The Unredacted Inmarsat Satellite Data for MH370

Inmarsat’s Mark Dickinson holding the satellite data in an interview with CNN

Since we first learned of its existence, we’ve been asking for the complete record of the communications data between MH370 and Inmarsat’s satellite network. In May 2014, Malaysia released satellite data logs, but they were incomplete: fields of data were missing, and only a small number of data records from before the flight was made available. When pressed for the complete logs, Inmarsat and Malaysia both claimed the data had to be released by the other.

We now have what we believe is the complete record of communications between airframe 9M-MRO and the Inmarsat satellite network, from March 7, 2014, at 00:51 UTC, until March 8, 2014, at 01:16 UTC. This time period includes the previous flight from Beijing to Kuala Lumpur.

The satellite data was shared with me by a relative of a Chinese passenger on MH370. The data was given to him by Malaysia Airlines with the following email text:

Please find attached the Inmarsat data, for your info. Please note that these are raw data as you have requested. The authorities agree to release the data, on condition that:

  1. We will not translate the data into any meaningful information as the data is proprietary to Inmarsat. The Malaysian Investigation team does not have any experts to translate these data into any meaningful information.
  2. We will not translate the data into any other language, including Mandarin.
  3. These data are complete and obtained from Inmarsat. Please do not manipulate the data.

I know, by having these data, you will have more questions, but I have to say that we are providing these data to satisfy your request, but we cannot answer any questions on the data because we too, cannot understand it. Only the experts from Inmarsat can.

 Hope you understand.

 Thank you

I suspect the data will confirm some assumptions, and will raise even more questions. I hope the data can help us learn more about the disappearance.

298 Responses to “The Unredacted Inmarsat Satellite Data for MH370”

  1. buyerninety says:

    Although not relevant to the above current topic, this article discussing the experiences
    of some of the Chinese relatives of MH370 passengers is probably worth being more widely
    known by VictorIs readers;
    http://www.marieclaire.com/culture/news/a13397/malaysia-airlines-lost-flight-anniversary/

  2. ALSM says:

    My comments on the new, complete Inmarsat log are here: https://goo.gl/Th7QH1

    ALSM

  3. Ge Rijn says:

    @ROB

    On your previous comment about the cellphone in Flight-mode.
    Flight-mode blocks all possible transmissions, so I don’t think this could cause a log-in somewhere. Rather a log-off but no log-in.
    At least the log-in on WeChat indicates Z’s cellphone was not switched off while it should have been one minute before take-off.
    And it probably further indicates the FO was the PIC during take-off and climbing to altitude till the first FL350 message from Z. to ATC.
    After this I assume the FO was relieved by Z. and Z. took over control.
    The second FL350 message was done by Z. 8 seconds after the final ACARS data transmission.

    I would be curious if there was also a log-off registered by WeChat indicating the phone was switched of or put in to Flight-mode and if not this could mean the connection with WeChat prolonged during take-off and the early stages of the flight.

  4. Ge Rijn says:

    @ROB @others

    Speculating about Z’s WeChat connection, using a Chinese server and encrypted messages, could it have been used to inform a Chinese contact about the faith of the plane under the name of ‘CMB’?

    How could this ‘CMB’ have known at 3:30 in the early morning of 9 march 2014 there were 153 Chinese passengers on the plane and state that the plane was never going to be found if a contact was not informed by someone who had this information at that time?

  5. Nederland says:

    @Ge Rijn

    I think it was the other way round. Zaharie was PIC up to the point when (shortly before) the WeChat login occurred. After that, Fariq was PIC (for an unknown period of time). The login happens to fall in the time period when the change in PIC also seem to have occurred. This is because of the change of voice associated with the speaker who communicated to ground.

  6. DennisW says:

    @ALSM

    Nice. Thanks.

  7. buyerninety says:

    @ALSM
    I’m not at an .xlsx enabled application device ATM.
    I seem to remember that MH370 communicated to the satellite using the low-gain antenna,
    then via the high-gain antenna – also, communicated initially with the POR, then switched
    to the IOR.
    If this is correct, could the ~1600 data point anomalies be the result of these occurences?

  8. ROB says:

    @GeRijn

    Possibly I got confused. I was reading the earlier discussion iro whether or not a cell phone switch off causes a WeChat logon. As Mike Gilbert has pointed out, what exactly does FI mean when it states “Zaharie last WeChat login 40:02? Does it actually mean 40:02 was the last time Zaharie was logged on to WeChat? ie, he logged off (or his cell phone was powered off at that time? It’s yet to be clarified, think.

  9. Niels says:

    @ALSM

    In the SUlog I noticed for both (attempted) phone calls towards the end:
    Significant BER in combination with reducing C/No as well as reducing Rx power level. (No significant change in BFO)
    Would you know how to explain this?

  10. Richard says:

    Here is a link to a complete decoding of all Initial and Subsequent Signalling Units from Hex to Ascii.

    The first octet of each Subsequent Signalling Unit gives the message order, e.g. C8, C7, C6, C5, … etc., which is not always followed.

    In each row of 12 octets, the first 2 octets are header and the last 2 octets are check sum.

    I have therefore only translated the middle 8 octets in each row.

    https://www.dropbox.com/s/8ngsielzp0fegct/ISUs%20and%20SSUs.pdf?dl=0

  11. Victor Iannello says:

    @Richard and @ASLM: Thank you for jumping on this so quickly and providing us with information.

  12. Joseph Coleman says:

    @Richard

    In your translation why is there the words “Pet Shop Boys”?

  13. Joseph Coleman says:

    @Richard

    This might be a silly explanation, but could it part of an Advert broadcast notice, because they were perhaps going to be live on tour in Malaysia later that year?

  14. Don Thompson says:

    @Joseph,

    The transmission containing the “Pet shop boys” string is that noted in the FI as pertaining to the IFE BITE channel. It’s an air-to-ground message.

  15. edmg says:

    Yeah, it’s just something the IFE sent to the ground when it started up.

  16. Don Thompson says:

    ATSB did make this comment in the 2014 (Jun , revised thru Aug) Definition of Underwater Search Areas report:

    The 1825 and 0019 SATCOM handshakes were log-on requests initiated by the aircraft. A log-on request in the middle of a flight is not common and can occur for only a few reasons.

    Must not have looked at the prior flight, MH371, ZBAA to WMKK. The AES executed 22 Log Ons between OFF at ZBAA and ON at WMKK.

  17. ALSM says:

    buyerninety:

    Re: “I seem to remember that MH370 communicated to the satellite using the low-gain antenna, then via the high-gain antenna – also, communicated initially with the POR, then switched to the IOR.”

    9M-MRO used the LGA and POR sat at 15:55:58.765 to log off. From 15:59:56.414 on MH370 used the HGA and IOR sat exclusively.

    Re: “If this is correct, could the ~1600 data point anomalies be the result of these occurences?”

    I do not understand what you are referring to. There were no “…1600 data point anomalies …” that I am aware of.

  18. ALSM says:

    Niels:

    Re: “In the SUlog I noticed for both (attempted) phone calls towards the end:
    Significant BER in combination with reducing C/No as well as reducing Rx power level. (No significant change in BFO) Would you know how to explain this?”

    The non-zero BER threshold was C/N0 = ~52 dB. For a data rate of 21000 b/s, that is about a 9db Eb/N0. I’m not up to speed on the FEC and other coding used on the C channels, but 9 dB is about right. (Don will be able to provide details I believe.) The system starts with a high AES EIRP (higher is needed for the 21,000 b/s channel speed). Then the GES sends commands iteratively to gradually reduce the AES uplink EIRP in order to minimize the required s/c downlink EIRP. The steady reduction to the min requirement and then slight increase is apparent in the following graph: https://goo.gl/QZ7lnV

  19. Joseph Coleman says:

    @edmg
    @Don Thompson

    I understand guys thanks, a least over the years there wasn’t any conspiracy Nut, trying to link this to their song “Go West”.

  20. edmg says:

    @Joseph Coleman

    There’s a story about the Pet Shop Boys conspiracy theory, but it will have to wait until the definitive history of MH370 is written. Mostly because I don’t remember whose theory it was.

    It was a good one, though, at the time.

    And it was “A New Life”, not “Go West”.

  21. ALSM says:

    Niels:

    I should have also noted that a BER of ~200 in 10^-6 (better than 10^3) is excellent for a typical coded voice channel.

    Mike

  22. Brock McEwen says:

    @Victor: thanks for forwarding. Have Inmarsat publicly vouched for the authenticity of this data?

  23. sk999 says:

    I downloaded the .xlsx file, and it seemed to retain the original metadata. The original timestamp on the file is 03:31 UT on Mar 8, 2014, barely 2 hours after the final transmission. According to this article:

    https://corpcommsmagazine.co.uk/features-and-analysis/view/what-lessons-can-communicators-learn-from-malaysia-airlines-flight-mh370

    from an interview with Chris McLaughlin, originally Inmarsat was asked for the data by SITA. It would appear that this spreadsheet is the file that was prepared at the time and sent off. According to McLaughlin, Inmarsat did not itself examine the data further until the afternoon of the next day (Sunday). Presumably it was Alan Schuster Bruce who did so, although he is not mentioned by name in the article.

    It should be possible to fully model the BTO and BFO (which will require modeling the anti-Perth contribution more accurately, but I don’t the eclipse effect comes in – right?) Will take a bit of time to set it up.

  24. Victor Iannello says:

    @Brock McEwen: I have not formally asked Inmarsat, the ATSB or Malaysia whether the data is authentic. Considering the level of detail, the quantity of data, and the consistency with everything released to this point, I would be very surprised if the data was not authentic. That said, I would welcome any efforts on your part to obtain a formal statement from the authorities about the authenticity of the data.

  25. Victor Iannello says:

    @sk999: I agree that the creation time/date for the file is very interesting.

    That’s no small task to get the BTO and BFO models right for POR in addition to IOR. I appreciate your willingness to try.

  26. Don Thompson says:

    @sk999

    I’d expect the POR pilot AFC reference to originate at Paumalu teleportin Hawaii (if that helps).

  27. sk999 says:

    OK, I was just thinking of modeling the IOR data. Are there enough transmissions through that satellite to make it worthwhile? It looks the the SATCOM bounced between the two satellites during the flight. I don’t have TLEs for 3-F3 on Mar 7, and my space-track account expired ages ago, so I’d have to start over there. The location of the pilot signal transmitter is obviously important, assuming that the anti-Perth effect applies to it as well. The biggest amount of work will be validating the anti-Perth model (which I haven’t looked at in 3 years), pulling out yet more weather data, and putting together a realistic flight path. If the ACARS positions reports encoded in the data packet can be turned into readable tables, that would be a big help.

  28. Peter Norton says:

    > Mick Gilbert says:
    > Regarding deployment of the portable unit and antenna extension, it
    > would appear that there are a variety of mounting and securing set-ups
    > […]
    > In both instances, however, note the plastic cover.
    > […] from the MAS Safety and Emergency Services Manual it would appear
    > that their portable ELTs were fitted with a cover. If it is the same
    > as that shown in those photos then it encompasses the folded whip
    > antenna.
    > […] there also appears to be a lanyard connecting the beacon to the mounting bracket.
    > Regardless of the configuration of the portable ELT on 9M-MRO one
    > thing that is clear is that it is not designed to self-deploy.

    Agreed. Great research!

    The other type of ELT I referenced in my previous posting (“[…] designed to float on the surface of the water. The water usually dissolves a paper/cardboard collar holding a folded antenna in place. When the collar releases, the antenna springs up and the transmitter starts operating”) seems more useful to me than MH370’s ELT type, as it would more likely activate upon ditching.

    Generally speaking, it’s frustrating for me that ELTs are so unreliable.
    The FI report categorizes ELT failures in 173 accidents over the past 30 years as follows:

    Operated effectively: 39 (22,5%)

    Submerged: 1 (0,6%)
    Terrain shielding: 1 (0,6%)
    Internal failure: 5 (2,9%)
    Damaged: 11 (6,4%)
    Battery failed: 14 (8,1%)
    Other: 21 (12,1%)
    Not activated: 22 (12,7%)
    Unknown: 59 (34,1%)

    Given a less than 1 in 4 chance of successful ELT transmission upon crash, MH370’s lacking ELT signal unfortunately doesn’t seem to allow to draw any deduction. (If anything, the data suggests that no ELT signal would actually be the normal case and an ELT signal would be the exception.)

    > Mick Gilbert says:
    > while condensation would not be an issue for an internally mounted
    > ELT, incorporating water activation would mean ELTA would be designing
    > and manufacturing two quite different ELTs for fixed and portable
    > applications. […] That one unit/two applications approach has a number
    > of advantages for both suppliers and users.

    IMO usefulness / fitness for purpose trumps commercial considerations (such as “it would be easier to design/manufacture”).

    (following your argument, the same ELT would also be used for the life rafts)

  29. Brock McEwen says:

    @Victor: okay; anything to help. What is the full name and title of the MAS rep who sent the email? I’d like to cite the most upstream source you have when communicating with Inmarsat.

  30. ALSM says:

    Brock:

    I confirmed through an official channel that the file we have is identical to what ATSB received in March 2014.

  31. Paul Smithson says:

    @sk999 – on the subject of weather, do you have access to historic ECMWF wind data?

  32. Marion Ravenwood says:

    Indy says congratulations on acquiring the back side of the headpiece to the staff of Ra. Now let’s find the Ark itself.

  33. Victor Iannello says:

    @Brock McEwen: The name of the MAS official was stripped from the email forwarded to me by the Chinese NOK.

    Based on the info that Mike supplied, I’m not sure it is worth your effort. But that’s your call.

  34. sk999 says:

    Paul Smithson,

    R.E. ECMWF, not I do not.

  35. Tamas Feher says:

    > edmg says (June 12, 2017 at 6:40 pm)

    Huh, did anyone seriously consider a Pet Shop Boys conspiracy theory? Imagine if it had been Depeche Mode with “Enjoy the silence”…

  36. Victor Iannello says:

    @Don Thompson said, The AES executed 22 Log Ons between OFF at ZBAA and ON at WMKK. to demonstrate that ATSB’s statement that The 1825 and 0019 SATCOM handshakes were log-on requests initiated by the aircraft. A log-on request in the middle of a flight is not common and can occur for only a few reasons. seems contradictory.

    I plotted the BFO for MH371 to show how many times the AES switched between IOR and POR. The variation in BFO is smallest in the middle of the flight, where turns and altitude changes are less likely. Also, under normal conditions, there appears to be no abnormal BFOs caused by an inflight log-on.

  37. Victor Iannello says:

    [I have removed a previous comment of mine until I can verify the timestamp information from the satellite data file.]

  38. edmg says:

    I suspect Inmarsat just created logs of all the data related to that aircraft, and didn’t actually look at them in any detail until the next day. There was already a search going on in the China Sea, so they had no reason to do so until someone decided to see if they could do anything to help locate it for the searchers.

    In normal circumstances, you’d have expected it to be found by Sunday afternoon, by spotting wreckage or oil slicks. Only when it wasn’t found was there any reason for someone to take a close look at the logs.

  39. DennisW says:

    @Victor

    I anxiously await the plot of BFO error over time for MH371. Similar to figure 5.4 in “Bayesian Methods…”. I am too lazy to do it.

  40. Victor Iannello says:

    @DennisW: With the scarce data we have, it’s not trivial to do, and other things I’m working on have higher priority.

  41. Niels says:

    @ALSM
    Mike, thanks for explaining the reduction of Rx power level and C/No during phone calls. While I don’t get all the details yet I understand the system deliberately reduces/optimizes AES uplink radiated power.

  42. Niels says:

    @VictorI
    The timestamp info for the Inmarsat data file is indeed important. I hope you can somehow verify it.

  43. Victor Iannello says:

    @Niels: In a previous comment, Mike said, “I confirmed through an official channel that the file we have is identical to what ATSB received in March 2014.” That would imply that the creation date is correct.

  44. TBill says:

    @edmg
    In one of the major MH370 TV documentaries, Inmarsat talks about their timing but if I recall correctly by late Sunday they were trying to get word thru their contacts to MY that the flight lasted hours longer than thought.

    @all
    I have some gut reaction to the comment to the NoK ” but we cannot answer any questions on the data because we too, cannot understand it. Only the experts from Inmarsat can”

    …but is has been over 3 years since the accident, and Inmarsat has not displayed any reluctance to answer such questions to be of help. It implies
    that many obvious avenues of investigation have not been done. Sorry to feel that way.

  45. ALSM says:

    Re: File Creation date

    I have confirmed via an official channel that the file creation date was 8 March 2014 03:21 UTC. Confusion about the date and time may arise because of the way Microsoft records and displays this date/time in Excel Files. The files contain the creation time and date in UTC. However, the date displayed on a users computer (in Properties) is displayed in then current local time. Thus, depending on where you are when you open the file, you may see a different date and time in Properties.

    On my computer in Colorado (currently UTC-6), the properties for the file Victor shared shows 7 March 2014 9:21PM. This is the same as 8 March 2014 03:21 UTC. This is about 3 hours after MH370 crashed, and 10 hours after it went missing.

  46. DennisW says:

    @Victor

    “it’s not trivial to do, and other things I’m working on have higher priority.”

    I used to have people for that, but no mas.

  47. Brock McEwen says:

    @Victor: that’s too bad. But no sweat: what was the name of the relative of who supplied you with the redacted email? I’ll ask that person for the MAS rep’s name.

    Now that it is being widely reported, Inmarsat should go on camera and vouch for it. As a fellow pro bono researcher into the bizarre disappearance of MH370, I’m sure you agree we’ve all had enough dodgy “leaks” to last us a lifetime.

    Thanks in advance.
    Brock McEwen

  48. HB says:

    Does anyone know
    (1) the exact list of system’s condition to generate the message “unresponsive AES” (eg total loss of signal, invalid/corrupted AES ID, AES ID valid but no data … etc.)
    (2) how is the normal log off sequence and why this message is not generated for the other log off events?
    (3) there is an apparent conflict with the Factual Information stating that the last ACARS message, basically the fax from MAS ODC should be repeated every 2 min until 18:43:33 UTC p47 if no response from the AES. Why was 18:43:33 UTC mentioned – i cannot correlate this to the log? Any idea about that?
    (4) there is a statement also in the factual information p43 saying that MAS should contact the plane via SATCOM if ACARS is silent for 30min. This apparently was not done at 17:37 ish: was this clarified with MAS as part of the investigation?
    (5) is the there another log of satcom interface activities (print outs etc) on MAS side?
    (6) does the satcom support multi tasks for instance can a fac be sent during an ACARS engine condition monitoring download which takes several seconds?

  49. DrB says:

    @all,

    Here is the 12:50 POR Log-on BFO data plotted on top of the other log-ons:

    https://twitter.com/DrBobbyUlich/status/874836356324872193

    It is a perfect match to the OCXO transient curve.

  50. ALSM says:

    Brock:

    Why are you pursuing the source at MAS? Hasn’t it occurred to you that this source is better left unknown to the public? We already know it came from Inmarsat via MAS via NOK via VI and we know the file is authentic (confirmed same as ATSB’s file copy). Let’s encourage more cooperation between MAS and NOK. Exposing sources does not help anyone. Let’s get on with the analysis of the new data.

  51. Warren Platts says:

    In addition to “pet shop boys” there are other seemingly strange messages, e.g., 16:06:32.405 there is a reference to some sort of special load:

    NOTOC MESSAGE SPECIAL LOAD NOTOC FLIGHT DATE EDNO MH0370 /08 08MAR14 FROM/TO AC/REG KULPEK

    This shortly followed by this seemingly bizarre message at 16:06:38.905 regarding lack of evidence of something:

    THERE IS NO EVIDENCE THAT ANY DAM = 9M-MRO C1U QTY/TI IMP DESCRIPTION PEK —

    followed by a description of over 4 tonnes of mangosteens (the exotic tropical fruit).

    Then at 16:06:44.405 this seemingly ominous message caught my eye:

    AGED OR LEAKING PACKAGES CONTAINING DANGEROUS GOODS HAVE BEEN LOADED ON THE AIRCRAFT THIS STATION. END ACARS NOTOC [NOtice TO Captain] !

    Aged or leaking packages containing dangerous goods were loaded! This wouldn’t have anything to do with the disappearance, would it? Or so it was my first thought.

    However, probably the two messages should be combined:

    “THERE IS NO EVIDENCE THAT ANY DAMAGED OR LEAKING PACKAGES CONTAINING DANGEROUS GOODS HAVE BEEN LOADED ON THE AIRCRAFT THIS STATION”

    I guess this is just a standard message routinely put at the end of special load manifests. So nothing here. Just putting this out in case anyone else gets confused.

  52. Paul Smithson says:

    @ALSM, @DrB

    That is indeed a striking similarity in BFO curve shape/amplitude between the cold start POR logon 12:50 and the 18:25 logon.

    It suggests that the “overshoot” value at 1825 is peak or close to peak of the curve. This pattern also pertains for the transient curves that Dr B has plotted from the Holland paper.

    However, Dr B’s curve fit for 1825 logon places the “high” value half way up the spike toward the overshoot peak, rather than at the peak.

    What do we think is the “proper” curve fit to the 1825 pattern? Can it tell us, what was the “real” underlying BFO value at 1825-1828 if the transient drift is excluded?

  53. Andrew says:

    @Warren Platts

    RE: “I guess this is just a standard message routinely put at the end of special load manifests.”

    That’s correct. The phrase ““THERE IS NO EVIDENCE THAT ANY DAMAGED OR LEAKING PACKAGES CONTAINING DANGEROUS GOODS HAVE BEEN LOADED ON THE AIRCRAFT THIS STATION” is a standard declaration on every NOTOC (Notification to Captain). The NOTOC is a form given to the Captain before departure, notifying him/her of any dangerous goods loaded on the aircraft.

  54. ROB says:

    @Warren Platts

    Hi Warren. So, probably no dangerous good on board. 😉

  55. DennisW says:

    @DrB

    The 18:25:27 BFO error was zero.

  56. HB says:

    Actually not exactly correct. DG goods are controlled such that if something happens the consequence remains within the packaging so DGs do not present a hazard. That is why it is required to inspect the packaging and goods with defective packaging are not allowed on board. This follows IATA DG Regulations. For instance if batteries in significant quantities contain acid (corrosive Cat 6) if any leakage is observed or packaging damage is observed, they will not be transported. Nothing abnormal here.

  57. Niels says:

    So there are strong indications that this complete record of communications was shared with SITA early on March 8th. For me this raises several questions:
    – for whom did they ask for this data / with which parties did they share it?
    – what happened then between March 8th and say March 12th with the data through this SITA “channel”? I mean, you don’t have to be Einstein to see from the file that the AES was active for many hours after the aircraft had disappeared from radar. SITA and their “customers” should have been aware of this early on.

  58. Victor Iannello says:

    @Niels: I’m not sure that the complete record was shared with SITA on March 8th. The timeline seems to be something like this:

    1. Plane disappears at 17:21 UTC on March 7.
    2. SITA asks Inmarsat for data after disappearance.
    3. Inmarsat supplies SITA with initial 1.5 hours of data.
    4. Inmarsat creates complete data log at 3:21 UTC on Saturday, March 8, possibly from a facility at Perth.
    5. Inmarsat (London) begins to study complete logs in afternoon of Sunday, March 9, and sees 7 hours of activity after disappearance.
    6. Inmarsat confirms 7 hours of activity on Monday, March 10.
    7. Inmarsat notifies SITA of 7 hours of activity on Wednesday, March 12.

    If McLaughlin’s narrative is correct, it appears that although the complete data logs were created just hours after the final satellite communications, SITA was unaware of the complete logs until several days later. What was initially shared with SITA were not the full communication logs.

  59. Niels says:

    @VictorI
    Assuming the timestamp is as I saw it (and as Mike confirmed and explained, so 3:21 UTC March 8th not March 9th) I find this two-stage release of data to SITA unlikely. In other words I doubt your points 3 and 4 are correct.

  60. Victor Iannello says:

    @Niels: Yes, I typed the wrong timestamp date. Yours is correct. I fixed it in the comment to avoid confusion.

    The timeline I proposed was based on McLaughlin’s comments. I have no way to know if the comments were accurate. I will say that when I submitted questions to him by email, I received a terse response indicating he had no intention to respond to independent inquiries. Whatever information Inmarsat shared with me was not from him.

  61. Mick Gilbert says:

    @DrB

    A fair bit to a lot of the discussion on the technical aspects of OXCO BFO error is beyond my grasp but would I be correct in inferring that the OXCO was likely to have been colder at the 1825 log-on than it was at the 1250 log-on? If that is the case, given the location of the SDU in passenger cabin, can we reasonably infer that the passenger cabin was colder at 1825 than it was at 1250?

  62. Brock McEwen says:

    @ALSM: I am simply trying to get everything above board. I am sick and tired of privately leaked evidence, when officially disclosed evidence should be the standard.

    We need full names and titles of Inmarsat executives under-signing this data, or else it is not yet worthy of our trust. This is not a controversial point. At least, it isn’t in my industry.

    What is the name of the official to whom you spoke, claiming a perfect match with what the ATSB received? I am happy to follow up directly with that person.

  63. ALSM says:

    Brock:

    The file is authentic. I verified it with an official that has a copy of the same file and we compared properties. They match. That should be sufficient. Focus on the new data.

  64. Brock McEwen says:

    @ALSM: what is the name of that official? Thanks in advance for your transparency.

  65. ALSM says:

    Mick and Bobby:

    We do not know what the ambient temperature was in the cabin at 18:25. It could have been lower than normal (which is typically +25C w/ AC on). But there is no data to tell one way or the other. It cannot be reliably estimated from the shape or timing of the BFO transient alone.

    That said, I agree with Bobby that it would take longer for the oven to reach the set point temp (~+75C) if the ambient at 18:25 was very low, say -25C. That would nearly double the time to reach the set point temp. In such a case, the last SUD BITE check to Pass probably would be the oven temp, in which case it is more likely for the first transmission to occur while the transient is still “on the way up”. If the oven reaches the set point before other BITE tests pass, it is more likely that the first transmission will occur after the peak, on the way down.

    But the timing of the first transmission following power on is a function of several software and hardware factors. It should be noted that the time difference between the first two BFOs (142 and 273 Hz) circa 18:25 was only 7 seconds. If the first transmission had occurred 10 seconds later, the 18:25 BFO transient would not have had a first value that happened to be close to steady state on the way up to the peak. It would have been near the overshoot peak. There are at least a handful of scenarios besides a super cold cabin that could explain this timing.

    In sum, Bobby may be right. I have often speculated about the cabin going cold. If the left bus was off, and maybe the right bus too, it could have been cold. Decompression is another possibility. But we don’t know. Perhaps Thales would be willing to consider running a test in the environmental chamber to see if transient usually does look like 18:25 when the AES is very cold at power on.

  66. Mick Gilbert says:

    @ALSM

    Mike, so you are saying that there are a couple of additional variables that mean we cannot reasonably infer that the passenger cabin was colder at 1825 than it was at 1250?

  67. DrB says:

    @Mick Gilbert,
    @ALSM,

    There are three observed facts that strongly argue that the initial OCXO temperature at SDU power-up leading to the 18:25 log-on was the coldest of all the log-ons studied:

    (1) The initial transmission at 18:25 (the log-on request) occurred at the earliest time in the BFO transient curve, and this is the only log-on event in which the first BFO error is near zero (instead of being the largest error observed, which happens in all other cases).

    (2) The peak amplitude of the best-fit transient curve is the largest of all the log-on examples. Generally speaking, colder start-up temperature results in larger observed overshoot.

    (3) The 18:25 event is the only one occurring in flight, and after a likely 1-hour power-off period.

    ALSM believes the SDU OCXO temperature was the same sitting on the tarmac at 12:50 as it was in flight at 18:25. I don’t. I believe the -45C outside air temperature cools the aircraft skin considerably below the cabin temperature. I believe the SDU is located between the cabin ceiling and the upper skin of the aircraft. I have no idea how it is insulated, nor how it is heated or cooled, but it seems quite possible to me that an unpowered SDU would get passively cooled below the cabin temperature because of the proximity of the cold aircraft skin. Perhaps the loss of AC power to many equipment items (maybe the entire left AC bus?) affected the external heating/cooling of the SDU when it was unpowered. I would also ask if heat is supplied from an external source to the SDU when the unit is cold and unpowered. Maybe it is not. I invite anyone with knowledge of external heating/cooling of the SDU to address my questions above.

  68. DennisW says:

    @DrB

    The BFO value at 18:25:27 was virtually perfect. How do you explain that?

  69. DrB says:

    @ALSM,

    I did not see your response to Mick until after I posted my response.

    You said:” Perhaps Thales would be willing to consider running a test in the environmental chamber to see if transient usually does look like 18:25 when the AES is very cold at power on.”

    I have already suggested to Ian Holland that he do exactly that. I have no word back yet on whether that might happen.

    I was not considering a cold passenger cabin at 18:25. That might also affect the SDU temperature then. I think there was discussion on this point in the past. I seem to recall the cooling rate depended on the decompression rate.

  70. DrB says:

    @DennisW,

    You said:”The BFO value at 18:25:27 was virtually perfect. How do you explain that?”

    Your memory must be failing you since I have previously answered that question from you about five times (and the last time I promised never to do it again). Please consider how you can add signal instead of noise to this discussion. If you can’t follow the answer, then ask a question about the part you don’t understand. You keep doing the same thing but expect a different result. Here’s some good advice for participating in group activities: Lead, follow, or get out of the way. Please pick one.

  71. Mick Gilbert says:

    @DrB
    @ALSM

    Thank you for your answers.

  72. Andrew says:

    @DrB

    RE: “I believe the SDU is located between the cabin ceiling and the upper skin of the aircraft. I have no idea how it is insulated, nor how it is heated or cooled, but it seems quite possible to me that an unpowered SDU would get passively cooled below the cabin temperature because of the proximity of the cold aircraft skin. Perhaps the loss of AC power to many equipment items (maybe the entire left AC bus?) affected the external heating/cooling of the SDU when it was unpowered. I would also ask if heat is supplied from an external source to the SDU when the unit is cold and unpowered. Maybe it is not. I invite anyone with knowledge of external heating/cooling of the SDU to address my questions above.”

    The SDU and various other SATCOM components are located on the E11 equipment rack, immediately above the cabin ceiling near the door 3 area. The fuselage is insulated by fibreglass/mylar insulation blankets mounted next to the fuselage skin, as shown in the following photo:
    B777 insulation blankets

    The SATCOM equipment is cooled by the aft equipment cooling system, which uses the lavatory/galley ventilation fans to draw cabin air through the equipment. The exhaust air then discharged overboard through the aft outflow valve. A backup SATCOM cooling fan operates if the lavatory/galley ventilation fans fail. There is no other cooling or heating.

    If the SDU was unpowered, then I imagine it would stabilise at a temperature similar to that of the cabin. The cabin itself would have to be extremely cold to significantly lower the temperature of the SDU.

  73. ALSM says:

    Bobby, Andrew, Mick:

    Thanks for the insulation Photo Andrew.

    Bobby, we are generally in agreement about the cold start BFO transient. However, one thing you suggested above is not accurate.You wrote: “Generally speaking, colder start-up temperature results in larger observed overshoot.”

    This is only true if the start up temp is within a few degrees of the setpoint. The overshoot will be the same for any initial difference greater than 5C or so. If the initial difference is greater than that, the loop will start in a first order mode which means the heater will be on at maximum, and stay on at max until close to the setpoint. Thus, the rate of change and overshoot will be about the same for all initial conditions except when the power comes on while the oven is already very close to the setpoint, in which case the loop is already switched to second order, and the heater authority is dialed back from max. Been there. Done that.

  74. DennisW says:

    @DrB

    “Your memory must be failing you since I have previously answered that question from you about five times (and the last time I promised never to do it again).”

    I do remember your promise, but I don’t remember your answer to the question.

  75. DennisW says:

    @DrB

    “Here’s some good advice for participating in group activities: Lead, follow, or get out of the way. Please pick one.”

    Good advice according to DrB.

    I find it strange that in Dr. Holland’s paper “The Use of Burst Frequency Offsets in the Search for MH370” that the term “overshoot” does not appear even once. Holland had the benefit of peer review for his paper including the manufacturer of the AES who is part of the SSWG. Holland actually discards the 18:25:27 BFO as an outlier (as you probably read). Apparently, even the manufacturer does not understand how their equipment works.

    I will follow your “good advice” above and take the “lead”. I am calling your overshoot/transient theory nonsense.

  76. Paul Smithson says:

    @ALSM. Thank you for that information.

    Could you also inform us what is the typical duration of the transient distortion since I believe the x-axis was not quantified in Holland’s paper?

    If your thesis is correct, why should duration / magnitude of transient distortion depend on power-off duration for any power-off where temperature has gone outside of the “5C or so”?

  77. DrB says:

    @Andrew,

    Thanks for posting the photo and information on the SDU thermal control. It seems that an unpowered SDU in steady state would have to be at a temperature in between the lavatory/galley air and the aircraft skin and probably much closer to that of the galley air (assuming the exhaust fans were working).

    @ALSM,

    You said: “. . . the rate of change and overshoot will be about the same for all initial conditions except when the power comes on while the oven is already very close to the setpoint . . . .”

    I agree. It appears to be “about the same” and yet not “exactly” the same for cold starts at somewhat different “cold” temperatures.

    Part of that difference might be related somehow to the difference in the time required to complete the other BITE tests, especially loading the System Table. Certainly this time variability affects the observed BFO pattern. Yet we not only observe time-axis shifts, but also see a significant difference in the best-fit peak overshoot value. For instance, although the peak measured BFO errors at 18:25 (~110 Hz) and at 12:50 (~125 Hz) are somewhat similar, the best-fit peak overshoots are considerably different (~215 Hz at 18:25 versus ~125 Hz at 12:50). Why does this happen?

    I also note that the 12:50 log-on has a smaller peak overshoot than Holland’s Log-on #3, so the 12:50 Log-on (= Holland’s Log-on #6) is not unique among the non-flight log-ons.

    While I think we understand why the short-unpowered-time log-ons have smaller overshoots, the 12:50 and 18:25 events also have peak (fitted) overshoots that differ in amplitude by ~40%. This cannot be due to BITE timing variation. The peak overshoot value will depend on the velocity error when the servo is switched to second order. What could cause the OCXO heating rate (= “velocity” of the control parameter – temperature – in degrees / second) to vary between those two cases? I would guess that one explanation could be if the physical internal SDU temperature (i.e., the portion of the SDU surrounding the oven) were also changing at different rates in the two cases while the crystal is being heated inside the oven using the first-order loop. That would happen if the SDU were colder at 18:25 than at 12:50, which is almost certainly the case in my opinion, although the difference might only be a few degrees. I have never personally experienced a commercial flight where the cabin air was not noticeably cooler at high altitude (as it was at 18:25) than it was when parked at the gate (as it was at 12:50).

  78. Victor Iannello says:

    @DrB, @ALSM, @DennisW: Do we know the sign of the slope of the temperature-frequency curve for the OCXO under consideration? Does the sign change near the operating temperature? Understanding this relation would help us determine whether the approach of the frequency to its final value is a damped overshoot or a simple exponential decay.

  79. ROB says:

    @DRB

    You said “I have never personally experienced a commercial flight where the cabin air was not noticeably cooler at high altitude (as it was at 18:25) than it was when parked at the gate (as it was at 12:50)”.

    I hope you will for forgive me for now injecting a little levity into the discussion, but evidently, you weren’t one of my fellow passengers on the BA trans Siberian flight from Tokyo to Heathrow on 1st May 2006, where we roasted comfortably for 11.5 hours at altitudes of up to 39,000ft, or on a British Airtours Tristar charter flight from Gatwick to Heraklion on 16th June 1984, when due to the eccentricities of the Tristar’s air con design, my feet froze while my head boiled. (end of levity)

    However, as Andrew has pointed out, the pressure cabin metal skin is fully insulated throughout. This would include the area in the immediate vicinity of the E11 equipment rack where the SDU is housed. So, if the logon characteristics at 18:25 do in fact point to a significantly lower SDU ambient temperature at 18:25 than at 12:50, then the most likely explanation, in my opinion, is that the cabin had been depressurized for an appreciable period prior to the 18:25.

  80. DrB says:

    @ Victor,

    The sign is positive everywhere within the transmission-enabled band. Higher temperature = higher BFO.

  81. DrB says:

    @ROB,

    You may be correct. I was reluctant to take that last step (leap?). However, environmental testing of the SDU at low temperatures could tell us if the 18:25 log-on can only be matched with very cold/thin cabin air. On that point I will say that reduced cabin air pressure will also lower the SDU temperature.

  82. Victor Iannello says:

    @DrB: Do you have a reference for this OCXO? If true, then the transient temperature profile represents an overshoot.

  83. Ge Rijn says:

    I’m following this specialists discussion with great interest.
    Finally independent investigation possible on the original Inmarsat-data.
    And it seems already a key-point of the puzzle has been found with comparing the log-ons.

    Just to put a thought in the pocket I think Dr.B mentions something important about temperature and air-pressure.
    Indeed the isolating properties of thin air at ~30.000 feet would be much higher tha on sea/land-level.
    Warming-up something with air as an intermediate would take longer in a decompressed cabin at high altitude.
    Maybe the air-pressure is even more decisive than the air-temperature.
    Don’t know if this could have significant consequence in those values.
    Just a thought.

  84. ALSM says:

    A few facts:
    Honeywell Frequency Reference Oscillator (OCXO) Assembly 81771-MBE
    Honeywell Frequency Reference Oscillator Schematic 81771-WDME
    Racal part number 81771-MBE
    SC cut crystal ovenized oscillator manufactured by CQE.

    The SC cut is chosen for a near flat turnover. The turnover point is ~90C.

    Good luck tracking down any additional info. I spent a fair amount of time at it 3 years ago. The manufacturer has been out of business for awhile. No spec sheets found.

    More details here: https://goo.gl/1Zx0WZ

  85. Victor Iannello says:

    @ASLM: So for Cut 0, we know that the frequency dependency on temperature is small at the operating temperature, by design. But do we really know if the slope of the curve is positive or negative? Bobby thinks it is surely positive. I am trying to understand why.

  86. ALSM says:

    Victor:

    All SC cut crystals exhibit an S shaped curve. They all cross at the inflection point around +92C. OCXO designers select a cut that provides zero TC at the turnover point. The turnover point is always selected so as to minimize the required oven temp for a given maximum ambient operating temp. The oven needs to operate about 20C above the maximum ambient temp spec. In this case, the maximum ambient operating temp is +55C, so I’m guessing the set point temp is +75-80C. For all SC cuts, for a temperature below +92C (see chart in Fig 7), the TC is positive below the setpoint temp. Nobody uses the turnover points above +92 unless the required ambient operating temperature is very high (like +90C). Again, referring to Fig 7, the best SC cut for a 75C oven is about +0.7 degrees (point where red line crosses 75 degrees).

  87. DrB says:

    @VictorI,

    If I may add to Mike’s information, the set point is at +75 C and the turnover point is at +90 C. In other words, the slope does not go through zero and become negative until the temperature exceeds 90 C. The operating band (near the set point) is always on the positive slope. See Figure 7B on page 6 in this reference for an SC-cut temperature curve and explanation:

    https://drive.google.com/file/d/0BzOIIFNlx2aUbERKaE03QThWMnc/view?usp=sharing

  88. DrB says:

    @DennisW,

    In this situation one function of “a” lead person (not necessarily “the” lead person) is to propose a theory that explains all the data (or at least as much of the data as other theories can explain). The “data” in this case are the 18:22 radar location/track/speed and the 18:25-18:28 BTOs/BFOs. I have demonstrated and published an explanation that is entirely consistent with all the data. You have not.

    You take the position that because the first BFO at 18:25:27 is close to the “expected value”, that means my theory is somehow incorrect. To the contrary, my theory shows under what conditions a small apparent BFO error can occur. In this case, a turn to the North (which is determined by the BTO data) happens to increase the BFO by roughly the same amount (about +25 Hz) as the transient error for the log-on request ( about -30 Hz).

    You have not presented an explanation of why the 18:25:27 BFO (the log-on request) is good but the one 7 seconds later at 18:25:34 (the log-on acknowledge) is either bad or due to an aircraft maneuver during that 7 seconds. You can’t fall back on the “it’s unreliable” excuse (and neither can Ian Holland). In my opinion, NONE of the BFOs are unreliable – just not understood by everyone. In my opinion, no plausible B777 maneuver can produce a +130 Hz change in BFO in 7 seconds when cruising, as is needed to explain the second BFO (at 18:25:34) if the first one has no warm-up transient error (because the SDU had not lost power). If the first one is not affected by a warm-up transient, then none of them can be affected. I am still waiting for you (or Oleksandr or anyone else) to present a maneuver that will explain all the BTOs and BFOs from 18:22 (by calculation) to 18:25-18:28 (by measurement) other than a 15 NM right offset and an SDU warm-up transient occurring together. I won’t hold my breath.

  89. Victor Iannello says:

    @ASLM, @DrB: Got it. As long as the operating point is less than the turnover point (TC=0) and even further from the inflection point (TC<0), then TC should be positive. So it seems the observed transient behavior is indeed on overshoot of both temperature and frequency.

  90. TBill says:

    I would say thin air could result in either high temp or low temp.
    The high temp possibility comes when the thin air is less conductive of heat away from the heat source. The low temp possibility comes if the the thin air is cooler.

  91. ALSM says:

    Victor & Bobby:

    Yes, I think we are all nearly on the same page. But to be clear, the inflection point is always near +92C for any SC cut. For cuts <0 degrees, there is no turnover point. All cuts 0 degrees, there will be two turnover points, one below and one above the inflection point. The one below +92C will have a positive TC below the turnover point, a negative TC between the 2 turnover points, and a positive TC above the higher turnover point. Typically, for operation up to +55C, an SC cut crystal will be specified to have a cut of ~+0.7 degrees, resulting in a lower turnover point of ~+75C. The goal is to match the turnover point to the set point temp so that the operating point at steady state is right on the “top” of the curve where the TC is zero. Obviously, every OCXO is slightly different, so some will end up with a tiny positive TC at steady state, and some will have a tiny negative TC at steady state. But in both cases, the TC will be positive approaching steady state from initial power on.

    In summary, the the TC will be positive for all cases except possibly the last few Hz of the stabilization profile. There can be a tiny ripple at the end if the set point is exactly on top of the crystal curve, and the thermal control loop has a tiny bit of ripple closing in.

  92. Richard says:

    Hi All,

    In the linked file below, I have decoded all the ACARS Messages from the Signal Unit data included in the full set of Inmarsat Satellite Data.

    https://www.dropbox.com/s/8ngsielzp0fegct/ISUs%20and%20SSUs.pdf?dl=0

    This includes the previous flight of 9M-MRO to MH370, which was MH371 from Beijing (ZBAA) to Kuala Lumpur (WMKK). The ACARS Messages include all the standard OUT, OFF, ON and INTO Messages, Position Reports, Engine Health Monitoring (EHM), Fuel Usage, Weather En-Route and Maintenance Messages after landing.

    The ACARS Messages allow you to check the Fuel Usage in each Engine, Health of each Engine and the Aircraft Systems function on the previous flight and compare this data with the start of the MH370 flight.

    There was a long list of Maintenance Messages after landing from Beijing:
    MSG 2412080 – DB ELMS P210 ELECTRONICS UNIT
    MSG 2123004 – DB CABIN TEMPERATURE CONTROLLER (R CHAN1)
    MSG 2714012 – DB PRIMARY FLIGHT COMPUTER (RIGHT)
    MSG 3158592 – DB CPM/COMM(M001) IN RIGHT AIMS
    MSG 3230530 – DB BRAKE SYSTEM CONTROL UNIT (PRIMARY CHAN), DB BRAKE SYSTEM CONTROL UNIT (SECONDARY CHAN)
    MSG 2112138 – DB ELMS P210 ELECTRONICS UNIT
    MSG 2360591 – DB SATELLITE DATA UNIT (LEFT)
    MSG 2342610 – DB OVERHEAD PANEL BUS CONTROLLER (R), DB OVERHEAD PANEL BUS CONTROLLER (L)
    MSG 2355050 – DB AUDIO MANAGEMENT UNIT

    You can also see the data from the Engineers trying out the SDU between flights.

    I also detected a few minor typos in the Malaysian Factual Information in their reporting of the EHM.

    For the EHM Parameters, please refer to the Malaysian Factual Information pages 133 and 134 (linked below).

    https://www.dropbox.com/s/ghbkuv1xvpvf61a/EHM%20Parameter%20Key%201.png?dl=0

    https://www.dropbox.com/s/72lfqyvzp08xq8l/EHM%20Parameter%20Key%202.png?dl=0

    This file was checked against the Malaysian Factual Information pages 135 and 136 as well as pages 168 to 180.

    There appear to be a few typos in the Malaysian Factual Information pages 135 and 136.

    Then again, there might be a few typos in my file, although the decoding process was mostly automated.

  93. Paul Smithson says:

    @Richard,

    Quite an achievement. Thanks for sharing. At last we have the means to do a proper validation/calibration of the fuel models using the previous flight data. I trust Dr B will be on to it faster than you can say Pet Shop Boys.

    With regard to the numerous fault/error messages. That’s intriguing. I wonder how many of these would show up as caution messages to the pilots? @Andrew – does this number of fault/error messages surprise you?

    Perhaps we may be seeing multiple symptoms of common fault/error? I see from the discussion linked below that multiple messages might be explained by common fault at bus controller: “something went wrong in the bus controllers and the first thing the maintenance system “noticed” was that it couldn’t see the component that runs through them. Maintenance message logic is *extremely* sensitive to timing when it comes to the order effects show up.” http://www.airliners.net/forum/viewtopic.php?t=768565 that multiple messages

    Perhaps somebody with access to the 777-200 Fault Isolation Manual could look those up and see if they can make sense of them.

  94. Paul Smithson says:

    ps: Is “audio management unit” not the piece of kit required for functionality of ALL radios, VHF and HF? Just saying 😉

  95. Andrew says:

    @Paul Smithson

    RE: “@Andrew – does this number of fault/error messages surprise you?”

    No, the maintenance computer generates quite a few messages at the end of every flight, but most of them are minor and of no consequence. Unfortunately I don’t have access to the FIM so can’t decode the messages in Richard’s file. However, anything that had a ‘flight deck effect’, such as an EICAS warning, caution, advisory or status message should have generated a separate ACARS report at the time of the occurrence. The fact there were no such reports suggests the messages were of no consequence to the operation of the aircraft.

  96. Victor Iannello says:

    @Richard: Thank you for your hard work.

    @All: Here are three lines of inquiry based on information extracted so far from the complete satellite data set:

    1. The reason for SATCOM’s numerous switching of links between the IOR and POR satellites during flight MH371 (previous to MH370).
    2. The mean and standard deviation of the BTO offset for a log-on request. Inmarsat has advised using a mean value of 4600 μs with no advice about the variation. However, the data from MH371 suggests a mean of about 4550 μs with a significant variation. This could impact both the location and error we associate with the 7th arc.
    3. The meaning and significance of the maintenance messages generated when 9M-MRO landed in Beijing.

  97. George Tilton says:

    @Victor
    “1. The reason for SATCOM’s numerous switching of links between the IOR and POR satellites during flight MH371 (previous to MH370).”

    You can find the reason in:
    “Aeronautical Air-Ground Data Link Communications” by Mohamed Slim Ben Mahmoud, Christophe Guerber, Nicolas Larrieu, Alain Pirovano, Jose Radzik

    The pertinent information is in the Satcom Section 1.3.3 paragraph 1.3.3.4

    “AMSS does not implement actual handover procedures as in mobile networks.
    However, it may be necessary for an AES to move from one satellite to another
    one (or from one regional beam to another one). One of the following events may
    trigger the handover procedure: P-channel degradation (detected either by loss of
    clock synchronization for more than 10 s or failure of log-on renewal), satellite
    below elevation handover threshold with another satellite being at least 1° higher
    than the log-on satellite for more than 10 s and user command. As a result, AES
    proceeds with log-off (except P-channel degradation) then connects to the new
    GES applying the procedure presented in Figure 1.18.”

    Here is a link to a preview:
    https://reader.bookshout.com/books/834451/book_data/preview

  98. ALSM says:

    George/Victor…

    From ZBAA, the AES antenna pointing angles were as follows:
    AZ EL
    IOR 243.7 19.7
    POR 109.1 12.9

    The elevation angle was better (>1 degree) for the IOR bird all the way to WMKK. For these relatively low elevation angles (especially to the POR s/c), there may have been some multipath issues on the ground at ZBAA. However, I would not expect multipath to be much of an issue above 10 degrees while in the air. The coding should take care of that. P-channel degradation is one possible cause. Collisions on the random access R channel is another.

  99. Richard says:

    Hi All,

    I have been asked by some of you for an Excel spreadsheet version of the data that I published earlier today as a .pdf at 02:35am.

    Here it is:

    https://www.dropbox.com/s/ra3x6lzba91kybz/ISUs%20and%20SSUs%20Export.xlsx?dl=0

  100. Richard says:

    @Paul Smithson

    The reason for sharing the data is so that multiple people can pursue different lines of investigation simultaneously.

    As Victor has pointed out already, there are a number of different avenues to explore.

    Anyone of us can only properly analyse a limited number of ideas at one time.

    The fuel analysis is only one aspect out of many, that can now be re-analysed based on the actual facts from the previous flight. The engines were not changed between flights MH371 and MH370 on 7th March 2014. The serial numbers are identical for both flights and they are still ESN 51463 Left and ESN 51462 Right.

    It would be good to have a thorough analysis of the Engine Health Monitoring at 01:34:19 UTC and 01:43:52 UTC as compared to 16:41:58 UTC and 16:52:21 UTC.

    An analysis of APU usage would also be helpful.

    An analysis of the Maintenance messages between the flights may be very revealing.

    Even an analysis of the In-flight Entertainment Equipment (IFE) messages regarding the “Pet Shop Boys” might tell us something we do not know.

  101. George Tilton says:

    @ALSM
    I wonder if the switching between satellites on the previous flight was a symptom of an impending hardware failure in the SATCOM or power feed to the hardware…
    Did MAS maintenance do any diagnostics on the SDU during the lay-over or just the normal between flight stuff?

  102. HB says:

    it is remarkable that a number of these systems in the maintenance messages, have been suspected to be isolated or failed or modified during the flight. Is that a coincidence or is there a possible link?

  103. Paul Smithson says:

    @Richard. “The reason for sharing the data is so that multiple people can pursue different lines of investigation simultaneously.” Indeed – and many thanks for making this important contribution.

  104. ALSM says:

    George: The frequent s/c swaps look more like network issues, not AES hw or sw issues. I would like to know what the sdu squawk was about.

  105. DrB says:

    @ALSM,

    You said: “In summary, the the TC will be positive for all cases except possibly the last few Hz of the stabilization profile. There can be a tiny ripple at the end if the set point is exactly on top of the crystal curve, and the thermal control loop has a tiny bit of ripple closing in.”

    This would explain why some SDUs show overshoot and some do not, depending on the exact temperature setpoint made by electronic adjustment in the temperature control system. This is an issue for thermal testing of SDUs to understand the 18:25 BFO transient. Since we don’t have the 9M-MRO SDU to test, one must be used that behaves similarly or is adjusted to behave similarly (with overshoot).

  106. ALSM says:

    Bobby:

    I don’t think the last few seconds and Hz matter. It’s the first few minutes that we need. I would set up the whole AES (except antenna) in the chamber and run several power on tests at different temps. They should measure the SDU 10.08 MHz OCXO frequency vs. time every second, recorded with a counter locked to a rubidium (or similar) standard. Also record the timing of all the logon related events using a simulator like the one manufactured by Square Peg, or just put an antenna outside and use the satellite/GES. Do the test at +25C, 0C and -30C. The goal would be to see if the OCXO warm-up transient timing changes relative to the R600 logon timing, as a function of cold start ambient temp.

  107. ALSM says:

    Bobby:

    I would add…If Thales would agree to conduct this test, it could potentially provide a very important clue. If the tests confirm that the SDU was probably VERY cold at 18:25, then we could narrow down the onboard scenarios that might have taken place on the aircraft between 17:21 and 18:25.

  108. Peter Norton says:

    > Andrew says:
    > The fuselage is insulated by fibreglass/mylar insulation blankets
    > mounted next to the fuselage skin, as shown in the following photo:

    Here is a higher resolution photo:
    http://airchive.com/blog/wp-content/uploads/2013/04/EVERETT-2013-777-FUSELAGE-INTERIOR.jpg

  109. Ge Rijn says:

    @ALSM

    Just in case it could be usefull I add this thought to your suggestion of testing the OCXO/SDU at different temperatures:
    Could it also be usefull to test the OCXO/SDU in a simulated high altitude (~35.000ft) thin air/low pressure environment?
    I mean in such an environment the transference of heat would probably be less efficient cq. slower?

    I did some interesting reading on OCXO’s. I put it here for others to read also:

    https://www.vectron.com/products/literature_library/ocxo.pdf

  110. David says:

    @Ge Rijn. Thanks. p7 Orientation effect, “When the physical orientation of an oscillator is changed, there is a small frequency change…, due to
    the change in stress on the crystal blank resulting from the gravitational affect upon the crystal supports.”

    As an extension of that pulling g in a high descent rate theoretically could have an effect.

  111. ALSM says:

    David; Ge Rijn

    I do not believe the gravitational effect is significant at 00:19 compared to the descent rate induced BFO change. I remember looking at this in depth 2-3 years ago. Can’t find any notes now, but I remember the result. According to Greenway:

    “The magnitude of these frequency shifts is determined by the quartz crystal’s acceleration or “g-sensitivity” vector and the characteristics of the applied acceleration force. The range of typical g-sensitivities for bulkmode quartz crystals can span several orders of magnitude, from less than 1×10-10 per g for a carefully made precision SC cut…”http://www.greenrayindustries.com/library/accsens_sglpage.pdf
    This report infers <1Hz/g at L band.

    Heat transfer within the OCXO is almost entirely via conduction and radiation, not convection, so I don't think a lower pressure environment would have much effect on the warm-up timing. The crystal itself would be in a hermetically sealed nitrogen filled package.

    That said, if Thales had a thermal-vacuum chamber, they could check the frequency vs. pressure easily. That test could be suggested, if they have not already done so during certification.

  112. David says:

    @ALSM. Gravitational effect. Thanks for that.

  113. Victor Iannello says:

    @Andrew said, However, anything that had a ‘flight deck effect’, such as an EICAS warning, caution, advisory or status message should have generated a separate ACARS report at the time of the occurrence.

    Here are ACARS messages for UA930, which was a B777 flight from San Francisco to London that returned to San Francisco after experiencing problems. Notice that there is the following message:
    MSG 2360181 I 2020 02SEP97 CL I 1 PL
    DB SATELLITE DATA UNIT (LEFT)

    Satellite data Fault 2360181 seen here must be more serious than Fault 2360591 seen on MH371 as it generated an ACARS message, as did many other faults.

  114. ROB says:

    @ALSM, DrB, All

    ALSM said “Bobby, I would add…If Thales would agree to conduct this test, it could potentially provide a very important clue. If the tests confirm that the SDU was probably VERY cold at 18:25, then we could narrow down the onboard scenarios that might have taken place on the aircraft between 17:21 and 18:25”.

    Indeed! It could be one of the most significant clues to emerge to date. Because, if it can be shown that the SDU must have reached a very low temperature prior to the 18:25 logon, it would be persuasive evidence of deliberate cabin depressurization. If the aircraft had suffered an accidental depressurization, the crew would have immediately carried out an emergency descent. Obviously,this didn’t happen. In fact, the aircraft maintained its cruising altitude and followed what looks suspiciously like a preplanned flight path, avoiding Indonesian airspace.

    If the aircraft carried out an offset manoeuvre at around 18:25, while just out of primary radar range, it would show that the PIC was still a PIC at this time, ie still fully in control.

  115. Ge Rijn says:

    @AlSM

    Thanks explaning. I later read in another paper the same as you state.
    Most heat transfer is by conduction.
    Although in this particular OXCO there is some minor heat transfer by convection with air as an intermediate between the heater and the resonator: page 56 – 59, figure 4.6 and 4.7.

    https://etda.libraries.psu.edu/files/final_submissions/3071

    But as you say the OCXO used in 9M-MRO is a sealed unit filled with nitrogen then the air pressure around it would make no difference for the volume or density of the nitrogen inside won’t change.

  116. Ge Rijn says:

    @VictorI

    On the ACARS message from UA930:

    ‘DB SATELLITE DATA UNIT (LEFT)’

    Does this mean there were two SDU-units on that plane (LEFT and RIGHT)?
    And if so could this have been the same for 9M-MRO?

  117. Victor Iannello says:

    @Ge Rijn: I’ve seen no evidence to suggest there were two SDUs on 9M-MRO. A reference to LEFT doesn’t mean there was a RIGHT.

  118. buyerninety says:

    @Andrew
    I’ve seen a suggestion the EFB can display Airplane Condition Monitoring Function (ACMF)
    messages. Does it also contain a look-up table that lists all the message numbers & their
    causes and effects?
    (Asking because we don’t have access to a Honeywell ‘Aircraft Diagnostic Systems Engineer’.)

    @Andrew
    With ADIRU inoperative, the SAARU provides heading & attitude, but not correction for wind
    acting on the airframe and not correction for Coriolis effect – Yes?
    The FMC makes no correction for Coriolis effect – Yes?

    ________________________________________________________
    (@VictorI
    Maintenance messsages, vide;)
    http://elektroarsenal.net/fault-reporting-and-correlation.html

  119. DrB says:

    @ALSM,

    You said: “I don’t think the last few seconds and Hz matter. It’s the first few minutes that we need.”

    I’m not sure I understand your reference to the last few seconds. My point was that according to Ian Holland’s paper, each SDU has a different transient curve. Presumably this depends on the exact temperature set-point relative to the crystal characteristic. I think that if you want to convince anyone that you are faithfully recreating the BFOs after start-up from various temperatures, you need to use a SDU that has a similar shape of transient error as the SDU in 9M-MRO, especially so during the first 2-3 minutes after LOR.

    Another benefit of thermal testing would be a more accurate estimate of the time power was actually applied to the SDU prior to 18:25.

    It seems possible to use the log-on BFO error pattern as a thermometer of the SDU/OCXO temperature at power-up. We just need to calibrate that thermometer with lab testing of a similar unit.

  120. ALSM says:

    Bobby:

    Good suggestions.

  121. Don Thompson says:

    @Victor,

    Reading the timeline of the ACARS msgs related for the 2 Sep 1997 UA930 event, and the subsequent maintenance release for the re-scheduled departure, my interpretation is that N788UA lost power to the L AC Bus (due to GCU issue first noted at 20:13) subsequent to take off (at 20:12). The maintenance messages indicate detection of a power interruption by the various systems.

    An inadvertent/non-normal loss of an engine IDG power supply while airborne is not mitigated by the ‘no break power transfer’ (NBPT) design of the electrical system. The NBPT design is intended to mitigate interruption when transferring supply from two concurrently active sources when on the ground.

    At 20:20, a related SDU message is generated (but not relayed air-to-ground until 20:22). If the power break exceeded the SDU’s capability to hold-thru, a reboot will have ensued.

    I note also that the airphones were placarded as inop since 4 Aug 1997.

    Don

  122. DennisW says:

    @DrB

    “You have not presented an explanation of why the 18:25:27 BFO (the log-on request) is good but the one 7 seconds later at 18:25:34 (the log-on acknowledge) is either bad or due to an aircraft maneuver during that 7 seconds. You can’t fall back on the “it’s unreliable” excuse (and neither can Ian Holland). In my opinion, NONE of the BFOs are unreliable – just not understood by everyone. In my opinion, no plausible B777 maneuver can produce a +130 Hz change in BFO in 7 seconds when cruising, as is needed to explain the second BFO (at 18:25:34) if the first one has no warm-up transient error (because the SDU had not lost power)”

    I agree. I have no explanation for the 18:25:34 BFO value. I remain troubled by the accuracy of the 18:25:27 value. My only admonition to you is that your explanation requires a bit of luck. That does not mean it is incorrect.

    I also agree that testing by the manufacturer would be an excellent idea, and I suggested the same thing myself. I can’t recall if it was here or in a private email to Victor. Holland’s paper could have been so much more valuable with that information.

  123. Mick Gilbert says:

    @ROB

    Re: “It could be one of the most significant clues to emerge to date. Because, if it can be shown that the SDU must have reached a very low temperature prior to the 18:25 logon, it would be persuasive evidence of deliberate cabin depressurization.

    As was pointed out to me when I was contemplating the likely cabin temperature subsequent to a depressurisation event, the end-state temperature is largely a factor of how rapidly the depressurisation occurs. Deliberately depressurising the airplane by manually selecting the outflow valves to open is unlikely to depressurise the airplane rapidly enough to achieve a very low temperature.

    Re: “If the aircraft had suffered an accidental depressurization, the crew would have immediately carried out an emergency descent.

    The nature of the accident, the presence of concurrent or rapidly escalating issues, the capacity of the crew to react and the status of airplane systems would all impact upon whether an emergency descent was initiated and whether the airplane responded.

  124. Andrew says:

    @buyerninety

    RE: “I’ve seen a suggestion the EFB can display Airplane Condition Monitoring Function (ACMF)
    messages. Does it also contain a look-up table that lists all the message numbers & their
    causes and effects?”

    The Tech Log function of the EFB might display maintenance messages associated with existing flight deck effects, but I’m not sure because our aircraft don’t have an EFB. However, all existing maintenance messages and associated information can be displayed on the Maintenance Access Terminal (MAT) located next to the second observer’s seat in the flight deck. The maintenance engineers interrogate the MAT during every turn around to check for faults. That said, I don’t think it allows you to look up faults that haven’t already been detected by the system. I believe you’d need the Fault Isolation Manual (FIM) to find that information.

    RE: “With ADIRU inoperative, the SAARU provides heading & attitude, but not correction for wind
    acting on the airframe and not correction for Coriolis effect – Yes?
    The FMC makes no correction for Coriolis effect – Yes?”

    Correct. The SAARU heading information is only valid for 3 minutes after the ADIRU fails. The heading must then be updated periodically by entering the standby compass magnetic heading in the FMC POS INIT page.

  125. Mick Gilbert says:

    @Andrew

    G’day Andrew. You said, “The SAARU heading information is only valid for 3 minutes after the ADIRU fails. The heading must then be updated periodically by entering the standby compass magnetic heading in the FMC POS INIT page.

    What happens if the heading is not updated manually? How might the lack of updates to the heading affect an airplane flying to a CMH?

  126. Andrew says:

    @Mick Gilbert

    The SAARU’s initial heading is based on the ADIRU heading immediately before the failure. After the failure, the SAARU has no way of compensating for changes in the magnetic variation; it is simply a glorified directional gyro with no magnetic reference. Consequently, its heading output must be periodically updated using the magnetic compass. If that isn’t done, then the SAARU heading output will be in error by an amount equal to the change in magnetic variation since the ADIRU failure.

  127. Andrew says:

    In other words, in terms of heading reference, it behaves the same way as a basic gyro compass in a Cessna 172!

  128. Victor Iannello says:

    Don Thompson said: An inadvertent/non-normal loss of an engine IDG power supply while airborne is not mitigated by the ‘no break power transfer’ (NBPT) design of the electrical system.

    Are you saying that you believe that a flameout of the left engine, with the right engine running and the tie breaker functioning, would cause the SATCOM to reboot?

  129. David says:

    @Mick Gilbert. “Deliberately depressurising the airplane by manually selecting the outflow valves to open is unlikely to depressurise the airplane rapidly enough to achieve a very low temperature.”

    Turning off bleed air also would accelerate depressurisation and surely the aircraft would get cold then even if some electric heaters remain, that is if not shed/deselected?

  130. DrB says:

    @Don Thompson,

    What about a failure of the left IDG rather than left engine flame-out? Would that cause the SDU to reboot?

  131. Mick Gilbert says:

    @Andrew

    Thanks for those answers regarding the SAARU heading output. Would it be correct to say that absent any manual heading updates, the CMH output by the SAARU will start to approximate CTH?

  132. Mick Gilbert says:

    @David

    Even with bleed air off the rate at which you can deliberately depressurise the airplane is not likely to be rapid enough to cause a really substantial drop in temperature. I’m pretty sure it was Gysbreght and TBill who had provided the background information about the rate of depressurisation and some end-state estimates (I’ve been trawling through old posts to see if I can find their original comments without success so far).

  133. David says:

    @Mick. I hope they will correct me as needed but I imagine they were talking about adiabatic expansion and conduction through the skin plus heat added from bleed air during that expansion. If the outside air is cold as at altitude, bleed air is turned off, if other internal heat sources contribute comparatively little and it is at night, according to Newton the temperature inside will drop exponentially to that outside.

  134. Andrew says:

    @Mick Gilbert

    RE:“Would it be correct to say that absent any manual heading updates, the CMH output by the SAARU will start to approximate CTH?”

    If the variation at the point where the ADIRU failed is small, then the magnetic heading at that point will approximate the true heading. If the magnetic variation subsequently increases but the SAARU heading is not adjusted to compensate, then the SAARU heading will continue to approximate the true heading.

  135. Andrew says:

    @Mick Gilbert
    @David

    RE: “Even with bleed air off the rate at which you can deliberately depressurise the airplane is not likely to be rapid enough to cause a really substantial drop in temperature. “

    The sudden loss of cabin pressure will cause a drop in temperature. If the air conditioning packs are ON, the cabin temperature should recover somewhat as the air conditioning system automatically pumps in warm air in an attempt to regulate the temperature. However, if the packs are OFF, then the temperature will continue to decrease as cabin heat is lost to the outside atmosphere.

  136. Mick Gilbert says:

    @David

    My thought process was very similar to yours but what I was ignoring was the fact that we’re not just talking about a tube of air, we’re talking about an insulated tube of air with a lot of contents (seats, panelling, passengers, etc) that are all at or above 25°C to start with. Further, we’re talking about only an hour or so between any likely depressurisation event and the 1825 UTC log-on.

    @TBill

    I think that you provided some estimates of end state temperatures for a few different depressurisation scenarios; might I trouble you please to repeat those.

  137. ROB says:

    @Mick Gilbert

    Sorry but you a wrong on both counts. If the outflow valves are selected open and at the same time, the aircon pack inflow valves are selected closed, the cabin will depressurize to ambient in a remarkably short space of time, a few minutes only.

    An accidental, sudden depressurization could in theory happen quickly enough to prevent the crew from taking emergency action, I am careful to say in theory because such a situation has never happened on a large, civil airliner. In every case, the crew were able to take appropriate action before they were overcome by hypoxia. The Helios accident is often cited when discussing this subject, but it is not commonly appreciated that the Helios accident was not a full depressurization. The faulty outflow was only failed open 12% of full open setting, and the air pack inflow valves remained open throughout.

    Loose the yoke of denial, and see the situation for what it was.

    With MH370 what is significant is the sequence of events following the takeover – the events do Not in any way indicate or suggest an aircraft flying out of control with the crew slumped on their seats. On the contrary, the aircraft is observed to make a series of controlled manoeuvres, as it maintained altitude and speed, and carefully observed FIR boundaries, as it flew its way with the help of selected waypoints, to a position whereby it could turn south into the SIO (hopefully) undetected, and continue in a straight line until fuel exhaustion!

  138. buyerninety says:

    …”depressurization could in theory happen quickly enough to prevent the crew from
    taking emergency action”…

    I don’t think Mick specifically said anything about sudden depressurization.

    …”because such a situation has never happened on a large, civil airliner.”

    But for the training/quick thinking of only the co-pilot, this example would
    disprove that assertion ( – some might suggest this example supports Mick’s line of
    enquiry is not unreasonable…);
    https://www.fss.aero/accident-reports/look.php?report_key=1059

  139. buyerninety says:

    @Andrew, Thankyou for that info.

    I notice David said;
    “Turning off bleed air also would accelerate depressurisation and surely the aircraft
    would get cold then even if some electric heaters remain, that is if not shed/deselected?”

    Is the aircraft air heated by ‘electric heaters’, or is it heated by something like a
    ‘heat exchanger’ transfering the heat from the hot air (sourced from the engine{s})
    pressurized bleed air?

    (Note, for clarity, the question disregards any references to heating by the electric
    water heaters (galley), seats (pilot) heating or cockpit windows, etc.. )

  140. Mick Gilbert says:

    @ROB

    Thank you, you actually neatly illustrate my point for me; deliberately depressurising the airplane is unlikely to depressurise it rapidly enough to achieve a very low temperature.

    By your own estimate, the fastest possible deliberate depressurisation utilising flight deck controls would take “a few minutes“. A 7.5 psi depressurisation over a period of a few minutes is not particularly rapid and it would not cause significant cooling, certainly not to the extent contemplated to achieve a very cold ambient temperature for the SDU (~ – 25°C) in only 60 minutes or so.

    If we look at an accidental depressurisations on the other hand, such as QF30 where only 41 seconds elapsed between the hull breach and depressurisation to ambient, the cabin gets colder much faster. Accordingly, if the SDU is determined to have been very cold at the 1825 UTC log-on then an accidental depressurisation is more likely to have been the cause.

    Regarding the contention that “because such a situation has never happened on a large, civil airliner” that it is theoretical, well, the history of aviation safety is replete with examples that bear out the adage that there is a first time for everything; BOAC781, TK981, TWA800, UA585, AF4590 are just a few accidents where the fact that “such a situation has never happened on a large, civil airliner” didn’t matter a jot – theoretical turned to practical in the twinkling of an eye, killing a good many people in the process and in most cases confounding aviation safety experts for quite some time.

    As for that hackneyed spiel that the airplane made a “series of controlled manoeuvres … there were three; the initial turnback which was entirely consistent with the response to an inflight emergency, the turn from south of Penang that may well have been the result of an LNAV Discontinuity Error and an incapacitated crew, and the final turn south;
    … as it maintained altitude and speed … did it? that is highly conjectural, there is evidence that the airplane’s speed and altitude at the turnback and back across the Malay Peninsula varied and, by some accounts, varied considerably, and there’s nothing to suggest these variations didn’t continue as it tracked up the Straits of Malacca;
    … and carefully observed FIR boundaries, … really? the initial turnback took the airplane across the Kuala Lumpur/Bangkok FIR boundary at least twice;
    … as it flew its way with the help of selected waypoints, … really? which waypoints? there is no conclusive evidence that the airplane navigated by any waypoint, selected or otherwise, subsequent to the turnback;
    well that narrative was found wanting on the first day it was articulated and time and evidence have done it no favours.

    Since we’re handing out gratuitous advice, apart from exercising some courtesy, you might also try picking up the glasses of objectivity; the view is a little bit more complex but it is far more instructive than blinkered myopia.

  141. ROB says:

    @Mick Gilbert

    You’re talking complete cobblers. Do yourself and the rest of us a favour and get yourself a thinking brain 😊

  142. ROB says:

    @Buyerninety

    Thank you for the valuable clarification. Welcome, as ever.

  143. Victor Iannello says:

    @Mick Gilbert said: there is evidence that the airplane’s speed and altitude at the turnback and back across the Malay Peninsula varied and, by some accounts, varied considerably

    Which accounts? Other than early reports that were later corrected, I’m not aware of evidence that the altitude might have varied considerably.

  144. Don Thompson says:

    @Dr B & Victor,

    I wrote that N788UA, UA930, “lost power to the L AC Bus (due to GCU issue first noted at 20:13)“.

    Each GCU controls its Bus Tie Breaker (for supply from the opposite engine IDG, or APU, via the Bus Tie Bus) and its Generator Control Breaker (for supply from its IDG). A GCU fault will revert both the BTB and GCB to open.

    If both the GCUs are operating normally, and an IDG supply fails, the switchover is not ‘no break’ but occurs within the hold up period that is defined for the SDU. From AMM, Chapter 24,

    When the ac system changes from one power source to another in the air, it does break power transfers. On the ground, it does no-break power transfers. The system momentarily connects two power sources to one bus so there is no interruption in power. A GCU electronically adjusts an IDG speed to match the power of the two sources to do a no-break power transfer. For transfers between external power and the APU, the APU controller adjusts the APU speed to do the no-break power transfer.

    N788UA experienced a GCU issue, temporarily isolating the L AC Bus. It appears N788UA’s crew manually intervened to restore supply to L AC Bus.

    There’s nothing in the N788UA example to contradict the accepted sequence of events for 9M-MRO: first engine flamed out; that engine’s GCU closed its Bus Tie Breaker; AC bus supply restored within the avionics hold up period.

  145. Victor Iannello says:

    @Don Thompson: OK. It’s the GCU fault that caused the problem on N788UA.

  146. buyerninety says:

    VictorI
    Perhaps Mick is refering to the FI, page 3 – altitude variations, 35700, 32800, 31100 to 33000 ft.

  147. ROB says:

    @Mick Gilbert

    I’ve calmed down enough to give you a relatively reasoned response to you criticisms.

    When I said such an event has never happened to a large civil airliner, I was referring to a case if sudden depressurization that incapacitated the crew which the crew before they were able to don oxygen masks and initiate an emergency descent, in an otherwise still controllable aircraft. I was not referring to cases of rupture of the airframe or airframe disintegration that resulted in the aircraft crashing out of control. If such a fate had befallen MH370, the wreckage would have been found in the South China Sea.

    Your criticism has no credibility. You did not read my post. Your’s is a classic case of myopia, which is the very attribute you have kindly bestowed on me.

    The remainder of your critique is simply not worth the effort of a response.

  148. ROB says:

    @buyerninety

    Alternatively, perhaps Mick was just being economical with the truth, to suit his narrative.

    FYI, the apparent altitude excursions derived from the primary radar tracking, and reported in FI, were more likely due to insufficient calibration in elevation of the radar antennas, nothing more.

  149. Victor Iannello says:

    @buyerninety: I would not consider those to be considerable variations. In fact, those variations are probably barely detectable by the military radar.

  150. Victor Iannello says:

    @Rob: I’ve yet to see a narrative that doesn’t have major holes or inconsistencies. After three years, we’re still left with choosing the best among some questionable scenarios.

  151. sk999 says:

    [2nd try – typo in email the last time]

    All,

    The following is a draft report on flight path reconstruction and BTO/BFO calculations for MH371 (IOR points only). Bottom line – lots of outliers, but no large Mumbai-like drift.

    https://docs.google.com/document/d/14hleZyx1pUPL44yaeHKt6jnSQ3DbgRq2zibbKkFLq2c/edit?pref=2&pli=1

    [Caveat Emptor – since this is a draft, I reserve the right to change anything in the future, including the bottom line.]

  152. Victor Iannello says:

    @sk999: That’s an impressive and very helpful piece of work. Thank you for completing it and sharing it.

  153. Paul Smithson says:

    @sk999. Very nice work. Interesting to see that BTO “drift” mid-flight of ~ -40 microsecs. Do you attribute that to the BTO drift +/- 10 microsecs described in DSTG or to errors on the satellite position predictions?

  154. Victor Iannello says:

    @sk999: In the DSTG Bayesian analysis, it advises that the BTO outliers can be corrected with a term N*7820 μs, where N is an integer. I have found this correction to work for all the points I have tested, i.e., the corrected points fall within the expected deviation from the mean of the sequential burst. It would be interesting for you to recreate your plot with the outliers corrected.

    Why this is significant is it provides justification for using the outlier “log-on acknowledge” BTO value at 00:19:37, which should be more accurate than the “log-on request” BTO value at 00:19:29 because the standard deviation for an anomalous R1200 channel value is 43 μs versus the 62 μs observed for an R600 channel. In addition, the log-on request at 00:19:29 has to be corrected by the offset of 4600 μs, which might also have an associate variation.

  155. ROB says:

    @Victor

    That’s fine. I look forward to studying SK999’s work when I get to a desktop PC.

    There is still much useful work to be done iro fuel flows and path reconstructions West of the Malacca Strait, but I will have to leave that to the “experts”

  156. DennisW says:

    @sk999

    Thanks!

  157. TBill says:

    @Mick
    I have previously given the following numbers for inside AIR temp:
    (for depressurization from FL060 to FL350 from 70 deg F)

    Explosive Decomp (Adiabatic): -85 deg F (-65 deg C)
    Rapid Decomp (Isentropic): -27 deg F (-33 deg C)
    Slow (isothermal): No change

    But this above is only the instantaneous temperature drop of the air inside the cabin. What happens next is the air warms back up because of temperature exchange with the surroundings (seats, PAX, walls, etc.). So I have suggested that the final temperature of the air in the cabin might be 10-20 deg F (5-10 deg C) lower than before depressurization.

    In the chemical process industry we have fairly complex decompression calculation models that take into account the heat exchange with the surroundings. This is done for equipment safety reasons to calculate how low the temperature could go for embrittlement of valves and piping during a loss of pressure. Therefore I have suggested Boeing may similarly have complex models for aircraft temperature upon depressurization.

    Separately I have estimated depressuring rate could be quite rapid if the outflow valve was set to OPEN and the bleed air was set to OFF. In other words, the outflow valves are quite big if they are full open. This contradicts the recent book that said depressurization would be very slow. But the book was apparently not contemplating a worst case intentional depressurization scenario.

  158. Ge Rijn says:

    @TBill

    To shime in on this with a question at the end.

    In any case, explosive, rapid or slow, also after a ‘slow’ full decompression, the decompression will be complete in a few minutes.
    I read considerable drop in temperature is already happening with a leaking door seal (Helios flight report; earlier leaking problems).
    With explosive or rapid decompression, temperatures will drop instandly or quite fast to well below zero C (you mention -65C and -35C).
    And with the bleed air/heating closed this drop will be even faster.

    I don’t think the stored heat in the plane’s materials and in the body’s of the people in the plane can compensate a -65C or -35C to a +5 or 10C level. But if so this could only have been during several minutes for this stored heat would be lost very soon.

    Imo in any case if the plane got decompressed fully at ~35.000ft shortly after the IGARI turn the temperature drop would have become unbearable/unsurvivable within at max. an hour if no emercency descent was made (which was not as far as we know now). Probably much shorter if also the bleeding air was shut off. So I assume the latter did not took place.
    For this would also affect the PIC imo in which he could not have made a controlled flight till at least 18:25 or even till Penang.
    But this is just my guess.

    What is your view on this?
    Do you see any option for a controlled flight till 18:25 or Penang under full decompression and shut-off bleed-air? Or after an explosive decompression without emergency descent?

  159. TBill says:

    @buyerninety
    The American Trans Air Flight 406 depressurization incident report confirms that a combination of OPEN outflow valves and reduced bleed air can quickly decapacitate the PAX.

    This contradicts the recent book “THE CRASH OF MH370” by retired airline pilot James Nixon. Nixon essentially claimed that it was not reasonably possible to intentionally and rapidly depressurize commercial airliners. He said it would be a very slow process and therefore not a reasonable possibility for decapacitating MH370 PAX (paraphrasing here).

    Thank you for the Link. I feel cabin air temperature and pressure should be monitored and reported by ACARS if they are out of bounds.

  160. TBill says:

    @Ge Rijn
    I am not an aircraft engineer, but I feel the extreme low temperature would probably be a transient. I believe the Captain could optionally selectively bring warmer bleed air into the cockpit. Remember the thinner air does not have as much heat transfer ability to cool off your body.

    In the Helios incident, there was a widely circulated and bogus fake news report that the PAX were frozen in their seats. Upon autopsy the PAX were apparently only unconscious at the time of the crash. So there is this tendency to assume deep freeze conditions, but I don’t think so. Now if you left the “heat” off for a long time, then the inside of the aircraft would eventually cool off to ambient. Sounds to me like a pilot could in theory take a few moments to decapacitate the PAX and bring the bleed air back on.

    This is one reason why the powers that be (airline industry) have no interest in finding MH370 and potentially showing the public what could be done by a pilot. We do not know what happened on MH370, and many would like to keep it that way.

  161. Victor Iannello says:

    @All:

    Duncan Steel has created files in ECEF and LLA coordinates for the IOR and POR satellites for 25 hours starting at 0:00 UTC on March 7, 2014, in 1 second increments. He is making those files available here:
    https://www.dropbox.com/sh/ogh8o1z5sm3n5un/AAB9n7O44ZmzC7zF0fFE_mOsa?dl=0

    In Duncan’s words:
    Both [IOR and POR data sets] are in one-second steps (hence large files: around 6MB), based in the SGP4 integrator, backwards from orbits with epochs on March 09 [sic] as defined by USSPACECOM hence Spacetrak hence agi.com (whence I induct them into STK). Screen grabs as PNG files in the above Dropbox location indicate which reference orbits I used. The essence here is that I am assuming that there were no station-keeping thrusts (i.e. artificial orbit changes) applied to either satellite between the start of the period of interest (beginning of March 07) and the epoch for each satellite orbit (65-71 hours later).

    For those that are using @sk999’s excellent orbital model, I have fit the POR satellite data supplied by Duncan to the orbital model. I find an excellent fit with the following parameters:

    POR parameters:
    lsat = 178.078 deg
    i = 1.0354 deg
    e = 0.000519
    rs = 42164.2 km
    uto = 6.418 hr
    utp = 0.0689 hr

    Victor

  162. Joseph Coleman says:

    @Victor

    Is there any data from the SU log not included between row 5821 and 5822?

  163. Mick Gilbert says:

    @Victor

    Re: “Which accounts? Other than early reports that were later corrected, I’m not aware of evidence that the altitude might have varied considerably.

    As well as the PSR variations referred to by buyerninety there is also Gysbreght’s extrapolation of the Figure 4.2 data from Bayesian Methods in the Search for MH370.

  164. Victor Iannello says:

    @Joseph Coleman: I don’t understand your question. To the best of our knowledge, the SU log is complete. What leads you to believe that data is missing?

  165. Don Thompson says:

    @Joseph Coleman

    Line 5818: 08:02:27 was the final transmission from the AES before the aircraft was powered down after completing the ZBAA-WMKK MH371 service.
    Lines 5819-5821: 09:01:28, 09:01:39, and 09:01:49, show the failed Log On Interrogations sent by the GES (failed, as the aircraft was powered down).
    Line 5822: 12:50:19 is a Log On as consequence of the aircraft being powered up.

    Why might you expect any AES activity to be logged between 08:02:27 and 12:50:19?

    Don

  166. sk999 says:

    Victor,

    Thanks very much for those orbital elements.

  167. Andrew says:

    @buyerninety

    RE: “Is the aircraft air heated by ‘electric heaters’, or is it heated by something like a
    ‘heat exchanger’ transfering the heat from the hot air (sourced from the engine{s})
    pressurized bleed air?”

    Two air conditioning packs cool the bleed air from the engines to a usable temperature. Each pack cools the hot bleed air by passing it through two heat exchangers and an air cycle machine. The cabin is divided into zones that can be individually temperature-controlled. The pack outlet temperature is determined by the zone that requires the coldest air. If the other zones require warmer air, then hot trim air from the bleed system is mixed with the conditioned air that supplies those zones.

  168. Mick Gilbert says:

    @ROB

    The remainder of your critique is simply not worth the effort of a response.

    That’s essentially my view on much of your post. You set out saying that if the SDU was “very cold” (~ – 25°C) at the 1825 UTC log-on then that pointed to a deliberate depressurisation. I pointed out to you that the rate of depressurisation that can be achieved by manual means from the flight deck made that unlikely. You said I was wrong while at the same time providing information that supported my position and contradicted yours.

    When I get something wrong, and I frequently do, I usually accept that I was wrong, thank the person who has corrected me and apologise for any confusion that I may have caused.

    As to the matter of events that have “never happened on a large, civil airliner.” I was making the point that simply because something has never happened before doesn’t mean that it can’t. You criticised me for not reading your post, you clearly haven’t read mine. The examples of “first of” accidents were not direct analogies for MH370; I wasn’t suggesting that MH370 suffered the same fate as BOAC781 or TWA800 and it is disingenuous to misrepresent me as doing so.

    As to “… perhaps Mick was just being economical with the truth … ” perhaps you might be a little less economical with courtesy? I share what research I have on MH370 regardless of what camp it might support; for example, I most assuredly would not have pointed out that the 10N and 45S1-2 data points from the Captain’s flight simulator can be joined by a simple CMT of 180° if I was being “economical with the truth“. My interest is in dispassionately and objectively assessing the data. I have never discounted the possibility of malicious deliberate action with regards to MH370; events may have unfolded in that fashion but equally they may not have. As Victor has pointed out any number of times both the malicious deliberate action and escalating inflight emergency narratives have weaknesses. If you want to settle on malicious deliberate action and ignore the deficiencies in that narrative then that’s fine and beaut … for you. Don’t expect others to do so and kindly refrain from ad hominen attacks and insults when they don’t.

    I don’t think that anything further can be reasonably gained by continuing this particular exchange so I won’t be.

  169. Mick Gilbert says:

    @TBill

    Thank you for repeating that information on depressurisation scenarios and likely temperature drop.

  170. Victor Iannello says:

    @Mick Gilbert said: As well as the PSR variations referred to by buyerninety there is also Gysbreght’s extrapolation of the Figure 4.2 data from Bayesian Methods in the Search for MH370.

    As I said, the altitude variations from military radar were modest, and probably at the limit of the altitude resolution of the equipment.

    As for Gysbreght’s analysis, if I recall correctly, he assumed the total energy, i.e., potential plus kinetic energy, was constant after 17:21. While that might be an acceptable approximation for a short-lived transient such as a zoom-climb, the only way that approximation would be accurate over the course of an hour would be if the engine thrust exactly balanced the drag. (Perhaps that’s what you are proposing?)

    Not to mention that I believe that speed graph from the DSTG report is wrong, as it doesn’t even match the known speed before 17:21 when ADS-B coverage was available.

  171. Victor Iannello says:

    @Mick Gilbert said: As Victor has pointed out any number of times both the malicious deliberate action and escalating inflight emergency narratives have weaknesses.

    It would be an interesting exercise if contributors here were first asked to choose what they believe is the most likely scenario, and then had to make a list of every weakness of that scenario.

  172. Victor Iannello says:

    @sk999: Regarding the orbital elements, I did not derive the parameters from the TLE for POR. Rather, I looked at the 26 position vectors on the hour over the course of 25 hours and determined the orbital elements that minimized the square errors of all the position components as compared to the STK prediction.

  173. DennisW says:

    @Victor

    “It would be an interesting exercise if contributors here were first asked to choose what they believe is the most likely scenario, and then had to make a list of every weakness of that scenario.”

    I think it is equally important to have a set of boxes that a scenario has to check. Focussing on weaknesses is a good thing. Focussing on strengths is equally or perhaps even more important, IMO. There will always be weak links, but there is a lower threshold for a minimal checking of boxes.

    Most people would never even consider the questions:

    1> Why the flight to Beijing?

    2> Why the almost nonexistent Malay response?

    3> Why the coordinates on Shah’s simulator?

  174. Mick Gilbert says:

    @Victor

    Re: “I would not consider those [altitude variations of 35700, 32800, 31100 to 33000 ft.] to be considerable variations. In fact, those variations are probably barely detectable by the military radar.

    Victor, a few points if I may:

    1. If the PSR altitude data in the FI was derived from military radar then the variations seen significantly exceed the height accuracy range of the radars, they are most assuredly “considerable” and well within the detection capabilities of military radar. The Selex AMS RAT-31DL at Western Hill and the Martello S-743D at Gong Kedak are designed from the get go as military-grade primary radars; tracking a target like a B777 (at least twice the radar cross section of a Tu-22M Backfire) at 35,000 feet at ranges of less than 100 nm would have been a doddle; it would have taken only three sweeps to lock it up and they would have then tracked it to the horizon. They are designed to track targets above 100,000 feet so extreme elevation/slant range tends not to be an issue; the target elevation from Gong Kedak would have been less than 10°, well within the capabilities of the S-743D. If the FI PSR altitude data came from military radar then the airplane almost certainly changed altitude.

    2. I suspect that MH370 was not tracked by Malaysian military radar at any time rather the FI PSR data comes from the civilian radars at Kota Bharu and Butterworth (a joint use Alenia-Marconi ATCR-33S). The incomplete tracking, short traces and possible slant range issue just north of Kota Bharu (where the target elevation would have exceeded 20°) are far more likely to be associated with civilian rather than military radar.

    3. While there may be differing views on the basis and accuracy of the altitude variations in the PSR data and Fig 4.2 interpretation they are nevertheless “accounts” of variations in the altitude of the airplane, as is Gysbreght’s extrapolation of the Fig 4.2 data. You can dismiss, ignore, rationalise and/or explain those accounts away but there is no way that anyone can say conclusively that MH370’s altitude did not vary as it tracked back across the Malay Peninsula.

  175. ROB says:

    @Mick Gilbert

    That’s absolutely fine my me.

  176. Joseph Coleman says:

    @victor
    @Don

    Thanks for answering the question in detail.
    I presume after the last failed Logon interrogations from GES there is no other hourly interrogations.

  177. Don Thompson says:

    @Joseph Coleman,

    You asked, “I presume after the last failed Logon interrogations from GES there is no other hourly interrogations.

    Yes, after a defined number of failed interrogations the GES marks the AES as logged off & initiates no further contact with the AES.

    In the AMS(R)S specification to hand, dated 2007, it’s defined that after a period of no datalink activity the GES will make five Log On Interrogations at 10 sec intervals. In the Log we see 3 interrogations. The Log On Interrogation is a GES Management Function: that Inmarsat may have changed the function to make 3 rather than 5 failed interrogations, since 2007, is of no consequence here.

    Don

  178. Victor Iannello says:

    @Mick Gilbert: You can dismiss, ignore, rationalise and/or explain those accounts away

    Yes, “explain” is the right word.

    I believe that radar-derived altitudes are inaccurate. Early reports, later dismissed due to inaccurate radar data, had the plane climbing and diving near IGARI. The RMP report included radar data showing an unrealistically high and steady altitude of FL447 near Penang.

    As for any analysis that derives altitude by assuming total energy is constant, I believe that is only a valid approach over a relative short period of time.

    No, I can’t conclusively say the altitude did not considerably vary while MH370 flew back over Malaysia. However, I am not aware of credible evidence that it did. That was my point.

  179. sk999 says:

    All,

    I added a link to the complete ACARS data, now nicely formatted. Dry weight was 175.5 tons, about 1 ton more than MH370.

  180. Don Thompson says:

    @SK999

    Thank you for producing a consolidated & tabulated FMS position report.

    A gross weight of 490320 is repeated over four reports, perhaps a cut and paste error?

    It has allowed me to make a quick check on the initial conditions (IC), climb (CL), and enroute ER) ACMS/engine status reports which are recorded through the report. The ER reports appear to be released after Top of Descent.

    Don

  181. Joseph Coleman says:

    @sk999

    Great work, I just don’t understand why MAS didn’t just release the easier interpretation of the MH371 ACARS, it’s from the same plane. Perhaps because it wasn’t a specific request.

    @Don
    Thanks again for another detailed explanation.

  182. Paul Smithson says:

    @sk999. How do the fuel burn numbers on MH371 compare to the model that you had derived and validated against the flight brief (fuel plan). Do you still get near-perfect fit, or is your model a bit on the high/low side?

  183. ROB says:

    @sk999

    Yes, very impressive work.

    Do you think there is any future possibility of the systematic BFO errors being understood to the point that they permit the actual navigation parameters (as supplied to the SDU) to be reconstructed, and accordingly the actual flight path?

    You point out that both ECON and LRC modes require a complete flight plan, including an end of descent. Might make these modes less likely on the night in question

  184. TBill says:

    @sk999 per @ROB
    …also, sk999 I was struck by the statement made somewhere that only the MH371 BFO’s from level flight could be fit well. Is that true? and is there a way to correct those “off” points by realizing the actual flight path and seeing why the deviation (Predicted vs. Actual) is happening?

  185. Andrew says:

    @ROB

    RE: “You point out that both ECON and LRC modes require a complete flight plan, including an end of descent. Might make these modes less likely on the night in question”

    That’s not strictly true. VNAV only needs the end of descent point to calculate the descent path and a top of descent point. VNAV works fine in the cruise even if an end of descent point has not been defined by entering an arrival/approach in the FMC. ECON and LRC cruise modes are still available.

  186. ROB says:

    @Andrew

    Thanks for clarifying the situation on VNAV and availability of ECON and LRC.

    However, it’s still clear to me how ECON and LRC cruise function in the absence of a destination point in the FMC. As I understand things, VNAV has to be engaged to enable the FMC to perform a step climb during the flight, to optimize cruise altitude and speed. But I cannot see how the FMC could perform this function without having a target point to aim for.

    Thanks

    Rob

    Rob

  187. ROB says:

    @Andrew

    Re 2nd para of my prev post: of course I meant “it’s still not clear to me”.

    Thanks,

    Rob

  188. Victor Iannello says:

    @Rob: With VNAV engaged, the FMC does not automatically perform a step climb. In VNAV and ECON or LRC modes, the FMC advises the optimum, recommended, and maximum altitudes, as well as when to perform an altitude change. If LNAV is not engaged, or if LNAV is engaged but there are no altitude restrictions associated with waypoints, then the Top of Descent (TOD) is not calculated in VNAV. As Andrew said, the plane can continue at ECON or LRC speed without a descent plan.

  189. ROB says:

    @Victor

    Thank you for the explanation. Much appreciated.

  190. ALSM says:

    Rob/Victor:

    Re “With VNAV engaged, the FMC does not automatically perform a step climb”

    In fact, no plane can “automatically” perform a step climb regardless of the NAV mode. Step climbs must be authorized by ATC in realtime, so they cannot be programmed to occur in advance.

  191. Andrew says:

    @ROB

    Apologies for the delay in replying.

    Expanding on Victor’s comments, the FMC breaks the flight into different phases (eg. climb, cruise, descent) and optimises those phases separately in a bid to minimise the overall cost of the flight. The derivation of the cost function used by the FMC can be found in the Airbus publication Getting to grips with the cost index, but it boils down to:

    τ = ΔF + (CI x ΔT)

    where
    ΔF = trip fuel
    CI = cost index
    ΔT = trip time

    Over a given sector length (ΔS), the cost function can be expressed as:

    τ/ΔS = (FF + CI)/Vg

    where
    FF = fuel flow
    Vg = ground speed

    For a given sector length and cost index, the FMC only needs to look at fuel flow and ground speed to determine the optimum altitude and cruise Mach no., using the aircraft’s weight and wind/temperature data. It then calculates optimum step climb points by using the step size entered in the FMC. That step size defaults to ‘ICAO’ (4,000 ft), but can be zero (ie no step) or any multiple of 1,000 ft up to 9,000 ft.

  192. DennisW says:

    @all

    What we need is a new fuel range analysis. Never mind that Boeing was a member of the SSWG.

    Does that not matter to anyone? Do we really believe we can out analyze the manufacturer of the aircraft when it comes to fuel range or flight dynamics/parameters. Why are we beating this drum?

  193. ROB says:

    @Andrew

    Many thanks. it’s a lot clearer to me now. In particular, step climbs are optional when flying ECON/LRC on a long haul sector, not obligatory. Furthermore, a top of descent point is not a necessary prerequisite.

  194. ROB says:

    @DennisW

    Agree, a new fuel range analysis would be very welcome at this point. Although Boeing was a member of the SSWG, they are primarily a commercial organization. Their first priority is to stay in business – going the extra mile to find MH370 is not high on the list of corporate responsibilities.

    I think it was rather telling that when the search was called off, and they (Boeing) were pressed for a response, their only reply was that if the crew choose to disengage ACARS, then there was very little that could be done (in so many words)

  195. Mick Gilbert says:

    @ROB

    Re: “I think it was rather telling that when the search was called off, and they (Boeing) were pressed for a response, their only reply was that if the crew choose to disengage ACARS, then there was very little that could be done (in so many words).

    It’s interesting that Boeing would make any comment as to possible crew actions when their company position is usually to not comment on open investigations. Do you have a link or reference for that reply from Boeing suggesting that the crew chose to disengage ACARS?

  196. ROB says:

    @Mick Gilbert

    No Mick, sorry but I can’t remember the actual item where I got that from. However, I do remember them being reported as saying it at the time.

  197. Mick Gilbert says:

    @ROB

    Okey doke, thanks, but it would have been sometime around 17 January this year when the search was officially suspended?

  198. buyerninety says:

    Andrew said to Oleksandr (last month);
    “The SAARU sends inertial data on the center flight controls ARINC 629 bus. The SAARU data goes to:
    – Left and right AIMS cabinets
    Left, right, and center primary flight computers (PFCs)
    – Left, right, and center autopilot flight director computers
    .”

    @Andrew, I’m somewhat hazy on what’s occuring here – I would have assumed the aircraft ‘present
    position’ was held by/at the AIM(s), and forwarded from there to other logical units for their
    use/calculation/display.

    An example question I could ask is;
    Can GPS (update of position) Updating be selected to be inhibited in only certain of these units,
    (which might allow, for example, the pilots to see what the affect is on aircraft {displayed} position
    in one PFC {with GPS ‘in the mix’}, compared to the aircraft {displayed} position in another PFC {with
    GPS inhibited -‘out of the mix’} )??

    If the above question is not how it works, could you comment; if GPS position updating is selected to
    be inhibited, is it inhibited in AIM(s) + every unit, or only in some of these units?

  199. buyerninety says:

    @Andrew, I forgot to specify in the above question, for simplicitys sake
    in your answer, please regard the ADIRU as non-operational.

  200. Joseph Coleman says:

    @ Mick Gilbert
    @Rob

    “Boeing said it accepted the conclusion of the authorities leading the search”

    http://mobile.reuters.com/article/idUSKBN1510GO

    Not sure if Boeing actually said this, I’ve not seen any official statement from them.

  201. Victor Iannello says:

    @Andrew said: That step size defaults to ‘ICAO’ (4,000 ft), but can be zero (ie no step) or any multiple of 1,000 ft up to 9,000 ft.

    Minor question: Would a pilot more likely choose RVSM (2,000 ft steps between FL290 and FL410) instead of ICAO as the step size? Or just choose 0 so that the FMC doesn’t assume a stepped climb in its fuel and time calculations?

  202. ROB says:

    @Mick Gilbert

    As I said, I don’t remember the particular report, only that I remember reading it. 😁

  203. ROB says:

    @Mick Gilbert

    If I may add a bit more (I write a post, I sit back and think about the problem, as you do and then I remember something else I wish I’d said at the time).

    Although I can’t remember the particular report, I do remember being taken aback at the time. I remember thinking “how could they?, what a cop out, to absolve themselves in that way”.

    I wasn’t surprised that Boeing referred specifically to ACARS being switched off by the crew because to be honest, the world and his wife already knew that’s what happened. In a press conference just after the plane disappeared, the Malaysian government said that ACARS was switched off and the radar transponder(s) was switched off. They have never changed their story, right to this day.

  204. DennisW says:

    @Rob

    I do believe Boeing acted in good faith relative aircraft performance estimation. Certainly their tools and knowledge base are better than anything we could hope to replicate here.

    Just trying to stir the pot back to path and terminus considerations.

  205. TBill says:

    @ALSM @Victor
    “In fact, no plane can “automatically” perform a step climb regardless of the NAV mode. Step climbs must be authorized by ATC in realtime, so they cannot be programmed to occur in advance.”

    That is interesting.

    Victor I know in PSS777 we can go in LEGS screen and set speed and altitude for each leg, but so far I can’t get the aircraft to adhere to that pre-set change instruction.

  206. ROB says:

    @DennisW

    I hadn’t got you down as a stirrer 😆

    For myself, I was hoping the recently released ACARS data on the previous flight might might lead to a more definitive MRC.

  207. Victor Iannello says:

    @TBill: In a VNAV climb, the altitude won’t exceed what is in the altitude window of the MCP or whatever altitude restrictions are associated with waypoints in the route. (The altitude in the MCP can be set to whatever altitude is cleared by ATC, for instance, to follow a plan yet ensure compliance with ATC instructions.) It may be possible (I haven’t run this down) to schedule automated climbs by associating altitudes with waypoints and ensuring the selected MCP altitude exceeds the waypoint altitudes. When you did your PSS777 test, was the MCP altitude properly set?

  208. ventus45 says:

    @Andrew,

    As a side issue to the step climb questions, I wonder, if there is any way, or any mode, in which the FMC could, either “default to”, or be “programmed to”, “indulge” shall we say, in a classic “optimal cruise climb” as Gross Weight decreases with fuel burn ?

  209. TBill says:

    “When you did your PSS777 test, was the MCP altitude properly set?”

    Not sure…but I may have been trying to exceed the altitude. One thing I had in mind was pre-programming the MH370 wind data to SIO in thin layers and then using pre-programmed minor altitude changes to change the wind along the route.

    But also more importantly it would be interesting to know what could be pre-programmed ahead of a long ghost flight, besides LNAV of course.

  210. ALSM says:

    Re Step Climb comment…I should make clear that I was not referring to what the equipment may or may not be technically capable to doing. I’m not sure about that. I was pointing out the operational limitations. So, regardless of the technical capability, step climbs can only occur following a request to ATC and clearance to proceed. They do not automatically approve. Sometimes, especially when it is very busy, step climb requests are delayed or modified to a different FL. This is what I meant when I stated it is not automatic.

  211. Andrew says:

    @buyerninety

    RE: “I’m somewhat hazy on what’s occuring here – I would have assumed the aircraft ‘present
    position’ was held by/at the AIM(s), and forwarded from there to other logical units for their
    use/calculation/display.

    An example question I could ask is;
    Can GPS (update of position) Updating be selected to be inhibited in only certain of these units,
    (which might allow, for example, the pilots to see what the affect is on aircraft {displayed} position
    in one PFC {with GPS ‘in the mix’}, compared to the aircraft {displayed} position in another PFC {with
    GPS inhibited -‘out of the mix’} )??

    If the above question is not how it works, could you comment; if GPS position updating is selected to
    be inhibited, is it inhibited in AIM(s) + every unit, or only in some of these units?”

    Just to be clear, the inertial data the SAARU sends does NOT include position information because the SAARU does not compute position. If the ADIRU fails, the SAARU will output inertial data including pitch & bank angles, heading, accelerations, etc. That data is output on the centre flight controls ARINC 629 bus and used by those systems that need it.

    Also, the PFCs are part of the primary flight control system (ie the ‘fly by wire’ flight controls); they have nothing to do with navigation or the display of navigation data and they do not use GPS data. Navigation is handled by the Flight Management Computing Function (FMCF) within AIMS. Normally, the FMCF uses a combination of ADIRU, GPS and radio position data to compute aircraft position (primarily the GPS because it is the most accurate). If the ADIRU is not available, then it will continue to use the GPS data if it’s available. Inhibiting the GPS in the FMC only affects the FMCF’s use of the GPS data. As far as I’m aware, it is not possible to inhibit GPS data on one side only.

    @Victor

    RE: “Would a pilot more likely choose RVSM (2,000 ft steps between FL290 and FL410) instead of ICAO as the step size? Or just choose 0 so that the FMC doesn’t assume a stepped climb in its fuel and time calculations?”

    Yes. The pilot would normally choose whatever step size suits the airspace. Some one-way routes even allow 1,000 ft steps. Another way of doing it is to enter a step size of 0 and then manually enter the step altitudes against the waypoints, as per the operational flight plan. It’s also possible to enter the next required step altitude at 1R on the VNAV CRZ page. The FMC will then compute the optimum point to climb.

    @Ventus45

    RE: “I wonder, if there is any way, or any mode, in which the FMC could, either “default to”, or be “programmed to”, “indulge” shall we say, in a classic “optimal cruise climb” as Gross Weight decreases with fuel burn ?”

    No. The minimum step size that can be entered in the FMC is 1,000 ft.

  212. buyerninety says:

    @Andrew said;
    “Inhibiting the GPS in the FMC only affects the FMCF’s use of the GPS data.”

    The above statement might be open to misinterpretation, so to clarify, could you answer this;

    Assuming GPS ‘position updating’ has been inhibited, (and allow a non-short time period has
    passed), what position is being passed from the AIM(s) to the SDU? (- the position with GPS
    derived input ‘in the mix’ or the position that is calculated without GPS ‘in the mix’?)
    Again, consider the ADIRU has been/is non-operational.

  213. buyerninety says:

    @Andrew said;
    “If the ADIRU fails, the SAARU will output inertial data including pitch & bank angles, heading,
    accelerations, etc.

    OK, could you clarify, does the SAARU sense accelerations on the airframe caused by wind, i.e.
    does it calculate & adjust for wind action on the airframe?

  214. Mick Gilbert says:

    @ROB

    Re: “I wasn’t surprised that Boeing referred specifically to ACARS being switched off by the crew because to be honest, the world and his wife already knew that’s what happened.

    I am very surprised that Boeing would offer any comment as possible causation regarding an open investigation, particularly given their typically taciturn approach to commentary regarding MH370 from day one. If you come across a reference for the Boeing statement I’d be grateful if you could let me know.

    The very simple fact of the matter is that nobody (or their wife) can know what caused the transponder signal failure and the interruption to the SATCOM connection and ACARS transmissions. You can speculate till the cows come home but as Tom Sawyer observed, saying so don’t make it so.

  215. buyerninety says:

    URrrggg, Mick, (hopefully, to greatly short-circuit the liklihood of discussion of this
    matter of) who it was that said ‘the crew chose to disengage ACARS‘, – it was probably actually Tim Clark, President of Emirates Airline, in an interview with Der Spiegel, and
    duly translated?/mistranslated?/reported?/misreported? in the DailyFail ;
    http://www.airliners.net/forum/viewtopic.php?f=3&t=577505&start=50#p8745633

  216. Mick Gilbert says:

    @buyerninety

    Thank you for that reference. I have the transcript of the Tim Clark – Andreas Spaeth interview of November 2014 (that’s more than two years earlier than when ROB thinks Boeing made their statement) and while Clark makes a number of observations about ACARS he doesn’t say that he suspects the crew of deliberate malicious action. In fact, Clark states that while he believes that “probably control was taken of that aeroplane” you’re left with the impression that he believes it was someone other than the flight crew;

    “… the events that happened during the course of its tracked flight will be anybody’s guess of who did what and when. I think we need to know who was on this aeroplane in the detail that obviously some people do know, we need to know what was in the hold of the aeroplane, in the detail we need to know, in a transparent manner.

    Clark goes on to largely discount pilot suicide;

    If you eliminate the pilot on a suicide mission, I’m sure you could have put the aircraft in the South China Sea, rather than fly it for seven hours. So if he was on a suicide mission, he would have done it then.

  217. ALSM says:

    Don Thompson and I have been looking at fuel flow information in the “unredacted Inmarsat Logs”. My summary follows here:

    https://goo.gl/fx4xdh

    Credits to Richard Godfrey and Don Thompson for cracking the HEX codes and Gysbreght for suggesting the 6xxx figures were fuel flow (WF values).

  218. sk999 says:

    ALSM et al,

    Nice work on L/R fuel burn. I finished a preliminary fuel burn model for MH371 (while I was at 37,000 feet this evening – quite appropriate, one would think) and hope to have written up in a few days.

  219. Irthe Turner says:

    Here is my 2 cents on a few earlier posts. No intention on my part to step on anyone’s toes. The main reason flights west e.g. to Jeddah or Amsterdam, would not have worked IMO is because 3 in the cockpit is a crowd. Chances of a single person taking control of the aircraft would drop exponentially. Secondly, downing an aircraft with mainly PAX from either of these sovereign countries, lined with very deep pockets, may have resulted in never ending searches and/or these sovereigns taking control of the investigations. Not desirable if the aircraft was meant to disappear for good.

    If the PIC is indeed the offender in this tragic event, he would know MY, MAS, ATC and the area like no other. Response times could be attributed to initial denial anything was amiss, lack of practice as it relates to emergency protocols etc. The intent seemed to be to cause mayhem and confusion as to what was really going on. It worked.
    The turn at IGARI and subsequently flying back over mainland MY is very risky when you have 238 PAX/crew with workable cell phones. It would take just 1 phone to connect to a cell tower to relay some kind of message and this is definitely possible even at high altitudes. If PAX/Crew were not immediately incapacitated, it begs the question were they a threat to the PIC in any way? The FO would have been well aware what was happening which is probably why his cell phone was switched on and was later detected over Penang.

    Even the most seasoned bankers – those that are well versed in the complexities of executing international wire transfers – would be struggling to extort funds from any government successfully.

    Electronic wire transfers require proper authorizations, a minimum 4 eye principal, security tokens, signatories and sometimes call-back procedures to verify the legitimacy of large transfers. This is because the people that own the money will go to great lengths to protect it. Trying to move large sums of money out of any bank account is no walk in the park even if books of fiction makes one believe otherwise.

    The safe return of PAX/crew is only likely if the PIC was willing to sacrifice himself for the sake of all others. The only fate awaiting him would be immediate arrest, extradition and subsequently the death penalty. These countries dish out the death penalty for much lesser offences every day. It is doubtful he would have been hailed a hero and his family would not only have been deeply shamed in MY society but branded by his actions for the rest of their natural lives. It seems the intention was to make the evidence disappear and, to date, it worked.

    You guys are the smartest guys in the room! You will find MH370.

  220. Paul Smithson says:

    @ALSM and others you acknowledge.

    Nice work. What a stonking PDA on the right engine. From the few data points in the summary table it looks like R:L slightly greater than 1.03 and this could obviously be refined using all MH371 data points. With slightly lower initial fuel in R tank and total flight time of ~7.6hrs, you would expect MH370 fuel exhaustion R to occur about 15-17 mins before left. Very interesting. I recall that Dr B’s fuel numbers from ATSB (from MH371?) also indicated much greater PDA R:L.

  221. ROB says:

    @Irthe Turner

    Right on man! That’s slightly more than 2cents worth, though. The vexed question of the PAX mobile phones has been much discussed, as you would guess.

    It’s one reason why the pilot kept his plane at FL350. Picture this, the Captain makes the cabin crew go carefully through the plane, making sure all personal equipment is switched off or on flight mode to prevent “interference with the onboard navigation systems” with the promise that they will be able to use them once they are established in the cruise. That chance never comes, however, because as soon as the FO is locked out of the cockpit, the Captain depressurized the cabin. It’s the element of surprise that overtakes them all. It’s the element of surprise that lets the Captain get into the SIO undetected. Well almost. Things didn’t go 100% to plan.

  222. Andrew says:

    @buyerninety

    RE: “Assuming GPS ‘position updating’ has been inhibited, (and allow a non-short time period has
    passed), what position is being passed from the AIM(s) to the SDU? (- the position with GPS
    derived input ‘in the mix’ or the position that is calculated without GPS ‘in the mix’?)
    Again, consider the ADIRU has been/is non-operational.”

    The SDU receives ‘raw’ ADIRU position information via the AIMS. If the ADIRU is inoperative then the SDU does not receive any position information.

    RE: “OK, could you clarify, does the SAARU sense accelerations on the airframe caused by wind, i.e.
    does it calculate & adjust for wind action on the airframe?”

    Amongst other things, the SAARU outputs backup heading, track, along track & cross track horizontal acceleration and flight path acceleration. I’m not sure what you mean by ‘adjust for wind action’. The SAARU simply calculates the inertial data and outputs it for other systems to use; it doesn’t ‘adjust’ anything.

    If the ADIRU is inoperative, the only lateral AFDS mode available is HDG SEL. The pilots must steer the aircraft by selecting headings to maintain the desired track.

  223. Andrew says:

    @Victor

    A correction to my previous comment about the FMC default step size: The default step size can be ICAO, RVSM, or 0, as selected by the airline in the FMC’s airline modifiable information database.

  224. Don Thompson says:

    Concerning cellphone connect.

    The goal for a cell base station antenna is to provide maximum coverage at ground level, not 35,000ft. However, the planar array antennas used to form the radiated power into a shape resembling a segment of a squashed torus also produce anomalous lobes radiating at high elevations. Imagine the anomalous lobe as a ‘thick walled dish’.

    At 01:52:27MYT, the time of the cellphone network registration, 9M-MRO had initiated its turn around Penang Island. For approximately a minute, the flight path resulted in a consistent range from the BBFARLIM2 cell base station site.

    If the flight path coincided with the high elevation lobe, I suggest this enabled the phone registration with the network LBS.

  225. Don Thompson says:

    @Buyerninety

    It might be worth clarifying that AIMS is a platform and, itself, does not perform a function. The platform is a collection of processor, associated data bus comms modules, and inter-connections between the two AIMS cabinets. AIMS hosts many separate avionics functions, each of which is strictly partitioned within the hardware components of AIMS.

    When it’s described that a particular data path routing involves AIMS, that data path routing involves a specific function hosted on AIMS. It should not be inferred that the data path routing with AIMS involves distribution to very function hosted on the platform.

  226. ALSM says:

    Here is an updated summary of the MH371 fuel flow analysis:
    https://goo.gl/UJTm6J

    The data tends to confirm that the fuel flow sensors are quite accurate. At least, the sum of the left and right rate sensors agrees within 0.13% with the dFWT/dT.

  227. Paul Smithson says:

    @ALSM. That’s very nice. So now we have delta GWT (lbs), delta fuel remaining (kgs) and flow rate R, L, total that all appear to be mutually compatible/consistent.

    Now, @sk999, @Victor, @DrB: How do these empirical burn rates compare to your fuel flow models f(weight,altitude,M,temp)?

  228. buyerninety says:

    @Andrew, Don
    Yes, thanks, that has clarified the question of the IRS source for the SDU.

  229. DennisW says:

    @Don T

    Re: your latest post above

    Is there any doubt that the cellphone registration occurred?

  230. Ge Rijn says:

    @Don Thompson

    I wonder about this still regarding the altitude and distance the FO’s cellphone was detected.

    BBFARLIM2 cell tower is in a south western suburb of Georgetown to the north east of Penang. Surrounded in the south and west by hills with only a more open line of sight to the north east.
    In a city environment a cell tower has a detection range of a few km and on an elevation with undisturbed line of sight a max of ~25km.
    I think this max line of sight range could only have been to the north east with undisturbed line of sight.
    Which is hardly imaginable regarding the location of BBFARLIM2.
    To the south and west the hills would block its line of sight soon imo.

    What would the detection range/diameter of this anomalous ‘disk’ be from the BBFARLIM2 location?

  231. TBill says:

    @Irthe Turner @ROB
    I am in total agreement as far as most likely scenario. However, this is not much new info. Your analysis is essentially equal to what the book Goodnight Malaysian 370 by New Zealand pilot Ewan Wilson and Geoff Taylor said in July_2014.

    I would like to know if Wilson/Taylor have any updated thinking, but I feel their book stands the test of time. It has long been my adopted hypothesis.

  232. Ge Rijn says:

    @Don Thompson

    Like to add that in the flight tests they did afterwards there were detections by several celltowers but none was detected by BBFARLIM2 and as far as I remeber none at ~35.000ft.
    If I remeber it well the highest detection was at ~22.000ft.

  233. Victor Iannello says:

    @Ge Rijn: The tests that were conducted by Malaysia were at lower speeds, lower altitudes, and not exactly the same path that MH370 is believed to have flown near Penang. A connection was detected at the highest altitude flown (FL240), although not for the BBFARLIM2 tower. It would be very difficult to use these results to determine if a cell phone connect was possible for the speed, altitude, and path of MH370.

  234. Irthe turner says:

    @Rob, Tbill, Thanks. Not read the book yet but you piqued my curiosity. Will defitely pick it up.

  235. ALSM says:

    Ge Rijn:

    Information wrt Cell Phone Registration is here:

    https://goo.gl/j9PUAK

    The plane was within range (26 km), and the path was line of sight with sufficient margin for a connection.

  236. ALSM says:

    All:

    BTW…The cell phone tests conducted by the Malaysians were poorly conceived and poorly executed. It is almost trivial to show by link analysis that the cell phone could connect as long as the Doppler was not too high, and the B777 window loss was not over ~25 dB. The plane was only 26 km from the cell tower at the time of registration, within the 32 kn system spec. But none of their tests really addressed these two unknowns.

  237. Lauren H. says:

    @DennisW – Why the Beijing flight? In addition to the above suggested reason to restrict the flight deck to a 2 person crew, how about, “So S&R would waste a week looking in the SCS?”

    Possible caption for your cartoon, “..And they said we wouldn’t find this place either”

  238. DennisW says:

    @Lauren H.

    Yes, the two person crew is a biggie that I had not considered.

    @ALSM

    It is really hard to find decent experimental data on aircraft cell phone connections. I think it is somehow discouraged for the public domain. Used to be a that Google search would bring up the experiment of a couple of farmers from Ohio who tried to call their wives without success from a Cessna that one of them owned. The Malay experiment was not much better.

  239. DrB says:

    @Don Thompson,

    With the aircraft banking to the right turning around the south end of Penang Island, the person sitting in the FO seat with a cell phone in his pocket has an (almost?) unobstructed line of sight to the cell tower through the right side window. That seems like the most likely seat and time for a cell tower connection to have occurred during this flight, so it’s not surprising to me that it did occur.

  240. Don Thompson says:

    @DennisW

    Doubt? No, at least, not on my part. I can’t now recall what motivated me, earlier today, to post a comment about it.

  241. DrB says:

    @Paul Smithson,

    You said: “Now, @sk999, @Victor, @DrB: How do these empirical burn rates compare to your fuel flow models f(weight,altitude,M,temp)?

    My fuel flow model in level flight agrees within 1% with both the MH370 Flight Brief and with a prior flight (not MH371) for which ATSB has provided me with downloaded ACARS and FOQA data. I am checking the MH371 fuel flows now.

    The R Engine Fuel Flow Sensor agrees within a few tenths of 1% with the slope of the R Tank Fuel Quantity Sensor.

    The L Engine Fuel Flow Sensor reads about 0.5% lower than the slope of the L Tank Fuel Quantity Sensor.

    The ratio of R engine Fuel Burn Rate to L engine Fuel Burn Rate in cruise is 1.021 +/- 0.004, based on my analysis to date.

    Corrections to predicted fuel flows include weight, static air temperature, pressure altitude, tailwind, and Cost Index. The Cost Index correction for fuel flow and Mach relative to LRC tables depends in a complex way on Mach, weight, and altitude.

    Based on my analyses so far, the Cost Index of 52 may have been used for all analyzed flights, and the average PDA is very near the MH370 Flight Brief value of 1.5%. The L Engine PDA is near 0.4% and the R engine PDA is near 2.6%. The ratio of R/L FF Sensor readings is 1.033 +/- 0.002 in cruise. The ratio of R/L engine fuel burn rates is 1.021 +/- 0.004.

    I also found that the fuel parameters are quantized in different sizes of steps: Fuel Quantity in 100 kg, gross weight in 40 pounds, and Fuel Flow in ~7 kg/hr. It is much better to estimate total fuel weight using the gross weight and subtracting the Zero Fuel Weight than to use the TOTFW parameter. Unfortunately, the L and R and Center Tank fuel quantities are always rounded to the nearest 100 kg.

    The bottom line:

    (1) the Fuel Flow sensors are quite accurate (within 1%), but probably not quite as good as the Fuel Quantity Sensors,
    (2) when tailwinds are high, corrections must be made to the predicted Mach and Fuel Flow because the FMC reduces Mach (and thus Fuel Flow) depending on the tailwind and the Cost Index,
    (3) after applying numerous small corrections, the measured and predicted fuel flows are in excellent agreement, including the Boeing LRC FF table, the MH370 Flight Brief, and a prior flight of 9M-MRO some days earlier (and my MH371 comparison is not yet complete).

  242. Ge Rijn says:

    @VictorI @ALSM

    I commented on the ‘cellphone’-link @ALSM also provided on Jeff’s blog.
    I think the information in the link is not all correct.

  243. ALSM says:

    Ge Rijn:

    I replied on JWs blog. The base station coordinates stated are correct.

  244. Ge Rijn says:

    @ALSM

    I did the same in return.I guess my Dutch GoogleEarth works the same as yours..

  245. Victor Iannello says:

    @ALSM, @Ge Rijn: I’m not sure we are certain of the location of the cell tower. Don believes it is on the St John’s Ambulance building based on cell tower information he found. On the other hand, if you look at the center of the cell coverage sector from the map in the RMP report, the location does not appear to coincide with the ambulance building. I never resolved this. Perhaps somebody else has.

  246. Oleksandr says:

    @Andrew,

    I apologize for being silent for a while – a bit busy with relocation.

    You wrote: “If the ADIRU is inoperative, the only lateral AFDS mode available is HDG SEL”.

    According to FCOM, besides HDG SEL there are several more roll modes available when the ADIRU fails: HDG HOLD (if magnetic heading is entered as POS INIT page), rollout (irrelevant to our case), and the ATT. Do you agree that it is possible to set a bank angle of, say 10 deg (>5 deg), and leave the plane on its own? This will be the ATT mode, affected by wind (see my analysis of EY440 pattern) etc. In my understanding, there was no definite conclusion with regard to what happens if the plane is left at the bank angle of 0.1 deg: either the AFDS rolls to wind level, but cannot switch to HDG HOLD, or it keeps 0.1 deg because it ‘knows’ about the failed ADIRU.

    Regarding SAARU to SDU data flow. If I recall correctly, the diagram you posted mentioned inertial data, but not the ADIRU specifically. Furthermore, I recall later we figured out that the SAARU data are also sent to the AIMS, but the SDU was not listed as a recipient. This does not exclude, however, such a possibility, right?

    Re: “The SDU receives ‘raw’ ADIRU position information via the AIMS. If the ADIRU is inoperative then the SDU does not receive any position information.” I recall the SDU receives time data via GP channel, and the time is sourced from GPS. Also, while it was stated that the SDU receives time data, was it confirmed that the SDU cannot receive position data via the same GP channel, especially when inertial data is not available? Finally, it is not clear what would happen if only position data are not available to the SDU.

  247. PaxLambda says:

    The cell tower is perhaps this one:
    goo.gl/sTQqKG

  248. sk999 says:

    I finally got aroud to making a comparison of ACARS weather with GDAS. The following numbers are in the sense ACARS-GDAS:

    SAT: Mean difference 1.0, stdev 1.0

    E-W wind (knots): Mean difference -1.1, stdev 4.5

    N-S wind (knots): Mean difference -0.3, stdev 4.6

    These differences are consistent with prior estimates – in particular, the rms error in the GDAS wind (1-d) is of order 5 knots. For reference, Baysian Methods used a steady-state error of 5.7 knots.

  249. sk999 says:

    Addendum: SAT = Static Air Temperature (deg C).

  250. Mick Gilbert says:

    @ROB

    Re: “Picture this, the Captain makes the cabin crew go carefully through the plane, making sure all personal equipment is switched off or on flight mode to prevent “interference with the onboard navigation systems” with the promise that they will be able to use them once they are established in the cruise.

    What exactly are we picturing here? Turning off portable electronic equipment is part of the standard pre-take-off briefing. Passengers can turn their gear back on once the seat belts sign has been turned off, typically as the airplane passes FL150; that would have been about 10 minutes after take-off, about 20 minutes before they reached IGARI.

    Or are you suggesting that the Captain issued some sort of special instruction for passengers to keep their portable electronic equipment off for longer?

  251. Mick Gilbert says:

    @DrB

    With the aircraft banking to the right turning around the south end of Penang Island, the person sitting in the FO seat with a cell phone in his pocket has an (almost?) unobstructed line of sight to the cell tower through the right side window. That seems like the most likely seat and time for a cell tower connection to have occurred during this flight, so it’s not surprising to me that it did occur.

    Snap! That was also my conclusion and it suggests or is at least not incompatible with crew incapacitation.

  252. ALSM says:

    sk999: It should be noted that commercial aircraft have been supplying near real time met data to NOAA, ECMWF, etc. for decades. (I used to work with people at NOAA and NCAR that pioneered the idea.) Now it is SOP everywhere. They not only use the cruise level data, they also use the climb and descent vertical profiles (like radiosondes). Therefore, it is no surprise that the MH371 met data is in good agreement with GDAS. GDAS data is a product that is derived in part from commercial aircraft flying that route.

    https://amdar.noaa.gov/docs/bams/

    https://amdar.noaa.gov/FAQ.html

    https://www.wmo.int/pages/prog/www/GOS/ABO/

    Mike

  253. Andrew says:

    @Oleksandr

    Perhaps I should have been more explicit in my reply. HDG SEL is the only available lateral mode that might be used by the pilots to ‘steer’ the aircraft along a defined track.

    RE: “Do you agree that it is possible to set a bank angle of, say 10 deg (>5 deg), and leave the plane on its own? This will be the ATT mode, affected by wind (see my analysis of EY440 pattern) etc.”

    ATT mode engages when the autopilot is first engaged or when the flight director is first turned on in flight. If the angle of bank is between 5-30 degrees, the AFDS will hold the bank angle and will not roll wings level. If the bank angle is less than 5 degrees, the AFDS will roll wings level.

    However, I believe you’re considering a scenario where the ADIRU fails? If that occurs while LNAV is engaged and the aircraft is established in a turn, then the AFDS remains engaged in an “attitude stabilising mode”. That condition would be annunciated on the PFD by an amber line drawn through the LNAV mode annunciation. If the condition persists, my understanding is that the aircraft would continue in the “attitude stabilising mode” with a constant bank angle. ATT mode would not be annunciated on the PFD in that scenario.

    RE: “Regarding SAARU to SDU data flow. If I recall correctly, the diagram you posted mentioned inertial data, but not the ADIRU specifically. Furthermore, I recall later we figured out that the SAARU data are also sent to the AIMS, but the SDU was not listed as a recipient. This does not exclude, however, such a possibility, right?”

    It has already been stated, several times, that the SDU receives inertial position data from the ADIRU for beam steering. The SDU does NOT receive data from the SAARU.

    RE: “I recall the SDU receives time data via GP channel, and the time is sourced from GPS. Also, while it was stated that the SDU receives time data, was it confirmed that the SDU cannot receive position data via the same GP channel, especially when inertial data is not available?”

    Yes, the SDU receives date/time data from the L AIMS cabinet via the GP 1 data bus. The AIMS cabinets receive date/time data from the multi-mode receivers (ie GPS). The GP 1 data bus does not transmit position data.

    RE: “Finally, it is not clear what would happen if only position data are not available to the SDU.”

    The position data is used by the SDU for beam steering. I’m not sure exactly what happens if that data is lost, but if the SATCOM doesn’t fail immediately, I assume it will eventually ‘lose’ the satellite as the aircraft position changes.

  254. Victor Iannello says:

    @Andrew: Position, groundspeed, and track data are used for pre-compensation of the L-band Doppler shift. Even if the low gain antenna (LGA) is in use, which requires no beam steering, position information is required.

  255. Victor Iannello says:

    @Pax Lambda: Welcome to the discussion.

  256. sk999 says:

    ALSM,

    Yes indeed, I am aware that ACARS data are ingested by NOAA, etc. and your comments and references are very relevant. You stop short of asking the next question – how good are the wind models over the SIO, where commerical flights are few and far between? I have no good answer – it could be that the models are as much fancy as they are fact. Now I do have a scheme to find out …

  257. ALSM says:

    SK999:

    Yes, commercial flights are few and far between in the SIO. But…The GPS/MET demonstration project and later COSMIC systems that I worked on 1992-1997 have revolutionized weather prediction in data sparse areas like the SIO. There are quite a few radio occultation payloads in orbit now, but the one I managed to get launched on MicroLab-1 back in 1995 was the first. Radio occultation payloads observe the GPS L1/L2 signals as they “slice through the atmosphere” when a LEO payload rises or sets relative to the observed GPS satellite. Each LEO payload can observe several hundred vertical temperature and pressure profiles daily.

    With many LEO satellites now carrying GPS radio occultation payloads, the number of soundings in the SIO is now quite dense. From these soundings, the global models develop “geostrophic winds”. It is very cool technology, contributing to the accuracy of winds in the SIO.

    Short Video produced by UCAR & NSF: https://goo.gl/DvuZDH

    http://www.cosmic.ucar.edu/

    http://www.cosmic.ucar.edu/gpsMet.html

  258. DrB says:

    @sk999,

    Did you figure out the correct gross weight values for MH371 at 2:04, 2:34, and 3:04. I think Don T asked about this. My apologies if you posted something that I missed.

  259. Irthe turner says:

    @Rob, Am hopeful the data will eventually unravel this mystery. If one were to speculate you have to put yourself in the PIC shoes and think through what he would do and why. The ultimate end goal, disapearing forever, is key imo. I agree with you that eliminating the main threat from the get go is critical to the success of said end goal.

  260. Paul Smithson says:

    @sk999. re wind model error magnitude. A couple of months ago I used windytv.com which offers GFS and ECMWF to look at difference in (current) winds 250HPA every 1 degree of latitude along a path from NW Sumatra to 7th arc at 38S. Findings: neither systematically lower/higher than other. RMS difference on wind strength 8.0 kts, maximum difference 16kts. RMS difference on wind direction 15.3 degrees. No correlation between magnitude of difference between models and wind strength.

  261. ROB says:

    @Irethe Turner

    Yes, my approach has been exactly the same as yours, ie. to try and see the situation from the pilot’s point of view and ask the question “if I wanted to do this, how would I go about it”. Needless to say, the plan demanded getting as far into the SIO as possible, as efficiently as possible, which rules out any loitering in the area west of Sumatra. Every minute spent loitering, would increase the chance of being identified.

    Personally, I am confident he was aiming for the area centered on S38, E89 (the Bayesian Hotspot) because this flight path turns out to be synchronized with a low sun angle at flameout, in such a precise way as to rule out coincidence. The big question at the moment is “did he reach his intended terminal area, of did things go seriously wrong following the FMT, and the aircraft ended up on constant heading/track mode, and ended up closer to S35. The BTO favours S38, but the BFO is more persuaded toward S35.

    this

  262. buyerninety says:

    @VictorI said;
    “Even if the low gain antenna (LGA) is in use, which requires no beam steering,
    position information is required.”

    Wasn’t it determined long ago, (by Don or ALSM?) that the comms log indicated that it
    was the High Gain Antenna(s) that was/were in use?

  263. ROB says:

    @Mick Gilbert

    Ok, we have to address the question why (apparently) to cellphone messages got out during the flight, apart from on reported message while still on the ground. This in itself if a highly unusual situation, even for a redeye flight taking off after midnight.

    It’s a sad fact, but we will probably never know exactly what went on in the passenger cabin as this flight unfolded,but as the evidence stands at present, no messages got out after takeoff. The FO’s phone made a fleeting contact with the outside world as the plane skirted round Penang Island, but there was no message sent. We will never know if the FO was trying to contact the ground at that time, or if he was already dead by then. He would have switched his phone off prior to takeoff. My guess is that as soon as he realized what was happening, ie the Captain had locked him out of the cockpit, was depressurizing the plane and diverting it from it’s planned course, he being second in command would deem himself responsible for trying to make contact with the outside world. No other cellphones made contact as the plane passed Penang. What does that tell you? It tells me that the passenger and cabin crew were in no fit state to engage in the luxury of turning their phones on.

  264. ROB says:

    @Mick Gilbert

    That should read why No cellphone messages got out.

  265. IrntTurner says:

    Probably because cell towers are scarce in Taman Negara & Tanjung Mentung. I imagine Celcomm
    found the rate of mobile phone uptake by Lar Gibbons to be particularly disappointing.

    Mick, r u hearing a dogwhistle?

  266. Mick Gilbert says:

    @ROB

    Re: “we have to address the question why (apparently) [no] cellphone messages got out during the flight, apart from on reported message while still on the ground. This in itself if a highly unusual situation, even for a redeye flight taking off after midnight.

    Why do you consider it “highly unusual” that no cellphone messages got out during the flight? Given the restrictions on the use of portable electronic devices from doors close, the time of departure and the time at the destination, I would expect that no cellphone activity would be normal on the outbound leg.

    On the leg back to Penang, getting a cellphone connection would be problematic at the best of times. As some other contributors have observed, the banked right turn south of Penang opens up the line of sight from the right side of the airplane to the ground making that the most likely place and time to get a connection. And that’s what happened for one cellphone that might reasonably be expected to have been on the right side of the airplane.

    He would have switched his phone off prior to takeoff.

    We have no way of knowing that. Ask flight crew how many times they have forgotten to turn their cellphones off prior to departure; it’s not uncommon.

    No other cellphones made contact as the plane passed Penang. What does that tell you? It tells me that the passenger and cabin crew were in no fit state to engage in the luxury of turning their phones on.

    It tells me the pretty much the same thing. However, it tells me absolutely nothing about how or why that situation had come about.

    If you’re trying to look at this through the eyes of a malicious perpetrator, and if getting as far into the SIO as possible was important, why would you deliberately fly the airplane all the way down to Penang, with the increased risk of detection or someone making contact with a cellphone, and then back up the Straits of Malacca? Why wouldn’t you save time, fuel and greatly reduce the risk of detection by cutting straight across Malaysian airspace from near LOSLO/BIDMO direct to MEKAR?

    Re: “I am confident he was aiming for the area centered on S38, E89 (the Bayesian Hotspot) because this flight path turns out to be synchronized with a low sun angle at flameout, in such a precise way as to rule out coincidence.

    I don’t know what level of precision you’re talking about such that it rules out coincidence because there is only seven minutes difference for sunrise across the entire 2,000 kilometre expanse from one end of the possible impact zone – 27°S 100°E – to the other – 40°S 85°E.

  267. Victor Iannello says:

    @Rob said: Ok, we have to address the question why (apparently) [no] cellphone messages got out during the flight, apart from [one] reported message while still on the ground.

    We don’t know whether the cell phones of passengers connected to any cell towers. The RMP report is silent on this. The FI didn’t even mention the cell phone connect of the First Officer. (Why is still a mystery.) I have reason to believe that the Malaysian authorities never collected the cell phone numbers of passengers, which would be necessary to check for any cell phone registrations during the flight.

  268. Victor Iannello says:

    @buyerninety: During the flight, yes.

  269. Mick Gilbert says:

    The MH150/three crew member discussion is interesting but seems to be based on at least one false premise. While MH150, being over 8 hours but not more than 12 hours, would most assuredly require three flight crew, there seems to be an assumption that the third flight crew member would be on the flight deck; that is not the case.

    The purpose of carrying extra flight crew on longer flights is to ensure that the limitation on the maximum time (viz 8 hours) that two pilots can be in the seat on a single sector is not exceeded. In order to accomplish that there is a requirement that “there will be for the crew member resting a comfortable reclining seat, or bunk“, that is a seat in business class or in the crew rest area. Accordingly, securing the flight deck on MH150 would be no more difficult than asking the First Officer to go and fetch his relief from business class and then latching the flight deck door behind him; a process that is certainly less unusual and probably somewhat less affronting than asking him to play waitress and go and fetch a coffee.

    Even with an additional crew member on board, MH150 presents itself as an ideal flight to deliberately divert if mysterious murder-suicide ending with an undetectable impact in the Southern Indian Ocean was the plan. That is because MH150 flies through a radar and VHF blind spot over the Bay of Bengal; it starts on N571 about 50 nm past BIKEN and then extends for about 200 nm (the blind spot is actually shaped like a corridor roughly 150-200 nm wide running north-south). Quite coincidentally, if not conveniently for a malicious perpetrator, MH150 reaches that blind spot about 2 hours after take-off which is precisely when the first relief crew change would typically take place. It presents as an ideal opportunity for securing the flight deck and diverting the airplane at a time when there is no radar or VHF radio coverage. Furthermore, by turning to roughly 178° from around the middle of the blind spot, with the fuel onboard, you could fly straight down the corridor completely unobserved by radar and out of VHF range all the way to somewhere near 50°S 91°E.

  270. ROB says:

    @IrntTurner

    You said “Mick, ru hearing a dog whistle?”

    Woof! Woof!

  271. TBill says:

    @ROB
    “The BTO favours S38, but the BFO is more persuaded toward S35.”

    The path is really the only part of your argument I have trouble with. BTO only favors 38S based on ATSB-like assumptions of a quick FMT near Indonesia and straight level flight. I feel maneuvers are apparent up to at least a little after 19:41 and seems to me a descent and slow down after Arc5 to reduce visual profile as sunrise approaches.

  272. ROB says:

    @Mick Gilbert

    Mick, 7 minutes difference across the search area? Do me a favour mate. 1 hour would be more like it. The morning terminator was advancing across the ocean at latitude S37 at the rate of 13.5 miles/minute, so would have advanced westwards by 95 statute miles in 7 minutes.

    If you had actually taken the trouble to check it out (like I did) you would find the flight path with 186deg heading, the IGOGU/ISBIX/S38, E89 flight path ran almost exactly parallel to the advancing morning terminator, diverging from exactly parallel by less than 1deg during the final 2.5 hours of the flight. This alignment meant that at 00:19, the Sun was about 4deg above the horizon at sea level, at all points along the flight path, which meant the pilot only had to wait until fuel exhaustion to get the right lighting conditions for a ditching. He didn’t have to bother about adjusting the cruising speed as he wasn’t trying to hit a particular spot, he only had to wait for fuel exhaustion, and “bingo” (if you pardon the pun)

    With this flight path he would be in darkness for all but the last half hour or so, just enough time to make sure he wouldn’t be coming down next to an ore carrier, if they ever get down that far. No other flight path had this lighting arrangement while at the same time realized the range potential afforded by the particular fuel load. If you try to approach the sunrise zone from either the east or the west, it would be like trying to hit a fast moving target (moving at 13.5miles/min, and you would never do it.

  273. ROB says:

    @Victor

    The reason why the RMP only mention the FO phone contact was because it must have been the only one, in my opinion. If thee had been others, we would have heard something by now.

  274. ROB says:

    @Mick Gilbert

    Did you spot the deliberate mistake!? Who would ever have wanted to approach from the east? They would have been in daylight for far to long. He could have manoeuvred to approach from the east, but he would have had trouble hitting the spot at the same time as he ran out of fuel, and when you manoeuvre you unavoidably rate into the range potential.

  275. ROB says:

    @Mick

    My typing is atrocious. He could have manoeuvred to approach from the west, not from the east, but it would mean he didn’t get as far into the SIO, and he would have trouble hitting the spot with the right lighting conditions.

    I’m now going to give this a break, and take a walk in the country to clear my head. 😎

  276. Victor Iannello says:

    @Rob: If they didn’t collect passenger cell phone numbers from the NOK, the authorities would not know whether other cell phones registered on cell towers.

  277. DennisW says:

    @Victor

    It is also not clear how many of the Chinese PAX phones would operate on GSM or LTE. My guess is few to none. Most of the China telecom operators use some variation of CDMA. The PAX would need to rent phones in KL.

  278. ventus45 says:

    @Rob

    If you are interested in the terminator positions, at various altitudes, and how to plan to track parallel to the advancing terminator, and how to descend unseen until the last moment, I laid it all out, over 17 months ago now.

    See here: http://www.auntypru.com/forum/showthread.php?tid=28&pid=3569#pid3569
    and here: http://www.auntypru.com/forum/showthread.php?tid=28&pid=3585#pid3585
    and here: http://www.auntypru.com/forum/showthread.php?tid=28&pid=3656#pid3656

  279. lkr says:

    @ROB Don’t forget that 38S definitely causes problems with drift patterns [specifically W Australia], and 35S isn’t much better. There really isn’t a single solution that fits everything we know [or think we know].

  280. Victor Iannello says:

    @DennisW: Yes, I agree that a large portion of Chinese cell phones would be incompatible with GSM networks in Malaysia.

  281. Mick Rooney says:

    @Victor
    “We don’t know whether the cell phones of passengers connected to any cell towers. The RMP report is silent on this. The FI didn’t even mention the cell phone connect of the First Officer. (Why is still a mystery.) I have reason to believe that the Malaysian authorities never collected the cell phone numbers of passengers, which would be necessary to check for any cell phone registrations during the flight.”

    @Dennis
    “It is also not clear how many of the Chinese PAX phones would operate on GSM or LTE. My guess is few to none. Most of the China telecom operators use some variation of CDMA. The PAX would need to rent phones in KL.”

    Excellent points Victor and Dennis. Though, I will point out we only (at least me) have access to the RMP folders leaked to me and others. I don’t know what the PAX folder contains and whether detailed assessment was made of their cell phones. The released network data, analysis from several providers and tests carried out, *might* indicate that no effort was made to subject the cell phones of PX in the way cockpit and cabin crew were. Also, the RMP report remains a snapshot in time – 2014. I don’t know if the RMP, post 2014, pursued their criminal investigation into the period 2015-17.

    I tend to go with Victor’s assessment that they likely did not and we would have heard something by now. Remember, back then – 2014 – it was clear the focus was on the cap and FO.

    But the point Dennis makes is equally cogent – that even if it was (RE: GSM or LTE), the chances of a positive ping/cell phone detection significantly decreases.

    All that said, we did have PAX in the cabin who were business people, clearly technically-minded, and must have been well-travelled.

    But I hear the echoes of Jeff Wise’s comment – when I recently interviewed him for radio – that the (RMP) criminal investigation may have been somewhat “half-assed.”

  282. Victor Iannello says:

    @Mick Rooney: “Half-assed” is an understatement. The RMP report is shoddy, incomplete, and raises many more questions than it answers. The description of the cell phone testing is just one of many examples of this. However, the report does present evidence that is not documented elsewhere, such as the contents of the flight simulator files, and the record of the cell phone connect.

  283. Paul Smithson says:

    @Dr B. Thanks for your feedback on fuel. I look forward to hearing if there is any tweak once you have looked at the MH371 numbers more closely.

  284. Don Thompson says:

    Re cell phones

    I expect that the reality is that MY cellular operators do not routinely log the basic device registrations for service on their networks. Such data would present little, or no, value. Legislation in the UK, and I believe EU wide, requires mobile call data records (CDR) to be maintained for 18mths, that is, records of call history. MY’s legislation is unlikely to be different, in terms of the scope for data retention, if any exists at all. The cost of implementing a unique regime would be the issue, not the ability to sort/retain.

    My recollection of the RMP report details concerning IP connections from smartphones to social media accounts was that investigation was mostly inconclusive as the operators didn’t record IP address leases to specific devices. While the socmedia services logged inbound IP addresses, the connecting ‘link’ to device ID was absent.

  285. Victor Iannello says:

    @Don Thompson said: I expect that the reality is that MY cellular operators do not routinely log the basic device registrations for service on their networks.

    How do you explain the registration of the First Officer’s cell phone at Penang? And if this was event was recorded, isn’t it possible that other connections from aboard the plane were also recorded?

  286. Oleksandr says:

    @Andrew,

    Re: “However, I believe you’re considering a scenario where the ADIRU fails? If that occurs while LNAV is engaged and the aircraft is established in a turn, then the AFDS remains engaged in an “attitude stabilising mode”. That condition would be annunciated on the PFD by an amber line drawn through the LNAV mode annunciation. If the condition persists, my understanding is that the aircraft would continue in the “attitude stabilising mode” with a constant bank angle. ATT mode would not be annunciated on the PFD in that scenario.”

    Yes, I consider the ATT mode active under the failed ADIRU assumption, as otherwise the ATT would only be a transient mode. I have a long lasting interest in the ATT mode because there are trajectories confirming BTO and BFO data. And given that these trajectories are subjected to wind and Coriolis, and in addition they are consistent with the drift studies, I believe this is a remarkable coincidence. Note that the impact of the Coriolis is equivalent to flying in the constant gyroscopic heading mode – this what SAARU provides.

    Re: “It has already been stated, several times, that the SDU receives inertial position data from the ADIRU for beam steering. The SDU does NOT receive data from the SAARU.”

    Could you please remind where it was explicitly stated that the SDU receives position data from the ADIRU only? The SAARU does not output position data, but GPSs do provide, and I have not seen explicit statements that the SDU cannot use GPS data when the ADIRU fails. FCOM (11.10.10), for example, states that a pilot can select altitude source for the transponder: ADIRU or SAARU. If I am not mistaken in the diagrams and descriptions you posted it was never mentioned that the SAARU data are supplied to the transponders. The absence of evidence is not the evidence of absence (D.S.).

    Re: “Yes, the SDU receives date/time data from the L AIMS cabinet via the GP 1 data bus. The AIMS cabinets receive date/time data from the multi-mode receivers (ie GPS). The GP 1 data bus does not transmit position data.”

    Again, where was it explicitly stated? Absence of the position data in the description lists does not mean that such a data is not supplied via GP channel. Especially in case when ADIRU data is not available. From my point of view it would be logical for reliability reasons at no additional cost.

    Re: “The position data is used by the SDU for beam steering. I’m not sure exactly what happens if that data is lost, but if the SATCOM doesn’t fail immediately, I assume it will eventually ‘lose’ the satellite as the aircraft position changes.”

    I have not seen any proper study. If the SDU can keep in its memory the last position, in-principle it could use it. The error in steering angle would be negligible over several minutes of flight if not hours. This is in contrast to the attitude data. I don’t know how much is the tolerance error in the steering angle: perhaps somebody can comment on this topic.

  287. Mick Gilbert says:

    @ROB

    You’re quite correct, it’s an hour and 7 minutes from one end of the zone to the other. My mistake, I had missed the time zone change, my apologies.

    Isn’t the problem with this supposedly precision flight path along the terminator that when the impact point was searched thoroughly, the airplane wasn’t there?

  288. Don Thompson says:

    @Victor,

    The interaction with the FO’s phone involved the Location Based Service, it’s my understanding this service is involved in the process that provides cellular smartphone subscribers with useful supplementary information such as the nearest coffee/pizza/donut vendor when they execute a search on their device.

    Don

  289. Don Thompson says:

    @Oleksandr,

    “The error in steering angle would be negligible over several minutes of flight if not hours.”

    Seriously, no, not negligible.

    The expanded GES log, the subject of the post to which we are commenting, illustrates how sensitive the AES is to accurate tracking of a satellite.

  290. Oleksandr says:

    @Don Thompson,

    “Seriously, no, not negligible.”

    I think you are confusing contributions of the attitude and position, and respectively sensitivity to these two factors. For example, recall that the satellite also moves: displacement along Z-axis (in ECEF) ranged from 390 to 1206 km according to Tab. 3 from ATSB June 2014 report. Did this cause significant impact on the ability to transmit/receive signal due to the error in antennae steering? The answer is obvious – no. Similarly, the same impact (or its absence) could be expected in the case of airplane’s displacement of order 800 km, which is equivalent of 1 hour of flight.

  291. Donald says:

    @All

    Come on guys. As problematic as the RMP ‘report’ is, I’m a wee bit incredulous that we would not by now heard heard about any other pax or crew cell phone connects. Surely if they had occurred, word of such an occurrence would have leaked from NOK. And surely other investigative agencies have done their due diligence on the matter.

    There is clearly a significance to only the FO’s cell phone making a connect…the most logical being a desperate attempt to contact someone after being locked out of the cockpit. And he was 100% locked out as he posed the greatest risk to Z…and I think it’s safe to rule out a physical assault or pick your poison, as anything other than a locked out FO would be far to risky in regard to neutralizing the FO.

    Let’s not kid ourselves about the import of the cell connect, the lack of any other pax or crew connects, and the story this tells.

  292. Irthe turner says:

    Mick@, In a few hours, I will be on a 10 hour flight. Behaviour of a 3 person flightdeck crew is highly unpredictable. They have their own resting area which is definitely not in BC on a 777. Often all 3 are in the cockpit until well in the flight. They hang around the galleys and rotate a few hours rest when needed . The point being , you can’t plan on the unknown with 3 pilots on board.

  293. Mick Gilbert says:

    @Irthe turner

    Malaysia Airlines’ Ops is quite specific as to how the Captain is to organise crew resources and manage the timing of crew changes.

    When a set of flight crew consisting a relief Captain is required for a flight, the hours for the flight will be shared equally. For example, in a 9 hours flight, the hours will be shared as follows … :
    First 2 Hours – Capt A & Co-Pilot
    Next 3 Hours – Capt A & Capt B
    Next 3 Hours – Capt B & Co-Pilot
    Last 1 Hour
    – Capt A & Co-Pilot

    MH150 is a nine hour flight. It can’t be much clearer than that.

    Whether the relief crew sit in Business Class or the crew rest area would largely depend on the configuration of the airplane (not all have a crew rest station) but it would ultimately be at the Captain’s discretion whether the relief crew member is ensconced on the flight deck.

  294. Ge Rijn says:

    @VictorI @ ALSM

    On the ‘poorly conducted flight-tests’ regarding the FO cellphone connection. I think you might be right for those test were conducted on (far) lower speeds and altitudes than MH370 was supposed to fly on.
    But then; is it not the more suprising BBFARLIM2 did not detect or connect any cellphone during those tests?
    Chances would be far better than with a plane passing by at ~35.000ft/500knots.
    Still it did not happen even on far lower altitudes and speeds in the tests.
    But the FO-cellphone on MH370 did anyway.

    Imo this indicates MH370 must at least have flown on a lower altitude and probably also a more northern route passing Penang over land to make the detection of BBFARLIM2 possible.

  295. Andrew says:

    @Oleksandr

    RE: “The SAARU does not output position data, but GPSs do provide, and I have not seen explicit statements that the SDU cannot use GPS data when the ADIRU fails.”

    You won’t find such a statement. I have already provided you with a diagram that shows all the systems that interface with the SDU. The GPS is not depicted. Do you not think that such an interface would be shown or described somewhere in the text, if it existed?

    RE: “FCOM (11.10.10), for example, states that a pilot can select altitude source for the transponder: ADIRU or SAARU. If I am not mistaken in the diagrams and descriptions you posted it was never mentioned that the SAARU data are supplied to the transponders.”

    If I remember correctly, the previous discussion simply stated that SAARU data is output on the centre flight controls ARINC 629 data bus. It did not consider all the systems that might or might not use that data. Further, the discussion had nothing whatsoever to do with the transponder, so why would it have been mentioned? If you’d asked, I could have given you references from both the FCOM and the AMM that show the transponder can use SAARU data as an alternate source for altitude reporting, if selected by the pilots. The same does NOT apply to the SATCOM.

    RE: “Again, where was it explicitly stated? Absence of the position data in the description lists does not mean that such a data is not supplied via GP channel. Especially in case when ADIRU data is not available. From my point of view it would be logical for reliability reasons at no additional cost.”

    Again, you won’t find a statement that says “the GP 1 data bus does not transmit position data”. The manuals describe the type of data that is transmitted to the aircraft systems via the various data buses. They don’t explicitly state the data that is not transmitted. Again, do you not think that GPS data would be mentioned in the manuals if it were sent to the SDU via the GP 1 data bus?

    Consider what happens when the aircraft is on the ground with the ADIRU selected OFF. According to your theory, the SDU would be provided with position data from the GPS and other inertial data would be provided by the SAARU, correct? So SATCOM voice communications using the HGA should be possible on the ground when the ADIRU is selected OFF, correct? I can assure you, based on many years experience of operating the aircraft, that is NOT the case. SATCOM voice communications are NOT available when the aircraft is on the ground with the ADIRU selected OFF. Why?

  296. Mick Gilbert says:

    @Donald

    Re: “There is clearly a significance to only the FO’s cell phone making a connect…the most logical being a desperate attempt to contact someone after being locked out of the cockpit.

    How is that “the most logical” interpretation? It might be an interpretation that fits your preferred narrative but it is hardly “logical”. As DrB has observed the banked right turn south of Penang opens up the line of sight from the right side of the airplane to the ground making that the most likely place and time to get a connection, particularly from a phone that was located on the right side of the airplane. And it appears that’s what happened for one cellphone, the First Officer’s, that might reasonably be expected to have been on the right side of the airplane.

    Re: “And he was 100% locked out as he posed the greatest risk to Z…and I think it’s safe to rule out a physical assault or pick your poison, as anything other than a locked out FO would be far to risky in regard to neutralizing the FO.

    That doesn’t make much sense. If the First Officer was the greatest risk, why would you have him somewhere beyond your control? Locked out of the flight deck the First Officer would have been free to alert the cabin crew, try to alert the authorities or, worse still for a malicious perpetrator, access the Main Equipment Centre and start disabling critical flight deck support systems such as the flight crew oxygen system. Moreover, if you did determine that you were going to lock the First Officer out, why would you then bring the airplane with in cellphone range of Penang?

    Re: “Let’s not kid ourselves about the import of the cell connect, the lack of any other pax or crew connects, and the story this tells.

    Well, what story exactly does it tell? If the First Officer was locked out of the flight deck yet still able to use his phone as the airplane rounded Penang why weren’t any members of the extraordinarily experienced cabin crew (the average length of service was over twenty years, the Inflight Supervisor had 35 years experience and even the most junior member had 13 years experience) able to do the same? Had the First Officer been locked out as you’ve suggested, he would have been able to coordinate with and organise the cabin crew and you would have expected to see multiple cellphone registrations not just one.

    I think that a single cellphone registration from a phone that you’d expect to be on the right side of the airplane as it executed a banked right turn does tell a story; a story of happenstance (in that the First Officer had left his phone on) and his possible incapacitation while still in his seat.

  297. Victor Iannello says:

    Ge Rijn said: But then; is it not the more suprising BBFARLIM2 did not detect or connect any cellphone during those tests?
    Chances would be far better than with a plane passing by at ~35.000ft/500knots.

    Not necessarily. It depends on the radiation pattern of the cell tower antenna, and where the lobes are at various elevation angles. Because of the lobes, a higher elevation angle does not necessarily translate to a lower gain. As for speed, if the plane is traveling tangentially to the antenna, the line-of-sight speed is much lower than the groundspeed.

Leave a Reply