Russian Military Planes, Flying With Transponders Off, Provoke Alarm in Europe

FIR maps small
credits: left, Financial Times; right, SkyVector

 

In the latest in series of aggressive maneuvers by Russian military planes in European airspace, the Financial Times is reporting today that a Russian intelligence plane nearly caused a mid-air collision with a Swedish passenger jet on Friday while flying along a Flight Information Region (FIR) boundary with its transponder turned off.

An SAS jet taking off from Copenhagen on Friday was warned by Swedish air traffic control to change course to avoid a Russian military intelligence flight, said Swedish authorities.

Peter Hultqvist, Sweden’s defence minister, said it was “serious, inappropriate and downright dangerous” that the Russian aircraft was flying with its transponder — used to identify its position — switched off. He told Swedish reporters: “It is remarkable and very serious. There is a risk of accidents that could ultimately lead to deaths.”

The incident is the latest in a series involving Russian military aircraft over the Baltic Sea this year. In March, an SAS airliner came within 100 metres of a Russian military aircraft shortly after take-off from Copenhagen, Swedish television reported.

In the most recent incident, the Swedish and Danish military detected the Russian aircraft in international airspace on radar and warned the SAS flight, said to have been bound for Poznan, Poland.

A story about the incident in WAtoday links to a YouTube clip of ATC audio combined with speeded-up playback the commercial flight from Flightradar24.com, which indicates that the incident took place near the boundary between two FIR zones, Sweden and Rhein-UIR, with the Russian plane flying west to east along the boundary.

As I wrote in an earlier post, military pilots have been known to fly along FIR boundaries with their transponders turned off as a means of escaping detection. In what may or may not have been a coincidence, after it deviated from its planned course to Beijing, MH370 flew along the FIR boundary between Malaysia and Thailand with its transponder turned off. The pilot in Friday’s incident may have been testing NATO air defense systems to see how well the technique might work over busy Europeans airspace.

351 thoughts on “Russian Military Planes, Flying With Transponders Off, Provoke Alarm in Europe”

  1. @airlandseaman: thanks for posting. Initial comments:

    (FYI: I remain a novice at actual signal data analysis, and am looking at this from a pure math/stats perspective. Hoping Bobby comments – he seems to have a bead on a layer of the onion that needs to be peeled out before applying this kind of analysis.)

    1) I replicated your trend line (equation)
    2) Your 72.51 trend line BFO is for 18:40:00; refining to 18:40:20 gives 72.65, for a variance of 13.94
    3) Using 5 data points to (in)validate a 6th is dangerous – particularly if the 6th is an endpoint*
    4) If I include all 6 data points, the 18:40 error drops to 6.99; I think this is the test Bobby applied
    5) If 13Hz error = reject, I’ll need it patiently explained to me how the “decompression scenario” got published

    * For example, suppose that, instead of using [19:41..00:11] to test the fit of 19:40, I instead used [18:40..22:41] to test the fit of 0:11. Using your method, the trend line misses the 00:11 reading by 10.55Hz; comparable to your 13.94Hz for 18:40. How do we know it’s not the 00:11 value that’s the outlier?

  2. Brock:

    There are no outliers here. All good data. The first engine did not run out of petrol until after 0011, so the 1941 – 0011 data is all indicative of a fairly straight path at a near constant speed and direction. Extrapolated back to 1840, we see that the final turn was very likely still in process at 1840. 14 Hz is not an error or an outlier. It is the result of a different direction at 1840…probably ~200-210 degrees. Hope this helps explain.

  3. @Rand:

    Thank you for that post. It says a lot. And it means a lot. And it confirms some things. And I hear the (understandable) glimmer of despair in your writing. And I get it.

    But you know what? The ‘crowd’, wherever it’s to be found, is energized. And not stupid. Whatever is going on behind the curtain, ‘they’ will win if people just throw up their hands and stop looking. And stop caring. In fact, I’d submit that may be the objective.

    But we can’t give up.

    No investigations of any kind in Malaysia doesn’t mean that there aren’t people investigating something somewhere else. We just may not be aware of it. And perhaps we aren’t meant to be.

    But something isn’t right.

    Think about it: if you’re the owner of a troubled flagship airline (one that’s now being re-organized) you want to do everything possible in the wake of a disaster to show the world that you’re turning over every rock and care about finding out what happened.. Because if nothing else, it would inspire confidence with passengers and investors. Moreover, in the 21st century, a widely-used commercial aircraft with a great safety record doesn’t just disappear (with all indications that it was intentionally diverted) and the plane’s owner (read: Malaysia) is nonchalant about investigative follow-up. Particularly when that country hasn’t lost one, but TWO of its airplanes within months of each other.

    Recall that Indonesia’s police chief called a press conference to announce that he knew what happened to MH370. And his counterpart in Malaysia immediately jumped in to say that he (in Indonesia) was misquoted. https://twitter.com/KBAB51/status/511152093017563136 There were apparently multiple news organizations present that heard the same message, but Indonesia’s police chief said nothing further after being ‘corrected”. And that all went away.

    So Malaysia isn’t interested in further investigation — and they’re also not interested in anyone else (at least Indonesia) talking about the whereabouts of this plane.

    Do people also note that the investigations of MH370 and MH17 have strange goings on in common?

    You know what it feels like? Somebody’s trying to apply pressure on Malaysia (and I would include in that of a political variety), teach them a lesson (read: payback) or Malaysia (and others) have been been silenced.

    Rand, while we may occasionally disagree or misunderstand each other, we are buddies (forever) and we’re on the same side. I hope you know that I know where your heart is. The point of debate (however heated it might get) is not to win, but to clarify — and to discover. And I believe we are both, along with everyone else here, sincerely making an attempt to do that.

    I also know there are some out there who would like nothing better than for us and the rest of the ‘crowd’ to wave the white flag and walk away.

    But all I keep thinking about are the souls on that plane and their next of kin. And if I were either one of those, I’d hope that some kind stranger would care enough to fight for me. Or the soul of me. Because we live in a crazy, dangerous, unpredictable world. And there, but for the grace of (insert your deity here or none), go we.

  4. Never wave the white flag the truth must be known. There must be something in the early days that gives the truth in some coded way.

  5. Mike,

    We are not yet connecting on a key point. I will try to do a better job of explaining, using your crisper terminology.

    What I am saying is that your last comment has the same statistical validity as the following statement:

    “the 1840 – 2241 data is all indicative of a fairly straight path at a near constant speed and direction. Extrapolated forward to 0011, we see that a (new) final turn was very likely already started by 0011. 10.55 Hz is not an error or an outlier. It is the result of a different direction at 0011…probably ~[whatever a 10.55Hz delta translates into] degrees.”

    I’ll post a spreadsheet to show all calcs. The key concept is that, if I fit a straight line to the FIRST five points of the six in your series (1840..2241), I get a lovely trend line of

    y = 674.985940533x – 28,150,877.3244987

    …with a robust-looking R^2=0.991. If I extend THAT line to the RIGHT to test whether the SIXTH (0011) reading indicates a turn, I’ll get roughly the same answer you did when you extrapolated out to the LEFT: trend line of 91.45 vs. data of 102 = variance of 10.55 Hz – my gosh, the plane MUST have been 12% of the way into a NEW FMT (NFMT?) at 0011!

    (Maybe it WAS…)

    Why such a different conclusion? Because the data is both SPARSE and NOISY. There are large random errors buried in those scant few values – elsewhere discussed – and using any 5 to draw conclusions about the 6th is as inappropriate as what’s known in the business as “over-fitting” (adopting a best estimate that is herky-jerky, simply because it fits the noisy data with the most precision). For the exact same reasons.

    Bobby’s path runs essentially along the trend line you’d get if you used ALL 6 data points:

    y = 711.281417256x – 29,664,610.5880994

    This is the trend line I used to show that a more appropriate characterization of the 1840 variance was only 6.99 Hz.

    And I remain perplexed as to why we’re being asked to cast aside Bobby’s scenario for a [proposed-vs-BFO-indicated] bearing variance several TIMES less glaring than a scenario you yourself have publicly defended as still more likely than, say, the only place currently being deep-sea searched (by Go Phoenix)…

  6. I’m in the process of moving so have not yet read this latest article from our gracious host Jeff. I just want to take a moment and wish all the bloggers, IG Group, ATSB, data tweakers, analysts, mathematicians, searchers, and other “peanuts” a very blessed holiday season.

    To the MH370 families,I wish you the same. Never give up hope. May the New Year bring the answers you all so rightly deserve. I’m sure I speak for all when I say we think of you often, we haven’t deserted you, and will continue to help you to find out the fate of your loved ones.

  7. @Brock: Bobby needs a convincing argument as to why a great circle path is followed after overflying a waypoint immediately following his FMT. That’s not what the manuals say.

  8. Airlandseaman:

    “The C6 median of 51 obs at 184020 are 13.80 Hz higher than the linear fit to the 5 R4 values between 1941 and 0011.”

    “The first engine did not run out of petrol until after 0011, so the 1941 – 0011 data is all indicative of a fairly straight path at a near constant speed and direction. Extrapolated back to 1840, we see that the final turn was very likely still in process at 1840. 14 Hz is not an error or an outlier. It is the result of a different direction at 1840…probably ~200-210 degrees.”

    Bobby has shown that the airplane was not turning during the 51 obs at 184020.

    Isn’t the logical conclusion then that there probably was another turn between 1840 and 1941?

  9. Brock:

    Thanks for your analysis. I don’t have a problem with your spreadsheet per se, but it is a “math exercise in a vacuum” so to speak. It ignores everything else we know about the case. As discussed so often, we have a fundamentally ill-posed inverse problem that must be constrained by all the available contemporaneous data and some crucial assumptions. In particular, we know from radar data the airplane was headed ~ 290 degrees until at least 1822, and we know that all the models that assimilate the complete BTO and BFO data set put it on a heading of ~190 degrees after 1840. Constrained by these observations and facts, and the virtual certainty that there was no one alive to make a turn at 0011, it seems pretty obvious which of the 5 out of 6 points between 1840 and 0011 to fit to a trend line, and the implications for the FMT timing.

  10. @airlandseaman @flitzer-flyer @guardedDon

    Wow, thank you so much for the speedy and precise on the spot information about the altitude data. I see that nearly everybody on reddit was truly misled into this trap.

    just to confirm, al documents available confirm your view, that there is no value of Zero possible in the ADSB-burst.

    One tiny reservation has to be put on this because it depends on the configuration of the altitude parameters, that is up to the airlines in the different countries. While all advisories use the word “desirable” for the including of the GPS-altitude (GNSS), they do not use the word mandatory. So there is a very remote chance that the MAS configured its ADS-B in a way that it did not use the GNSS. But since the connections to australian customs are so close they will probably have used the same configuration ike the australians, which are what you say.

    Now the puzzle of the Zero value in the last two ADSB-Burst recordings get even mor mysterious.

    I think, since in hindsight we know that these were the crucial seconds where the fate of MH370 was decided we should look for any iregularities by that time. So the missing altitude data could be the “smoking gun” we are looking for. Because when there are problems with data we should think about electronics and electronis bay and what happens to data if there is a process for shutting down comms that originates from the E/E-Bay, or even an electronic hack or a bad software fault, where the control of the plane gets lost.

    So i would like to ask if anyone got a timetable by the second here, that states when exactly the last voice message ended, when exactly the secondary radar data ended and which values have been displayed there.

    Although the GPS-altitude is notoriously unreliable and nobody really likes to use it, it should have been a very strange coincidence that the A/C got into GPS satellite shadow exactly at the time, when the switch to HCM Frequency should have been operated.

    The clearification of this altitude issue could finally end all speculations about pilot suicide, because it might come out, that this faulty value could never have been created by the pilots.

    The consquence could also be, to try to replicate the mistake in a simulator and find a clue where the problems started.
    Just to mind: this would be the second unexplainable shortcoming of the electronic system after the unexplained logon.
    (thank you for the documentation. I went through that already back in April in search for the communication format of the pings, because i still wonder, why the GPS position was not transmitted with the pings, even if it was not recorded in the INMARSAT files , but then i did not have the ADSB-data)

  11. Gysbreght:

    Bobby has claimed it turned earlier, but I, and many others do not agree that he has “proven” that theory. We continue to discuss that question with Bobby on a back channel, and I believe we are making progress. The differences in our assumptions have been noted and discussed here by Victor and Richard G. I won’t repeat it all, except to note that all the information gathered from (1) 777 manuals, (2) 777 line pilot interviews and (3) a 777 simulator trainer all confirm that the AP will continue on the last heading once past the final WP has been passed, NOT a great circle. (Bobby has correctly noted that it would be possible to write AP code that would allow continuation on a great circle, but that is not how the 777 AP works.) So path models that assume a great circle path are simply inconsistent with all available information on AP behavior. See for example the following snip from the Continental 777 Manual, Page 1029:
    https://www.dropbox.com/s/49z386jb61ra022/CON_777_Manual_PG1029.JPG?dl=0

  12. @airlandseaman: as should have been obvious from the commentary, I’m not suggesting the “first 5” version is equally intuitive. That version merely illustrates the statistical weakness of your conclusion; if I WANTED to see a FMT starting at 0011, this noisy data would certainly accommodate me. And if we’re going to inject a dose of reality into this “vacuum”, how about the fact that E84 (18:28 FMT) was missed in the surface debris search, whilst E88 (18:40 FMT) was SCOURED?

    @F_F: I’m arguing not for Bobby’s scenario, but for searching out to E84, i.e. for ALL scenarios whose FMT was at or near 18:28. Because ruling them out (by overinterpreting noisy data) transforms E84-E87 from a “hotspot” (searched first) into a “blindspot” (searched last, or never). That the JIT appears (in its Oct.8 ATSB update) to have cleverly PRESERVED this zone’s blindspot status is deeply troubling.

    @Gysbrecht: Yes, this I am willing to concede – I proposed last week the FMT may have occurred in 2 stages – say, 90deg at 1828, 15deg some time shortly after 1840 (i.e. once clear of Sumatra). This looks intuitive to me on a map. But both Bobby’s AND this (GhysBrock?) scenario predict an impact point west of E85. If the signal data is authentic (…), they should have been covering that base from the moment an 1828 FMT was deemed plausible.

  13. Brock:

    I have often acknowledged that the E84 scenario is not impossible. It’s just less likely. It’s like the search width arguments. A recommendation to search *first* in a relatively narrow band around the 7th arc…say ±20NM… does not mean 370 “must be in a narrow band”. It means that you have a higher probability of finding it sooner there, and if it does not turn up there, obviously, expand the width. But you don’t begin the search 100 NM out from the arc and work your way back to the center. Same with the longitude argument. 89E is certainly not THE location on the arc. It is an estimate that is believed by many experts, including ATSB experts now, to be the most likely. So, it’s logical to start there. If nothing is found, then moving further to the SW is probably the best direction for expansion. Put quite simply, you can’t start everywhere at once! Why start at 84E when the AP does not follow a great circle route as required for the 84E scenario?

  14. @Brock McEwen

    hi Brock i got a message to transfer to you from pigdead on reddit:

    Quote

    von pigdead gesendet 1 Tag zuvor
    Dear WMRS,
    Since you are on the site, and I cant post (my job, totally unrelated), could you tell him its been pretty obvious to me why they moved the search area. They couldn’t get to it in the life span of the pingers with search ships. Given all the delays it was going to be too embarrassing to arrive on day 40 when pingers were likely dead. So they changed to an area that they only just got to (IIRC they heard acoustic pings on day 30). Im not claiming credit (deficit), but I posted that fact about a day before they changed the search areas.
    He has been chasing this bone for quite a while now, just want to relieve his pain 🙂

    unquote

  15. airlandseaman:

    Thanks for your reply. I never believed that the AP would maintain a great circle route after passing the last active route waypoint, but there could have been an active route waypoint after 1840 hrs. Also the absence of evidence of human activity after 1825 is not evidence of absence.

    In the graph you posted 17 December at 6:21 PM the FFB for the C6 cluster at 184020 appears to be 88 + 58.71 = 146.71 Hz and the FFB for the 5 R4 values between 1941 and 0011 is apparently 150 Hz. Do I understand that correctly?

  16. CosmicAcademy

    I do not understand your statement: “Although the GPS-altitude is notoriously unreliable and nobody really likes to use it…”. I have never heard such a claim. GPS altitude is very accurate and reliable, particularly with WAAS enabled (~7m). It is used for precision IFR approaches!
    One of my colleagues …actually several…have done an excellent job combing through the various ADS-B data sources and tools available for analysis thereof. The bottom line is that there is nothing mysterious about the “missing altitude and position data at 1721. It was also missing many other times during the flight. That’s just the way FlaightRadar24 and other ground based ADS-B tracking nets work. See the following for more details:
    https://www.dropbox.com/s/mfebeporb2kvbbh/ADS-B_Radar.xlsx?dl=0

  17. @airlandseaman

    this is the experience of a pilot on reddit

    “GPS data is not perfect, disappears at times and becomes invalid often due to satellite positional geometry being less than optimal. Also, altitude is not GPS’s strongest suit of data.”

    doesnt matter in this discussion anyway , thanks for the link will go through that tonight and be very happy if there be no mystery about these values

  18. Gysbreght:

    Yes, here are all the current best estimates for FFB. We are still waiting for some Square Peg and Thales feedback requested via ATSB, promised by not received so far. I do not expect the requested input to materially alter these bias estimates. The information requested may reduce the noise, but probably not the bias.

    Channel unit and type Mean FFB (Hz)
    R4 149.86
    R8 154.45
    T8 154.35
    T10 154.17
    R11 150.62
    T12 153.82
    C6 146.71

  19. Hi Gysbreght, airlandseaman.

    Gysbreght: I would be interested to know your thoughts on a couple of my posts in this and previous threads re FMT. In my opinion the time of FMT between 18:25 and 18:40 is just one of numerous possibilities.

    airlandseaman: I agree it appears that MH370 followed waypoints from the time it reached the Malacca. Moreover, the snapshot of radar image indicates that MH370 changed heading according to waypoints. However, it also appears that before ~18:00 and after 19:40 it did not follow waypoints. Why?

    I am skeptical about the timing of FMT because one of my models (‘mechanical’) gives surprisingly good agreement with BTO/BFO data and information provided by Kate (the sailoress). The minimum residuals correspond to the altitude of 3100 m if 23:15 BFO is included, and 4700 m if not. Low altitude also perfectly explains absence of data from the Indonesian radar. However, I do understand that the whole flight at the altitude of 3-4 km at 250 m/s air speed would result in too much fuel burnt. The only possibility I see is to add changing altitude into the model, but I do not know what assumption could be made with regard to it.

    Regards,
    Oleksandr.

  20. airlandseaman:

    “14 Hz is not an error or an outlier. It is the result of a different direction at 1840…probably ~200-210 degrees.”

    Interestingly, for BFO=88 Hz, FFB=146.71 Hz, Yap’s calculator gives a latitudinal velocity of 413.4 kt. Assuming 450 kt GS corresponds to 203 degrees true track.

  21. @airlandseaman
    wel thank you again. that excel file is overly convincing. no doubt whatsoever , the values can be forgotten.
    Seems that the the author of the thread on Reddit left many entries out, and so it happened to give a wrong picture.
    eating chalk
    🙂

  22. @airlandseaman: the IG’s bedrock principle is that the search should focus on the intersection of the performance limit and the 7th arc as a top priority. I had to fight through the JIT’s customary obfuscation, but I think we are all now finally approaching consensus that the range of PLAUSIBLE intersections is E84-E88. I tell you a third time: start at E88 if you think 1840 is a MORE plausible FMT than 1828. But search WEST from there. Fugro is searching EAST, and Go Phoenix has now spent months searching a zone NOBODY here thinks contains the plane. What a colossal waste of resources.

    @pighead via CosmicAcademy: thanks – it’s relief enough just to hear I have company in the “JIT searched for months where they KNEW MH370 was NOT” club. (I’ll send you a club t-shirt as soon as I find a way to squeeze the full name onto it.) Presumably, the acoustic ping authenticity was likewise fabricated as part of this operation.

    I agree that “theatrics to conceal embarrassing incapacity/poverty” was my initial assumption, too – and it remains a strong candidate – but why not just ADMIT this, now that the plot has blown up in its face? Until they ADMIT this, I must continue to hammer away at the inconsistencies in their story, in the hopes the MSM will some day muster the guts to call them out on it. Because you have to admit: there are far more nefarious reasons for faking a search.

    And even if you are 100% correct: I won’t feel “relief” until the JIT explains publicly to passengers’ loved ones that the raising and dashing of their fervent hopes was deemed an acceptable cost in pursuit of the much more important goal of making the JIT look heroic.

  23. Brock McEwen:

    re: “I tell you a third time: start at E88 if you think 1840 is a MORE plausible FMT than 1828. But search WEST from there. Fugro is searching EAST, and Go Phoenix has now spent months searching a zone NOBODY here thinks contains the plane. What a colossal waste of resources.”

    Well, it could be a bit NE or SW, but I agree with your sentiment about the waste of time looking far to the NE. Even that is not impossible. Just very unlikely.

  24. Richard Cole:

    C6 was calibrated using a “bootstrapping method” (Sid’s temonology). Basically, we calibrated all the channel unit/channel type pairs using ground truth (Gate C1 and ADS-B data) between 1600-1707. That gave us a calibration for R4 values. Then we looked at the relationship between the 2314 cluster of C6 data and the trend line for the R4 values. This is not as exact as calibrating with ground truth, but it is good enough to allow the use of thr C6 data with better accuracy than assuming a nominal FFB of 150 Hz.

    https://www.dropbox.com/s/umfqve8kkqxukfs/2014-12-17_Trend%20Line%20Calibration.JPG?dl=0

  25. @Brock,
    @airlandseaman,
    @Flitzer_Flyer,

    Let me address two topics on which you have commented. The first is the 18:40 BFO data and the timing of the Final Major Turn (FMT). The second is the default horizontal navigation mode past a waypoint discontinuity.

    18:40 BFO DATA AND TIME OF FMT

    1. The information contained in the BFO data sequence at 18:40 is contained in two basic parameters – the mean value and the slope.

    2. First, let me address the slope. As I have shown in my previous post, the slope during the 61 second extent of the 18:40 BFO data will be large and negative if a turn was underway during this period of time. It will exhibit at least a ~10 Hz drop over one minute (see my previous post and also the post by Victor providing the -10 Hz value). The slope in the actual BFO data is at most a rise of 1 Hz or less. Thus the 18:40 BFO data slope is quite inconsistent with the BFO model turn prediction and disproves the notion that a turn was underway at that time.

    3. Next, let me address the mean value of the 18:40 BFO data. Mike Exner has argued that the displacement of the 18:40 mean value above the trend line that can be fit to the subsequent data indicates that a turn must have been underway at that time. This conclusion is false and demonstrates a lack of understanding of the component terms that contribute to the BFO measurements. First, the fact that the subsequent BFO data appear to be rather linear is due to the fact that the uncompensated satellite Doppler coincidentally happens to be fairly linear for a number of hours as it heads south during its diurnal motion. For example, from 19:41 to 00:19 the satellite Doppler toward the aircraft changed from about 0 Hz to +58 Hz. However, during the entire period from 17:20 to 19:41, the same Doppler component only varied between -3 Hz and 0 Hz. So it was essentially flat until 19:41 and then it started ramping up afterwards at ~ 12 Hz/hour. If I fit a trend line to that Doppler component from 19:41 to 00:11, it would predict a value at 18:40 about 12 Hz below its correct value, because of the change in slope that occurred at about 19:41. Indeed, that is approximately what the total BFO trend line also shows, and this is, in fact, caused by the satellite, not by an aircraft turn.

    4. My BFO model, with a single turn at 18:28 and a bearing of ~192 degrees afterwards, predicts a BFO of 87 Hz at 18:40 using the C6 FFB of 146 Hz proposed by Mike Exner. The measured mean BFO value is ~88 Hz. The excellent agreement of the predicted and measured BFOs demonstrates there is no inconsistency in the 18:40 BFO that requires a FMT to be underway at that time. A southward path fits quite nicely.

    5. The simplest scenario that matches the BFO data (including at 18:40) is a single turn between 18:27 and 18:39. More complex scenarios involving multiple turns, or a turn and an accelerating climb, can also be devised to match the 18:40 BFO data, but these appear less plausible to me. One thing is quite clear: no turn was underway at 18:40.

    LNAV MODE AFTER FMT

    1. My main proposed route is a great circle through, and after, the waypoint WITN (Maimun Saleh Airport). It assumes that the FMS would continue a great circle upon passing the last entered waypoint. This may be an incorrect assumption.

    2. So far, further digging has revealed two B777 manuals that say different things on this subject. One says the last “heading” is maintained. Another says the last “track” is maintained. Neither one says the same great circle path would be maintained.

    3. In my opinion, it is more likely that waypoint(s) were used to set the final course, but it is possible that it was done using the MCP to set a course in one of four possible modes (constant heading or constant track with either a magnetic reference or a true reference). I will also point out that magnetic reference is normally used except in polar regions (thus NORM is the usual reference switch position rather than TRUE).

    4. I further note that all models, including the IG’s recent V13.1, make assumptions regarding the horizontal navigation mode. V13.1 assumes one waypoint (ISBIX) followed by a true track at 185.6 degrees. In the same vein, I evaluated true track routes in my original white paper and found they also provided very steady speeds.

    5. If the IG believes that the FMS definitely requires a constant heading mode after the last waypoint is passed, then why does their V13.1 model use true track rather than constant heading (probably with a magnetic reference) after passing ISBIX?
    6. Because of the westerly cross winds, all routes using “heading control” end up east of the “track control” routes.

    7. Similarly, all magnetic reference routes end up east of true reference routes.

    8. Similarly, using a true track route instead of a great circle route moves the end point eastward in this case.

    9. Using a true track route rather than a great circle after passing WITN shifts the end point on the 7th arc about 1-2 degrees east. The average speed is reduced by ~6 knots, which matches cruising speed expectations better. The Equivalent Still Air Distance traveled is also reduced by ~55 NM. This range reduction allows engine PDAs to be ~1.6% larger and is more likely to be reachable with available fuel.

    10. In summary, there are three arguments that would favor a (mostly) true track route south over a great circle route: (a) agreement with at least one B777 manual, (b) speeds closer to Long Range Cruise, and (c) a total distance traveled consistent with more realistic engine PDAs.

    PARAMETRIC STUDIES UNDERWAY

    1. I am currently evaluating routes by varying a number of parameters: horizontal navigation mode, altitude, inclusion of 00:11 BTO in fit, and speed optimization variable.

    2. I am evaluating at least two horizontal navigation modes – great circles and true tracks. I will also take a look at heading modes, but I am not optimistic on any of these.

    3. I will do fits constrained to pass through WITN and also fits that are unconstrained by waypoint(s).

    4. I will do constant altitude trials at 35,000 ft and at 40,000 ft.

    5. I will do fits to BTOs from 18:25 to 22:41 and from 18:25 to 00:11. My previous results were done jut through 22:41, but I will add the 00:11 point and see what impact that makes.

    6. I will do trials using three speed optimization methods: (1) constant TAS, (2) constant Mach number, and (3) the LRC profile modeled by Victor Ianello.

    7. The total number of trials to be done is at least 48. It will take a while.

    8. The key parameters of all fits will be shown in tables that allow convenient comparisons to be made. Stay tuned.

  26. @ Bobby Ulich

    that all looks and Sounds great, your analysis has sharpened Focus,.. GREAT..

    The issue is, I have The velocity algarithom for you to use, via limited from Publicity and understanding, its sensative in Nature..

    The other thing is… Thge timing walk back and forth, I will also provide a sharper scope upon the velocity algarithom your need to better understand what the BFO and BTO is actually doing as designed.

    It’s not that your analysis is Bad, It just needs the information and design scope, that is not available to the public.

    Also, yes, there is a turn upon 1827 and 1839, but its Not in Location You are thinking its at, & There is Proof in the
    Pudding…

    Going to bed now, Will catch up Later and email you the information if you like to consider the Facts and the evidence…

    @FelineNut The Chaotic Cat ;-)>

  27. 1/2

    @Falken:

    Thanks so much for the kind words and for the Foo Fighters!

    All:

    About three weeks ago you’ll recall I mentioned getting a tip about an upcoming article re MH370 that was to appear in Le Figaro. Le Figaro has not materialized. But what did pop up yesterday is a piece by Marc Dugain in Paris Match:

    Excerpt (translated from the French):

    “In the case of flight MH370, silence has replaced the excitement of the first hours with a rapidity which suggests that transparency is not the priority in this case.

    Many spies are frustrated writers. Hence the use of metaphors designed to discourage us to go further and make us understand, in the end, that ‘it is a highly sensitive matter of great complexity,’ and that ‘it is better to let time do its work rather than trying to speed it up, with all the risks that come up.'”

    http://t.co/XgN8lwvqGU

    Today, Dugain made quite a statement on France inter (radio). That’s the next post.

  28. @Bobby: I think the parametric study that you do will be helpful. It might be worthwhile to share some interim results here or by email so that models can be compared before you do all 48 cases, especially for the LRC speeds.

    You said that I posted something about a 10 Hz/min change in BFO in the turn. I don’t recall ever doing that calculation. Perhaps it was somebody else.

    Victor

  29. @nihonmama: I am glad that somebody is shining light once again on the sighting in the Maldives. Although there may be questions about whether this plane was indeed MH370, the timing makes it likely that the plane was somehow related to the MH370 incident. Here, we have multiple people seeing an airliner flying low and in a southerly direction. So either the plane seen in the Maldives was not MH370 (as I have postulated in my Banda Aceh scenario) or the satellite data is completely wrong.

  30. 2/2

    TODAY — Marc Dugain’s Paris Match follow up: “possibility that the Americans stopped the aircraft”

    Amongst other (some paraphrasing):

    -He was threatened by ‘English’ foreign agents/intelligence.

    -It’s impossible that Diego Garcia didn’t see MH370.

    -Malaysia and the US are involved – and it appears they do not want the investigation advanced.

    -MH370 was hijacked and destroyed before it crashed.

    – People in the Maldives were ignored.

    (French) http://t.co/wt7bd5zOR8

  31. In recent days I’ve tried to make a sensitivity analysis for the 6th ping ring (00:11) crossing, as a function of (error in) residual Doppler (D, in Hz), ground speed (GS, in kts), and track (T, in degree). An important assumption made is that there is zero ROC @00:11. Basically I’ve calculated D along the ping ring and looked at the position (P, in latitude S) with best fit to 29 +- 7 Hz. Note that I’ve used my own simple D calculator based on Henrik’s equations, using just a spherical earth model and an older ping ring KML file set obtained from Duncan. The values are therefore indicative and meant for methodology development rather than for exact location purposes. As the variation in D with 7 Hz possible error is the dominant contribution to the variation in P, I realize that proper error analysis would be essential. And I don’t know the size and nature of the different error contributions to the estimated 7 Hz error in BFO. Advice is welcome.
    An independent check with a different BFO calculator as well.

    In a first iteration I took:
    GS = 450, 520
    T = 170, 190
    D = 32, 46
    Which gives the following (GS, T, D, P):
    450, 170, 32, 29S
    450, 170, 46, 39S
    450, 190, 32, 29S
    450, 190, 46, 39S
    520, 170, 32, 30S
    520, 170, 46, 40S
    520, 190, 32, 30S
    520, 190, 46, 40S

    In a second iteration I took for GS @ 00:11 the average GS (two values) needed to travel from IGOGU @18:40 to the two limits in P, as to make a more self-consistent set of parameters. Required GS estimated using Skyvector.
    GS 380, 540
    T 160, 200
    D 32, 46

    Which gives the following set (GS, T, D, P):
    380, 160, 32, S27
    380, 160, 46, S37
    380, 200, 32, S27
    380, 200, 46, S37
    540, 160, 32, S30
    540, 160, 46, S40
    540, 200, 32, S30
    540, 200, 46, S40

    Note that to translate these values in a 7th ping ring position one has to add about 1 degree south to the calculated latitudes.

    There are a few observations that can be made from these values:
    – The uncertainty in the ping ring crossing is mainly caused by the (assumed) large possible error in D
    – The crossing of 6th ping ring is only weakly sensitive for local GS and track

    While the 540 kts average GS is probably unrealistic high, the combination 540, 160, 32, S30 suggests that there possibly exist flight paths which combine a late turn (long after 18:40) with high average speed and fit to the BTO/BFO data (within margins of error). Such flight paths would theoretically make a landing scenario as far north as Carnicobar possible. Such a scenario would require a descent @18:40 to reconcile the BFO value with a NW course at that point in time. Although such a scenario is not very likely, it would be important to verify if such a flightpath would be possible.

  32. @Bobby: You said “In summary, there are three arguments that would favor a (mostly) true track route south over a great circle route: (a) agreement with at least one B777 manual, (b) speeds closer to Long Range Cruise, and (c) a total distance traveled consistent with more realistic engine PDAs.”

    I am glad we are reaching agreement here. These are among the reasons the IG chose this mode of navigational control.

  33. @Victor:

    Couldn’t agree more.

    A ginormous airplane just shows up (flying low) over Kuda Huvadhoo (that never happens), numerous witnesses (WSJ:“We have talked to five to six islanders [individually] and their story is the same”) and all of that is dismissed.

    If nothing else, THIS should have been the signal to everyone that the ‘narrative’ warrants scrutiny.

    The rubber meets the road in the Maldives.

  34. Bobby:

    Taking them in reverse order…

    1. I agree with Victor. The parametric study would be a good idea. I would like to see a simple “two dimensional study”, one that that looks at the turn time as one variable (say 1822, 1827, 1832, 1837, and 1842), and for each turn time, a case for it turned “south” controlled by (a) a true heading hold and (b) normal (mag) heading hold. You could add from there. Include winds, fuel, etc. as constants.

    2. It sounds like we are in substantial agreement now on the LNAV Mode possibilities after FMT. I agree that it is possible an arbitrary WP somewhere far south could have been dialed in, but it would be so much more like a pilot to dial something in closer at hand, and let it “flywheel” to the end as described in the manuals, if any WP was entered. Just a pilot hunch.

    3. Regarding the attempts to calibrate the C6 FFB, it sounds like you are in general agreement that we can and should “bootstrap” the R4 ground truth based calibration (~150 Hz) to the C6 calibration (~146 Hz) along the lines (no pun intended) I have suggested. We still differ in our interpretation of the data around 1840, but there is a lot we agree on here. I will give more thought to your first derivative arguments. I note Victor disclaims any confirmation from him about a 10 Hz rate. And we all know how noisy differentiation can be. I don’t see any merit in trying to infer a slope from noisy data over a 1 minute period. But I will look at it again.

  35. Re the Maldives —

    Please recall (posted here previously) what journo Alberto Riva said on 03.18.14:

    “But the fire plus emergency diversion theory, as compelling as it is and similar to other known incidents, leaves one question unanswered. If the pilots tried for a landing at Langkawi and missed because they became incapacitated, the autopilot would have kept them flying straight and level on the last compass heading. (Which would have taken MH370 more or less over Kuda Huvadhoo, by the way.)

    So, either someone was entering those waypoints in the flight management computers, or the computers had been programmed earlier to send the plane there. The former option is not consistent with an unconscious or dead crew. The latter makes no sense for a crew in a dire emergency, looking for the closest place to land — unless one wants to believe the improbable and now-debunked scenario that hackers were steering the plane.”

    http://www.twitlonger.com/show/n_1sf4ubu

  36. And shortly after the Maldives seeing such low fly aircraft I read some where they heard a loud explosion type noise. Wondering if it blew up or crashed in one of the inner Maldives islets south of the eye witnesses.

  37. Hello Bobby,

    I agree with your argument against the turn @18:40. Probability of the call occurring simultaneously with the turn makes it even less likely.

    However, I am not very clear why you discard possibility of a series of turns (your point #5 “less plausible”). 18:40 BFO may not be in line with other BFOs simply because of different heading at that time. This could be, for example, if the aircraft was moving towards some other waypoint @18:40.

    With regard to the turn at 18:27, I am wondering whether SDU “power up” and possible turn are coincidental events. Could the turn be associated (caused) with a possibly simultaneous “power up” of FMS?

    Regards,
    Oleksandr.

  38. @Victor,
    @airlandseaman,

    BFO SLOPE AT 1840

    Victor, you are correct. It was Gysbreght who posted the following:

    “Gysbreght
    Posted December 11, 2014 at 4:14 PM

    Yap’s BTO&BFO Calculator shows that if the track changes from 220 degrees at 18:39:55 to 184 degrees at 18:40:56, the calculated BFO reduces by 10 Hz.”

    My model results agree with Yap’s model results. If a turn was underway then, the BFO would drop by ~10 Hz during the minute-long sequence (even at a very slow turn rate). The statistics of the data exclude a slope of even a significant fraction of 10 Hz, so they do not allow a turn in progress at that time.

    C6 BIAS

    Mike, your 146 Hz makes the 23:14 BFO data line up with the rest. That looks like a good value to use.

    PARAMETRIC ROUTE STUDY

    Following up on Mike’s suggestion of today, my original white paper in September shows the results for the four lateral MCP navigation control modes, including turn times, average speeds, speed variations, end point locations, etc. for a full range of bearings covering the region of interest along the 7th arc.

    DEFAULT NAV MODE AFTER WAYPOINT DISCONTINUITY

    I don’t know if we are in agreement just yet as to exactly what the FMS does when it reaches a waypoint discontinuity (is it track or heading or whatever was last used?). It does not appear to keep following the same great circle. It would seem from at least one of the manuals that a constant heading would be held. Is it constant in true coordinates or a constant magnetic heading or dependent on the switch position? In either case, I don’t recall seeing any heading route fits (except my own and perhaps some of the ATSB’s). All other routes either assume some waypoint(s) and/or a constant true track.

  39. @Bobby: In the past, I have studied the four cases of magnetic v. true and track v. heading for LRC speeds. I could only find reasonable fits to the BFO and BTO data for true track navigation after the FMT. That is not a conclusive result because there might have been viable cases that I did not consider.

  40. @ Bobby:
    I think there is still a deal of confusion over possible B777 roll modes, and the resulting flight tracks.

    Just because there are buttons on the panel that allow the selection of a Magnetic [NORM] heading reference, or alternatively TRUE, does not mean that it is possible to fly tracks based on this reference.

    Here is a little excerpt from the Honeywell B777 FMC Pilots Guide – –

    “Only the active waypoint course can be referenced to magnetic north because the ADIRU can provide magnetic variation only for present position. All subsequent waypoint courses are displayed as true courses.”

  41. @ Dr. Bobby Ulich

    “So far, further digging has revealed two B777 manuals that say different things on this subject. One says the last “heading” is maintained. Another says the last “track” is maintained.”

    The Lauda Air Operations Manual for the 777 (prepared by The Boeing Company in Seattle), on page 11.42.7 has similar text as the Continental Manual (text provided by airlandseaman):

    “If the airplane passes the last active route waypoint (or offset) or the last waypoint prior to a route discontinuity, LNAV maintains the current heading and a scratchpad message displays.”

    Can you please provide a reference to the B777 manual that says the last “track” is maintained?

  42. Nihonmama/Victor – plane sighting – the speed that the Maldivian govt shut that down is not surprising. They are as thuggish as they are authoritarian and increasingly hard line Islamic. The Maldives are also a financing hub for global terror and their remarkable finance laws seem designed to accommodate this. It looks idyllic but it’s degenerating fast into an Islamic totalitarian state. They had no solid grounds for the kneejerk shutdown of those testimonies, but the more you read you see the inclinations, and, possible motivations.

    https://www.opendemocracy.net/5050/moh%C3%A1csi/is-islamic-state-cover-for-government-policy-in-maldives

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.