought, if it blocks when not in neutral position.

  362. David says:

    @Don. Thanks. I included the R because my diag (Davis) does not have back up battery for that either.

    Yes I appreciate the elevator would have gone into ‘blocking’. The question is what instruction it gets when it comes out, like other L2 (and R?) PCUs.

  363. DennisW says:

    @VictorI@all

    A long soak in the tub produced nothing. My conclusion is there is nothing left. I did reflect on the fact that we are standing on the shoulders of giants – Laplace, Fourier, Hamilton,…( a very long list) and that is a bit depressing since we have contributed nothing in the process of chasing dollars.

  364. Andrew says:

    @David

    RE: “As a general point, do you think the ACEs have memory after electrical power loss?”

    If there’s a fault on the PCU, the ACE failure monitor de-energises the solenoid(s) to put the PCU in the applicable ‘failed’ mode (ie bypass, blocking, blocking/damped, damped), depending on the control surface. However, if there’s a total power loss caused by a loss of power to the power supply assembly (PSA) or a PSA failure, as in this case, the solenoid(s) de-energise(s) simply because there’s no power, not because there’s a ‘fault’ in the PCU. If power is subsequently restored, the solenoid(s) is(are) re-energised and the PCUs are back to normal, assuming there are no other faults.

    That’s similar to what happens when the aircraft is on the ground and de-powered. All the PCU solenoids are de-energised when the power is removed and re-energised when power is subsequently restored. The PCUs are then ready to operate in normal mode as soon as hydraulic pressure is available.

    @Don Thompson

    RE: “The right elevator ‘no DC power’ failsafe is blocking mode. Interesting thought, if it blocks when not in neutral position.”

    Good question! When electrical power fails after the dual flame-out, the ROB* PCU initially goes to bypass mode, because its blocking solenoid remains powered by the R ACE. If the R ACE subsequently loses power then both the ROB & RIB PCUs immediately go to blocking mode. If the elevator is deflected at the time, then I assume it must remain deflected unless power is restored. That could make things ‘interesting’!

    *ROB = right outboard, RIB = right inboard

  365. Andrew says:

    @David

    Re my post above, the second sentence should read: If there’s a total power loss to the ACE and its associated PCUs, as in this case…. Forget I mentioned the PSAs; the L2 ACE obviously isn’t powered by any of them.

  366. David says:

    @Andrew. Thanks for both. I will leave it there.

    On Don Thompson’s ‘interesting thought’ the two blocking solenoids both are de-energised only when the elevator panel is ‘near to faired’, if needs be when air loads (for both or the last) have moved it to that position.

    That begs the question as to flutter in the interim and the source of the continuing solenoid power.

    @Don. The FCDC battery is good for a minute ‘during PSA switching’ so maybe that will keep the right ACE alive while awaiting APU start if ‘switching’ does not extend the minute for that.

  367. David says:

    @Andrew. I meant to quote a reference about the air loads; AMM 27-30-00 p48, supposing the 200ER & 300ER to be the same here.

  368. Andrew says:

    @David

    RE: “That begs the question as to flutter in the interim and the source of the continuing solenoid power.

    Exactly. If electrical power to the L2 and R ACEs is lost, the blocking solenoids in both PCUs would be de-energised immediately and the elevator would be locked. The elevator would only move back to the faired position if power were available to keep the solenoids energised long enough for that to occur. Perhaps there’s some capacitance built into the system?

  369. Shadynuk says:

    @Dennis RE: “There is absolutely no way to know where the terminus is with the data we have.”

    I think that is absolutely true. I continue to be perplexed that many people seem to ignore the indeterminate nature of this problem. They continue to pile on the assumptions to arrive at conclusions with no way to verify the assumptions. What is more puzzling is that it seems some people are willing to bet money on this.

  370. Andrew says:

    @David

    RE: “@Don. The FCDC battery is good for a minute ‘during PSA switching’ so maybe that will keep the right ACE alive while awaiting APU start if ‘switching’ does not extend the minute for that.”

    That would normally be the case, but what if the APU doesn’t start within the one minute capacity of the FCDC battery? The APU might hang, suffer some other fault, or the aircraft might have been dispatched with the APU unserviceable.

  371. HB says:

    @Don. Got the issue now thanks.
    @David , Andrew

    In any case the manual refers to loss of hydraulic pressure in my opinion which is the worst case failure condition.
    There is no signal from the lvdt/ACE to the HYDIM. Contrary to one of your earlier post, i think, the HYDIM is the master pressure controller triggering the start of additional pumps including the RAT. It says so in one b777 manual. The ldvt s only tell the recpective ace if the pcu worked or not. The hydraulic system does not care about the integrity of its users, it manage itself only.
    To go back to the question of roll or no roll, it all depends if the demand from the users are significant in the immidiately following the second engine failure to deplete the residual pressure and exceed the windmill flow condition. I think not but it can be calculated simulated by the manufacturer. The reason i think not is that the main consumers are hydraulically locked in fly mode and there is no main demand expected. I guess only from the rudder but i m not a pilot. In any case if roll occurs it will take time to reach this condition.

  372. HB says:

    I think there is no clarity of what it means by deenergising. Deenergising does not mean that power to the solenoids cannot be restablished. Typically, the power to the solenoids is given by the controller is all logical conditions are met. So long as the power is available to the ACE, If conditions are met in the ACE, the solenoids should renergise. ACE should be on battrey backup. Ie, needs to be comfirmed but if the pressure is reestablished it shoud auto-reset unless a permissive interlock prevents it. In this case the windmilling condition should autoreset the pcu interlock logic monitoring differential pressure in my view. The fails safe conditions for all users are carefully studied such that if there is a major failure, the degarded condition is not catastrophic. I am pretty sure this has been studied.

  373. Andrew says:

    @HB

    RE: ”Contrary to one of your earlier post, i think, the HYDIM is the master pressure controller triggering the start of additional pumps including the RAT.”

    I agree. Which post are you referring to?

  374. Andrew says:

    @HB

    RE: ”To go back to the question of roll or no roll, it all depends if the demand from the users are significant in the immidiately following the second engine failure to deplete the residual pressure and exceed the windmill flow condition. I think not but it can be calculated simulated by the manufacturer. The reason i think not is that the main consumers are hydraulically locked in fly mode and there is no main demand expected. I guess only from the rudder but i m not a pilot. ”

    The autopilot and TAC will disengage when the PFCS reverts to secondary mode following the second engine failure. There should be very little demand on the hydraulic systems, assuming there are no manual control inputs.

  375. Richard Godfrey says:

    @DennisW @Shadynuk Re: “There is absolutely no way to know where the terminus is with the data we have.”

    I agree.

    However, there is a lot we do know, for example:
    1. MH370 was between 4°S and 16°N at 19:41 UTC on the 2nd Arc.
    2. MH370 continued on a fairly constant flight path until 00:19 UTC.
    3. MH370 aligned with all the BTO and BFO satellite data between 19:41 UTC and 00:19 UTC.
    4. Aligned with the Boeing 777-200 ER aircraft performance data.
    5. Aligned with the Rolls Royce Trent 892 Engine fuel performance data.
    6. Wind Speeds and Directions en-route.
    7. Air Temperatures at Altitudes en-route.
    8. Location of debris finds in the Southern Indian Ocean.
    9. Not found in the original 120,000 sq. km. ATSB search area.

    What we do not know is, for example:
    1. Whether a Pilot was in Command for all the flight.
    2. Whether there was a system malfunction of any kind.
    3. The Location and Timing of the final major turn southwards.
    4. The selected Navigation mode.
    5. The selected Speed schedule.
    6. The selected Altitude and step climbs, if any.

    Without making any assumptions regarding what we do not know, it is possible to describe a set of solutions that match what we do know. It is also possible to assign a probability to each possible solution.

    Victor subjectively puts the chances of Ocean Infinity’s success in finding MH370 at around 67%. I have put it higher at 80%, Dennis even higher. The answer is neither 100%, nor 0%. Ocean Infinity have also made their own risk analysis.

  376. David says:

    @Andrew. Yes, there must be power from somewhere for that solenoid. I will continue looking, mainly from curiosity.

    Re the FCDC battery, Don Thompson had mentioned it but I had not responded about it.
    I was looking at both L2 and R being de-powered without an APU hold-up.

    Now distracted but will add to the above tomorrow if anything worth mentioning emerges.

  377. Rob says:

    @All
    I’m caught behind the curve as usual; the ATSB tell me the Malaysian DCA are running the OI search. That would make sense, and explain why Malaysian observers are on board SC

  378. Ge Rijn says:

    @David

    Maybe helpfull to your searching:

    https://www.google.nl/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0ahUKEwjt07uZ2f_YAhXJD8AKHZRGCAUQFggxMAA&url=http%3A%2F%2Fwww.davi.ws%2Favionics%2FTheAvionicsHandbook_Cap_11.pdf&usg=AOvVaw16ZNNQ7yux9MXStcHf5qQw

    Chapter 11.4.5 I found particularry interesting:

    “There are three individual power systems dedicated to the Primary Flight Control System, which are collectively referred to as the Flight Controls Direct Current (FCDC) power system. An FCDC Power Supply Assembly (PSA) powers each of the three power systems. Two dedicated Permanent Magnet
    Generators (PMG) on each engine generate AC power for the FCDC power system. Each PSA converts the PMG alternating current into 28 V DC for use by the electronic modules in the Primary Flight Control System. Alternative power sources for the PSAs include the airplane Ram Air Turbine (RAT), the 28-V DC main airplane busses, the airplane hot battery buss, and dedicated 5 Ah FCDC batteries. During flight, the PSAs draw power from the PMGs. For on-ground engines-off operation or for in-flight failures of the PMGs, the PSAs draw power from any available source.”

    Now what I know about Permanent Magnet Generators is they keep generating power as long as they are turning. Their magnets are permanent so independent on any external electrical sourse to make them supply power. They are also variable speed generators. So I assume they will generate considerable power also when an engine is windmilling.
    Enough to keep the PSA’s and ACES doing their job if all else fails to do it?
    I’ll leave this to you and other experts.
    Hope it’s helpfull.

  379. Victor Iannello says:

    @Shadynuk said: I continue to be perplexed that many people seem to ignore the indeterminate nature of this problem. They continue to pile on the assumptions to arrive at conclusions with no way to verify the assumptions. What is more puzzling is that it seems some people are willing to bet money on this.

    The problem is indeterminate in the sense that a precise solution is unobtainable with the data at hand. That doesn’t mean we can’t bound the problem and conduct a search within those limits. @Richard Godfrey made a nice summary of the key data we are using to do this.

    Most (but not all) here agree that searching to around 28S at +/- 25 NM gives better than even odds of finding the debris field. Assessing economic opportunities using probabilistic methods is what quants like the backers of OI do for a living. Let’s suppose it costs $150K per day to conduct the search. At 90 days, the financial obligation would be $13.5M. I have no idea what the numbers actually are, but with a potential payout of $90M, the risk/reward calculation could be quite favorable. There is no reason to assume an ulterior motive other than the potential for a favorable return on their investment. Altruism and personal interests could be other factors that might lower what an acceptable minimum return on investment might be.

  380. HB says:

    @David, According to below ref, solenoids are powered from the flight control DC power via the ACEs.
    para 3.1.2 and Figure 7 (https://citemaster.net/get/02baa57e-f4b3-11e3-b859-00163e009cc7/yeh98_777-fbw.pdf)

    @Andrew, I agree, I think the only condition to generate roll would be pilot input

    @All. i want to clarify which specific PCU hydraulic conditions will trigger a bypass. Triggering bypass on low pressure is unnecessary as the pcu will already be free to move. I think based on this drawing (Figure 7 of http://www.ijarest.com/papers/finished_papers/150511112042.pdf) bypass is only triggered on abnornally high differential pressure (such as caused by blocked return line) or in case of ACE electronic failure and not on low pressure. Ie, I suspect that in case of low pressure, the ACU/PCU would NOT switch to bypass. ie there is no condition to lead to PCU bypass for the flaperon for the case of both engine failure ???

  381. Ge Rijn says:

    @HB

    To shime in. If all you and others (like me) suggest is consistent, an uncontrolled phugoid glide after second engine flame-out as VictorI mentioned in his june 2017 article could be possible.
    But the final BFO’s then conflict this possibility. Which then only leave a pilot controlled fast descent and whatever glide after.

    I’ll keep advocating OI should not constrain themselfs on a +/- 25Nm width entirely.
    To go up north till ~28S within this limits I surely endorse but hey should search the section between 33S and 32S much wider between 95S and ~97.5S.

  382. TBill says:

    @DennisW @Shadynuk
    Re: “There is absolutely no way to know where the terminus is with the data we have.”

    @Richard
    However, there is a lot we do know, for example:
    1. MH370 was between 4°S and 16°N at 19:41 UTC on the 2nd Arc.
    2. MH370 continued on a fairly constant flight path until 00:19 UTC.

    I do not really agree that he path is unknown. To me the BTO/BFO strongly suggest the crash base case site is 180 South path to 32-35 South. Only if we say the BFO cannot be used as an accurate indicator is the path less certain. Bayesian analysis apparently was looking for a perfect fit to Arc5/6 based on a passive pilot, and thus missed 180S mark, possibly due to active pilot or else somehow the mathematical predicted passive track was wrong at the end (eg; high winds).

    Richard-
    (1) OK OK I see now…4 deg South I can buy that.
    (2) Fairly constant path…I agree and I have also adopted that assumption but Jean-Luc has not, and I reserve the right to consider non-constant path if necessary. But what I would say is, a relatively straight level path fits the data well (probably “perfectly” from at least Arc3 to Arc5) so that is an assumption we can make. I would say from Arc 5.5 to the end it would be hard to see maneuvers in the data.

  383. Ge Rijn says:

    .. between 95E and ~ 97,5E..

  384. Donald says:

    @All
    fa’
    \]4`

  385. Ge Rijn says:

    @Donald

    Please explain.

  386. Shadynuk says:

    @Ian @Richard @TBill Thank you for your comments. I understand (mostly) and agree (mostly). I think it is the ‘large assumptions’ that puzzle me the most: Did Z really just fly off following a failed negotiation to ‘hide’ the plane in the most difficult spot he could imagine. Why? Some have said, ‘to embarrass’ the Malaysian. Is it really possible to embarrass those guys? Are they duly embarrassed? It seems they are no worse off for this.

    If there was a negotiation, would Z really have given up all hope as early as 18:40ish and executed the infamous ‘FMT’? Such negotiation would have been difficult and volatile and perhaps he would have elected to remain within safe landing distance of a reasonably ‘safe haven’ as long as possible. As you all know well enough, such assumptions significantly affect a potential terminus and it seems the ‘hard data’ can still support a terminus as far north as 10S. Although some would say the debris locations rule that out. Perhaps. (But there I go – making assumptions to rationalize a terminus!)

    In any case, I hope your (and others) analysis will help OI to find the site even though finding the wreckage may not solve more than a small part of this great mystery.

  387. IR1907 says:

    @shadynuk,

    There was no negotiation. There was no communication between Ahmad Shah and the Malaysian government. I don’t know why people keep pinning hope on this speculation.

    I have been saying for years that Shah’s purpose was to simply vanish without a single trace and HE HAS BEEN SUCCESSFUL. The plane is STILL MISSING after nearly 4 years.The recovered debris is just a gift from mother nature. Doesn’t that tell you something ? Using common sense is necessary sometimes.

    Usually during an accident there will be some kind of detail left giving major clues to the location of the aircraft but in the case of MH370 we have absolutely nothing! I am flabbergasted that people do not see the obvious.

  388. Rob says:

    @IR1907

    “Usually during an accident there will be some kind of detail left giving major clues to the location of the aircraft but in the case of MH370 we have absolutely nothing! I am flabbergasted that people do not see the obvious”

    Well said IR1907. I’ve often thought the same thing.

    It seems as if there are two fundamental types in the population: those who think rationally, eg as would a homicide detective, and those who still secretly want to believe in Father Christmas.

    Problem is, the pro fantasy brigade tend to drown out the rational pragmatists, which is why they are still looking for the plane. Quit sad, really

  389. Shadynuk says:

    IR1907 “HE HAS BEEN SUCCESSFUL” ‘Successful’ at what? He is dead. The passengers are dead. The NOK are still grieving. There has been no published note of responsibility, no demands, nothing gained by Ahmad Shah. The Malaysians carry on as usual. It is more like he failed totally. Was he really that naïve?

    I had not mentioned an ‘accident’ – you did. I of course do not know that there was a negotiation (perhaps there was not) – but you cannot be certain that there was not. When people claim absolute certainly concerning something about which they cannot be certain – it is time to take note.

  390. IR1907 says:

    @Shadynuk,

    He has exposed Malaysian government incompetency and corruption to the whole world. Ask yourself, how much of a change you wanted for Malaysia Before March 2014? And how much you think or care about Malaysia now compared to then ?

    Ahmad Shah single handedly exposed the whole corrupt bunch in his country. If he didn’t take the souls of 238 other people with him i would call him a patriot or a hero perhaps.

  391. IR1907 says:

    @Rob,

    To be honest, i have nothing against people trying to explain a logical scenario for the behavior of the PIC of 9MMRO (like what Dr.Ulich is doing) but it should not have ulterior motives (cheerleading for the family of the culprit).

  392. Victor Iannello says:

    @IR1907: Let’s think about this. If the goal was to draw attention to Malaysian incompetency and corruption, why didn’t he leave a message or some other form of publicity? If getting attention was his only goal, he did it in a way that still has people guessing about whether a mechanical failure caused the disappearance.

  393. IR1907 says:

    @Victor,

    Shah was a very intelligent man. He knew what kind of impact a silent disappearance would have on the public opinion. The people would demand their government for more information about their missing airliner and this is where it gets interesting.

    Disappearance with a note would put the opposition under immense pressure and question their honesty.

  394. Shadynuk says:

    @IR1907 Perhaps that was his thinking. A ‘Pyrrhic Victory’ at best. Naive?? My vote still says we know little about motivation. I do note that it does seem like the Malaysians knew almost immediately what had happened and began their mitigating actions (propagating unverified information, delays in initiating SAR,..) They win – pretty much everyone else looses.

    Thanks.

  395. Rob says:

    IR1907

    Agree, and we all owe a debt of gratitude to Dr Bobby and his dedication. Don’t get me wrong, I am no sycophant, but his work is in a class apart.

    I will confess to suffering from paranoid thouhts at times. I regularly have to supress wicked and groundless thoughts of conspiracy. Conspiracy between the M********s and the A*********s to supress the facts. Silly, I know. But then again, the longer this circus keeps running….

  396. Rob says:

    @Victor

    “If getting attention was his only goal, he did it in a way that still has people guessing about whether a mechanical failure caused the disappearance.”

    Victor, if there is anyone out there still flirting with mechanical failure, or conspiracy or negotiation scenarios (eg blackmail) gone wrong, they need to seek impartial and sympathetic professional advice without delay.

    This man was a cruel, calculating psychopath, and must have been secretly planning his outrage for months, if not for years. If the scheme was going to have the intended, poisoning, effect, there had to be no obvious clues left behind, only subtle suggestions or indications that would feed suspicions, but would never permit closure.

  397. Michael John says:

    Dare I ask what kind of mechanical failure could explain Mh370 disappearing & does the ISAT Data not rule out this possibility? I still believe in at least 2 different events happening which could explain why no single event makes sense.

    Is it fair to point the finger at Shah without evidence to support those accusations?

  398. Nederland says:

    How can terrorist activity be ruled out? The French seem to assume just that.

  399. Victor Iannello says:

    @IR1907: Sorry, but I have trouble reconciling a pilot that wants to draw attention to a cause against the Malaysian government with somebody that creates a “silent disappearance” in a way that allows the government to control the narrative. I think we are missing important parts of the story.

  400. sk999 says:

    Regarding uncorrected errors in track mode (true or magnetic – shouldn’t matter), Bobby Ulich has pointed out that both ADIRU drift and imperfect response of the aircraft to follow rapidly changing wind conditions (even in the presence of no ADIRU drift) will contribute to cross-track errors. Some measure of the latter might be provided by the DSTG model for heading errors for LNAV routes. In that model, the rms heading error at any given time is 0.1 deg; integrated over 5.5 hours, the cross-track error might be of order 4 nm.

    A different approach is to asked what types of errors are needed to have an impact. A CMT route at speed 430 knots makes a better fit to the BFOs than a CTT route, but it has a worse fit to the BTOs – my best-fit route has an rms error of over 50 microsec. Suppose there were an uncorrected sinusoidal cross-track error, what amplitude, period, and phase would be required to improve the match? To test this possibility, I considered an error term with a fixed 2 degree amplitude (rms 1 degree) and found that the best parameters were a period of 4.2 hours and a zero-crossing time of 22.6 UT. The rms BTO error dropped to 32 microsec, close to the DSTG determined rms of 29 microsec.

    This sinusoidal error term achieves a maximum offset of 10 nm and maximum speed of 16 knots; one can think of this speed as being a mis-determination of the cross-track wind speed by this amount.

    This error seems large compared to what pilots report in the field for ADIRU drift (e.g., accumulated errors at the end of a flight) but are within the range of what is tolerated before ADIRU replacement is recommended. The end-of-flight is over 8 hours after the ADIRU was aligned in KL. It is also a bit large compared to the DSTG model of heading errors (although the rms is only 6.5 nm). Perhaps the combination of ADIRU drift and random heading errors could do it.

  401. Victor Iannello says:

    @Rob said: Victor, if there is anyone out there still flirting with mechanical failure, or conspiracy or negotiation scenarios (eg blackmail) gone wrong, they need to seek impartial and sympathetic professional advice without delay.

    I believe the contributors here are arguably the best informed of any group that is not part of the official investigation, and yet there is a large diversity of opinion within our group, independent of my own opinions. To be honest, you do worry me when I see you repetitively state your opinions and facts.

  402. Victor Iannello says:

    @sk999 said: Some measure of the latter might be provided by the DSTG model for heading errors for LNAV routes. In that model, the rms heading error at any given time is 0.1 deg; integrated over 5.5 hours, the cross-track error might be of order 4 nm.

    I lost you. In LNAV mode, the cross-track error is constantly measured and minimized by calculating the aircraft’s (inertially, GPS, and navaid-derived) position versus the target (lateral) path. Why would there be accumulated error due to heading error?

  403. Ulric says:

    @Victor: Yes we are missing important parts of the story. The mere fact that since the event there have been no demands, accusations or further releases of information by anyone with some sort of motivation for these events means that we are only guessing.

    There are two possibilities. Either there are people out there that know something and the information has been so effectively suppressed that nothing has leaked or, all the information there ever was is still in the wreckage. For me, both possibilities are terrible to contemplate.

  404. sk999 says:

    Victor,

    Lateral distance error accumulates in MCP track mode, not LNAV mode. I am assuming that the rms heading error is the same in the former as in the latter; it is the the target that is different (specific heading needed to achieve a particular track angle in the former; specific position at the end of a leg in the latter.)

  405. Paul Smithson says:

    @sk999. If you haven’t done so already, could I ask you to share your CMT route: start lat/long/time, initial heading T/M, speed profile, altitude and the results of this – lat/long at each of the ping times?

  406. TBill says:

    @IR1907
    You think a lot like me.

  407. Brian Anderson says:

    @Victor, SK999

    I suspect track errors accumulate in MCP track mode [both CTT and CMT] because there is not a clearly defined reference point somewhere further along the track. For CTT the reference is presumably a track angle compared with a geodesic. For CMT I have no idea what the reference could be. A constant track angle compared with what . . . leading to where? Those errors do not occur in LNAV because the desired end point is known, although there might be small deviations from a precise GC track to get there.

    CMH I can understand, because that would be a very normal and often used mode, for example when following ATC directions during an approach or a clearance from an airport or control zone. I can’t imagine when CTT or CMT particularly would be used, or even useful.

  408. Victor Iannello says:

    @sk999: Thanks. Now I understand your reasoning.

  409. Victor Iannello says:

    @Brian Anderson: We are all in agreement that errors don’t accumulate in LNAV mode as they do in other modes where an angle (be it track, heading, magnetic, or true) is the target.

  410. sk999 says:

    Brian Anderson,

    The ADIRU is always outputting both true and magnetic track (and heading) at all times while in flight. You dial in a number into the HDG/TRK window. Depending on the setting of the NORM/TRUE switch, the AFDS steers the plane so that the appropriate output of the ADIRU matches the number that was dialed in. No need for geodesics or reference points.

    R.E. lateral drift, I’m sure I did the calculation wrong, so think of it as just an order-of-magnitude number.

  411. DennisW says:

    @all

    I agree with Victor. The notion of a premeditated dissappearance with no hint of causality makes no sense at all. As Victor says, it allows Malaysia to control the narrative, and does nothing relative to the legacy of the event. Suicide has no basis whatsoever.

    Think back to the the Malay government response in the hours after the plane went missing. The response was a non-response, and controlled by people high up in government. Some say even the PM. This is not normal since there are protocols in place to define how the response should have been carried out and by whom. The normal search and rescue agencies were nowhere to be seen. Even ICAO was critical of it, and that is very unusual for that agency.

    The reason for the non-response is simple. People knew the circumstances of the diversion. Perhaps not the direction of travel. It was expected that Shah would land the aircraft somewhere, and appropriate action would be taken at that time. It would be foolish in those circumstances to initiate a search or an interception.

    Also keep in mind the known political leanings of Shah. He was solidly in support of the opposition party and his interactions with his soul mate, Tim Pardy, are well documented. Likewise it was well known in Malaysia that Najib and his buddy Jho Low embezzled hundreds of millions of dollars from government coffers. Jho Low’s antics are well documented relative to several celebrities – Paris Hilton, Miranda Kerr, Leonardo DiCaprio,… Jho Low remains missing to this day, and his whereabouts are the subject of internet speculation. Several governments including the US are investigating, and several arrests have been made. There is a ton of dirt out there.

    The simple explanation supported by all the circumstantial evidence is that Shah hijacked the plane. What is not known for certain is who the other players were on both sides of that equation, and what terms the hijack team was demanding.

  412. IR1907 says:

    @Dennis,

    I agree that Shah hijacked the plane. What i don’t agree with is some kind of sophisticated negotiation and demands.

    Let’s keep it simple and easy. All the evidence points to mass murder by Ahmad Shah. It was not really a typical suicide. It was mass murder.

  413. IR1907 says:

    @TBill,

    Please explain.

  414. DennisW says:

    @IR1907

    Why do you think Shah hijacked the plane? I still have the childish (not necessarily politically correct) belief that people do things for a reason.

  415. DennisW says:

    @IR1907

    An exception to doing things for a reason is the Amazon AI engine. I just one clicked another one of their suggestions. I had not an ounce of premeditation before receiving that email. I may block their emails in the future. I have no other solution I can think of.

  416. Andrew says:

    @Brian Anderson

    RE: “CMH I can understand, because that would be a very normal and often used mode, for example when following ATC directions during an approach or a clearance from an airport or control zone. I can’t imagine when CTT or CMT particularly would be used, or even useful.”

    The CTT/CTM modes are rarely used in day-to-day operations except for short-term track keeping. CMT might be used to steer the aircraft during an instrument approach such as a VOR, if the approach isn’t in the FMC navigation database. CTT might be used in similar circumstances in polar regions. The modes are useful if the aircraft’s navigation capability is degraded and LNAV is not available, eg dual FMC failure, but that’s a rare occurrence.

  417. sk999 says:

    Paul Smithson,

    I published the complete CMT route back in Jan 2016, albeit with more detail than you wanted. I’ve put up a much more compact version of the route – not identical, but pretty close – which you can fetch at the top link here:

    https://docs.google.com/document/d/14hleZyx1pUPL44yaeHKt6jnSQ3DbgRq2zibbKkFLq2c/edit?pref=2&pli=1

  418. Brian Anderson says:

    @Andrew,

    Thanks. Yes, I understand those possibilities. But how [and why] would MH370 get into those modes for a track into the SIO?

    Is it possible that the actual track turned out to be a CTT or CMT “look alike”, because of some fundamental failure of the ADIRU or IRS which made position data unreliable?

  419. David says:

    @Ge Rijn. Thanks for that. About the PMG windmilling possibility, I have raised that earlier and it was ruled out. The PMGs for this are in the backup generators and I believe they will drop off line at idle. It may be that their PSA AC-to-28V DC converters are not designed for the voltage or frequency below that. @Andrew might be able to confirm the idle drop out.

    BTW, a PMGs provides power to the three PSAs and their ACEs but the latter does not include the L2 ACE. The others are battery/RAT(centre only) supported but not the right (except by FCDC battery) or L2.

    @HB. “According to below ref, solenoids are powered from the flight control DC power via the ACEs.”

    The ref URL did not work on my contraption.

    Yes the solenoids are powered by the ACEs. Those of the elevators have two solenoids operating their blocking valves one on their by-pass valve. Those valves are activated when the solenoids’ power is selected off or the ACEs lose power. The question is what happens when ACE power is lost, the description of the blocking valves being that they will be activated when air loads have faired the elevator panel concerned in about its neutral position. Thence, what keeps the solenoids from activating while the panel moves to this, if the ACE concerned has lost power?.

    @Andrew has suggested that a capacitor might. The FCDC batteries fill the gap during power supply switching, nominally though for a minute only and the L2 ACE is not powered by one. Maybe the panel can be depended on to take up the faired position quickly. Maybe, maybe.

    Dunno as yet. However I think we can be confident that there is an answer since the design would not have passed muster otherwise.

    I am curious though as to how long it would take since on the face of it the elevator panel is unprotected from flutter whilst its actuators are not hydraulically locked.

    BTW you have alluded to lock being pressure dependent, ie, “Triggering bypass on low pressure is unnecessary as the PCU will already be free to move”. While there is a hydraulic supply of any pressure the actuator will be locked when in the required position. When not it will be locked by blocking, for surfaces vulnerable to flutter at high speed. Those such as the flaperons, invulnerable within the flutter boundary, are free to rotate if both PCUs are inoperative and therefore by-passed.

  420. HB says:

    @Ge Rijnn RE”… others (like me) suggest is consistent, an uncontrolled phugoid glide after second engine flame-out as VictorI mentioned in his june 2017 article could be possible.”

    Correct me if i am wrong, Victor’s article assumed a roll to occur to match the BFO data but the straight plugoid scenario did not match the BFO data and timing of events. All i can say is that the expected hydraulic system response does not seem to support a roll.

  421. sk999 says:

    Andrew,

    Just as a matter of interest.

    Have you ever been given the instruction: “Fly runway heading”?
    Have you ever been given the instruction: “Fly runway track”?

    Would CMH be of use in the first instance?
    Would CMT be of use in the second instance?

    Read lots of angst from others over what the instructions actually mean.

  422. Andrew says:

    @sk999

    RE: ‘Have you ever been given the instruction: “Fly runway heading”?
    Have you ever been given the instruction: “Fly runway track”?’

    Yes to both! The instruction to ‘maintain runway heading’ causes lots of angst because it has different meanings in different parts of the world. Some countries, such as the UK, have removed the phrase ‘runway heading’ from their lexicon because is does cause so much confusion. In most countries, ‘maintain runway heading’ implies that a drift correction should be applied to maintain the runway track. However, in the USA, Canada and Australia, ‘maintain runway heading’ means the runway heading should be maintained without drift correction.

    RE: ‘Would CMH be of use in the first instance?’

    Only in the USA, Canada or Australia (as far as I know!).

    RE: ‘Would CMT be of use in the second instance?’

    Yes, in all cases.

  423. Victor Iannello says:

    @Andrew: Is CMT that rare? When I look at SIDs, I often see “TRK” directives. For instance:

    WITT SID BEDA2B: RNW 35 HDG 350 UNTIL 2000 TRK 205 INTERCEPT RADIAL 266 TO FIX BEDAX FIX BEDAX

    WMKK SID PIBOSA: RNW 32R TRK 326 UNTIL 569 FIX OVERFLY VKL09 AT OR ABOVE 1200 TRK 050 INTERCEPT RADIAL 027 TO FIX PIBOS FIX PIBOS

  424. sk999 says:

    Andrew,

    Many thanks! It would seem that “heading” and “track” should be construed as “terms of art” as defined in law:

    “term of art – noun – a word or phrase that has a SPECIFIC or PRECISE MEANING within a given discipline or field and might have a different meaning in common usage'”

    Alas, it seems we are not there yet.

  425. Victor Iannello says:

    @HB said All i can say is that the expected hydraulic system response does not seem to support a roll.

    I’d say we have not yet identified a configuration with no pilot input that results in a roll that would reproduce the measured BFO values. Perhaps the roll was due to a failed re-start of the left engine, but there have been persuasive arguments that there would not be enough fuel.

    In the past, I discussed the possibility of a nose-down input to create the 0.65g downward acceleration. Another possibility is the control wheel was rotated, or aileron trim was applied, either which would produce a high speed graveyard spiral. The high loading on the right flaperon from the high speed could have damaged the trailing edge and led to separation.

  426. Andrew says:

    @Victor

    RE: ‘Is CMT that rare? When I look at SIDs, I often see “TRK” directives.’

    It’s certainly not uncommon to maintain a track as part of a SID, but the track would normally be flown in LNAV mode, not TRK SEL.

  427. TBill says:

    @IR1907
    I agree with most of your statements starting with “I am flabbergasted that people do not see the obvious.” At this point of course everything has to be qualified with “apparent most likely explanation.”

  428. sk999 says:

    Andrew says,

    “It’s certainly not uncommon to maintain a track as part of a SID, but the track would normally be flown in LNAV mode, not TRK SEL.”

    Wow, another thanks. That tiny detail has been bugging me for 2 years. Scratch it off the list.

  429. Andrew says:

    @sk999

    You’re welcome! That obviously assumes the SID is in the navigation database. It would be rare for that not to be the case.

  430. David says:

    @All. There has been interest in whether objects spotted from the air on 28/29th March 2014 were investigated by surface ship. I have posted an extract of an HMAS Supply’s Report of Proceedings from that time, which confirmed she had investigated various objects but without giving their positions.

    The messages under are a follow up. They add little though it is possible they will be of interest to some. I hope you can get the sense amongst the acronyms.

    https://www.dropbox.com/s/dy2rnrm04j6rwy0/28%2C%2029th%20March%202014%20surface%20search%20areas.pdf?dl=0

    Some points are:
    • 28th March (message sent morning of 29th local).
    The changed search area for the 29th is outlined.
    • 29th March(sent afternoon 29th). Mentioned possible lifejackets, in the NW corner of the search area and other objects the previous day which were in the SE corner but that is all we have on positions.
    The IL-76 mentioned is a Chinese aircraft. Of the 6 ships, 5 were for China. I leave FOI information seeking from there to others.
    From the ATSBs Operational search for MH370 of October 2017, 16 items were spotted on 29th joining the 52 of the 28th.

    AMSA update 30 of 30th March said, “A number of objects were retrieved by HMAS Success and Haixun 01 yesterday. The objects have been examined on the ships and are not believed to be related to MH370. The objects have been described as fishing equipment and other flotsam.”

    AMSA report 25 which briefly describes some aerial sightings is here:
    https://www.amsa.gov.au/news-community/news-and-media-releases/search-operation-malaysia-airlines-aircraft-update-25

  431. HB says:

    @David
    i was previously refering to the Flaperon PCUs. For the elevator the configuration is different. I think if you manage to open those links it will become all clear.

    RE: “The question is what happens when ACE power is lost” The solenoids will de-energise and valve will close. Note each block valve has two solenoids for fault tolerance (but i guess still the same power source from the same control loop). Note at same time, the PCV will close and the bypass will close. For reseting, just need to re-establish power to the ACE and all will go as normal.

    RE”While there is a hydraulic supply of any pressure the actuator will be locked when in the required position.”. For the flaperon PCU, it is not like that, there is only one bypass valve. If signal/power is lost it goes to bypass, it the signal is reestablished it goes back to normal position. It is basically autoreseting.

  432. Andrew says:

    @David

    RE: “About the PMG windmilling possibility, I have raised that earlier and it was ruled out. The PMGs for this are in the backup generators and I believe they will drop off line at idle. It may be that their PSA AC-to-28V DC converters are not designed for the voltage or frequency below that.”

    The PMGs in the backup generators are directly connected to power converters within the PSAs. I assume they would continue generating power at windmilling RPM, but I don’t know if that power would be within the tolerances required by the power converters.

    RE: “I am curious though as to how long it would take since on the face of it the elevator panel is unprotected from flutter whilst its actuators are not hydraulically locked.”

    In bypass mode, the elevator’s movement would be damped by a restrictor within the mode selector valve. Nevertheless, the air loads should force the elevator back to the neutral position very quickly, probably within a few seconds.

  433. DennisW says:

    @all

    A big moon tomorrow morning along with an eclipse. I will try to watch it.

    I always shake my head at the cosmology crowd who claim to know how the universe started yet they have found no dark matter in 50 years of looking, and can’t explain why the moon looks bigger on the horizon.

  434. Andrew says:

    @HB

    You seem to have missed part of the discussion. The elevator blocking mode was discussed earlier and @Don Thompson asked what happens if both elevator PCUs lose power while the elevator is deflected? If the PCUs lose power as a result of a power failure, then presumably the blocking solenoids would de-energise immediately and the elevator would be locked in position, unless there is some other facility that would allow the elevator to move back to the neutral position. I suggested that perhaps the blocking solenoids have some kind of capacitance built in that would keep them energised long enough for the elevator to move back to neutral in bypass mode. Any ideas?

  435. HB says:

    @Andrew,
    I did not miss that. I think the will block in any case even if in not the desired position.

  436. David says:

    @HB. “Note each block valve has two solenoids for fault tolerance (but i guess still the same power source from the same control loop)”.

    In the elevator there are 2 PCUs powered from different ACEs. Please take a look at the Davi reference Ge Rijn posted, which might help.
    The two blocking solenoids of both are powered from the two ACEs, one each. If a PCU or ACE fails that de-energises both its by-pass valve, its blocking solenoid and one of the other PCU’s blocking solenoids. The elevator then is controlled by the remaining good ACE and PCU, the failed PCU being in by-pass and both blocking solenoids remaining in the unblocked position, one solenoid being powered in each.

    If the other PCU fails, its ACE will de-power both remaining blocking solenoids when that elevator panel is close to neutral, faired. Both PCUs will then be blocked and in hydraulic lock, the elevator panel near neutral. If the second ACE fails the same will happen though where the power will come from in some circumstances to hold the blocking solenoids while awaiting elevator panel neutralisation is moot.

    I had surmised that if the elevator faired quickly and in all circumstances (even when the elevator is developing high lift), FCDC batteries would have enough stamina to power those solenoids. @Andrew has confirmed that the fairing should be quick. However should the outboard right elevator ACE, L2, lose power it has no FCDC battery back up. Also should any ACE fail but not from losing power, and its neighbour has failed already, the possibility of a blocked elevator not in neutral arises, at least in theory, which becomes @ Don Thompson’s ‘interesting question’.

    My, “While there is a hydraulic supply of any pressure the actuator will be locked when in the required position.” Your, “For the flaperon PCU, it is not like that, there is only one bypass valve.”
    That is true, I did not assert otherwise and how many does not matter. The PCU will not be in by-pass unless the ACE so orders. In previous discussion it has become apparent that it would do that because the PCU actuator is not following orders from its ACE: it is not in the position required. At that point the ACE will take power off the by-pass solenoid. Until that happens the PCU will be in hydraulic lock irrespective of supply pressure.

  437. David says:

    @Andrew. Missed yours to HB and some of mine superfluous is therefore. It took a time to write unfortunately.

    About the PMG windmilling possibility, advice to me was that it would not happen. I thought it was from you! Months ago. Obviously not. I have discovered nothing in the books about that. I mentioned a day or so ago that the ATSB had indicated the manufacturer, Boeing presumably, had said there would be no electrical power generated by windmilling. They may have included PMG, maybe not. However I note as above that the L2 ACE receives no PMG power anyway.

    “Nevertheless, the air loads should force the elevator back to the neutral position very quickly, probably within a few seconds.” As you have implied and I have detailed to HB, as it stands a few seconds, even a split second, will not be fast enough. I am unsure that in a pitching manoeuvre the delay might be just those seconds anyway. That elevator panel could end up locked while well off neutral. Still, really, that cannot be the way of it.

    @Ge Rijn, I take back my assertion about PMG generation. Do not know. The left and centre PSAs have battery backup so it would not matter there so much but the right engine windmilling could keep the right PSA going if it did generate enough and of the right specs. Intuitively though I would be mildly surprised if it did.

  438. Andrew says:

    @David

    The backup generator won’t produce power at windmilling RPM, but the PMG is a very simple generator that might produce some power. That said, I don’t know if the power would be usable by the power converter in the PSA. There’s a good chance it would not be usable.

  439. Ge Rijn says:

    @David

    I found this on the PMG’s of the B787:

    “The loss of Hydraulics and Electrics result in power from the advanced Permanent Magnetic Generators (PMG’s) which produce power even if the engine is wind milling. If the PMG’s fail, flight controls are powered by the 28-volt standby bus.”

    It seems to me this also counts for the B777 but cann’t find confirmation on this.

    It’s inder ‘flight controls’ at the following link:

    http://www.flight.org/boeing-787-a-pilots-perspective

  440. Ge Rijn says:

    @HB

    On your comment: “Correct me if i am wrong, Victor’s article assumed a roll to occur to match the BFO data but the straight plugoid scenario did not match the BFO data and timing of events.”

    I basically said the same in my comment you referred to (if you read the next sentence in that comment you did not copy/past). In @Victor’s june 2107 article a long phugoid glide was not considered only mentioned.
    A scenario like this could not fit the final BFO’s/high speed descent it seemed.
    That probably would only leave pilot-input if it could be proven (minimum required) hydraulic pressure would be maintained after second engine flame out and also electric supply by the PMG’s, batteries and RAT to the ACEs and solenoids.

  441. Don Thompson says:

    @Andrew, David, et al

    The PMG AC output is converted to 28V DC by a power converter within the FCDC Power Supply Assembly (PSA). We’re clear on that. A PSA has multiple sources: the corresponding DC bus (L, C, or R), the PMG power converter, its FCDC battery, hot battery bus (L & C only).

    The PSA’s controller includes an interface to the Flight Controls Buses, function as yet unconfirmed.

    The objective of my review is to drill down for a more accurate picture of the PFCS conditions during the RAT-APU-RAT cycle that initiates the end-of-flight phase. That process appears to have contradicted some of the assumptions previously made for the developing state of the PFCS control surfaces, and that’s a good outcome in itself. It remains that reports from flight simulator sessions describe ‘upset’ effects developing from that cycle of electrical system conditions, as do the BFO measurements at 00:19.

    I’ll refer to Nederland’s comments for the EOF controlled/not controlled debate, I’m not particularly interested in that (yet). @Ge Rijn: glad to see you’re onboard with some fact-finding.

    The component design of the PSA provides much redundancy, however, at a systems level predictability is an important goal. For example, can the PSA be relied upon for an extended period of engine windmilling in the fuel exhaustion scenario?

    Switching to the hydraulics and PFCS: on review of the 777 QRH, the procedure for a engine driven HYD system low pressure warning is to turn off the EDP on that system. That QRH action ensures a predictable outcome: that one system is definitely inoperable; the PFCS continues to function but is slower to react with less motive power to drive the control surfaces. I began my review to explore the effect of an pre-emptive OFF selection for the hydraulic demand pumps, its even feasible that a pre-emptive selection to OFF for the L & R hydraulic EDPs is possible. Other pre-emptive actions are assumed during the earlier phase of the flight, its quite possible that pre-emptive actions where made subsequent to the 18:25 Log On, or the FMT.

  442. Victor Iannello says:

    @Andrew said: It’s certainly not uncommon to maintain a track as part of a SID, but the track would normally be flown in LNAV mode, not TRK SEL.

    Of course. Constant heading and constant track legs can be part of a route and flown in LNAV mode. Thank you.

  443. Andrew says:

    @Ge Rijn

    RE: “It seems to me this also counts for the B777 but cann’t find confirmation on this.”

    Be careful; the B787 electrical system is very different to the B777. The article you cited says the PMG in the B787 is ‘advanced’. That might imply a different design to the B777.

  444. Victor Iannello says:

    @Ge Rijn: The backup generators are designed over a 2:1 speed range. While this is adequate for the operating speed range from idle to full speed (assuming the power takeoff for the gearbox is not on the low pressure spool), it would not be adequate for windmilling speed. From Designing The Boeing 777 Electrical Power System:

    A variable speed direct drive alternator was selected for this system to minimize the size of the engine installation. There is one backup alternator per engine directly driven from the engine gearbox at approximately 14,000 to 28,000 rpm. The output power of this alternator, at a frequency of 933 Hz to 1867 Hz, is converted to a constant 400 Hz by an electronic converter in the fuselage.

    The generator appears to have 8 poles to get this speed and frequency relationship.

  445. Ge Rijn says:

    @Andrew

    Yes I see. This might be misleading. The Trent 1000 engines might have PMG’s with higher output f.i. I still cann’t find conclusive details on this but still searching.
    I can hardly imagine the B777 would not have the same kind of windmilling-redundancy with its PMG’s to power the PSA’s and ACEs but this is no proof ofcourse.

    Found a detailed article which describes the system including the PMG’s but nothing about windmilling output:

    https://www.google.com/patents/US20070083301

  446. TBill says:

    @Victor
    “Of course. Constant heading and constant track legs can be part of a route and flown in LNAV mode.”

    I don’t think I know how to stay in LNAV…is there some way to enter a heading instead of a waypoint?

  447. Ge Rijn says:

    @VictorI

    Yes but the alternator you referre to is not a PMG.
    Following your reference from the article just after it states:

    “Also incorporated within this machine are two permanent magnet alternators which are used as isolated independent supplies for the electronic flight controls system.”

    It’s these altenators that supply exclusively the electronic flight controls during flight (and also during windmilling in a B787) and cannot be switched off by any means as I understand it.

  448. Ge Rijn says:

    I eventually found the output of the B777 PMG’s;

    777 Electrical System
    “Traditional” Hybrid – 115Vac & 28Vdc
    • Power Sources:
    – Two – 120 kVA, 115Vac, 400Hz engine driven generators
    – One 120 kVA, 115Vac, 400Hz Auxiliary Power Unit (APU) driven
    generator
    – Four 950 W Permanent Magnet Generators (PMG) integrated into the
    two backup generators
    – One 7.5kVA Ram Air Turbine (RAT)
    – Main, APU, and flight controls batteries

    So 950W per PMG. 1900W per engine.

    Link:

    https://www.google.nl/url?sa=t&rct=j&q=&esrc=s&source=web&cd=19&ved=0ahUKEwiPjr6TsILZAhUhLcAKHc5ACm44ChAWCFgwCA&url=https%3A%2F%2Fwww.ece.cmu.edu%2F~electriconf%2F2008%2FPDFs%2FKarimi.pdf&usg=AOvVaw3OShzWRunBIRZPV5-BZ2oV

  449. Victor Iannello says:

    @Ge Rijn: OK. But please help me to understand something: Irrespective of power from the PMGs during windmilling, is there any reason to expect that 28 VDC power might be interrupted?

  450. Don Thompson says:

    @Andrew, Victor

    The 777 engines each pack 6 PMGs: 1 for excitation on IDG, 1 for excitation of the backup generator, 2 for the FCDC PSAs, 2 for the engine FADECs.

    In the absence of manufacturer specification, reverting to PPRUNE (anecdotal, accepted), the Trent with N3 power take-off for the accessory gearbox needs 8% RPM to maintain adequate PMG output.

  451. Victor Iannello says:

    @Don Thompson: Are you assuming the PMGs on the B744 behave like those on a B772?

  452. Rob says:

    @All

    A few thoughts….

    If you want to catch a crook, it’s good if you can think like a crook. Ever wondered why there are so many bent cops around.

    The way to make progress is to put yourself in his shoes and imagine what he would do.

    How the pilot covered his tracks:

    As the aircraft passed beyond radar range at 1822, it was following air route N571. About two minutes later, the pilot initiated a 15nm right offset manoeuver from N571. The timing of the manoeuver is particularly significant because it suggests the pilot wanted to ensure that the manoeuver couldn’t be tracked by the primary radar on Penang Island. This offset manoeuver, combined with an FMT timed to commence 25nm before reaching waypoint IGOGU, an FMT that would put the aircraft on a pre-chosen course toward a waypoint known only to the pilot himself, would make any attempts at subsequent flight path reconstruction extremely difficult, because it would be virtually impossible to associate the flight path with any known navigation waypoints. The pilot had decided in advance, to use geographic coordinates S41.00 E88.00 as a manual waypoint.

    The pilot seems to have planned for a particular sun angle at fuel exhaustion, and had worked out in advance that if the aircraft flew in a straight line towards S41.00, E88.00, the fuel would run out shortly after local sunrise – the Sun was about 6 degrees above the horizon at fuel exhaustion. Coordinates S41.00 E88.00 would be easy to remember and easy to enter into the FMC as a manual waypoint and importantly, would be just beyond the maximum range of the aircraft. In the event, the fuel ran out about 209nm before reaching the waypoint.

    How the pilot synchronized fuel exhaustion with sun angle:
    It’s no easy task. Advance planning is essential. If you time it wrong, you can either end up flying the final hour or so in full daylight and risk being eyeballed from below, or you can run out of fuel while it’s still pitch dark. Neither outcome enviting to a pilot bent on covering his tracks. Only by using a manual waypoint do you have the degree of flexibility needed for planning the flight path. Firstly, you need to estimate both the time of fuel exhaustion and the approximate distance the aircraft could be expected to cover from FMT onwards. You then have to identify the area in the Southern Indian Ocean where the Sun will be at the desired elevation as the fuel runs out, and then finally you need to select a suitable latitude/longitude manual waypoint that is located safely beyond the estimated fuel range, a set of easily remembered coordinates that will direct the flight path toward the area you have duly identified.
    On March 7th, the southward flight path ran almost exactly parallel to the advancing sunrise line, both were angled at about 7 degrees west of the due south line, permitting the aircraft to fly under cover of darkness until about 30 minutes before fuel exhaustion. A flight path that runs roughly parallel to the advancing sunrise terminator line makes it much easier to synchronize fuel exhaustion with sun angle.

    If your desire is to maximize distance reached into the SIO by flying in a straight line after FMT, then there are only a couple of periods in the year when the lighting conditions would be considered acceptable for the MH370 flight. Early October to late November is one period, late January to Mid March is the other. In southern summer the prospective flight path would terminate uncomfortably close to New Amsterdam Island, while in southern winter it would terminate uncomfortably close to the west coast of Australia. If you leave it till late March, the sunrise line will by then be oriented due north to south and consequently, the FMT would need to be later, over the Andaman Islands. If you let it go beyond late November, the sunrise line would then be nearly 20deg to the vertical (and increasing) which would push the FMT back into the Malacca strait, and the aircraft would have to overfly the Banda Aceh peninsula.

    How the pilot made good his getaway:

    At 1829 the aircraft location was N7.17 E95.702, Airspeed M0.84 at FL350, and it was at about this point that the pilot a gradual climb from FL350 to FL400 and allowed the airspeed to bleed off from M0.84 to M0.82. The slowdown to M0.82 was completed by about 1835, the climb to final cruising altitude FL400 was completed by about 1844. The aircraft then maintained a constant Mach 0.82 cruise at FL400 until MEFE.

    The aircraft continued on the 15nm offset from N571 until 1835.50, at which time the FMT toward manual waypoint S41.00 E88.00 was initiated. Groundspeed was 486Kts. The FMT was initiated when the aircraft had 25nm to go before arrival at a position that would have been directly abeam of waypoint IGOGU. The FMT was begun at position N07.565, E94.902 and completed 2.65 minutes later at N07.417, E94.631, final bearing 186.692447 deg. The aircraft turned through 108 deg, at a rate of 0.68 deg/sec.

    At 1840 the climb to FL400 was in its final stages. The 1940 BFOs permit a climb rate of 300fpm at this point.

    The total distance covered between 1829 and MEFE timed at 0017.50 is 2789nm. By comparison, if the aircraft had flown from 1829 to MEFE entirely at FL400, at Mach 0.82, fuel exhaustion would have occurred one minute later at 0018.50* and the aircraft would have travelled 2795nm. (ref Dr Bobby Ulich’s fuel model) This is a particularly significant finding – it reveals that the endurance/range actually achieved by 9M-MRO between 1829 and 0017.50 is within 1 minute and 6nm of the estimated theoretical maximum distance that 9M-MRO could have covered in the specific atmospheric conditions on the night of 7th March 2014, when flown at optimum airspeed and altitude. Dr Bobby’s fuel model applies specifically to 9M-MRO and for the atmospheric conditions pertaining on the night of the disappearance. The fuel model actually predicts a MEFE time of 0019 for M0.82 at FL400. The difference is possibly due to a slight difference in the assumed Mach 1 values at FL400, making Dr Bobby’s M0.82 fractionally slower than mine.

    So that’s how he did it. Why not close in on him now:

    The flight path described here possesses the three important characteristics: 1) Distance flown is very close to the maximum distance that could be achieved on the night in question. 2) Aircraft flew on a great circle path toward a manual waypoint with coordinates S41.00 E88.00. 3) Estimated time of fuel exhaustion appears to have been carefully synchronized with sun angle. To have just one of these characteristics alone would not necessarily confer special status, but a path having all three….

    This flight path both determines and is defined by the ping arcs. Consequently all other potential flight paths crossing the 7th arc, such as at those at latitude S35 or S30, are constrained by the flight path described here, although none of them have the three special properties the S38 path has. But why should the other candidate flight paths have to share the same ping arcs the S38 path? Surely, all other candidate paths would be free to generate their own individual arcs? Of course they would. However, the unavoidable fact is they never had the opportunity to generate their own arcs, because they never existed.

    At fuel exhaustion, the aircraft had reached S37.61 E88.66. A 16:1 glide starting from 38,000ft altitude (RH engine having flamed out 7 minutes previously) would terminate at about S39.20, E88.36. when allowing for estimated average headwind during the 20/25 minute descent, the still air range being shortened by about 7Kts. Pilot would want to stay on the same bearing and wring out as much mileage as he could before finally calling it a day.

    A convenient way of defining the necessary search area would be to trace out a circle of 20nm radius, centred on S39.20 E88.36. Total area is 4,300 sq km, and would be 99% certain of containing the wreckage.

  453. DennisW says:

    @Rob

    Your story has a significant negative attribute. It is wrong. It simply cannot explain the debris findings and lack of findings. You will get nowhere here by ignoring the debris.

    Your narrative also lacks motive or causality. Resting your case on Shah being a mad man is simply not compatible with his behavior prior to the flight.

  454. Ge Rijn says:

    @VictorI

    I referre to my earlier posted link:

    “Two dedicated Permanent Magnet Generators (PMG) on each engine generate AC power for the FCDC power system. Each PSA converts the PMG alternating current into 28 V DC for use by the electronic modules in the Primary Flight Control System.

    Alternative power sources for the PSAs include the airplane Ram Air Turbine (RAT), the 28-V DC main airplane busses, the airplane hot battery buss, and dedicated 5 Ah FCDC batteries. During flight, the PSAs draw power from the PMGs.”

    For on-ground engines-off operation or for in-flight failures of the PMGs, the PSAs draw power from any available source.”

    From this it seems to me there will be no interuption in power supply to the PSAs, ACEs and PCU-solanoids as long as the engines are windmilling at sufficient RPM. Back-up will be the RAT, hot battery, 5Ah FCDC batteries and the APU if it comes on line.

    The APU must have come online after second engine flame-out I think for only running on PMGs, RAT and batteries would have load-shedded the SATCOM and the final log-on would not have taken place.

  455. Ge Rijn says:

    @Don Thompson

    Thanks for your included comment to me.

  456. lkr says:

    JW reports [correctly!] the Malaysian Official Site has update on the search, showing no results for the first area scanned. Also, a relief to some here, they show an extension of the official search area to approximately 29N full-width.

  457. Lauren H. says:

    @DennisW – Its a Super (meaning closer to the earth) Blue (second full moon in one month) and just for you on the US West coast a “Blood” (meaning its red color coming from the sunlight being eclipsed by the Earth) Moon and don’t forget to try bending over and looking at it between your legs.

  458. Victor Iannello says:

    @lkr: I don’t see anything in that report that is newsworthy, and I wouldn’t use that graphic image as a reference. We already knew the search would be extended past the 25,000 sq km from the details about the payments terms. We also have been saying here that the extended area would likely be centered on the “13790 μs – 20,000 ft” 7th arc at +/- 25 NM. I think that is more likely than the boundaries of the extended area shown in the recent report.

  459. Victor Iannello says:

    @Rob said: A convenient way of defining the necessary search area would be to trace out a circle of 20nm radius, centred on S39.20 E88.36. Total area is 4,300 sq km, and would be 99% certain of containing the wreckage.

    Future comments of this length and on this subject matter need to be posted by reference to a link. You are using way too much bandwidth on multiple renditions of a theory that has no traction here. With the sparse evidence at hand, anybody that is 99% sure of their solution is not being objective, to say the least.

  460. Paul Smithson says:

    @Richard Godfrey. With reference to you earlier post
    “However, there is a lot we do know, for example:
    1. MH370 was between 4°S and 16°N at 19:41 UTC on the 2nd Arc.”

    How is it that we can be so confident of the latitude at 1941? What’s the basis for this “known known”? Time & distance from LKP? BFO confidence interval? Other?

  461. Don Thompson says:

    @Victor

    Am I “assuming the PMGs on the B744 behave like those on a B772?

    Not exactly, in considering reply I was thinking about idle speed: the “RB211 Trent 892” N3 100% rating is certified as 10,611rpm (sea level) while the RB211-524 family N3 rating is certified as 10,610rpm. Power take off for the accessory gearbox is via the N3 shaft.

    The PMGs may perform differently, but each of these engine types employs PMGs for its FADEC.

    Another PPRUNE sourced quote, from a Trent MM:

    Trent 500, the Power Supplies
    (1) The primary source of electrical power for the FADEC system is an EEC dedicated alternator (4044KS). Three-phase power from the alternator is regulated (internally in the EEC) and supplied for each EEC Channel.
    (2) The alternator can supply sufficient power to the FADEC system to control the engine (and windmill starting with the hydraulic pump de-energized) at speeds higher than 5 percent N3.

    N3 speed of 5-8% required from windmilling to keep PMG output up and the FCDC PSA circuits powered.

    Irrespective of the status of the FCDC PSAs, only the C1 and C2 TRUs are powered from the RAT to provide 28V DC power. The L and R 28V DC buses are lost on RAT only power generation.

    IF the PMGs hold up on windmilling, the electrical status for the PFCS remains reasonably intact, certainly over the few minutes we’re concerned for in this case. Only the PCUs connected to the L2 ACE lose all DC power. Among those PCUs, however, is the outbd R elevator PCU: it goes to blocking mode and its actuator can no longer move.

    Returning to the expected, or predictable, behaviour: as the aircraft speed falls, it would be expected that horiz stab trim input follows from the flight crew to set a descent. If no stab trim commands are made??

  462. Paul Smithson says:

    @sk999. Thanks. I think I recall trying to open the previous version and failing to do so because I couldn’t deal with a file extension. I look forward to having a look at your summary version.

    @all. I am intrigued that the electrical situation at second flame out is starting to (possibly) look more nuanced. My preferred solution may become more plausible as a result – namely controlled flight beyond 7th ping ring. This would, of course, require the BFOs at 0019 to be anomalous – for which I also don’t have an explanation.

    I cling to the remote possibility that the power cycle on the SDU (final reboot) might be related to a brief hiccup in phase-matching a power switchover to alternative source [or something of the sort]. And that this might – even more sketchy speculation – have been triggered by the first engine flameout rather than the second if left side power had been deliberately isolated. And the combination of those circumstances might make a fairly lengthy, straight-ish, controlled-ish end of flight possible.

    This in turn would allow an early FMT solution to still be correct – with a terminus well beyond the searched area at 38.5S

  463. Victor Iannello says:

    @Paul Smithson: The final BFO values and the drift model results are significant hurdles for this scenario.

  464. HB says:

    sorry if i miss some previous discussions/points as i have little spare time.

    @Don: RE “The PMG AC output is converted to 28V DC by a power converter within the FCDC Power Supply Assembly (PSA). We’re clear on that. A PSA has multiple sources: the corresponding DC bus (L, C, or R), the PMG power converter, its FCDC battery, hot battery bus (L & C only). ”
    to add to that it is stated in section 11.4.5 and Figure 11.3(http://www.davi.ws/avionics/TheAvionicsHandbook_Cap_11.pdf) that in case of both engines off, the PSA draw power from any available source including Ram Air Turbine (RAT), the 28-VDC main airplane busses, the airplane hot battery buss, and dedicated 5 Ah FCDC batteries (for L and C).

    @David: I agree that that failure mode exists. It cannot be avoided. It is either that or running the risk to experience flutter. Worst case, the ultimate degradation is you have the elevator on the other side (although not ideal but this is called graceful degradation).

    For the case of MH370, it will be odd to have another 2 or 3 independent failures to the already unlikely sequence of events.

  465. Don Thompson says:

    @HB: concerning the supply to PSA, thank you, I had all options covered.

    @Paul S: intrigue and nuance. However the many flight simulator trials demonstrated that the aircraft has a propensity to bank in the absence of any controlling input.

  466. Rob says:

    @DennisW

    The negative aspect can be attributed to the negative act, no more no less.

    I’ve discussed the debris many times. You now my conclusions on that subject.

    OI’s financial backers have deep pockets. They will need deep pockets.

  467. David says:

    @Don Thompson. Thanks for the info on the PMGs likely generating at low revs. In the short term the only difference this makes at fuel exhaustion is the right ACE on line. In the longer term saving battery drainage could become an issue, though the RAT should see to that.

    I am unsure where this continuing discussion is leading.

    About your, “Among those PCUs, however, is the outbd R elevator PCU: it goes to blocking mode and its actuator can no longer move”, I do not think so Don, not with the right elevator inboard PCU still operative, its ACE powered by a right back-up generator PMG in that scenario.

    It will go to blocking mode only when both solenoids on its blocking valve are de-powered.

    However what I believe will happen as I outlined to HB is that on loss of the L2 ACE only one of its blocking valve solenoids will be de-powered, together with its by-pass valve. Thence it will go into by-pass, the inboard PCU being powered still. At the same time its ACE will remove power to one of the inboard PCU’s blocking valve solenoids, to which it is connected. Only should the inboard’s right ACE or its PCU fail will the power to the one remaining active solenoid on both PCUs blocking valves be cut, the right ACE providing power to its remaining blocking valve solenoid and likelwise being connected to the one remaining in the outboard’s. The PCUs then will go to blocking mode simultaneously, though as we have discussed, not necessarily immediately.

  468. DennisW says:

    @Rob

    I’ve discussed the debris many times. You now my conclusions on that subject.

    Actually you have never discussed the debris. You simply reject all the drift analytics because they do not support your conclusions. That is not at all professional. If you reject the debris analytics you should have a reason besides their negating your point of view.

    I really try to be reasonable with cogent arguments (many of them not provable), but I would never reject the analytics of other qualified professionals without an extremely good reason.

  469. Andrew says:

    @Don Thompson
    @David

    RE: “Only the PCUs connected to the L2 ACE lose all DC power. Among those PCUs, however, is the outbd R elevator PCU: it goes to blocking mode and its actuator can no longer move.”

    I agree with David’s comment. If the R ACE is still powered, it will keep one coil of the ROB PCU blocking solenoid energised. The ROB PCU bypass solenoid will be de-energised and the PCU will be in bypass mode. The R elevator will then be powered by the RIB PCU in normal mode (assuming no other failures). If the RIB PCU subsequently fails, the blocking solenoids in both PCUs will be de-energised and the elevator will be locked in position. That might occur before the elevator returns to neutral.

    @HB

    RE: “Worst case, the ultimate degradation is you have the elevator on the other side (although not ideal but this is called graceful degradation).”

    That ‘graceful degradation’ could mean the elevator authority is very much reduced, depending on the position of the failed elevator. That’s not so ‘graceful’ from a pilot’s perspective!

  470. ventus45 says:

    @Rob
    Re your post January 31, 2018 at 11:53 am

    Your reasoning parallels mine from long ago, written up over many posts, on Auntypru Forum, in the thread titled:
    “MH370 – time to think of it as a criminal act”.
    http://www.auntypru.com/forum/showthread.php?tid=28
    Send me an e-mail

  471. DennisW says:

    @Ventus

    Your reasoning parallels mine from long ago

    Funny shit. Referring to Rob and reasoning in the same post is incomprehensible to me.

  472. TBill says:

    @Victor @DrB
    Looking at the SLOP/CP:
    Problem I have is PSS777 apparently does not have the SLOP/CP function.

    But an expedient SLOP substitute, or proposed alternate path, is assuming (after MEKAR) the pilot keyed in 0796E 0894E BEDAX (or ISBIX).

    I’ll have to key in a Ft/Min descent at 1840, but if the A/P knew I wanted to go down to FL270 at 0894E, I suppose the automated descent timing/turn could be close.

  473. DennisW says:

    @TBill

    Does it really matter? The OI search strategy is well-defined. I am sure of that. The discussions here are largely mute. I am content to sit back and watch what happens. I am done with the analytics, and Ami is happy about that. Give it a rest until we see what is “discovered”.

  474. TBill says:

    @DennisW
    We have two recent papers DrB and Jean-Luc so it implies some front-end changes to flight paths.

    It was not too long go Victor doubted the 18:22 radar point. Now ATSB indicates that was solid radar, and the SLOP/CP is back in style.

  475. HB says:

    @Adrew, RE:”That ‘graceful degradation’ could mean the elevator authority is very much reduced, depending on the position of the failed elevator. That’s not so ‘graceful’ from a pilot’s perspective!”

    I agree. Note the same applies to the ailerons having a hydraulic lock. Introducing this lock inherently introduces this failure mode which could be critical if it occurs during the time at risk where the elevator is in the undesirable position. There is no feature to delay the block valve action for a number of scenarios. The rest is a matter of whether the failure probability is acceptable or not.

    Not sure whether all this discussion is leading. But in the Case of both engine failures, i don’t think that this issue will happen.

  476. Nick says:

    Constructor has been ‘dark’ now on AIS for several hours. Could mean something, could mean absolutely nothing.

  477. David says:

    @Andrew, Don Thompson. I think the answer to an elevator panel being exposed to risk of flutter while it awaits air loads getting it close to neutral is that in so called by-pass mode the PCU is in fact damped. The TM and AMM make no mention of this though the Davi manual does indicate both rudder and elevator PCUs can be damped, under Fig 11.3.

    Rudder PCUs have a ‘damped’ solenoid valve, very similar in function to the ‘blocking’ in the elevator though the mode selector valve is different. The elevator PCUs have no damped solenoid.

    However whereas the flaperon in by-pass has no damping restrictor the diagram of the elevator does show one. This is in its mode selector by-pass duct and is similar to that in the rudder damped mode yet unmentioned in the description of operation. (Not the only error in that.)

    Thus while the elevator has a by-pass solenoid valve and the manuals describe such a mode in fact there isn’t – that mode is damped. That is consistent with what is described in Davi under Fig 11.3, which makes clear this is to dampen flutter.

    Still, none of this answers where the solenoid power comes from in a blocking interregnum.

  478. HB says:

    @David, from the diagram the selector valve for the rudder allows flow in the damped condition but has a rectricted orifice and allows some flow to go through. For the elevator, the selector valve in blocked condition does not allow any flow at all.
    There is alse separate bypass valve in both case that allows to selector to allow full flow of hydraulic oil.

    i am not sure where the confusion for the solenoid power. it is a spring back valve maintained open by two solenoid coils each powered from their respective ACE.

  479. Andrew says:

    @David

    RE: “I think the answer to an elevator panel being exposed to risk of flutter while it awaits air loads getting it close to neutral is that in so called by-pass mode the PCU is in fact damped.”

    You beat me to it while I was composing my reply, but I’ll post it anyway seeing as I spent so much time writing it!

    @HB

    RE: “But in the Case of both engine failures, i don’t think that this issue will happen.”

    Maybe, maybe not. The jury is still out on whether the PMGs can supply useable power to the PSAs*. For the moment, let’s say they can’t. In that case, the L2 ACE will lose power immediately after the second engine failure and the ROB elevator PCU will be in bypass mode. The R elevator will be driven by the RIB PCU, with the R ACE and the RIB PCU powered by the R PSA’s FCDC battery. If that battery goes flat before the APU starts generating power, the R ACE and the RIB PCU will also lose power, causing both PCUs to go to blocking mode. Going back to @Don Thompson’s original question, what happens if the elevator is deflected at the time?

    * For those suffering acronym overload:
    PSA = power supply assembly
    ACE = actuator control electronics
    ROB = right outboard
    RIB = right inboard
    PCU = power control unit (a fancy name for hydraulic actuator)
    FCDC = flight controls DC

    In relation to the elevator blocking mode, you previously commented:

    I agree that that failure mode exists. It cannot be avoided. It is either that or running the risk to experience flutter.

    I don’t know if the elevator PCUs will immediately go to blocking mode in the event of a power failure, or not. However, I don’t agree with the logic that an immediate change to blocking mode is required to prevent flutter. The bypass mode of the elevator PCUs is different to that of the other control surfaces, because there is a restrictor in the elevator PCU mode selector valve that damps the movement of the elevator if both PCUs are in bypass mode. Accordingly, the elevator bypass mode could be described as a ‘bypass/damped’ mode. That damping would surely help to suppress flutter, especially in the short time it would take for the elevator to move back towards the neutral position if it were deflected.

    Consider the failure scenario where one elevator PCU is failed and is in bypass mode. The same elevator is then driven by the remaining PCU in normal mode. If the remaining PCU subsequently fails due to a hydraulic or electrical failure in the PCU, its controlling ACE will keep the blocking solenoids on both PCUs energised long enough for the elevator to move back to the neutral position. Why is that allowed to happen if flutter is such an issue? According to your logic, wouldn’t the blocking solenoids be de-energised immediately?

    RE: “There is alse separate bypass valve in both case that allows to selector to allow full flow of hydraulic oil.”

    That is not correct in the case of the elevator mode selector valve. The bypass position of that valve has a flow restrictor that damps the movement of the actuator.

  480. Richard Godfrey says:

    @Paul Smithson

    The basis of the statement, MH370 was between 4°S and 16°N at 19:41 UTC on the 2nd Arc, is the intersection of the circle defined by the range from the Last Known Position at the maximum speed of a B777-200 ER until 19:41 UTC, with the 2nd Arc as defined by the satellite data at 19:41 UTC.

  481. David says:

    @HB. In your earlier post you said, ‘Note the same applies to the ailerons having a hydraulic lock.
    Introducing this lock inherently introduces this failure mode which could be critical if it occurs during the time at risk where the elevator is in the undesirable position. There is no feature to delay the block valve action for a number of scenarios.”

    The aileron PCU has the three modes; normal, by-pass and blocking/damped. The blocking mode is there for deliberate lock-out above cruising speed, ie ACEs powered. By-pass is there like the flaperons – in case of a PCU hydraulic or electrical failure, so that its companion can continue to drive the control surface.

    By-pass/damping according to the AMM, “..happens when the PCU has no power or the PCU has a failure at or above cruise speed.” In that mode the aileron cannot lift but it can lower, damped. This would replicate the elevator if the aileron happened to be deployed down at the time of double ACE failure, though the concern about the left elevator, apparently vulnerable, is lessened by PMG power most likely powering the right ACE at fuel exhaustion.

    @Don Thompson, Andrew. The aileron lock out function above cruise, in effect a straight block, is described thus in the AMM, “With flight at or above cruise speed, the PFCs command, through the ACEs, the ailerons to fair and to lock out.” This sounds very like the elevators in character, the point being that the fairing is commanded. I am developing the impression that the elevator PCU design envisages them being commanded in the same way, on PCU failure, elevator position neutrality being awaited, ACE(s) apparently presumed to be operative.

    Perhaps as above PMG power to the right ACE reduces the double ACE failure risk to the acceptable in those circumstances.

  482. David says:

    @Andrew. Like you the above was written without having seen your most recent.

  483. Barry Carlson says:

    Here is a little story that could well have an unexpected sequel.

    During the various search phases for AF447 in the Equatorial South Atlantic, there were numerous vessels utilized in the various search phases. The initial Phase One search was for the ULB’s and used 2 vessels. Also included was the tracking of naval vessels during the recovery of bodies and other floating debris. Then followed the Phase Two search using both AUV’s and towed side-scan sonar, and the final Phase Three search using the Alucia and AUV’s.

    All the above events were covered in various social media blogs, e.g. PPRUNE, by those tracking the vessels using Satellite – AIS.

    It wasn’t until after the searches had successfully ended and the DFDR/CVR were recovered that the BEA became aware that all the vessel movements over the previous 2 years had been shared via social media/blogs etc..

    Since then there have been two other notable though short term missing aircraft searches in the marine environment, and more recently the ongoing MH370 search conducted by the ATSB on behalf of the Malaysian government, and dutifully recorded via S-AIS. The resurrection of the MH370 search by Ocean Infinity on a ‘No Find, No Fee’ basis involved a deal / contract with the Malaysian government. We know some of the ‘numbers’ associated with that contract, but the small print as usual is missing in action.

    The new search using the Seabed Constructor and its fleet of AUV’s has received similar attention, and a number of persons have been publishing S-AIS tracks and deducing the numbers of AUV’s utilized and their tracks. That being the case, those who are interested in the progress of the search have been able to keep abreast of developments.

    Sadly, at 20180131-1856UTC the S-AIS reports from the Seabed Constructor stopped, while S-AIS reports from vessels nearby and in range of a participating satellite’s orbital path were still able to be tracked.

    Regulation 19 of SOLAS Chapter V – Carriage requirements for shipborne navigational systems and equipment – sets out navigational equipment to be carried on board ships, according to ship type. The regulation requires AIS to be fitted aboard all ships of 300 gross tonnage and upwards engaged on international voyages, cargo ships of 500 gross tonnage and upwards not engaged on international voyages and all passenger ships irrespective of size. The requirement became effective for all ships by 31 December 2004. The technical requirements follow ….

    It’s the Maritime equivalent of Aviation’s Transponder / Secondary Radar combination and its associated TCAS, plus Mode B.

    The Master of each vessel is responsible for compliance, and in this case the Norwegian state is the controlling authority.

    Go figure who’s behind this strange turn of events?

  484. haxi says:

    @Barry Carlson, trying to link SC’s AIS silence to the Norwagian state is…unfair. After all, Fugro vessels have been tracked for years, and they should know that. Why switch off after beginning the mission for days?

  485. Mick Gilbert says:

    @Richard Cole
    @Richard Godfrey
    @Barry Carlson

    Gentlemen, it’s obviously quite frustrating that we’ve lost the AIS feeds for Seabed Constructor. Hopefully that will be restored.

    In the meantime, I notice that DIAGRAM 2 of the MH370 OPERATIONAL SEARCH UPDATE #1 released by the Malaysians shows a dark line running through the length of the the shaded “data coverage” area. That line appears that it might be where Seabed Constructor returned to after scanning the inner (northern) band of the priority (site 1 area 1) search area. I’m starting to think that the line may represent a narrow swathe for which there is no data.

    Richard C, you had noted that SC appeared to have launched and was shadowing an AUV that was probably not using its side-scan sonar given the variations in speed. Do you think that the AUV tasked with that swathe on the original sortie may have had a problem with the look-down echo-sounder and they are now re-scanning that swathe? Might that circular manoeuvre prior to launching the AUV have been to ensure that Seabed Constructor was accurately aligned with the “missed” swathe?

  486. Victor Iannello says:

    @Mick Gilbert: I also noticed that “line” in the completed area. I don’t know about your reason for the circle manoeuver, but I think the rest is true.

  487. Victor Iannello says:

    OI already released to Malaysia an area that was scanned, reviewed, and cleared. Based on that, they are likely just filling in a line that was missed. I don’t think they are investigating a previously-identified contact, as Malaysia says none were identified. I don’t know the reason why they turned off the Sat AIS.

  488. Richard Godfrey says:

    The only reason for going dark that is justified in my view, is that Ocean Infinity has found something.

    They need time to double check and then time to notify the NOK first.

  489. Don Thompson says:

    @Mick, all

    Constructor’s speed is consistent with the AUV performing an MBES survey. The AUV’s MBES instrument is used to ‘fill-in’ the SSS nadir area under the craft, MBES backscatter provides imaging for this narrow under-vehicle swath where the SSS is blind. The narrow linear ‘gap’in the Diagram 2 of the Msia Report possibly correlates to that nadir scan failure, or even due to an issue with one side of a SSS receiver array.

    To reacquire this narrow swath, Constructor followed the AUV closely so as to maintain an accurate survey line with constant nav updates using the acoustic USBL command link between AUV and ship. The small variations in Constructor’s speed probably reflect the AUV’s terrain following. The detailed bathymetry shows this area of seabed isn’t flat. The AUV will be ascending-descending to maintain constant survey altitude.

  490. Don Thompson says:

    @David, Andrew, HB

    Thanks for the detail on the cross connect on elevator PCU function.

    @David wrote, “I am unsure where this continuing discussion is leading.“. Clarification for the PFCS behaviour. Also, refer to last paragraph above

  491. Don Thompson says:

    @Mick, all

    For clarity, my opening line above should read:

    After 2018-01-31 03:20, Constructor’s speed…

  492. Victor Iannello says:

    @Richard Godfrey: It’s possible they are investigating a contact located in the inner primary area, an area that was not yet cleared and released to Malaysia. However, past history shows the scanning will generate many contacts of interest that are not the debris field. It may be that OI has chosen to go silent every time there is a contact they are investigating.

  493. Richard Godfrey says:

    @Victor Iannello

    If Ocean Infinity go dark every time they are investigating something, then it will make it easy to spot.

  494. John Halabarec says:

    @Mick Gilbert
    @Victor Iannello

    Might the circular manoeuvre been a quick “practice run” to ensure they could a steady path before they began tracking the AUV to the south, as it would be the first time in this mission where this sort of exercise had been performed.

  495. Victor Iannello says:

    @Richard Godfrey: The might go dark when they don’t want to “announce” the precise location of a contact that warrants further investigation. Perhaps it is to ensure that investigating a contact will not compromise independent review of the data by DOS, for instance. So going dark may be for quality assurance more than anything.

  496. Richard Godfrey says:

    @Victor Iannello

    Transparency is very important for a project such as MH370.

  497. Victor Iannello says:

    @Richard Godfrey: So is quality assurance while the data is still under review. If going dark gives us higher confidence in the review process, I support the decision. All hypothetical, of course.

  498. Ge Rijn says:

    I can imaging it can be distracting for OI and people on SC that every move they do is followed carefully by the community.
    The ‘every move you make, every step you take, I’ll be watching you’ syndrome..
    Probably many of OI and those onboard will also read comments here and else.
    To keep everyone focussed on the task it could be helpfull to go dark at times.

  499. Rob says:

    @DennisW

    “I really try to be reasonable with cogent arguments (many of them not provable), but I would never reject the analytics of other qualified professionals without an extremely good reason.”

    Dennis, you are either gullible or mischievous. I’m not sure which, so I will have to give you the benefit, and say gullible.

    Those CSIRO priority zones are nothing but a desparate clutching at straws. The ATSB needed to get out of a fix, a fix mostly of their own making, when they couldn’t find the plane. They cut short the original search effort in that part of the 7th arc. There had been dissention in the ranks of the SSWG members and from outside too (me included) concerning the prospects of finding anything from S35 northwards. In promoting the 25,000 sq km zone is nothing more than a retrospective tidy up, keeping their fingers crossed, but I don’t think lady luck is going to cooperate.

    The reason, the only reason, they never found the aircraft in the original search is because they didn’t go far enough downrange of the DSTG hotspot. When they finally get round to S39.20 E88.36, then all will miraculously become self evident.You will hear people saying “yes, now I always had doubts about their decision to go for S35, but I didn’t say anything at the time because it would be breaking ranks. Get with the program, they said, there’s safety in numbers, you know. The more people believe something, the more likely it will be true”.

  500. Rob says:

    @Ge Rijn

    “Probably many of OI and those onboard will also read comments here and else.”

    Poor devils, think what they’ve been going through! It must be bad enough just being out there in the middle of nowhere, going round and round in circles, but to have people talking about you, criticising you, chewing the fat and arguing almost to the point of fisticuffs, over where you should go next. Totally intolerable. The Captain had a choice; order a blackout or face a mutiny.

  501. Paul Smithson says:

    @Richard G. Thanks for your response on evidence base for latitude limits at 1941. Am I right in thinking that you are treating 1822 as LKP rather than 1802?

  502. Ge Rijn says:

    I’ve searched all I can on the B777 PMGs output when windmilling. Nothing decisive yet. The B747 PMGs delivere sufficient power to the primary flight controls (they even don’t have a RAT) and the B787 also does when windmilling (at sufficient speeds).
    I can hardly imagine the B777 with its ETOPS clearence won’t do the same.

    But no proof still.

    Consequence could be if all or most crucial ACEs and solenoids were kept powered after second engine flame-out the plane could have ended up outside the +/- 25Nm area without any pilot-input.
    The steep descent could be coincidental at the time of those final BFO’s but this makes no sense in this regard.

    An active pilot induced descent at this time (final BFOs) makes a lot more sense considering this all.

  503. Ge Rijn says:

    @Rob

    You’ve got my point.

  504. Don Thompson says:

    @Ge Rijn wrote “Consequence could be if all or most crucial ACEs and solenoids were kept powered after second engine flame-out the plane could have ended up outside the +/- 25Nm area without any pilot-input.

    kinda like teacher asking Jimmy to recite his alphabet:

    “Sure ma’am, A to Z, don’t we all know what’s in between?”

    747’s, except -8, don’t have a RAT. Four engines were deemed adequate. Damned unlucky to lose all 4 altho not unheard of, and I have watched a pilot, in a sim, successfully deal with loss of three.

  505. Victor Iannello says:

    @Rob: You ignore the real reasons that nobody is following your advice: the investigators believe the drift models, the BFO data, and the debris analysis all run counter to a glide and impact at 39S latitude. All three evidence sets would have to be misinterpreted for your scenario to be a possibility. If you would like to present reasons why the three evidence sets have not been properly interpreted, please do so. Otherwise, it’s time to move on.

  506. TBill says:

    @all
    Does anybody have a better map of the Broken Ridge sea floor at 32S…Google Maps seems very vague there. We should probably suggest Google update with the MH370 bathy data.

  507. Ge Rijn says:

    @Don Thompson

    You have a problem which is commonly associated with narcissism I have to conclude. Main chararteristics are grandiosity, devaluing anyone you consider is ‘beneath’ you and lack of empathy. Heared it before? I’ve tried my best but showing this makes you even more predatory. The signs of a true narcissist to me.

    I’m done with you. BTW you only contribute figures and numbers. Nothing new, original or substantial gathered from your own thoughts. Most taken from others. What I bring in, you ridicule and keep ridiculing like a classic narcissist.

    I told @VictorI I woud stop bickering on you but this I can not leave unanswered for you don’t stop. This is my final comment to you unless you take notice and redeem yourself

  508. Don Thompson says:

    @TBill,

    Diamantina Escarpment & Broken Ridge are illustrated here with the 7th arc, the boundaries of the 25,000km² area and the previously side-scan surveyed area shown. The map includes the entire extent, NW and SE parallel to the 7th arc, of the Geoscience Australia bathymetry data in the vicinity the Diamantina Escarpment.

    GEBCO appears to name the region south of the Diamantina Escarpment as the Kerguelen Plateau, and north of it as Broken Ridge (a plateau, rather than an escarpment).

  509. Andrew says:

    @Ge Rijn

    RE: “The B747 PMGs delivere sufficient power to the primary flight controls (they even don’t have a RAT)”

    Are you sure about that? As far as I’m aware the PMGs on the B747 provide power for the generator control units (ie IDG excitation current, etc). There are also separate PMGs (referred to as ‘dedicated alternators’ in the B744 or ‘permanent magnet alternators’ in the B748) that provide power for electronic engine control, as @Don Thompson mentioned earlier. None of those PMGs provide power to the flight controls. The PMGs for the electronic engine control are required to provide power at windmilling RPM to allow the in-flight re-start of a windmilling engine.

    BTW, the B748 does have a RAT, but only for backup hydraulic power.

  510. Richard Godfrey says:

    @Paul Smithson: “Am I right in thinking that you are treating 1822 as LKP rather than 1802?”

    My apologies! I should have pointed that out.

    Yes, I used the 18:22 point as LKP.

  511. Don Thompson says:

    @Ge Rijn,

    I offered a little clarity on the 747 electrical configuration & teased you on yet another reinforcement of the “…ended up outside the +/- 25Nm area without any pilot-input.” mantra. Lighten up.

    Recently, I’ve been attempting to explore alternative interpretation that might align with the scant emprical data we have: the time of the final two satcom events, and the recorded BFO associated with those two events; plus the derived times for fuel exhaustion, APU, and so on. I may, or may not, decide to describe what I think the conditions were prior to or, during those events.

    Our host chastised @Rob earlier today for again repeating his mantra, did I not read you committed to lay off on yours?

  512. Paul Smithson says:

    @Victor. I think you might be overstating 39S being strongly counter-indicated by drift models. I do understand that “in general” they indicate areas further north and that “in general” beaching in Aus is more likely the further south you go. HOWEVER, please bear two things in mind. First, when the drift models were first being reviewed & refined by CSIRO they reached the conclusion that the DSTG hotspot was “not incompatible”. Second, when you look at the drift origin probability map you will see that there is a “window” around 38-39S – in the same way as there is at 35S. The reason behind this is a strong eddy/micro gyre that was present in both positions at the time of the incident – which would have resulted in initial drift trajectories N and W rather than E. There are other reasons to believe that the original hotspot is counter-indicated: a) null result on search b) poorer BFO fit. But I think it is fair to say that drift models alone do not count strongly against that location on the arc. Happy to hear your views/rebuttal.

  513. Victor Iannello says:

    @Paul Smithson: It’s not just the lack of debris in WA. It’s also the timing of the debris recovered in Africa. The timing of the discoveries of Roy and the right flaperon don’t match 39S latitude.

    I didn’t include the null result of the previous search and the poorer BFO fit because @Rob’s scenario included a glide past the search area, and the poorer BFO fit might still be acceptable with some oscillator drift. The other three evidence sets I think are stronger.

  514. TBill says:

    @Don Thompson @Nederland
    Those are some wonderful maps, thank you. Looks like there are some blank spots where OI will have to be careful. It seems generally consistent with Google maps as far as deepest spots (5000m = 16500 ft). I don’t know for sure, but I assume the hardest spot to find MH370 would be on the ragged sides of mountain ranges.

  515. Ge Rijn says:

    @Don Thompson

    Oke. On this level I can respond to you. It’s not about a mantra. It’s about your attitude to me.
    The case about the PMGs and the attitude of the plane after second engine flame-out can be of huge importance on the EOF scenario and the current search effort.

    If the steep descent could not have been caused by electric/hydraulic failure on flaperons, elevator, RAT, batteries, APU, or other causes it could only have been caused by pilot-input.

    And then there is the possibility this PIC recovered from this steep descent and glided the plane outside the +/- 25Nm area.

    This is what I keep stressing unless proof turns up this is impossible.
    I haven’t seen any yet.

  516. Victor Iannello says:

    @Ge Rijn: If that’s your goal, I will say that without the flight data recorder, it will be impossible to definitively prove whether or not there was pilot input after fuel exhaustion. Anything that occurs without pilot input can occur with pilot input. You have made it abundantly clear that you believe a glide past 25 NM is possible. It’s impossible to prove this wrong. It comes down to subjectively assigning probabilities, and we’ll never reach 100% consensus on these. From a practical point of view, only if the +/- 25 NM search fails is there reason to argue about this. We are, at a minimum, months away from that point. Let’s move on.

  517. Don Thompson says:

    @Ge Rijn.

    You wrote: “it could only have been caused by pilot-input.

    If you imply that pilot input is necessary after fuel exhaustion to cause the rate of descent indicated by the 00:19:38 BTO, then I categorically refute that. There are conditions that may have been established earlier in the flight, the AFDS compensated for those conditions, fuel exhaustion caused caused the AFDS to disengage, and a rolling, descent then began.

  518. Don Thompson says:

    @TBill,

    Google’s bathymetry data is typically not better than 1km resolution. The Geoscience Australia downloadable bathymetry for the 7th arc is 45m resolution.

  519. HB says:

    @Andrew,
    i double checked and correct, the bypass selector position has also a restricted orifice for the elevator. But it is not specified to be a damping mode (see below)

    RE “I don’t know if the elevator PCUs will immediately go to blocking mode in the event of a power failure, or not. ”
    I keep arguing that it will go to blocking mode immediately. Assuming that all the digital output signals from all the ACEs are deenergised for whatever reason (total power failure including battery failure, concurrent electronic failure, etc.), all the solenoid valves will spring back to the deenergised position (by spring action and not from another power source) immediately. For the elevator, the deenergised position is blocked mode and the selector valve will spring back to the blocked condition immediately.

    The same will apply here if you lose the hydraulic pressure, the selector will spring back to blocked on loss of pressure on the hydraulic pilot line (no power or ACE logic required for that, it is all mechanical)

    RE: “there is a restrictor in the elevator PCU mode selector valve”
    Yes, correct

    “that damps the movement of the elevator if both PCUs are in bypass mode. Accordingly, the elevator bypass mode could be described as a ‘bypass/damped’ mode. ”

    I would say no unless Boeing confirms the opposite. To be effectively damping to prevent flutter, it has to be sized/designed adequately. Since the terminology does not use “Damping”, i doubt it will be adequately designed to stay in this mode for a long time. I would assume that bypassing for too long could lead to flutter as tail control surfaces are very prone to flutter. (see below)

    RE: “However, I don’t agree with the logic that an immediate change to blocking mode is required to prevent flutter.”

    I see your point of view but my undestanding is as follows:
    Flutter in a nut shell: increasing the speed will get you closer to flutter criteria, lowering the natural frequency for the same speed will also get you closer to the flutter criteria.

    When bypassing is applied, it lowers the natural frequency (as you introduce more degrees of freedom) and the control surfaces on the empenage are prone to flutter in such case. Obviously, time in this mode should be minimised.

    Compared to bypassing, damping mode yields the same natural frequency as bypassing and it only attenuates the oscilation amplitude (which could still overshoot the design conditions depending on the design) it does not change the natural frequency away from the possible flutter conditions, ie still prone to flutter but the amplitudes are damped. It may not be possible to damp the amplitudes sufficiently to avoid damage and in that case you also want to minimise the time in this mode.
    When Blocking is applied, the natural frequency is back to normal and you are immediately away from the flutter conditions. Ie to manage flutter, it is better to move away from the natural frequency range than to manage the amplitudes.

    RE: “Why is that allowed to happen if flutter is such an issue? According to your logic, wouldn’t the blocking solenoids be de-energised immediately? “.

    where possible, you want to maximise the chance to block in a friendly position and build redundancy to minimise the chances of blocking directly in an unfriendly position. That’s my understanding of the design intent.
    Now if flutter is less severe than blocking one elevator, the design needs to be reviewed. I dont have the response to which one is more severe.

  520. DennisW says:

    @Rob

    Dennis, you are either gullible or mischievous. I’m not sure which, so I will have to give you the benefit, and say gullible. 🙂

    Definitely not intending to be mischievous in this case. I do feel bad relative to telling TBill to “give it a rest” earlier. Everything we do is still potentially worthwhile, and there is no reason not to keep digging and thinking.

    I guess my feeling is best described as “frustrated”, and it sometimes shows up. I am not very good at saying the “right” things which is pretty much expected in this day age.

  521. Rob says:

    @DennisW

    I know I must be insufferable at times. Sorry, but my social skills are deficient, and social skills are not entirely learned, they are partly innate. You cannot offend me. I am practically offend proof. I hope I haven’t ever offended you. If I have, I apologise unreservedly

    Nearly everything we do here is of real value one way or another, but sometimes it’s advantageous to play devil’s advocate. I have learned a great deal by bouncing ideas off of people here. Crowd sourcing is great in an otherwise intractable problem like MH370, as long as the disruptive types are kept at bay.

    So let’s rock on. Things could get really interesting.

  522. Mick Gilbert says:

    @Richard Godfrey
    @Victor
    @Don Thompson

    Thank you for the follow up on my initial observation/enquiry.

    Don, apologies for not addressing you in my original post, I forgot that you’re one of the go-to guys for the underwater stuff.

  523. Rob says:

    @Victor

    Sorry if I’m a pain. I happen to think I’m right on the drift study issue, but I acknowledge I could turn out to be wrong, but at the moment I personally think I’m on the right track. Either way, what’s the harm in healthy debate, where people are not afraid to speak their minds. It’s always a bad move when people stifle free debate. OI have three months to find the plane, so there’s a better than even chance they’ll find it, especially with a financial incentive.

    There are still lots of unresolved issues. For example it could turn out that chasing down BFO errors was pointless. Interpretation of the debris itself and what it tells us about how the aircraft hit the water, is still very much open to debate. The final BFO could still come back to bite us.

    When they find the plane, I will be overjoyed. Wherever they find it, I will be overjoyed like everybody else here.

    We have one particularly insidious common enemy, called confirmation bias. I am not immune, and neither are you.

  524. sk999 says:

    The map included with the Jan 30 Malaysian Gov’t search update report comes with an embedded coordinate grid that appears accurate. For those interested in digitizing the search zone, I’ve created a rectified version of the map here (top item, as usual):

    https://docs.google.com/document/d/14hleZyx1pUPL44yaeHKt6jnSQ3DbgRq2zibbKkFLq2c/edit?pref=2&pli=1

    The search zone appears to be considerably wider than +/- 25 nm – thus, Seabed Constructor’s foray “outside the search zone” was still within the one depicted on this map.

  525. lkr says:

    @sk999:

    Yes, that’s the same map that JW picked up yesterday. The scale bar clearly shows a uniform width of +/- 25 nm from 7th arc, all the way to 29N.

    BTW, it’s reasonable to believe that the official Malaysia progress reports are via the two Malaysian officers embedded the OI ship. Will OI be reporting independently either to ATSB or via their onshore assets?

  526. Andrew says:

    @HB

    RE: “I keep arguing that it will go to blocking mode immediately. Assuming that all the digital output signals from all the ACEs are deenergised for whatever reason (total power failure including battery failure, concurrent electronic failure, etc.)…”

    Sure, but the assumption is that ‘all the digital output signals from all the ACEs are de-energised’ immediately. What if that’s not the case when power from the PSA suddenly fails?

    RE: “I would say no unless Boeing confirms the opposite. To be effectively damping to prevent flutter, it has to be sized/designed adequately. Since the terminology does not use “Damping”, i doubt it will be adequately designed to stay in this mode for a long time. I would assume that bypassing for too long could lead to flutter as tail control surfaces are very prone to flutter. (see below)”

    Perhaps ‘damped’ is the wrong word, but the restrictor would certainly slow the movement of the elevator in bypass mode. I did not say the PCU is ‘designed to stay in this mode for a long time’. It would only need to stay in bypass mode for the short time it would take for the elevator to be blown from its deflected position back to the neutral position (‘near to faired’ according to the AMM) by the air loads. The PCU would then go to blocking mode and the elevator would be locked. That doesn’t seem conducive to flutter.

    RE: “…you want to maximise the chance to block in a friendly position and build redundancy to minimise the chances of blocking directly in an unfriendly position. That’s my understanding of the design intent.”

    If that’s the design intent, then why not allow the elevator to move back to the neutral position in both the cases I described previously?

  527. Andrew says:

    @Mick Gilbert
    @Don Thompson
    @Victor
    @David

    RE: Mick’s question: “Are we saying that the synoptic is definitely not an accurate representation of system and control status on RAT power only?”

    A quick trip to the B777 simulator has helped answer that question. The simulator was set up at FL350 with a low fuel quantity and both engines were allowed to flame-out due to fuel exhaustion. The following images show the flight control and hydraulic system synoptics after the flame-out, with the engines windmilling and the RAT operating:

    Flight Control System Synoptic

    Hydraulic System Synoptic

    In a nutshell, the L & R hydraulic systems remained powered with the pressure around 3,000 psi. Apart from a few spoiler pairs, all flight controls, including the left flaperon, remained powered by at least one PCU. The L2 ACE was inoperative, as expected.

    In case you’re wondering, the aircraft rolled to the right. The system pages available at the instructor’s station showed that the left aileron was deflected -3° (ie down), while the right aileron was deflected -1°. Both flaperons were at 0°. I have no idea why the ailerons were deflected slightly down, let alone why the left was deflected more than the right.

  528. Andrew says:

    RE: My last comment “I have no idea why the ailerons were deflected slightly down, let alone why the left was deflected more than the right.”

    On that point, both ailerons were at +1° before the engines flamed out.

  529. ventus45 says:

    @Andrew
    Re: “On that point, both ailerons were at +1° before the engines flamed out.”

    Is that normal ?
    Why would they be both at +1 in level flight at FL350 ?
    Was it for gust aleviation, or is that normal for spanwise lift distribution control, or something else ?

  530. Victor Iannello says:

    @Andrew: Can you describe what happened next? Did the roll stop at some bank angle? Or continue into a steep descent? Phugoid behavior?

  531. Andrew says:

    @ventus45

    RE: “Is that normal ?”

    It might have something to do with wing bending relief, as you suggested, but I’m only speculating. I’ll have to dig around for some more information.

    @Victor

    The roll stopped at about 20° AOB and the aircraft exhibited a phugoid behaviour in pitch. There was no steep descent. In this case, both engines failed at more or less the same time (within a second or two). Unfortunately I didn’t have time to explore it further.

  532. Andrew says:

    @Victor

    Further to my comments above: There was no restart of the left engine and the APU did not start either, so it seems the modelling in this case does not take into account the possibility of residual fuel in the tanks once the fuel state goes to zero,

  533. Victor Iannello says:

    @Andrew: OK. The aileron asymmetry was light, and was balanced by the dihedral effect at 20°. At this bank, I would expect phugoid behavior.

  534. Ge Rijn says:

    @Andrew

    Could it be the RAT caused the right banking?
    While the RAT is sticking out of the right side of the fuselage I think this would cause a momentum on the plane pushing it slightly to the right.
    Very interesting you did this simulation which seems to confirm hydraulic pressure and electrical power to the ACEs (except L2) and all essential PCUs is maintained after engines flame-out without the APU starting.

    Can you tell how long you maintained the simulation/flight after engines flame-out?

  535. sk999 says:

    lkr, you say, “Yes, that’s the same map that JW picked up yesterday.”

    Superficially yes, in detail no. JW’s version (the bigger of the two that he offers) is a hacked, size-reduced .jpg version. Mine is extracted directly from the Malaysia gov’t PDF file. Full resolution. (or at least, that is how it started before I carried out my “recification”.) A matter of taste, I suppose.

    You ask, “Will OI be reporting independently either to ATSB or via their onshore assets?”

    I have no more information than you do. My understanding is that the ATSB is on the sidelines, but what the heck do I know?

  536. Andrew says:

    @Ge Rijn

    RE: “Could it be the RAT caused the right banking?

    In this case I’d say it was the asymmetry between the ailerons. The left aileron was deflected downward slightly more than the right.

    RE: “…seems to confirm hydraulic pressure and electrical power to the ACEs (except L2) and all essential PCUs is maintained after engines flame-out without the APU starting.”

    Yes, but I would still expect the FCDC battery powering the R ACE to go flat after a short time if the APU doesn’t start. When that happens, the R elevator and a couple of additional spoiler pairs should become inoperative. I only had time to watch the aircraft descend for about five minutes, but during that time there was no change to the status of the ACEs or the flight controls. I can’t say if that’s realistic or a function of the simulator modelling.

  537. Andrew says:

    @Ge Rijn

    Further to my comment above about the FCDC battery: The battery is rated to support an emergency load of 30A for one minute. In this scenario, I think the load would be less than 30A, given there are no control inputs occurring. Consequently, the battery would probably last longer than the nominal one minute quoted in the manuals. How long, I don’t know.

  538. HB says:

    @Andrew, I agree that it will be better to transit to bypass before blocking but it is not the case right now for worst case scenarios such as both ACE failures. I dont know whether a design solution exists to allow that for all the scenarios.

    RE: “Sure, but the assumption is that ‘all the digital output signals from all the ACEs are de-energised’ immediately. What if that’s not the case when power from the PSA suddenly fails? ”

    Good question. Assuming worst case of failure of 3PSAs, the ACE-C is fed from its one battery according to Davi’s Figure 11.3 and the PCUs outputs analog and digital will remain powered until this battery is depleted (ie the graceful degradation scenario i described and not so graceful for pilots and yes in that case one of the elevators could be in an undesirable postion). Ie all the pcus for L and R will go to deenergised condition immediately, elevator will go the blocked condition immediately.

    The battery is totally independent from the PSAs and power is not routed from the PSAs. If that battery is depleted or power to the PSAs is not reestablished, then all the outputs for ACE-C will go to deenergised conditions immediately.

    Now if at anytime, the power is reestablished to any ACE’s, the respective PCUs will revert automatically to normal (energised condition – no memory is required). This may take some time as ACE powering may have some self-diagnostic checks.

  539. Michael John says:

    Whilst a ditching is judged to be highly unlikely it is interesting how most of the debris found to date appears to be from the right hand side of the aircraft & there is talk of Mh370 banking to the right at fuel exhaustion.

    Can anyone simulate what would happen if Mh370 contacted the water in a ditch like fashion with a slight right hand bank? I am curious to see what debris would be produced & whether that would match the debris found to date.

  540. lkr says:

    @sk999:

    Sorry to get in your hat again. And thanks for the work you did to rectify the projection from the Official MH370 site. I did not say you based your map on JWs, rather that both yours and JWs were based on the Official search map.

    What I was referring to was your comment that the search zone in this map is “considerably wider than +/- 25nm. All 3 versions of this map clearly show a scale [at lower right]. And all at eyeball range, show a search band 50nm wide at least from Zone 2 [yellow] to the northern end at 29S. I might give a you couple more nm at the north end of Zone 2, but that looks like an artefact of the original projection..

    The question about OI having a separate public voice was meant to be directed at large… Probably moot — if I were in their shoes, I’d be happy to have the Malays sending out weekly updates. And of course archiving all data onshore in Australia..

    Again, peace!

  541. Paul Smithson says:

    @Andrew. I realise that your sim test was not specifically designed to test this. But do your results suggest that the pitot heat also would not be lost? If that is the case, would we not remain in normal law?

  542. David says:

    @Andrew. At power failure of the L2 ACE, the left aileron outer PCU’s by-pass and blocking solenoids should be de-energised resulting (IMO) in that PCU allowing downwards movement but no upwards. That would be followed by the inboard PCU failing after centre system hydraulic failure, its ACE selecting by-pass. That would allow 15 degrees of upward aileron lift by aerodynamic force were it not for the outboard PCU putting a stop to that.

    The AMM version of that supposes that at ACE failure just the by-pass valve solenoid will be de-energised, the outboard PCU then being in by-pass. Both PCUs being in by-pass will then allow the aileron upwards movement.

    To me that does not envisage that ACE failure can be a power failure, mirroring the elevator parallel we have discussed. Like that, the appearance is that the designers have not foreshadowed ACE power failure in the design of either control surface.

    I think my version is supported by you not having noticed a sharp left roll. However I should emphasise that is what would happen in my interpretation of the AMM were the L2 ACE to have a secret power supply (such as your capacitor) available on AC power loss; and that would help explain what might have prompted a quick left bank if so. Certainly that should be the outcome were a left hydraulic failure substituted for a power failure to the L2 ACE in the above scenario.

    Going back to your simulation, the right aileron would not have been subject to like failures.

    I had conjured up the possibility that with engine sequential failure and consequent rudder correction of yaw, on restoration of the inboard PCU on RAT deployment that might then be subject to a residual command left by the A/P for a little right bank, causing the slight asymmetry you noticed. That would be consistent with the outboard PCU being in either blocking damped or by-pass. However that does not explain the 2 deg downwards shift of the right aileron.

    Spoilers 1,7,8 and 14 being out is interesting, they relying on the same control and hydraulic sources as the only right flaperon PCU still operating together of course the left aileron’s inboard PCU.

    Incidentally, not to muddle things (!), switching of PSA power supply can take a while if the FCDC batteries are designed to last for a minute. If ACE output were not seamless during that……

    Thank you for you simulation and advice as to its outcome. Among other things it puts to bed the notion that windmilling engines might deliver enough bleed air to run the centre ADPs.

    You say there was no relight so there was no help with that aspect. I assume from what you say that no residual fuel was the cause of that and the lack of an APU autostart. Interesting that there is no software standardisation in simulation when I would have expected they ought to meet the same non-airline-specific training criteria. This scenario might be beyond that I suppose. @Don Thompson would be up with that.

    Your experience though seems to support the notion that while the descent rates of the BFOs might be replicated, that would not be at around the two minutes from MEFE mark.

  543. Paul Smithson says:

    @Andrew. Sorry, pls disregard my earlier. I see now from the flight control synoposis that it clearly indicates flight control mode SECONDARY.

    Did reversion to secondary happen instantaneously at second flame-out/spool-down?

    As regards bank/spiral inducements – your simulated scenario may not have the thrust asymmetry prior to second flame-out if the two went down at more or less the same time, so you wouldn’t have rudder trim in the way that we expect would be the case for 9M-MRO, right?

  544. Rob says:

    @Michael John

    “Whilst a ditching is judged to be highly unlikely it is interesting how most of the debris”.

    Judged to be highly unlikely by whom? If you are just repeating what others have said, you are merely serving to perpetuate a myth. That will benefit no one, in the long run.

    The ATSB conveniently decided to rule out a ditching, citing the retracted flap and the apparent high descent rate as evidence.

    The ATSB went on to conjure up a dfresh search zone, on the back of this so-called finding, with the willing cooperation of CSIRO, giving the myth a veneer of respectability. However, it remains an unsubstantiated assertion, a myth.

    What the ATSB seemingly chose to ignore is that there is more than one way to ditch an aircraft. The normally accepted way is to follow emergency procedures and ditch with flaps down a la Sully, and if you are lucky, and help is at hand, you and your passengers get to live another day.

    The other kind of ditching is the one most respectable people would rather not bring up. A ditching carried out by a pilot gone rogue, a pilot who has already killed all his passengers, and is now intent on hiding all evidence that he and his plane ever existed in the first place. After he’s flown as far as possible into the remote reaches of the SIO (M0.82 at FL400 is unfortunately no myth either), he glides as far as possible then dumps his plane in such a way as to get it under control water as efficiently as possible.

    He manages to get the plane to literally drop out of the sky, flat on it’s belly, to crack open the fuselage without shredding and distributing it and it’s contents far and wide. The aircraft naturally stalls in the process, and the right wing hits the water first. The right wing is bent upwards and breaks like in Boeing’s test to destruction. A tiny piece of the aileron washes to shore, so badly smashed, it’s hardly recognisable. Two originally mating parts of RH outboard flap fairing no7 also wash up minus their outboard sides. part of the outboard flap washes up, as does the flaperon. Bits of the left wing are conspicuous by their absence, only two pieces from the outboard wing, two pieces that also were in very close proximity when mounted on the wing, the outboard flap trailing edge section looking is a lot better shape than the right hand aileron fragment.

    An uncontrolled, high energy impact? I hardly think so.

    The Malaysians are organising the search effort. They couldn’t organise a piss-up
    in a brewery They only take advice from ATSB. So what chance of success? You tell me.

  545. Don Thompson says:

    @Rob wrote “The Malaysians are organising the search effort.

    Thankfully, the Malaysian authorities are organising nothing of consequence.

    The simple facts are that Malaysian authorities have agreed a success based fee contract following an unsolicited offer from Ocean Infinity, Malaysia has instigated an ‘operations room’ for the duration of the contract (whatever that means), and put two observers on Seabed Constructor. None of that can be construed as “organising the search“.

  546. Mick Gilbert says:

    @Andrew

    Thank you for undertaking that simulation. I think that most of us understand that a ‘quick trip to the sim’ in fact takes a not inconsiderable amount of planning and effort to say nothing of time, so again we’re indebted to you.

    Any old how, to the findings. I suppose the first thing to note is that some of the conditions for your simulation differed from what we suspect the conditions were for MH370, the near concurrent flame-outs being the most obvious. Presumably after 10 minutes or so with one engine out the TAC would have been engaged, the airspeed on second engine flame out would havs been a bit lower and there would have been a bit more nose down pitch. Would the time with one engine out prior to the second engine flaming out have impacted the overall state of the various hydraulic systems? and might that have impacted what happened post-second engine flame-out?

    The 2° aileron asymmetry subsequent to the engines flaming out is interesting. It seems that whatever which way, subsequent to both engines flaming out and while the airplane is powered by the RAT only it throws up some sort of control surface asymmetry. Like ventus45 I’m also interested in the uniform +1° deflection prior to the engine flame-outs. Do you think that uniform +1° deflection would be impacted by a period of asymmetric thrust as we suspect happened with MH370?

    Looking at the flight control synoptic, am I seeing a smidge of right elevator deflection? Might this reflect the effect that you have been discussing with Don et al.

    What’s the cause of the spoiler failures? Near as I can tell the only common system for the 1, 5, 7, 8, 10, 14 spoilers is the centre hydraulic system and that should be powered by the RAT.

  547. Rob says:

    @Don Thompson

    If it’s not the Malaysians organising it, then who is?

    Is it the ATSB behind the scenes, or Victor and the IG?

    Surely it’s not left to poor old OI themselves?

    I would do a much better job myself, believe me. At least there would then be a sporting chance of finding the plane. 😉

  548. Don Thompson says:

    @Rob

    Thank you for that insightful reply.

  549. Rob says:

    @Don Thompson

    Don’t mention it. 🙈🙉🙊

  550. TBill says:

    @DennisW
    “I do feel bad relative to telling TBill to “give it a rest” earlier. Everything we do is still potentially worthwhile, and there is no reason not to keep digging and thinking.”

    No prob.
    The following path VAMPI MEKAR NILAM 0796E 0894E looks like a possible close alternative to the SLOP/CP…it is a least a very handy flight sim short-cut.

  551. Ge Rijn says:

    @Andrew

    Different from your simulation regarding the few seconds between first and second engine flame-out at FL350;
    If (as calculated by some) the first engine flame-out on MH370 occured ~10 minutes before the second flame-out would the plane not also have automatically descended to a more favourable altitude for one engine INOP (~25.000ft)?

    I mean could the altitude at second engine flame-out logically be on this lower altitude in such a case if the first engine flame out was at ~35.000ft?

  552. Michael John says:

    @Rob

    I am in full agreement with your thoughts on ditching. However it isn’t generally well received by the participants in this forum therefore I like to tread carefully. I have satellite imagery from the 16th of March 2014 at a very high resolution. Regardless of the logic behind it the imagery does look very much like a plane. It would also support a ditching scenario. It matches Mh370 in shape & size. However the location is an issue. For data & depth reasons. I’m still holding my belief on it until the plane is found however….

    You can see the images here:

    https://drive.google.com/folderview?id=1mHY9uJQbf3WUbrzWOJSqpEbRNzaO28QM

  553. Victor Iannello says:

    @Michael John: Please infect other corners of the internet with Tomnod imagery. There’s no place for that here.

  554. Michael John says:

    @Victor

    Thanks for proving my point. 🙂

  555. lkr says:

    There seems to be an outbreak of emojis on this heretofore virgin website. For some reason, the outbreak is mappable to the only current contributors who are willing to plant a flag in a 25nm circle.

    I don’t mind reading your ideas, but please, no more emoticons here!

  556. Rob says:

    @ikr

    “There seems to be an outbreak of emojis on this heretofore virgin website. For some reason, the outbreak is mappable to the only current contributors who are willing to plant a flag in a 25nm circle.

  557. HB says:

    you guys made me study the b777 power distribution and i really wanted to avoid that. I found some interesting points (maybe known to you already for long time but some new to me) in some documents.
    One interesting reference here (http://ads-b.ca/777/B777-Electrical.pdf but i went through other references too)

    1) correct to assume that the B777 power distribution allows autotransfer between power sources without interuption and this has been extensively tested

    2) Correct to assume that the L MAIN DC, R MAIN DC lose power immediately upon two-engines failure. The R-PSA will lose power shortly after that (time to deplete the associated PSA battery).

    3) the PSA battery purpose is only to ensure power continuity during switchover. It has nothing to do with providing a delay to PCU actuation.

    4) could not find data for the ACE-C battery stated in davi’s reference. Not sure if that battery exists and what is the sizing case.

    5) the PSA-L receives also backup power from the PSA-C and vice versa as well as hot battery bus

    6) SATCOM issue. I noted that the
    * SATCOM
    * TCAS
    * Radio HF and datalink are load shed
    * centre tank override / jettison pump (center tank and auxiliary fuel cannot be used)
    * external lights
    * cabin temperature control (degraded)
    * etc.

    are all AUTOMATICALLY load shed in case of both IDGs and APUs are not operating because the backup generator cannot cope with these loads.

    I so far understood that isolating these loads required left bus isolation however another way to shudown these loads include:
    * both engine failures and delay in APU to start
    * DGs and APUs set to off (no need to touch the backup generator)(my preferred one if you want to avoid detection)

  558. Rob says:

    @Michael John

    Thank you for your support, suddenly the world isn’t quite such a lonely place.

    Fascinating pics I agree. Not surprised to learn that location is a problem, location always seems to be a problem with pics like those. There was a particularly nice one off the coast of Vietnam, in the early days. It never fails to amaze me (being serious now) how resourceful people can be when the opportunity arises. If only such talent were focused on finding the plane….. Anyway, it kind of renews one’s faith in the human race, doesn’t it

    Man is a clever little monkey, if you pardon the pun.

  559. DennisW says:

    @Rob

    Thank you for your support, suddenly the world isn’t quite such a lonely place.

    ?????

    The world is full of whackos in my experience.

  560. Don Thompson says:

    @HB

    The electrical system is quite logical, multi-layered.

    Your points 3, 4 & 5 aren’t quite correct. Each PSA has its own FCDC battery. Each PSA has two PMG feeds, the L PSA inputs are ganged to a L engine PMG, the R PSA inputs are ganged to a R engine PMG, the C PSA gets a PMG feed from a second PMG on each engine. L & C share an alternate 28V DC feed from the hot battery bus, whereas the R PSA isn’t provided with similar.

    It’s not the case that SATCOM, etc, are ‘load shed’ if engine IDG & APU aren’t available. If the IDGs & APU gen are lost there is no supply to the L & R Main AC buses, anything load supplied directly off these buses is simply lost. The intention for the Backup Generators is to maintain supply to the Backup Generator Converter which, in turn, is the alternate supply to the Transfer Buses. ‘Backup’ implies an alternate supply to the Transfer buses which are normally supplied by the L & R Main AC buses. The Transfer Buses are the critical 115V AC buses for avionics (except SATCOM!).

    Load shedding is a more granular control of power by ELMS to specific loads, typically, on the L & R Main AC buses & the direct branches/subsidiaries of those buses.

    Hope this helps.

  561. Don Thompson says:

    @Rob wrote “Man is a clever little monkey, if you pardon the pun.

    Good sensors, but variable quality output when processing received data.

  562. lkr says:

    @Dennis

    Just do the Venn diagram for “clever little monkeys”, emoticons and whackos. Not much daylight.

  563. Paul Onions says:

    @HB

    Load shedding logic is provided which disconnect non-essential airplane loads if an overload conditions exists during either ground or flight operations to protect the power generation sources. Theses loads are listed in order of priority. The first load listed will shed first.

    1. Door 4L Galley GLR
    2. Door 4R Galley GLR
    3. Door 2 Galley GLR
    4. Door 1 Galley GLR
    5. L AC Utility bus
    6. R AC Utility bus
    7. E/E Cooling vent fan ELCU
    8. Door 4L Galley ELCU
    9. Door 4R Galley ELCU
    10. Door 2 Galley ELCU
    11. Door 1 Galley ELCU
    12. L Lower & Fwd Upper Recirc fans
    13. R Lower & Aft Upper Recirc fans
    14. L Lav/Galley fan
    15. R Lav/Galley fan
    16. C2 Hydraulic pump ELCU
    17. L Hydraulic pump ELCU (on ground only)
    18. IFE Equipment (P110)
    19. IFE Equipment (P210)

    Note: IFE can be load shed by the power management panels P110 and P210.

  564. sk999 says:

    Gysbreght (on JW blog) noted that, at nederland’s geoscience link given above, there is a high quality map of the radar data, with the first major turn now properly depicted; once can even see evidence of the 10 second sampling period.

    https://www.dropbox.com/s/pnq6lxvni7888w2/Screenshot%202018-02-02%2009.04.45.png?dl=0

    A couple of glitches in the radar data seem to correlate with waves in the DSTG Kalman filter reconstruction, suggesting that speed and heading were actually more stable than the DSTG figure suggests.

    Right at IGARI, the plane turned from 25 to 38 deg heading, then turned more to 65 deg, then started the major turn left at 17:22:47, 2 min, 16 sec after passing IGARI.

  565. Brian Anderson says:

    The website is worth looking at too. There are a number of short interactive video clips illustrating a variety of seabed features. Amazing detail. Would be interesting to see OI renditions to compare.

    Brian

    https://geoscience-au.maps.arcgis.com/apps/Cascade/index.html?appid=038a72439bfa4d28b3dde81cc6ff3214

  566. Andrew says:

    @David

    RE: Yesterday’s comments.

    Both ailerons would have been locked out before the dual engine failure and subsequent loss of electrical power. In that configuration, the aileron PCU bypass solenoids are energised, but the blocking solenoids are de-energised by a command from the aileron lockout function of the PFCs. The PCUs are still in normal mode.

    When the L2 ACE subsequently loses electrical power as a result of the dual engine failure, the bypass solenoid in the LOB aileron PCU is also de-energised. The blocking solenoid remains de-energised and the PCU is then in blocking/damped mode. As I see it, the blocking solenoid of the LIB aileron PCU would already have been de-energised by the aileron lockout function. Neither PCU would be in bypass mode.

    I was also surprised by the loss of the #1, 7, 8 & 14 spoilers, which are all powered by the C hydraulic system. I had previously thought that the RAT would power all the primary flight controls powered by the C hydraulic system. On further investigation, the manuals aren’t entirely clear on that point. The FCOM states: The RAT, when deployed, provides hydraulic power only to the primary flight control components connected to the center hydraulic system. However, the AMM and Training Manual both say (my emphasis): The RAT hydraulic pump supplies hydraulic power for some of the center hydraulic system flight controls (TM 29-20-00, p.82) without specifying which controls. Still, I don’t think it matters for the purposes of defining the aircraft’s likely flight path following fuel exhaustion, as all the spoilers would have been faired due to either a lack of control input or a failure of the spoiler PCU/ACE.

  567. Andrew says:

    @Paul Smithson

    RE: “Did reversion to secondary happen instantaneously at second flame-out/spool-down?”

    Yes, as soon as the electrical power was lost. That occurred about 5 seconds after the second engine failed.

    RE: “As regards bank/spiral inducements – your simulated scenario may not have the thrust asymmetry prior to second flame-out if the two went down at more or less the same time, so you wouldn’t have rudder trim in the way that we expect would be the case for 9M-MRO, right?”

    No, the TAC had no time to react between the two engine failures. If one engine failed before the other, as suspected in the MH370 case, the TAC would have applied a rudder input after the first engine failure. That rudder input would have been removed as soon as the second engine failed. Although it takes about 5 seconds for the electrical power to fail and the PFCS to revert to secondary mode, the TAC would start removing the rudder input as soon as the thrust differential between the engines began to reduce. Any residual rudder input from the TAC would have been very quickly removed as soon as the TAC dropped out after the PFCS reverted to secondary mode. However, any manual rudder trim input applied by the pilot would remain.

  568. Andrew says:

    @Mick Gilbert

    RE: “Would the time with one engine out prior to the second engine flaming out have impacted the overall state of the various hydraulic systems? and might that have impacted what happened post-second engine flame-out?”

    I doubt it. In my simulation, the airspeed got as low as 210 KIAS during the phugoid that developed after the engines failed. All three hydraulic systems held up when the speed reduced. In the simulation I conducted last year, the speed got back to about 180 KIAS at one point and the hydraulics were still pressurised.

    RE: “Do you think that uniform +1° deflection would be impacted by a period of asymmetric thrust as we suspect happened with MH370?”

    Probably. Prior to fuel exhaustion, the TAC would have applied a rudder input to counteract the asymmetric yaw. If the autopilot were engaged in LNAV or HDG/TRK SEL, it would have commanded aileron inputs to maintain the desired track or heading.

    RE: “Looking at the flight control synoptic, am I seeing a smidge of right elevator deflection? Might this reflect the effect that you have been discussing with Don et al.”

    I’m not sure. The screen at the instructor’s station showed the deflection was zero.

    What’s the cause of the spoiler failures? Near as I can tell the only common system for the 1, 5, 7, 8, 10, 14 spoilers is the centre hydraulic system and that should be powered by the RAT.

    Good question! See my comments to David above.

  569. Mick Gilbert says:

    @Andrew

    Thank you for those answers. My question about the impact of the delay between engine failures was more to do with whether a period of single engine operations would have any impact on the hydraulic systems. In simple terms, when you lose an engine is there any degradation to the left, right or centre hydraulic systems?

  570. Andrew says:

    @Ge Rijn

    RE: “If (as calculated by some) the first engine flame-out on MH370 occured ~10 minutes before the second flame-out would the plane not also have automatically descended to a more favourable altitude for one engine INOP (~25.000ft)?”

    The AFDS initially tries to maintain the selected altitude, assumed to be FL350. However, there is insufficient thrust to maintain FL350 on one engine. Consequently, the aircraft speed would reduces while the AFDS maintains the altitude. Once the speed reduces back towards the stick shaker speed (ie just above the stall), the AFDS stops trying to maintain the altitude and the aircraft then descends with the speed just above the stick shaker speed. That whole process takes quite some time (at least 5 minutes) to play out. The aircraft keeps descending until the thrust produced by the remaining engine is sufficient to maintain level flight, assuming there is sufficient fuel remaining before that engine also flames out. The ATSB estimated the aircraft could not maintain any altitude above FL290 on one engine.

    RE: “I mean could the altitude at second engine flame-out logically be on this lower altitude in such a case if the first engine flame out was at ~35.000ft?”

    Yes. That’s what the ATSB thinks happened, although it probably didn’t descend as low as FL250 before the second flame-out. They estimated that the right engine could have failed up to 15 minutes before the left and that the aircraft began descending during that time, before the left engine flamed out.

  571. Andrew says:

    @Mick Gilbert

    RE: “…when you lose an engine is there any degradation to the left, right or centre hydraulic systems?”

    No, assuming there are no other failures and the electrical system is configured normally.

  572. David says:

    @Andrew. “Both ailerons would have been locked out before the dual engine failure and subsequent loss of electrical power.”

    I see. I was assuming speed below cruise, which I assume (again) was the MH370 case.

    I also assumed that the centre hydraulic system failed before the RAT came on line. However you say, “As I see it, the blocking solenoid of the LIB aileron PCU would already have been de-energised by the aileron lockout function. Neither PCU would be in bypass mode.” If the centre system failed before the RAT was deployed, I think the inboard PCU would be in the blocked/damping mode too, as distinct from lockout. Still, since that would prevent upward aileron movement anyway it makes little difference, particularly with no control input.

    If MH370 was beneath cruise and the centre system failed quickly I think what I theorised earlier would apply. However taking it that your simulation confirms that windmilling will preclude loss of left and right hydraulics, then 15 deg left aileron up-float on double by-pass would no longer be a possibility. It is strange though that this effect of windmilling has received no obvious attention earlier, given that it scrubs the flaperon asymmetry thesis.

    Yes I agree that the spoilers have no practical significance but I am always on the lookout for suspicious simulator activity including an explanation of the kick of the controls in ALSM’s, which straightened out a right bank. That remains unexplained to my knowledge and is part of my reason for looking into ACE/PCU what-ifs.

    @Mick Gilbert. I think you can cross spoilers 5 and 10 off your list since they are under L2 ACE control.

  573. HB says:

    @Don
    For 5 you are correct. There is a loop PSA-C battery bus hot battery bus PSA-C and PSA-C and PSA-L share the same feed from the hotbattery bus.

    Why are you saying 3 is not correct? The quote from that document is
    “Each PSA aslo uses a dedicated battery to the related flight control DC bus. The batteries have limited capacities and are incorporated to supply power for brief periods during PSA power source transfer” Ie for power source transfer and not for another purpose.

    @Paul, Don re Satcom. Right these loads are under load shedding control but in that case it will simply power down. Turning off the IDGs and APU? is that not the easiest way to depower those loads and go dark while effectively maintaining the flight controls under backup generator? or there is an interlock to prevent the pilot to switch off the three power sources at the same time? It seems intuitively easier than isolating the left bus.

  574. Andrew says:

    @David

    RE: “I see. I was assuming speed below cruise, which I assume (again) was the MH370 case.”

    According to the aileron lockout schedule, in normal mode at 35,000 ft the ailerons are locked out unless the speed is below about 200 KCAS. The speed might well have decreased below 200 KCAS in the MH370 case, if the aircraft was maintaining level flight with one engine inoperative. The ailerons might therefore have become unlocked for a time.

    However, once the PFCS reverted to secondary mode, the aileron lockout would only depend on the flap position. With the flaps retracted the ailerons would once again lockout. I think the aileron PCU status post-fuel exhaustion would have been as follows:

    LOB – blocking/damped mode, due to L2 ACE failure
    LIB – locked out (assuming C hyd system available, otherwise blocking/damped mode)
    RIB – locked out (unless R PSA FCDC battery runs flat before APU comes on-line, in which case, blocking/damped mode)
    ROB – locked out

  575. Andrew says:

    @David

    Correction: In secondary mode the aileron lockout depends on the flap and slat position. The ailerons are locked out when the flaps and slats are retracted.

  576. David says:

    @Andrew. The lockout schedule you reference is in the AMM. It is preceded by a page which is quite specific. In secondary and direct modes, “When the flaps and slats are retracted, the ailerons lock out,” just as you say.

    The AMM I refer to is about the -300ER as you know. The Training Manual, applicable to the -200ER, has no page such as you posted or that which I quoted from above. I attach what it says below. Please disregard the fifth page, which is irrelevant.

    The descriptions in both are hard to follow and interpret and I attach a functional diagram of the aileron PCU in case that is of interest to any brave enough to follow this discussion. Oddly the descriptions are similar in the AMM.

    I accept that you almost certainly you have it right and raise this only in case there is some difference between the two 777 models, though I think your training notes would suggest not.

    Finally, it seems to be a matter of whether the reversion to secondary beats the L2 power loss to the punch, though it will be decisive with inboard PCU failure. If the outboard PCU has already gone to blocked/damped will reversion to secondary retrospectively change that?!!!. However I intend that as a rhetorical question only….

    https://www.dropbox.com/s/2lkiqhands4gcw8/Aileron%20PCU%20states.pdf?dl=0

  577. HB says:

    @Andrew, David, Don
    FCDC Batteries performance here (www.enersys.com/WorkArea/DownloadAsset.aspx?id=25769804050)
    it is interesting to note that the perfomance is singificantly reduced if the ambient temperature is reduced. Not just a little. See the graphs.
    So in low ambient temperature (eg following depressurisation or cabin termperature control being degraded beforehand), i wonder whether these batteries will have sufficient capacity to keep the power transfer uninterrupted.

    This may have implications on ACE-R (and R-elevator blocking in undesired position) going back to @Andrew’s concern about battery depletion before APU kicks in. Saying that if APU manages to start after the R-elevator is in blocked condition, I would expect the ACE-R to reestablish power to the solenoids and hence control of the R-elevator.

    Also, somehow relevant-ish: inflight elevator power off test going bad (https://skybrary.aero/bookshelf/books/1341.pdf)

  578. Richard Cole says:

    @sk999

    >The map included with the Jan 30 Malaysian Gov’t search update report comes with an embedded coordinate grid that appears accurate. For those interested in digitizing the search zone, I’ve created a rectified version of the map here.

    This has been very useful, thanks for doing the remapping.

  579. DennisW says:

    @all

    It is customary for me to carefully analyze all the “prop” bets associated with the Superbowl. I post the two best picks here as a public service. Of course, I can/will be measured on these picks. Much more dangerous than sticking a pin in the map.

    1> PINK’s National Anthem time over 120 seconds. @-180

    2> Trump tweets during the game over 5.5. @-115

  580. HB says:

    @Andrew,
    to add more meat on the elevator blockage issue, i found in the manual that there is a function called “Elevator Offload Protection” that automatically moves the horizontal stabiliser if the elevator is not on neutral position for more than 2 sec min. The movement to the stabiliser will move the elevator to the neutral position. so if the elevator is in blocked condition, this should somehow compensate and at the same time you have the flutter protection offered by the blocked-in condition of the elevator. What is your view on this? will this function really help as per the manual saying?

  581. Paul Smithson says:

    @sk999. That depiction of the DSTG’s filtered output of the radar is very interesting. How does it compare with your earlier “digitised track”. It this slide is a mercator projection, which it looks like – then it ought to overlay on to google earth with less distortion than trying to overlay google earth image onto google earth.

  582. sk999 says:

    Paul Smithson,

    Unfortunately, once again, we have a map with no coordinate grid. I digitized it a couple of different ways using distant islands and spits of land as calibration points. In detail, the path falls about 3 nm away from my previous tracks. I shifted the new path to align as best as possible with the old ones, and it’s an OK match.

    The bottom line is that the distance traveled in going around the bend is now longer compared to the old route (speicifically, if one integrated the old route explicitly as drawn, not worrying about right-angle bends.) My first try gives 10 nm, but don’t quote it yet.

  583. Andrew says:

    @HB

    RE: ”i wonder whether these batteries will have sufficient capacity to keep the power transfer uninterrupted.”

    I’d be very surprised if the batteries were unable to provide sufficient power during power transfers, despite their degraded performance at low temperatures. I very much doubt the regulators would allow the complete loss of power to the flight control components during any reasonable failure scenario.

    The battery is capable of supporting an emergency load of 30 A for one minute, presumably at the ideal operating temperature of 23°C. The typical load is probably much less than 30 A, so the battery would either last longer than the nominal one minute while operating at the ideal temperature, or last one minute at a much lower temperature.

    Also bear in mind that most of the flight control components, including the FCDC batteries for the L & C PSAs, are located in the MEC. The electronic components in that compartment generate a lot of heat, so the temperature is unlikely to drop to an extremely low level, even if the cabin depressurises. The battery for the R PSA is located in the forward cargo compartment. That area is heated by warm air from the forward equipment ventilation system, so again, I think it’s unlikely the temperature would drop to an extremely low level.

  584. Andrew says:

    RE: ”i found in the manual that there is a function called “Elevator Offload Protection” that automatically moves the horizontal stabiliser if the elevator is not on neutral position for more than 2 sec min. The movement to the stabiliser will move the elevator to the neutral position. so if the elevator is in blocked condition, this should somehow compensate and at the same time you have the flutter protection offered by the blocked-in condition of the elevator. What is your view on this? will this function really help as per the manual saying?”

    The elevator offload function only operates when the PFCS is in normal mode. In the dual engine failur scenario, the PFCS will revert to secondary mode when electrical power is lost and the elevator offload function will be inoperative.

  585. Andrew says:

    My previous comments should have been addressed to @HB.

  586. HB says:

    @Andrew, then i can see why you are worried about that block mode. But i think it is necessary for flutter protection. Something to bring up to the manufacturer i guess.

  587. Ge Rijn says:

    @Andrew

    In all your simulation-flight shows contradicting results to the Boeing and @ALSM simulations.
    Your flight developes a bank/roll to the right which stops at 20degrees.
    Left flaperon keeps actuated on one PCU and does not move upwards.
    All essential PCUs keep actuated. APU did not start after fuel exhaust within ~5 minutes.
    The plane does not develope a spiral roll and dive but only descends in the ~5 minutes you were able to watch progess.
    It seems to me the plane would just keep descending straight on in a phugoid motion but you had not the time watch it go further.
    Which is a pity I think. 10 minutes more would have been nice I think.

    May I ask; did the plane initiate a turn to the right also when it started banking to the right?

  588. Ge Rijn says:

    @Andrew

    I mean did the plane keep its heading during the banking attitude.

  589. TBill says:

    @sk999
    The IGARI switchback turn has a funny shape that could be data quality or could be real.

    Here is my first attempt at matching the shape:
    https://docs.google.com/document/d/1M5Z_idoFCwjWYw1GCpgMsoJaCbXFLBI9L5mHoac9YME/edit?usp=sharing

  590. Victor Iannello says:

    @TBill: Did you manually set the bank limit? Or leave it in auto?

  591. Victor Iannello says:

    @Ge Rijn: In all your simulation-flight shows contradicting results to the Boeing and @ALSM simulations.

    I don’t think that’s true. When Mike simulated the dual flameout with no rudder trim, the plane slowly rolled to the right, stopping at about 20°. (For small control surface asymmetry, the roll stops at the bank angle at which the roll rate induced by the asymmetry is balanced by the dihedral effect.) At that time, the left engine restarted, which caused a steeper bank, induced by the yaw from the thrust asymmetry. Without the engine restart, my guess is the two simulations were similar.

  592. TBill says:

    @Victor
    I tried to leave in auto, partially because on PSS777 I have a hard time finding the sweet spot to click the bank angle. So the intent was auto but I’d have to repeat it to confirm.

  593. Victor Iannello says:

    @TBill: If you want to use PSS777 to find the minimum turn radius, you should choose the maximum bank angle limit of 25°. Or calculate it by hand and compare to the radar plot.

  594. Andrew says:

    @HB

    RE: ”But i think it is necessary for flutter protection.”

    I agree that blocking mode is necessary for flutter protection, but I don’t agree that it’s necessary the instant that electrical power is lost. If that were the case, why would the manufacturer allow both PCUs to initially go to bypass mode in the event of a dual PCU or ACE failure (as opposed to a loss of power) when the elevator is ‘not near to faired’? According to the AMM’s description of blocking mode

    If there is a failure with the second PCU or its ACE when the elevator is near to being faired, all PCU solenoids de-energize. When it is not near to faired, the energized blocking solenoids on each PCU stay energized.

    As soon as the airloads move the elevator panel near to faired, both blocking solenoids de-energize. At this time, a hydraulic lock in the PCUs keeps the elevator in the faired position.

    It might well be the case that both PCUs do go to blocking mode immediately upon power loss; I just don’t agree with the logic that an immediate reversion is necessary for flutter protection.

  595. Andrew says:

    @Ge Rijn

    RE: In all your simulation-flight shows contradicting results to the Boeing and @ALSM simulations.“

    The simulations that have been conducted by various parties have ALL had different results. Boeing’s first simulations all went left, but some of their more recent simulations went right. All of the simulations ALSM witnessed also went right, as did mine. The APU started in ALSM’s simulations and there was also a brief engine relight, as Victor mentioned. Neither of those occurred in Boeing’s simulations, or mine.

    Correct me if I’m wrong, but I don’t think the ATSB ascribed the left roll seen in Boeing’s simulations to the upward movement of the left flaperon.

    RE: ”…did the plane initiate a turn to the right also when it started banking to the right?”

    Yes, the aircraft entered a right turn. It did not maintain heading.

  596. Victor Iannello says:

    @Andrew said: All of the simulations ALSM witnessed also went right, as did mine.

    Mike witnessed three simulations. For two of them, right rudder trim was added, which induced a turn to the right. In the third, there was a dual flameout without rudder trim, and the plane began a slow roll to the right (before the left engine relight). I wouldn’t use the first two simulations as evidence that the simulated plane would naturally turn to the right without the rudder trim that was added.

  597. Paul Smithson says:

    @sk999. On my computer (macbook) a lat/long grid does appear very clearly at certain zoom levels. And it is a square grid – hence my comments on it being a mercator projection.

    Furthermore, I offer the following very tentative observations.

    1. I’m not sure that this really tells us where/when the turn began, only that a smooth turn has been interpolated where before there was a right angle.
    2. The Penang turn has some discontinuities. The trend of the “incoming” from the centre of the peninusular predicts passing very close to the south of Penang island (very nearly over the airport itself). The track in the immediate vicinity of S Penang (presumed from Butterwork and/or Penang) seems to cause a discontinuity – with an “outward” shift of range (or perhaps just an elevation/slant-range anomaly?). Then we get back to a pretty straight track on the “outward” towards 1802. It looks to me that forward extrapolation of “inbound” and backward extrapolation of “outbound” could be explained by a single turn, much closer to Penang. Previous trace seemed to require a “multi-stage” turn with at least 3 changes of course. This is obviously conjecture – but I think it is a conjecture that is more clearly supported by this graphic.

  598. Victor Iannello says:

    @sk999, @Paul Smithson: After inquiring, I was told by an Australian investigator that the radar data at the turn near IGARI should be viewed suspiciously, as evidenced by the unrealistic extreme dip in speed in the DSTG figure, something I have long suspected. I suspect the data was carelessly created by Malaysia to fill in holes in the radar coverage. I also suspect that most of the speed oscillations were artefacts of incorrect timestamps and filtering. I really wouldn’t put too much stock in the new radar figure. Any data originated from Malaysia has to be viewed in a very critical light because we don’t know the details and limitations of the data.

  599. sk999 says:

    Paul Smithson,

    I agree r.e. Mercator projection. Greenland is bigger than Australia – the classic, smoking-gun proof.

    Victor,

    The DSTG figure suffers from the fact that a) we don’t know the details of the algorithm used; and b) numerical integration of the speed and heading produce a route that doesn’t even match the secondary radar and ADS-B tracks immediately before IGARI. I do not think the DSTG results can be used to assess the credibility of the data that went in.

    A comparison of the Geoscience figure with the prior NTSB version may be more telling – they match pretty well (including the gap between tracks BD764 and BE144) with the exception of the newly-revealed turn. You may be correct that this part was (inartfully) added by the Malaysians or someone else, thus implying the NTSB figure should have shown a gap where, instead, it shows a segment indicating a right-angle left turn.

    One new thing we do learn from the Geoscience figure is the reason for the 18:01:49 point – track BE144 makes a sudden jump right and ends, showing that the tracker process had failed. Presumably the tracker was never able to connect together all the plots data shown on the Lido image, which is why none of these data were releasedd.

  600. David says:

    @Victor. The elephant in the room.

    With simulations apparently not supporting the timing of BFO final descent rates at the required 2 mins from MEFE, or thereabouts, search width is now shaky. The remaining possibility is that a relight could explain it but that remains hypothetical.

    If it would not we are between a rock and a hard place: the BFOs are wrong or the two minutes is. If the two minutes are wrong, the basis for the log-on is.

    To me the next step is a simulation of a 3020 lb/hr fuel supply to an engine at altitude after a double flame out and at a realistic speed, to determine for how long it would need to run to cause those BFOs and so the minimum residual fuel needed.

    That is, as a first cut leave aside the possibility of intermittent supply or relight failure.

    If simulators cannot do that (with RAT deployment and APU start/L2 ACE repowering) a Boeing analysis would help, though that would take a while.

    For discussion with a view to me, or someone, raising it with Malaysia if seen a sensible. However it may be that the time for that has now passed.

  601. Don Thompson says:

    @David,

    If one considers that intervention to the normal configuration of the overhead P5 Electrical Panel took the AES offline, and restoration of the configuration brought about the Log On at 18:25, is it not possible that a second intervention to the configuration of the P5 Electrical or Hyd/RAT Panel anticipated the fuel exhaustion conditions?

  602. David says:

    @Don. You are raising human intervention in the final log-on scenario unless I misunderstand. While what you say provides paints where he might go to do that, can you be more specific?

  603. TBill says:

    @Victor
    “@TBill: If you want to use PSS777 to find the minimum turn radius, you should choose the maximum bank angle limit of 25°. Or calculate it by hand and compare to the radar plot.”

    OK you are correct as usual…my link above is updated for 25 deg bank.

    Approx. 15.8 nm dia (normal turn) and down to 10.5 nm dia with the apparent IGARI-shape (which I used a steep climb during turn to mimic the shape).

  604. Victor Iannello says:

    @sk999: Presumably the tracker was never able to connect together all the plots data shown on the Lido image, which is why none of these data were releasedd.

    A Malaysian official recently told me that the Lido image was a composite of captures from different planes, none of which were MH370. If true, I don’t understand why the timestamps in that image match MH370 and why the image was shown to the NOK. It does raise questions about the manoeuver some of us have introduced to match the 18:25-18:28 satellite data with the radar position at 18:22.

    It’s for this reason and others that I don’t spend a lot of time any more trying to figure out the meaning of the radar data. Because we don’t know the details and the limitations of this data set, and maybe never will, there are still many questions.

  605. Victor Iannello says:

    @David said: the BFOs are wrong or the two minutes is. If the two minutes are wrong, the basis for the log-on is.

    There are other possibilities. The electrics and hydraulics might have been abnormally configured, as Don alluded to. Or, the simulations do not capture the actual system interactions after the second flameout. Or, there was pilot input. Or some combination of these three possibilities. I would go for one of these possibilities before throwing out the final BFO values.

  606. TBill says:

    @David
    Geoff Thomas’ recent article, last sentence:
    “Searchers will want to see who is in the cockpit – the captain or someone else.”

    …well maybe somebody thought about that, not sure who is in the cockpit is who had set up the controls for that final moment

  607. sk999 says:

    Victor,

    You keep referring to information provided individuals such as an unnamed “Australian investigator” and now an unnamed “Malaysian official”. Are these individuals one that you can trust to provide reliable information? Do you have more information to share? If so, please share it. Do not hide it in reserve simply to use as a counterpoint.

    If none of the aircraft in the Lido image were MH370, then the final radar point at 18:22:12 must be fiction. If so, then why would the Malaysians have positively identified it as the “Last Air Defense Radar Point” on May 1, 2014?

    Aside – A RAT 31-DL has both primary and secondary radar – right? If so, then a primary capture from some plane ought to match up with a secondary capture, thus identifying it as a known plane – i.e. not MH370. So why the confusion? Is it simply because a RAT31-DL is … Italian radar?

  608. DennisW says:

    @Victor

    Some time ago I computed the BFO error as a function of heading error. It came about to about 10Hz per degree near 38S on the 7th arc. While a degree of error is big relative to sensor update latency at typical turn rates and inertial update rates, the turn rates at 00:19 would hardly be typical (3 degrees per second). It is very conceivable that the contribution to the 00:19 BFO values due to sensor update latency is not insignificant.

  609. David says:

    @Victor. It is possible a relight could provide an answer. Currently there is no other theory on offer.

    Certainly it would be impossible to prove the descents at two minutes didn’t happen but that is not where the onus lies.

  610. TimR says:

    @DennisW “Some time ago I computed the BFO error as a function of heading error.”

    Would it be possible to compute the BFO error as a function of heading error near 10S on the 7th arc.

  611. DennisW says:

    @TimR

    Sure. I did it at 38S for obvious reasons. Tomorrow being Superbowl, I am not going to look at 10S soon.

    Just made another prop bet, BTW, that Belichick wears a hoodie at -115.

  612. Victor Iannello says:

    @sk999: The Australian investigator is very knowledgeable about the Bayesian analysis. He shared my concern about the quality of the radar data near IGARI as evidenced by the big dip in speed, although for the DSTG’s analysis, it really didn’t matter because it doesn’t effect the posterior distribution of the impact position. That’s not news because in the report it says as much. As for the Malaysian official, I believe he said what he was told by others regarding the Lido Hotel image radar data. I didn’t say I am sure that what he said was accurate. It does however provide another reason to question any of the radar data after 18:02, including the capture at 18:22. I have no idea how civil aircraft targets with transponders could be misidentified. Barring new information, I classify the Lido data as uncertain. You may choose to treat it differently for your purposes.

  613. Ge Rijn says:

    @Andrew @Victor

    On:

    “RE: In all your simulation-flight shows contradicting results to the Boeing and @ALSM simulations.“

    I was pointing to contradicting results in general: no APU start, no engine re-light, no spiral dive developing right or left. Rudder, flaperons, elevators in neutral/faired position. No initial turn to the left (most Boeing sims).

    Did not know @ALSM’s simulation with neutral rudder als stopped banking to the right at 20 degrees till the engine re-light made it turn and bank to the left. Similarity with @ Andrew’s sim is indeed remarkable in this regard.

    On:
    “Correct me if I’m wrong, but I don’t think the ATSB ascribed the left roll seen in Boeing’s simulations to the upward movement of the left flaperon.”

    No, but it has been long taken as almost fact (in discussions here and else) that the left flaperon PCUs would both go in bypass and move upwards soon after second engine flame-out.
    Initiating a left bank/turn.
    Your simulation shows this did not happen. All essential PCUs kept activated/powered without a APU-start during the ~5 minutes you were able to watch the flight after second engine flame-out.

    I would like to know if @ALSM’s similar simulation (with neutral rudder) showed the same hydraulic and electric configuration as yours after second engine flame-out (before the re-light).

  614. Victor Iannello says:

    @Ge Rijn: In Mike’s simulation, the left engine relit, causing the plane to continue to roll to the right, not left.

  615. Ge Rijn says:

    Yes ofcourse. Sorry. I confusingly had a right engine re-light in mind..
    Still I like to know what the hydraulic/electric configuration was before the re-light after second engine flame-out.
    Is this known?

  616. Andrew says:

    @Ge Rijn

    RE: “Still I like to know what the hydraulic/electric configuration was before the re-light after second engine flame-out. Is this known?”

    It’s not evident from Mikes’ videos, but perhaps he has other information.

    If you need further confirmation, have a look at the following video of a B777 simulator exercise, where both engines are deliberately shut down shortly after take-off and the pilot successfully turns back and lands on the reciprocal runway. It’s hard to see all the EICAS messages, but there’s a HYD PRESS PRI C1 caution at 0:35, followed by a HYD PRESS SYS C at 0:58. There don’t appear to be any messages related to the L or R hydraulic pumps or system pressures, suggesting the left flaperon remained powered throughout the exercise.

    https://youtube.com/watch?v=lblxOxMgtU4

  617. sk999 says:

    Victor,

    Thanks for your reply. However, it provides no useful information. I repeat – without knowledge of algorithms employed by the Bayesian Methods analysis (particularly since it could not reproduce portions of the flight path for which we have multiple independent sources of information that confirm each other), BM is uninformative regarding the quality of the input data. I challenge your “Australian Investigator” to demonstrate otherwise.

    Your Malaysian official seems to be repeating hearsay information. Can you find someone more informative?

    Not saying that your sources are wrong, but it is time to stop passing along “he said, she said” drivel.

  618. Ge Rijn says:

    @Andrew

    Thank you.
    Did you see this one too? It’s flying only on batteries and RAT with clear messages and explanation:

    https://www.youtube.com/watch?v=0egMmRivZfo

  619. DennisW says:

    @sk999

    I respect your post and point of view, and I certainly respect you. However, we are not dealing with a pure science issue anymore. The non-scientific nuances have to be weighted. Obviously, we have failed miserably at the pure science approach. We have to take these tidbits of info, weight them, and act accordingly.

  620. Ge Rijn says:

    @Andrew

    Is this sim-video accurate? Also no left-flaperon in upwards position..

  621. Mick Gilbert says:

    Seabed Constructor came back onto the board about an hour ago. As of 04/02/2018 03:57:20 UTC it was at 35° 53.151′ S 094° 16.699′ E on a course of 076 at 13.1 knots. It looks like it’s well on its way towards Fremantle, Western Australia.
    Doubtless, either or both of the Richards will have something to add.

  622. sk999 says:

    DennisW,

    As John Paul Jones allegedly once said, “I have not yet begun to fight!”

    The SS Central America was found at the edge of the 95% (2-sgima) zone. We have a long way to go before we reach that with MH370.

  623. Andrew says:

    @Ge Rijn

    RE: “Is this sim-video accurate? Also no left-flaperon in upwards position.”

    Thank you, I hadn’t seen that video. I don’t think it is accurate, because it shows several HYD PRESS SYS EICAS messages, starting with HYD PRESS SYS C and culminating in HYD PRESS SYS L+C+R. I would expect to see the HYD PRESS SYS C message for a short time until the RAT deployed, but not the others, because the windmilling engines should keep the L & R hydraulic systems pressurised. The messages probably appear in the video because in this case the PMDG software shows the failed engines with zero RPM (N1 & N2 are both 0.0). I don’t know if the PMDG software always does that for a failed engine, but it is not realistic for the fuel exhaustion case where the engines have flamed out.

  624. HB says:

    @Andrew,
    RE “I just don’t agree with the logic that an immediate reversion is necessary for flutter protection.”

    This is a very interesting point. Flutter hazard is not an immediate concern if the elevator can be damped for a short period. That period is still a time at risk despite the amplitudes being damped. From a risk point of view, i agree with you, it should be acceptable to delay the blocking mode and in the current configuration it is not delayed.

    For me the worst case is loss of hydraulic pressure (ie leakage of the tubing to the PCU) to a single PCU on the elevator. This will inherently drop the pilot pressure to the selector which will spring back to blocked mode irrespective of the solenoid status(single jeopardy case) and nothing can be done to override that, the pressure is gone. Although, the selector will briefly transit to bypass, it is not garanteed that the time in bypass will be sufficient to allow a neutral position.

    The take away here is that the manufacturer has considered that blocked mode was safe and probably not considered secondary consequences of this default position. Now if you say that the “Elevator offload protection” cannot be relied on for all the situations, the elevator defaulting to blocked mode is a serious safety concern to report.

    It maybe Boeing did not consult the pilots for that one. It is an issue to bring back to them.

  625. Andrew says:

    @Ge Rijn

    CORRECTION:

    I said “I would expect to see the HYD PRESS SYS C message for a short time until the RAT deployed, but not the others, because the windmilling engines should keep the L & R hydraulic systems pressurised.”

    The HYD PRESS SYS C message should remain after it appears because the pressure is sensed in a part of the hydraulic system that is not pressurised by the RAT. I would not expect to see the others if the engines are windmilling.

  626. DennisW says:

    @sk999

    We often get inputs from various sources that request anonymity. You have to respect that. I was simply defending Victor. He is not a secretive guy by any means, but he has to respect his sources.

  627. Andrew says:

    @HB

    RE: “From a risk point of view, i agree with you, it should be acceptable to delay the blocking mode and in the current configuration it is not delayed.”

    Thank you. That’s why I was wondering if there is some way the ACE or PCU might delay the de-energising of the blocking solenoid if electrical power to the ACE is lost. I don’t believe there’s a ‘secret’ power source as David suggested, but there might be some capacitance built into the blocking solenoid or ACE that would keep the solenoid energised for a short time, allowing the elevator to move ‘near to faired’ under air load. The solenoid would then de-energise and the elevator would be blocked. There is no mention of such a feature in the AMM, however, I don’t believe the manual addresses the power failure case. It would be interesting to find more detailed documentation about the workings of the ACE and PCU. I’ve tried searching for patent documents, but have not uncovered anything so far.

    RE: “For me the worst case is loss of hydraulic pressure (ie leakage of the tubing to the PCU) to a single PCU on the elevator. This will inherently drop the pilot pressure to the selector which will spring back to blocked mode irrespective of the solenoid status(single jeopardy case) and nothing can be done to override that, the pressure is gone.”

    There are several check valves within the PCU that should stop the loss of pilot pressure in the event of a leak in the hydraulic supply. The ACE failure monitor would de-energise the bypass solenoid and put the PCU in bypass mode as soon as the actuator stopped responding to command inputs.

    RE: “The take away here is that the manufacturer has considered that blocked mode was safe and probably not considered secondary consequences of this default position. Now if you say that the “Elevator offload protection” cannot be relied on for all the situations, the elevator defaulting to blocked mode is a serious safety concern to report.

    It maybe Boeing did not consult the pilots for that one. It is an issue to bring back to them.”

    I’d be very surprised if the loss of electrical power had not been considered by the manufacturer in the design of the PCU. My guess is there’s something in the system that’s not obvious from the manuals, or that the range of likely movement of the elevators during flight is reasonably small and that the remaining elevator provides sufficient control authority in the event that one of them becomes blocked while deflected. That might mean reduced pitch authority, but still sufficient to safely land the aircraft.

  628. David says:

    @Mick Gilbert. SC altered course? Their noon position reported 05:42 UTC, an hour ago, course 147 deg, 13 knots according to Marine Traffic. Position 35.798 S, 94.733 E, so in fact a little north of east from your position 1:45 ago as would be expected. Maybe a glitch.

  629. Richard Cole says:

    Here is the back-projection of Constructor’s course since AIS data was reenabled at 01GMT on 4th Feb. On the possibly naive assumption that the course was not changed, it seems to have been working back at its start position in the far south of the search area.

    http://goo.gl/DYYw8t

  630. Mick Gilbert says:

    @David

    That’d be a glitch I think David, I’m not seeing any change of course to 04/02/2018 07:02:10 UTC. They’re Freo bound.

  631. Richard Godfrey says:

    At the current speed of 12.8 knots, Seabed Constructor will arrive in Fremantle on 07/02/2018 at around 17:00 UTC (08/02/2018 00:00 Local Time).

  632. Richard Cole says:

    The back-projection of Constructor’s course crosses a point in the southern search area visited several times earlier. I had marked it previously as ‘Calibration Target site?’. Suspect this was the last point visited before the departure to Perth, but the reason is not understood.

    http://recole.org.uk/MH370/4-2-18b.jpg

  633. HB says:

    @Victor, Don, David
    RE: “There are other possibilities. The electrics and hydraulics might have been abnormally configured, as Don alluded to. ”

    I think it would be very unlikely for the hydraulics. But for the electrical systems, i was actually thinking the same that’s why i started investigating further the electrical system.

    From the documentations it is possible to turn off the two IDGs and open the APU breaker (as opposed to just isolating the left bus) from the overhead panel. In that case, the electrical systems will be powered from the backup generators but the IFE/Satcom/etc will be off. The IFE/Satcom/etc will be off because only one backup generator can supply power to the converter and therefore there is a limit on how much power is available to the L/R Main Buses. This is what i mentioned in a previous post. I mentioned “load shedding” which is not the correct terminology. I believe this is a simple load balance electrical interlock. it is not possible for the backup generator to supply loads to the IFE/Satcom/etc.

    Now if you agree, it is possible that the first depower of Satcom is due to this and when power was restored to IDG1/2, the APU breaker was left open or stuck. On fuel exhaustion (assuming that this is what happened), the APU did not start leading to ACE-R and ACE-L2 PCUs defaulting to their designated fail safe positions.

  634. Andrew says:

    @HB

    RE: “In that case, the electrical systems will be powered from the backup generators but the IFE/Satcom/etc will be off. The IFE/Satcom/etc will be off because only one backup generator can supply power to the converter and therefore there is a limit on how much power is available to the L/R Main Buses.”

    That is not correct. The backup generators can only power the transfer buses; they cannot power the L & R Main AC buses. If the IDGs & APU generator are selected OFF, the SATCOM is de-powered simply because it has no power supply. It has nothing to do with load shedding or ‘load balance electrical interlock’.

  635. HB says:

    @Andrew

    RE: “My guess is there’s something in the system that’s not obvious from the manuals”. On the PCU signals, i am not sure but i doubt so, a fast response (no delay) from the solenoids is necessary for for instance the case of PCUs not moving at the same speed which could cause stuctural damage. If a timer is implemented, it will have no effect in case of power losss. If a capacitor or equivalent is implemented, this could delay the bypass actuation when required.

    RE: pitch authority is not jeopardised, maybe but it may be worth bringing this up to them. I think this is a valid point to discuss with them.

  636. Ge Rijn says:

    @HB @Andrew

    Some thoughts;
    At 0:11 the SATCOM was still powered, so at least the left engine/IDG was running. Possibly the right engine was already flamed-out at this time.
    And maybe the APU was switched to OFF before 18:25 as a precaution and never switched to AUTO again.

    Did the other simulations (Boeing/@ALSM) showed an APU restart? I cann’t remember this has been mentioned in those simulations (excusse me if they did). I remember the brief left-engine restart in @ALSM’s simulation.

    I can imagine residual fuel would move to one side of the left and right main tanks when the plane started banking to 20 degrees supplying some of the fuel inlets again triggering a re-start of one of the engines (or even both) and triggering the log-on.

    Or a PIC switched the APU from OFF to AUTO somewhere after the first engine flame-out before the second engine flame-out and the APU kept running for a short while after second engine flame-out triggering the log-on.

  637. Andrew says:

    @HB

    RE: “On the PCU signals, i am not sure but i doubt so, a fast response (no delay) from the solenoids is necessary for for instance the case of PCUs not moving at the same speed which could cause stuctural damage. If a timer is implemented, it will have no effect in case of power losss. If a capacitor or equivalent is implemented, this could delay the bypass actuation when required.”

    I agree that it would not be wise to delay the action of the bypass solenoid. However, please re-read my post; I referred specifically to the blocking solenoid, not the bypass solenoid:

    “…there might be some capacitance built into the blocking solenoid or ACE that would keep the solenoid energised for a short time…”

  638. Victor Iannello says:

    @sk999: Not saying that your sources are wrong, but it is time to stop passing along “he said, she said” drivel.

    I’ll pass along information as I please. You are free to ignore it or go elsewhere. Last I checked, you are not the moderator.

    When a knowledgeable person intimately familiar with the Bayesian analysis tells me that the radar data near IGARI is questionable because the speeds are unrealistic, to me that is very relevant, and I don’t consider that “drivel”. After all, that person had direct access to the 10-second radar data supplied by Malaysia, and his comment was made to me on that basis.

    As for the Malaysian source, my point is only that the information we have about the Lido Hotel image is inconsistent. The image was presented to the NOK as radar captures of MH370. The timestamps match MH370. Yet Malaysia now does not acknowledge the image, and a Malaysian official recently said the captures were other planes. Collectively, it makes no sense. Rather than ask me for another source, I suggest that you work to find somebody in Malaysia that has more information and pass along that information.

  639. Victor Iannello says:

    @Ge Rijn said: And maybe the APU was switched to OFF before 18:25 as a precaution and never switched to AUTO again.

    The three positions for the APU switch are OFF/ON/START. When power is lost to both transfer busses, the APU will start regardless of the position of the switch.

  640. HB says:

    @Andrew, sorry to pick on the details but i want to get an accurate picture

    RE: “That is not correct. The backup generators can only power the transfer buses; they cannot power the L & R Main AC buses. ”

    I think it is correct. I summarise below my point.
    Technically, the L MAIN AC is connected to the L-XFR bus by a tie-in breaker (L TBB) which is normally closed and i think will remain closed all the time unless there is an electrical fault. Without any interlock, power will go up to the L MAIN AC and then to the Utility bus. If it does, the loads on the utility bus will draw additional current and there is a risk of voltage drop and eventually total black out. To prevent this, there must be an interlock to open the breaker L UB between the L-Main and the utility bus or to isolate the specific loads on the utility bus that are not necessary (these include the Satcom, etc.). I expect this to be part of the load transfer logic to start the backup generator.
    This logic is described in (http://ads-b.ca/777/B777-Electrical.pdf under the Backup Generator section).
    The implication is that whatever loads on the L-Bus which are not cuts will remain powered. i don’t know what else in from that bus.
    Ie some breakers have to be opened automatically, the question is which ones in particular? L TBB, L UB or the individual loads on the utility bus?

    You can refer to the simplified single line diagram in this link (https://www.scribd.com/doc/105499684/B777-Quick-Reference).

  641. Ge Rijn says:

    @HB @Andrew

    On my comment;

    “Or a PIC switched the APU from OFF to AUTO somewhere after the first engine flame-out before the second engine flame-out and the APU kept running for a short while after second engine flame-out triggering the log-on.”

    Thinking twice that ofcourse would not have triggered the final log-on for power would be uninterupted when the APU was already running at second-engine flame-out.
    Considering @Andrew’s simulation that would only leave a brief engine re-start left or right (or both) as cause of the final log-on?

    Or a unknown intervention from a PIC? But then the I assume a 20 degrees banking/roll would not have occured cause the PIC would have kept the plane level. It’s mind-boggling to me.

    But anyway, what happened before the 0:19 log-on caused by a PIC or something else, the log-on is a fact and it indicates a steep descent near the 7th arc.
    It might be impossible to prove a PIC caused the log-on unless there is no other possibility?

  642. Andrew says:

    @HB

    RE: Technically, the L MAIN AC is connected to the L-XFR bus by a tie-in breaker (L TBB) which is normally closed and i think will remain closed all the time unless there is an electrical fault. Without any interlock, power will go up to the L MAIN AC and then to the Utility bus.“

    The following is taken from the AMM:

    The backup generator converter controls the automatic operation of the TBB and CCB. The converter gets a switch position signal from the backup generator (BACKUP GEN) switch.

    When the switch is in, the converter controls the TBB and CCB. The TBB and CCB are normally not energized. A spring holds the TBB closed when it is not energized. A spring holds the CCB open when it is not energized.

    The converter energizes the TBB and CCB to permit the backup system to supply power to the transfer buses. The TBB opens when it energizes. The CCB closes when it energizes.

    When the switch is out, the TBB and CCB are relaxed.

    Power Transfer

    The left transfer bus normally gets power from the left IDG, through the left GCB and left TBB. When the converter senses no power at the left main ac bus, it opens the left TBB and closes the left CCB. This permits power to go from the converter to the left transfer bus.

    If necessary, the converter supplies power to the right transfer bus through the right CCB.

    If the backup generator system is powering one or both transfer buses, the respective TBBs are automatically opened, separating the transfer bus(es) from the main AC bus(es). The backup system does NOT EVER power the main AC buses.

  643. Don Thompson says:

    @HB:

    The Backup Generator Convertor controls the TTB (Transfer Bus Breaker – Main to Transfer bus) and the CCB (Converter Circuit Breaker – Backup Generator Converter to Transfer bus) and does so by sensing the load side of the main (IDG) Generator Circuit Breaker.

    Supply to the transfer bus is exclusively one, or the other, source: no chance of ‘back feed’ from the Transfer Bus up to the corresponding Main AC Bus. Well… very, very, very unlikely: the de-energised states of the TBB and CCB are opposite with TBB closed and CCB open.

    SATCOM is not supplied off the L Utility Bus, it’s the secondary branch of L Main AC Bus.

  644. Don Thompson says:

    @Andrew,

    Simultaneous posts, independent, no collaboration behind the scenes.

    @HB considers that an intervention to the normal state of the Hydraulics systems, prior to MEFE, would be “unlikely“.

    No less likely than a similar intervention to the Electrics systems, as is most likely during the period 17:08 through 18:25.

    Whatever fits the 00:19:29 and 00:19:38 BFO records is most likely.

  645. HB says:

    @ Andrew, Don,
    Right, that clarifies which breaker is opened. So in the scenario of both IDGs isolated and APU breaker opened from the overhead panel, there will be no power to the
    * SATCOM
    * TCAS
    * Radio HF and datalink are load shed
    * centre tank override / jettison pump (center tank and auxiliary fuel cannot be used)
    * external lights
    * cabin temperature control (degraded)
    * etc.

    And based on what you say all the loads are are supplied directly from the Main buses (and sub buses other than the transfer bus).

    I am questioning effectively whether the plane could have flown under backup power after the first loss of communication. Then before the first relogon, power to the IDGs could have been established. For the last BFOs, if the APU breaker was not properly closed, it is credible to have a case of both engine failure with no power from APU.

    i am not sure how long can a plane fly under backup power generator but this may be another scenario than the postulated Left Bus isolation scenario with the right IDG providing power.

    PS: i am not saying power will go back up. That was to illustrate thate there is an electrical interlock.

  646. Victor Iannello says:

    One scenario I have considered is that at 17:21, three switches were operated: L BUS TIE (AUTO to ISLN), L GEN CTRL (ON to OFF), and BACKUP GEN L (ON to OFF). However, at 18:22, only the L BUS TIE was switched back to its normal position (AUTO). During the period 17:21 to 18:22, the L MAIN bus would be de-powered, which of course explains the offline behavior of the SATCOM, and allows normal operation of the A/P as the L XFR and R XFR busses remain powered. At 18:22, when the L BUS TIE is set to AUTO, the L MAIN bus would be powered by the R MAIN bus, and the SATCOM begins to start-up.

    Suppose the L GEN CTRL and BACKUP GEN L switches remain OFF, and at around 00:17, the right engine flames out with no fuel remaining in the right tank. Now you have a condition where there is a small amount of fuel remaining in the left tank, and the left engine is still running and producing thrust, but the L MAIN and R MAIN busses are de-powered, along with the L XFR and R XFR busses. With no pitot heat, the flight control mode changes to SECONDARY. After a minute, the APU starts, drawing fuel from the remaining fuel in the left tank, the L MAIN, R MAIN, L XFR, and R XFR busses are re-powered, and the SATCOM starts to power-up. However, the flight control mode remains in SECONDARY. With the left engine producing thrust and the flight control mode in SECONDARY, there is no TAC to move the rudder to compensate for the thrust asymmetry of only the left engine running. (Even with TAC, the yaw correction is only partial.) The thrust asymmetry will cause yaw, which will induce a banked descent to the right, and the start-up of the APU will not help things because the flight control mode will remain in SECONDARY. This could easily cause a 0.65g downward acceleration with no pilot input and within the two-minute time window of the flame out of the right engine. In fact, my main reservation about this scenario is that the steep descent rates should be reached earlier than two minutes after power is lost to the L MAIN bus.

  647. Victor Iannello says:

    @Andrew: Is it possible to simulate the scenario I outline above?

  648. Tim says:

    @victor,
    The scenario you’ve described above is one that needs to be considered as a plausible possibility. As someone who does not support any nefarious human input— I prefer the crew O2 bottle rupture scenario, the immediate failure of the left bus and it’s associated switches fits well as they are all in close proximity to the crew O2 bottles.

    But I need to find a reason for the L BUS TIE to automatically connect at 1822z—-anyone got any ideas?

  649. Victor Iannello says:

    @Tim: There is no way for the state of the L BUS TIE switch to automatically change from ISLN to AUTO. It requires manual intervention.

  650. el_gato says:

    @Victor Iannello

    You said: “The image was presented to the NOK as radar captures of MH370.”

    Do we know this for sure? I was not aware, might have missed it.

    Thank you for sharing the info, it’s an important issue IMO.

  651. Victor Iannello says:

    @el_gato: On March 21, 2014, at the Lido Hotel in Beijing, Malaysian military officials presented the radar image to Chinese families with the explanation that MH370 at 18:02 UTC was captured by radar over Pulau Perak and was last captured at 18:22 UTC. Here is a translated report that appeared in China.

  652. Rob says:

    @VictorI

    Remember that a backup generator can only power the transfer busses, not the main AC busses, The AC busses power the transfer busses, but not vise versa. So what would be the point in disconnecting the backup generator And you would need a plausible explanation as to why the bus tie and Left IDG were disconnected when the aircraft went dark and changed course, and reconnected when it coincidentally flew out of radar range. Your scenario just seems a bit contrived.

  653. HB says:

    @Victor,
    To me, technically feasible (on the electrical side). There is no reason to switch off the backup generator. A possible scenario is that this L backup generator failed to start on demand which is not unusual for a back up generator.
    What about the RAT?
    I don’t quite understand the ‘no pilot heat’ to Secondary. How specifically secondary is triggered?

  654. Victor Iannello says:

    @Rob: We believe the satellite data suggests the left bus was isolated and later re-connected. I am simply suggesting another sequence of left bus switching that accomplishes the same thing.

    If, for instance, there was a desire to power down the SATCOM and the CVR, those three (L BUS TIE, L GEN CTRL, and BACKUP GEN L) switches would need to be operated. Meanwhile, restoring power to the left main bus could be achieved by switching the L BUS TIE from ISLN to AUTO (although there will be load shedding in the L UTIL bus).

  655. Victor Iannello says:

    @HB: Switching off the left backup generator could be related to the CVR. I don’t think the RAT would unlock if the right transfer bus remained powered.

  656. Ge Rijn says:

    @VictorI

    Oke. So the APU will start anyway regardless if the APU-switch was in OFF position and there was fuel (becoming) available after second engine flame-out to the APU.
    But then this fuel could possibly only have become available later after the second engine flame-out (not directly, causing the start-up of the APU after one minute) when the plane moved to its bank of 20 degrees and residual fuel was moved to the right side of the left main tank becoming available to the APU fuel inlet.

    @Andrew’s simulation does not show an APU-start or engine re-light before or within the ~5 minutes he watched after second engine flame-out.
    Maybe this could have happened (just) after this ~5 minutes?

  657. Tim says:

    @victor: I’m assuming the L BUS TIE switch is still in the auto position. For some reason the L MAIN BUS was initially isolated due to arcing/shorting/ voltage fluctuations. Then at 1822z the L BUS TIE Automatically reconnected the L MAIN BUS when conditions had settled down.

    I’m just not sure of the automatic function of the L BUS TIE. Does anyone have any references from the AMM?

  658. Don Thompson says:

    @HB wrote “L backup generator failed to start on demand which is not unusual for a back up generator.

    The Backup Generators, one on each engine, supply the Backup Generator Converter, which in turn supplies the Transfer Buses. Failure of the BGC would require both Backup Generators to ‘fail to start’.

  659. Victor Iannello says:

    @Tim: Sorry, I can’t help you develop your theory, which I believe is very unlikely. You’re on your own.

  660. Victor Iannello says:

    @Ge Rijn: If you are asking me about the scenario I proposed, you have misunderstood the basic premise. I hypothesized that the right engine, drawing from the right tank, exhausted its fuel, while the left engine continued to run with the remaining fuel in the left tank. Like the left engine, the APU drew fuel from the left tank. The APU automatically started after the right engine flamed out because the left and right transfer busses went dead.

  661. HB says:

    @Victor,
    on right engine shutdown? xbuses will be down and RAT will deploy right? any effect of this? will the mode revert to Secondary still?

  662. Victor Iannello says:

    @HB: 1) The RAT doesn’t power the transfer busses, and there is delay for the RAT to be unlocked. 2) Even if somehow the transfer busses were re-powered, e.g., from the APU (via the main busses), once the flight control mode degrades to SECONDARY, it is necessary to cycle the PRIMARY FLIGHT COMPUTERS switch from AUTO to DISC to AUTO to return to NORMAL mode.

  663. Brian Anderson says:

    @Victor,

    That’s an interesting sequence of possible events.
    Based on various analyses of fuel burn and the difference in fuel flow [PDA] to the right and left engines, I had been of the view that the right engine failed first at around, or even before 00:11. Then the left engine continued to run for perhaps another 6 minutes. If the TAC did not reasonably compensate for the asymmetric thrust for that time then I would be surprised if the aircraft had managed to remain in anything close to level flight for that duration.

    My references describe the TAC operation in various engine failure scenarios. In a number of scenarios the TAC still leaves some yaw adjustment to be made by the PIC, primarily to assist the pilot in identifying which engine has failed. However, this quote seems to apply in the cruise condition . . .”To provide an indication of the engine failure, TAC does not initially fully compensate for the yaw. Within a few seconds, sufficient rudder is input by TAC to center the control wheel”.

  664. Victor Iannello says:

    @Brian Anderson: We don’t know exactly how the plane was flown from 18:22 to 19:41, and it is therefore difficult to know exactly when fuel exhaustion occurred. In the scenario I proposed for consideration, the right engine flamed out at 00:17. The left engine might not have flamed out until impact. If the left IDG and backup generator were isolated, there would be no TAC after the first engine flamed out, and yaw-induced bank would occur. Having both engines running until 00:17, and then a possible turn under power of one engine, could be why we see relatively high average groundspeed after 00:11.

  665. Andrew says:

    @Victor

    RE: ”Is it possible to simulate the scenario I outline above?”

    Yes, it is possible to simulate the scenario you described. The L autothrottle servo motor would also lose power as soon as the right engine failed, so the left engine thrust would not increase to compensate for the loss of thrust. Given the aircraft’s light weight at that point, the thrust setting of the left engine would be relatively low, so the aircraft might not ‘depart’ quite so quickly. Even if it did, perhaps that might just mean the rates of descent were at the higher end of those suggested by the BFOs.

  666. Andrew says:

    @Ge Rijn

    RE: “@Andrew’s simulation does not show an APU-start or engine re-light before or within the ~5 minutes he watched after second engine flame-out.
    Maybe this could have happened (just) after this ~5 minutes?”

    Many, if not most, simulators don’t model the residual fuel in the tanks after fuel exhaustion. In those simulators, the APU will not start after the fuel quantity indicates zero fuel remaining. The United simulator (witnessed by ALSM) appears to be an exception.

  667. Andrew says:

    @HB

    RE: “I don’t quite understand the ‘no pilot heat’ to Secondary. How specifically secondary is triggered?”

    The pitot heat system is supplied with 115V AC power by the L & R transfer buses. If that power is lost (as it would be in Victor’s scenario), the primary flight control system (PFCS) automatically reverts to secondary mode. The PFCS does that because the air data from the pitot system is considered unreliable due to the possibility of icing.

    If power is subsequently restored, the PFCS will remain in secondary mode until the PRIMARY FLIGHT COMPUTERS Disconnect Switch on the overhead panel is manually cycled to DISC and then back to AUTO.

  668. Don Thompson says:

    Victor has proposed an intervention, prior to fuel exhaustion, that involves the aircraft electrical system.

    I renewed the discussion of hydraulics system effects with an idea that the electrical system function might have been restored at 18:22 (or thereabouts) so as to initiate the Log On at 18:25 and that an intervention might have then been made to the hydraulics systems so as to ensure a predictable ‘upset’ at fuel exhaustion.

    If the L & R ‘Primary’ and Demand pump selectors are set to ‘OFF’, those systems lose pressure, regardless of N3 shaft speed for Power Take Off to the accessory gearbox. However, the Centre system continues to function with its AC Motor Pumps (& potentially Air Driven Motor Pumps) to supply hydraulic power to the PFCS surfaces.

    If the L & R Hydraulic systems are inop & without pressure the L flaperon will fair, invoking a left wing down roll. While the AFDS is engaged, it will compensate for that L flaperon effect. Until Fuel Exhaustion and the AFDS disengages.

    So, two scenarios will accelerate an ‘upset’ at fuel exhaustion: one in the electrical system; the other in the hydraulic system.

    Timing is the key.

  669. David says:

    @Victor. I believe this has been simulated. The ATSB in its Search and Debris Examination Update of 2nd November, 2016 describes how, “In an electrical configuration where the loss of engine power from one engine resulted in the
    loss of autopilot (AP), the aircraft descended in both clockwise and anti-clockwise directions.”

    In those which are orange and white at fig 6, page 7 of that report, the right turners, by my rudimentary measure the turn radius at log-on is about 18 km, 11 NM, which by my estimation (earlier post) indicates insufficient angle of bank to result in the required downwards acceleration.

    Since this simulation apparently entailed no relight/residual fuel uncertainties (thought the ATSB has it in mind that the right could relight) it may be a fair representation.

  670. Don Thompson says:

    @David

    Thank you for the reminder and pointer to ATSB’s findings.

  671. Rob says:

    @David

    IRO the subject of angles of bank, would you happen to know what angle would be required to produce a turn radius of 11.5nm, in level flight TAS 486kts, FL350 (neglecting wind)?

  672. David says:

    Corrections to my last: page 13 vice 7, ‘though’ in place of ‘thought’.

  673. Rob says:

    @David

    Re my previous post on angle of bank, sorry, I should have also said at 208 tonnes weight.

    Thanks

    Rob.

    Got to turn in now

  674. David says:

    @Rob. In time I hope. Weight and FL make no difference. Subject to horizontal acceleration only (0.3 g), 16.7 deg.

  675. TBill says:

    @Rob
    “…turn radius of 11.5nm, in level flight TAS 486kts”

    FWIW I reported yesterday (link above) in FS9 I saw about 15.8nm dia @ 25 deg bank angle at about 500 TAS, so you’re talking a tight turn.

  676. Mick Gilbert says:

    @Rob

    Re AOB and turn radius I’m pretty sure that
    Radius of turn = V²/g*(Tan AOB)
    so I’m thinking that 16.7° AOB should get you there or thereabouts. The rate of turn will be about 0.67 deg/sec (~40’/sec).

  677. ventus45 says:

    @Victor,

    A long time ago now, back in mid 2017, Malaysia announced that the “Final” report had been written, and promised that it was going to be released late in 2017.

    True to form, (when it comes to Malaysian promises), surprise surprise, that did not happen.

    No one (it seems) has made any comment on this fact (as far as I am aware).

    It would seem that the Ocean Infinity odyssey has totally consumed everyone’s attention since mid 2017.

    It is now 00:20 UTC 5th February 2018, precisely 31 days to the 4th anniversary of the loss of MH370.

    Are we to expect the final on that date, another obfuscated interim, or nothing ?

    A penny for your thoughts.

  678. Nederland says:

    @ventus45

    The final report would need to be published within a year of completion of the search. As it stands, the search is ongoing. If the final report is published, and the wreckage subsequently found, that would be very strange.

  679. Andrew says:

    @Tim

    RE: “I’m just not sure of the automatic function of the L BUS TIE. Does anyone have any references from the AMM?”

    If the Bus Tie switches are selected to AUTO, the position of the bus tie breakers (BTB) is controlled automatically by the generator control units (GCU). If both IDGs are operating normally, the BTBs are open and the buses are isolated from each other. If one IDG trips or is selected OFF, the IDG is disconnected from the bus by the associated GCU. The GCU also closes the on-side BTB, allowing the associated AC bus to be powered by the other side. The GCU will not re-connect the IDG unless the generator control switch is cycled (or selected ON if it was OFF). A bus fault might also cause the GCU to disconnect the IDG from the bus. In that case, the GCU will not close the BTB and the bus would be isolated. Once isolated, the BTB will not close unless the fault is removed, the IDG remains off-line and the BUS TIE switch is cycled.

  680. TBill says:

    @Rob @Mick
    OK yes diameter vs. radius
    I saw about 20 nm dia at probably about 20 deg Turn, 10 nm dia seems directionally consistent with Mick’s calc

  681. TBill says:

    Correction above 10nm radius

  682. Andrew says:

    @ventus45
    @Nederland

    ICAO Annex 13 only stipulates that the Final Report be released “as soon as possible and, if possible, within twelve months”. Annex 13 also states: “If the report cannot be made publicly available within twelve months, the State conducting shall make an interim statement publicly available on each anniversary of the occurrence, detailing the progress of the investigation and any safety issues raised.” So far, Malaysia has complied with its obligations in that respect, with the issue of Interim Statements on each anniversary of the aircraft’s disappearance. I would expect another Interim Statement to be released on 8 March this year

    I assume the Malaysians had prepared a Final Report after the search was abandoned in early 2017, on the expectation that it would not be resumed. However, publication of that report was probably delayed after Ocean Infinity began making overtures about resuming the search. I expect the Malaysians will release another Interim Statement on 8 March this year, for the following reasons:

    1. If Ocean Infinity discovers wreckage by 8 March, publication of the Final Report will likely be delayed until the wreckage is recovered and further analysis conducted.

    2. If Ocean Infinity does not find anything by 8 March, publication of the Final Report will likely be delayed until after the search is completed.

  683. Victor Iannello says:

    @David: Yes, I am aware of the Boeing simulation. I still think it is worthwhile to conduct a new simulation so that we are sure of the exact conditions and can get a description of what occurred, as we know little about the Boeing simulations. For instance, it would be helpful to understand why the turn changed directions, if that does occur in the new simulation. Also, if a scenario with a single engine and no TAC does not produce a banked descent with a high downward acceleration, then the engine restart scenario is not likely to produce a high downward acceleration.

  684. Brian Anderson says:

    @Victor,

    The earlier fuel burn calculations and later information on the assumed PDA obtained by Dr B both seem to support the view that the right engine would fail first, and that the left engine would continue to run for something like 6 minutes, on about 300Kg of fuel.

    The precise timings of the engine failures can be debated, but the time interval between them seems pretty solid. That is a long time to run without any TAC, unless there is someone up front managing the rudder pedals.

    Where does the . . ” two-minute time window of the flame out of the right engine” . . . in your scenario come from?

  685. David says:

    @Victor. “…it would be helpful to understand why the turn changed directions, if that does occur in the new simulation.”

    I think that their two directions are right and left engine failures the other still running as distinct from a descent reversing.

    “..if a scenario with a single engine and no TAC does not produce a banked descent with a high downward acceleration, then the engine restart scenario is not likely to produce a high downward acceleration.”

    Based on that Boeing simulation, it wouldn’t.

    Nice to have that confirmed as I expect it would be. If not, fuel flow intermittency, restriction, and relight uncertainties would need consideration. With the first being beyond simulation or analysis, that would all just be left dangling.

  686. HB says:

    @Andrew, thanks, just to understand the timing sequence.
    Is the detection based on electrical relay sensing power or on temperature sensor? and could you confirm if there is a timer/delay logic? or the action is immediate, ie before APU has a chance to start?
    Not related to this scenario but also if you could confirm if there is a logic such as 2oo3 or any sensor will trigger secondary?

  687. Richard Godfrey says:

    On 5th February 2018 at 06:50:52 UTC Seabed Constructor was at 34.5971°S 101.2324°E on a course of 076°T at 12.0 knots.

    At the current speed of 12.0 knots, Seabed Constructor will arrive in Fremantle on 07/02/2018 at around 21:00 UTC (08/02/2018 04:00 Local Time).

  688. Andrew says:

    @HB

    RE: “Is the detection based on electrical relay sensing power or on temperature sensor?”

    Detection is based on a relay sensing the power supply to its associated pitot probe. If the power fails, the relay supplies a ‘failed’ signal to the respective pitot air data module (ADM). The ADM sends the signal to the primary flight computers (PFCs) and various other systems via the flight controls ARINC 629 bus. The PFCs revert to secondary mode on receipt of the failed signal.

    RE: “…could you confirm if there is a timer/delay logic? or the action is immediate, ie before APU has a chance to start?”

    The action is immediate.

    RE: “…if you could confirm if there is a logic such as 2oo3 or any sensor will trigger secondary?”

    The PFCs only revert to secondary mode when all three sensors (ie L, C & R) indicate ‘failed’.

  689. Ge Rijn says:

    @VictorI

    On your comment to me:
    “If you are asking me about the scenario I proposed..”
    No, I was replying to an earlier comment from you (about the APU-switch).

    I’m still thonking about your interesting simulation-proposal.

  690. Andrew says:

    @Richard Godfrey

    RE: “Seabed Constructor will arrive in Fremantle on 07/02/2018 at around 21:00 UTC (08/02/2018 04:00 Local Time).”

    Make that 05:00 local time! Perth is UTC+8.

  691. Ge Rijn says:

    ..still ‘thinking’ not ‘thonking’.. I’m not that smart as you all know

  692. Richard Godfrey says:

    Many thanks Andrew, for the correction to local time in Perth.

  693. Rob says:

    @David
    @TBill
    @Mick Gilbert

    Thank you guys for the AOB information. Much appreciated.

  694. Victor Iannello says:

    Brian asked: Where does the . . ” two-minute time window of the flame out of the right engine” . . . in your scenario come from?

    The two minutes is the time between the flame out of the right engine (hypothetically at 00:17 UTC) and the SATCOM log-on (at 00:19 UTC). In the scenario under consideration, power is lost to the main and transfer busses after the flame out of the first (right) engine, not the second (left) engine. The two minutes it is not the time difference between the flame out of the two engines, which as you say, should be greater.

  695. Victor Iannello says:

    @David said: “..if a scenario with a single engine and no TAC does not produce a banked descent with a high downward acceleration, then the engine restart scenario is not likely to produce a high downward acceleration.”
    Based on that Boeing simulation, it wouldn’t.

    Based on Mike Exner’s simulation, it would. The (failed) restart of the left engine caused an extreme bank.

  696. sinux says:

    @Don Thompson
    Re pre-emptive action to ensure consistent EOF outcome.

    Wouldn’t it be easier for the PIC to :
    – disconnect the autopilot
    – trim the aircraft in a banking/pitching up or down/etc…(pick what ever you like) attitude
    – reconnect the autopilot

    The plane flies straight and level while on autopilot.
    When it reverts to secondary, the plane is sent in steep spiral/descent/etc due to the trim set before.

  697. Tim says:

    @Andrew: Many Thanks re BTB operation.

    So if the L MAIN BUS had been isolated due to a fault, it would take manual intervention to cycle the BTB to bring the L MAIN BUS online. Thanks.

  698. Victor Iannello says:

    @sinux: Whatever aileron or stabilizer trim is added will be cancelled by the A/P when it is engaged. However, the rudder trim, which can be added with the A/P engaged, will persist. If rudder trim was added so the plane is mistrimmed, the A/P would add opposing aileron action, and the plane will fly less efficiently due to the yaw and bank. When the A/P disengages, the aileron input added by the A/P will be removed, the rudder trim will remain, and the plane will experience yaw-induced bank, which could be significant if the mistrim is large. So accepting the higher fuel burn, adding rudder trim is a way to induce a steep descent at the end-of-flight.

  699. HB says:

    @Victor, from fuel starvation to loss of power, there will be a few seconds due to spinnng inertia. Remember the power source is an IDG. In that time the auto pilot and TAC will be active for a few more seconds. On loss of xbus the hydraulics pcus except those controlled by ACE R and ACE L2 will be maintained in fail last position for some time as discussed above.

  700. Ge Rijn says:

    Perhaps an interesting link to others too on the PMGs and overall B777 electric power supply.
    The article suggests (to me) the PMGs will maintain to provide power in case of “..short of deploying the RAT that should occur only in the direst emergency”:

    https://books.google.nl/books?id=8XFwAAAAQBAJ&pg=SA7-PA18&lpg=SA7-PA18&dq=pmg+b777&source=bl&ots=BV6SXjvIWY&sig=gmLBvLHOYvL11tGQO7_3gtrKpNs&hl=nl&sa=X&ved=0ahUKEwiw6tbQuo7ZAhUJB8AKHaZBAb0Q6AEISDAH#v=onepage&q=pmg%20b777&f=false

  701. Victor Iannello says:

    @HB: I believe that when the transfer busses are unpowered and the flight mode degrades from NORMAL to SECONDARY, any input from the TAC goes to zero (but rudder trim persists). If you are implying that the TAC input remains during the period of time that one engine is operating and the main and transfer busses are down, I don’t think that’s true, but I will yield to those that have studied this more closely than me.

  702. Victor Iannello says:

    Read this only for its entertainment value.

  703. Mick Gilbert says:

    @Victor

    Thanks for that. Entertainment value indeed. Did you notice that the story is filed under ‘Travel – Updates – Incidents’.

  704. Ge Rijn says:

    @VictorI

    Fantastic media attention! Keep them going! It keeps MH370 under international attention which is important whatever they speculate or say.

    Probably the crew just only needed a break after weeks at see..

  705. TBill says:

    @sinux
    …That’s what I am thinking too, philosophically.

  706. HB says:

    @victor, yes that is what i was thinking. The control will try to compensate and overshoot the desired rudder position then fail last. Potentially explaining the clockwise and then counterclockwise issue. Saying that i m not sure whether a simulator could be that sophisticated to model such transients.

  707. Victor Iannello says:

    @Ge Rijn: At this point, either OI will find the debris or it will not, regardless of unfounded accusations leveled at OI, which I do not think help solve the mystery.

  708. Don Thompson says:

    @Ge Rijn wrote ‘The article suggests (to me) the PMGs will maintain to provide power in case of “..short of deploying the RAT that should occur only in the direst emergency”

    Incorrect, the excerpt reads “All AC generators include a PMG to bootstrap the excitation system. […] the B777, there are a total of 13 PMGs/PMAs…”

    We have accounted for the 13 PMGs : 1 on APU generator (as excitation), 6 on each engine (1 for IDG as excitor source, 1 for backup generator as excitor source, 2 for FADEC as 28V AC supply, 2 for FCDC PSAs as 28V AC supply).

    The PMGs only supply minor loads. The RAT electrical generator provides 7.5kVA that is absolutely necessary as a Standby Power source for the C1 and C2 TRUs that provide 28V DC.

    However, Moir’s book does ignore the RAT’s 2 PMGs (for excitation and the RAT GCU).

  709. Ge Rijn says:

    @VictorI

    Completely agree. On the other hand the supporting community still has opportunities to influence the strategy of OI while it’s still in its first stages.
    I think that’s why the discussion here is still very important and can still have influence on the ungoing search.
    Flexibility is key to succes regarding OI I think.
    If they are willing to adapt there strategy on the way when new evidence or very plausable scenarios show up, their chance of succes would increase.
    In a way I trust OI will take this approuch.

  710. Victor Iannello says:

    @Ge Rijn: OI’s motivation is primarily economic, not political. They will use whatever information they believe will help them succeed.

  711. Ge Rijn says:

    @Don Thompson

    I still did not find proof the PMGs provide sufficent power to the ACEs and PCUs under windmilling.
    But the quote in the article suggests to me they do (when the RAT won’t deploy).
    In @Andrew’s simulation power to essential ACEs and PCUs kept maintained for ~5 minutes at least after second engine flame-out (till @Andrew had to quite watching) without the APU starting.

    Do you have proof the PMGs on a B777 are not capable of sustaining power to the essential ACEs and PCUs after second engine flame-out?

  712. Don Thompson says:

    @Ge Rijn,

    Perhaps I misunderstood your point, ‘The article suggests (to me) the PMGs will maintain to provide power in case of “..short of deploying the RAT that should occur only in the direst emergency”

    I read the quote, from you above, as implying the PMGs may be regarded as an alternative to the RAT and its supply the C1 & C2 TRUs. That is certainly not the case. The RAT deploys automatically when the L and R Transfer Buses lose power for more than 15s so as to restore supply to the C1 & C2 TRUs.

  713. Ge Rijn says:

    @Don Thompson

    RAT would only deploy in case of all engines flame-out I assume.
    The article suggests (to me) that even in the very unlikely case the RAT does not deploy the PMGs would deliver sufficient power to the primary flight controls. Perhaps I misunderstand the quote.
    But you did not answer my question.
    Do you have proof the PMGs are not capable of sustaining power to the essential ACEs and PCUs after second flame-out (under windmilling)?

  714. Don Thompson says:

    @Ge Rijn

    In the case of the electrical system, the RAT deploys if the Transfer Buses lose power for at least 15 seconds.

    To your specific question: I have no emphatic proof that a FCDC PSA will (or won’t) sustain supply to its PFC, ACE, and PCUs if its supplied by only one PMG and that PMG is being driven by a windmilling engine. Look back through the entire set of comments above & there is no emphatic answer to that.

    It’s been stated that the PMGs produce 950W, at 28V DC that is approximately 33A (PF unknown). Andrew stated above that the FCDC Battery is capable of delivering 30A for one minute. Assume the PSA load, comprising the PFC, ACE, and PCUs, is in the order of 30A.

    I cannot be sure that a PMG, spinning at low revs, would sustain its full load.

    However, regardless of what the ‘article’ suggests, in the event that the aircraft is running on ‘standby power’, that is using Main Battery or RAT as source, the concern for sustaining FCDC power relates only to the Right FCDC PSA.

    Note that the C FCDC PSA is supplied by PMGs in each engine whereas the L & R PSAs are supplied by only one ‘onside’ PMG. That is possibly a clue.

  715. David says:

    @Victor. About engine yaw asymmetry I do not think the ALSM experience counters the Boeing. After log-on the Boeing in its right turns and some left tightened too but also not in time.

    However that remains a small sample.

    There are phugoids also to consider but again apparently too late in development.

    A speculative deduction from the most recent Boeing simulations is that previous simulations (no relights, no engine thrust imbalance after A/P disengagement) were inconsistent with the magnitude of the final BFO descents, or at least inconclusive, and hence the ordering of the investigation of different scenarios; and then the First Principles Review’s enthusiasm for the outcome.

    In other words the extra simulations were not tidying up of loose ends and that scenario is central.

  716. Victor Iannello says:

    @David: No, the scenario that Mike observed with the relight is very different from the Boeing simulation. In Mike’s simulation, the roll rate was high due to the relight, and the plane would have entered in a spiral dive except that there was an unexplained abrupt correction that leveled the wings. That said, we are trying to understand complex system dynamics with a limited understanding of the systems, a limited understanding of how well each simulator models those systems, and a limited understanding of what actually occurred in each simulation. There is also the possibility of pilot input. If we accept the BFO values are valid, the plane impacted close to the 7th arc, barring an extraordinary recovery by the pilot. Unless somebody has access to a Level D simulator and is willing to conduct careful experiments and record and analyze the results, it will be difficult to carry this much further.

  717. David says:

    @Victor. OK, thanks.

  718. Andrew says:

    @HB

    RE: “…from fuel starvation to loss of power, there will be a few seconds due to spinnng inertia. Remember the power source is an IDG. In that time the auto pilot and TAC will be active for a few more seconds. On loss of xbus the hydraulics pcus except those controlled by ACE R and ACE L2 will be maintained in fail last position for some time as discussed above.”

    I agree with the first of your comments above; the IDG(s) will continue producing power for about five seconds as the engines run down after they fail. The primary flight control system (PFCS) will remain in normal mode and the AP and TAC will continue to function normally during that time.

    However, I’m not sure what you’re implying in your second comment, or in your follow up comment “The control will try to compensate and overshoot the desired rudder position then fail last.”

    The rudder is powered by three PCUs. The PCU that is supplied by the C hydraulic system might go to bypass mode if the C hydraulic system pressure drops before the RAT deploys, but the other PCUs will remain in normal mode, powered by the L & R hydraulic systems and the L1 & R ACEs respectively. It will not ‘fail’.

    The TAC will continue to operate during the few seconds between fuel exhaustion and loss of electrical power. When electrical power is lost, the PFCS will revert to secondary mode and the AP and TAC will disengage. At that point, the TAC input will be removed and the PFCS will drive the rudder back to its trimmed position, as Victor mentioned. The rudder remains powered throughout.

  719. HB says:

    @Don, Ge Rijn,

    RE: “I have no emphatic proof that a FCDC PSA will (or won’t) sustain supply to its PFC, ACE, and PCUs if its supplied by only one PMG and that PMG is being driven by a windmilling engine.”

    Four PMGs (two on each backup generator assembly) provide power to the FCDC regardless whether the backup converter is swtiched on or off. The backup generator assembly is not mechanically disengaged from the engine shaft. Only the IDG continuous speed drive shaft can be disengaged (if disengaged by the staff, it cannot be reengaged). ie power will be generated from the four PMGs.
    Each PMG provide power over a very wide range of speeds. In wind mill condition of the right engine, the 2PMGs on the backup generator assembly will generate some power proportionally to the speed of the engine (i think it is directly proportional). The two PMGs on the running left engine will provide normal useful power at the speed of that engine.
    If too much load is drawn, they load will be shutdown according to the load shedding logic and if not the load will trip.
    It will be useful to know what if the power of each PMG at cruise speed (then work out how much power they generate at windmill).
    When the xbus is depower the RAT will be deployed after some time. Onece it is deployed the PMG only need to provide power to the ACE-R (ACE-L2 will be automatically shutdown as part of the load shedding logic). The ACE-R is mainly an analogue device and won’t consume much power maybe 300-500 W (guesstimate – based on battery rating ~800W but it will be less) but definitely less than 1kW. The ACE-R is expected to remain powered.
    In case of two IDGs down, it will be borderline maybe you get 100W-200W guestimate) out of these at low speed and should be enough at high speed. It would be interesting to get the manufacturing data for each. If not sufficient, the battery will discharge but slowly as the PMGs still produce useful power.

  720. HB says:

    @Andrew, Victor,

    It the TAC input removal drives back the rudder to trimmed position, ignore my second comment. I expected the last input to be retained.

  721. Richard Godfrey says:

    On 6th February 2018 at 06:33:43 UTC Seabed Constructor was at 33.5433°S 106.8249°E on a course of 075°T at 12.2 knots.

    Seabed Constructor is estimated to arrive in Fremantle AMC on 07/02/2018 at around 21:00 UTC (08/02/2018 05:00 Local Time).

  722. haxi says:

    @All,

    Malaysia just released the second weekly report: http://mh370.gov.my/en/mh370-underwater-search-2018

    The report didn’t explain why OI’s AIS signal was off, but it did say two points of interest had been identified and later classified as geological.

  723. Richard Godfrey says:

    If you have trouble downloading the second Malaysian Report, it is also available here:

    https://www.dropbox.com/s/mn7h7oioaygyc11/MH370%20Search%20Weekly%20Report%202.pdf?dl=0

    In case you missed it, the first report is also available:

    https://www.dropbox.com/s/rolfwhuwi6what9/MH370%20Search%20Weekly%20Report%201.pdf?dl=0

  724. David says:

    @Richard Cole. Re your SC back projected track of 4th Feb, Mick Gilbert posted her position at 03:57 on 4th as 35° 53.151′ S 094° 16.699′ E. The Malaysian 2nd weekly report has her leaving the search area at 0200 UTC on the 4th, two hours earlier. At 13 knots she would have been well outside any part of the search area then, much less the bottom of it; yet that evidently is where she a started.

    Some poetic licence as to the extent of the search area in the Malaysian report do you think or the Russians tied a rocket to her?

  725. ventus45 says:

    The positions of the two geological features are:
    https://pbs.twimg.com/media/DVV2CL-VoAAlRUq.jpg:large

  726. Victor Iannello says:

    @All: Some of you might have had difficulty reaching the site today. There was a glitch with the domain name renewal process. All should be fine now.

  727. Ge Rijn says:

    @VictorI

    Can we allready conclude the CSIRO/Griffin hotspot areas turned out empty?

  728. Victor Iannello says:

    @Ge Rijn: I am not sure that all of the points of interest in the primary area have been cleared. The Malaysian report is a bit cryptic. I am trying to learn more, and I’ll share what I can.

  729. Rob says:

    @VictorI “I am trying to learn more, and I’ll share what I can.”

    That’s good, and rest assured you can rely on us to keep whatever you find, strictly in house. Meanwhile, it’s mum’s the word.

  730. Ge Rijn says:

    @VictorI

    Yes, it’s good to be carefull especially now I think. At any time if they find the plane they won’t tell the world right away unless they are absolutely sure.

    To note: @Gysbreght made some interesting analysis on the Boeing simulations imo posted on JW. Further detailing the problem we’ve been discussing here for a while now from different angles.

    I also think it comes down in the end to your conclusion from yesterday.
    Whatever happened before 0:19 if we take the final BFO’s as facts proving a high speed descent then there are only three options left:

    1: An unpiloted scenario with an unknown hydraulic/electric/engine configuration that should explain the too early high speed descent according the Boeing- and other simulations. Ending up close to the 7th arc. Maybe your latest simulation proposal?

    2: The high speed descent was induced by a PIC and the log-on was triggered by a brief APU or engine start during the descent which ended close to the 7th arc.

    3:The same as two but then the PIC made an extraordinarry recovery and glided the plane outside the +/-25Nm area somewhere.

    When the final BFO’s represent a (high speed) descent from ~20.000ft I think a recovery could have been established at ~15.000ft.
    Limiting the possible gliding distance to about ~50Nm +/- the 7th arc.
    I still believe this possibility should be considered seriously by all and OI in specific.

  731. Richard Godfrey says:

    @Rob

    There is no way of keeping information published here “in house”.

    According to the internet statistics from alexa.com, Victor’s website reaches around 12,360 users with an average of 27,210 pageviews each month. There are a lot more readers, than commenters.

    When the website went off the air briefly earlier today, I had people emailing me from the UK and Australia asking me if I knew what happened (and I am just a commenter).

  732. lkr says:

    @GR:
    “When the final BFO’s represent a (high speed) descent from ~20.000ft I think a recovery could have been established at ~15.000ft.
    Limiting the possible gliding distance to about ~50Nm +/- the 7th arc.
    I still believe this possibility should be considered seriously by all and OI…”

    Those like you and Rob who prefer piloted glide endings seem to assume that the glide phase must be at right angles to the 7th arc. If the original heading is maintained, you could get your +40nm DOWNRANGE, but no need to look an equivalent distance uprange.to the south and east. If the bearing in the glide phase is NOT postulated to be a continuation of the previous hours’ flight, then any deviation in track during a glide would end closer to the arc, so even granting a glide possibility, the probability should drop rapidly as you exceed the the +25nm, and fall to near-nil outside the -25nm. Unnecessary to say also, that any lower elevation than 15000 puts terminus pretty much in the +/- 25nm zone. Or that you’re giving OI a useful bet..

  733. Rob says:

    @Richard

    What I said was meant to be taken with a pinch of salt. I was only joking!

    Anyway, it’s good that were still talking.

    I promise I will keep anything Victor tells me in confidence (I wish) to myself.

    We are in for an interesting few weeks BTW.

  734. Ge Rijn says:

    @Ikr

    ~15.000ft is just an indication. A ~50Nm glide a maximum glide range from around that ~altitude. I take it wide. It’s about the point.
    And I consider a radius from a point on the 7th arc. A PIC could have gone anywhere within tis radius with this scenario.
    Indeed the lower a recovery was made and a glide started the shorter the radius. I hope they find the plane within ~+/-25Nm.

    It’s just that this possibility cannot be ruled out yet as @VictorI also mentioned. And with growing odds during the latest discussion the end of flight scenario could probably be best explained by pilot-input this possibility becomes more realistic too.

    All I try to avoid is this possibility will be refuted by OI too also as long it’s not off the table yet conclusively.

  735. Victor Iannello says:

    @Ge Rijn said I still believe this possibility should be considered seriously by all and OI in specific.

    Yes, you have made that clear. Again and again.

    If the search proceeds to 29S or further north at +/- 25 NM without a discovery, then we will have to consider whether to search wider, scan areas already scanned, or search further north along the arc. We are months from reaching that point. For now, I think the best option is to continue with the current plan.

  736. Richard Cole says:

    @David

    >Some poetic licence as to the extent of the search area in the Malaysian report do you think or the Russians tied a rocket to her?

    The first datapoint on my map was at 01:07GMT on 4th Feb, at 02:00GMT it was
    just crossing the line of longitude through the eastern point of the primary search area, perhaps that’s what they were referring to. Hard to interpret generalised statements in a precise way. No rockets were needed.

  737. Rob says:

    @ikr

    You obviously haven’t bebouten reading my posts. Perhaps when you saw Rob’s name you switched off like everybody else, dismissing his posts as evangelical rants.

    Well it’s time for a correction as they say on Wall Street.

    The ATSB said they were optimistic about finding the plane in the original search area. Misplaced optimism. Then, with CSIRO’s help, identified the 25,000sq km as a good bet, CSIRO identified priority zones, based on rehashed drift studies and grainy satellite images. You lot bought that again. Well, let me break it to you now, there is isn’t a snowball’s chance in hell of finding the plane there, either. Becoming a frustratingly familiar story isn’t it. Groundhog day again in the SIO.

    And to add insult to injury, Victor is pinning his hopes on areas even further north, to ever more irrational territory. Victor still cherishes the idea of loiter between 1829 and 1941. Confirmation bias is an insidious enemy.

    The pilot was intent on flying as far into the SIO at his fuel will take him. The pilot was intent on ditching in a way that would the plane a rapidly as possible, leaving a little floating debris as possible. He needed daylight to make sure he could carry out his evil plan, and be sure nobody would be close enough to see him do it. His flight path was carefully arranged to remain in darkness until the last half hour. Do you have any conception of how unlikely that flight path synchronisation with sunrise is a coincidence? If you had read my previous post describing the planning required to rendezvous with daylight, at S38, you would be left with only one conclusion – it couldn’t possibly be a coincidence. The flight path to S38 achieves the greatest distance possible on the available fuel, as M0.82 at FL400 (ref Bobby Ulrich’s fuel model) Another coincidence? No, not a coincidence. The aircraft turned 25nm before it got to IGOGU (strictly 25nm before arriving abeam of IGOGU, as aircraft was offset 15nm from N571 at the time) echoes of simulator practice on 2nd Feb and flew toward N41 E88 exactly. Another coincidence? You are surely not serious.

    Victor dismissed a great circle path to S38 with fuel exhaustion at 0017.50 as technically impossible. Dreaded confirmation bias at work again

    Ikr, when the fuel ran out, the pilot glided his plane, and kept on the same bearing, to get as far south as possible before ditching. That’s why he wasn’t found in the first search. He was gliding on RAT power only, that’s why the flaps were retracted.

    The ATSB adopted the line that retracted flaps pointed to an unpiloted descent. Wrong! The ATSB adopted the line that the final BFO’s pointed to an unpiloted descent. Confirmation bias, no less!

    OI are about to waste their own money and kill any future hope of finding MH370, in following “expert advice” of ATSB.

    Watch it unfold.

  738. Rob says:

    @ikr

    That’s S41 E88, not N41 E88. Auto complete spoiling the best efforts.

  739. David says:

    @Richard Cole. “… at 02:00GMT it was
    just crossing the line of longitude through the eastern point of the primary search area, perhaps that’s what they were referring to.”

    Thanks Richard. Probably the 0200 position was over 8 hrs from her start therefore.

  740. Victor Iannello says:

    @Rob said: And to add insult to injury, Victor is pinning his hopes on areas even further north, to ever more irrational territory. Victor still cherishes the idea of loiter between 1829 and 1941. Confirmation bias is an insidious enemy.

    More accurately, I don’t know what occurred between 18:22 and 19:41. Depending on the BFO drift that might have occurred, the BTO and BFO data allow paths up until around 26S. A descent/loiter remains a possibility, but not at all a certainty. That’s not confirmation bias at all.

    Victor dismissed a great circle path to S38 with fuel exhaustion at 0017.50 as technically impossible. Dreaded confirmation bias at work again.

    More accurately, I said an impact that far south requires a long glide past the 7th arc, and is not consistent with the drift models and the BFO. That’s not confirmation bias at all.

    If you read my post, I put the success probability at 67% if +/- 25 NM wide search area is extended along the 7th arc to 26S. I leave open the possibility of a glide past the search area, a path further north, or a missed debris field. I believe the search is following the search strategy that will most likely succeed. However, I don’t guarantee it.

    If there is any confirmation bias, it’s yours. You repeat opinions and facts and cherry pick the data. You’ve made your points many times. There is no reason to repeat them. If you are inclined to do this, in future comments please reference a link that others can follow if they want.

  741. Don Thompson says:

    @David,

    ABC: I think the term for that is weak news. Congratulations to Paul Kennedy for his promotion to the CEO position at Fugro. Is the rectangular object among the the coal & anchor detritus even a box? That image is at great resolution, there’s an anemone(?) living on it.

  742. Victor Iannello says:

    @David: The speculation is amazing. I believe that Seabed Constructor chose to go silent so that the world wouldn’t get excited as it tended to an equipment issue. The AIS data we have showed that SC traveled in a circle and then begin to retrace an AUV’s path that in the coverage graphic from the first Malaysian report was represented by a dark line, evidently due to bad or missing data. It’s silly to think that going silent has anything to do with exploring shipwrecks or treasure hunting.

    It’s ironic that in an attempt to reduce speculation about the position data, going silent actually might have increased the excitement.

    I hope Ocean Infinity is able to explain exactly what occurred so that this matter is put to rest. Until the facts are made public, outrageous speculation will drown out reason.

  743. David says:

    @Don Thompson. “….even a box?” Looks to be iron judging by the detritus around it.
    Apropos nothing that is an admiralty pattern anchor, stock separated and lying at the shackle end I think.
    Just the one though.
    Maybe fouled but the cable missing.

    @Victor. “The speculation is amazing.” Yes, there is a lot of interest.

  744. David says:

    @Don Thompson. Correction. Second anchor:

    https://d3q3olqy33wgar.cloudfront.net/FE0019_anchor.jpg

  745. DennisW says:

    @Victor

    If you read my post, I put the success probability at 67% if +/- 25 NM wide search area is extended along the 7th arc to 26S.

    I agree with that number – it is identical to my own guess of 70%. Determining probabilities is very difficult. Even the smartest people in the world, quants who work at Wall Street and Vegas, struggle with it for high complexity events. The recent Superbowl results highlight that fact. I actually found some interesting arbitrage opportunities among the Superbowl bets. Very unusual for quants, especially on Wall Street. Vegas took it in the shorts this time. I imagine some of my friends who took the Vegas quant route will be returning to the Valley (stock option land). Even Nate Silver no longer plays in that sandbox.

    https://photos.app.goo.gl/5dTpycENbIXzSSKz2

  746. Mick Gilbert says:

    @David
    @Victor

    Victor, I agree with you that the speculation is amazing but it is most assuredly not unexpected when there is an information void. And, for better or for worse (probably the latter in this case), this information void seems to be of Ocean Infinity’s making. I can sort of understand the rationale that going dark was to avoid speculation that they had found the debris field. But if you game out going dark versus not, all going dark does is generate more speculation across a broader range of possibilities. If we could see where the Seabed Constructor was sailing, the speculation would be essentially one dimensional; we’d be asking what was it doing? A relatively simple practical question to answer; backfilling some gaps in the initial search, investigating points of interest.

    Because of the decision to go dark the speculation is now multi-dimensional, extending well beyond the practical questions – where was Seabed Constructor sailing? and what was it doing? – to motivation – why did you choose to go dark? what did you have to hide? will you continue that practice?

    All that OI have managed to accomplish is swapping out a couple of relatively easier questions for a raft of much harder ones. That’s probably not a good strategy for a company seeking to develop its public image and reputation.

    Any old how, you’d expect that there will be some sort of announcement about the blackout when they reach Fremantle (if not before). It will be interesting to see how OI handle the next phase of the PR/media campaign.

  747. Barry Carlson says:

    @Mick Gilbert,

    “That’s probably not a good strategy for a company seeking to develop its public image and reputation.”

    More likely the “strategy” is nothing more than abiding by the terms of the contract entered into with the Malaysian government. There is bound to be a “no operational statements to media or public” gagging clause in it.

    In fact all that Ocean Infinity have had to say, is to provide references to the Malaysian Reports.

  748. TBill says:

    @Victor
    “It does however provide another reason to question any of the radar data after 18:02, including the capture at 18:22.”

    But don’t we have the most recent ATSB report? I believe ATSB reported MH370 at the 18:22 point. While it does seem possible the 2014 Lido presentation to the NOK might have taken some liberties to help tell the story to that audience, I had hoped the ATSB report made the 18:22 point more solid.

    Also the SLOP/CP seems to be a fantastic fit to the 18:22 thru 18:28, which we must give you credit for seeing how well that proposed maneuver fit the radar/satellite interface data, back in 2014 or so.

    Parenthetically, as you know, I am wondering if instead of a SLOP, it could have been a course change to 1090E via 0796E from NILAM, which looks very similar to a 15nm SLOP/CP…I am not thinking MH370 went all the way out to 1090E, just got on that operating line, with 0894E as a possible hopping off waypoint (at lower altitude).

  749. DennisW says:

    @TBill

    The events between 18:22 and 19:41 are impossible to determine – which I concluded over two years ago. Why keep beating on a dead horse?

  750. David says:

    @Mick Gilbert. “It will be interesting to see how OI handle the next phase of the PR/media campaign.”

    As @Barry Carlson says it might not be up to them.

    As you say, “I can sort of understand the rationale that going dark was to avoid speculation that they had found the debris field. But if you game out going dark versus not, all going dark does is generate more speculation across a broader range of possibilities.”

    To my mind the Malaysian Government’s priority would be to try to avoid anything which stirs the relatives unnecessarily/prematurely. Whether prompted by the Malaysians or OI, for SC to go dark to investigate anomalies such as these last two should inure us all, observing the likely frequency of that: two anomalies so far. To stop for a ROV visual will then be unremarkable in a frequent routine rather than limiting the darks to just that and each of those creating excitement and risking disappointment. So to spread it across a broader range of possibilities might have its advantages?

    They might prefer to leave all this vague. Were they to come up with specifics that might itself stir some and prove unwise in net. Even to discuss it might not help.

    To me there is no right answer, just a least worst.

  751. Richard Godfrey says:

    On 7th February 2018 at 07:21:13 UTC Seabed Constructor was at 32.4818°S 112.3748°E on a course of 075°T at 11.8 knots, with a 167 NM to go.

    Seabed Constructor is estimated to arrive in Fremantle AMC on 07/02/2018 at around 21:30 UTC (08/02/2018 05:30 Local Time).

  752. HB says:

    @Victor, Andrew,
    I am not entirely convinced of that pitot heater will lose power completely. Power will be reduced significantly, that s for sure.
    In your scenario, the r backup generator will kick in and generate power at whatever speed the shaft is rotating including windmill condition. The backup generator is capable of providing power over a wide range of speed. The load shedding will kick in to maintain load balance. There is obviously a transient condition here to consider.

  753. Nederland says:

    @TBill

    It seems to me that the Butterworth base in Penang had switched off their digital radar backup at around 18:01.

    Neil Gordon confirmed that the DSTG has not received data sets for the flight beyond 18:01 and that this is probably because no digital data backup exists for the later segment.
    http://jeffwise.net/2016/09/07/exclusive-interview-with-top-mh370-search-mathematician-neil-gordon/
    But they have received data sets up to 18:01:
    “The radar data contains regular estimates of latitude, longitude and altitude at 10 second intervals from 16:42:27 to 18:01:49. A single additional latitude and longitude position was reported at 18:22:12.” (DSTG) – note the use of the word “reported”

    The FI mentions only two locations past 18:01, the one is the time “registered over Pulau Perak” (perhaps a written note as no digital backup exists for that segment), the other the “observed” disappearance 10 nm behind MEKAR (“The tracking by the Military continued as the radar return was observed to be heading towards waypoint MEKAR, a waypoint on Airways N571 when it disappeared abruptly at 1822:12 UTC”).

    In visualisations of the MH370 track, the path from 18:01 to 18:22 is therefore a straight line, crossing the island of Pulau Perak.

    The switching from digital backup to manual observation of the radar in real time seems to coincide with the ATC phone call to the Malaysian military at around 18:00 UTC as reported by Reuters.

  754. Andrew says:

    @HB

    RE: “I am not entirely convinced of that pitot heater will lose power completely. Power will be reduced significantly, that s for sure.”

    The pitot heat system DOES lose power in the event of a dual engine flame-out, as evidenced by the EICAS message HEAT PITOT L+C+R. The Dual Eng Fail/Stall checklist procedure requires the crew to wait for the HEAT PITOT L+C+R message to disappear (indicating the APU has started and is generating power) before cycling the PRIMARY FLIGHT COMPUTERS switch to restore flight control normal mode.

  755. HB says:

    @Andrew, this part is well understood, however, in reality there will be some power from the backup generator under windmill condition and the pitot heaters are not big loads. My point is there is a question mark on whether the instrument will sense loss of power in the first place. Do you know the relay set points?

  756. Andrew says:

    @HB

    RE: “My point is there is a question mark on whether the instrument will sense loss of power in the first place.”

    There is no question mark. The EICAS message appears following a dual engine failure because the probe heat sensor relays have detected power failures in all three pitot heat systems.

  757. HB says:

    @Andrew, it appears on what? The simulator? This is a difficult transient to build in a simulator. The scenario here is with windmill.

  758. Don Thompson says:

    @HB

    Perhaps you have an oversimplied perception of the Backup Generator and the Backup Generator Converter. The Backup Generator Converter supplies L, R or both Transfer buses. The Backup Generator Converter is the controller for both the L and R generators, applying an exciter field to one, or other, generator if a set of criteria are satisfied. The criteria include under/over voltage and over/under frequency. It’s abundantly clear that the Backup Generator Convertor does not provide supply to the transfer buses if the engine is windmilling.

    Load shedding is a specific function of ELMS, it is not intended to control loads supplied off the Transfer buses.

  759. Victor Iannello says:

    @HB: I think that windmill effects (rotational speeds, drag) would be fairly accurately modeled in a simulator.

    Previously, I provided a reference that says the backup generator operates over a rotational speed ratio of 2:1. When windmilling, the rotational speed of the high pressure spool would be much less than 50%. It is very unlikely to produce power.

  760. HB says:

    @Victor. I understand that but power will be produced although reduced. Then it is a matter of load balancing. For the simulator, do you have a reference to that? I think only a load analysis taking into account relay set point and loads shedding can tell. Even if you go to 25pc, ie 5 kVA, it may be enough after load shedding.

    @Don, with all due respect it is clear based on what? As i said this will cope with wide range of speed/frequency. Only low voltage on the xbus is an issue here if loads are not shed or loads dont trip. If the heaters are part of that load shedding logic it would be clear. The heaters are not big loads.
    It will be useful to have the detailed single line diagram if someone can share.

  761. Andrew says:

    @HB

    The ‘reality’ is that the transfer buses are unpowered after a dual engine failure. A whole bunch of systems lose power, as indicated by the slew of EICAS messages that appear. One of the immediately obvious indications is the failure of the FO’s displays, which are powered by the R transfer bus, via the C2 TRU and the FO flight instrument bus. Do you seriously believe the simulator does not model that behaviour correctly? If so, then you’d better contact Boeing, the simulator manufacturers, the civil aviation regulators and airlines to let them know they’re wrong and have been training pilots incorrectly all these years.

  762. Don Thompson says:

    @HB

    Clear” based on the experiences of a number of this forum’s contributors who have conducted trials in Level D simulators, one an experienced 777 pilot, the Boeing technical documents, and Victor’s note above (windmilling results in less than idle rpm on the PTO shaft to the accessory gearbox).

  763. TBill says:

    @DennisW
    “The events between 18:22 and 19:41 are impossible to determine – which I concluded over two years ago. Why keep beating on a dead horse?”

    …above Victor basically says from 18:02 forward we don’t really know (because Lido slide was reportedly not MH370), which seems a little controversial to back off the 18:22 point (since I believe ATSB seems to accept it).

  764. DennisW says:

    @TBill

    Yes, I had pretty much accepted the 18:22 position as well.

  765. Victor Iannello says:

    @TBill, @DennisW: I’d say there is more certainty about the 18:02 position than the 18:22 position. I’m not advising to totally disregard the 18:22 point. It’s possible the Lido Hotel image shows MH370 captures and the Malaysian officials have reasons for not acknowledging that.

  766. Ge Rijn says:

    It seems to me there is a misunderstanding here.
    Each backup generator consist of one ~20kVa generator and two PMGs.
    The two PMGs (per engine) directly supply power to the Primary Flight Control DC system. Completely independent of the IDGs and 20kVa backup generators (which both loose power after flame-out) as I understand it.

    The PMGs will keep delivering power when the engines are windmilling.
    On the B747 and the B787 they provide sufficient power to keep the primary flight controls working when engines are windmilling (according information found).

    The question was- and is do they also provide sufficient power on a B777 when windmilling to ACEs and PCUs etc.
    This question has not been conclusively answered yet as far as I know.

  767. Victor Iannello says:

    @Ge Rijn: The recent comments were to address @HB’s following comment: I am not entirely convinced of that pitot heater will lose power completely. Power will be reduced significantly, that s for sure.

    Pilot heat is supplied from the transfer busses. The backup generators in turn supply power to the transfer busses. Your question about the PMGs is a different one.

  768. Richard Godfrey says:

    The DSTG report states that there was a series of radar data every 10 seconds up until 18:01:49 UTC.

    There was a completely separate and single additional latitude and longitude position that was reported at 18:22:12 UTC.

    The two data sources need not necessarily to have been the same source.

    The fake Lido image uses the 18:22:12 UTC data point and is an attempt to join up the dots from 18:01:49 UTC.

  769. Ge Rijn says:

    @VictorI @HB

    I assumed @HB was talking about the PMGs in the backup generator powering the pitot heat under windmilling.
    It’s clear to me the 20kVa backup generators stop producing power after flame-out.

  770. Ge Rijn says:

    @HB

    Maybe you want to explain to me what you meant when you talked about backup generators in this regard.

  771. Joseph Coleman says:

    @Richard Godfrey

    Is there any reference in any official reports (Malaysia or ATSB) as to the exact co-ordinates at 18:22:12UTC (not a just 200NM NW of wherever or 10nm after MEKAR ) if so could you or anyone please provide a link. I know there is no official reference as to the exacts of how this approx position came to be. It’s just a request for knowledge, no hidden agenda

    Thanks in advance

    Joe

  772. TBill says:

    @Joseph Coleman
    Below is ATSB final report.

    This is what I am saying, ATSB is saying they knew it was MH370 by comparison to EK343 (unnamed below). I am just saying 18:22 point seems semi-solid, even if Lido slide is not totally valid. The other implication to me is ATSB probably knows EK343 exact route and timing out past LAGOG. which I’d love to know the exact data.

    “Radar data shows the aircraft then headed to the northwest, eventually aligning with published air route N571 from IFR waypoint VAMPI. The validity of this section of the radar data was verified using the track of a commercial flight that followed N571 about 33 NM behind MH370. The aircraft continued to the northwest until a final radar position for the aircraft was recorded approximately 10 NM beyond IFR waypoint MEKAR at 1822:12 (Figure 3). There were no reports of the aircraft being detected by any radar after this time. Key events are summarised in Table 3.”

  773. Richard Godfrey says:

    @Joseph Coleman

    There is no official reference giving the exact position of the the last primary radar contact at 18:22:12 UTC. to my knowledge.

    We have all calculated a point 10 NM beyond waypoint MEKAR on flight route N571.

  774. Victor Iannello says:

    @TBill: Yes, the ATSB, the DSTG, and Malaysia all seem to say different things about the radar data after 18:02. If I had to guess, I’d say the Lido radar data is good and Malaysia has reasons to not acknowledge it. But that’s just a guess, which was my point.

  775. Nederland says:

    The ATSB final report does not explicitly say either that there are digital radar backups available past 18:01 as MH370 was heading to a northwesterly direction already by that time, up to which the existence of radar data is confirmed. Certainly, no co-ordinate has ever been published, suggesting that it does not exist because it would be such an important data point.

    The Bayesian method book likewise speaks of the final position as a report rather than as recorded data (p. 18).

    “The final reported position from radar was at very long range from the sensor and there was a long time delay between it and the penultimate radar report. This report is at long range and it is likely to have rather poor accuracy because the angular errors translate to large location errors at long range.”

    This also, to my mind, confirms that MH370 was indeed specifically tracked in real time because the operator switched from short to long range.

    Perhaps MH370’s route relative to EK343 was also reported by the radar operator, otherwise how could this be known if no recorded data exist?

  776. ALSM says:

    SC will arrive dockside at AMC CUF #4 around 9-10am local this morning. Looking forward to meeting crew and others gathered here in Fremantle.

  777. HB says:

    I was talking about the backup generator and how the associated loads to the xbus. Thanks for confirming this is not based on electrical load analysis. The PMG is another matter which i commented earlier.

    Unfortunately, i cannot comment on the rotational speed in windmil condition as this is not my domain but at low speed, you will still get some power out of it provided that this is still within the acceptable frequency range which is quite wide. The 2:1 ratio has no relevance to this topic. The 50% mentioned is relevant (i expect a lower value though).

    Basically, it is mounted on a variable speed shaft and the more rotation more power less rotation less power. What is the operating range? Then what is the min and corresponding shaft speed? and is that shaft speed include windmill condition?

    It should still provide essential loads at low speed (question mark for windmill speed) it is not just a backup at cruise speed.

    Since you have all the documentation, could you mention what is the rotational speed at which it will be shutdown, the frequency range and power range? and what are is the loads list for the xbus with their values associated with the xbus?

    @Andrew, in the electrical distributions, loads will be shed to the non-essential as you all know. Then, there will also be a sequential shutdown in the essential loads based on relay set points and the load demand/ load balance in this particular scenario and the individual load protection. Motors with underload protection and bigger loads will drop first, if you have unbalance in loads, the unbalance side will drop first. Also Two TRUs will switch to RAT which will reduce the loads (not clear yet on the timing). If it is not a preconfigured load shedding system, the sequence will be based on relay coordination and load balance/configuration. Personnally i don’t think a simulator (nevermind how sophisticated it is) can capture that, it will assume the worst case that all loads are dropped and RAT will be deployed after x sec.

    Going back to the pitot tube heater, the signal triggering the Secondary will be based on 0 voltage while the relay settings in the distribution will be higher. If it is a small load like i think may well be a load to drop last. This is quite important as the loads will be unbalanced in Victor’s scenario.

    All the above to say, i still have question marks about this topic. A test or dynamic simulation would be interesting.

  778. Mick Gilbert says:

    @Nederland

    ‘… the Butterworth base in Penang …’

    Butterworth isn’t in (on?) Penang, it’s on the mainland. Do you mean Western Hill?

    And what is a ‘… digital radar backup‘?

  779. Nederland says:

    @Mick Gilbert

    AFAIK, the Butterworth base is located in the Malaysian state of Penang, but not on the island, so in, rather than on, Penang, sounds about right to me.

    With digital radar backup (the term used by Neil Gordon) I mean a digital recording of co-ordinates and other data several times a minute, such as was made available to Australia from take off up to 18:01.

  780. Victor Iannello says:

    @HB: Several of us have tried to explain various reasons why the backup generator is not expected to provide power during windmilling. The reference I cited gives both the shaft rotational speed and output frequency range, indicating a turndown ratio of 2:1. There is no load shed for the transfer bus, which is what feeds the pitot heaters. We are talking in circles. I don’t see any new facts coming to light. I suggest we all move on.

  781. ventus45 says:

    @ALSM
    Have a good look time onboard.

  782. ventus45 says:

    Seabed Constructor Arriving in Fremantle.
    She is approaching the Outer Pilot Boarding Point Now.
    https://pbs.twimg.com/media/DVd8Rn0V4AAtKqF.jpg:large

  783. DennisW says:

    @ALSM

    Looking forward to meeting crew and others gathered here in Fremantle.

    Buy them a round from me, and email Victor the bill. I’ll get it.

  784. Andrew says:

    A beautiful day here in Perth today:
    https://www.transport.wa.gov.au/imarine/swanbourne-beach-cam.asp

    (Swanbourne is a few km north of Fremantle)

  785. Mick Gilbert says:

    @Nederland

    AFAIK, the Butterworth base is located in the Malaysian state of Penang …

    Fair enough. I’m still curious as to what radar you are talking about; there are two – Butterworth and Western Hill. The radar at the RMAF base at Butterworth is an Alenia-Marconi ATCR-33 air traffic control radar; it has a range of about 60 nm and its primary radar function provides no altitude discrimination. The radar operated by RMAF No 310 Squadron radar station at Western Hill is a Leonardo (formerly SELEX) RAT-31DL 3D air surveillance radar; it has a range of about 270 nm.

  786. Hi Victor,

    Been following you for some time now.

    Quick question, is this still relevant or has it been ruled out?

    https://www.youtube.com/watch?v=ksfInwSHVwU&index=21&list=FLDynrb3hmpXSScn4yNLP6Xg

  787. Victor Iannello says:

    @David: I’m not sure that explanation is correct. More soon.

  788. Barry Carlson says:

    @David,

    The standard of journalism has certainly dropped! Each of those outlets with an online presence has tagged onto the coat-tails of a competitor, and dutifully reported on unsubstantiated Twitter dribble. Or better still, made up their own fictional garbage around it.

    Even the ABC has fallen into the same trap.

    Thankfully, the Australian Marine Complex at Henderson is a ‘closed’ facility, but even that will have a downside and the conspiracy mill will probably get into top-gear mode.

  789. HB says:

    @Victor
    Please provide that reference again. If the generator shutdown at speeds higher than windmill, fair enough, otherwise please read again my comment carefully.

  790. David says:

    @Barry Carlson. Yes I see apparently she went to Henderson. Doubtless we will hear how that thickens the plot yet more.

    Below, Liow has been interviewed but if you listen to it he does not say exactly what the report claims. I do not hear him say that OI turned off their position indicating because, ” …they have every right to do so”. I did hear him say that OI did so because they were, “..on their way back…”; in which he seems to have been misinformed. They left that out.

    I cannot quite make out what he actually says about the reported, “…items picked up from the seabed, over the past few weeks”. I have my doubts their interpretation is what he said, though that would be news if so.

    https://www.thestar.com.my/news/nation/2018/02/08/mh370-hunt-going-smoothly-liow-dca-will-provide-more-details-on-the-missions-updates/

  791. David says:

    @Victor. “…..More soon”.

    Curious, in light of Minister Liow’s inteview.

  792. Joseph Coleman says:

    @Richard Godfrey

    Many Thanks

    @TBill

    As regards to your previous comment, I understand that based on the paragraph you’ve mentioned from the ATSB Final report, that you have the idea that the Officials have another form of data for EK343. That may or may not be true. You’d have to ask them. We have to understand that the officials should be able to trust the composers of the extraction of the radar data from its raw format (specifics details unknown) to distinguish the difference between the two radar hits from the recordings.

    There may be different ideas about the SOM track after Penang to the perceived 18:22:12 UTC hit, but without specific knowledge of the full extraction process and any additional errors if any in that process to its conclusion, they are just ideas.

  793. Nederland says:

    @Mick Gilbert

    My understanding is that the Western Hill radar, not the Butterworth approach radar, is the Malaysian military radar mentioned in the FI for the last segment of the flight and is also operated by Butterworth. The Butterworth approach radar can not be the source for the data recording up to 18:01 because of its limited range and because the data include altitude.

    I am not sure whether the early turn back at IGARI was also captured by this radar or whether it was captured by a different military radar unit (such as in Gong Kedak on the east coast?)

  794. Richard Cole says:

    @David

    I thought that Liow was pretty clear in the interview:
    1. Items have been found on the seabed in the last few weeks
    2. The ship will turn off the AIS when it returns to the search area

    There was a piece on ABC radio Sydney at 18:30AEDT about a fight between civil and military accident investigators in Malaysia to put 6 more staff (if I recall correctly) on the ship, to secure the flight recorders if they are found

  795. Richard Cole says:

    @David – more

    …that didn’t chime with a continuation of a general search of the wide search area.

    Liow’s body language looked pretty happy, I thought.

  796. David says:

    @Richard Cole. The report referred to items,”picked up from the seabed” which I took it meant recovered as distinct from discovered. He says,”..you know, one of those items they pick up over the last couple of weeks”. Yes that interpretation I thought, and agree, is unlikely. Most probably he was referring to detection of the anomalies.

    On the other, what I reckon he says is, “…not lost could locate the ship. Not true just switch off actually the signalling because OI is on the way back to the port…” That was his explanation. I did not hear him say that the ship would turn it off on return to the search.

    I agree he projected happiness with the search.

  797. David says:

    @Richard Cole. Thanks for the ABC interview. That should keep the wires humming.

    It implies that the Malaysian Air force is intent on controlling the investigation from two directions.

    Yet at the same time I think the DGA is in process of becoming a statutory body at the direction of Government, ie independent of it.

    ?

  798. Julia says:

    It’s very worrying news if true that the Malaysian military will be in sole charge of investigating the black boxes when found.From what I understand Malaysian military planes failed to intercept MH370 when it was spotted flying in the wrong direction from the official flight path. This is well documented.

  799. sinux says:

    @David
    Re:Liow Interview
    He isn’t explaining why the ship switched off the AIS.
    He is saying, the ship couldn’t have disappeared, because it’s on its way back to port (paraphrasing).

  800. Mick Gilbert says:

    @Nederland

    No 310 Squadron at Western Hill runs the radar, 310 Sqn and 321 Sqn at Gong Kedak fall under the No 1 Sector Operations Centre at Butterworth. Refering to the Western Hill radar as ‘Butterworth‘ on the basis that’s where the SOC is is confusing particularly as Butterworth is where the Butterworth/Penang International ATC radar is located. If you want to understand the military radar set up Don Thompson authored an excellent piece on the topic some time back – Malaysian Military Radar Surveillance Capabilities.

    Regards the FI, as to which radar they are refering to under ‘Malaysia Military Radar‘ is conjectural, theoretically it could have been Western Hill. The traces illustrated on pp.10-11 of the FI are from the DCA (civilian) radar at Kota Bharu and the Butterworth/PIA ATC radar.

  801. Don Thompson says:

    Kelly Perazzolo asked “is this [CTBTO] still relevant or has it been ruled out?

    That specific hydroacoustic event was recorded 1hr15min after the final satcom burst was recorded at the Perth Ground Earth Station.

    If one considers that the event recorded by the CTBTO HA01 array may have originated near the 7th arc, at a latititude between S32º and S26º, the time taken for the impulse to travel through the ocean’s deep sound channel is in the order of 18 minutes (distance from arc to HA01 varies with latitude). There’s is an obvious disparity between the time of the satcom event, directly attributable to the aircraft, and the estimated origin time of the hydroacoustic event.

    A team at Curtin University, WA, investigated the CTBTO HA01 recordings and correlated with their recordings from the Australian IMOS network. The University’s conclusion was that the recorded event originated much further north than the MH370 area of interest and was of geological origin. That specific event recorded by CTBTO’s HAO1 array has been set aside as not relevant.

  802. Victor Iannello says:

    @Julia: The reason for some in Malaysia wanting to replace accident investigators with Air Force personnel on Seabed Constructor remains a mystery. It might be related to OI’s statement in The Economist story that the blackbox if found would be given to the ATSB for analysis, and some in Malaysia would prefer that Malaysia controls access and interpretation of that data. Perhaps we’ll get clarification from Malaysian officials.

  803. Julia says:

    Thanks for this @Victor. I envisage a long struggle for the truth ahead but one step at a time.

  804. Ge Rijn says:

    The search area is in international waters. The Malaysian Government or their military have no authority there.
    OI can do what pleases them best I think.

  805. Nederland says:

    @Mick Gilbert

    OK, but if I understand you correctly, then the Western Hill military radar is still operated from Butterworth.

    At any rate, I can think of no other reason why the military radar mentioned in the FI (quite possibly the one above) was digitally recording up to 18:01 but not thereafter, even though the flight was still being tracked and the radar observed.

    The illustrations in the FI are clearly underneath the heading ATC (civilian) radar. The section on the military radar is particularly obscure.

  806. ventus45 says:

    Put these in your tool box.
    Two very useful spreadsheets, for converting UTM to Lat/Long and vice versa.

    Created by: Richard Stanaway
    Quickclose Pty. Ltd.
    (Freeware, QUICKCLOSE Software – PO Box 1364 Carlton VIC 3053 AUSTRALIA)
    email: quickclose@quickclose.com.au
    Version 1.0 20th July 2009
    http://www.quickclose.com.au/PNG94Ellipsoid_to_PNGMG94.xls
    http://www.quickclose.com.au/PNGMG94_to_PNG94Ellipsoid.xls

    Screenshot of both sheets (side by side) for the coordinates of Image-1 and Image-2 from Malaysian Weekly Report #2

  807. Victor Iannello says:

    @Ge Rijn: OI has contractual obligations to Malaysia. We don’t know all the details. OI can’t necessarily do as it pleases.

  808. Ge Rijn says:

    @VictorI

    Yes we don’t no the details of the contract. But in international waters OI has the final say what happens to the black boxes or the plane if they get found, whatever the contract.
    Malaysia can put 7 or 20 military on SC but they cann’t force them to hand over the black boxes if they find them. Unless under gunpoint.
    I assume OI does not allow weapons on their ship.

  809. Victor Iannello says:

    @Ge Rijn: OI will likely not deliberately breach their contract with Malaysia, as it would put their payment in jeopardy.

  810. TBill says:

    @Nederland
    As far as EK343, I assume the investigators may have contacted Emirates for detailed records of the aircraft path and timing. If not, then it was not much of an investigation.

    Questions-
    (1) Does military radar see Transponder Flight ID?
    (2) Can the Transponder Flight ID be changed to a different flight no.?
    (3) Can the ADS-B be turned off to be silent on FlightRadar24? If so then perhaps TCAS was turned on at 18:25

  811. Ge Rijn says:

    @VictorI

    We’ll see. Breach of the contract could be the replacement of the two Malaysian military onboard SC by seven others (fighter pilots..)
    If I was OI I won’t like it and reject it.
    In the end OI has souverinity on every decision they like to make.
    The cards are in their hands not in Malaysia’s.

  812. Richard Cole says:

    @David

    >I did not hear him say that the ship would turn it off on return to the search.

    At 01m49s Liow says:

    “..so when they are back, on the way back to the…, they will switch off the signal, this is normal, so the public will not, should not be alert[ed?], over the loss of signal, searching for the mission MH370”

    IMO he is saying AIS will, or might, be disabled when Constructor returns to the search area.

  813. Ge Rijn says:

    Nothing special I think. The occasional storm in a glass of water.
    I guess CSIRO/Griffins’s hotspots have turned out empty.
    It’s on to the next move nearing ~32S.

  814. Nederland says:

    @TBill
    – AFAIK –
    (a) a military radar is normally both primary and secondary, so yes, the transponder signal should be picked up on the secondary radar
    (b) I believe it is possible to manually alter the flight ID in the cockpit, at least you can delete it, this seems to be what happened in the case of MH370
    (c) There are different settings on the transponder, you can find more information on this here:
    http://www.bendixking.com/HWL/media/Pilot-Guides/006-18301-0000_1.pdf

  815. Nederland says:

    According to German-language newspaper Der Spiegel, the captain of the Seabed Constructor has specifically asked the Norwegian shipping board to temporarily permit turning off the AIS tracking system because of exceptional circumstances. The reason quoted is that media and social networks have put too much pressure on the operation, potentially jeopardising its safety. This information comes from the spokesperson of the Norwegian shipping board.

    http://www.spiegel.de/wissenschaft/technik/mh370-suchschiff-verschwand-tagelang-von-der-landkarte-a-1192453.html

  816. DennisW says:

    @Nederland

    The reason quoted is that media and social networks have put too much pressure on the operation, potentially jeopardising its safety.

    That makes a lot sense, actually.

  817. Andrew says:

    @Ge Rijn

    RE; ”The search area is in international waters. The Malaysian Government or their military have no authority there.
    OI can do what pleases them best I think.”

    @Victor already mentioned OI’s contractual obligations to Malaysia. There’s also Annex 13 to the UN’s Convention on International Civil Aviation. In this case, Annex 13 gives Malaysia control over any wreckage that may be found outside the territory of any State:

    ”5.3 When the location of the accident or the serious incident cannot definitely be established as being in the territory of any State, the State of Registry shall institute and conduct any necessary investigation of the accident or serious incident. However, it may delegate the whole or any part of the investigation to another State by mutual arrangement and consent.“

    ”5.6 The investigator-in-charge shall have unhampered access to the wreckage and all relevant material, including flight recorders and ATS records, and shall have unrestricted control over it to ensure that a detailed examination can be made without delay by authorized personnel participating in the investigation.”

  818. David says:

    @Richard Cole. “IMO he is saying AIS will, or might, be disabled when Constructor returns to the search area.”

    Thank you Richard. My mistake.

    Taking what he says literally they will have it switched off all the time on station not just during investigations.

  819. Richard Cole says:

    @David

    The Der Spiegel article implies they turned AIS off as soon as permission was received from the Norwegian Authorities, not due to any change in their operations. If so, they will continue with it off, when in the search area.

  820. David says:

    @Richard Cole. Thank you. My translator does not make that clear, as might be the case for others.

    A sample. The first line was translated as, “Who wants to watch thick pots, must no longer hold his nose in the wind. Instead of drafty spread ships long ago by the domestic sofa can be traced all over the world.”

    I do not want to watch thick pots right now really.

  821. Nederland says:

    As a German native speaker, I translate the first sentence as follows:

    “Whoever wants to observe massive ships, has no longer expose his nose to the wind. Rather than tracking vessels from stormy shores, it has long since become possible, over the whole wide world, to do this from your own couch.”

    The article says that the Norwegian authority approved the request to turn off the transponder on 1 February. It is not clear when OI submitted that request.

  822. Nederland says:

    “Auf Anfrage des SPIEGEL erklärte Steinar Haugberg, Sprecher der für das Schiff zuständigen norwegischen Schifffahrtsbehörde, zwar sei der Einsatz des AIS-Systems auf Schiffen dieser Größe durch internationale Regeln vorgeschrieben. “Das AIS kann aber nach Ermessen des Kapitäns unter bestimmten Bedingungen abgestellt werden.”

    Das Unternehmen Swire Seabed, Besitzer der “Seabed Constructor” habe sich in der Tat bei der Behörde gemeldet und darum gebeten, das AIS abstellen zu dürfen. Die Firma habe sich auf anhaltenden Druck durch Medienhäuser und Social Media berufen, der Besatzung und Offiziere des Schiffs von ihren Aufgaben abgelenkt und den sicheren Betrieb der Operation behindert habe. Man habe dem Antrag am 1. Februar zugestimmt – unter anderem unter der Zusicherung, dass die Sicherheit anderer Schiffe in der Region nicht gefährdet wird. ”

    Upon inquiry of Der Spiegel, Steinar Haugberg, spokesperson of the Norwegian shipping board responsible for the vessel, explained that although international rules make it mandatory for vessels of this size to use the AIS system, “the AIS can, however, at the discretion of the captain be switched off under certain circumstances”

    He says that the company Swire Seabed, which owns the “Seabed Constructor”, has indeed approached the authority and asked for permission to switch off the AIS. The company has quoted persistent pressure by publishing houses and social media, which they said has detracted staff and officers of the vessel from fulfilling their tasks and put the security of the operation into jeopardy. They had approved that request on 1 February – provided that, amongst other precautions, the safety of other ships in the area is not endangered.

  823. Nederland says:

    “Thick pots” is actually quite funny because Der Spiegel is based in Hamburg and “dicke Pötte” is “plattdeutsch” (northern German dialect) for massive ships.

  824. David says:

    @Nederland. Caution if you order a large ship from Hamburg then.

  825. DennisW says:

    @Nederland

    Good stuff, although I am not fan of Hamburg. I prefer Munich where one can routinely see a Lamborghini Diablo parked at the curb. No one would park their Lamborghini in San Francisco.

  826. Rob says:

    It will ratchet up the suspense.

    DennisW

    There’s a coincidence, been watching a travel programme on the telly about San Francisco, seems a nice friendly place. Very easy going. I’m sure your lambo would be quite safe, he said tongue in cheek.

  827. Mick Gilbert says:

    @Nederland

    … if I understand you correctly, then the Western Hill military radar is still operated from Butterworth.

    No, that is not correct. The Western Hill military radar is operated from Western Hill.

  828. DennisW says:

    @Rob

    There’s a coincidence, been watching a travel programme on the telly about San Francisco, seems a nice friendly place. Very easy going. I’m sure your lambo would be quite safe

    Sure, you can be walking along pier 14 holding your daughter’s hand when she gets shot (and killed) by an illegal alien who gets off with a wrist slap. Cali is f’ing sick (Pelosi is a great example). I am looking at leaving – convert assets to bit coins and move to South America.

  829. Don Thompson says:

    Lamborghinis, in the US.

    Is that even a thing?

  830. Mick Gilbert says:

    @David

    I love the way News Corp have adopted that semi-blurred photo of a Malaysia Airlines B737 in the 2012 livery as their standard accompaniment for MH370 stories.

  831. TBill says:

    The good news is debris from the shipwrecks is visible, contrary to some speculation back in 2014 of several feet of “mud” that could be covering the debris

  832. Lauren H says:

    I try to make a mental note of any information that is significant in determining the impact location. Here is one;
    Above, @Andrew stated that after right engine failure, the left servo-motor would fail and auto-throttle would NOT increase left engine thrust to try to match the previous total thrust. This fact will affect my pin location because it changes both range and endurance. It also might affect @Brian’s decel rate and the path required to fly from the 6th to 7th arc at a lower speed as well as the altitude after left engine flame out.

  833. Victor Iannello says:

    @Lauren H: Be careful. That statement was for a particular electrical configuration in which the left IDG and the left backup generator were isolated.

  834. Nederland says:

    @Mick Gilbert

    Two credible journalistic sources, Reuters and the Australian ABC, reported that the Western Hill radar is indeed operated from Butterworth, Penang.

    “Air Force chief Rodzali Daud went to the air base in Penang on March 9, where the plane’s final radar plot was recorded.”

    https://www.reuters.com/article/us-malaysia-airplane-investigation/exclusive-malaysia-starts-investigating-confused-initial-response-to-missing-jet-idUSBREA3A0NS20140411

    4 Corners (ABC) also say that while the radar itself is indeed located on the island of Penang rather than on the mainland, it is still operated from the Butterworth base (from around min 29:30):

    http://www.abc.net.au/4corners/lost-mh370/5464818

  835. DennisW says:

    @Don

    Lamborghinis, in the US.

    Is that even a thing?

    I don’t know anyone who owns one. I have seen a few.

    Personally I can’t drive by a high school parking lot without noticing that all the vehicles in there are better than what I am driving.

  836. Mick Gilbert says:

    @Nederland

    Sorry you lost me when you quoted the ABC as a reliable source.

  837. Don Thomspon says:

    @David,

    Thank you for the reminder. Fremantle Ports’ vessel movements database published Constructor’s arrival & departure schedule more than a week ago.

  838. David says:

    @Don Thompson. Confirmation.

  839. Ge Rijn says:

    So the search is about to resume.
    Than we can safely conclude the CSIRO/Griffin hotspot areas have turned out empty. One theory and area down again. Up to the next area.

  840. Rob says:

    @Ge Rijn

    “One theory and area down again. Up to the next area.”

    Exactly which theory are you referring to?
    Perhaps it was the one that was not so much a theory but a precise pinpointing of three impact zones based on a detailed study of localized ocean currents found to be flowing on the morning of 8th March 2014, combined with revelatory satellite photos depicting tsunami debris fields, by professional, highly respected marine scientists?

  841. HD says:

    Does anyone know how OI’s 90 day search contract is worded?

    Is it an exact 90 day period…ending April 20? Or does the clock stop during these re-supply trips to AU?

  842. IR1907 says:

    ”, which they said has detracted staff and officers of the vessel from fulfilling their tasks and put the security of the operation into jeopardy. ”

    OI staff really cares what mike chillit and a few other lowlife trolls have to say?

  843. Nederland says:

    @Mick Gilbert

    Do you have a link to the paper on Malaysian military radar?

    It would appear odd that Butterworth has access only to an approach radar.

  844. Don Thompson says:

    @nederland

    TUDM Air Defence Surveillance

    The air traffic management function at Butterworth, also covering Penand International, exploits a Selex ATCR-33 radar system. It is a joint military-civil function, similar to other Malaysian aerodromes, Kuantan and Labuan. It is networked into the civil ATM system,

    The air defence surveillance function is something entirely different, the RAT-31/DL radar head is sited on Western Hill on Penang Island. The RAT-31/DL is networked into the MADGE C² system which permits the air situation to be displayed at any of its operations centres.

    The ATM and ADS functions in TUDM fall under separate commands.

  845. Mick Gilbert says:

    @Nederland

    You can find a link to Don’s paper here.

    Re: ‘It would appear odd that Butterworth has access only to an approach radar.

    Not at all. Not every air force base (or even every fighter base) is a microcosm of an integrated air defence system (surveillance (ground-based and airborne), interception/fighter assets, air-to-air refuelling, command, control and communications). The integration of all of those capabilities occurs at a level above and separate to the air bases. However, each air force base with aircraft assets must be functional as an aerodrome. To that end, it must have an air traffic control capability and that almost invariably means an area and approach radar. ATC is a base function, air surveillance is not. So, not every air base has a surveillance asset tied to it but it will almost invariably have an ATC radar attached to it.

    The organisation of the Royal Malaysian Air Force with regards to air defence is pretty straightforward. The air space to be defended is divided up into regions and then the regions are divided into sectors. Each sector has an operations centre (SOC) that takes surveillance feeds from its radar assets and that then applies its assigned interceptor/fighter assets as it sees fit. The SOCs don’t ‘operate‘ the assets that are assigned to them, the individual squadrons do that. So you have the Sector 1 SOC (which is located at Butterworth) taking a surveillance feed from the surveillance radar at Western Hill (which is operated by No 310 Squadron) making decisions about the deployment of its assigned fighter assets (typically the F/A-18Ds that are operated by No 18 Squadron).

    To avoid confusion, if you are going to use a geographical name to refer to a radar you would use the name that is relevant to the location of the radar head.

  846. David says:

    @Victor. Another foray into end-of-flight.

    Earlier I put my view that the ATSB’s 2016 scenario of, “…an electrical configuration where the loss of engine power from one engine resulted in the loss of autopilot (AP)”, most likely was central to their deduction that Boeing simulated descents were consistent with the final BFOs.

    What that scenario does is introduce a prolonged yaw asymmetry unlimited by the small quantity of residual fuel and the vagaries of relight accessing that fuel: in short the residual fuel and its supply become irrelevant to the final log-on. At AC loss the left engine would need to rely on APU DC pump supply but since there would be ‘normal’ fuel left in its tank the limited APU supply could be supplemented by tank suction fuel, the two maintaining its pre-right-engine-flame-out power (auto-throttle being inoperative) and not a reduced and intermittent thrust of a relight.

    While the Boeing simulations are inconsistent with the turn radius that could be expected in a steeply banked descent, if the aircraft rolled onto its back as in your June simulation that might be an explanation – but in any case I put that aside for my purposes here.

    There is no other explanation as to how the aircraft could have met with those descent rates and that quickly, around 2 minutes from AC power loss, other than perhaps @Don Thompson’s hydraulic scenario. That reintroduces flaperon asymmetry as a possibility but as yet without a rationale as to how the scenario would come to be, so for now I put that aside also.

    Within these limitations, to me the ATSB scenario, to which you have drawn attention recently, now offers strong potential evidence of that electrical configuration being the key. To remind others, the left engine (or even the other way round, the right if there were some reason to believe that would run out of fuel first) would have its IDG and backup generator both inoperative, obviously for reasons other than the engine being shut down, the right powering the aircraft electrically. On the right’s failure the aircraft would be without autopilot and so free to yaw, roll and, mostly, pitch. The RAT would deploy unless it had already (see later) and the APU would auto-start, that rebooting the SDU for the final log-on.

    A later left relight using residual fuel would not bring a second log-on since APU fuel, which includes fuel in its line, would outlast that for relight and in any case its IDG would be off line.

    A possibility raised earlier is that the scenario would risk reaching the 7th arc from the 6th, though I think all it requires is a little extra fuel in the tanks to delay the right’s fuel exhaustion by a few minutes, the left’s by realisation of the yaw, roll and pitch the aircraft needed for the final BFO descent rates and the timing of that, around 2 minutes. I assume these few minutes and two could be accommodated within overall fuel consumption accuracies and outcomes and so would not be decisive.

    What would be rewarding is to find a reasonable rationale that connected that electrical configuration with the earlier log-on. An example would be selection of a total power loss at or shortly after the IGARI turn, followed by restoration of the right IDG at the time of the first in-flight log-on. The interim effects on aircraft handling would be loss of the autopilot, and the L2 ACE, the latter mattering little. The left and right hydraulic systems would remain engine powered. The centre hydraulics system would be powered by a mix of RAT and air driven pump. The backup generator PMGs would keep the PSAs, PFCs and ACEs powered other than the L2, presuming that the PMGs would generate with the generators off line.

    If of course the right backup generator were not selected off at all, the aircraft would have autopilot available throughout and that would not need to be selected on later with the right IDG. The RAT would not have deployed early but air driven pumps would keep centre hydraulics powered anyway, supposing AC motor pumps would be unsupported by the backup generator. If the PMGs did depend on their backup generators being on the lack of the left two would lead to the L1 ACE also being inoperative after the minute of so its battery would last for. Again there would still be enough PCUs left operating the flight controls.

    Main battery drain in these two instances would be overcome by either that right backup generator or RAT. Fuel would be saved by the air conditioning packs being inoperative between IGARI and the first log-on.

    I lean to the lot being selected off, the restoration of the autopilot being the purpose when the right IDG was switched back on, that coincidentally leading to SDU reboot – the significance of that not being apparent at the time.

    The point I make is that accepting this end-of-flight scenario as quite likely would provide a platform for rational backtracking.

  847. Niels says:

    @Mick Gilbert
    It appears to me that Butterworth is more than a SOC.
    “The FPDA provides defence co-operation between the countries, establishing an Integrated Air Defence System (IADS) for Peninsular Malaysia and Singapore based at RMAF Butterworth under the command of an Australian Air Vice-Marshal (2-star)”
    From:
    https://en.m.wikipedia.org/wiki/Five_Power_Defence_Arrangements

    It would be interesting to know the totality of the radar data which is “integrated” and who would have access to it.

  848. Mick Gilbert says:

    @Niels

    Butterworth is more than just a SOC, it is an air force base that hosts a range of operational and support units. Currently the RMAF have four flying squadrons based there plus a raft of maintenance and support units plus elements such as 1 SOC and IADS. Up until 1988 it was a Royal Australian Air Force base (we still maintain a small presence there), between 1941-1957 it was a Royal Air Force base.

    The Integrated Air Defence System (IADS) is essentially the last vestige of the Five Powers Defence Agreement (FPDA), which was set up in 1971 as a loose consultative arrangement involving Britain, Australia, New Zealand, Malaysia and Singapore. 45 years ago IADS had responsibility for the defence of Malaysian and Singaporean airspace and was responsible for developing the region’s air defence capabilities but not any more. These days it’s mainly a coordinating and planning body. How we Aussies managed to hang onto the two star command slot is anyone’s guess.

  849. David says:

    @Victor. My above, 4th last para first sentence. Please delete the last part leaving, “If of course the right backup generator were not selected off at all, the aircraft would have autopilot available throughout.”

    There would be no need, unmanned, for the right back up to be on at the end of flight in this scenario even if its PMGs did not deliver otherwise. The right IDG would keep ACEs other than the L2 powered via DC busses.

  850. David says:

    @Victor and others. My apologies. To the last word of my last add, “…busses, the L2 being ELMS powered.”

    God bless the rest and any who read it.

  851. Don Thompson says:

    During 2001 the acronym IADS was redesignated to refer to the ‘Integrated Area Defence System’. The change reflected the increased role of land and naval forces, alongside airborne assets, in exercises and defence planning.

    This notion that Malaysia’s Air Defence Surveillance system is networked across borders is not substantiated by any information, whatsoever. However, it is frequently the case that the RAAF deploys aircraft that operate out of Butterworth.

  852. Niels says:

    @Mick Gilbert, Don Thompson
    Thank you for explaining.
    Irrespective of the operational details of the IADS, it is hard to imagine that the passage of MH370 close to such an important base went unnoticed. The obvious question why there has not been a reaction has never been answered.
    Mick, regarding “your” air vice-marshal at Butterworth: Perhaps they did not lift him from his bed when MH370 passed virtually over their heads, but it must have been known to him soon afterwards.

  853. Nederland says:

    @Don Thompson, Mick Gilbert

    Thank you for that link and further clarification.

    I still maintain that the information provided by Reuters and 4corners, although of course simplified, is essentially correct.

    “The air defence surveillance function is something entirely different, the RAT-31/DL radar head is sited on Western Hill on Penang Island. The RAT-31/DL is networked into the MADGE C² system which permits the air situation to be displayed at any of its operations centres.”

    I take it that the MADGE C2 system can be operated nationally from any military operation centre (including Butterworth):

    http://www.airforce-technology.com/news/newstrs-madge-c2-system-royal-malaysian-air-force/

    Hishammuddin has confirmed on camera that the ATC (in Kuala Lumpur) notified “the military” about missing MH370, confirming the earlier report by Reuters. He did not give time and place, but Reuters said this was Butterworth base and around 18:00 UTC (presumably Western Hill operation centre is not itself manned at night during weekends).

    According to the FI and DSTG publications, there are digital position data (including altitude) available for the accident flight up to 18:01, but although the flight was further tracked and the radar “observed” (apparently in real time), there are no digital position data available after 18:01 and this could perhaps indicate that the digital radar backup was no longer active/switched off. How else to explain this?

    The Lido image was then apparently fabricated to have something to show to the next of kin, but it never became part of any official publication. Instead, published maps showing the radar record provide exact locations, including margin of error, for the track up to 18:01, but only a straight line from the 18:01 position to the “reported” final position 10 nm behind MEKAR at 18:22.

    Interesting.

  854. Rob says:

    @All

    Take another look at Victor’s map of the new search area. See how the CSIRO priority zones are sandwiched tightly in between the outer limit of the area already scanned by Phoenix, and the boundary of the area originally marked out for it’s attention. Quite a neat trick conjured by an organisation (CSIRO) who found themselves between a rock and a hard place. Totally unprofessional. They are now totally discredited, as are the ATSB.

    “One theory and area down again. Up to the next area.”
    Exactly which theory are you referring to?
    Perhaps it was the one that was not so much a theory but a precise pinpointing of three impact zones based on a detailed study of localized ocean currents found to be flowing on the morning of 8th March 2014, combined with revelatory satellite photos depicting tsunami debris fields, by professional, highly respected marine scientists”

    Any more for any more? Any more further along the 7th arc?

    What was the likelihood of finding it within the 25,000sq km zone. Do I have any advance on 85%?

    What a total, unmitigated farce it’s turned into.

  855. sk999 says:

    “…there are no digital position data available after 18:01 …”:

    Where did anyone say that? All we know is that any such data were not supplied to the DSTG.

    I am no expert on radar systems, but the following article (plus many related pages on the same website) provides useful background information on the distinction between “plots” and “tracks”:

    http://www.radartutorial.eu/10.processing/sp20.en.html

    “Plots” are individual detections of a target from a single pass of the radar antenna. Plots from mulitple passes are joinged together to create “tracks.” The data supplied to the DSTG were tracks. From the Final Report –

    Table 6: “Malaysia Air Defense secondary surveillance radar data. Track BD764 7 March 2014 1642:07–1728:37 UTC recorded at 10 second intervals.”

    Table 7: “Malaysia Air Defense primary radar data. Track BE144 7 March 2014 1729:09–1802:59 UTC recorded at 10 second intervals.”

    The Ldio radar image is labeled, “MILITARY RADAR PLOT FROM PULAU PERAK TO LAST PLOT AT 02:22H”

    The term “plot” is not used casually – it has a specific meaning in the context of radar. Here is a youtube video from Day 5 where Daud refers to “last plot after several intermediate primary plot”

    https://www.youtube.com/watch?v=Pl0CW5m9OX0

    It would appear that the RAT-31DL tracker did not combined the “plots” of MH370 as it passed over the Strait of Malacca into any coherent track, which would explain why the data were not provided to the DSTG.

    “The Lido image was then apparently fabricated …” Sorry, there is not a shred of positive evidence to support this supposition.

  856. Mick Gilbert says:

    @Nederland

    Re: ‘I still maintain that the information provided by Reuters and 4corners, although of course simplified, is essentially correct.

    Then I shall leave to your reading and viewing. Didn’t Reuters have MH370 cutting straight across the Malay Peninsula from IGARI to VAMPI and then to GIVAL and IGREX at one point? As for 4 Corners, it was once a quality program but these days they are hauled before the Australian Communications and Media Authority on such a regular basis for ‘failing to make every reasonable effort to ensure that factual content of the program was accurate‘ I think that they have assigned car parking spaces at ACMA.

    Re:’Hishammuddin has confirmed on camera that the ATC (in Kuala Lumpur) notified “the military” about missing MH370, confirming the earlier report by Reuters. He did not give time and place, but Reuters said this was Butterworth base and around 18:00 UTC

    Civilian notification to the military would not have been directly to a SOC. The military have this pesky little thing called Chain of Command. The notification from the civilian Kuala Lumpur Area Air Traffic Control to the military would have gone to the RMAF Region 1 Regional Air Defence Operations Centre (1 RADOC) which is also located in Kuala Lumpur.

    As to timing, I think 1800 UTC would have been unlikely for the principal reason that at 1804 UTC KL AATC was still trying to sort out where the airplane was with HCM AATC.

    If you go to the FI, p.97; “…at time 1804:39 UTC [0204:39 MYT] KL ATCC Radar controller informed HCM ACC; “…reference to the company Malaysian Airlines the aircraft is still flying, is over somewhere over Cambodia”. And thirty one minutes later, at time 1835:52 UTC [0235:52 MYT] MAS Operations Centre informed the position of the aircraft was at latitude N14.9 0000 and longitude E109 15500, which was somewhere east of Vietnam and this information was relayed to HCM ACC.

    At 1930 UTC [0330 MYT] MAS Operations Centre called in and spoke to Radar controller, admitting that the information from the ‘flight tracker’ was based on projection and could not be relied for actual positioning or search.

    I’d contend that it would be highly unlikely that KL ATCC would have notified the military any earlier than 1930 UTC [0330 MYT] because up till then they believed that the airplane was still flying along its planned route and was somewhere off the coast of Vietnam.

    The Lido image was then apparently fabricated …

    If you were going to fabricate a radar plot showing an airplane tracking up the Straits of Malacca to placate the NOK why would you leave a whacking great hole in the guts of the plots? And have such obvious errors as the Butterworth radial and distance label?

    Any old how, as none of this conjecture assists in refining a likely terminus for the airplane in any way, I’ll just tap out.

  857. TBill says:

    @Rob
    Victor gave 33% for finding in the 25000 km2 search zone, much of which still has to be searched. Most of the sentiment here was the CSIRO hot spots were lesser chance due to being too far off of Arc7.

    Clearly hopes are riding on the the crash being reasonably close to Arc7. I personally feel it could be inside Arc7 whereas the search has tended to focus outside Arc7.

  858. Victor Iannello says:

    @TBill: Yes, I estimated only 33% for the entire 25,000 sq km, of which only about 5,000 sq km have been searched (7,500 sq km if you include what was scanned beyond the designated 5,000 sq km). If the added confidence from the satellite images balance the reduced confidence from the distance from the 7th arc, that would mean my prior estimate of the success probability was around 6.6%. Despite the confidence expressed by CSIRO on their priority locations, I think we would have been quite lucky to find the plane where they have searched. We have to have patience and allow more seabed to be scanned before we come to any conclusions.

  859. Julia says:

    Four Corners were pretty good interviewing Hishamuddin immediately after the disappearance of MH370 almost four years ago.I have the link to the interview somewhere. Hishamuddin was asked specifically about why the military did not intercept the plane when they spotted it deviating or having deviated from the planned flight path. Most of you will have seen this interview I imagine and as the interview does not contain details of Butterworth etc (just vague non committal answers and observations by Hish) it may not be relevant to your current discussion..

  860. Victor Iannello says:

    @Julia: Malaysia can’t make two consecutive statements that don’t conflict. It’s not easy to determine what is true and what is not.

  861. Don Thompson says:

    @Mick, Nederland

    The notification from the civilian Kuala Lumpur Area Air Traffic Control to the military would have gone to the RMAF Region 1 Regional Air Defence Operations Centre (1 RADOC) which is also located in Kuala Lumpur.

    According to Hugh Dunleavy who, in March 2014, was the Director of Commercial at MAS: “Malaysia’s air traffic control and military radar are in the same freakin’ building. The military saw an aircraft turn and did nothing.“. The building is on Subang Airport.

  862. Niels says:

    @Julia
    Here is one link to the interview:
    https://m.youtube.com/watch?v=2XQPCwY5SHk

    In my observation Hishammuddin escapes the question about why there was no interception by going “ad absurdum”:
    “If you’re not going to shoot it down, what’s the point of sending it (a fighter) up?”
    See also:
    https://sg.news.yahoo.com/hishammuddin-now-says-military-told-keep-eye-mh370-025118812.html

  863. TBill says:

    @all
    Re: Winds below 22-South
    What were the winds below 22S? I am seeing up to 60-70 knots at FL350 altitude, but most flight paths seem to quote a smaller number. I also notice that at lower altitude, the winds seem to be much less, so if MH370 had descended somewhat, possibly that explains why the actual flight path may have been straighter than the wind assumptions would suggest.

  864. Ge Rijn says:

    @Rob

    OI could not have afforded not to search the CSIRO/Griffin hot spot areas even if there were no hot spots. And CSIRO/Griffin had reasonable arguments too, given the data and circumstances.
    Any others are hardly any better (although some a bit better I think).
    This area had to be searched anyway. Next move.
    The more area is searched the better the probability gets.

  865. DennisW says:

    @Neils

    Your second link above is interesting, and I recall reading it previously. Certainly the involvement of high level Malays early in the diversion seemed strange to me both then and now. That would not be the case in the US where the response to such an event would be handled by professionals not politicians.

  866. TBill says:

    @Neils

    Was looking over some old Nov_2014 posts and you seemed to be among the first adopters of the 18:40 descent vs. FMT option.

  867. Nederland says:

    I have basically said everything before.

    Yes, it is confirmed only that the DSTG did not receive data records beyond 18:01, but how likely is it that Australia spends several millions of tax payers money, but does then not ask for crucial radar data.

    The suggestion that Malaysia might not have recorded data past 18:01 either is not mine, it’s the one by Neil Gordon in the above link.

    Also, the FI does not say there are recorded radar data available beyond 18:01 either, on the contrary the language swaps to “observed/registered”, whereas for the early segment the word “recorded” is regularly used.

    No one has ever confirmed that any plots are recorded past 18:01. The Lido image, for example, specifically points to the two plots at or near Pulau Perak and the final plot, that’s exactly the same data that you can also find in the FI as “observed/registered” plots (not records).

    http://tmfassociates.com/blog/wp-content/uploads/2014/03/Beijing-event.jpg

    Note that plot is used in singular (Pulau Perak plot, final plot, no track in between)

    In the above, Victor has mentioned a source claiming that there is no recorded track of MH370 past 18:01 and that the Lido image instead shows a different flight (perhaps Emirates 343?). I was commenting on that.

    The problem with all early statements on the radar data past 18:01 is that all that information was contradictory and this does not seem to suggest that Malaysia had an easy time simply checking recorded data. The early Reuters report reflects those contradictions and does not mean the reporting was wrong.

    I am not Australian, so I can’t comment on the general trustworthiness of the program, but the point is that it provides a recording of an interview with Hishammuddin.

    Even if Hishammuddin was not correct to say that ATC contacted the military, this does not change the recording of radar data as outlined in the above.

    The information MH370 was over Cambodia (likewise hard to believe) came in not before 18:04. SOP, however, requires to contact the military within minutes after radar contact is lost or the flight disappears from the screen (i.e. at 17:21). It would be logical that ATC contacted the military long before 18:04.

    Considerations like that could indeed inform the search for MH370. I have earlier suggested a path, assuming that MH370 felt reasonably “safe” (i.e. from interception) in Malaysian airspace because of a predictably poor response, but did not do so with regard to crossing Indonesian airspace within radar range. This would mean a late FMT and an end of flight further north than the area currently searched (perhaps around 31S).

  868. DennisW says:

    @Nederland

    This would mean a late FMT and an end of flight further north than the area currently searched (perhaps around 31S).

    I agree. My current best estimate is a latitude of 5N to 6N at 19:41 with a terminus near 30S.

  869. sk999 says:

    Nederland,

    “… how likely is it that Australia spends several millions of tax payers money, but does then not ask for crucial radar data.”

    Good question, but since neither you nor I is Australian, it would be best to have a native answer that question. Remember also that spending decisions are sometimes driven, not by logic, but by the “use it or lose it” principle.

    “The suggestion that Malaysia might not have recorded data past 18:01 either is not mine, it’s the one by Neil Gordon in the above link.”

    Here’s what Neil Gordon was actually asked and his answer:

    Q. “Was there any explanation or speculation on why a plane would be detected at that point but not before or after?”

    A. “I GUESS it was that they’ve just got snapshots off the radar screen. I’m SPECULATING here but I would imagine they’ve recorded a video of the screen but they don’t necessarily have a digital backup of the measurements.”

    Gordon has no idea of the distinction between “plots” and “tracks”. Further, neither you nor I nor Neil Gordon is an authority regarding the operation the RAT-31DL as configured and employed by the RMAF.

    “Also, the FI does not say there are recorded radar data available beyond 18:01 either, on the contrary the language swaps to ‘observed/registered’, whereas for the early segment the word ‘recorded’ is regularly used.”

    Interesting – that language would be consistent with Gordon’s speculation of a recorded video. But video of what – plots data? Tracks data? Something mix of the two?

    “In the above, Victor has mentioned a source claiming that there is no recorded track of MH370 past 18:01 and that the Lido image instead shows a different flight (perhaps Emirates 343?). I was commenting on that.”

    Let me engage in wild speculation. In addition to primary radar,a RAT-31DL has SECONDARY RADAR. Is it possible that EK343 had a working transponder, and that the RAT was able to match up the secondary radar position of EK343 with the primary radar target, and that it was able to distinguish EK343 from another primary radar target without a matching secondary target that was a whopping 38 nm in front?

    “SOP, however, requires to contact the military within minutes after radar contact is lost or the flight disappears from the screen (i.e. at 17:21)”

    That’s a new one to me. You have a reference?

    “It would be logical that ATC contacted the military long before 18:04.”

    Logical – to you, perhaps. Not to the Malaysians.

    https://www.reuters.com/article/us-malaysia-airplane-investigation/exclusive-malaysia-starts-investigating-confused-initial-response-to-missing-jet-idUSBREA3A0NS20140411

    “A sixth source, a senior official in the civil aviation sector, said the plane’s disappearance had exposed bureaucratic dysfunction in Malaysia, which has rarely been subject to such international demands for transparency. ‘There was never the need for these silos to speak to one another. It’s not because of ill intent, it’s just the way the system was set up,’ the official said.”

  870. ventus45 says:

    @sk999

    Let me engage in wild speculation. In addition to primary radar,a RAT-31DL has SECONDARY RADAR. Is it possible that EK343 had a working transponder, and that the RAT was able to match up the secondary radar position of EK343 with the primary radar target, and that it was able to distinguish EK343 from another primary radar target without a matching secondary target that was a whopping 38 nm in front?

    Not speculation, wild or otherwise. Obviously it does, could, and would. It would be a pretty piss-poor system if it didn’t !!

  871. DennisW says:

    @sk999

    My suspicion is that Malaysian ATC did not realize the plane was missing until Vietnamese ATC asked them about it. That may have actually delayed any diversion based activities since it took more time than planned for Malaysia to understand/confirm that the plane was not on the flight path to Beijing.

  872. sk999 says:

    “My suspicion is that Malaysian ATC did not realize the plane was missing until Vietnamese ATC asked them about it.”

    Actually, I think we can say that that is established fact, based on the ATC transcriptions.

    “That may have actually delayed any diversion based activities since it took more time than planned for Malaysia to understand/confirm that the plane was not on the flight path to Beijing.”

    At this point you have lost me. The speed with which the plane flew back across Peninsular Malaysia speaks loudly and clearly – the intent was to “get the hell out of Dodge.” The introduction of delay would be consistent with that intent.

  873. DennisW says:

    @sk999

    At this point you have lost me. The speed with which the plane flew back across Peninsular Malaysia speaks loudly and clearly – the intent was to “get the hell out of Dodge.” The introduction of delay would be consistent with that intent.

    The flight time from the initiation of the diversion to the FMT at 18:xx, some 1.5 hours or so, was planned (in my scenario) as negotiating time before a go or no-go decision was made for the FMT to nowhere. Just my thinking. No basis in anything factual.

  874. Barry Carlson says:

    @All,

    Seabed Constructor departed from the Australian Marine Complex at Henderson, WA shortly after 11-1700UTC and passed abeam of Fremantle at 1815UTC heading north in the shipping channel before turning west to pass north of Rottnest Island.

    At 12-0248UTC SC was at -32.2820 113.4285 heading 256.9°T at 13.045kts.

    Indications are that the vessel is heading back to the “calibration target”, identified by Richard Cole, near -36.310 91.785. With 1104NM to go, the ETA is about 15-1530UTC or 15-2130LT. Following a calibration check, I suspect the calibration target will be recovered and relocated closer to the next area of interest – wherever that may be.

  875. TBill says:

    @Nederland
    have you had a chance to review the MH370-CAPTIO.net proposed path? It’s is a little like your path up to 1941 except they do not try to fly as far out as Sabang radar coverage, like you and I had proposed.

  876. DennisW says:

    @all

    It is a bit slow in here at the moment so I don’t feel bad about making a short rant on geometric algebra.

    Let’s start with the notion of a velocity vector. We instinctively place the tail of the vector on what we believe to be the traveling object location. This is a “feel good” notion. The velocity vector contains information on magnitude and direction only. There is nothing in the vector parameters that defines it’s position in R^3. Is it harmless to place the tail of the vector on the traveling object? We all do it. The reality is that it is meaningless.

    Likewise the notion of angular momentum. We are conditioned to use a cross product which produces a vector perpendicular to R and v. Is there any reason to think there is some magic force in this direction. Of course not. If we are working in two dimensions the notion of a cross product is meaningless. Angular momentum in the geometric algebra sense is defined as an area. In R^3 it is an oriented area.

    And so it goes. We are hobbled by Gibbs. It is the very reason why most physicists, even with advanced degrees, cannot deal with tensor algebra. We have been brain washed by our education, and cannot easily recover. Tensors are a very natural evolution in geometric algebra when the dimension increases beyond R^3.

    I strongly advise everyone to give geometric algebra a hard look. I used it for the MH370 path, and found a weak preference for 30S. I cannot explain why this is so, and my colleagues advised me to back away from it. Which I did. It is probably due to some uncompensated ellipsoidal warping, but I cannot find the error.

  877. Richard Godfrey says:

    On 12th February 2018 at 07:06:32 UTC Seabed Constructor was at 32.4774°S 112.3699°E on a course of 257.4221°T averaging 12.9475 knots, with 1,935 km to go to reach the initial calibration point at 36.3092°S 91.7895°E.

    Seabed Constructor is estimated to arrive back in the search area at the initial calibration point on 15/02/2018 at around 16:00 UTC (15/02/2018 23:00 Local Time).

    As Barry Carlson has suggested above, we expect the calibration target to be recovered and moved to the start location of the Malaysian Secondary Search Area at around 34.85°S 93.75°E.

    The Malaysian Secondary Search Area is of particular interest as a MH370 route on 180°T to the South Pole as suggested by Victor Iannello at a Constant Mach of 0.8 would have ended at around 34.3°S 93.7°E.

  878. Nederland says:

    @sk999

    Neil Gordon has to be careful about what he says in public. He is naturally using a lot of verbal modifiers throughout the interview, not just in that passage. It still comes down to the belief that no recorded data exist past 18:01 and that the DSTG has therefore not received any rather than that the data exist but have been withheld. Certainly, no position or altitude data for that segment has been confirmed, or even communicated, by Malaysia. All that is mentioned in the FI or elsewhere is the “registered” time over Pulau Perak and the “observed” time before abrupt disapperance 10 nm behind MEKAR, which really is an estimate rather than an exact position. The ATSB had to guess the alitude during this segement of the flight. However, the FI makes it also clear that “The tracking by the Military continued as the radar return was observed to be heading towards waypoint MEKAR”, suggesting many more plots than just the one were “observed”, but no recorded data is available.

    I guess you don’t have to be an authority in specific radar technology to assume that if the system is able to record full data in 10 seconds intervals up to a distance of 80 nm, it would be able to record some plots thereafter also. This does not have to be some sinister explanation, perhaps the operators were swichting from short to long range detection.

    On actions to be taken when an aircraft disappears from radar, see, for example, the excerpts in the FI, p. 94

    “when an aircraft fails to make a position report when it is
    expected, commence actions not later than the ETA [estimated time of arrival] for the reporting point plus 3 minutes” … “notify the RCC that the Uncertainty Phase exists” … “inform appropriate radar units (civil and military) of the circumstances”

    After half an hour (before 18:00 UTC), you have full overdue action and uncertainty phase.

    On p. 101, you can read about the need to initiate NOTAM actions (i.e. information of adjacent ATCs and/or military of the missing flight via network call, which would have been particularly helpful in tracking the flight, i.e. Indonesia/Australia)

    “Position reports” from the airline operation centre can play no role in these considerations, it is all about loss of communication/radar.

    The last paragraph from Reuters you quote relates to Malaysian crisis management in dealing with international media in the days after the disappearance.

    @TBill

    I had a look at that paper. There has been an official paper discussing the statistical likelihood of straight paths from ring 3 to 7, concluding that there were probably few if any changes in direction and/or speed, but the whole flight path up to ring 3 is completely unknown to us and unaffected by that consideration. I find it more likely MH370 flew around Sabang radar and tried to avoid Indonesian airspace. The Cocos Island path works only if a wild maneuvre is assumed for each crossing of the arc, which at the moment looks a bit like a stretch.

  879. Paul Smithson says:

    @Richard G. At the risk of annoying you further I just have to say this because it sticks in my craw so much. Thanks (and I mean it) for the updates on SC position, course, speed and destination/ETA.

    But with a position to 1/10,000th of a degree we would need to know which bit of ship we are talking about and have a timestamp of fractions of a second. Ships on passage do not hold course more accurately than +/- a degree or so at any one time [ref your 257.4221°T] and the notion of a four-decimal speed in knots is laughable even if it is an “average”. You obviously know that more numbers does not mean more accurate so I am completely flummoxed why you continue to do it.

  880. Richard Godfrey says:

    @Paul Smithson

    The average course to date is being compared to the average course required to reach the calibration point. The difference is in the 3rd decimal place. I calculate to 4 places, so that errors are not introduced when making statements such as SC is on course to the calibration point.

    I use average course, because the instantaneous course can vary substantially with currents and windage.

    As you clearly do not understand navigation systems, why not simply ignore my posts or use rounding.

    BOD supplies data to 4 decimal places, Marine Traffic to 6 decimal places.

    I choose to use the precision provided.

  881. Niels says:

    @DennisW
    “Certainly the involvement of high level Malays early in the diversion seemed strange to me both then and now.”

    I understand your point, considering also that they did not consider the tracked aircraft as hostile and it was in the middle of the night.

  882. Niels says:

    @TBill
    Re the “old nov_2014 posts”:

    It seems I’m out of phase with the majority view all the time:
    Currently I think it is slightly more likely that the 18:40 BFOs indicate an early turn south (which was possibly followed by some delay between 18:40 and 19:41).

  883. Victor Iannello says:

    @Niels: I don’t know what the majority view is because I have not taken a poll, but my current belief is that it is difficult to surmise what occurred between 18:22 and 19:41 with any level of confidence. However, termination points further north along the 7th arc are consistent with a delay, as you say.

  884. Victor Iannello says:

    @Richard Godfrey, @Barry Carlson: Why would Seabed Constructor be interested in an old calibration target that is not close to where we thought was the next area to be scanned?

  885. Richard Cole says:

    @Victor: ..interest in old calibration target…

    I have the same view, if they really wanted it back why not pick it up when they were down there a few days ago?

    IMO, either Constructor will turn off its AIS when close and then change course to the next search area (for ‘privacy’), or they intend to widen the primary search area outside the 7th arc.

  886. Paul Smithson says:

    This poster who “knows nothing about navigation” just happens to have a skipper’s ticket and 20k+. Just saying.

  887. Paul Smithson says:

    *20k refers to sea miles skippered

  888. Richard Godfrey says:

    @Paul Smithson

    Then you should know that 1 degree change in course extrapolated after 1935km means you are out by around 33 km.

    Instead we are offered precision that allows us to be 3 m out.

    What is wrong with that?

  889. Richard Godfrey says:

    @ Paul Smithson

    I also have a Yacht Master and International Skipper’s certificate.

    Just saying ….

  890. Richard Godfrey says:

    @Victor @Richard C.

    All I can say is they are heading back toward the start point of the Primary Search Area and not the start point of the Secondary Search Area.

    Why do you think that is?

  891. Victor Iannello says:

    @Richard Godfrey: I have some theories, but I can’t be sure. Perhaps it’s related to the dark line shown in “Site 01/Area-01 (southern leg)” in Diagram 1 from the last Malaysian report.

  892. TBill says:

    @Nederland @Neils
    I do not endorse the end-point of MH370-CAPTIO.net at the moment, but I feel the paper makes some key points, namely that descent at 1838 makes perfect sense as a preparation for crossing over the other flight paths from N571.

    As you say Nederland, the biggest argument I have against Xmas Island is that straight, constant path seems perfect fit for Arc3-Arc5, and straight with varied speed/altitude may be perfect fit for Arc2 to perhaps Arc6.

    The simplest compromise I see, 18:38 is the descent and about 1843 at 94East is the FMT due south. Seems very good fit. Looks a lot like DrB’s path but IGOGU is a little too early for the FMT.

  893. Ge Rijn says:

    @VictorI

    To shime in.
    I see no use in recovering an old ‘calibration target’ at this time in the search.
    It would be a waste of precious search-time. They can recover those targets at any time after the search if they like. There must be another reason.

    I assume they talked in Fremantle and decided to search wider.
    It’s a bit my wishfull thinking too I admitt.
    To me it would be a sign OI is also seriously considering a pilot-controlled descent, recovery and glide outside the 25Nm area.

    The previous discussion about the Boeing/@ALSM and @Andrew simulations have shown a scenario without pilot input cannot explain the final descent/BFOs on the premises those uncontrolled scenarios were based on.

    Overall, it seems most are still reasoning with a not-active pilot after FMT in mind. This mind-set is basically not logical considering the flight till ~18:40 (clearly pilot-controlled) and statistically.
    There really is no reason/evidence at all to support this assumption.
    So I still don’t understand why so many cling to this.
    Whishfull thinking too?

    All in all I think the data (Inmarsat, drift, debris) show the difference between controlled or uncontrolled would not be huge.
    It’s the difference between an uncontrolled steep descent after flame-out and a controlled one.
    With the latter a possible recovery and glide up to ~50Nm from the 7th arc.

    This is the width the ATSB/Fugro took searching till ~35S. I guess they had their reasons to take it this wide.

    I hope OI/SC does the same at least between ~33S and ~32S/~95E/~97E. With emphasis on the east part. It’s about the Broken Ridge +5000m trenches.
    The most difficult terrain. I hope it doesn’t scare them off.

    Just my opinion again. It’s about finding the plane.
    I know my tenancy in this is irritating at times. But my intention is the same as yours and all here: finding the plane and conclusive answers to what happened and why.

  894. DennisW says:

    @Ge Rijn

    My statistical preference is to search to 28N or so at +/- 25km. If the plane is not found then consider widening the search. There is no compelling reason to go wide early, IMO.

  895. Victor Iannello says:

    @Ge Rijn: If they decided to widen the search area, it would make more sense to start at the inner leg, which has not yet been searched out to 25 NM from the 7th arc. However, I think there is little chance they would widen the search much past 25 NM when there is so much area within the 25 NM limits that has not been searched.

  896. Victor Iannello says:

    @DennisW: I didn’t mean to repeat what you said. We were composing comments at the same time.

  897. Marijan says:

    Do you remember the Kraken Aquapix SAS: https://mobile.subseaworldnews.com/2017/09/11/kraken-aquapix-sonar-system-for-ocean-infinity-auvs/
    My guess is that at least some of the Hugins will get those installed. As it can be seen from the pictures in the second Malaysan report, Hugns are using the SSS/MBE (at least some of them).

  898. Marijan says:

    Maybe they are calibrating SAS if they were installed during the last visit, but this just a speculation.

  899. DennisW says:

    @Marijan

    How does the swath width on the SAS system compare with the SSS? My quick look shows that the SAS has about half the swath width as the SSS, but that might not be accurate.

  900. Ge Rijn says:

    @DennisW @VictorI

    My ‘confirmation-biased’ concern (I’m aware) is that OI/SC will skip this box between 33/32S and 95/~97E only searching within +/-25Nm.
    If they would only search the east part till ~97E I think it would not make a lot of difference in time left searching.
    I think better do it as you can than later when weather and time limits deteriorate.
    It would probably take them a month to search those Broken Ridge trenches besides the 7th arc.

    Still time left to search till ~29N within +/_25Nm.
    To skip the possibility the high speed descent at 00:19 was controlled and a recovery and glide were possible is irresposible now.
    The Boeing/@ALSM/@Andrew simulations show an uncontrolled end-of-flight are nearly impossible. @Andrew’s flight continued for at least 5 minutes after second engine flame-out with a limited bank of 20 degrees.
    They either correct a developing bank early (@ALSM), or end at 20 degrees. Or take wide turns after second engine flame-out (Boeing).

    The data show a steep descent assumed ~2 minutes after second engine flame-out. This contradicts all the recent previous findings.
    If so the conclusion can only be a pilot-induced descent and end-of-flight scenario.

  901. TBill says:

    @sk999
    “…Let me engage in wild speculation. In addition to primary radar,a RAT-31DL has SECONDARY RADAR. Is it possible that EK343 had a working transponder, and that the RAT was able to match up the secondary radar position of EK343 with the primary radar target, and that it was able to distinguish EK343 from another primary radar target without a matching secondary target that was a whopping 38 nm in front?”

    Well, I hope so too, but they shoudl have called E