More secret MH370 documents released

Mick Rooney, aka @Airinvestigate, has released further documents from the secret Royal Malaysian Police investigation into the disappearance of MH370. I asked him if he could tell me anything about how the documents were sourced or why they are being released now, but he says that he is bound by an oath of confidentiality not to discuss further.

Those familiar with recent events surrounding the case might be able to hazard a guess.

Here are the files:

data-from-flight-simulator-computer

This 14-page document includes technical information about the data found on Zaharie’s flight simulator hard drives. It appears that the machine crashed multiple times in the months before MH370’s disappearance. The document also includes a log of when the flight sim was played, the last time being on March 15, 2014, a week after the plane disappeared (presumably this reflects activity by investigators.) Prior to that, the sim had last been played on February 20, two weeks before the disappearance. This suggests that Zaharie was not using his flight simulator to practice vanishing in the weeks before his disappearance.

data-from-prelim-exam-report-translated-from-malay

This 7-page document seems to have been machine-translated from Malay, and appears to describe a preliminary investigation of the computer hard drives by a Malaysian police technician. It lists the various hard drives found with the flight-sim computer. Among the information recovered were passwords and account information for Zaharie’s hobbies and interests, as well as information about an online bookstore, Zaharie’s various social media accounts, and online shopping. Of particular note, investigators found a deleted folder labeled “777TwinTower” which contains pictures of a Malaysia Airlines plane flying toward the Kuala Lumpur city center. Given widely held suspicion that Zaharie took MH370 on a suicide flight, and that fact that terrorists flew two planes into New York’s twin towers in 2001, this will no doubt raise eyebrows. However, this document notes that: “These images have been taken from the computer screen to play a simulated airplane. The assessment believed that the owners of these computers have taken one of those images for the purpose of being used as an icon on the account.” That is to say, an innocent interpretation of this folder and its contents would be that Zaharie, a proud Malaysian 777 pilot, wanted to create an image of his plane flying past an iconic Malaysian landmark.

After a section discussing the seven deleted points from the flight simulator, which have been much discussed in this forum, the report concludes with a brief Summary: “The results of the examination of the goods were found that no any activity outside the common. The overall computer use to host gaming Flight Simulator only. Nor has any information source which directly indicates there any plans to eliminate MH370 found.”

sim-data

This 31-page document appears to contain all of the saved data in the seven above-mentioned flight simulator points. Hopefully independent flight simulator experts will look it over and render an opinion for the rest of us who lack the expertise to properly grapple with it.

Overview

How does this new information alter our understanding of the MH370 mystery?

For me, it is noteworthy that so little incriminating information was found on any of Zaharie’s computers, even (especially) among the deleted files. The way we use computers these days, they are essentially extensions of our brains. Any passing fancy that drifts through our head is likely to be reflected in our internet search history, in notes we write to ourselves, and so on. When Andreas Lübitz was in the throes of his final mental dissolution, he spent a great deal of time online reading about mental disorders and researching ways to commit suicide. It’s all right there to be seen. Yet on Zaharie’s computer there is nothing. Indeed, he seems to have been spending his time prior to the disappearance doing things like making instructional DIY home-repair videos and pretending to fly an antique DC-3 airplane. Not, it would seem, the behavior of someone contemplating his imminent extinction.

In the light of this newly released information, it is easier to understand why the Malaysian police came to the conclusion that nothing about Zaharie’s behavior points to him  being the culprit.

384 thoughts on “More secret MH370 documents released”

  1. All,

    I have prepared a new study that should have been done ages ago – listing and intercomparing all the available datasets regarding the flight path of MH370 from takeoff to last radar contact. Others have done similar studies in the past comparing specific datasets, but here I try to combine everything in a consistent fashion. A lot of digitization of figures was needed. I include links to all the data files used in the study.

    As usual, here is a link to an index of all my reports; the latest one it at the top.

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

  2. These few leaked pages allready contain some quite compelling data and disturbing information that Z. could well have been the culprit from the start (which was also declared early on as he being declared the prime suspect).

    What else does the 1000 pages report shows in this regard?
    Is this what everyone tries to avoid publishing? Is it all too obvious?
    Are the owners of a copy scared of legal steps against them if they bring it all to the light? Who do they protect? Themselfs or others?
    Is this the prime cover up from MY?
    The disgrace and financial consequences that one of their own people and pilots was responsible?

    The report should be published completely.
    Not to judge about a culprit but to help find the plane on behalve of the NoK and the truth.
    That should stay the ultimate goal.

    BTW the ‘debris-storm’ is over for quite a while now, which could tell something too; not a big debris field but a small one.

  3. @sk999:

    Thanks for the link to the index of your MH370 reports. I particularly liked your work on ‘RADAR’ with the comparisons of all available sources.

  4. @Gysbreght

    –“Therefore all that can be said about the three sets of [SimVars.0] and [Engine Parameters.n.0] data is that they were saved from simulator sessions that started from conditions that were derived from the same ancestor *.FLT file.”

    Wow.

    On the one hand, we have Occam’s razor. A simple explanation derived from evidence which requires very few assumptions, if any, in order to reach a conclusion:

    – 7 .flt files…DELETED…from MK25
    – three files each share a PAIR of identical parameters
    – those same three files roughly suggest the flight of MH370
    – file “3” shows a 20 degree left bank and normal turn-rate passing through 255 HDG (a virtual FMT performed by the A/P at 20 deg limit)
    – there are ZERO airways headed south anywhere in the vicinity of 10N90E (file “3”)
    – files “4” and “5” show a zero-fuel descent down to low altitude over a remote part of the SIO
    – MK25 – and all of the DELETED files discovered on it – belonged to a guy who disappeared on a 777 which flew from the vicinity of sim point “3” to the SIO, roughly on the way to sim points “4” and “5”

    On the other hand, we have you. Many necessary assumptions, a complex, labyrinthine set of conditions, what-if-this and what-if-that, and a fetish for extreme coincidence.

    Your first instinct is to question the nature of the leak itself: “Why is Florence de Changy…?” Then, after you dig around for awhile, you publish a very narrow and low-probability scenario which, under a very specific set of assumptions, would invalidate any connection between the three files.

    Congratulations. You are the anti-Occam. Let’s see how you handle Victor’s paper next week.

    In the meantime, If I hijack a heavy jet and disappear with hundreds of innocent people, I’ll politely ask you NOT to defend me.

  5. Nice work @sk999. Are you going to put this up on Reddit too? Might be useful for feedback and discussion.

  6. @Matt Moriarty, you wrote:
    “– files “4” and “5” show a zero-fuel descent down to low altitude over a remote part of the SIO”

    Did you mean “zero-fuel” climb?

  7. @Gysbreght

    That’s strange because what your guy says in that article exactly matches what I just wrote:

    “Point 3 is the important difference from the RMP report. The heading of that is 255 and NOT 104 which would actually take it back towards Malaysia.”

    You don’t think I noticed how .flt shows HDG as either + or – from 0? From the sim data at point 3, you see a heading of -104.53 and a HVelWorld showing -0.0154693xxx rad/sec. Converted to deg, this is a turn rate of 0.89 deg/sec to the LEFT.

    That is a snapshot in time showing a momentary heading of 255 while the plane is turning to the left. Exactly what I wrote above.

    Your boy also writes: “That means that there is a 7th point not published & I can only assume that’s because it didn’t suit the agenda of the leakers & more than likely indicates there is NOTHING sinister in Capt. Shah’s simulator.”

    Well, they’ve all been published now and there’s actually 8 of them. Three separate points on rwy 32R at WMKK and five in the air.

    And your point is?

  8. @Matt Moriarty,

    You previously said: “It is known far and wide that the endurance shown by the timing of the 7th arc coincides perfectly with the book endurance of a normal LRC flight of an intact 777 that weight and fuel load.”

    In response, I note that Mark twain once said “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.”

    I think this quote captures perfectly what may be going on in the case of MH370 fuel consumption. I hope you and other pilots and investigators might be in a position to evaluate my thesis.

    Simply put, my detailed fuel calculations appear to demonstrate that the 7th Arc is actually NOT reachable in the center of the ATSB’s Current Search Area (at ~37S). It all comes down to a missing footnote.

    Let me demonstrate my thesis using the following figure based on my calculations:

    https://drive.google.com/file/d/0BzOIIFNlx2aUcERONkFuNkt0ZGM/view?usp=sharing

    This figure is a plot of (average-engine) Performance Degradation Allowance (PDA) versus post-FMT flight level. The two red dashed lines are the approximate limits of PDAs for used engines (in this case Rolls Royce Trent 892) with PDAs ranging from approximately +1.5% to +3.5%. My route fitter adjusts the PDA so that second-engine fuel exhaustion occurs at 00:17:29 (per ATSB), so all the points shown in this plot embody a constant endurance. Acceptable route solutions must have a PDA between the two red dashed vertical lines. Smaller PDAs imply insufficient fuel to match the known endurance. Larger PDAs imply excess fuel and expected endurance beyond 00:17. There are numerous details concerning the fuel model that I cannot describe here, but I am willing to share them with anyone who is interested.

    The problem I have been addressing for quite some time is to identify routes that have acceptable PDA and also match the expected RMS residual error values for BTO (~33 microseconds), BFO (3-4.5 Hz), air speed / tailwind (< ~5 kts for all routes), and crosswind (< ~5 kts for heading routes only). The acceptable tailwind and crosswind errors are based on estimates of the uncertainty in the GDAS wind data (which are derived from comparisons of direct, simultaneous windspeed measurements with GDAS data).

    I have been exploring post-FMT combinations of speeds [Long-Range Cruise, ECON 52 (the flight plan), Maximum-Range Cruise, and Holding (maximum endurance) ] and lateral navigation methods (Great Circle, True Track, Magnetic Track, True Heading, and Magnetic Heading).

    The five nearly vertical curved lines in the plot represent different speed modes and conditions, and all are for the same True Track route ending at 37.6S, which is at the center of the ATSB’s +/- 40 NM wide search area.

    Start with the black dashed curve near the center of the plot, which is LRC using the Boeing fuel flow table for the 9M-MRO aircraft (B777-200ER) with the Rolls Royce Trent 892 engines. As you said, Matt, this curve indeed falls inside the acceptable PDA window for flight levels from 300 to about 370. The problem is, this curve is (probably) VALID ONLY AT STANDARD ISA TEMPERATURES. I say “probably” because there are no specific temperature notes on this particular Boeing table of LRC fuel flows and speeds (at least for the version I have). However, the analogous Boeing LRC table for the GE90-94B engines on B777-200ERs does include several notes on temperatures. A note in the table header says “STD TAT” meaning Standard (i.e., ISA = International Standard Atmosphere) Total Air Temperature (TAT), so we know in this case the fuel flow table is for ISA temperatures only. An important footnote on this same table says “INCREASE/DECREASE FUEL FLOW 3% PER 10 C ABOVE/BELOW STANDARD TAT”. The reason this temperature dependence of fuel flow is critical to MH370 is because the air temperatures at cruising altitudes on this night (7-8 March 2014) were elevated above the ISA temperatures by an average of about +11C. The 3%/10C TAT is equivalent to 3.42%/10C SAT (Static Air Temperature = Outside Air Temperature), so +11C would increase fuel flows by 3.7%. This is a large effect and, if this same effect is also applied to the Trent fuel flow table, it reduces the range and endurance by a little more than 100 NM and 15 minutes. It also reduces the PDAs (i.e., the engines must be 3.7% more fuel efficient) needed to achieve the known endurance compared to the fuel flows which are uncompensated for non-standard temperatures.

    I believe the temperature effect on engine fuel flows applies to all turbofan engines, including the Rolls Royce Trent 892s used on 9M-MRO. Here is a heuristic argument in support of this. For an absolute Static Air Temperature (=SAT) that is different from the ISA temperature (=ISAT), the local sound speed varies with SQRT(SAT/ISAT). Similarly, since the True Air Speed (TAS) for ECON speeds and Holding are actually commanded as Mach numbers (or KIAS), the TAS will also vary with the square root of the Static Air Temperature. We also know that the fuel flow varies with Mach number and therefore with TAS. From the Boeing B777 fuel mileage graph, from MRC to LRC the speed increases by 1.7% but the fuel flow increases by 2.4%. So, roughly speaking, the fuel flow increases about 2.4%/1.7% = 1.4% for every 1% increase in speed. An 11C increase in SAT on that night increases LSS and TAS by SQRT[ (273-43) / (273-54) ] = 1.025 = 2.5%. Multiplying this 2.5% speed increase by this factor of 1.4 (%FF/%TAS) results in a 3.5% increase in fuel flow. This is close to my calculated average of 3.7% from my fuel model. My conclusions are that the temperature dependence of the fuel flow is an inherent property of turbofan engines, and it applies to the Rolls Royce Trent 892s used on 9M-MRO (footnote or not).

    The assumed conditions Boeing used to compute the range limits shown by the ATSB in their report were not explained in detail, so I don’t know whether or not this temperature effect was included. I also don’t know the post-18:22 route limits (and possible FMT locations) they assumed. As a result, I cannot do any meaningful comparison between my “performance limit” results and Boeing’s calculations.

    The impact of the increased fuel flow at MH370’s elevated temperatures is shown by the solid black line in the figure, which is for LRC with the temperature compensation applied to the fuel flows. As expected, this solid black line is shifted 3.7% to the left (i.e., to lower PDAs = more efficient engines) compared to the dashed black line (uncompensated LRC). See the two black-filled circles at FL358 (with a terminus at 37.64S degrees, as shown in the labels). The one on the right has a PDA of +1.8%, which is acceptable. The one on the left (temperature compensated fuel flow) is at PDA = -1.9%, which is unacceptable. In this case (the solid black line), all LRC routes with used engines have insufficient endurance to reach the 7th Arc. The solid orange line is for ECON 52 (the flight plan), and the solid green line is for MRC (both with the temperature compensation applied). As expected, the fuel flows for these two cases are less than for LRC, and so they are to the right in the PDA plot (i.e., higher PDAs are allowed = less efficient engines are possible for the same 00:17 endurance).

    Note in the figure that when the fuel flow temperature compensation is applied, no LRC, ECON 52, or MRC route (i.e., the black, orange, and green solid lines, respectively) fall within the acceptable range of PDAs. That is why I have concluded that MH370’s fuel was insufficient to reach the 7th Arc for straight and fast routes (such as True Track and Great Circle). Looking at the four points plotted on or very near the solid black (LRC) line on the far left, you see the actual route fits for Great Circle (ending at 37.83S), True Track (ending at 37.64S), True Heading (ending at 35.23S), and Magnetic Track (ending at 34.32S). The numbers next to the symbols on the plot are the 00:19 (7th Arc) latitudes of the best-fit routes. This demonstrates that no LRC, ECON 52, or MRC route has an acceptable PDA, no matter what lateral navigation method is used. There is simply insufficient fuel on this particular night with typical used engines.

    The purple, blue, orange, and green dashed lines running up and to the right are the candidate route results for the four navigation methods: Great Circle, True Track, True Heading, and Magnetic Track (no candidate route has been identified for Magnetic Heading).

    The solid blue line on the right side of the plot is for Holding (= maximum endurance) speed with fuel flow temperature compensation applied. We can see that in this case (True Track with no descent) it intersects the PDA limits (the red dashed vertical lines) near FL370 to FL390, so there may be Holding speed routes that match the expected PDA range at these flight levels. Looking at the solid blue line (Holding with fuel flow temperature compensation) we see several route fits on or near it for various lateral navigation methods. The blue circle at the top of the plot is the True Track fit, and in this case the fuel is also insufficient to reach 37.6S (because its PDA is less than the acceptable range). This is not surprising because of the slower Holding speed. In other words, using Holding speed there is insufficent time to reach the main (+/- 40 NM wide) ATSB search area, and using MRC/LRC there is sufficient time but not sufficient fuel.
    The blue square on the right side of the plot is the Magnetic Track / Holding route ending near 34.2S. In this case the PDA is too high, meaning there is excess fuel. Additional fuel consumption would be required for this case to become a viable candidate, such as a major descent (costing 1-2 % in fuel), or assuming the APU were running and electrically loaded for most of the period after diversion (costing about 2% in fuel), or assuming excess drag, or possibly some combination of these effects. These effects may be possible but seem unlikely to have occurred.

    A lateral navigation mode which produces a route with a PDA in the acceptable range without extended APU operation or excess drag is True Heading using Holding speed. It is possible, although some may think otherwise, the default FMC mode upon reaching a route discontinuity is True Heading (not Magnetic Heading). I will note here that I have extensively searched for an acceptable Magnetic (NORM) Heading route without clear success. An acceptable True Heading route does exist, although it is not as precise in determining the terminus as the non-heading routes. The reason for this is rather simple – we just don’t know the wind along the route very accurately. That is why I have modified my route fitter so that tailwind and crosswind errors are allowed to perturb the route calculated using the GDAS tailwinds and crosswinds, so long as their RMS values are consistent with the GDAS wind errors (which turn out to be somewhat larger than I had expected – about 5 knots RMS). The impact of significant wind errors is imprecisely-determined intersections with the 7th Arc for heading routes. The heading route solutions tend to define a limited zone along the 7th Arc rather than a precise point. Even so, I think this True Heading result may prove to be useful in defining a revised, limited search area.

    The True Heading route at Holding speed is shown in the figure as a blue diamond. It has an acceptable PDA of 3.2% and ends near 35.26S. This route is actually quite simple. Shortly after the end of the radar track at 18:22 (at LRC and FL360), 9M-MRO makes a jog to the right of airway N571 by about 12 NM. It continues flying NW (at 296 degrees true) parallel to N571. This maneuver matches all the 18:25 – 18:28 BFOs and BTOs (with the single exception of the 273 Hz BFO which ATSB says to ignore). After the jog, at about 18:32 a small step climb was made from FL360 to about FL370, which is near the optimum for the current aircraft weight and speed. The next step is a left turn toward nearby ANOKO ending at 18:39:25, just before the 18:40 phone call begins. The new course is 185.5 degrees true. At the same time as the turn begins, the speed is reduced from LRC to Holding. Perhaps this turn and slow-down was the first step in the ANOKO Two Charlie standard arrival at WITT (Banda Aceh). All the necessary pilot inputs were made before 18:38. No further changes were made to the auto-pilot. When ANOKO was reached, the FMC defaulted to a True Heading of 183.6 degrees and continued this until fuel exhaustion. The terminus is at 35.26S 92.52E. The average TAS from 18:22 to 00:11 is 455 kts, and the air miles traveled from 17:07 to 00:17 is 3,293 NM (ESAD). I also note that this route passes only ~2 NM west of Kate Tee (near NOPEK). The altitude would have been ~FL370, the time of nearest approach would have been about 18:46, and the aircraft may have been visible between ~18:41 and ~18:51. If this route is correct, 9M-MRO would not have been the aircraft she saw at low altitude but rather the one she described as being at high altitude and moving north to south nearly overhead.

    To summarize, candidate route solutions do exist with acceptable PDAs, and, based on my work to date, they appear to be limited to Holding speed with a True Heading between flight levels approximately 370 – 390. They seem to fall within the ATSB’s narrower search area to the northeast of the +/- 40 NM area (this narrow area has apparently been searched out to +/- 20 NM in the lower portion and -5/+15 NM in the upper part). Clearly ATSB thinks there is a reasonable probability for these NE extensions or they would not have collected sonar data there. Perhaps the aircraft is in this region, just outside the limited width already scanned (and I would guess more likely inside the arc). If my fuel model calculations are correct, I think the simple reason the ATSB’s search has been unsuccessful within the +/- 40 NM wide search area is because there is an insufficient combination of fuel and time to reach it on this particular night.

  9. @DrBobbyUllich

    It’s a long piece and I won’t be able to get to it today, but I certainly will get to it.

    In the meantime, please know the quote you attribute to me was regarding endurance. Not range.

    Get back to you after I’ve read your post.

  10. @sk999

    Link to satellite position and velocity code is broken, so I wrote a quick and dirty Python script to implement the equations you have in satellite.pdf. The z values come out perfectly. Everything else is fubar. You sure the equations for x,y,vx,vy, and vz are correct?

    Sorry to be a pest. Just too lazy to go to the source material.

  11. @Jeff

    you wrote:

    “@DennisW, You wrote, “As a matter of policy I do not use BFO values with a missing or a bogus BTO value.” Lucky thing that your policy happens to jettison the data that nullifies your theory.”

    Jeez, Jeff. How long have I been blogging here? You know I don’t screw up. The 18:40 BFO value (even if correct) requires some arm waving to reconcile. I am just not in the right mindset to screw around with it. I’ll take Victor’s explanation, and leave it at that. He has the aircraft at the same place, roughly same track, and slightly lower ground speed than I do at 19:40.

    Not to say I am in love with the path I plotted recently. It is just another path as far as i am concerned. I am certainly not evangelizing it, but I reserve the right to evangelize it at any time in the future.

  12. DennisW,

    It would help if you clarified which document you are reading. I presume it is the one linked to as “Satellite Orbit”. There is a link within that to a file “AIAA-2006-6753Code.zip” at “centerspace.com” that is broken. Is this what you are referring to? If so, the file can be retrieved from celestrak.com. Unfortunately the World Wide Web does not have the longevity of clay tablets such as those produced during the Babylonian era.

    Two years ago I was quite positive that all my equations were correct. They are all coded up in routines that I call “satpos” and “satvel”, which I have been using ever since. (For historical reasons, I use TCL, not Python, but it shouldn’t matter). The equations are simply a generalization of the epicyles of Ptolemy (circa 100 AD). I am seriously considering releasing the code for all to use, including cross-comparing with Python code.

  13. @SK999

    Yes, I think we are talking about the same document – how to compute satellite x,y,z,vx,vy,vz. I’ll take another pass at it in the AM. Probably I am not grouping terms as you intended.

    If I knew where the satellite was and its velocity components at 18:40, I might take a pass at reconciling the 18:40 BFO, and get Jeff off my ass. Maybe not.

  14. @All, the ATSB said there were no pilot inputs at the time of fuel exhaustion and subsequent high rate of descent. How do you know this? If the SDU/AES logged off before due to power interruption, could this not have occurred again when fuel was exhausted? Could this be the cause of the anomalous BFO if that occurred? I am trying to understand why the aircraft could not have glided another 170km or thereabouts, also detaching flaperon, after the last unsuccessful ISAT logon.Apologies if this was covered before as I must have missed it

  15. @MattM
    I am trying simulate the moonset and sunrise times of MH370. But FS2004 is apparently inaccurate for the Moon after 2003 – I cannot see the Moon at all in the Sim for 8-Mar-2014. So if I want to simulate Moonset like MH370, then I have tell FS2004 that the date of flight is 11-March-2003, which has approx. the same Moonset times as 8-March-2104. So then for SIO sunrise I would need to change the sim date to 8-Mar-2014 to get the same sunrise time as MH370.

    So this is one possible reason the first part of the flight from KLIA to the Andamans in the sim could seem disconnected with the later points from Andamans to the SIO.

  16. @Dennis W are you not using the Henrik Rydberg satellite position/velocity interpolation? Or did you come up with your own from the TLE?

  17. @Paul

    I was unaware or Henrik’s work. Been using the table provided in the ATSB report and an interpolation for 23:15.

  18. @sk999

    Ok. As a last resort I had to read the document. Did not realize your x,y,z coordinates were rotated relative to ECEF. Duh! Do I have to do everything around here? 🙂

    BTW, thanks for your post. A treasure trove of info.

  19. @Dennis W you’ll find it here
    http://www.aqqa.org/category/models/
    download BSMv7-9-4.xlsx. under tab PAR5. As it seems you haven’t perused this spreadsheet before, you are in for a treat. Its a thing of beauty (if that’s what pushes your buttons) and a tribute to Barry Martin.

  20. @Paul S

    Thanks! I got sk999’s formulation working just fine. I swore I would never use a Microsoft Product again once I retired. It is difficult for me to even look at other people’s Excel spreadsheets.

    As an aside, Google Sheets, is extremely useful. Lots of built in function calls that scrape the WEB for all sorts of info. I find the function calls to Google Finance particularly helpful. In my corporate forced MSFT days I had to build my own WEB scraper.

    As another aside, in the more upscale NorCal communities realtors won’t take promo pictures inside a house with a PC visible. It would reflect poorly on the owners.

  21. @Matt Moriarty

    Interesting you came up with the Dordrecht Bassin (Hole). A fiction I suggested a long time ago based on the idea a pilot who planned this ‘vanishing’ of MH370 would also choose an end point which was most suitable.

    The Dordrecht Hole is most suitable with its depth of ~21.000ft. And indeed the SIO sim-fix could have served as a ‘distant waypoint’ for a great-circle route.

    VictorI came with his great-circle McMurdo-route which also crossed the Dordrecht Hole and intersected the 7th arc around ~28S.

    Problem still is 280Nm is too far away for a glide from the 7th arc.

    Do you see any solution to overcome this problem?
    Is it only a range/endurange/fuel problem?

  22. @DrBobbyUlich:

    You wrote to Matt Moriarty Nov.20 at 3:48 PM –

    Simply put, my detailed fuel calculations appear to demonstrate that the 7th Arc is actually NOT reachable in the center of the ATSB’s Current Search Area (at ~37S). It all comes down to a missing footnote.

    Let me demonstrate my thesis …

    Without knowing where you start your fuel calculations I find it difficult to follow your thesis. You could help me by providing UTC, fuel on board, latitude, longitude, and altitude for your starting point. Thanks in advance.

  23. @Matt Moriarty @Gysbreght

    I was thinking maybe it could be usefull to put the question the other way around.

    To reverse-engineer/calculate an FMT from a specific possible crash area like maybe the Dordrecht Hole or another area.
    So to not take an unknown FMT as a starting point but an assumed crash area.

    What conditions would be necessary? What speed, altitude, fuel, auto-pilot etc.?

    Where should the FMT must have been if calculating this way?

  24. @Ge Rijn: “maybe it could be usefull to put the question the other way around”

    No, I don’t think that is useful.

  25. Gysbreght said;
    – file “3” shows a 20 degree left bank and normal turn-rate passing
    through 255 HDG (a virtual FMT performed by the A/P at 20 deg limit)
    – there are ZERO airways headed south anywhere in the vicinity of
    10N90E (file “3”)
    ..@buyerninety..

    Not sure what your question was.. of course, whether the (sim)aircraft
    passed through 255 is not a stateable fact, it’s an assumption
    what is known is that 255 or 256 will point you straight at Colombo.
    (Keep the bank going, you’ll shortly be headed towards Male, another
    destination MAS flys or flew to;
    http://routesmap.blogspot.com.au/2011/09/malaysia-airlines-routes-map.html
    )
    And as there are zero airways leading south at Coord 3, it would be
    reasonable to assign an equal, or greater, probability that the (sim)
    aircraft was going where it was headed.
    Incidently, I remember a long while ago, thinking about the Altitude
    value for Coord 5, (that we now know is 3999.99) – I seem to
    remember that MS flight sim had a level of 4000 ft where the scenery
    changed in detail?, or maybe it was colors? -anyway, if you are testing
    to find out why your sim crashes, one methodology would be to reset a
    parameter just at, or about, where you know a change occurs, start the
    sim and then exceed the parameter – (like 3999.99 to immediately 4000).

    Making a provisional guess, it seems that the two FLT files in Table 7,
    Data-from-Flight-Simulator-Computer.pdf , dated 8/12/2014 06.49pm and
    dated 20/2/2014 11.06am , which are identified as ‘Program Generated
    Temporary Flight’ may be used for resets, as this webpage mentions;
    “I think it actually saves the program generated temporary flight”…
    (files)…”It’s used for Reset Flight”.
    Cheers

  26. @buyerninety: “Not sure what your question was…”

    I replied to MM’s assertion that “file “3” shows a 20 degree left bank “. You had provided a link to the Paul Howard blog, who argues correctly that 20 degrees bank could be either left or right. File “3” gives a positive number, and the usual sign convention uses positive values for right bank angles, and left bank are denoted as negative values. While it is of course possible that FS9 or PSS uses an unusual sign convention, there is no authoritative reference for that.

    The Paul Howard blog also has a map from SkyVector where the aircraft course is approximately parallel to an airway towards the Maldives, contradicting MM’s assertion that “there are ZERO airways headed south anywhere in the vicinity of
    10N90E (file “3”)”.

  27. CORRECTION: For “airway towards the Maldives”, please read airway towards Colombo (Sri Lanka).

  28. DennisW,

    Great that you were able to make use of that old satellite orbit paper. My equations were derived from a proper dynamical model for the satellite motion. As I recall, Henrik developed a set of equations independently at about the same time, and while he was making an empirical guess about what terms to use in his equations, they were prettly close to what I was using, so his equations worked just about as well.

    R.E. code release, it would be about 10,000 lines (plus about 200,000 lines of data files). Everything that I use to compute flight paths, BTOs, BFOs, plus code for running nonlinear least squares solutions. The holdup is making figures.

  29. @sk999

    As you probably know, when it comes to releasing code, no good deed goes unpunished. Every time I have done that I ended up with a maintenance/Q&A chore. No amount of documentation and release notes can spare you. Best to let sleeping dogs lie sometimes.

  30. Dr. Bobby:

    As always, I’m impressed by your thoughtfulness and thoroughness. I’m still looking at your PDA/temp assertions. I can’t find that 3% increase for every ISA+10 figure in either my GE or my RR manual and it doesn’t sit right with me. I’ll get back when I have something better for you.

    To me the surest way to get a handle on the actual difference between the performance tables and 9M-MRO (with its unique PerfDeg) on that particular night would be to work backward from the 17:07 ACARS report and compare to the manual. Have you done that already?

    In the meantime, I’ve drawn your route out of ANOKO several times and, unless I’m grossly misreading that paragraph, I cannot get that route to touch the 19:40 arc at all.

    I drew your 12nm north offset from N571. I made your left turn near IGOGU. I entered ANOKO on a hdg of 185.5 true and departed it on 183.6 and the closest I could get it to the 19:40 arc was about 25nm. 455 kts gs also has it crossing the 20:40 arc 40 mins early.

    My Google Earth also showed the bearing from ANOKO to the terminus you describe as 182.8 true, not 183.6, but maybe that’s splitting hairs.

    The path between 18:25 and 19:40 has been the bette noir of all of us from the minute this puzzle got laid on the table. Did it bite you in the butt here, too?

  31. @Gysbreght

    Positive in pitch is down and positive in bank is LEFT. The authoritative reference for that is Victor Ianello from this past summer when he posted the spreadsheet to Duncansteel and he will weigh in even more authoritatively on the subject of .flt files very soon. You’re gonna love it.

    “Contradicting MM’s assertion that there are no airways headed south…”

    Somebody get Gys a compass.

    @buyerninety

    It’s not an assumption. Its an observation that the lateral acceleration showed at turn rate of .89 deg/sec while in a 20 degree bank.

    @TBill

    I’m not surprised. There are many idiosyncrasies, especially in the way it saves data to .flt files. We’ll learn more about those soon. Happy moonrise in the meantime!

  32. @Matt M @all

    I am sensing some convergence on this problem. Of course, there are stubborn hangers on with “skin” in a different game, but the pointers are becoming more and more consistent. We are looking at a terminus at 30S +/- based on the Shah sim data, the drift analytics, and a consideration of motive.

    Anything South of ~30S is a non-starter at this point, and the ATSB is wasting their time (and taxpayer money).

  33. @Paul S

    “and a tribute to Barry Martin.”

    Is this not the same Barry Martin who championed the notion of different BFO biases for different channels in the Inmarsat data. No plausible reason was given for the results, and the statistics used were inappropriate for the problem statement.

  34. @DennisW

    I agree. It would be great to see the same thing happen with the path between 18:25 & 19:40.

    @GeRijn

    The fact that the plane came down shy of Dordrecht I chalk up to the inevitable imperfections in executing any plan. There’s always a point where the go/no-go decision is made. When he looked at the good weather in Beijing and realized he wouldn’t be able to load extra fuel without implicating himself in the post-mortem, it probably pushed him toward no-go. But then you weigh that against a whole list of other factors, whatever they were, and somehow the decision was made to proceed.

    He wasn’t certain he’d make it past Indonesia without an interception. It was probably his biggest worry, sending the loss of the really deep water down the list. He clearly had no idea that he’d be tracked with BTO/BFO, so once he was a half-hour into the final leg, he knew he’d pulled it off and by that point, another 10,000′ of water wasn’t that critical.

  35. One request to the Shah did it camp.

    What happened to the passengers and to Fariq? Being tricked out of the cockpit would have alerted him and everybody else within the first half hour after the turnback. Did he kill Fariq or did he just lock him out of the cockpit? Why was the portable ELT not activated? Why was only Fariq’s phone active over the mainland? Even the installed ELT was accessaible after removing two roof panels and can thus be activated manually, Why was it not used?

    Had Shah the guts to kill all the passengers instantly after the turn around without anotjer helpfull hand? Was he even the guy being able to perform such a mass killing and continue living for further 6 hours?

    We heard the story about the passengers being held as ransom for political negotiations. When took those political negotiations place, with whom, by what means and for what purpose? What evidence is there that such negotiations were even planned, were attempted or were performed? Wouldn’t such failed negotiations become public to show the ruthlessness of the governement to achieve the desired political aim?

    What routing would such a rouge captain choose while negotiating and why ended the whole flight somewhere around 30° defree south? Why not hold in the vicinity of possible landing destinations and give up, when the negotiations failed and land somewhere?

    Is the above described initial intention mainly favoured by Dennis not completely Bullocks when we consider the Sim points as pre planned definition of the grave for all POB?

    Are we now back to the suicidal mass morder scenario?

  36. @DennisW, you wrote: “We are looking at a terminus at 30S +/- based on the Shah sim data, the drift analytics, and a consideration of motive.”

    Can you remind me what that motive is? I thought it had evaporated.

  37. @DennisW, A group of people telling themselves that they’ve figured out the problem is what bought us a 120,000 sq km search area.

    There is a difference between a hunch and a hypothesis.

  38. @Matt Moriarty: “The authoritative reference for that is Victor Ianello from this past summer when he posted the spreadsheet to Duncansteel”

    When Victor Ianello posted the spreadsheet to Duncansteel this past summer, he had no authoritative reference on Microsoft Flight Simulator other than the RMP report. Let’s hope that soon he will be better informed than he was last time.

  39. @Gysbreght, @Matt Moriarty, I think it’s important to keep front of mind that Victor has an axe to grind and is using selective release of the RMP report to spin his agenda.

    Recall that he spun me by giving me just a few pages of the report which, taken out of context, supported the interpretation that he wanted to promote. This interpretation turned out to be inaccurate.

    If he wants to move the discussion forward he should release the whole report so that the community can work through the data, debate, and form its own conclusion.

  40. @Gysbreght

    the motive was to get around Indonesia and to land somewhere

    why he/they didn’t succeed is another question

    @JW

    more like a group of people totally disregarding common sense&logic

  41. @RetiredF4

    You said

    “What happened to the passengers and to Fariq? Being tricked out of the cockpit would have alerted him and everybody else within the first half hour after the turnback. Did he kill Fariq or did he just lock him out of the cockpit? Why was the portable ELT not activated? Why was only Fariq’s phone active over the mainland? Even the installed ELT was accessaible after removing two roof panels and can thus be activated manually, Why was it not used?”

    My response to that: Yes he locked him out of the cockpit and then depressurized the cabin. There would have been panic and confusion. Fariq would have very quickly guessed what was happening _ the plane had been taken over by a rogue captain. Fariq as the next senior crewman would have considered it as his responsibility to try and alert the outside world as to what had taken place, so he would naturally have switched on his mobile. Fariq would have had access to a portable oxygen cylinder. the cabin crew would have been trying to tend to the desperate passengers. in these circumstances one can understand why no one else syitched on their mobiles, and why the portable ELT were not activated (assuming they would work from inside the cabin). The fact that only Fariq’s mobile was on when the plane passed Penang Island, is very strong evidence that the captain depressurized the plane.

    You asked “Had Shah the guts to kill all the passengers instantly after the turn around without anotjer helpfull hand? Was he even the guy being able to perform such a mass killing and continue living for further 6 hours?

    Simple answer is Yes.

    You said “We heard the story about the passengers being held as ransom for political negotiations. When took those political negotiations place, with whom, by what means and for what purpose? What evidence is there that such negotiations were even planned, were attempted or were performed? Wouldn’t such failed negotiations become public to show the ruthlessness of the governement to achieve the desired political aim?

    My considered reply: any theory purporting negotiations to have been attempted, id complete bullocks. Absolutely no evidence for it. In fact it is totally preposterous.

    You said “Is the above described initial intention mainly favoured by Dennis not completely Bullocks when we consider the Sim points as pre planned definition of the grave for all POB?”

    Simple answer, Shak would not have been foolish enough to leave SIM points for the actual flight path taken. Any SIM points left on the computer would have been left to confuse, IMHO.

    You said “Are we now back to the suicidal mass morder scenario?”

    Answer, As far as I can see, the suicidal, mass murder was always the only realistic scenario on the table.

  42. @DennisW

    You said “I am sensing some convergence on this problem. Of course, there are stubborn hangers on with “skin” in a different game, but the pointers are becoming more and more consistent. We are looking at a terminus at 30S +/- based on the Shah sim data, the drift analytics, and a consideration of motive.

    Anything South of ~30S is a non-starter at this point, and the ATSB is wasting their time (and taxpayer money).

    I reply, there are none so blind as refuse to accept the scientific evidence! Dennis, are you conveniently ignoring the CSIRO drift analyses? 30S is completely out of the question. 38S is the most likely starting point. “Yes, but why haven’t they found the plane yet?” I hear you belatedly reply. Well simply because they overlooked it with their inappropriate equipment/methods (possibly its lurking in one of the shadowed areas, yet to be searched with an AUV) or he managed to glide sufficiently south (by accident or design) of the search area.

  43. @Matt Moriarty

    Yes in a scenario like this I can imagine his prime concern would have been interception by Malaysian Air Force or maybe even more by the Indonesion Air Force for they are notorious for their alertness and quick, progressive (to not say aggresive..) interceptions.

    I think that if this was his prime concern he would have chosen to turn late towards the SIO to avoid Indonesian airspace/radar/FIR as much as possible and without wasting time and fuel with a loiter or further turns and more risk of detection.

    In this regard avoiding Indonesian radar with a steep descend just after FMT to low altitude and fly on a while like this with (necessarily) lower speed would make more sence IMO.
    Climbing back to altitude again when he felt save enough all without altering his heading towards his ‘distant waypoint’.

    Maybe this could also explain the ‘time-problem’ between 18:25 and 19:41?

    Then maybe no complicated turns and loiters are needed.
    Just one great-circle route from a late FMT to the SIM-6th-‘distant waypoint’ towards the Dordrecht Hole without ever reaching it.

Comments are closed.