Crash Debris from Lion Air JT610 Provides Clues about MH370

Lion Air flight JT610, with 181 passengers and 8 crew, was climbing out of Jakarta on a flight to Pangkal Pinang (Indonesia) when control was lost at around 5,000 ft. Soon after, the Boeing 737-MAX 8 aircraft crashed into the Java Sea northeast of Jakarta. Although we don’t know the cause of the crash, there were some anomalies noted on the previous flight related to sensor disagreements for speed and altitude which required maintenance. Some suspect that these previous issues might have been related to the crash. Luckily, the flight data recorder (FDR), commonly known as a “black box”, was recovered, and the approximate location is known for the cockpit voice recorder (CVR), the other “black box”, so there is a good probability that the cause of the crash can be determined.

Investigators are now claiming they have recovered 69 hours of data from the FDR, which would be sufficient to analyze JT610 as well as the preceding flight which had the anomalous behavior.

The last ADS-B data that we have from Flightradar24 has the aircraft at an altitude of 425 ft, a groundspeed of about 360 knots, and a descent rate of 30,976 fpm. That translates to an approximate true airspeed of 472 knots and a descent angle of about 40 deg. That suggests the aircraft impacted the sea with very high energy. Similarly, the final BFO values for MH370 suggest a downward acceleration of about 0.7g over 8 seconds, reaching a descent rate of about 15,000 fpm. Unless a pilot was at the controls and skillfully recovered from this descent, MH370 also impacted the sea with high energy. Therefore, the debris produced from the JT610 crash gives us some indication of the types of debris probably produced from the crash of MH370.

Indonesia’s National Search and Rescue Agency (BASARNAS) is in charge of the rescue and recovery operation for JT610, which is in water about 100 feet deep, with efforts reportedly hampered by strong underwater currents and limited visibility. Already one volunteer diver has lost his life while recovering body parts.

The video at the top shows floating debris for JT610. If the objects shown are truly representative of the main field of floating debris, it is evidence that a high speed impact produces only small floating parts spread over a fairly limited area. Now admittedly, a B777 is considerably larger than a B737, and the floating debris field should be easier to find. However, the surface search for MH370 in the Southern Indian Ocean (SIO) from the air began weeks after the disappearance, and the dispersive effects of waves and currents in the SIO are strong. The combination of a dispersed field and small parts might explain the failure for the air search to detect floating objects along the 7th arc. The small size of the floating parts might also explain why satellite images along the 7th arc have not spotted aircraft debris.

Despite the likelihood of small floating debris, the underwater searchers for MH370 expect to find a fairly substantial debris field (bigger than 100 m) and substantial, distinguishable objects such as the landing gear and engines, consistent with the debris field of Air France 447. This is also consistent with the parts of JT610 that have already been found on the seabed. For instance, the picture below shows an engine and part of the landing gear of JT610.


To locate parts on the seabed, BASARNAS is using a combination of technologies, such as multi-beam echo sounders (MBES), side-scan sonar (SSS), magnetometers, and remotely-operated vehicles (ROV). The SSS technology has been the workhorse for the subsea search of MH370, used in both the towed vehicles and the underwater drones.

The video below shows divers helping to retrieve debris from the seabed.

The debris recovered from the crash of JT610 helps explain why no MH370 floating debris was spotted by air and by satellite along the 7th arc, and why we remain hopeful that it will be detected on the seabed with sonar sensors once the correct search area is selected.

194 Responses to “Crash Debris from Lion Air JT610 Provides Clues about MH370”

  1. airlandseaman says:

    Thanks for this post on JT610, Victor. The 69 hrs of FDR information will likely help them understand what caused this crash. But regardless of the cause, the surface and ocean floor debris fields are providing valuable insight into what a high energy impact looks like, with implications for MH370.

  2. Victor Iannello says:

    @airlandseaman: Agreed.

  3. Victor Iannello says:

    For reference, this story contains seabed photos of AF447 as released by the BEA. The condition of the landing gear and engine is similar to JT610.

  4. DrB says:

    @Victor Iannello,

    The small diameter of the engine hulk has me thinking it might be the APU rather than a main engine.

    The very small sizes of floating debris also makes me think the Pemba flap and the flaperon came off in the air. There don’t seem to be any intact flight control structures left floating in this crash, and maybe not for MH370.

  5. Richard Godfrey says:

    @Victor

    Please find attached a link to a Google Earth map of the final track marked in red, the location of the 2 flight recorders marked R1 (recovered) and R2 (located) and the 3 boats currently on site marked with a B.

    https://www.dropbox.com/s/yqlwji4g9g556em/JT610%20Crash%20Site.png?dl=0

  6. Richard Godfrey says:

    @Victor

    I should have said the 3 boats on site recorded by Marine Traffic.

    There is a much larger number of boats on site, which are not sending satellite position updates.

  7. Don Thompson says:

    @DrB

    This 737 MAX was powered by a CFM LEAP-1B engine, its LP fan has 18 carbon composite fan blades. The hub, facing the camera in Victor’s still image, has 18 slots and other close up images show remnants of the carbon composite blade roots still in the hub. Even the compressor casing, forward of the combustors, is torn away.

    Definitely an engine, not APU.

  8. Victor Iannello says:

    @DrB: I doubt it is an APU. A single stage centrifugal compressor and a two-stage axial turbine would be more typical of an APU, such as the Honeywell 131-9D. In the photo above, multiple axial stages (each with its own disk) and what looks like two spools is apparent.

    Yes, I agree that flight control surfaces might have separated from MH370 before impact.

  9. Victor Iannello says:

    @Richard Godfrey: Thank you. The FDR and CVR were right where you’d expect them to be.

  10. DennisW says:

    @Victor,

    Yes, I agree that flight control surfaces might have separated from MH370 before impact.

    I am aggressively skeptical of that hypothesis, as well as the dive followed by a glide scenario. The French were fairly clear on water damage relative to the flaperon trailing edge. How else would you explain it except flutter which was not mentioned in the French report?

    The two week interval relative to the SIO search is huge.

  11. Peter Norton says:

    > Victor Iannello: “sensor disagreements for speed and altitude on the previous flight”

    I have a bad feeling that we see a repeat of AF447 and QZ8501 with an unsuccessful stall recovery and an unreliable airspeed condition as in AF447.

    Although possibly unrelated to JT610, I am still bewildered every time I am reminded of the many lessons-not-learned and parallels between AF447 and QZ8501, for example the dual side-stick input and the exact same linguistic misunderstanding between the pilots:

    « The example of miscommunication between the pilots was when the plane was in a critical stalling condition, the co-pilot misunderstood the captain’s command “pull down”; instead of pushing the airplane’s nose down […], he pulled the stick back […]. Because the captain was also pushing the stick forward and because Airbus has a dual-input system, the two stick inputs cancelled each other out, which led to the plane remaining in a stall condition until the end of the black box recording. (See the similar side-stick control issue in the Air France Flight 447 accident.) » (wiki)

  12. haxi says:

    @Victor,

    Is there a way to gain access to the satellite data (if any) of JT610 to see whether high-energy impact would result in negative BFO values like the -2 value with MH370?

  13. Andrew says:

    @TBill
    @Victor

    RE: “I take it a B777 does not have a STS system, but a 737MAX does? Perhaps due to higher power engines relative to the size of the aircraft.”

    The B777 FBW flight control system has a speed stability augmentation function within the primary flight computers. The B737 MAX has a conventional flight control system; only the spoilers are FBW. Consequently, it has a separate speed stability augmentation system, known as the Speed Trim System or STS.

    The B737 STS is designed to improve the aircraft’s manual handling characteristics under certain flight conditions (low gross weight, aft centre of gravity, high thrust). It does so by changing the incidence of the tailplane, which increases the control column force and tends to restore the aircraft back to its trimmed speed. The STS ensures the pilot must apply a significant control column force to increase or decrease speed by changing the pitch attitude.

  14. Andrew says:

    @DrB

    RE: “The small diameter of the engine hulk has me thinking it might be the APU rather than a main engine.”

    As Victor and Don have already mentioned, the photo is definitely an engine. The engine core is quite small once the fan, cowling and ancillary bits are stripped away, as you can see in the following photo of a LEAP 1B engine core under assembly:

    https://www.aero-mag.com/wp-content/uploads/2017/08/234234234234234234-e1504174772170.jpg

  15. Victor Iannello says:

    @haxi: I doubt it. JT610 would have to have a terminal with the same Doppler pre-compensation algorithm, and there would have to be a SATCOM transmission at the time of the descent.

  16. Victor Iannello says:

    Reuters is reporting that JT610 had faulty airspeed indication.

  17. TBill says:

    @Andrew
    Interesting that 737 MAX is not FBW.

  18. airlandseaman says:

    JT610 would not have needed satcom for that short, local flight. Even it it did, the AES would not have been transmitting much (if anything after wheels up) during a 13 minute flight.

    Bobby: Re: “The very small sizes of floating debris also makes me think the [MH370] Pemba flap and the flaperon came off in the air. ”
    The only 610 photo I’ve seen so far that comes close to a flaperon or flap segment is here: http://bit.ly/2PEQN1V It looks like it might be a wingtip. What ever it is, it is the largest piece of the wing (or tail) I have seen that is not related to the landing gears or engine. I think we will learn a lot more from JT610 about the implications for MH370. But the preliminary photos offer more evidence that the Pemba flap and the flaperon did separate prior to impact.

    Dennis: The French analysis was flawed in several ways. They speculated that the flaperon probably did not separate in the air based on a critical assumption. They assumed that if the flaperon separated in the air, it would descend in a stable, leading edge (LE) down attitude, reaching a very high speed and crash into the water LE down, causing much more damage to the LE. They provided no data or information to back up this claim. The simple fact is that a relatively flat object will tend to fall like a leaf, not the way the French assumed. Time to toss some more models out the door at 1000 feet.

  19. DennisW says:

    @Mike

    I have no opinion on how a flaperon separated in flight would fall. Your leaf analogy sounds plausible to me. My issue with the flaperon is the trailing edge damage. I can’t conceive of how it could have occured except by flutter or contact with the water while still attached to the wing. You would think the French analysts could identify flutter damage versus water contact damage.

  20. TBill says:

    @Victor
    Re: Lion Air
    I see some articles saying the Lion Air ELT was recovered, but I have not heard if the ELT signal was heard, which we would say probably not for this type of crash.

  21. Andrew says:

    @TBill

    RE: “Interesting that 737 MAX is not FBW.”

    The 737 MAX is simply an incremental update, designed to keep the aircraft competitive with the A320 NEO in terms of efficiency. The major changes include advanced, fuel efficient engines; aerodynamic tweaks to decrease fuel burn and FBW spoilers to reduce weight. I suspect that a full FBW flight control system would have been too expensive and taken too long to develop.

  22. Shadynuk says:

    @DennisW Re: “I can’t conceive of how it could have occured except by flutter or contact with the water while still attached to the wing.”

    You can see similar damage here without water contact and no mention of flutter:
    http://code7700.com/accident_china_airlines_006.htm

    (I hope I never get a flight with those guys driving. Basic mistake – trust your instruments, not your inner ear.)

  23. David says:

    @Andrew. In the 777 the autopilot disengages on loss of transfer bus power. It was conventional wisdom on blogs that this was because of pitot tube heat loss, a consequence of that, immediately after which airspeed data could be trusted no longer.

    A couple of months ago I believe you said you had found that what prompted A/P disengagement, while still due to possible input data corruption, was more complex than just this pitot heat loss, though without giving details.

    My (loose) speculation is that the BFOs could be affected by whatever the cause of the unreliable data is, for 2 reasons. First the assumed 1 min for the SDU to warm up after APU auto-start on engine run-down may be shorter than restoration of reliable data to it generally. Second, if pitot tube heating still has a part in this, a question is whether after a minute’s cooling while the APU comes on line, heating time of a minute while the SDU comes on line will be sufficient.

    I note that there is no pitot tube temperature sensor to say when its output can be considered reliable or not. In other words SDU BITE might not test for integrity of all data.

    Any comments please?

  24. Don Thompson says:

    @TBill

    Yes, recovery of a piece of debris, identifiable as the ELT, has been reported. The colour of the ELT may have attracted the attention of a diver or ROV pilot searching for the CVR. The recorders and the ELT are located at the rear of the fuselage.

    On the day of the accident, BASARNAS did confirm that no transmission had been received by their SARSAT terminal.

  25. Don Thompson says:

    @David

    The position and attitude data requried by the AES is sourced from the ADIRU, but the inertial reference part, not the air data psrt of that system.

  26. David says:

    @Don Thompson. My thanks for your patience. I know you have expounded on this before.

    What underlies my interest is why pilots are automatically denied the autopilot in a double flameout until the APU fires up or an engine relights and gets back above idle. Powering it by DC or standby AC would leave it engaged unless deselected by deliberate switching or control movement. A pilot might be busy with other things.

    A possible reason for not doing that, relying on the same ADIRU data that the SDU/AES does, is that it is not ‘good’ enough. That leads to my query as to what the SDU BITE will accept as good enough.

  27. Victor Iannello says:

    @David: The SDU relies on latitude, longitude, groundspeed, track, pitch, bank, and heading to steer the antenna and pre-compensate for Doppler shift. The SDU doesn’t care about air data. On the other hand, stability enhancement and envelope protection need air data. Based on this, it’s no surprise that with the loss of full capability of air data, the flight control mode degrades from NORMAL to SECONDARY, and envelope protection and the autopilot are lost, while the SDU functions in the usual fashion.

  28. David says:

    @Victor. Fair enough. Thank you.

  29. Peter Norton says:

    > TBill says:
    > @Peter Norton: Thank you I will give it a try when I get a chance…not sure if that
    > will work when I have an aircraft flying in flight sim. I guess I can toggle back and forth.

    Please let me know when you have tried.

  30. Andrew says:

    @David

    RE: ‘A couple of months ago I believe you said you had found that what prompted A/P disengagement, while still due to possible input data corruption, was more complex than just this pitot heat loss, though without giving details.’

    I did give details! On 25 August, I commented:

    ‘The loss of power to both AC transfer buses would have caused the loss of a number of systems and sensors that are required for the operation of the PFCS in normal mode. Consequently, the PFCS would have automatically reverted to secondary mode, resulting in the loss of the AP and TAC, etc. In the past it was thought that the mode reversion was caused by the loss of pitot heat. However, recent advice indicates the reversion is caused by the loss of other sensors required by the PFCS. The loss of pitot heat would only cause a mode reversion if the system detected invalid air data.’

    On 26 August, you asked:

    ‘Are the transfer buses the source of power for the “other sensors” or are the main buses? Any more on which they are?’

    I replied:

    ‘I don’t have a comprehensive list, but my understanding is that they are mostly position sensors associated with actuators controlled by the L2 ACE. The L2 ACE is only powered by the L 28V DC bus, which is normally supplied by the L AC transfer bus (via the L TRU).’

    In short, the loss of pitot heat alone does not cause reversion to SECONDARY mode unless the air data actually becomes invalid. The reversion to SECONDARY mode is caused by the immediate loss of several other sensors that provide data required by the PFCS. However, as Don and Victor have already mentioned, the data required by the SDU is not affected.

  31. David says:

    @Andrew. Thanks for that refresher.
    15 demerits to me.

  32. Victor Iannello says:

    Regarding JT610, Geoffrey Thomas is reporting that sources within Lion Air say the aircraft’s tech log shows a record of a problem with “elevator feel and control” on the last flight the day before the tragedy. This is in addition to false airspeed indications.

    It would seem that those two problems are related. For instance, an incorrect airspeed input to the STS might cause incorrect trim and incorrect elevator feel, making it harder to manually control pitch.

  33. TBill says:

    @Victor
    If STS comes on when A/P is off, and A/P is off due to bad airspeed data, it does not really make total sense that STS would use the same bad airspeed data. But it does sound like something could be wrong with STS system, or pitch/elevator controls. Igf you are heading for water surface, might as well try to pull off a Sully-style dicth, unless it was more like Air Alaska with no control of flight surfaces.

  34. Richard Godfrey says:

    @Victor

    You stated in a comment on the previous post “I am not seeing an obvious way to use the drift data to discriminate impact latitude along the 7th arc.”

    Please find links below to 3 annotated drift maps from various start latitudes of around 22°S, 26°S and 31°S.

    All 3 take 477 days ± 8 days from the 7th Arc to Madagascar. All 3 end up at a latitude of around 18.8°S ± 2.3° on Madagascar.

    All 3 follow a similar transoceanic track, following the start phase:

    https://www.dropbox.com/s/gpa4xft9kcudgd3/Drift%20Map%2021.9666S%20103.5739E%2021.1447S%2048.4794E%20477d%20Annotated.png?dl=0

    https://www.dropbox.com/s/nb1k6de5irl56hn/Drift%20Map%2025.6655S%2099.6084E%2018.2956S%2049.4172E%20484d%20Annotated.png?dl=0

    https://www.dropbox.com/s/qx3hjw2j8ulctdy/Drift%20Map%2030.7332S%2097.5208E%2017.1433S%2049.5997E%20469d%20Annotated.png?dl=0

    I can appreciate the reason for your statement above. I think we need a much broader based statistical analysis from the simulated drift models to be able to draw any conclusions from the drift analysis. However, the drift models must include all the likely or confirmed MH370 discoveries made so far, from Mossel Bay, South Africa to Pemba, Tanzania. The northern bias in David Griffin’s model, almost excludes the southern most find at Mossel Bay, South Africa and makes you wonder how Pemba, Tanzania was the only discovery north of Madagascar.

  35. Victor Iannello says:

    @Richard G: Yes, looking at it from 40,000 ft…for a broad range of impact latitudes, a lot of debris is predicted to land in Madagascar. A lot of debris was found in Madagascar. It was also the best searched shore in Eastern Africa. Beyond that, we are trying to make sense of what seem like outliers.

  36. DennisW says:

    @Victor,@Richard,

    I don’t believe it is necessary to consider a broad range of latitudes, or do any more work on the drift analytics. Richards’ work has shown (to my satisfation) that latitudes North of 20S are very unlikely based on the debris we have identified so far (including outliers). We have searched +/-25km of the 7th arc from 38S to 25S. All that is left is 25S to 20S at +/-25km. I have shown that there is greater than an 80% probability of finding the wreckage with an underwater search of this area. A conclusion I can now safely put in the “martingale” category. Richard has already done the “heavy lifting”.

    There is the issue of reconciliation with the group advocating a wider search width. I have no interesting in that reconciliation, but you are a peacemaker. Those people are not going to be quieted by additional drift analytics. My advice for smart people would be to focus on the search width so that a convincing case can be made for why and how a future should be conducted.

  37. Victor Iannello says:

    @DennisW: I’m not trying to reach a consensus, which will never occur. I am trying to give everybody an opportunity to make their case to maximize our collective knowledge.

  38. Victor Iannello says:

    @TBill: The A/P might have been manually disengaged, not automatically because bad air data was detected. But that doesn’t explain why the pilots were not able to set pitch and power and stabilize the plane. With faulty “elevator feel” and/or runaway trim, that task might be more difficult than we realize. They might have also had to deal with faulty stall warnings, which would be loud and very disconcerting. We’ll know soon enough.

  39. DennisW says:

    @Victor

    My definition of reality – things that people universally experience the same way. A flat earth was once a reality like dark matter today. Reality should not be confused with truth or speculation (which dominates posts here).

  40. Niels says:

    @VictorI, RichardG
    I’m wondering if there are ways to better validate the “Godfrey” and the “Griffin” drift models, by experimental data. To check for a possible “bias” and to check how well the dispersion of a cluster of objects would be captured.
    Could this be done based on historic drifter data, or would a specific experiment be needed. For example, what would we learn if 20 undrogued drifters would be released beginning of March 2019 from 32S and 20 from 24S on the 7th arc and followed in time (and compared to the models).

  41. Victor Iannello says:

    @Niels: The only variation between Richard G’s model and the historical drifters would be due to annual variations, as the historical drifters define the spatial and temporal drift fields.

    @Richard G: Here’s a fundamental question: When you are using multiple historic drifters to calculate the drift velocities, do you average the speed and direction separately, or do you average the velocity vectors? I suspect you average the velocity vectors, i.e., velocity components.

  42. Barry Carlson says:

    @All,

    Drift modeling, unfortunately, is relying on historical and/or after the fact drifter buoy data, which IMO isn’t a compelling methodology in determining a specific drift scenario. Thousands of theoretical ducks left to their own devices will go whichever way the air/water interface pushes them, and it doesn’t take long before they have spread far and wide. Drifter buoys behave in a similar fashion, but we don’t have thousands of them – just a few.

    In short, the actual sea level at any given time is subject to the prevailing atmospheric pressure and the surface wind friction caused by the differential isobaric pressure. The air/water interface surface current so created will be modified by any existing subsurface currents, up-wellings etc. due to temperature inclines, and often this disturbance can turn into a gyre. Overall, the conditions being experienced at any point in time are extremely fluid (no pun intended) and due to the nature of the beast, basically impossible to pin down, let alone quantify after the fact.

    On top of the previously mentioned conundrum, is the difficulty in estimating the performance of assorted pieces of floating debris in respect of how each item reacts to the affects of Stokes drift and additional leeway due to windage. Errors made in assumptions for the aforesaid become accumulative, and over 100’s of days small ones become large.

    Essentially, every drifter buoy and/or piece of floating debris could announce to the world, “Know me well, for I shall pass this way only once.”

    Having said the above, I respect the efforts that have been and are currently being made to determine from whence the debris that has been identified as from MH370 actually originated.

  43. Niels says:

    @VictorI
    Yes, so for the “Godfrey” model the experimental validation would need to cover several years to get a feel for the annual variations.

    A first indication for these variations (however less convincing than obtained from experimental data), could perhaps be obtained by running the “Griffin” model several times for different starting years.

  44. Andrew says:

    @Victor

    RE: “It would seem that those two problems are related. For instance, an incorrect airspeed input to the STS might cause incorrect trim and incorrect elevator feel, making it harder to manually control pitch.”

    The leaked maintenance log page I referred to earlier indeed shows the ‘FEEL DIFF PRESS’ light illuminated during the previous flight. A malfunctioning elevator feel system might make pitch control more difficult, as you mentioned, and that could make an unreliable airspeed event harder to manage. However, the B737 elevator feel system uses a dedicated pitot system for dynamic pressure sensing, with pitot tubes mounted either side of the vertical stabiliser. That system is independent of the pitot-static system used to derive airspeed, so the two problems might not be related. Either way, it seems the aircraft had some significant gremlins that were not adequately resolved before the accident flight.

  45. airlandseaman says:

    Bloomberg article on JT610 here: https://bloom.bg/2qwPBjb

  46. Victor Iannello says:

    @airlandseaman: That’s not good. It sounds as though they are converging on a bad AoA sensor.

  47. DennisW says:

    @all

    Confessions to make here.

    I have never had a pilots license, but have flown quite a lot. My proprety did not have a long enough straigt-away, so you had to run around in a circle a few times to gain speed before take-off. Landing was OK, but you had to drop it in pretty quickly.

    When I read about AF447 or JT610 I can only shake my head (no disrespect, Andrew). Good grief, fly the f’ing airplane. What is the problem with pilots these days?

    I have the same opinion about ABS, particularly on motorcycles of which I am an affectionado, and have raced them for over two decades. I have some 15 of them, none newer than 2009 simply because non-ABS is no longer an option. If I want a computer controlling some portion of my life, I will elect not to participate in that portion (and I am really good with computers).

  48. Andrew says:

    @DennisW

    RE: “When I read about AF447 or JT610 I can only shake my head (no disrespect, Andrew). Good grief, fly the f’ing airplane. What is the problem with pilots these days?”

    No offence taken; seasoned pilots have been warning about that kind of thing for donkey’s years. Automation is a fabulous tool that has improved safety in many areas, but that has occurred at the expense of good old-fashioned stick and rudder piloting skills. Many (most?) airline pilots today barely get enough hands-on flying to remain proficient, let alone improve their skills, especially those that fly long-haul. Some airlines insist their pilots make the maximum use of the automation by engaging the autopilot immediately after take-off and only disengaging it shortly before landing. That’s all very well and good, except when something goes wrong and the pilot suddenly needs to fall back on his or her manual flying skills. The problem is exacerbated by the increasing number of low-time pilots that are moving into the right hand seat at many airlines, particularly the low-cost carriers. I fear it will only get worse if the forecast global pilot shortage starts to bite and cockpit experience levels plummet.

  49. DennisW says:

    Thx Andrew,

    I regretted the post almost the moment I made it.

    I really want to do and encourage intelligent things going forward. IMO, that is to continue North on the 7th arc at the +/-25km width.

    Sorry to be boring and sorry Victor, I have nothing else to contribute here.

  50. Richard Godfrey says:

    @Victor

    You asked “When you are using multiple historic drifters to calculate the drift velocities, do you average the speed and direction separately, or do you average the velocity vectors? I suspect you average the velocity vectors, i.e., velocity components.”

    When calculating each day a new drift speed and direction, there might be up to 20 drifters within the 1° of latitude and longitude spatial granularity and 1 month temporal/seasonal granularity. Each drifter will be moving with a different speed in a different direction and changing speed and changing direction all the time within the artificial bin created by the granularity definition.

    If I would take the average speed and average direction separately, it would distort the result giving an average point to point trajectory within the artificial bin. This would make drift simulations significantly faster than actual drifters. Drifters do not travel at constant speeds in constant directions, not even for a short time frame. An average point to point speed and direction is misleading. Of course the raw GDP data is only every 6 hours, so I do not know what goes on in between successive data points.

    In my view, you have to average the velocity vectors within the artificial bin created by the granularity definition. I did quite some testing comparing actual drifters to the simulator results and within the constraints listed below, I was satisfied with the results.

    The simulator is obviously not perfect. (1) I am averaging a small number of 6 hourly data points for each new calculation. (2) I am averaging data points between the years 2000 and 2018, not just for the years between 2014 and 2016, when the MH370 floating debris was actually in passage. (3) I am averaging only 280 undrogued drifter data sets. (4) I do not account for a storm passing through the simulator track. (5) The average drifter data will reflect storms that particular drifters encountered, although this tends to get averaged out.

  51. Victor Iannello says:

    @Andrew: On the bright side, whatever issues there are should be quickly identified and resolved. Clearly, Boeing has liability. It will be interesting to see, after the final report is written, whether there is a determination that pilot error contributed to the crash.

  52. Niu Yunu says:

    ———–
    DennisW: “@Donald, Being the gatekeeper of the whacko category I inform you that you are IN.”

    DennisW: “@TBill, I think you are full of shit.”
    ———–

    @Victor: Sorry, but why do you tolerate such language on your blog ?

  53. Victor Iannello says:

    @Niu Yunu: I try to encourage respectful behavior, but I also don’t apply a heavy hand until the language or accusations become a distraction to the discussion. I sometimes get accused of both too much and too little moderation.

  54. airlandseaman says:

    Andrew: Given a full down elevator command from the new 727 trim system(s), how much stick pressure would be necessary to over ride the trim? Is it even possible?

    Faced with a sudden, unwelcome full down elevator trim situation…probably some negative Gs, multiple alerts, and other bad jumbi going on… how long do you think it would take the average commercial pilot to pull the right CBs or switches to regain manual control?

  55. TBill says:

    @Niu Yunu
    My interpretation was that I had won that debate.

    MH370-Possible Reasons for Intentional Descent at End of Flight

  56. Niu Yunu says:

    @Victor: Thank you for the feedback. I understand. I just wondered why nobody said anything. I’m certainly just oversensitive. Sorry.

    @TBill: best riposte 🙂

  57. airlandseaman says:

    JT610 final seconds: http://bit.ly/2F7mbCj

  58. airlandseaman says:

    The previous JT610 linked plots had some scale color errors, corrected here: http://bit.ly/2PdUITW

  59. Richard Godfrey says:

    @All,

    The link below gives an update to my drift results for beachings in the Southern Indian Ocean west of 65°E from various start latitudes near the 7th Arc in a similar format to David Griffin’s Fig. 5a:

    https://www.dropbox.com/s/bjdh6wonx2a7yzg/Drift%20Analysis%20Results%20West%20of%2065E%2007%20Nov%202018.pdf?dl=0

    There were 155 beachings out of a total of 190 simulator runs. 35 simulator runs ended up in mid-ocean gyres or heading out into the Southern Atlantic Ocean having rounded South Africa.

    The link below gives the list of beachings in excel form, including the days spent in transit for each simulator run:

    https://www.dropbox.com/s/bwjn3zm5ig0angk/Beaching%20Locations%2007Nov2018.xlsx?dl=0

  60. David says:

    @ALSM. Thanks for the JT610 plot.
    The 23:31:45.4 7500 ft/min roc looks spurious, particularly noting instantaneous acceleration to 9000 ft/min rod.

  61. airlandseaman says:

    David: I agree that it looks spurious. FR24 confirmed several spurious “transmission errors” in the previous flight’s ADS-B data set. Remember, the transponder modulation is a form of amplitude pulse modulation, which is more susceptible to transmission interference and error than more robust coded digital modulation schemes, like PSK.

  62. Andrew says:

    @ALSM

    RE: ”Andrew: Given a full down elevator command from the new 727 trim system(s), how much stick pressure would be necessary to over ride the trim? Is it even possible?”

    A lot. I doubt it would be recoverable at high speed.

    ”Faced with a sudden, unwelcome full down elevator trim situation…probably some negative Gs, multiple alerts, and other bad jumbi going on… how long do you think it would take the average commercial pilot to pull the right CBs or switches to regain manual control?”

    It shouldn’t take long, provided the pilot realises what’s happening. I suspect that was made more difficult in this case because of the stall indications and IAS disagreement. The pilot could have momentarily stopped the runaway STS by using the control column trim switches to trim nose up. However, the runaway STS would have started trimming nose down again 5 seconds after the pilot released the trim switches. The longer term fix is to select the STAB TRIM CUTOUT switches to CUTOUT to shut down the electric trim system, as recommended by Boeing in its latest FCOM bulletin. The cutout switches are located on the centre pedestal, within easy reach of both pilots. On the 737, the pilot could then manually trim the stabiliser using the stabiliser trim wheel, also located on the centre pedestal.

  63. Peter Norton says:

    @Andrew: Thank you for the links.

    “the AOA sensor can […] cause the jet to enter an aggressive dive”¹

    Is this as bad as it reads or am I over-reacting ?

    ———-

    Dominic Gates: “If the automatic trim system is fed false angle of attack (AOA) measurements, it will go awry. If the sensor tells the system the AOA is too high, the automatic system begins moving the tail to point the nose down. Even when the pilot corrects that by pulling the column […] the system will begin to pitch the nose down again.”²

    Boeing FCOM bulletin: “… can occur during manual flight only”

    • I’m surprised that even when supposedly flying “manually”, there is so much automation that can prevail over the manual inputs – even when applied with force – and force your plane into the ground.

    ———-
     
    John Cox (former pilot and chief executive of aviation consultancy Safety Operating Systems): “On no other 737 is there a system based on the angle of attack that will move the (horizontal tail) trim. That is unique to the MAX. I was surprised that a single angle of attack indicator could cause the activation of this system.”²

    AD: “if an erroneously high single angle of attack (AOA) sensor input is received by the flight control system, there is a potential for repeated nose-down trim commands of the horizontal stabilizer. This condition, if not addressed, could cause the flight crew to have difficulty controlling the airplane, and lead to excessive nose-down attitude, significant altitude loss, and possible impact with terrain.”

    • What happened to redundancy ?

    • And why is there no failsafe to do NOTHING in case of AOA sensor disagreement ?

    ———-

    Boeing FCOM bulletin: “Initially, higher control forces may be needed to overcome any stabilizer nose down trim already applied.”

    ALSM: “Given a full down elevator command from the new 727 trim system(s), how much stick pressure would be necessary to over ride the trim? Is it even possible?”

    Andrew: “A lot. I doubt it would be recoverable at high speed.”

    • Then why is this advised by Boeing ?

    ———-

    ALSM: “Faced with a sudden, unwelcome full down elevator trim situation…probably some negative Gs, multiple alerts, and other bad jumbi going on”

    Seattle Times: « Cox said he considers the MAX safe and would fly on one tomorrow without hesitation. […] Bjorn Ferhm, a former jet-fighter pilot and an aeronautical engineer […] believes the MAX is as safe as any 737. »²

    Victor Iannello: “Clearly, Boeing has liability.”

    • How is this “safe”, if from onset of the problem you have a mere ~10 seconds to identify the problem and hit both stab trim cutout switches, or your plane is beyond recovery and lost ? (And in these few seconds all hell breaks loose in terms of warnings, see the 9 items on p.2 of the bulletin, which makes calm thinking next to impossible.) This doesn’t read very safe to me:

    « the sensor may feed false information about this angle to the flight computer, which in turn triggers other errors. With the sensor falsely indicating that the nose is too high, when it isn’t, it causes an automatic system response that “trims” the horizontal tail of the plane to [put] the plane’s nose down. At the same time, it causes an indicator of the minimum speed to tell the pilot that the plane is near a stall, which also causes the pilot’s control column to shake as a warning. And the airspeed indicators on both sides of the flight deck disagree. The pilots can use extra force to correct the nose down trim, but the failure condition repeats itself, so that the nose-down push begins again 10 seconds after correcting. “If the nose is trimmed down on an aircraft, it becomes difficult for the crew to hold it,” said the person briefed on Boeing’s bulletin. “The nose is turning itself down and they are having to fight it. It takes a lot of effort to keep it from diving. Especially if you have a crew that’s confused and doesn’t know what’s going on.” […] Pilots are typically trained on how to handle a runaway trim situation, said the person briefed on the Boeing bulletin, but that’s with everything else working as it should. In this case, the control-column shaking, the stall warning and the air-speed-indicator disagreement all combine to create confusion and keep the pilots very busy. »³


    ¹ theaircurrent.com
    ² Seattle Times
    ³ Seattle Times

  64. Andrew says:

    @Peter Norton

    RE: “• I’m surprised that even when supposedly flying “manually”, there is so much automation that can prevail over the manual inputs – even when applied with force – and force your plane into the ground.”

    The STS helps with speed stability and is also part of the stall protection system. If a high AoA is detected, the STS is designed to trim the stabiliser nose down to increase the amount of rearward force the pilot must apply to further increase the AoA.

    “• What happened to redundancy ?

    • And why is there no failsafe to do NOTHING in case of AOA sensor disagreement ?”

    I don’t know. According to John Cox, the AoA input to the STS is a new feature that was introduced with the 737 MAX. The investigation might explore the basis for its certification by the FAA.

    “• Then why is this advised by Boeing ?”

    Because INITIALLY the pilot would be able to overcome the nose down trim by applying more force on the control column. Applying rearward force on the control column is only one of the steps that is needed to contain the failure.

    “• How is this “safe”, if from onset of the problem you have a mere ~10 seconds to identify the problem and hit both stab trim cutout switches, or your plane is beyond recovery and lost ? “

    The system doesn’t trim fully nose down in the first 10 seconds; it’s not that fast. However, if the pilot doesn’t do anything, the system will keep putting in more nose down trim until it reaches the nose down limit. Cox believes the aircraft is ‘safe’ because: “Countering the problem, if it arises in flight, should require only a standard pilot procedure — one that was newly emphasized by Boeing in its safety bulletin Tuesday.”

  65. Peter Norton says:

    @Andrew: thank you

    Andrew: “The system doesn’t trim fully nose down in the first 10 seconds; it’s not that fast.”

    How much time does the crew have to react?
    I.e.how long would you say it takes from the onset of the problem (first STS trim down) to full nose down trim (which I assume unrecoverable at this flight level) ?

    Andrew: “The investigation might explore the basis for its certification by the FAA.”

    Hopefully. It seems kind of a self-certification:

    “If the accident investigation ultimately were to find fault with Boeing’s design of the nose pitch control system, it would raise questions about the safety-certification process, much of which is delegated by the FAA to Boeing itself. Boeing on Wednesday declined to comment.” (Seattle Times)

    Personally, as a pilot, what is your thought about what we are reading here?
    • “I am not afraid/worried, because countering the problem requires only standard pilot procedure.”
    or more like
    • “This is creepy. Can lead to a really dangerous situation and urgently requires a software update.”
    ?

  66. Richard Godfrey says:

    Planes plummeting is getting a habit …

    https://www.google.de/amp/s/www.bbc.co.uk/news/amp/uk-46137445

  67. Victor Iannello says:

    @Peter Norton: We need more facts and a better understanding of Boeing’s design philosophy. It might be that if one AoA sensor indicates stall, and the other does not, the designers believed that automated protection against stall has higher priority than protection against an unplanned descent, as it would be easier for the crew to recover. After the AF447 crash, I think both designers and crew put a large emphasis on the avoidance of stall when sensors disagree.

  68. TBill says:

    @Andrew @all
    Re: Lion Air
    Did the pilots ever have have a chance to turn the AutoPilot on, or did the problem prevent that?

    While we are talking automated systems, the new system I want if is possible is logic to prevent aircraft diving into terrain.

    Regarding instrumentation, in industry we often use 3oo5 (3 out of 5 voting) for temperature readings required for automated emergency shutdown.

  69. DennisW says:

    @TBill

    Multiple sensors are a good thing, but the methodology of sensor fusion is complex.

    How did you combine the outputs of your five temperature sensors? Average? Something more complex such as picking the average of the three sensors having the smallest variance from the ensemble mean? I could go on and on…

    Darwin’s boat, the HMS Beagle, had 22 chronometers on board during the 2nd voyage. It is not known how Fitzroy combined the chronometer readings, but over the five years of the second voyage the time determined by the Beagle was in error by 33 seconds (8.5nm error). An amazing achievement considering the technology at the time, and the environment on board the Beagle. (Fitzroy was said to be dissappointed).

    Sorry for the distraction, but sensor fusion, Kalman filtering,…are pet topics of mine.

  70. Tim says:

    @All,
    So it would seem that the STS, unless it is isolated by the ‘Stab cutout’ switches could keep applying more and more incremental nose down trim. The pilots would have to pull back more and more putting great strain on the Stab screw jack. By the sudden dive at the end I suspect it finally gave way.

    What do you think?

  71. Richard Godfrey says:

    @All,

    Below is a link to an academic paper by Casner et al. 2013 entitled “The retention of manual flying skills in the automated cockpit”.

    Given 3 different problems in a simulator, the results were:

    (1) 98% of pilots verbalised there was a problem.
    (2) 75% of pilots cross-checked other instruments.
    (3) 69% of pilots deviated from the flight path (altitude, heading, approached stall).
    (4) 77% of pilots diagnosed the problem.

    It appears that pilots retain their basic manual flying skills, even when they are not used that much. However, only 13% of pilots, at least sometimes, spend time with the flight director, autopilot, and flight director all off, to keep their skills sharp.

    It appears that pilots lose their cognitive skills as a result of cockpit automation. More frequent and serious problems were observed, when a failure was introduced in the simulation. In the event of a failure, 69% of pilots deviated from the planned flight path.

    Eventually, 77% of pilots diagnosed the problem, but the question is whether the deviation from flight path had already taken too long, before diagnosis and remediation occurred.

    The stated purpose of cockpit automation is to afford pilots more time to concentrate on following the progress of the flight. The study shows that pilots spend 20% of their time during automated flight thinking about other things, whereas it is less than 7% of their time during manual flight when using instruments.

    https://www.dropbox.com/s/3np40ntawtjopvf/The%20retention%20of%20manual%20flying%20skills%20in%20the%20automated%20cockpit.pdf?dl=0

  72. Richard Godfrey says:

    @All,

    The link below gives an update to my drift results for beachings on mainland Africa west of 42°E from various start latitudes near the 7th Arc in a similar format to David Griffin’s Fig. 5b:

    https://www.dropbox.com/s/0b5b27vmka0y3d6/Drift%20Analysis%20Results%20West%20of%2042E%2008%20Nov%202018.pdf?dl=0

    Out of a total of 190 simulator runs, 78 beachings were on mainland Africa, 73 beachings on Madagascar and 38 simulator runs ended up in mid-ocean gyres or heading out into the Southern Atlantic Ocean having rounded South Africa.

    The link below gives an update to my drift results for beachings in the Southern Indian Ocean west of 65°E from various start latitudes near the 7th Arc in a similar format to David Griffin’s Fig. 5a (many thanks to @DrB who spotted a mistake in yesterday’s comment, where I had included 3 particles as beachings that ended mid-ocean):

    https://www.dropbox.com/s/bjdh6wonx2a7yzg/Drift%20Analysis%20Results%20West%20of%2065E%2007%20Nov%202018.pdf?dl=0

  73. Peter Norton says:

    > TBill: @Andrew @all
    > Did the pilots ever have have a chance to turn the AutoPilot on, or did the problem prevent that?

    I read the Boeing bulletin as saying as long as AOA data is erroneous, A/P is not available (p.2, abridged cit.):

    « pilots are reminded that an erroneous AOA can cause some or all of the following indications and effects:
    • inability to engage A/P
    • automatic disengagement of A/P
    • AOA DISAGREE alert »

    Sounds intuitive to me, since A/P always disengages in case of insufficient required data to fly on.

    But if so, why would Boeing make a point pointing out that this “AOA failure condition can occur during manual flight only” (p.1, emphasis NOT mine), when you end up flying manually either way due to A/P disengagement ?

  74. Peter Norton says:

    Victor Iannello: “It might be that if one AoA sensor indicates stall, and the other does not, the designers believed that automated protection against stall has higher priority than protection against an unplanned descent, as it would be easier for the crew to recover.”

    At low altitude?
    If AOA disagree after takeoff, design a dive?

    Victor Iannello: “After the AF447 crash, I think both designers and crew put a large emphasis on the avoidance of stall when sensors disagree.”

    If the dive is a design intention rather than a malfunction, the emphasis might have to return to avoidance of terrain.

  75. Victor Iannello says:

    @Peter Norton: I don’t know. A stall at low altitude is also very bad. As I said, we have to better understand the design criteria.

  76. Andrew says:

    @Peter Norton

    RE: “How much time does the crew have to react?
    I.e.how long would you say it takes from the onset of the problem (first STS trim down) to full nose down trim (which I assume unrecoverable at this flight level) ?”

    The STS uses the autopilot stabiliser trim, which has a range of motion of 0.05 to 14.5 units. My understanding is that with flaps retracted, the system trims at the low speed rate of 0.2 units/sec. The time it would take to reach the nose down limit obviously depends on the position of the stabiliser at the onset of the problem, but let’s say it starts at around 5 units. At 0.2 units/sec, it would take about 25 secs to trim fully nose down, assuming continuous operation. However, the Boeing FCOM Bulletin states that the pitch trim system trims the stabiliser nose down ‘in increments lasting up to 10 seconds’, which suggests that it won’t trim continuously to the nose down limit.

    RE: “Personally, as a pilot, what is your thought about what we are reading here?
    • “I am not afraid/worried, because countering the problem requires only standard pilot procedure.”
    or more like
    • “This is creepy. Can lead to a really dangerous situation and urgently requires a software update.”
    ?”

    I think it’s a problem that needs to be investigated and a fix found to prevent it happening again; however, you can put me in the first category – ‘I am not afraid/worried’. I don’t think it’s so serious that it requires grounding of the MAX fleet. I believe the steps taken by Boeing and the FAA to alert pilots to the problem and the existing procedures that should be used to deal with it are appropriate. There are plenty of things that can go wrong with an aircraft that could have catastrophic consequences if not handled correctly. The key is to make sure that pilots have the necessary information and training to deal with such problems.

  77. Andrew says:

    @TBill

    RE: “Did the pilots ever have have a chance to turn the AutoPilot on, or did the problem prevent that?”

    At this point, we don’t know. As @Peter Norton mentioned, the problem might have prevented them engaging the autopilot, or caused the autopilot to disengage if it was already engaged.

    RE: Regarding instrumentation, in industry we often use 3oo5 (3 out of 5 voting) for temperature readings required for automated emergency shutdown.”

    It sounds as though the stall protection system on the MAX only requires one of the two AoA vanes to detect a high AoA. Going to 2oo2 might have unintended consequences that make the system less safe in terms of stall protection. Let’s say the aircraft has a high AoA that is not detected by one of the AoA vanes. In a 2oo2 system the stall protection would not activate. I doubt that would get past the regulators!

  78. Andrew says:

    @Tim

    RE: “So it would seem that the STS, unless it is isolated by the ‘Stab cutout’ switches could keep applying more and more incremental nose down trim. The pilots would have to pull back more and more putting great strain on the Stab screw jack. By the sudden dive at the end I suspect it finally gave way.”

    That might be the case. The ADS-B data suggests that something catastrophic might have occurred in the last few seconds. Perhaps we’ll know more when the Indonesians release their preliminary report.

  79. Andrew says:

    @Richard Godfrey

    RE: “Below is a link to an academic paper by Casner et al. 2013 entitled “The retention of manual flying skills in the automated cockpit”.”

    Thanks for posting the paper; I am guilty of oversimplifying the skills retention problem in my earlier post. I think Casner et al are correct in that it’s the pilot’s cognitive skills that deteriorate, which leaves the pilot with little spare capacity to process other problems when confronted with a manual flying task.

  80. Don Thompson says:

    Andrew wrote “you can put me in the first category – ‘I am not afraid/worried […]’.

    I concur with Andrew’s view.

    It would be wrong to think that onset of the condition described by the FAA’s AD and considered to have led to the loss of PK-LQP was somehow insidious: the disparity in L & R ADIRU data (IAS/AOA DISAGREE flags) would result in a Master Caution alert and, if the stabiliser trim did ‘run away’, the sustained spinning of the trim wheels on either side of the throttle quadrant would have been unmissable. The standby attitude display, separate from the main L & R flight deck displays, provides a third instrument reference.

    Like PK-AKC/QZ8501, PK-LQP appears to have suffered from an interrmittent technical problem that persisted over a number of previous sectors and was not resolved by maintenance. The LCC business model relies on minimising time on the ground, maximising airframe utilisation. NTSC and the Indonesian regulator must ensure that maintenance is given primacy for serious serviceability issues, and that ‘tested/no fault found’ is not an acceptable practice.

  81. DennisW says:

    @Don

    I totally disagree with you and Andrew which is unusual.

    For Boeing to describe a procedural work around for what is obviously a design defect is unforgivable IMO. It is reminiscent of the Toyota unintended acceleration problem (which killed 89 people and injured 52 people). Fix the problem! Toyota did not send a notice to Lexus owners instructing them to put the car in neutral, turn off the ignition, and apply the brakes. Toyota did end up paying $1.2B in financial penalties because they knew the source of the problem, which was rare, and they continued to manufacture defective vehicles.

    IMO, the 737 MAX should be taken out of service until the issue is properly corrected and verified to be corrected. Boeing should be held liable for the damages sustained because of the problem.

    Sure, taking the plane out of service will cause a lot of pain to Boeing’s customers. That is just the way it should be.

  82. Barry Carlson says:

    @Don Thompson,

    … and that ‘tested/no fault found’ is not an acceptable practice.

    I take from that an implication to consider other dependencies that may have conspired in giving the pilots reason for writing up the fault being dealt with, should also be investigated, and they be noted along with the result(s) of any testing signed off too.

  83. TBill says:

    @DennisW
    As far as 3oo5 readings for automated emergency action, we are looking for temp runaways, so we are not too concerned about the average. We are just looking for sky high temperature readings. But an open thermocouple can throw a bad high reading, so it needs to be more than one data point.

  84. Andrew says:

    @DennisW

    RE: “IMO, the 737 MAX should be taken out of service until the issue is properly corrected and verified to be corrected. Boeing should be held liable for the damages sustained because of the problem.”

    I suspect we’ll have to agree to disagree on that point. For the record, I was not advocating that Boeing publish a work around procedure and leave it that. I said ‘…it’s a problem that needs to be investigated and a fix found to prevent it happening again’. Do I think the fleet needs to be grounded in the interim? No. At this point in the investigation we don’t know exactly what happened, or if the crew’s action (or inaction) contributed in any way. There are perfectly good procedures for dealing with an errant stabiliser trim that have been around for a very long time – flip the cut-out switches, fly the damn aircraft and trim manually.

    Many years ago I had an incident shortly after take-off in a Falcon 900. The aircraft’s stall protection system activated when we retracted the slats & flaps, resulting in lots of warning noises in the cockpit and the outboard slats continuously cycling in and out. Confusing at first, but it was obvious that we were nowhere near a stall condition, so we simply extended the slats & flaps (which stopped the warnings) and came back and landed. We admittedly did not have a runaway trim problem, but we did what we were trained to do and FLEW THE AIRCRAFT. It turned out that our problem was caused by a faulty AoA vane that had taken in some water after the aircraft was parked overnight in heavy rain.

  85. DennisW says:

    @Andrew

    Yes, it is OK to disagree, and I do see your point of view. My issue is based more on protocol than the specifics related to this incident.

  86. MH says:

    Yes the new B737max’s should be grounded until the problems are corrected
    The work around aren’t proven to prevent another crash.

  87. Peter Norton says:

    @Victor Iannello: I concur. Although I also sympathize with DennisW RE:Toyota.

    @Don Thompson: RE “unmissable, not insidious”: I think nobody contended that. I see the opposite problem: a situation where you get overwhelmed with the aural and visual errors listed on p.2 of the Boeing bulletin.

    • You have a stall warning/stick shaker.
    • You have uncommanded, hard to fight nose down trim.
    • maybe even a third type of error we don’t know yet

    I imagine the situation to be quite confusing. Are you stalling? (In which case the nose down trim would be helpful.) Or is the automatic trim system going rogue? Or is something else going on (in addition)? Which errors/instruments do you trust?

    Reading the last paragraph here, I could see how it would be difficult to handle the situation.

  88. Peter Norton says:

    @Andrew: Thank you for the provided timing.

    > @Andrew: “the Boeing FCOM Bulletin states that the pitch trim system trims the
    > stabiliser nose down ‘in increments lasting up to 10 seconds’, which suggests that it
    > won’t trim continuously to the nose down limit.”

    You mean after 10 seconds it makes a pause and continues later on?
    How long would that pause take ?

    Within the context of the described scenario and countermeasures, I had interpreted this information as “[only] if the pilot fights back the nose down trim for 10 seconds using the electric stabilizer trim switches, the trim system stops, but will restart 5 seconds after releasing the switches, [otherwise it will trim continuously to the nose down limit]”.

  89. Andrew says:

    @Peter Norton

    RE: “You mean after 10 seconds it makes a pause and continues later on?
    How long would that pause take ?”

    I’ve never flown the B737, but that’s how I interpret the FCOM Bulletin, which states:

    ‘In the event of erroneous AOA data, the pitch trim system can trim the stabilizer nose down in increments lasting up to 10 seconds. The nose down stabilizer trim movement can be stopped and reversed with the use of the electric stabilizer trim switches but may restart 5 seconds after the electric stabilizer trim switches are released…It it possible for the stabilizer to reach the nose down limit unless the system inputs are counteracted completely by pilot trim inputs and both STAB TRIM CUTOUT switches are moved to CUTOUT.’

    I don’t know how long it would pause between the 10 second increments. My understanding is that the STS stops immediately if the pilot uses the electric trim switches, but restarts 5 seconds after they are released.

  90. Andrew says:

    @Peter Norton

    RE: “I imagine the situation to be quite confusing. Are you stalling? (In which case the nose down trim would be helpful.) Or is the automatic trim system going rogue? Or is something else going on (in addition)? Which errors/instruments do you trust?”

    Yes, it certainly would be quite confusing. That’s where pilots need to be disciplined and quickly assess if the aircraft is still flying and which instruments are correct. If the aircraft has the correct pitch attitude and the correct thrust, then it is unlikely to be stalled if it was flying normally immediately before the warning occurred. The pilots can easily check which airspeed indications are correct by cross-checking the indications on their PFDs with the standby airspeed indicator.

  91. Peter Norton says:

    @Andrew:

    >> Peter Norton: “You mean after 10 seconds it makes a pause and continues later on?
    >> How long would that pause take ?”
    >
    > Andrew: “I’ve never flown the B737, but that’s how I interpret the FCOM Bulletin,
    > which states: […]
    > I don’t know how long it would pause between the 10 second increments.
    > My understanding is that the STS stops immediately if the pilot uses the electric trim
    > switches, but restarts 5 seconds after they are released.”

    I know, I have read the Bulletin, but I don’t find it overly logical that if the pilot does not “fight back”, the STS stops itself after 10s and takes a pause, possibly longer than the 5s pause it grants WHEN the pilot does “fights back” …

    > Andrew: “That’s where pilots need to be disciplined and quickly assess if the
    > aircraft is still flying and which instruments are correct. If the aircraft has the
    > correct pitch attitude and the correct thrust, then it is unlikely to be stalled if it
    > was flying normally immediately before the warning occurred.”

    FR24 data suggests JT610 had not been flying normally but at varying altitude before the crash, possibly due to the crew fighting the trim system in intervals.

    > Andrew: “The pilots can easily check which airspeed indications are correct by
    > cross-checking the indications on their PFDs with the standby airspeed indicator.”

    According to Skybrary“the standby airspeed indicator […] will not provide reliable information if the fault is within the pitot static system”. This may not have been the case for JT610, but AF447 sent an ACARS message at 02:10 (4 min before final transmission) indicating a fault in the pitot-static system.

    > Andrew: “Yes, it certainly would be quite confusing.”

    Even more so when you have to tune out that a few hundred people will die if you fail to stop whatever is occurring within probably less than 60s. It’s hard to think under that pressure.

    > Brian Anderson: “aircraft can be flown quite safely without any airspeed indication.
    > The standard procedure is to set the pitch [angle] to the appropriate value, and
    > set the power levers to the appropriate position, then if everything else is OK the
    > aircraft will fly straight and level.”
    >
    > Andrew: “Absolutely. A well trained, competent pilot should be able to fly and land
    > the aircraft quite safely with an unreliable speed indication. The procedures are
    > well defined and the required pitch attitude/thrust tables are readily available in
    > the pilot’s QRH.”

    I wonder why there is such stark disagreement among professional pilots. In the aftermath of the AF447 crash, Bruno Sinatti, president of Alter, Air France’s third-biggest pilots’ union, said: “Piloting becomes very difficult, near impossible, without reliable speed data”.

  92. Richard Godfrey says:

    @Andrew

    The AOA DISAGREE alert is only if the AOA indicator option is installed.

    Do we know whether this option was installed on PK-LQP?

  93. Andrew says:

    @Peter Norton

    RE: “FR24 data suggests JT610 had not been flying normally but at varying altitude before the crash, possibly due to the crew fighting the trim system in intervals.”

    Of course the aircraft wasn’t flying normally before it crashed, but it looks to me as though the aircraft climbed out pretty much normally until it reached an altitude just below 2,000 ft, which might be when the problems started.

    RE: “According to Skybrary“the standby airspeed indicator […] will not provide reliable information if the fault is within the pitot static system”. This may not have been the case for JT610, but AF447 sent an ACARS message at 02:10 (4 min before final transmission) indicating a fault in the pitot-static system.”

    Large airline aircraft (including the 737 MAX) typically have three independent pitot-static systems; one for the Capt, one for the FO and another for the standby instruments. A failure in one system should not affect the other systems.

    Pitot probe icing, as occurred in AF447, is one of the few things that can affect all three pitot-static systems. In the AF447 accident, the ADR1 (ie Capt) airspeed data was only invalid for 29 seconds before it recovered, while the ADR3 (ISIS or standby) data was invalid for about one minute. It’s not known how long the ADR2 data was invalid, because it wasn’t recorded on the FDR, but it probably recovered at about the same time. Consequently, the pilots of AF447 probably had valid airspeed data on all their instruments within about one minute of the first indication of a problem. If they had done nothing more than maintain the aircraft’s attitude and thrust setting, it’s very likely that 228 people would not have lost their lives.

    RE: “Even more so when you have to tune out that a few hundred people will die if you fail to stop whatever is occurring within probably less than 60s. It’s hard to think under that pressure.”

    It might sound callous, but pilots don’t spend their time fretting about the hundreds of people sitting behind; they are more concerned with keeping the aircraft safe. If it’s safe, then by extension everyone on board (including the pilots!) will be safe.

    RE: “I wonder why there is such stark disagreement among professional pilots. In the aftermath of the AF447 crash, Bruno Sinatti, president of Alter, Air France’s third-biggest pilots’ union, said: ‘Piloting becomes very difficult, near impossible, without reliable speed data’.”

    Well he would say that, wouldn’t he? Piloting does become very difficult, near impossible, without reliable speed data if the aircraft isn’t flown correctly. That’s why there is now far greater emphasis on unreliable airspeed (UAS) procedures during training. We regularly practice UAS procedures in the simulator and I can assure you the aircraft can be safely flown and landed without airspeed data. A number of years ago there was an incident involving a B747-400 at the airline where I work. The aircraft was on its first flight after heavy maintenance and lost all airspeed information shortly after take-off. That crew followed the procedures and landed the aircraft quite safely.

  94. Andrew says:

    @Richard Godfrey

    RE: “Do we know whether this option was installed on PK-LQP?”

    I haven’t seen anything official, but I did read a report that said the aircraft did not have the AOA indicator option installed.

  95. Paul Smithson says:

    @David @Andrew. I’d like to circle back to the issue of AP disengagement at end of flight.

    Originally, we had understood that loss of pitot heat was the trigger, moreover that it was loss of power to pitot heat and not drop in temperature. From Andrew & David’s earlier exchange “the loss of pitot heat alone does not cause reversion to SECONDARY mode unless the air data actually becomes invalid”. Question: other than icing conditions, is there any reason to expect the air data to become invalid? If icing conditions are unlikely until you reach cloud / vapour below (say) FL200, would you not expect air data to remain valid as long as the aircraft is at altitude?

    Second, Andrew mentioned that “..recent advice indicates the reversion is caused by the loss of other sensors required by the PFCS.” and that “my understanding is that they are mostly position sensors associated with actuators controlled by the L2 ACE. The L2 ACE is only powered by the L28V DC bus, which is normally supplied by the L AC transfer bus.” Question: In event of loss of power after flameout, are these circuits among those that failover to power from hot battery?

    If those actuator sensors do receive backup battery power and airspeed data will not go bad at altitude even without pitot heat, then I don’t see that we have good reason to expect near-immediate loss of AP and reversion to SECONDARY at flameout. That in turn opens the door to possible un-manned controlled flight: whether an unpowered “glide” of sorts or powered flight with left engine still providing thrust but not electrical power under “alternate electrical configuration”

    In considering these possibilities, please suspend belief (if you can) in traditional interpretation of the final BFOs.

  96. Peter Norton says:

    > Andrew: “pilots need to […] quickly assess if the aircraft is still flying and which
    > instruments are correct. If the aircraft has the correct pitch attitude and the correct
    > thrust, then it is unlikely to be stalled if it was flying normally immediately before the
    > warning occurred.”
    >
    > Peter: “FR24 data suggests JT610 had not been flying normally but
    > at varying altitude before the crash, possibly due to the crew fighting the trim system
    > in intervals.”
    >
    > Andrew: “Of course the aircraft wasn’t flying normally before it crashed, but it looks
    > to me as though the aircraft climbed out pretty much normally until it reached an
    > altitude just below 2,000 ft, which might be when the problems started.”

    What I was trying to say is that AF447 was, as you described, “flying normally immediately before the warning occured”. On JT610, however, starting at 23:22 (~1min into the flight) until the crash at 23:32 there were many ups and downs, which could have been disorientating and unsettling, given that the trim system possibly developed a life of its own. This in combination with all the alarms going off and the stick shaker might have created an overall situation in which a stall wasn’t dismissable out of hand. Plus the nose down trim would also be consistent with a stall.
    I could see why one could be confused and have doubts in this situation.

  97. Andrew says:

    @Paul Smithson

    RE: ”Question: other than icing conditions, is there any reason to expect the air data to become invalid? If icing conditions are unlikely until you reach cloud / vapour below (say) FL200, would you not expect air data to remain valid as long as the aircraft is at altitude?”

    The air data should remain valid unless the aircraft encounters icing conditions. That is not very likely at high altitude.

    RE: ”Question: In event of loss of power after flameout, are these circuits among those that failover to power from hot battery?”

    No.

    RE: ”I don’t see that we have good reason to expect near-immediate loss of AP and reversion to SECONDARY at flameout.”

    The PFCS does revert to SECONDARY when AC power is lost at flameout. The only question was why that occurrs.

  98. Paul Smithson says:

    @Andrew. Thanks for the quick response. “The PFCS does revert to SECONDARY when AC power is lost at flameout”. How do we know that definitively?

  99. Richard Godfrey says:

    @Andrew

    I am sure that you are a disciplined and well trained pilot.

    I am sure that your colleagues are disciplined and well trained pilots.

    It is the others that I am concerned about, especially when experiencing a “wow” situation and having a short time to diagnose and remediate the situation.

    To quote you, two days ago, “Erroneous angle of attack data causing the aircraft to trim nose down, together with spurious stall warning and airspeed disagreement. Wow…”

    Why is the AOA DISAGREE indicator an option?

    Surely it should be mandatory, if a system such as the STS relies on a single AOA sensor input.

  100. Victor Iannello says:

    @Paul Smithson asked: “The PFCS does revert to SECONDARY when AC power is lost at flameout”. How do we know that definitively?

    Boeing has confirmed that the loss of both IDGs and backup generators would cause a flight control degradation to SECONDARY because some critical data, such as total air temperature, would be lost.

  101. Paul Smithson says:

    @Victor. OK, I’ll take that (reversion to secondary and loss of AP) as read, then. I’d also like to probe the assumption that in secondary mode without
    manual input the aircraft will quickly become unstable. From what I can see online, control surfaces are still automatically adjusted without pilot input to keep the plane in equilibrium, albeit in a slightly cruder manner and without envelope protection.

    “Secondary mode Boeing Secondary mode is somewhat similar to the Airbus Alternate Law. When the PFCs can not support Normal mode operation due to internal faults or to loss of information from other aircraft systems, they automatically revert to Secondary mode. Reversion to Secondary mode results in the loss of the autopilot and the pilots must control the aircraft manually. The ACEs still receive pilot control inputs and send the appropriate signals to the PFCs. However, due to the degraded mode of operation, the PFCs use “simplified” computations to generate the flight control surface commands. These commands are sent back to the ACEs from whence they are sent to the flight control surfaces in the same manner as during Normal mode operations.

    Aircraft handling qualities are affected by the simplified computations or PFC control laws that are utilised in Secondary mode. While all flight control surfaces remain operative, the elevator and rudder are more sensitive at some airspeeds. The following functions are inoperative or degraded during Secondary mode operations:

    autopilot
    auto speedbrakes
    envelope protection
    gust suppression
    tail strike protection
    thrust asymmetry compensation
    yaw damping

  102. Andrew says:

    @Paul Smithson

    Our company test pilots have also confirmed the above behaviour. They regularly simulate the complete loss of AC electrics during maintenance test flights, to check the RAT auto deployment and APU auto start.

  103. Paul Smithson says:

    @Andrew. Thanks. By “the above behaviour” I’m assuming you are referring to the reversion to secondary mode and AP drop-out. What do your simulation tests reveal about the stability of the aircraft without AC electrics, AP disengaged, control mode secondary and no pilot inputs?

  104. Andrew says:

    @Richard Godfrey

    Thank you. My ranting was not meant to imply that every UAS event will have a happy ending. Some pilots clearly continue to be caught unawares, especially when confronted with other problems. Nevertheless, it is possible to fly the aircraft safely without airspeed information.

    I don’t know why the AOA DISAGREE caution is not available without the AOA indication option.

  105. Victor Iannello says:

    @Paul Smithson: You are asking about what the “simulation tests reveal about the stability of the aircraft without AC electrics, AP disengaged, control mode secondary and no pilot inputs?”

    This is what Boeing attempted to answer by conducting the end-of-flight simulations. From a previous post:

    If we consider that the end-of-flight Boeing simulations were representative of the actual conditions of MH370, we have three possibilities:

    1. If the flight was uncontrolled after fuel exhaustion, and if we ignore the final BFO values, the plane impacted no farther than about 15 NM from the 7th arc.

    2. If the flight was uncontrolled after fuel exhaustion, and if the descent rates suggested by the final BFO values are valid, then the aircraft impacted no more than about 8 NM from the 7th arc.

    3. If the flight was controlled after fuel exhaustion, an efficient, stable glide starting at about 40,000 ft could have achieved a distance of 120+ NM from the 7th arc.

  106. Paul Smithson says:

    @Victor. Fair enough. Do you happen to know what would happen if the aircraft had earlier dropped into secondary and AP disengaged due to left side power issues BUT AP was manually re-engaged at that earlier point? Does it drop out afresh at final flameout of does the earlier manual re-engagement “latch” it into an ON condition?

  107. Victor Iannello says:

    @Paul Smithson: I’m not sure I understand your question, but each time both transfer busses simultaneously lose power, the flight control degrades to SECONDARY, and the mode can only transition to NORMAL if the power returns to a transfer bus AND the PRIMARY FLIGHT COMPUTERS switch is cycled to DISC and to AUTO.

  108. Andrew says:

    @Paul Smithson

    RE: ”By “the above behaviour” I’m assuming you are referring to the reversion to secondary mode and AP drop-out. What do your simulation tests reveal about the stability of the aircraft without AC electrics, AP disengaged, control mode secondary and no pilot inputs?”

    Yes. In my experience in the simulator, the aircraft rolls to the right and descends.

  109. Richard Godfrey says:

    @Andrew

    You stated “I don’t know why the AOA DISAGREE caution is not available without the AOA indication option.”

    Fair enough! You are not an employee of Boeing in charge of their sales tactics.

    My question was poorly worded, my apologies.

    What I should have asked was:

    “As a pilot, familiar with Boeing aircraft, do you agree, that if a system such as the STS relies on a single AOA sensor input, that it should be mandatory to install the AOA indication option, so that in the event that one of the AOA sensors fails, the pilot is given an alert that AOA DISAGREE and can quickly diagnose the problem?”

    I think it is bad practice that Boeing offer the AOA indication as an option. It should be mandatory.

    Having worked on the design of aircraft systems, where a stringent single and double failure analysis has to be performed, I would regard this as a design (or at least a commercial) failure from Boeing and potentially a case of gross negligence.

  110. Peter Norton says:

    > Richard Godfrey says:
    > @Andrew, I am sure that you are a disciplined and well trained pilot.
    > I am sure that your colleagues are disciplined and well trained pilots.
    > It is the others that I am concerned about, especially when experiencing a “wow”
    > situation and having a short time to diagnose and remediate the situation.
    > To quote you, two days ago, “Erroneous angle of attack data causing the aircraft to
    > trim nose down, together with spurious stall warning and airspeed disagreement. Wow…”

    I didn’t state that because I found it obvious, but now that it’s there, I want to add for the record, that this is also where I was coming from.

    I also agree with Richard Godfrey on the comment directly above. It seems slightly weird to offer essential security features as options.

  111. Ventus45 says:

    This seems to indicate that OI will be not be extending the search for ARA San Juan beyond the contracted 60 actual search days.
    https://pbs.twimg.com/media/Drl4SOkXcAElWiu.jpg
    The part in heavy type:-
    “nos maniferaron que el buque no retornara a puerto alguno en Argentina una vez finalizados los 60 dias operativos y que seremos trasbordados a un buque de la armada”
    Appears to translate (google) as:
    “they made sure that the ship did not return to any port in Argentina once the 60 days were over and that we will be transferred to a navy ship”

  112. DrB says:

    Ocean Infinity will be ending their initial 60-day search phase for ARA San Juan on 16 November. Then Seabed Constructor will sail directly to South Africa for refurbishment. The Argentine Government has extended their contract for 120 calendar days, from ~ 16 November until mid-April 2019. Ocean Infinity plans to return in February and perform a second 60-day search for ARA San Juan. It is unlikely any MH370 search could be performed by OI this (southern) summer, even if funding were to become available.

  113. Andrew says:

    @Richard Godfrey

    RE: “As a pilot, familiar with Boeing aircraft, do you agree, that if a system such as the STS relies on a single AOA sensor input, that it should be mandatory to install the AOA indication option, so that in the event that one of the AOA sensors fails, the pilot is given an alert that AOA DISAGREE and can quickly diagnose the problem?””

    Yes, I agree. It seems that AOA input to the STS is a new feature of the 737 MAX. I am surprised that the AOA DISAGREE caution is an option, given that the system appears to require a ‘stall’ indication from only one AOA vane to activate the stall protection. I don’t know the logic behind the certification of this new feature and I hope that aspect is covered by the accident investigation. Perhaps we’ll see a recommendation that the AOA DISAGREE caution be made mandatory, or perhaps the FAA will do so of its own accord. We might also see a requirement for other modifications, which might include removal of the AOA input to the STS if it is found to be a major problem. Time will tell.

    Of interest, a Transavia B737-800 experienced a very similar problem with a faulty AOA vane earlier this year. That aircraft was fitted with the AOA indication/DISAGREE option. The crew of that aircraft safely returned to the departure airfield, but they did not have to contend with the STS problems that seem to have occurred in the Lion Air accident, as the aircraft was not a MAX variant. The Transavia incident is still under investigation by the French BEA.

    https://www.aeroinside.com/item/11033/transavia-france-b738-at-paris-on-feb-8th-2018-ias-aoa-and-alt-disagree
    http://www.avherald.com/h?article=4b4f3b2b&opt=0
    https://www.bea.aero/en/investigation-reports/notified-events/detail/event/alarme-sonde-dincidence-defaillante-en-montee-initiale-demi-tour/

  114. DennisW says:

    @DrB

    It is unlikely any MH370 search could be performed by OI this (southern) summer, even if funding were to become available.

    There will never be another search for MH370.

  115. DrB says:

    @DennisW,

    You said: “There will never be another search for MH370.”

    I think there will be, but it won’t be government funded, and it won’t happen soon. At some point OI will have competitors who might want to best OI’s null result. It’s also possible that OI will find ARA San Juan in the near term and recoup their investment in that effort, in which case OI might then decide to go back and look for MH370 if they felt there was a reasonable chance of success within the scope of a new search that they could afford.

  116. Victor Iannello says:

    @Andrew: Why do you suppose that AoA is an option and not standard? My guess is that some companies might believe the information is superfluous and confusing since the PFD already displays the stall margin via the speed barber pole and via the pitch limit indicator (PLI).

    If the AoA is not indicated, the logic for not displaying AOA DISAGREE might be that because there are not multiple AoA indications to interpret, there is no need to warn about disagreement. Instead, there is a warning that the indicated speeds do not agree.

    But as discussed, this also makes it harder to diagnose the runaway STS condition when of the two AoA sensors disagree.

    It will be interesting to learn of Boeing’s design criteria, and whether there will be changes to displays and/or functionality.

  117. Andrew says:

    @Victor Iannello

    RE: “Why do you suppose that AoA is an option and not standard? My guess is that some companies might believe the information is superfluous and confusing since the PFD already displays the stall margin via the speed barber pole and via the pitch limit indicator (PLI).”

    I think the main reason is simply because dedicated AOA indicators have not been used previously in many commercial aircraft, even though military fast jet aircraft have had them for years. Consequently, most commercial pilots have never been trained how to use them and some companies see them as superfluous when other instruments provide an indirect indication of the aircraft’s AOA. I guess it all comes down to the airline’s perception of cost vs benefit.

    The following article from Boeing’s AERO magazine discusses the use of AOA indicators in commercial aircraft:
    http://www.boeing.com/commercial/aeromagazine/aero_12/aoa.pdf

    RE: “If the AoA is not indicated, the logic for not displaying AOA DISAGREE might be that because there are not multiple AoA indications to interpret, there is no need to warn about disagreement. Instead, there is a warning that the indicated speeds do not agree.”

    That’s probably correct. It will be interesting to see if there’s a change of thinking given this accident and the potential impact of a faulty AOA vane on the control of the 737 MAX aircraft.

  118. Andrew says:

    The following is a presentation by the late Captain Warren Vanderburgh about control malfunctions and flight instrument anomalies, copied from the JT610 thread on PPRuNe. It’s about 20 years old, but still very relevant today:
    https://youtu.be/WfNBmZy1Yuc

  119. Richard Godfrey says:

    @Victor,

    You asked “Why do you suppose that AoA is an option and not standard?”

    What I said yesterday in my comment at 5:26 am was “The AOA DISAGREE alert is only if the AOA indicator option is installed.”

    The Boeing FCOM states “Additionally pilots are reminded that an erroneous AOA can cause some or all of the following indications or effects:

    AOA DISAGREE alert (if the AOA indicator option is installed)

    …”

    Please see the Boeing FCOM bulletin:
    https://pbs.twimg.com/media/DrbMs-VVsAA9NgB.jpg:large

  120. Richard Godfrey says:

    @Andrew

    Many thanks for your thoughtful and informative reply to my question regarding the AOA indicator option.

    http://mh370.radiantphysics.com/2018/11/04/crash-debris-from-lion-air-jt610-provides-clues-about-mh370/#comment-19891

  121. Richard Godfrey says:

    @Andrew, @Victor,

    The Boeing AERO magazine article Andrew posted is very informative on the subject of a separate AOA indicator.

    I understand why the AOA indicator or gauge on the Primary Flight Display (PFD) is an option, when you already have Stick Shaker (stall warning), Pitch Limit Indicator (AOA margin to stall warning) and Speed Tape Indicator (airspeed relative to speed limits).

    I appreciate Victor’s point that an erroneous AOA can cause up to 9 indications or effects, but I think of these the AOA DISAGREE alert is the most helpful in quickly diagnosing the problem.

    Therefore, I do not understand why Boeing only present the AOA DISAGREE alert if the AOA indicator option is installed. As Andrew says, that may be one outcome of the Lion Air accident investigation, to change the AOA DISAGREE alert to mandatory, whether or not the additional AOA indicator option on the PFD is installed.

  122. Richard Godfrey says:

    @Victor

    My apologies, I misinterpreted your question to Andrew: “Why do you suppose that AoA is an option and not standard?”

    On first reading I had understood – what evidence do you have that the AoA indicator is an option and not standard?

    On second reading I had understood – why do you think Boeing offers the AoA indicator as an option and not standard?

  123. Victor Iannello says:

    @DennisW said: There will never be another search for MH370.

    It comes down to whether the value proposition for the search changes. That requires either a reduction in the cost to search per unit area, or new knowledge or insight to limit the search area. On this blog, we’ve focused on the latter, arguably with no success. The former is not likely to occur for some years, and will require step change in technology the way OI’s underwater drones were a step change compared to the towfish.

  124. DennisW says:

    @Victor

    Yes. Another factor is the value associated with finding the wreckage. I don’t believe anyone who might write a check believes it was a aircraft failure issue, hence nothing to be learned in that domain. FDR will likely show the plane was flown to location where it was found. CVR will likely be blank. What is the driver for finding the wreckage? Some people will probably say closure for the NOK.

    The answers to this mystery lie in Kuala Lumpur IMO.

  125. Victor Iannello says:

    @DennisW: So how do we advance the ball forward? Or do we just give up?

  126. DennisW says:

    @Victor,

    I honestly feel we have a better handle on the aircraft location than ever before. Richard’s drift work puts a very reasonable limit of 20S on the 7th arc. If we can convince ourselves that +/- 25nm is also reasonable, that combination may prove attractive to OI or a similar entreprenuer.

    I have no other thoughts on your question.

  127. Victor Iannello says:

    @DennisW: I can make the case that we have the best chance to find the plane by searching at a width of +/-25NM to the north of 25S latitude. If there were no pilot inputs after 19:41, that’s where the plane is. If there were pilot inputs after 19:41, the search width might increase to +/-125 NM, the latitude range along the 7th arc would be large, and the search becomes impractical using current technology, in my opinion.

    So, how do we convince ourselves that there were no pilot inputs after 19:41?

  128. DennisW says:

    @Victor

    The BFO values at 00:19:XX have me convinced there were no pilot inputs at or after that time. There is nothing to suggest otherwise, and supporting data (no IFE log-in) that the plane plane came down shortly after 00:19:37. The APU shutdown is a fabrication (it is possible) to suggest a pull out and glide.

    If the above is true the previous search up to 25S at +/-25nm can be deemed valid subject only to a failure to detect the wreckage (I give that a very generous 10% probability, weighted by the 13 out of 18 degrees searched). 18 degrees being the range of latitude from 38S to 20S I am proposing.

    Richard’s work has convinced me that a terminus North of 20S is very unlikely. Based on this logic it is more than 80% probable that the wrechage is located between 25S and 20S latitude on the 7th arc, and within +/- 25nm of the arc.

    I have presented all the above in previous posts. There has been almost zero feedback except from the dive followed by glide enthusiasts who weigh in with the possible APU shut down. Sure, they could be right. More likely they just want to stir their pot. I can think of no way to make the dive only scenario any more compelling than it is (and it is pretty damn compelling, IMO).

  129. Richard Godfrey says:

    @All,

    Victor posted a flight path ending further north on the 7th Arc at 21.9666°S, back in June 2018.

    http://mh370.radiantphysics.com/2018/06/08/an-mh370-flight-path-ending-further-north-on-7th-arc/

    I have now completed the drift analysis centred on a possible MH370 end point at 21.9666°S 103.5739°E.

    The results are shown in the 3 links below (overview, table and 25 simulation runs):

    https://www.dropbox.com/s/3ygmt7ycx8ukpvk/Array%20End%20Points%20Victor.pdf?dl=0

    https://www.dropbox.com/s/54y0hrjjoesw5em/Array%2021.9666S%20103.5739E%20100%20NM.xlsx?dl=0

    https://www.dropbox.com/sh/ij76gp2bp2orjwq/AADaLrPQUchZ2SmFK1ihOdFNa?dl=0

    Comparing the new results with those previously published, using a simple scoring system, gives the following:

    End point at 31.6°S scores 7.

    End point at 24.0°S scores 37.

    End point at 21.9666°S scores 25.

    With an uplift of the Flaperon speed to Reunion of 15%, gives the following:

    End point at 31.6°S scores 7.

    End point at 24.0°S scores 35.

    End point at 21.9666°S scores 21.

    With an uplift of the Flaperon speed to Reunion of 30%, gives the following:

    End point at 31.6°S scores 23.

    End point at 24.0°S scores 10.

    End point at 21.9666°S scores 13.

    The data to beachings on mainland Africa, demonstrates bifurcation around Madagascar from an end point near the 7th Arc at 24.0°S and 21.9666°S, but not from 31.6°S.

    The drift angle for the Flaperon to Reunion based on theoretical calculations by Pengam was 18° or 28° left of wind depending on orientation. David Griffin’s empirical analysis showed a drift angle left of wind up to 30° and typically between 10° and 20°. The drift angle from the 7th Arc to Reunion from an end point at 31.6°S was on average 36.5° left of wind, for 24.0°S was on average 29.7° left of wind and from 21.9666°S was on average 33.1°S left of wind from the simulation runs.

    I conclude that a MH370 end point between 25°S and 20°S near the 7th Arc, in an area not yet searched, fits with the drift analysis. I do not exclude an MH370 end point at 31.6°S near the 7th Arc, but this area has already been searched.

  130. Richard Godfrey says:

    @Dennis

    I agree with you that there is a high probability that MH370 will be found between 25°S and 20°S within ±25 NM of the 7th Arc.

    The rate of descent analysis from the BFO data at 00:19:29 and 00:19:37 UTC shows an accelerating high speed dive in excess of 15,000 fpm. Both the Boeing simulations and the tragic accident of JT610 show that this rate of descent can develop in under 10 seconds and within 30 seconds of impact (much shorter for JT610 as the aircraft had only reached an altitude of maximum 5,475 feet).

  131. Andrew says:

    For those interested in JT610, the following was allegedly released to its members by the Allied Pilots Association, which represents the pilots of American Airlines. The information allegedly came from Boeing. I can’t confirm its authenticity.

    “MCAS (Maneuvering Characteristics Augmentation System) is implemented on the 737 MAX to enhance pitch characteristics with flaps UP and at elevated angles of attack. The MCAS function commands nose down stabilizer to enhance pitch characteristics during steep turns with elevated load factors and during flaps up flight at airspeeds approaching stall. MCAS is activated without pilot input and only operates in manual, flaps up flight. The system is designed to allow the flight crew to use column trim switch or stabilizer aislestand cutout switches to override MCAS input. The function is commanded by the Flight Control computer using input data from sensors and other airplane systems.

    The MCAS function becomes active when the airplane Angle of Attack exceeds a threshold based on airspeed and altitude. Stabilizer incremental commands are limited to 2.5 degrees and are provided at a rate of 0.27 degrees per second. The magnitude of the stabilizer input is lower at high Mach number and greater at low Mach numbers. The function is reset once angle of attack falls below the Angle of Attack threshold or if manual stabilizer commands are provided by the flight crew. If the original elevated AOA condition persists, the MCAS function commands another incremental stabilizer nose down command according to current aircraft Mach number at actuation.”

  132. Peter Norton says:

    @Andrew: Very interesting. Assuming it’s true, what is your take on it?

  133. Andrew says:

    @Peter Norton

    I don’t have much to add except that the MCAS on the 737 MAX seems to operate similarly to the STS found on the older variants, except that it has an AOA input and only operates when the flaps are UP. That might explain why the ADS-B data seems to show the aircraft’s flight path only became ‘wobbly’ after it reached an altitude just below 2,000 ft, which is about where the flaps would have been retracted.

  134. Mick Gilbert says:

    @Andrew

    The concerning passage in that alleged APA memo was the bit that followed the MCAS description;

    This is the first description you, as 737 pilots, have seen. It is not in the AA 737 Flight Manual Part 2, nor is there a description in the Boeing FCOM. It will be soon.

    If true, that’s somewhat astounding I would think.

  135. Peter Norton says:

    > Andrew: “That might explain why the ADS-B data seems to show the
    > aircraft’s flight path only became ‘wobbly’ after it reached an
    > altitude just below 2,000 ft, which is about where the flaps would
    > have been retracted.”

    Compare this to the problems on the previous flight(s):

    « The previous flight of the accident aircraft, JT043 from Denpasar, showed similar erratic values in altitude and airspeed immediately after takeoff. After 8 minutes the ADS-B reported values stabilized. An unverified tech log item for that flight […] remarked “Airspeed unreliable and alt disagree shown after take off. […] Identified that CAPT instrument was unreliable and handover control to FO.” […] Lion Air did confirm that the aircraft had a “technical problem the previous flight which had been resolved according to procedure”. »¹

    This reference was later replaced by:

    « The NTSC reported that an airspeed indication anomaly did occur on the accident aircraft’s last 4 flights. »²


    ¹ https://web.archive.org/web/20181029200558/https://aviation-safety.net/database/record.php?id=20181029-0
    ² https://aviation-safety.net/database/record.php?id=20181029-0

  136. DrB says:

    @Richard Godfrey,

    You said: “I conclude that a MH370 end point between 25°S and 20°S near the 7th Arc, in an area not yet searched, fits with the drift analysis. I do not exclude an MH370 end point at 31.6°S near the 7th Arc, but this area has already been searched.”

    Looking at your predictions near the 7th Arc, I cannot discern any significant, systematic differences between distances of 0 NM, 25 NM and 75 NM (or possibly more). Thus, so far, your drift predictions, with a limited number of trials, do not provide any discrimination of distance from the arc (and that may not even be possible), at least for separations < 75 NM. That said, if you cannot exclude 31.6 S as being an end point of MH370 within 25 NM of the arc, neither can you exclude 75 NM from the arc at 31.6 S. In simple terms, the drift models have not been shown to be capable of determining likely distance (if < 100 NM) from the arc, so they can’t be used to include or exclude end points based on distance from the arc, just, hopefully, based on latitude.

    It is important to understand that the debris beachings are generally low in probability, and, according to David Griffin's model, most of the debris never makes landfall in the Indian Ocean (he estimates only 1/3 beaches there). He uses a very large number of trials – 76,200 in fact. That does not inherently improve accuracy, but it certainly improves the precision of his predictions (by a factor of about 20X compared to your recent 199 trials).

    As you know, David Griffin’s model predicts different fractions going north and south of Madagascar than your model predicts. The problem is, we can't know from the debris locations what the real probability of landfall is, even in a relative sense, between north and south landfalls. We have only one item (the Pemba flap) that we know traveled north around Madagascar. Was it the only one, or were there many more debris which went to the north but were not found or were found but not reported? In the absence of a systematic search along the entire Africa coastline, I don't think we can conclude much from the number of debris reports, except to say that some went north and some went south. Most hit Madagascar, and a small error in the beaching model can have a large impact on how many pieces go around Madagascar and how many of those pieces beach in Africa. In particular, the rocky beach in the northern shoreline of Madagascar may be particularly difficult to model.

    I am in the process of comparing the two drift model predictions in great detail, and I will soon post some results intended to clarify and compare their predictions for a range of crash latitude This includes their geographical distributions, divided into 6 zones, and their predicted drift times compared to debris reports, also segregated into the same 6 zones.

  137. Andrew says:

    @Peter Norton

    I guess it depends on how you define ‘immediately’. The JT43 flight path seems pretty much normal until it reached about 1,500 ft. Again, that’s about the time the flaps would have been retracted:
    https://www.flightradar24.com/blog/wp-content/uploads/2018/10/JT43-Alt-Ground-Speed-Vspeed-First-11-Minutes.jpg

  138. DennisW says:

    @DrB

    That said, if you cannot exclude 31.6 S as being an end point of MH370 within 25 NM of the arc, neither can you exclude 75 NM from the arc at 31.6 S.

    It is true that the drift analytics cannot exclude distances greater than +/-25nm from the arc, but the weight of other evidence does, IMO. The rate of descent at 00:19:XX strongly suggests an impact near the 7th arc. The lack of an IFE log-in strongly suggests an impact near the 7th arc. There is nothing that supports your point of view. Why do you cling to it?

  139. Andrew says:

    @Mick Gilbert

    RE: “If true, that’s somewhat astounding I would think.”

    Yes, it is surprising. However, as I said to you privately, I don’t think it’s necessarily a critical factor in this accident given the similar behaviour of the STS stall protection, minus the AOA input.

    The crew of JT610 should have been aware of the previous problems and the maintenance work that had been completed. That history should have been a warning to them of the potential for similar problems to occur on their flight. Having accepted the aircraft, the Captain should have carefully reviewed the ramifications with the FO, together with the likely actions that would be needed if the problem did recur. The CVR, if it’s found, should allow the investigators to determine if that occurred.

  140. David says:

    @Andrew. “Having accepted the aircraft, the Captain….”
    Interesting that. Is that the way with airlines generally? To me it is the ground crew with their access to detailed manuals and, when required, professional engineers, including recourse to others outside the airline, who have prime responsibility for “technical airworthiness” and the clearing of a defect. That could entail local maintenance test flights with a maintenance test pilot and if needed engineers to observe, to confirm the defect has been cleared, before it is released.

    Expensive on an airline certainly.

    I would expect requirements like that to be regulated, not just left to an airline or indeed a pilot.

  141. DennisW says:

    @Andrew

    The problem existed over several previous flights, and there was little reason to assume that it was remedied by maintenance. The aircraft should have been taken out of service, and 189 people would still be alive.

    Sorry, I do not subscribe to your views on how the problem should have been dealt with then or how it should be dealt with now. That aircraft needs to be taken out of service until a fix is implemented and tested.

  142. Andrew says:

    @David

    RE: ‘To me it is the ground crew with their access to detailed manuals and, when required, professional engineers, including recourse to others outside the airline, who have prime responsibility for “technical airworthiness” and the clearing of a defect. That could entail local maintenance test flights with a maintenance test pilot and if needed engineers to observe, to confirm the defect has been cleared, before it is released.’

    Yes, of course that’s what should happen before an aircraft is released, although maintenance test flights are rare unless such a requirement is specified in a manual somewhere. However, an aircraft commander is not obliged to accept an aircraft unless he or she is satisfied with the work that has been completed. The commander is obviously not a maintenance expert, but nevertheless it is incumbent upon him or her to review the aircraft’s previous defects and the associated rectification work. If there are ANY doubts, the commander should consult the engineer that released the aircraft.

    As I said previously, HAVING ACCEPTED THE AIRCRAFT, a prudent commander would review what might happen if the problem recurs, together with any associated procedures. It’s called THREAT MANAGEMENT.

  143. Andrew says:

    @DennisW

    Stop putting words in my mouth. I DID NOT SAY THE AIRCRAFT SHOULD HAVE CONTINUED IN SERVICE. Given it’s history, it should have been taken out of service until the problem was properly investigated, with the help of Boeing if necessary. However, rightly or wrongly, the aircraft was released for service and the captain accepted it for the flight. Having done so, he should have carefully reviewed what might happen if the problem recurred.

  144. David says:

    @Andrew. In this case the corrective action was insufficient and the potential seriousness was evident. IMO the aircraft should not have been released once that condition became known, much less in continuation.

    Either the regulatory structure which allowed that was flawed or was not applied: for example the training/professional education of those involved (which should also be regulated) was insufficient.

    The aircraft designer and certifiers might carry responsibility for the condition but not for the aircraft’s repeated release.

    As you say the pilot might have taken sensible precautions under threat management but to me that is under his responsibilty for safe operation of the aircraft not for corrective action being adequate (‘incumbent’ to ‘review’).

    Those clearing the aircraft would be responsible for seeing there was a maintenance test flight to check out corrective action has been successful if that is uncertain and the consequences could be serious.

    It must be quite clear IMO where such responsibility lies.

  145. David says:

    @Andrew. I was not commenting on your most 11:39, not having seen your most recent.

  146. David says:

    Delete the first ‘not’ please.

  147. Andrew says:

    RE: “As you say the pilot might have taken sensible precautions under threat management but to me that is under his responsibilty for safe operation of the aircraft not for corrective action being adequate (‘incumbent’ to ‘review’).”

    To be clear, the commander does not review the corrective action to decide if the work is ‘adequate’. That’s the engineer’s domain, not the commander’s. The commander reviews the maintenance log to ensure that he or she understands what work has been completed and that the defect has either been rectified or deferred under the provisions of the MEL. In doing so, the commander must obviously rely on advice from the dispatching engineer, but can also seek advice from the airline’s engineering management if he or she is not satisfied.

  148. David says:

    @Andrew. Thanks.

  149. Wall says:

    Hey guys,

    I was wondering: from the moment MH370 went missing,
    how long have the engines kept running? How many hours?

    Wall

  150. Victor Iannello says:

    @Wall: The answer to this question and many more can be found in the Safety Report issued in July 2018.

  151. Perfect Storm says:

    DennisW says: “@DrB: The rate of descent at 00:19:XX strongly suggests an impact near the 7th arc. The lack of an IFE log-in strongly suggests an impact near the 7th arc. There is nothing that supports your point of view. Why do you cling to it?”

    Dennis, you begin to sound more and more like GeRijn who was banned from this forum. Evermore preaching, proselytizing and ad-hominem attacks and less and less sober, factual discussion from you.

    DrB already answered here – with his usual scientific rigor – the exact same questions you repeat over and over again.

  152. Victor Iannello says:

    @DennisW said: The rate of descent at 00:19:XX strongly suggests an impact near the 7th arc.

    I’d say the final BFO values are consistent with an impact near the 7th arc.

    The following remain possibilities:

    1) The APU ran out of fuel before the expected IFE log-on.

    2) The pilot arrested the steep descent, glided for some distance, and then entered into a second steep descent.

    Of these, (2) seems very unlikely, but that’s just my opinion. How do you eliminate it as a possibility without bias?

  153. DennisW says:

    @Perfect Storm

    DrB’s reply had no reference to anything scientific. He merely suggested that one guess is as good as another. My assertion is that my guess is better than his guess.

  154. Richard Godfrey says:

    @Perfect Storm

    Your argument against @Dennis is a touch hypocritical!

    “Argumentum ad Infinitum” is something you could accuse many of us of. The number of commenters sticking, come hell or high water, to their pre-conceived ideas is not a small minority.

    Unfortunately this has sometimes descended into “Argumentum ad hominem”.

    Your examples of @Ge Rijn (100% ad hominem) and @DrB (100% ad infinitum for 31.6S) are poor choices to compare and contrast with @Dennis.

    If you can’t take the heat, then get out of the kitchen.

    … and “ad infinitum”, MH370 is to be found between 25S and 20S!

    😂😳😉

  155. Victor Iannello says:

    @DennisW, @Richard G: I would have a very difficult time recommending a search area that implies the dive-glide-dive scenario to be true, unless other areas closer to the 7th arc come up empty. So perhaps it’s all a matter of priority. I’ve also made the case that if the dive-glide-dive scenario is true, then the possible range of latitudes for the 7th arc crossing is very difficult to define, because there would be little reason to believe the flight was completely automated after 19:41.

  156. Richard Godfrey says:

    @Victor

    I agree with you.

    Our priority is to check the search area between 25S and 20S at +/- 25 NM.

    If that fails to find MH370, then I agree to give up until a new generation of technology is implemented.

  157. DennisW says:

    @Victor

    How do you eliminate it as a possibility without bias?

    You cannot eliminate it as a possibility (remote as it is). You can eliminate it based on the fact that searching wider on the 7th arc is a non-starter. Entertaining DrB’s scenario is the equivalent of giving up and walking away.

    Ask yourself instead – what is the best way to proceed now? What is the best use of search resources deployed for a future search? The answer is obvious to me – you follow the path most closely aligned with what we know. At this moment that is also the most “doable” path.

    I was (as you well know) totally against starting an underwater search before debris was found based on the Inmarsat data. The Inmarsat data is garbage relative to anything other than the direction South and the rapid descent at the end of the flight. I (and knowledgable colleagues) knew that years ago, most people are just accepting that now. DrB, TBill, and Perfect Storm being notable exceptions.

    We have a very manageable area remaining based on the huge amount of work and expense already done and the drift analytics. I am not willing to toss that in the toilet based on alternative remote possibilities. You might even say I feel strongly about it.

  158. Victor Iannello says:

    @DennisW: Would you be so bold as to state what you believe is the probability of finding the MH370 debris field between 25S and 20S latitudes within +/-25 NM from the 7th arc?

  159. DennisW says:

    @Victor

    The very simplest way to look at it is to assume the debris is in the latitude range of 38S to 20S and within +/-25nm of the arc. Using the Metron AF447 criterion, the probability of finding the wreckage is 90%. The probability of missing it being 10%.

    The above does not assign any probability to the debris not being in the area searched. If one assigns a 90% probability (arbitrary at best) to the debris being in the search range, then the probability of finding it in a continued search of 25S to 20S becomes ~80% which is a number I think is very reasonable.

  160. Victor Iannello says:

    @DennisW: OK. 90% that the impact occurred within 25 NM of the 7th arc, and a 90% detection probability. What is your rationale for limiting the probability of the dive-glide-dive scenario to 10%?

  161. DennisW says:

    @Victor,

    I thought 10% was generous based on:

    1> BFO values

    2> No IFE log-in

    3> Unexplainable choice of flight dyamics

    Item 3> is particularly relevant. There is absolutely no reason for the initial dive if a glide was ulimately intended. We have already established that the RAT would do fine with a simple glide from altitude. What could possibly explain a dive-glide-dive choice?

  162. DrB says:

    @Victor Iannello,

    You said: ” I’ve also made the case that if the dive-glide-dive scenario is true, then the possible range of latitudes for the 7th arc crossing is very difficult to define, because there would be little reason to believe the flight was completely automated after 19:41.”

    In my opinion the “little reason” you throw aside would be the pilot flying the aircraft in the normal automated fashion, which was also done from VAMPI onward. I don’t agree that using that method is of low probability from VAMPI until MEFE.

  163. Richard Godfrey says:

    @Victor

    There is no historic precedent for an aircraft crash with a dive-glide-dive scenario.

    Out of 20,300 incidents in the ASN database, only 23 have suicide as a contributory cause.

    We are talking less than 0.1%.

    Even if you add the 328 occurrences of missing aircraft (such as MH370) and the 104 occurrences where the cause is undetermined, we only get 2.3%.

    Even if you add 44 occurrences of loss of situational awareness, 55 occurrences of alcohol/drug abuse, 3 occurrences of mental problems, you only get to a 2.7% probability.

  164. Victor Iannello says:

    @DrB: You are seriously mistaken. Your proposed route has the pilot changing from LNAV to a constant magnetic track, and then remaining in that automated navigational mode with no further pilot inputs for 5 hours until fuel exhaustion. To fly that period of time in that particular autopilot mode is not a “normal automated fashion”. To then believe a pilot would enter into dive-glide-dive maneuvers is proposing an absurd series of events. It’s all possible…but I’d consider this sequence to be very low probability.

    The only automated mode that would be “normal” over those distances and times would to be following one or more waypoints.

  165. Victor Iannello says:

    @DrB: You said “the pilot flying the aircraft in the normal automated fashion, which was also done from VAMPI onward”

    That’s your belief. What you call fact, I call conjecture. There are many reasons to doubt the Lido Hotel data, including the fact that the Malaysians refuse to officially acknowledge their existence, and during the investigation, they told the Australians the data doesn’t exist. Not to mention that it’s not possible to extrapolate the radar data to the satellite data without introducing a lateral offset maneuver in the 3-minute gap between the two data sets. Those are all excellent reasons for questioning the Lido Hotel data.

  166. Victor Iannello says:

    @DennisW: I note that Byron Bailey believed an initial high speed descent was necessary for the RAT to produce the required power. He argued it here on this blog. Now, we know that he is mistaken. However, he is a former 777 captain, and his knowledge base might be representative of other 777 captains, so perhaps the MH370 captain had the same mistaken belief.

  167. Victor Iannello says:

    @Richard G: I’m not sure I understand your point. What scenario do you believe is most likely?

  168. Richard Godfrey says:

    @Victor

    My point is that a dive-glide-dive scenario is an extremely low probability.

    My rationale is, that there is no historical precedent for a dive-glide-dive scenario in any aircraft crash.

    There are plenty of occurrences of a dive.

    There are some occurrences of a glide.

    But never a dive-glide-dive.

    I find such a scenario contrived to try and match certain end points.

    In my view we need to search between 25°S and 20°S up to ±25 NM from the 7th Arc.

  169. DrB says:

    @Richard Godfrey,

    MH370 is unique. You cannot compute its probability.

  170. Richard Godfrey says:

    @DrB

    You stated “MH370 is unique. You cannot compute its probability”.

    That is “Argumentum ad Absurdum”.

    The Aviation Safety Network (ASN) Database contains detailed descriptions of over 20,300 incidents, hijackings and accidents. Each of the 20,300 events are unique. That does not prevent the ASN publishing statistics and computing probabilities.

    Of the 14 pilot or hijacker murder suicide events, all were a powered dive into a target building, own home, target airfield and aircraft, mountains, ocean or until a mid air break up. MH370 was an unpowered dive after fuel exhaustion, into the ocean, in a remote location, at over 15,000 fpm, as shown by the BFO data.

    The dive-glide-dive scenario is a contrivance to get around the conflict between the BFO data and the negative search within ±25 NM of the 7th Arc.

  171. Richard Godfrey says:

    @Victor,

    All oceanographers agree that the most likely location for MH370 debris to beach is on the eastern coast of Madagascar. Out of my 200 recent simulation runs, 66 beach on the eastern coast of Madagascar, which is only 378 NM from Reunion and it takes on average 47.5 days for floating debris to travel from Reunion to Madagascar. The time between Reunion and Madagascar is almost constant, irrespective of the start latitude on the 7th Arc. The arrival time near Reunion, taken from these 66 simulation runs, that eventually beached in Madagascar is shown in the link below:

    https://www.dropbox.com/s/qhxxywj0fld8onh/Arrivals%20near%20Reunion%2012Nov2018.pdf?dl=0

    The start latitude near the 7th Arc closest to the 508 day trajectory of the flaperon is 24.9°S.

    An arrival in the Reunion area 30 days earlier is closest to a start latitude near the 7th Arc at 23.5°S.

    Excluding 7 outliers, the start latitude is highly unlikely to be north of 18.8°S or south of 31.8°S.

  172. Niels says:

    @RichardG
    The arrival time near Reunion you just posted; is that purely based on undrogued drifter data, so roughly excluding possible extra (direct) wind induced drift for the flaperon on top of the Stokes drift?

  173. Richard Godfrey says:

    @Niels

    The arrival time in Reunion is based on undrogued drifter data. Additional windage would mean an earlier arrival. As indicated, 30 days earlier would mean a start latitude on the 7th Arc of 23.5S.

  174. Niels says:

    @RichardG
    Thank you for clarifying. As a rough estimate for the uncertainties: 508 days x 10 cm/s is a lot in distance traveled.

  175. Richard Godfrey says:

    @Niels

    I am talking about the Southern Indian Ocean, not Stormy Bay Tasmania.

    Even 100 days earlier arrival in Reunion, would mean a start latitude on the 7th Arc at 21S.

    I have only claimed, we should search from 25S up to 20S.

    I find it interesting, when I say the Flaperon was in Reunion after 400 days, I am criticised that it arrrived far too early.

    I find it interesting, when I say the Flaperon was in Reunion after 500 days, I am criticised that I have forgotten windage..

  176. Peter Norton says:

    > Richard Godfrey: My point is that a dive-glide-dive scenario is an extremely
    > low probability. My rationale is, that there is no historical precedent for
    > a dive-glide-dive scenario in any aircraft crash.

    There is no historical precedent for MH370. What all theories have in common, is: none of them are based on any precedent. The known part of flight MH370 (until IGARI or 18:22, depending on what you admit as evidence) has no historical precedent either. And yet it occurred.

    Why would you demand that the end of an unprecedented event must have precedent ?

    It’s not valid to argue that the continuation of an unprecedented event would somehow be of “extremely low probability” because it lacks precedent.

    It’s like arguing it would be extremely improbable for a green alien to have yellow eyes, because there is no historic precedent for humans with yellow eyes.

    > DrB: MH370 is unique. You cannot compute its probability.
    >
    > Richard Godfrey: Each of the 20,300 events are unique.

    Few of the 20,300 events are completely unique. Most share a common cause with many other flights (cargo fire, birdstrike, midairs, CFIT, fuel exhaustion, spatial disorientation, etc). MH370, on the other hand, seems to be unique on every level.

    > Richard Godfrey: Of the 14 pilot or hijacker murder suicide events, all were
    > a powered dive into a target building, own home, target airfield and
    > aircraft, mountains, ocean or until a mid air break up.

    None of them lasted for 7½ hours.

  177. Peter Norton says:

    > DennisW: Another factor is the value associated with finding the wreckage. I don’t
    > believe anyone who might write a check believes it was a aircraft failure issue,
    > hence nothing to be learned in that domain. FDR will likely show the plane was flown
    > to location where it was found. CVR will likely be blank. What is the driver for finding
    > the wreckage? […] The answers to this mystery lie in Kuala Lumpur IMO.

    > DennisW: I view not finding wreckage dissappointing, but not a huge loss. I think the
    > CVR will contain no information, and the FDR will show that the plane was flown to
    > where it was found (not an aircraft failure).
    > What I don’t think the wreckage will provide is the answer to why the plane was
    > diverted. Since I am not a fan of the suicide motive, I am sure that answer lies
    > elsewhere and may eventually surface.

    @Victor:

    Is this blog’s mission limited to finding MH370, at which point it would be closed?

    Or, in case of DennisW’s scenario, will you all continue to investigate/discuss MH370 in a bid to solve key questions of MH370’s mystery (for example why it was diverted, as Dennis said) ?

  178. Richard Godfrey says:

    @Peter Norton

    You stated: “The known part of flight MH370 (until IGARI or 18:22, depending on what you admit as evidence) has no historical precedent either. And yet it occurred.”

    The flight route N571 is flown over 100 times each day.

    Hijackings, unfortunately have happened all too often.

    Pilot murder suicide has also happened several times.

    I have not “demanded” anything. YOUR WORD, NOT MINE. Do not put words in my mouth!

    Andreas Lubitz considered his plan for the German Wings for longer than 24 hours and the crash was only on the return flight. Why do you think a flight of over 7 hours proves anything unique?

    So what is your conclusion, following your tirade?

    Are you trying in your strange Argumentum ad Hominem, to prove a dive-glide-dive?

    If so, please deliver your rationale.

  179. Niels says:

    @RichardG
    I think a way forward could be to estimate error margins for the models we are using. This applies to the drift models as well as for the BFO model. At first we will possibly be shocked; next it could help to point out on which aspects of the models we should work to improve accuracies.

    For your model I’m not sure if we should put too much effort regarding the flaperon, as it is a single item with rather peculiar wind sensitivity. However, for the low windage items arriving at African / Madagascar shores, I think it is for example interesting to find out what possible errors are introduced by basing the drift model on historic data, as well as by the numerical methods used.

    Regarding the BFO model: I have been in brief contact with Ian Holland and he still refers to this “geographic dependency” for the “structured bias”. To me it indicates there is room for improvement in the BFO model that was used by DSTG.

  180. Victor Iannello says:

    @Peter Norton: The blog will remain open as long as I perceive it is adding value and I have the willingness to continue. It may continue after the aircraft is found. Or it may end before that time. We’ll see.

  181. Peter Norton says:

    @Richard Godfrey: I don’t think you have understood my comment. I don’t know how else to explain it. Maybe rereading it without feeling attacked personally might help, because

    > Argumentum ad Hominem

    there were none

    > I have not demanded anything.

    Sure. First 3 lines of the posting you were referring to.

    > Andreas Lubitz considered his plan for longer than 24 hours

    He considered it probably for weeks, months or maybe even years.
    The execution took only minutes. Not 7+ hours.

  182. Peter Norton says:

    @Victor Iannello:
    I didn’t want to inquire about this blog’s lifetime. I understand that this is of course hard to predict and a personal decision. I rather wanted to ask about its mission. Is it to help locate the wreckage or beyond that to solve the mystery?

    When DennisW wrote that the wreckage might not tell us enough, it occurred to me that I don’t know the answer to this question. I mean what would happen if, say, tomorrow the wreckage were found? Would the mission be considered accomplished, regardless of what is found or would it depend on what is found?

  183. Peter Norton says:

    (By “wreckage” I meant the crash site.)

  184. Richard Godfrey says:

    @Peter Norten

    You are continuing your Argument ad hominem.

    Now you accuse me of not understanding you (last resort argument). I understood you perfectly.

    However, you did not understand me, I asked if you would supply your rationale for a dive-glide-dive scenario.

    You failed again to provide your rationale, preferring to continue your personal attack!

    No wonder. you want Victor to close his blog.

    I rest my case.

  185. Peter Norton says:

    > Peter Norton: “The known part of flight MH370 (until IGARI or 18:22, depending on
    > what you admit as evidence) has no historical precedent either. And yet it occurred.”
    >
    > Richard Godfrey: The flight route N571 is flown over 100 times each day.
    > Hijackings, unfortunately have happened all too often.

    how many of those have “gone dark” (ACARS, XPDR, etc.) ?
    Do you really want to contradict that the circumstances of MH370’s disappearance have no historical precedent ?

  186. Peter Norton says:

    @Richard Godfrey: I would like to talk MH370, not engage in personal arguments please. I have not “accused” you of anything. You reaction gave me the impression that you didn’t understand my comment and I told you that. That’s all.

    > No wonder. you want Victor to close his blog.

    Come on, I have never said that in the least. I’m glad it exists.

  187. Victor Iannello says:

    @Peter Norton: If you look at the top of the page, it says “Helping to Solve the World’s Greatest Aviation Mystery”. If the mystery is solved, or we are not contributing towards solving the mystery, I’ll end it. This blog is not to promote a book, an article in a journal, or a media appearance. On the other hand, I suspect that finding the aircraft will not completely solve the mystery, and will in fact create many new avenues to explore that may be of interest to me and other readers here.

  188. Peter Norton says:

    @Victor Iannello: Thanks. I share your point of view and hope you can keep it open as long as possible.

  189. Hank says:

    @Peter Norton

    I have different theory that I raise from time to time but isn’t generally considered much.

    I assume that Z is planning a suicide mission and does not want the aircraft to be easily found. He plans to put the aircraft in the SIO. He plans to burn all of the fuel to minimize a fuel slick after impact. He plans to decouple any post flight simulation by Boeing of range at fuel exhaustion, by periodically performing a wide 360 clearing several times during the flight. He needs reasonable distance, but not maximum distance. The ATSB zones were biased toward long straight flights and intentional holding as not allowed – so ARC 7 most likely position was centered at 38S. With intentional clearing turns the ARC 7 crossing could easily be above 25S.

    Z also wants to put the plane underwater with a small debris pattern. I think that this might be best done with a vertical entry at relatively low speed. A dive from cruise altitude, leveling off just above the surface, bleed some speed off, then pull up into a loop with a near vertical entry. This takes daylight to execute. Lots of simulation by Z would be needed to be able to perform the maneuver. There may be other terminal maneuvers that could result in a clean, less than stall speed, vertical entry.

    A controlled vertical entry from inverted flight at 100 knots may be a good way to get the bulk of the aircraft wings and fuselage underwater without much surface breakup. If Z planned this including the holding on the way south he has been very successful.

  190. airlandseaman says:

    Hank: You have a vivid imagination.

    “I think that this [minimum debris] might be best done with a vertical entry at relatively low speed”? Good luck with that maneuver.

Leave a Reply