Radar Maybe Captured Fighter Jet Chasing MH370

Radar data from MH370 and another possible aircraft. (Click to enlarge.)

As readers here know, we have been re-visiting previously held beliefs about MH370 to better understand why the underwater search failed to find the plane. The ATSB continues to believe the assumptions that it and the DSTG used to define the current search area were correct. But if these assumptions were correct, the probability of finding the plane where it was searched was very high. Either the search team was extremely unlucky, or some of the long-held assumptions are incorrect.

It is in the spirit of questioning assumptions that I present a new possibility regarding the military radar data of MH370 above the Malacca Strait: I consider that the captures recorded by the military radar installations might be from two aircraft rather than just from MH370, as I show in the figure above. The motivation for this inquiry is the following facts that taken together cause some concern:

  1. The radar data shown in the Lido Hotel image shows an aircraft following airway N571 at a speed of around 500 kn. The last radar capture is at 18:22:12 just past waypoint MEKAR. (All times here are UTC.)
  2. If MH370 continued following N571 after the last radar capture, at the time of the log-on at 18:25:27, it would be traveling on a track of 296T.
  3. The BTO sequence during the SATCOM log-on suggests that MH370 was not following N571 at 18:25:27 through 18:28:15. However, the final BFO value suggests the plane was flying at around 296T and 500 kn, which would put it roughly parallel to N571.
  4. It is possible that the pilot performed a lateral offset manoeuver of around 12 NM between the last radar capture at 18:22:12 and the log-on request at 18:25:27. However, the manoeuver would have to be fairly well-timed to fit in that period of about three minutes.
  5. If we ignore the military radar data after around 18:02, we can find paths starting from this time that match the BTO and BFO data without the need to invoke the lateral offset manoeuver, as shown in the last post.
  6. The ATSB reports that the military radar data supplied by Malaysia concludes with a position and time at 18:01:49, and then after a 20-minute gap, there is a single capture at 18:22:12. This contradicts the many captures shown in the Lido Hotel image during this 20-minute period. The DSTG also reports that the final radar capture at 18:22:12 was not used to reconstruct possible flight paths.

To investigate this further, I considered a version of the Lido Hotel image that was studied back in May 2014 by IG member Bill Holland, and shown below. Bill enlarged the high-resolution version of the photograph, and he noticed that there were timestamps printed to the right of the targets. (The timestamps refer to local Malaysian time, which is eight hours ahead of UTC.) Although the timestamps were blurred and overlapping, using a timescale that he superimposed as a guide, he could determine what the values were for some of the timestamps. From this, he deduced that the plane’s speed was around 500 kn.

Enlarged Lido Hotel radar image with timestamps added by Bill Holland. (Click to enlarge.)

But there are features in this slide that was never explained. For one, there is a capture at 18:07:06 whose position is about 7.5 NM off of MH370’s path, and also about 2.3 minutes later than the surrounding points. Because this capture doesn’t match the path suggested by the other captures, most of us have assumed it represent a false target, and we have ignored it. Another feature is that there is an explained “hole” in the radar coverage between around 18:07:06 and 18:12:30. The Malaysian officials thought this gap in coverage was important enough to draw a white circle emphasizing it.

We can see in the figure that the radar captures that are to the west of the “hole” lie along airway N571 between waypoints VAMPI and MEKAR. We also see that these captures align with the unexplained capture at 18:07:06, suggesting this may represent the path of a second aircraft.

We can also see that the paths of the two aircraft cross in the “hole”. Perhaps the gap in radar coverage is due to the radar’s inability to distinguish and positively identify each target due to proximity. The intersection of paths occurs near the center of this circle, although the aircraft would pass this intersection point at different times.

There are some other interesting aspects of this hypothesis about two aircraft captured by radar. The path of MH370 that is shown in the figure corresponds to a track of about 296T, which is roughly parallel to N571 after MEKAR, yet with an offset to the right from N571 that produces a good match with both the BTO and the BFO data at about 495 kn. So, this theory allows us to match the BFO and BTO data without invoking a precisely timed lateral offset manoeuver. MH370’s path is already spaced to the right of N571.

Another interesting aspect occurs when we extrapolate the path of the second aircraft backwards in time. Doing so, we find that the path crosses a point just to the south of Runway 18 of Butterworth Air Field near Penang. This raises the possibility that the second aircraft departed from from Butterworth and chased MH370.  Richard Godfrey discovered that this point to the south of Butterworth falls very close to waypoint UPTOP.

 We can estimate the speed of the second aircraft by considering the timing and positions of the radar captures that have timestamps of 18:07:06 and 18:13:30, which correspond to the unexplained radar point to the east of the “hole” and another capture after waypoint VAMPI to the west of the “hole”. The distance is about 77 NM, which would mean the second aircraft was flying with a groundspeed of about 722 kn. Assuming a temperature offset from standard conditions of about 14K, this corresponds to Mach 1.07 at sea-level, and Mach 1.21 at FL350. This tells us the second aircraft was not a civilian aircraft. The speeds, although fast, are certainly within the speed capability of modern military fighter jets, including one of the Boeing F/A-18D Hornets based at Butterworth Air Field.

We can also use the timestamps of the targets to the west of the “hole” to determine that the groundspeed of the second aircraft was around 500 kn after it caught MH370. This suggests that the two plane were flying roughly side-by-side, albeit with an estimated lateral separation of about 18 NM.

If this hypothesis considered here is true, it would answer some important questions about the radar data. But it would also raise even more questions about how Malaysia responded to MH370 after it disappeared from civilian radar screens and flew back across the Malay peninsula and above the Malacca Strait. If the theory is correct, it also would raise important questions about why Malaysia chose to keep this high-speed chase a secret.

169 Responses to “Radar Maybe Captured Fighter Jet Chasing MH370”

  1. Brian Anderson says:


    I don’t have a “best estimate”. The data I was working with was that which was current in March 2015. There are too many unknowns to be precise, e.g. altitude at 00:11, speed at 00:11, fuel burn for the left engine after the right engine fails etc.

    It is possible to make some estimates for these things, and draw some possible conclusions, and also perhaps eliminate some possibilities. I think it is possible that the right engine failed as early as 00:09:30, but then the speed reduction would make it difficult [maybe impossible] to reach the 7th arc continuing in a straight line on the current track. If the assumption is a slower speed at 00:11 then I think one might conclude that the right engine must have failed later.

    The BFO at 00:11 suggests a slight descent was in progress at that time.

  2. Don Thompson says:

    RMAF has only 8 F/A-18D (twin seat) Hornets comprising 18 Squadron based at Butterworth. Not the ideal air interceptor, they might be stretched to make a supersonic dash to 250nm out over Str of Malacca & ensure sufficient fuel to return without AAR support. At some time between 2011 (contract announcement) and 2015 (completion) the aircraft & squadron underwent an upgrade program 80% of which was undertaken at Boeing, St Louis.

    RMAF’s Quick Reaction Alert capability (if it really has such a thing) has been described as provided by the Mig-29 equipped squadron at Kuantan AB.

    I will continue to seek plausible reasons for what radar artifact might result in the 18:07:06 target.

  3. PeterC says:

    Given the close proximity of housing to the Butterworth runway, if an FA-18 had have launched in the early hours, I think we would have heard about it long before now. Secondly, H2 was adamant in the ABC 4 Corners interview that they did not send up an interceptor. That segment was quite comical actually.

  4. MH says:

    What might not be an intercept but more of a planned escort of possibly MH370 and/or other aircraft involved in it’s disappearance.

  5. Gysbreght says:

    Just a left-over from the previous thread:

    DrB: ”What I really want to know is what the left engine FF goes to when the right engine shuts down if it were previously at Best Holding.”

    Andrew already replied to that question with a guess for the fuel flow at max CLB thrust at FL350 of “somewhere in the range 3500-4000 kg/hr”. My guess would have been slightly higher but I think Andrew’s guess is quite reasonable.

    BTW it is perhaps superfluous to explain that the thrust available from the engine at a certain rating, and the thrust required by the airplane to maintain level flight at a certain speed, are independent variables. For example, the blank fields in the FCOM PI.23 Engine INOP Holding table indicate weight/altitude conditions where the required thrust is greater than the available Max Continuous thrust.

  6. TBill says:

    The other aircraft we think was around was the Singapore AWACS type aircraft.

  7. TBill says:

    PS- I am giving that nice smooth black MH370 flight path the eagle eye…I do not think that path shows the same exact shape that the ATSB shows up to the 1802 point.

  8. AlexSiew says:

    Victor, I apologize for posting again. Please be advised as follows:

    1. Butterworths (Penang mainland) is the headquarters of the Five Power Defence Arrangement (FPDA) comprising the nations of Malaysia, Singapore, New Zealand, Australia and the United Kingdom. This FPDA operates an air defence system for the defence of Malaysia and Singapore called the Integrated Air Defence System (IADS) or something like that.

    2. For political reasons, the chief of the FPDA/IADS base at Butterworths has always been an Australian (Singapore cannot be seen to have a Malaysian having command over them and vice versa and both nations cannot be seen to have a Brit, their former colonial master, commanding over them). At all times, the five nations would have their respective various defence personnel stationed at the base at Butterworths. Angus Houston was once stationed there, as the head, in 1999-2000 from memory.

    3. So on the night in question, March 8, 2014, the Butterworths FPDA/IADS base was commandered by an Australian (an Air Vice-Marshal), with its ranks drawn from the 5 nations.

    4. The radar returns from this Butterworths base are piped in real time to the defence centers in Australia where the returns are monitored, according to reports in the Australian press at the time MH370 went missing.

  9. Victor Iannello says:

    @TBill: A G550 can’t fly at 722 kn.

    As for the radar path up to the 18:02 point, there are several versions of that path that are similar but not exactly the same, e.g., the Lido Hotel image, Fig 2 from the ATSB report, the integrated speed/track data from the DSTG report, and the path shown in the RMP report. Since the post discusses the Lido Hotel radar captures, and in particular the unexplained capture, the path shown approximately follows the captures shown in this image.

  10. Paul Smithson says:

    Some views on what the radar data tells us is this:-

    1. There is very clear discontinuity in ATSB/DSTG depictions of the radar track before/after Pulau Perak. Pre PP it is “wobbly”, post PP it is “straight”. Either that depiction is an extrapolated path or (if LIDO slide is to be believed) it was captured by a different radar that had less “noise” in its range estimates.

    2. The path depicted in the LIDO slide and its extrapolation back towards Penang look to me very much like a “straight line” (on track ~290) – at least from 1822 back to 99.7E. I don’t see any evidence that it shows an aircraft “joining N571”. Where is the distinct turn at VAMPI from ~290 to ~296? There might be other reasons for believing that the aircraft followed N571 but it does not seem to me that the radar data – such as it is – substantiates this.

    3. The turn at Penang likely occurred a bit closer to the Island than depicted by the ATSB/DSTG/RMP. This is because the RMP annotations show us that altitude was assumed FL447. This slant range would produce a correspondingly larger horizontal range, and the distortion should be greatest the more “abeam” of Western Hill that we are. Note that the Fariq phone capture position 175227 sits some 1.5NM closer to Penang than the adjacent radar positions. If you assume a ground speed for the aircraft and assume that the “radial” angles for RMP 175159, 175259, 175359 are correct, you can work out what the slant range (and corresponding horizontal range for altitude FL340-350) ought to be for the plane to traverse from one to the next in the 60s available. This also indicates that true positions lay ~1NM closer to WH than the positions attributed.

    4. With very few exceptions, notably close to Kota Bharu, the time/positions (from 1.1E, 1.1F, RMP, LIDO) are consistent with one another and consistent with attribution to a single track/aircraft flying at a “reasonable speed” (my estimate agrees with VI’s M0.84).

    5. As for the “second plane” conjecture, you can see outlier positions on the LIDO slide both north and south of the “principle track” depicted by the highest density of captures (dotted red). I don’t see any reason to seize on one or more outliers and suggest that this was a second, separate radar target.

  11. Brock McEwen says:

    @Brian: thanks very much. That’s in line with my model’s indications.

  12. Brock McEwen says:

    @all – this post is very different from my usual fare. It is the only time I will ever do this, so I beg the indulgence of Victor, and of the group.

    Below please find a link to access my gift to MH370 next of kin, to commemorate the 3rd anniversary of their suffering. Please share it as widely as you wish. I request especially that you pass this on to any next of kin of those on board MH370 with whom you are in contact, to ensure it reaches as many as possible.

    It is a poem: in latin, in english, and in song. I hope it breaks through barriers of language and culture, to express how deeply their pain has touched us, and how resolutely we stand with them.

    We will not rest until they are at peace.


  13. Ge Rijn says:


    Coming back on the celltower that captured the co-pilot’s phone.
    The RMP states the detection was made by Sector 2 of BBFARLIM2 base station which is located in Bandar Baru Air Itam.
    This township is surrounded by ~200 meter hills to the east and the south and up to ~700 meter to the west.
    I thought it was concluded back then the celltower who detected the call was based on the superstore in this township.

    Like to mention also the RMP states that in various test flights following the route south around Penang several cellphone connections were made but none by BBFARIM2 base station:


    I sure would like to know the location Don Thomson has identified the celltower to be where there are no hills that could limit/block it’s line of sight.

  14. Richard says:


    Many thanks for pointing out the Radar Data Point Anomaly at 18:07:06 UTC (02:07:06 MYT) and the fact that this Radar Data Point aligns with a direct track from waypoint UPTOP just south of Butterworth Runway 18 and with the waypoint VAMPI (please see link to map below). I estimate the position at 18:07:06 UTC (02:07:06 MYT) as 5.917237°N 98.721202°E. The estimated average ground speed from 18:07:06 UTC to 18:13:36 UTC (02:13:36 MYT) at 6.196413°N 97.530141°E is 674 knots (135.3 km or 73.056 NM in 390 seconds).

    I agree with your hypothesis, that if these 2 points were the same aircraft, then this was not a commercial jet but a military jet, as the speed was too high.

    The Radar Data Point Anomaly is both significantly out of line in position and out of sequence in time to the other Radar Data Points.

    The estimated average ground speed from 18:13:36 UTC to 18:22:12 UTC is 511 knots (138.3 km or 74.676 NM in 516 seconds). This fits with the estimated speed of MH370 at this stage and a military aircraft shadowing MH370.

    The track back from waypoint VAMPI to the Radar Data Point Anomaly at 18:07:06 UTC to waypoint UPTOP at Butterworth is marked with a red dotted line in the link below. A Hornet fighter jet scrambled as MH370 passed south of Penang Island at around 17:52:27 UTC with a top speed of 1,034 knots, could easily intercept MH370 at waypoint VAMPI. If the fighter jet shadowed MH370 from waypoint VAMPI, then this fits the average ground speed of 511 knots from 18:13:36 UTC up to 18:22:12 UTC.

    As you hypothesise, the Beijing Lido slide is possibly a combination of the real MH370 flight path up to the blank white circle followed by the fighter jet flight path shadowing MH370 beyond the white circle. The reason the white circle exists, may simply be because the radar data points would reveal that there were 2 aircraft involved at around the same time in this area.


  15. Richard says:

    CORRECTION: I estimate the position at 18:07:06 UTC (02:07:06 MYT) as 5.883851°N 98.721202°E. The estimated average ground speed from 18:07:06 UTC to 18:13:36 UTC (02:13:36 MYT) at 6.196413°N 97.530141°E is 679 knots (136.2 km or 73.542 NM in 390 seconds).

  16. Richard says:

    The Radar Data Point Anomaly at 18:07:06 UTC (02:07:06 MYT) is 40.51km or 21.875 NM from the Radar Data Point at 18:07:16 UTC (02:07:16 MYT) positioned at the eastern edge of the white circle.

    They cannot be the same aircraft, as these 2 points are only 10 seconds apart.

    At a closing speed of 168 knots (679 knots – 511 knots), this would mean an intercept in 7 mins 49 secs at around 18:14:55 UTC (02:14:55 MYT) just beyond waypoint VAMPI.

  17. Richard says:

    Here are 2 close ups of the Radar Data Point Anomaly at 18:07:06 UTC (02:07:06 MYT) and the Radar Data Point at 18:07:16 UTC (02:07:16 MYT).

    The first is from a Reuters high resolution picture of the Beijing Lido slide.

    The second is a screen shot from CCTV China Central Television film of the Beijing Lido Slide (Don Thompson kindly provided the link:


  18. Gysbreght says:

    If part of the primary radar track shows a chase airplane, where is the track of the airplane being chased?

  19. Victor Iannello says:

    @Ge Rijn said, “I sure would like to know the location Don Thomson has identified the celltower to be where there are no hills that could limit/block it’s line of sight.”

    In a previous response to you, I gave the exact coordinates of the tower atop the St John Ambulance building as well as the coordinates of MH370 at the time of the cell phone connect, and encouraged you to plot both in GE so you could see there were no hills to obstruct the path. I guess you didn’t.

  20. Victor Iannello says:

    @Paul Smithson said, “As for the “second plane” conjecture, you can see outlier positions on the LIDO slide both north and south of the “principle track” depicted by the highest density of captures (dotted red). I don’t see any reason to seize on one or more outliers and suggest that this was a second, separate radar target.”

    I don’t see any outliers that differ in time and position nearly as much as the one I highlight. Do you have a graphic to illustrate your point?

  21. Andrew says:


    Returning to the questions you posted in the previous thread, some of which may already have been answered:

    “Andrew, please enlighten me as to why you think the FF would be lower than 4750. I’m not understanding the cause.”

    The fuel flow decreases at higher altitudes due to the lower air density. A jet engine compressor running at constant RPM pumps a constant volume of air into the engine. At higher altitudes, the lower air density results in a lower mass flow of air through the engine and, consequently, less thrust. The fuel flow also reduces to keep the fuel/air ratio within the narrow range where the engine can operate efficiently.

    After an engine failure in the cruise, the remaining engine will only operate at maximum continuous thrust (CON) if the pilot executes a drift down or selects CON thrust via the Mode Control Panel. In the unresponsive pilot scenario, the engine will be limited to climb thrust (CLB). As I said previously, I think the fuel flow at FL350 will be in the range 3500-4000 kg/hr.

    “How long in advance of reaching the Holding pattern fix (at ANOKO) would the FMC (a) begin to reduce air speed from LRC to Best Holding speed, and (b) display the End of Offset error? I am guessing about 2 minutes for each.”

    The END OF OFFSET message is generated two minutes prior to the offset termination point, ie abeam ANOKO. The manuals don’t specify an exact time for the commencement of the deceleration segment prior to the holding fix, but two minutes seems to be a reasonable approximation.

    “Can anyone see anything wrong in my description of the sequence of pilot actions that would produce simultaneous End of Offset and End of Route errors at ~18:44 at a fix ~10 NM to the right (west) of ANOKO?”

    The description seems reasonable, although I question the assumption that the aircraft will maintain a constant true heading after the end of route, regardless of the HDG REF switch selection. The B777 FCOM states that the HDG REF switch “selects the heading reference for the PFDs, NDs, AFDS and FMCs”. That statement seems to indicate the position of the switch determines the heading reference used by the FMCs and AFDS. Given that the switch is usually selected to NORM except when operating in polar regions, the heading reference would be magnetic rather than true at the latitudes where MH370 was operating.

    Regarding the 12nm offset, if an aircraft operating in RVSM airspace needs to deviate off track in an oceanic region and can’t obtain an ATC clearance, the standard offset that should be flown is 15nm. Does that make any difference to your calculations?

    “What I really want to know is what the left engine FF goes to when the right engine shuts down if it were previously at Best Holding. I suppose the FF might also change as the aircraft slows down to the speed where it descends. I’m not sure if that descent begins before the left engine flames out ~9 minutes later. Does anyone have a guess as to the rate of change of air speed while in INOP and before a descent begins? Right now I’m using 5 KTAS/min while in INOP and 19 KTAS/min after the left engine quits.”

    I suspect the fuel flow on the left engine will be approximately 3500-4000 kg/hr, as stated previously. That won’t change much as the aircraft reduces speed. The speed reduction of 19 knots/minute mentioned by Brian Anderson seems reasonable. At that rate, the aircraft would start descending a little over three minutes following the failure of the right engine, assuming the aircraft was maintaining the best hold speed of 239 knots before the engine failure.

  22. Ge Rijn says:


    I did look for the StJohs ambulance building and its not in the township of Bandar Baru Air Itam but further north above the hill range that encloses Bandar Baru Air Itam to the east,west and south:


    The @sk999 MH370 5.220/100.289 coördinates I find about ~10km east-east-south from Butterworth above the Penang Plateau.
    Drawing a line just south of Kota Bharu over this Penang Plateau coördinates brings MH370 over Penang Island main land just south of Georgetown.

    I argue if the cellphone was only picked up by a celltower in an eastern township of Georgetown (more even if it was located on the StJohns ambulance building) MH370 did not flew south around Penang Island over sea but crossed the Island over land south of Georgetown.

    South of Georgetown there were no cellphone registrations if we believe the RMP.
    This only happened in the flight tests afterwards following the asummed radar track flight path.
    In those tests no registrations were made by BBFARLIM2 base station.
    This is telling IMO for it also indicates MH370 did not fly south around Penang Island but crossed it over land a bit south from Georgetown.

    Maybe the second radar track after VAMPI is MH370 and the first track before the gap is another plane?

  23. Victor Iannello says:

    @Andrew: To get you up to speed a bit, the behavior of the A/P when in LNAV mode and a route discontinuity is reached has been debated for years now. We know the plane remains in LNAV mode, but we question true vs magnetic and heading vs track. The only real way to discern constant heading from constant track is to measure the response of the plane as the wind is varied and see whether the track or heading remains constant. The only way to discern magnetic (NORM) from true is to measure the response of the plane as the magnetic variation changes and see if the track changes with changes in magnetic variation. Nobody that I know has done this experiment either in a Level D simulator or a real 777. And while the FCOM says one thing (constant heading), the Honeywell manual for the Boeing 777 FMS says another (constant track).

    I did do a series of experiments using the FSX PMDG 777 simulator, and determined that upon a route discontinuity, a constant true heading is maintained. The NORM/TRU switch only changes how the track is displayed. But of course, there are no guarantees that the PMDG model behaves like the true FMS. (The model does do surprisingly well at predicting other behaviors, such as what happens when a route discontinuity occurs at a hold fix while flying with an offset of the last leg of a route.)

    I can’t imagine that any 777 pilot would fly a plane in LNAV mode after a route discontinuity because it puts the plane in an indefinite state. You’d either enter additional waypoints to the route, or navigate using the HDG/TRK SEL roll mode.

    So unless we hear from somebody that has done the proper experiments in a qualified simulator, or from somebody within Honeywell that has intimate knowledge of the exact behavior, I think this issue will remain open.

  24. Paul Smithson says:

    @VI. The other outliers are shown clearly in your (Holland) graphic. It is the fact that they are outliers that permits the timestamps to be read since they don’t overlap with their neighbours. Examples (crosses in blue) are: 020233, 020123, 02171?, 0214?6, 02133. Or am I misunderstanding and the “blue” crosses are not radar captures?

    As mentioned in my previous post, if you ignore N571, look only at the concentration of radar fixes and fit a straight line to them, you will find a very nice straight line fit at around 290.3T.

  25. Victor Iannello says:

    @Andrew said, “Regarding the 12nm offset, if an aircraft operating in RVSM airspace needs to deviate off track in an oceanic region and can’t obtain an ATC clearance, the standard offset that should be flown is 15nm. Does that make any difference to your calculations?”

    That is very helpful. Thank you!

  26. Victor Iannello says:

    @Paul Smithson: The blue crosses are not radar captures. They are used to relate the actual captures to the indicated timescale.

  27. Paul Smithson says:

    @VI. No doubt you are right. But are you sure about that? Why (if they are simply there to relate captures to time-scale) are they positioned all over the place in relation to the centreline of the time scale?

  28. Victor Iannello says:

    @Ge Rijn: Let’s allow Don Thompson explain why he believes the St John Ambulance building was the location of the tower.

    Regarding the experiments performed to measure cell phone connects at altitude, they are inconclusive at best. The plane was at the wrong altitude, the wrong speed, and did not fly the correct route. As for whether there are any other cell phone registrations from MH370, I believe only crew phone numbers were checked. The passenger phone numbers were never collected by Malaysian investigators, according to one family member of a passenger.

  29. Victor Iannello says:

    @Paul Smithson: I don’t know why Bill chose to position the blue crosses exactly where he did. He probably thought it wasn’t important. But if you look at the high resolution versions of the Lido Hotel image, there are no targets that correspond to the position of the blue crosses, nor are there timestamps next to them.

  30. Mick Gilbert says:


    That is most certainly an interesting interpretation of the Lido Hotel slide. A few observations:

    Only 32 minutes elapsed between the turn back and when MH370 arrived to the south of Penang and there were only about 12 minutes between when Ho Chi Minh ATCC and Kuala Lumpur ATCC jointly realised that MH370 was not where it was meant to be and when it reached Penang.

    As Don has noted, the RMAF do not maintain a QRA capability at RMAF Butterworth. Even with a practiced QRA capability, getting a fighter away inside of 15 minutes from an alert is no mean feat. Without any sort of alert status or crew readiness, launching a fighter within 30 minutes of an alert on a Saturday morning at 2am would be extraordinarily difficult.

    The mechanics of launching and chasing high subsonic (M0.8+) targets are brutal. The F/A-18D Hornet can not achieve Mach 1+ with external stores other than the wingtip AIM-9 Sidewinders; that means no external tanks. For a Hornet, supersonic flight requires afterburner. With maximum internal fuel the Hornet can only sustain afterburner for about 14 minutes in total; that means a 9-10 minute supersonic chase, bingo fuel and a subsonic cruise home. It’s these sort of numbers that previously gave AWACS airplanes like the E-3 sitting on the edge of air engagements a lot of comfort; if their screening fighters ever looked like getting swamped they would simply turn and run and the bad guys wouldn’t have enough fuel to chase them down. It’s also the reason that the super-cruise capability (supersonic cruise without afterburner) of airplanes like the F-22 Raptor and the Eurofighter Typhoon makes them so potent.

    A sonic boom over or near Penang at 2am on a Saturday morning is not likely to have gone unnoticed.

    It’s difficult to imagine a circumstance where a military radar would lose contact with a B777, a target 15% larger than a B-52, but continue to track a fighter.

    Stupid is as stupid does but I can’t see even the RMAF as being stupid enough to present a slide evidencing a military intercept of a civilian target to grieving next-of-kin but, never say never.

  31. Victor Iannello says:

    @Richard: Thank you for your additional analysis and insights. The possibility of two planes is just a theory, and maybe a crazy one at that, but I felt it might be helpful to consider, even if it leads nowhere.

  32. Paul Smithson says:

    Thanks @VI. I’ll go back and have a look at that hi-res. I remain, however, of the view that the captures depicted fit nicely to a straight line without turning on to N571… It is strange how the eye is drawn to a certain interpretation of some noisy line/points as soon as some hypothetical lines are superimposed… Change the lines and the interpretation also changes.

  33. Victor Iannello says:

    @Mick Gilbert: Depending on when the fighter took off from Butterworth, supersonic flight might not have been necessary except for a short time (less than 5 minutes), and not over Penang. I estimated the average speed over a 6.5 minute period to be around Mach 1.07 at sea-level.

    To answer the question posed by you and @Gysbreght, I don’t think the Lido Hotel slide shows all radar captures over roughly 30 minutes. Rather, captures are grouped by what the radar system believes were the same target. So, I consider the possibility that the association was made incorrectly somewhere in the “hole”.

    If a fighter jet did follow MH370, it was probably not based on an alert after civilian ATC lost it for the reasons you mentioned. That doesn’t mean it didn’t occur. It just means there would have been earlier knowledge than we are led to believe.

    The RMAF might have shown the NOK what radar data that it thought was only MH370, and never intended to show the fighter jet. It might have later realized that targets were mixed. That would also explain why Malaysia won’t officially recognize this slide even after it was publicly displayed to both NOK and reporters. It would also help explain why the 10-second radar data supplied to the ATSB ends at 18:02 except for one sole point at 18:22.

    Again, I know the theory relies a lot on a single unexplained radar capture, and the theory could be wrong. If true, it raises many questions. But it also does tie together a lot of loose ends, including how to connect all the radar data with the satellite data without invoking a precisely-timed lateral offset. I haven’t seen another theory do that.

  34. Victor Iannello says:

    @Paul Smithson: Perhaps you should try regressively fitting the points to a line. That would remove any operator bias.

  35. sinux says:


    Woah! That’s the most thought provoking theory regarding MH370 I’ve read in a while. Potentially game changer. But before we get our hopes up :

    Is there any scientific/physical/technical reason why the 02:07:06 point was assumed to be a false target?

    Is there any technical reason for the apparent loss of density of points the closer the two tracks get? (like two out of phase returns that cancel out at the receiving end of the radar, if that’s even possible?)

    The absence of returns in the white circle was posited to exist because of occlusion by a hill, was it just a theory or has it been confirmed?

    And finally, how do you explain that the MH370 track is not seen on the radar west of the circle?

  36. Richard says:


    “If part of the primary radar track shows a chase airplane, where is the track of the airplane being chased?”

    My suspicion is the Beijing Lido slide is a composite including a standard background aeronautical chart overlaid with selected radar data points.

    The Malaysians did not want to show that there were 2 aircraft, so there are only selected radar data points.

    We do not see EK343 either.

    That is probably why there is a large blank white circle or hole. The radar data here would reveal that there were 2 aircraft and the 2 traces were so close together, it was difficult in the hurry to produce a slide for Beijing to separate out the 2 aircraft.

    In fact, they missed one of the radar data points that should not have been selected at 02:07:06 MYT.

    If the point at 02:07:06 MYT and 02:07:16 MYT are the same aircraft 10 seconds and 40.51 km apart, then the speed of that aircraft would be 7,875 knots!

  37. Victor Iannello says:

    @Richard. All possible.

  38. Victor Iannello says:

    @Alex Siew said, “The radar returns from this Butterworths base are piped in real time to the defence centers in Australia where the returns are monitored, according to reports in the Australian press at the time MH370 went missing.”

    Irrespective of press reports at the time of the disappearance (many of which we know were false), is anybody aware of whether this is true?

  39. Paul Smithson says:

    @VI – that is what I have done – albeit not with calculated RMS cross-track error but rather seeking the straight line that a) leaves errors on both sides b) doesn’t systematically “lean” one way or the other (from errors all on one side to errors all on the other).

  40. Victor Iannello says:

    @Paul Smithson: If you are willing, please share your results.

  41. Victor Iannello says:

    @sinux: Welcome to the discussion.

    To answer your questions, the capture at 18:07:06 was never officially discussed. In fact, the entire Lido Hotel image has never been officially acknowledged despite the fact it was shown to the NOK and reporters. Regarding the spacing of the data points near the white “hole”, there are previous theories that consider terrain masking, altitude changes, range windowing, and antenna radiation patterns, but nobody really knows. If the radar source was Western Hill on Penang Island, there should have been no terrain masking. As to why MH370 was not captured after the “hole”, it may have been and not shown in the slide due to confusion about which captures were which plane.

  42. Paul Smithson says:

    Sure, VI – I’ll do that. First I have to “declutter” my GE since it has so much crap on it!

  43. TBill says:

    Re: other air traffic
    If you look at the FR24 YouTube recording of MH370 to IGARI, there is other commercial air traffic heading up the Straights, north of N571. Another thing that @BuyerNinety and I have talked about is that SIA68 does not appear on the EK343 screen grab. We think that is because, comparing to KLedgerwoods SIA68 screen grabs, SIA68 must have been just slightly off-screen to the north. Because FR24 does not show flight sky-vectors, I am confused if the Singapore flights are allowed to cross into that white circle/restricted area, but to the eye it looks like they pass near there.

  44. Victor Iannello says:

    @TBill: Thank you. I think we have broad agreement that the data represented on the Lido Hotel slide was intended to show only a single aircraft, independent of the presence of other aircraft during that 30-minute period.

  45. DrB says:


    Thank you for responding to my many questions. I shall work on the offset to see if I can get 15 NM to fit the satellite data. Likewise I will find the impact of the INOP Fuel Flow on the PDA.

  46. Victor Iannello says:

    This was article about Blaine Gibson was recently published by Australia’s The Courier Mail.

    We owe a lot to Blaine, who took it upon himself to find aircraft debris despite the indifference shown by Malaysian officials. It is absolutely disgusting that anybody would threaten him because of his theories about the disappearance. His demonstrated ability to find debris multiple times should have squelched any naysayers. (Any doubts I had about the authenticity of his finds disappeared after he repeated his feats.)

    Further search blow for flight MH370 as ‘Indiana Jones crusader calls it quits

    THE AMERICAN adventurer who devoted years of his life to solving the MH370 mystery has called off his one-man mission after receiving death threats and being ignored by search officials.

    US lawyer and blogger Blaine Gibson found over two dozen pieces of debris in Mozambique and Madagascar, several items of which have been confirmed as coming from the missing Boeing 777.

    But just days out from the third anniversary of the doomed Malaysia Airlines’ flight, Mr Gibson said the personal cost of his self-funded crusade was too great for him to continue.

    Relatives of those on board have praised Mr Gibson’s efforts. However the modern day Indiana Jones has been the subject of anonymous death threats over his defence of MH370 pilot Zaharie Ahmad Shah.

    Other self-styled “investigators” have even suggested Mr Gibson “planted” debris; an allegation he finds offensive and defamatory.

    He also revealed his deep frustration and disappointment that evidence gleaned from the debris he and others had found, was not considered enough to warrant extending the official search.

    “They’re saying unless they know the location of the plane they’re not going to look for it,” said Mr Gibson.

    “That’s like saying they’re not going to look for the plane until someone else finds it. It’s comical and it’s tragic.”

    He said he believed there was enough new evidence for the search to continue in the new 25,000 square kilometre area identified by experts weeks before the official operation was suspended.

    “When the CSIRO came out with those (drift modelling) recommendations it felt so good because it seemed all that I had done was worthwhile,” he said.

    “The scientists had factored the debris in, they’d identified a new search and I really thought they were going to look there because the ships were still out in the water.

    “All they had to do was just come up with some extra money and go look there and they didn’t do it, and I was very disappointed.”

    After overseeing the complex and costly two-and-a-half year search for MH370, the Australian Transport Safety Bureau is now in the process of preparing a report on the operation.

    An ATSB spokesman said responsibility for the investigation remained with Malaysian authorities, who will release their own technical report before the year’s end.

    On Saturday, members of the Voice 370 group made up of relatives of the 239 people on board the Malaysia Airlines’ flight will launch a “crowd funding” initiative to raise money for a new search.

    The group also wants the private sector including airlines and aircraft manufacturers to step up and ensure the mystery does not remain unsolved.

    It will be three years on Wednesday that MH370 left Kuala Lumpur to fly to Beijing before suddenly veering off course, and heading towards the Southern Indian Ocean.

  47. Victor Iannello says:

    @DrB: A 15 NM lateral offset manoeuver will make it even harder to precisely fit in the time window between 18:22:12 and 18:25:27 unless you allow it to end a bit later.

  48. AlexSiew says:

    1. I agree with Richard: “…the Beijing Lido slide is a composite including a standard background aeronautical chart overlaid with selected radar data points…”.

    2. The question is who prepared this slide? The Malaysians or someone else?

    3. The Chinese CCTV link of the March 21, 2014 briefing at the Beijing Hotel where this slide was shown, provided by Richard courtesy of Don, has confirmed my long held suspicions that this slide was prepared by the international team that briefed the Malaysian PM before the PM made his speech on March 15, 2014 declaring, among other things, the 2 arcs (northern and southern) and that the last satellite signal was at 8.11am. http://www.malaysiakini.com/news/257165

    4. One of the other slides shown at this briefing on March 21st contained a statement that the military radar data had been shown to the Chinese team on March 17, 2014.

    5. Ackbal, the briefing officer can be heard on the video talking about the 6 handshakes.

    6. On March 16, 2014, Hishamuddin also spoke about the 6 handshakes. http://www.mh370.gov.my/phocadownload/Transkrip%2016%20Mac%202014.pdf

    7. On March 19, 2014, Hishamuddin expressly said they had received some radar data (from another country) and that the Malaysians would be sending a high level team led by Ackbal to brief the NOK at Beijing (which briefing took place on March 21, 2014 as recorded by the CCTV link). https://www.facebook.com/mh370tragedi2014/posts/228596664011829:0

    8. All indications are that the materials/slides presented at the March 21, 2014 briefing came from the international team (comprising of the FAA, NTSB, AAIB and Inmarsat) and had been given to the Malaysians on or around March 15, 2014.

    9. Malaysian own military data was different from what was shown on the Lido slide and was described in some detail at the press conference on March 12, 2014 where the Air Force chief Daud said the last blip was at 2.15am 200 miles off Penang, Zulkifli the army chief said they saw a blip near where MH370 had disappeared from and that the civil radar over at Kota Bahru also showed some blips and where the head of DCA Azharuddin said the plane was still posting at 1.21am but disappeared altogether at 1.30am. (see the link sk999 provided in a previous thread).

    10. The sporadic blips seen on Malaysian primary radar could merely have been caused by clouds as the weather at both Kota Bahru and Penang for the period in question was described as ‘mostly cloudy’. See the historical weather records at Weather Underground and the satellite photos on Tim Vasquez’ site.

    11. Finally, Appendix J-1A of Folder 4 (SKMM Analysis) of the ‘RMP Report’ gave a time of 1.41am for the purported turnback over at the South China Sea.

  49. Ge Rijn says:


    Can you explain why the unexplained 020706 radar point was changed from the deleted 020436 notation under it?
    I see in the enlarged Lido-picture the original underlying radar time notation is quite faint but it does not look clearly ‘020706’ to me.
    It makes me doubt the 020706 notation is correct.

  50. Victor Iannello says:

    @Ge Rijn: First, to be clear, I didn’t change the time notation. I believe that was done by Bill Holland, although I can’t be sure because I used the image found in a paper by Ron Belt.

    Here is another version of the image that was color-enhanced by airliner.net contributor BackSeater. Here is his comment.

    To me, it is very likely that 020706 is correct.

  51. Victor Iannello says:

    @PeterC said, “Given the close proximity of housing to the Butterworth runway, if an FA-18 had have launched in the early hours, I think we would have heard about it long before now. Secondly, H2 was adamant in the ABC 4 Corners interview that they did not send up an interceptor. That segment was quite comical actually.”

    Yes, H2O strongly defended the decision to not send an intercept, and (out of the blue) defended the decision to not shoot down the plane, saying that the United States would have. But considering how many times official statements from the Malaysian officials were later discovered to be false or misleading, I wouldn’t put much faith in his statement. In fact, his mannerisms and statements during that interview are so evasive and defensive that he appears to be hiding something.

  52. Andrew says:

    @Alex Siew:

    Regarding JORN:

    “JORN does not operate on a 24 hour basis except during military contingencies. Defence’s peacetime use of JORN focuses on those objects that the system has been designed to detect, thus ensuring efficient use of resources. ” https://www.airforce.gov.au/docs/JORN_Fact_Sheet.pdf

    “Based on the time of day that MH370 disappeared, and in the context of peacetime tasking, JORN was not operational at the time of the aircraft’s disappearance. Given range from individual OTHRs, the ionospheric conditions and a lack of information on MH370’s possible flight path towards Australia, it is unlikely that MH370 would have been detected if the system had been operational.” https://www.airforce.gov.au/docs/JORN-FAQS.pdf

  53. Andrew says:


    Thank you for the information about your experiments regarding the autopilot behaviour following a route discontinuity. I will try to source a definitive answer for you. It may take some time!

  54. TBill says:

    Re: EK343 and SIA68 FR24 Screenshot
    Following up on Alex Siew’s comments about clouds, I bumped into a new FR24 screenshot a little before 1822 (apparently) showing EK343 and SIA68 at a slightly earlier point in time:


  55. Victor Iannello says:

    @Andrew: Thank you for trying to get a definitive answer. It’s something we have been debating for years now.

  56. AlexSiew says:


    1. The ‘pamphlets’ in your link were produced post-MH370 and constitute proof positive of official Australian complicity in this cover-up going on.

    2. JORN is operated on a 24 hour basis and has always been operating on a 24 hour basis. One does not spend billions on a defence radar system and then decide to switch it on only a few hours a day to save electricity.

    3. The fundamental premise of a defence radar system is to provide all encompassing surveillance so as not to be caught out by a surprise attack by enemies. How would a defence radar system be of any use if it is on only for part of the time every day? What happens if the enemy strikes when this multi billion radar system is switched off to save a few pennies of electricity?

    4. JORN was designed not only to protect Australians but also Americans. Americans’ main global satellite and radar spying center Pine Gap, is located at Alice Springs, Australia.

    5. The Malaysians announced at 7.24AM Malaysian time, the plane’s disappearance. In Australia, 7.24AM MYT would already be 9.24AM.

    6. Even if we can say with a straight face we believe the post-MH370 blurbs about JORN not being operational on a 24 hour basis, do you think the Australians and the Americans at Pine Gap and the other joint Australian/American defense centers in Australia would not have the sense to switch on JORN on hearing a Boeing 777 had gone missing?

    7. At 7.24AM MYT, according to the official narrative MH370 was still in the air, at cruising altitude and speed, flying in the direction of Australia, heading towards west of Perth. From 7.24AM to 8.19AM the plane if the official narrative is true, would be right smack west of Perth.

    8. Those post-MH370 ‘pamphlets’ are contradicted by official Australian write-ups of JORN produced pre-MH370. As an example, the DSTG which is a department under the Australian Dept of Defence, has a write-up on JORN. Please take a look at the write-up and decide for yourself what is the truth about whether JORN was and is operational on a 24 hour basis. https://www.dst.defence.gov.au/innovation/jindalee-operational-radar-network

  57. DrB says:


    Was the air speed reduction seen in the simulator runs -19 KIAS/min or -19 KTAS/min?

  58. DrB says:


    On the default heading mode after an EOR error, I have queried ATSB. They have responded by repeating a less-than authoritative (in my opinion) answer from Honeywell. I believe that the PC-simulator result that Victor determined (after several valiant tries) is most probably correct for the B777 FMC. However, as Victor points out, it actually takes an extended experiment under the right condition of wind and location, in order to obtain a definitive result. This is not something that can be done in a 10-minute test. Others have already tried this and failed.

    I hope you can obtain a definitive answer from somebody, but brief experiences only seem to reinforce preconceived notions of how the FMC works. It requires a lengthy flight after the EOR in an area where the magnetic variation is changing fairly rapidly.

  59. Brian Anderson says:


    The IAS speeds were converted to TAS to derive the relationship.

  60. Brian Anderson says:


    The result was -19 KTAS per minute.

  61. Ge Rijn says:


    For what it’s worth here some additional pictures and comments about the Lido screenshot.
    It says someone called Bill found the time notation and translated it in 020436. It’s not changed yet in 020706 here.


  62. Richard says:

    @Ge Rijn

    Here is a link to the China Central TV Beijing Lido Slide Close Up Detail Enhanced:


    Here is a link to the Reuters Beijing Lido Slide Close Up Detail Enhanced:


    In my view, they both confirm 020706.

  63. Ge Rijn says:


    That makes three clear ‘020706’ pictures (including @VictorI’s link).
    I’m convinced now. It sure is not 020436 but 020706.
    At least we can assume we have a yet unidentified flying object here.

  64. Gysbreght says:

    An UFO?

  65. Ge Rijn says:


    I avoided to use capitals. This 020706 radar capture is the only and single capture that cannot be logically related to the sequence of all the other radar captures in the Lido-screenshot IMO.
    Compared to this sequence it shows itself as a single flying object which is yet unidentified.

  66. Mick Gilbert says:

    @Victor et al

    The 020706 plot is the second last capture before the “gap”. If the gap is the result of terrain masking (of say the RTAF Khok Muang radar) then you might expect a transitional clutter zone to exist as you approach the terrain boundary. Might something like that give rise to a refracted return or a return exhibiting an error in bearing?

  67. Victor Iannello says:

    @Ge Rijn said, “It says someone called Bill found the time notation and translated it in 020436. It’s not changed yet in 020706 here”

    That would be IG member Bill Holland, who I referenced in the post above.

  68. Ge Rijn says:

    @Mick Gilbert @others

    The Lido-screenshot is titled: ‘military radar plot from Pulau Perak to last plot at 02:22H”.
    Doesn’t this mean all the plots were made by a radar station on Pulau Perak?

    Is this radar station placed in and on top of the Pulau Perak lighthouse?
    To me it looks like it from this photo of the lighthouse:


  69. Victor Iannello says:

    @Mick Gilbert: There are many reasons for a false return, although using two different numerical tools, terrain masking at that location is not predicted for any known radar source.

  70. Andrew says:

    @Alex Siew:

    Believe what you will, but consider the following:

    1. JORN might well be capable of operating on a 24/7 basis, but the region is reasonably stable in a political sense and Australia certainly does not regard neighbouring countries as enemies. Consequently, there is currently no military threat and no need for JORN to operate 24/7.

    2. At the time of MH370’s disappearance, nobody knew the aircraft had headed south and the initial search efforts were focused on the South China Sea. It wasn’t until about a week later that investigators concluded the aircraft had diverted from its intended course and headed west across the Malaysian Peninsula, then continued on a northern or southern track for around seven hours. Given that time lag, the Australian Defence Force had no reason to suspect the aircraft might have headed anywhere near the Australian continent the night it disappeared, and hence no reason to ‘fire up’ JORN to start looking. Further, JORN does not ‘sweep’ like a conventional radar; the beam needs to be pointed towards a target area within the area of coverage. With no idea of where to start looking, JORN would be all but useless in the search for MH370.

    3. DSTG was involved in the development of JORN and certainly has reason to promote the system as providing a 24-hour military surveillance capability. However, the fact is that the system is now run by the RAAF’s 1RSU, with a relatively limited budget that does not currently support 24-hour operations due to the benign level of threat.

  71. Victor Iannello says:

    @Alex Siew: Since the disappearance, much has been written about JORN and why it did not detect MH370. And your argument that any country that has radar capability will use it 24 hours is demonstrably false. Unless you have new facts to present, let’s move on.

  72. DrB says:


    More pilot questions:

    1. If you were flying at FL350 and ECON 52 at IGARI when technical problems arose requiring a diversion, what speed and altitude would you select for the return to Malaysia?

    2. For the left turn around IGOGU from 296T to 180T (toward ANOKO), and with a 15 NM right offset, what bank angle or turn radius would the LNAV use? The latest possible turn would use the 25 degree bank angle limit which gives about 1.03 degrees per second (and 7.6 NM turn radius) at the LRC speed at FL360. However, this route flies well outside a circle around IGOGU with a 15 NM radius. I suspect the FMC actually flies the 15 NM radius circle centered on IGOGO. This only needs a 13.3 degree bank (at 0.52 deg/s) and it is a shorter, faster means of reaching the new 180T track 15 NM west of ANOKO while always maintaining a 15 NM separation from IGOGU.

  73. Ge Rijn says:

    @Mick Gilbert @others

    I just found there was no radar station at Pulau Perak in 2014.
    By now their probably is one operating there:


  74. sk999 says:

    Ge Rijn,

    Note that the new radar equipment is for the Malaysian Maritime Enforcement Agency, not the RMAF, and has a “… distance reach capability exceeding 30 nautical miles in good weather …” I presume the intent is to stop piracy in the Straits. Pretty useless for spotting unidentified aircraft.

  75. Ge Rijn says:


    It was just to be sure there was no radar station at Pulau Perak during 2014. Also a 30 miles radar would have spotted MH370 but we can savely put this option in the garbage-bin. Thank you.

    Leaves the question which radar station captured the track-points.
    Notation is all the same in format and (green) layout. Therefore it seems to come from one and the same radar screen.

    Western Hill station IMO would also show the radar track the whole way from Penang up to Pulau Perak and after. Though the first part is missing just like the ‘gap’ after Pulau Perak so it wasn’t Western Hill station IMO.

    Butterworth station could likely be hindered by the up to ~700m western hills of Penang masking this first part (gap) before the track to Pulau Perak.
    This leaves the gap after Pulau Perak yet unexplained though.

    To me it’s also unclear why the time and space between the noted radar points is increasing so much after ~020300 towards the gap while before they were so close together then completely vanishing at 020716.

    Reappearing at 021336 asif nothing happened continuing with close radar captures.
    Was there maybe a steep descend after ~020300 that again masked the Butterworth radar afterwards till 021336?

  76. TBill says:

    I know you have pretty much seen all the MH370 news articles, but pls check this one out just in case.

    Related to this thread, it gives an interesting account of what “reportedly” did happen that night at the radar bases. Basically saying the Malaysian military radar operators missed MH370 in real time, but within hours and after realizing MH370 was missing, they knew they had recorded an unknown aircraft that could have been MH370.

    The comment section has a translation of a purported Z FaceBook page, that I have not seen before (and cannot attest to it’s authenticity). But it is interesting too.


  77. Victor Iannello says:

    @TBill: Thank you for pointing to the article. The story has changed so many times it’s hard to know what to believe.

  78. buyerninety says:

    Readers may find some of the information (bank angle, FMC Winds, etc.)
    in this document, which was written for 777-300ER training, to be of interest.

  79. Brock McEwen says:

    @Victor: re: JORN:

    I am unaware of anyone trying to make the argument you just refuted. It is conceded there are SOME countries in this world who might turn their radar off occasionally.

    But that is not the topic at hand. The topic at hand is whether “Australia…the world leader in high-frequency, skywave over-the-horizon radar (OTHR) technology”

    …who was justifiably proud of its…

    “first comprehensive land and air early warning system (that) not only provides a 24-hour military surveillance of the northern and western approaches to Australia, but also…assists in weather forecasting making it possible to…give early warnings of cyclones and enable ships to save fuel by avoiding rough weather”

    …would either turn it off at night, or be insufficiently equipped to detect a jumbo jet zinging well inside its range. For HOURS.

    Looking at the preponderance of the evidence, it makes far more sense to suppose the after-the-event statements of impotence were designed either

    a) to give as little information as possible about capability, or
    b) to obscure the facts surrounding the curious case of MH370

    …than to suppose Australian taxpayers paid 180B for something that is either blindfolded at regular intervals, or is in fact utterly blind.

    It is one thing to fail to disclose radar detections to the PUBLIC. It is quite another to fail to disclose them to the SSWG.

    Same goes for Indonesian radar. That Indonesia has never supplied the SSWG with anything that might help guide the search has always been a huge elephant in the room. Either Indonesia is ripping off its taxpayers and is radar blind, they lied to the investigation, or MH370 never came within its radar range. Doesn’t it have to be one of these three?

    Having chosen our answer, we must now choose one of the three for Australia.

  80. Mick Gilbert says:


    Victor, you wrote, “… although using two different numerical tools, terrain masking at that location is not predicted for any known radar source.”

    It may be a semantical point but I’d argue that both models suggest that terrain masking from Khok Muang may have occurred. I say this based on the fact that both models show that Khok Muang’s coverage between the 238 and 241 radials falls away rapidly due to terrain masking; that aligns very accurately with the “thinning” of the radar plots leading up to the start of the gap on the Lido Hotel slide. That cannot be happenstance.

    Given the steep protrusion of nearby terrain (less than 25 km away) into Khok Muang’s sweep between the 241 and 255 radials it is very likely that that radar has clutter protection programmed in around that terrain; that is to protect the radar from being overwhelmed by clutter reflections associated with the terrain and its surrounds and it typically carves out a marginally larger “shadow” than yhe terrain itself. That sort of artificial constraint would provide the sort of “clean edge” we see in the Lido slide.

    Further, we know that anomalous propagation effects due to temperature inversions at altitude will have influenced radar performance on the night of 7-8 March.

    I am being pig-headed about this because I think that it is the best fit explanation for the Lido slide. I may well be wrong, time will tell.

  81. Andrew says:

    @Brock McEwen:

    Regarding JORN, there’s a vast difference between military capability and what actually occurs on a day-to-day basis. Australia has a fleet of E-7A Wedgetail aircraft that provide an airborne early warning & control capability. Do those aircraft operate 24/7, continuously monitoring the skies to give advance warning of an impending attack? NO. Australia also has several squadrons of F/A-18 Hornet fighters based at Williamtown and Tindal. Are any of those aircraft on instant readiness 24/7 to repel invaders? NO. Why? Because there are currently no threats that require the continuous deployment of such assets and, for the time being at least, Australia’s defence budget is better spent elsewhere. If a serious threat does arise, then the capability is certainly there to provide 24/7 coverage, but at the moment such a threat does not exist and neither did it exist back in 2014.

    JORN’s basic capabilities have been well documented and it is quite possible the system is far more capable than has been acknowledged. Nevertheless, the information available to the public indicates the system has a nominal range of 1,000-3,000km, depending on the ionospheric conditions, and the best coverage is achieved during the day time. MH370 may have been within range of the Laverton JORN site during the final stage of its southward flight, but only if the ionospheric conditions were suitable and only if the system was targeting that area at the time. Given that nobody knew the aircraft had headed south on the night of its disappearance, it seems extremely unlikely that JORN would have detected it, even if the system had been operational at the time.

    To be frank, this whole argument is quite pointless and will achieve nothing. Perhaps it’s time to move on, as Victor suggested.

  82. Victor Iannello says:

    @Mick Gilbert said, “I think that it is the best fit explanation for the Lido slide.”

    We’ll have to disagree on this point.

    @Brock McEwen: Regarding JORN and Indonesian radar, neither one of us will be able to definitively prove the other wrong, nor persuade the other with the evidence in hand. This was all discussed at length nearly three years ago, including lengthy discussions on Duncan Steel’s blog. You’ve stated your view. Unless you have new facts to present, I say we move on.

  83. Victor Iannello says:

    @Andrew: I responded to Brock before I saw your comment. Yes, we need to move on from JORN unless new facts surface.

  84. Brock McEwen says:

    @Victor, @Andrew:

    In my lifetime – because I was perfectly content to listen and learn, rather than speak – I have only ever made one comment re: JORN.


    The one just a few minutes ago, above.

    In response, I am told – not once, not twice, but THREE times – that it is time to move on. I find that interesting in and of itself.

    I agree, Victor, that even once is too often, if no new ideas are presented. To my knowledge, the parallels between Indonesian and Australian claims of radar impotence have never yet been made in this forum.

    But more importantly, the new idea requiring exploration is that, while the place they searched for 2.5 years was at least ARGUABLY out of range of JORN, this new path you are promoting leading to 30°S requires an excursion vastly more deep into its range. The plausibility of the official line reduces, and so our scrutiny must increase.

    Remember: despite the debris record strongly counter-indicating their own debris studies, the Aussies ran out the clock searching thousands of km offshore, instead of some place much more plausible. Even the CYA report that explained this still rules out anything within 2° of your path’s latitude. And even the staunchest supporter of the official search must admit that the REASON given for ruling it out (otherwise, the air search would have found surface debris) smacks of a bizarre and unfounded confidence.

    I think it is rational to wonder why the ATSB was so definitively ruling out everything east of “Arc 7 & 32°S”, yet citing relatively flimsy reasons. Perhaps the REAL reason is that search leadership would rather we NOT go back and take a harder look at those JORN claims…

  85. Brock McEwen says:

    2nd-last para, 1st sentence: clause before comma should read: “despite the zone being strongly counter-indicated by their own drift studies”. Apologies.

  86. Victor Iannello says:

    @Brock: Over the past three years, crash sites along the 7th arc from 40S to 20S latitude were considered. Australian officials have consistently said that JORN detection was not possible. Unless new facts or new evidence surfaces to the contrary, I am ending this discussion. We can’t rehash every MH370 conjecture that has appeared in the last three years.

  87. Andrew says:


    1. If you were flying at FL350 and ECON 52 at IGARI when technical problems arose requiring a diversion, what speed and altitude would you select for the return to Malaysia?

    It depends on the nature of the ‘technical problem’. In a time-critical situation (eg a fire) where the aircraft needs to land ASAP, then the crew would probably seek to fly at the maximum possible ground speed towards the diversion airfield. That would usually mean flying at MMO/VMO (M0.87/330 kt IAS) as long as possible. The best TAS would be achieved by descending to FL300, where MMO = VMO. At that level the TAS would be 508 kt, assuming ISA conditions. The wind would also need to be taken into account, but it was unlikely to be a significant factor in the case of MH370.

    In non time-critical situations, there is a wide choice of speeds and altitudes, again depending on the nature of the emergency and the distance to the diversion airfield. A depressurisation would ordinarily require a descent at MMO/VMO to 10,000 ft or MSA. Thereafter, the speed would probably be reduced to conserve fuel, depending on the distance to the diversion airfield. If there was structural damage associated with the depressurisation, the indicated speed would normally be limited to the speed at the time of the event in order to avoid further damage caused by increased air loads. In the case of an engine failure, the crew would normally descend at the drift down speed initially, while they complete the associated checklist and assess the situation. If terrain is a factor, the crew would then continue descending at the drift down speed to the one-engine inoperative level-off altitude. However, if terrain is not a factor, then the crew might choose to descend to a lower altitude and increase speed to the one-engine inoperative LRC speed or some other suitable speed, depending on the distance to the diversion airfield.

    Sorry I can’t be more precise. Do you have any particular scenario in mind?

    2. For the left turn around IGOGU from 296T to 180T (toward ANOKO), and with a 15 NM right offset, what bank angle or turn radius would the LNAV use? The latest possible turn would use the 25 degree bank angle limit which gives about 1.03 degrees per second (and 7.6 NM turn radius) at the LRC speed at FL360. However, this route flies well outside a circle around IGOGU with a 15 NM radius. I suspect the FMC actually flies the 15 NM radius circle centered on IGOGO. This only needs a 13.3 degree bank (at 0.52 deg/s) and it is a shorter, faster means of reaching the new 180T track 15 NM west of ANOKO while always maintaining a 15 NM separation from IGOGU.

    Good question and the short answer is I’m not sure without trying it in a simulator/aircraft. Unfortunately the manuals aren’t much help and in the real world we don’t normally see large offsets together with large track changes.

    Without the offset, the FMC calculates a bank angle that brings the aircraft as close as possible to IGOGU while completing a continuous turn from 296T to 180T. The maximum bank angle in that situation would be 25°, however the bank angle will be limited at high altitude (ie anything much above optimum) where the aircraft is thrust limited, to ensure it doesn’t fall out of the sky during the turn. In the case of MH370, the bank angle shouldn’t have been limited, because FL350 is below the optimum altitude for the aircraft’s weight.

    With an offset, however, the situation changes and I’m not sure if the FMC will endeavour to maintain the offset during the turn. It may simply look at the offset track and use that to calculate the bank angle, while ignoring the original track. Something else for me to try if I get the opportunity!

  88. Brock McEwen says:

    @Victor: I respect your wishes – it’s your blog, and you’ve more expertise then I. So pausing only to note that the degree of military blindness to rogue jumbo jets being posited – if true – raises more questions than answers, you’ve heard the last from me on this topic in this forum.

  89. DrB says:


    Thanks for responding to my questions.

    The military radar track data suggest an air speed very close to LRC / M0.84 at an altitude of about FL340-360. Certainly it was not MMO. Victor and I think an even FL is more likely than staying at FL350, since the track is now westward. Someone also suggested perhaps a 500 foot altitude increment. I would add that there is no definitive evidence of an altitude change later in the flight. Could have happened but I am pretty convinced it did not.

    Given that the speed is known to be LRC, can you infer any choice of altitude?

  90. Andrew says:


    FL340 and FL360 are both available for west-bound traffic on N571. FL360 would be closer to the aircraft’s optimum altitude and would certainly provide better range if the intention was to fly as far as possible. Given the aircraft’s lack of ATC clearance, a 500 ft increment would make sense. How about FL365?

  91. DrB says:


    Yes, I already looked at FL365. I found two issues: (1) the speed for the post-EOR route is a little too high at FL365, and (2) the fuel is noticeably under-consumed. Of course, it is possible that a slight descent occurred before the EOR error at ~18:44, but I don’t that is too likely. At least I can’t come up with a good reason to make a small altitude adjustment at some time between 18:22 and 18:44.

    I am also getting a slight difference between the nominal engine PDAs even at the FL360 route, which ends at 00:17:30 with ~1% of the 17:07 fuel load left in the tanks. I don’t know if my fuel model can ever be better than 1% error because I don’t know the temperature conditions for the previous several dozen flights. If I assume the temperatures for those flights is the same as it was for MH370, then I see about a 1% error in fuel remaining. This is conveniently subsumed by a 1% higher PDA so that MEFE occurs at 00:17:30. So, in summary, an average PDA that is ~1% higher than found from the previous flights is required for MH370 at Fl360.

  92. DrB says:


    Let me propose a straw-man end-of-flight scenario. Please tell me where I am going wrong:

    1. Assume the right engine suffers fuel exhaustion at ~00:08:30 at FL360 when the air speed is 420 KTAS / 237.3 KIAS / M0.7217 at SAT = 223 K.
    2. For the first 3 minutes, the KTAS declines by 19 KTAS / minute and the altitude stays at FL360.
    3. At ~00:11:30 with a current air speed of 363 KTAS, the aircraft begins to descend. Note that the descent begins after the 00:11:00 handshake, and there is no discernible effect on the BFO then. In addition, the impact of the slowing on the average leg speed from 22:41 to 00:11 is negligible.
    4. For six minutes (~00:11:30 to ~00:17:30) the air speed remains constant at ~363 KTAS while the rate of descent is ~1,000 fpm.
    5. At ~00:17:30 the altitude is now ~FL300 and speed is still ~363 KTAS. The left engine flames out.
    6. After 00:17:30 the air speed begins again to slow (by ~10?? KTAS/minute), and the rate of descent accelerates.
    7. At 00:19:29 the ROD is 4,400 fpm, the air speed is ~343 KTAS and the altitude is ~FL230.

  93. Mick Gilbert says:


    Re: Lido, Khok Muang and “We’ll have to disagree on this point.”

    Agreed, we will. But because life should be interesting and rewarding, I’ve got a bottle of ’88 Grange that says I’m right. It’s not a great Grange vintage but we’re not talking about the Higgs boson either.

  94. RetiredF4 says:

    A radar station on Pulau Perak does make a lot of sense from the military POV. It would have a perfect view on the air situation to the north inlet of the Malacca street, which otherwise could only be acchieved by airborne assets. Those are expensive to buy and to operate 24/7.

    Are there points which support the existence of a radar station on the island? Let’s speculate.
    What kind of task has the small military unit stationed on the rock? It is not need it to operate a light house. According to the antennas on the lighthouse it looks at least like a ELINT listening post. Why not combine it with a radar antenna orientated along the straight to the north-west, which does not necessarily have to be on the light house? It could not be a permanent system either, it could be a mobile tactical system brought in by helicopter. It could also be in the lighthouse itself, the walls of the lighthouse constructed to allow the passage of the radar signals.

    Does the Lido track give hints to a radar system on Pulau Perak at the time of the disappearance of MH370? I think so. Nobody has explained yet why the position of Pulau Perak sports a big unnamed circle, which seems to mark the no show area of MH370. If there was a radar on Pulau-Perak, this circle looks like the dead cone above the radar position. Would Malaysia like to communicate the existence and task of such a system? Not necessarily when the operation is combined with ELINT spying on their neighbours.

    @Victor Thanks for this blog and your work.

  95. Gysbreght says:


    No objections to your points 1 – 3.

    “4. For six minutes (~00:11:30 to ~00:17:30) the air speed remains constant at ~363 KTAS while the rate of descent is ~1,000 fpm.”
    The airplane will maintain 207 kt IAS, and the rate of descent at constant IAS corresponding to a horizontal acceleration of 19 kTAS/min (0.017 g) is about 600 fpm.

    5. At ~00:17:30 the altitude is now ~FL324 and speed is still 207 kt IAS (343 kt TAS ISA). The left engine flames out. The autopilot disengages and the airplane starts a phugoid descent.

    6. In the phugoid descent after 00:17:30 the indicated air speed varies sinusoidally around a mean value of 207 kt, while the rate of descent varies around a mean value of 2000 fpm, provided the airplane maintains wings approximately level.

    7. As the airplane descends to lower altitudes, the amplitude of the phugoid oscillations may reduce or increase, and the mean rate of descent will reduce proportionally to the reduction of TAS at constant IAS.

  96. Gysbreght says:

    @RetiredF4: The location of Pulau Perak is indicated on the Enlarged Lido Hotel radar image with timestamps added by Bill Holland.

  97. Victor Iannello says:

    @DrB & @Andrew: In FSX, I simulated the 15 NM right offset with a route of MEKAR-NILAM-IGOGU-ANOKO, with a hold at ANOKO, and created a KMZ file to drag into Google Earth. The A/P mode is LNAV,VNAV ALT with CI=85 at FL340, which produces TAS=GS=495 kn at the meteorological conditions, which should be about right. At the turn from IGOGU to ANOKO, the 15 NM offset is maintained, and the bank angle of the turn is about 25 deg. The plane begins a deceleration to holding speed just after the turn is completed. An END OF OFFSET message in the FMC appears before reaching ANOKO, and the plane continues on a constant heading after reaching ANOKO in LNAV,VNAV ALT and at holding speed.

  98. Victor Iannello says:

    @DrB: There appears to be disagreement about the speed at which a descent will begin with one engine. If A/P is disengaged, envelope protection prevents the pilot from trimming below the minimum maneuvering speed. I think here we have agreement.

    However, if the A/P is engaged, and it is in a mode in which it is maintaining altitude, the speed will drop to just above the stick shaker speed. At that point, envelope protection kicks in, and the nose is lowered and speed increases. The PMDG 777 model says that once the speed increases, the envelope protection disengages, and the A/P will again try to maintain altitude by raising the nose, which will again lower the speed, and engage envelope protection. So the altitude control and the envelope protection alternately cycle on and off, keeping the speed close to the stick shaker speed during the descent.

  99. DrB says:


    1. Victor, I don’t understand why your route turns back past 180 degrees at the end of the turn. Wouldn’t the turn just end at 180 degrees like I have shown here?

    Also a 25 degree bank is never needed for this type of turn that is using a 15 NM radius. It is only 13.3 degrees at 494 KTAS.

    2. A 15 NM SLOP is indeed possible to fit to the satellite data, but I think it requires two things:

    a. The 18:22:12 position is actually about ~13 NM past MEKAR, not 10 NM.

    b. There is some tailwind error increasing the ground speed.

  100. DrB says:


    Sorry for inserting the wrong link. To see my 15 NM plot, go here.

  101. Victor Iannello says:

    @DrB: Apparently, with the 15-NM offset, the PMDG 777 doesn’t “cut the corner”. The plane was in LNAV mode the entire time, and the maximum bank angle was around 25 deg. What you describe in your link is how turns look without the route offset. As for matching the BTO and the radar data with a 15-NM offset manoeuver occurring after 18:22:12 but before 18:25:27, you are threading a needle.

  102. TBill says:

    @Mick Gilbert
    On the side discussion about EK343, there is a Reddit thread on EK343’s 8-March-2014 flight details and it establishes EK343 KLIA take-off time as 1:29 MYT, which on the surface without doing a flight sim approximation, would seem consistent with the EK343 distance traveled in the two FR24 screen shots.

  103. TBill says:

    …fyi I would ask what simulation rate you are using in FSX…I always go to 1x for turns otherwise it can take the turn too wide.

  104. DrB says:


    I don’t think a 15 NM offset route can fit the 18:40 BFOs unless a 15 NM radius turn is used around IGOGU, or an error in the 18:22 position error larger than 3 NM is allowed.

    The timing of the lateral offset between 18:22 and 18:25 is tight, but it really only requires that the offset maneuver duration and the SDU time-to-LOR be fairly close (about 3 minutes each) and that the SLOP and the power-up occurred close together in time.

    The SLOP can start later than 18:22 if you allow the 18:25:27 BFO to be “contaminated” with the end of the SLOP maneuver. I’m on the road now, but when I get a chance I will try adding a frequency offset to the 18:25:27 BFO, refitting the OXCO frequency transient curve, and see if there is a magnitude limit on the route contribution at that time that still allows the other 6 log-ons to be well fitted. Maybe I can establish an upper bound on the route contribution to the 18:25:27 BFO. In my first analysis, I had assumed it was zero. If some route contribution is allowed, then the SLOP can begin later, closer to 18:25. Of course, the SLOP timing does not affect the turn radius needed around IGOGU.

  105. Victor Iannello says:

    @TBill: It was 1X at the turn. I accelerated time during the straight section.

  106. Victor Iannello says:

    @DrB: You might find interesting this image of the cockpit for the plane passing NILAM with a 15-NM offset to the right. The route is NILAM-IGOGU-ANOKO with a hold at ANOKO. If you look at the MFD (in NAV/MAP mode), the route is shown with and without the offset. In both cases, the route does not “cut the corner”, perhaps because the turn is greater than 90 deg. Also, if you look carefully, you will see a little circle showing where deceleration to holding speed will commence. Also, the circle around ANOKO indicates a holding pattern.

  107. Brian Anderson says:


    Some questions:

  108. TBill says:

    @Victor @DrB
    OK yes I think I see the similar wide on PSS777.

  109. Brian Anderson says:


    Some questions:
    1. What leads you to the assumption that the right engine failed at 00:08:30 ?
    2. Do you see a small [200ft/min] descent in the BFO at 00:11 ?
    3. At your assumed speeds is it actually possible for the aircraft to reach the 7th arc on the azimuth that it was tracking from 00:11 ?

    At the point of the left engine flame out the A/P disconnects and the TAC reverts to the pre-trimmed condition. The aircraft may commence a series of phugoid oscillations, but it may instead commence a shallow turn. Even if it begins with phugoids there is no guarantee that they will continue and the aircraft will stay wings level and relatively straight.

    The SIM tests showed pretty conclusively that a shallow turn will steadily evolve into an increasingly tighter and steeper turn. The BFO indicated descent rates about 00:19:xx are entirely consistent with a steep turning descent; in other words a spiral dive. [A spiral dive is not a vertical descent as so many seem to believe.]

    In fact, depending on some speed assumptions, I think it possible to show that with a turn to the left after the left engine flame out the aircraft is more likely to reach the 7th arc. [My April 2015 paper.]

  110. TBill says:

    @VictorI @DrB
    The only thing I see in FS9 if I manually change to ANOKO waypoint at exactly 16nM before reaching IGOGU, then that cuts the corner smoothly. But it is trial and error to find the perfect place to start the turn, otherwise you overshoot one way or the other. I am just using the actual waypoints without offset.

  111. Gysbreght says:

    @Victor: Does FSX always fly over a waypoint before it starts turning to the heading for the next leg?

  112. Victor Iannello says:

    @TBill: When you replace IGOGU with ANOKO, you are now commanding the FMC to fly “Direct To” ANOKO and your current position becomes a waypoint for calculating that leg. (That also cancels the offset.) That’s not really the same thing. For instance, the track won’t be 180T as you pass ANOKO.

  113. Paul Smithson says:

    @Mick Gilbert. Thank you for your thoughtful and well-phrased comment on Ben Sandiland’s article https://blogs.crikey.com.au/planetalking/2017/03/04/three-years-five-reasons-mh370-pilot-didnt/

    It never ceases to amaze me how the majority have opted to interpret the available evidence to infer that this was a nefarious act and not an accident.

  114. Victor Iannello says:

    @Gysbreght: No, that would not be realistic. In general, it starts the turn before the waypoint. I need to run some test cases to determine the algorithm it is using. I suspect it is the sharp turn and/or the large offset that causes the shape of the turn to be what it is for the turn from IGOGU to ANOKO.

  115. Mick Gilbert says:


    Re: EK343. Thanks for that uodate. Yes, it is clear that EK343 was where Richard had plotted it. My initial query was driven by my own lack of appreciation for the distances involved.

    @Paul Smithson. Thank you. I’d just like people to consider the evidence we have and think for themselves rather than mindlessly consuming some of the utter pap that passes for “expert” opinion in the MSM and accepting it unquestioningly as fact.

  116. Gysbreght says:

    @Brian Anderson:
    I’ll leave your questions 1 – 3 to DrB.

    ” The aircraft may commence a series of phugoid oscillations, but it may instead commence a shallow turn.” A phugoid is a pitch oscillation, and may exist with or without bank.
    ” Even if it begins with phugoids there is no guarantee that they will continue”
    That’s what I said.

    ” The SIM tests showed pretty conclusively that a shallow turn will steadily evolve into an increasingly tighter and steeper turn.”
    If you are referring to the SIM tests that ALSM witnessed, the outcome was obviously determined by the rudder trim applied. In the Boeing simulations shown in the ATSB report of Nov.2, 2016 only “an electrical configuration where the loss of engine power from one engine resulted in the loss of autopilot (AP)” or similar trim asymmetry resulted in “an increasingly tighter and steeper turn”. The ATSB has not been able to present a scenario that matched the BFO’s at 00:19:xx close to the time of the 7th arc.

  117. TBill says:

    @Victor @Gysbreght
    Do you know if I can do offsets on PSS777?

    I agree my path is not the exact same…I was inserting NILAM IGOGU ANOKO and NOPEK as my path, with no offset. And I was seeing what I could to match DrB’s flight curve shape to come straight down the 180S line. I tried a number of things, the only thing that seemed to work was to force the turn to start at 16 nM from IGOGU. yes by entering ANOKO.

    What I was doing is what PSS777 does itself: at 10 nM away from IGOGU, it switches to ANOKO. So I switched at 16 nM away instead, and that gave smooth curve ending at 180S.

  118. buyerninety says:

    @Mick Gilbert
    “I wanted to make this change clear to readers who have been following
    my work as I think it unfair that they be expected to spot that change
    between versions themselves.”
    Perhaps insert a changelog, even on the last page if necessary, with
    what part(s) of the document the change appears on, and a brief note
    as to the reason for the change,
    Page 2: ‘Golong’ River, changed to ‘Golok’ River , due to incorrect name.

  119. Mick Gilbert says:


    Yes, that’s a good idea. I have been unprofessionally remiss in managing version control and referencing (and spelling as it turns out – some of those foreign names play merry hell with spell check, not that that is a legitimate excuse).

  120. Victor Iannello says:

    @TBill asked, “Do you know if I can do offsets on PSS777?”

    That option is missing from the ROUTE window in the FMC. My guess is the answer is no. The PMDG 777 model is much closer to the FCOM.

  121. DrB says:

    @Brian Anderson,

    Here are the answers to your questions #1-3:

    1. What leads you to the assumption that the right engine failed at 00:08:30 ?

    That estimated time comes from my fuel model calculations. Knowledge is required of several parameters:

    (a) the ratio of fuel flow in cruise for the right and left engines (which I have from ATSB),

    (b) the fuel at 17:07 of 43,800 kg,

    (c) the fuel flow of the left engine after right engine flame-out (which has been addressed here in the last few days), and

    (d) the left engine flame-out time estimated as 00:17:30 by ATSB.

    2. Do you see a small [200ft/min] descent in the BFO at 00:11 ?

    No. My CTH route has a predicted BFO at 00:11:00 of 254 Hz, and the measured calibrated BFO is 252 Hz. No descent at 00:11 is indicated.

    3. At your assumed speeds is it actually possible for the aircraft to reach the 7th arc on the azimuth that it was tracking from 00:11 ?

    I fit the last leg from 00:11 to 00:19:29 by allowing the bearing to be fitted, forcing the BTO error to be zero, and achieving a TAS that matches the fuel model predicted average for that time period. That in turn depends on the assumed slow-down rate. The leg from 22:41 to 00:11 is at ~174 degrees true, and the 00:11 to 00:19 leg comes out at ~165 degrees. There is no problem reaching the 7th Arc at that azimuth. In my opinion, the aircraft was probably turning left at 00:19 and impacted inside the 00:19 arc.

  122. DrB says:


    There may be a third possibility for how a course is changed when a significant offset is being used. Here are the three cases:

    (1) The FMC may turn with a bank limit of 25 degrees before reaching the intersection of the 15 NM right offset N571 track at 296 true to the 180 true course through the fix 15 NM right (west) of ANOKO. The start of the turn is the latest and the path is the longest. The end of the turn is on a 180 true line through the ANOKO + OFFSET fix with no overshoot.

    (2) The FMC may follow the circular path from 15 NM right of IGOGU to 15 NM due west of IGOGU followed by a 180 true leg to ANOKO + OFFSET. This path only needs a 13 degree bank angle with a 15 NM turn radius. This method is shorter than #1 but longer than #3.

    (3) The FMC may fly to the 15 NM abeam position from IGOGU and then turn at 25 degree bank angle toward the 15 NM right position of ANOKO. This cuts across the corner and achieves the shortest path. Presumably it would intercept the due south line through ANOKO + OFFSET so that the path goes through that fix at 180 degrees true. This route is a little more complicated to calculate, but nothing that the FMC could not easily do.

    Victor, can you tell whether PMDG777 does one of these?

    Andrew, what does a real B777 FMC do?

    As an aside, if a real B777 FMC does #1, that would eliminate a 15 NM offset as being consistent with the radar and satellite data, in my opinion. #2 is (barely) consistent. #3 is consistent with some margin.

  123. TBill says:

    Just a thought on DrB’s question, if you had IGOGU/-6nM as the waypoint, it might start the turn earlier. Perhaps that is cheating, and not sure if you could enter that.

  124. Brock McEwen says:

    @Richard: I read your latest paper with great interest – thank you for continuing to do your best to offer closure for NoK.

    Three quick questions:

    1) You use the word ‘probable’ in both the title and the conclusion in connection with ‘Arc7 at 30S’. The report body implies precision, as in: “29S is too fast, and 31S is too slow, but 30S is JUUUST right..” (I am paraphrasing, obviously). So I am curious: by “probable” – and considering a range of, say, 29S to 31S – what numerical probability do you currently attach to this drift-indicated region? (I am not used to reverse drift analyses generating anywhere near this degree of precision, is why I ask.)

    2) Re: the para: “From a starting point on the 7th Arc at 30°S, the simulation shows debris arriving south west of Reunion after 487 days. This fits the timeframe of the Flaperon find after 508 days. The possible tracks show significant dispersion as the efficiency factor is varied, hitting Reunion and Mauritius but now including South Africa and Tanzania.” When the efficiency factor is varied sufficient to place ‘Roy’ in Klein Brak (not just ‘SA’ – Klein Brak, 1,455km around the horn from the NE start of the ‘SA’ shoreline) by day 655, on what day does debris THEN reach Réunion?

    3) Do you have path model versions (à la Version 13.x, 16.x, etc.) calibrated for either or both of your last two papers: “McMurdo=27s” and “Wilkins=30s”? If so, would you be so kind as to make them available to this forum?

    Thanks, and regards,

  125. DrB says:


    I’d be curious to see how PMDG777 flies a NILAM to IGOGU to NILAM route with a 15 NM right offset. Does it make a 30 NM diameter semicircle around IGOGU? Does the bank angle ever get higher than ~13 degrees?

  126. DrB says:


    I don’t understand why the route shown in your KMZ file overshoots the 180 degree track through the fix 15 NM west of ANOKO. I doubt the FMC would overshoot the desired track at all, much less the ~6 NM overshoot your plot indicates. Am I missing something here? Perhaps the PMDG777 simulator doesn’t follow the FMC accurately when there are route offsets, or is there something else amiss?

    Clearly the FMC has the ability to turn and follow a crossing course without overshoot given a maximum bank angle limit (nominally 25 degrees in A/P). Now if there is a small offset, say 1-2 NM, the same turn type will obviously work also without overshoot. However, with large offsets, that sort of late, high-bank angle turn wastes time and fuel. It is possible that once the offset exceeds the value of the usual turn radius (at 25 degree bank and at the current TAS), the bank angle limit is set so that a circle of radius equal to the offset is flown, and there is no overshoot. This sames time, fuel, and money. This is all conjecture on my part, but I can see a good argument that the turns for large offsets are treated differently conceptually to improve the flight efficiency.

  127. DrB says:

    @Brian Anderson,

    Thank you all for contributing to the discussion of the end-of-flight scenario. I will generate an end-of-flight model with ~1 minute time resolution listing FL, KTAS, KIAS, SAT, M, ROD, and FF for L/R/L+R engines, attempting to incorporate the comments and observations you have made. Actually there will be two models, I think. One will match Victor’s PMDG777 observations of a low speed being reached before descent, and one for Gysbreght’s 207 KIAS descent condition. Then perhaps we can better visualize the two cases and figure out how to reconcile the two.

    I’ll post tables and plots when I have something working. It might be a few days since I am travelling.

  128. Andrew says:


    Apologies for the delay in my reply. I’m now playing catch up!

    Regarding DrB’s straw-man end-of-flight scenario:

    “3. At ~00:11:30 with a current air speed of 363 KTAS, the aircraft begins to descend.”

    The aircraft should maintain altitude until the speed approaches or reaches the stick-shaker speed. Victor previously said the FSX simulator maintained altitude down to around 163 KIAS, or about 291 KTAS at FL360. I don’t have the data, but that sounds reasonable.

    I’ve revisited my previous comments regarding the aircraft’s behaviour as the speed decays following the first engine failure, and agree with Victor: when the aircraft enters envelope protection with the autopilot engaged, it should descend at a speed slightly above the stick-shaker speed. The B777 FCTM states the following:

    “As the speed approaches the minimum speed, the AUTOPILOT caution message appears, an amber line is drawn through the selected pitch mode and the flight director pitch bar is removed. At minimum speed, the stick shaker activates. Shortly after the stick shaker activates, the autopilot begins to descend from the selected altitude. The autopilot maintains a descent at a speed that is slightly above the minimum speed.

    I’m not sure about the PMDG 777 simulator behaviour described by Victor, where the aircraft cycled between altitude control and envelope protection. According to the manuals and my recollection of similar exercises in a Level D simulator, on entering envelope protection the aircraft should enter an attitude stabilising mode that keeps it descending at a more or less constant IAS, just above the stick-shaker.

    “4. For six minutes (~00:11:30 to ~00:17:30) the air speed remains constant at ~363 KTAS while the rate of descent is ~1,000 fpm.”

    The aircraft should descend at a more or less constant IAS, as described above. The TAS will decrease as the aircraft descends.

    “6. After 00:17:30 the air speed begins again to slow (by ~10?? KTAS/minute), and the rate of descent accelerates.”

    The autopilot will disengage after the second engine failure and at that point I would expect the aircraft to return to the trimmed state that existed as the aircraft slowed down past the minimum manoeuvring speed. The aircraft should, therefore, initially accelerate back to approximately the minimum manoeuvring speed (210-220 KIAS??). The TAS would initially increase, but would then start decreasing again as the aircraft continued descending.

    There seems to be some conjecture about the aircraft’s lateral behaviour following the second engine failure. The ATSB’s first simulator tests, mentioned in “MH370 – Definition of Underwater Search Areas”, dated 3 December 2015, found that “in each test case, the aircraft began turning to the left and remained in a banked turn”. Brian stated that “at the point of the left engine flame out the A/P disconnects and the TAC reverts to the pre-trimmed condition”. While the autopilot certainly disconnects, I have my doubts about the TAC reverting to the “pre-trimmed condition”. The TAC should disengage after both engines fail, due to the loss of engine running data; however, the way I see it, some of the rudder trim applied by the TAC should remain.

    When the TAC function is active, the primary flight computers send high rate rudder commands that directly move the rudder to counter any yaw caused by the difference in thrust. The TAC also sends rudder trim commands that move the rudder trim actuator to trim the rudder via the rudder pedals. The direct rudder commands cease when the TAC disengages, however, as I understand it, the rudder trim input is not removed. The amount of residual rudder trim probably depends on how long it takes for the system to recognise the second engine failure before the TAC disengages. I surmise that the yaw caused by the residual rudder trim input causes the aircraft to roll and enter a turning descent.

    This is already getting too long, so I’ll leave it there and return to the 15nm offset problem later!

  129. Gysbreght says:

    @Andrew: Thank you for the unambiguous statement from the FCTM. That settles it as far as I am concerned. It is also plausible that after the second engine flame-out, and the loss of autopilot and envelope protection that it entails, the speed will return to the minimum maneuver speed, which is the minimum trimmed speed.

    As to residual trim remaining after the TAC is lost, I would think that should be quite small, if anything. After engine flame-out the engine run-down takes a couple of seconds until the electrical generators drop off-line. During that rundown phase the TAC adjusts the trim input as the EPR drops. The amount of rudder is proportional to the momentary thrust difference. The TAC applies zero input when the calculated thrust asymmetry is 10%, i.e. before electrical and hydraulic power is lost and the TAC disengages.

  130. Andrew says:


    I agree, the amount of residual rudder trim is likely to be small, but it might be enough to cause a small amount of yaw and subsequent roll. The TAC uses the N1 speed signals from the EECs to calculate the thrust. The N1 signals are cross-checked against other engine data for accuracy. According to my B777 engineering notes, one of the conditions that will cause the TAC to disengage is “two engines not running in the air”. Unfortunately the notes don’t elaborate on how the system decides if two engines are not running. The notes also say the TAC function stops being active when the thrust difference between the two engines is less than 3% of the maximum rated thrust.

    The RAT will deploy after both engines fail, and will power the primary flight control components of the centre hydraulic system. In that configuration, the right flaperon will be powered by the centre hydraulic system, but the left flaperon will be left unpowered, as its PCUs are only powered by the left and right hydraulic systems. With no hydraulic power, the left flaperon PCUs will operate in bypass mode, allowing the flaperon to move freely in both directions. In flight, the air loads will cause the flaperon to move upwards about 10 degrees. That too also produce a rolling moment to the left.

  131. Andrew says:

    Oops – replace “too” with “should” in that last sentence!

  132. Victor Iannello says:

    @DrB asked, “Victor, can you tell whether PMDG777 does one of these?”

    I’m not sure what you’re asking because I already ran the simulation for the route with 15 NM of offset and provided the KMZ file. For a sharp turn, with or without offset, even for a “fly-by” waypoint type, there will be overshoot. The algorithm in the FMC clearly does not choose the path of minimum length. There must also be airspace considerations. I haven’t run enough test cases to determine what those are, although I would be surprised if the path construction at a turn as computed by the PMDG 777 FMC is very different than the real world.

  133. Brock McEwen says:

    @Victor re: “altitude control and envelope protection alternately cycle on and off”:

    Do you have any rough idea of / means to estimate…

    1) the period of this cycle, and
    2) the proportion of time spent in each mode?

    If we had these, both speed & altitude at engine 2 flameout become a determinable function of [speed & altitude at engine 1 flameout, and the time between them].

  134. Victor Iannello says:

    @Brock McEwen: Whether or not the altitude and stall control modes cycle on and off, the net effect is the same: The plane descends at a speed just above the stick shaker speed, which is about 162 KIAS at cruise altitude.

    Of course, this depends on the timing between the two engine flameouts to be long enough so that this speed is reached with one engine still running.

  135. TBill says:

    Interesting (to me) Arc7 graphic on Mike Chillit’s site, he shows UAE425, which I estimate could have been quite close to MH370 ditch location if MH370 ditched in the 20S to 22S vicinity. I have one suggested path that follows vector L894 (before that turn it follows the Iannello/Godfrey McMurdo path). I recently realized my L894 path is similar to Ed Baker’s proposed path, where he follows the heavy cloud line which thickens right at around L894 and south. Similar to Ed’s path, I got a BTO problem after Arc5, and probably a BFO problem, but I like the path anyways because avoiding UAE425 is part of the story line.


  136. Gysbreght says:

    The stickshaker is probably triggered by Angle-of-Attack, rather than a calculated ‘stick shaker speed’.

  137. Ge Rijn says:


    Perhaps allready a year ago I posted a weather satelite picture on @JW’s blog which shows the clouds that 7/8 March night. If I remember it well the area North and around 20/22S was heavily clouded like the whole West Australian ocean area till ~28S. Only after this ~latitude the skies were getting clear again to the South. Do you mean this cloud line?

    I suggest no plane flying above the clouds North of ~28S could have seen MH370 entering the ocean.

  138. Ge Rijn says:


    What I read earlier not only the right wing flaperon is actuated under RAT but also the left elevator (confirmation please).
    Could this be designed to compensate eachother keeping the plane level/controlable under RAT? It would make sence IMO.

  139. buyerninety says:

    (No disrespect to any poster here – below webpage may be helpful)
    @Gysbreght said;
    “As to residual trim remaining after the TAC is lost, I would think that should be quite small, if anything.”

    “If the TAC fails – any trim applied is removed (Trim centred)”

  140. Ge Rijn says:


    Another thing. I’m not a pilot but from what I understand of aerodynamics AoA and stall speed are very much dependend of the speed (and thrust) of the plane.

  141. DrB says:


    Victor, if I understand you correctly, you seem to be saying that the FMC will always overshoot a new course (say from 296T to 180T) even with zero offset. Perhaps I misunderstood what you meant. That contradicts what we have been told in the past (maybe from Don T) – that a turn begins before that intersection is reached and proceeds to become tangent to the new course with no overshoot. If anyone thinks this is incorrect, please say so. The question at hand is how the FMC handles a sizable offset. As I said before, I would be very surprised if there was any overshoot beyond the new course. Andrew, can you find the answer to this one?

  142. Brock McEwen says:

    @Victor: ok, thanks, I think I’ve now got it:

    First 10-11 minutes: “decel mode”: maintain alt, KTAS drops at ~19 kts/min from [cruising speed] to 162 KIAS ~ 275 KTAS

    Thereafter: “descent mode”: maintain KIAS (so still dropping KTAS, but much more slowly), alt drops at ~600 ft/min

    Since average KTAS required to traverse Arcs 6 to 7 is some 375 KGS (even if an immediate left turn towards perpendicular at engine 2 flameout is permitted), I agree with you that this second mode hardly seems worth our time.

    I couldn’t find it this AM – but didn’t the ATSB say their end-flight models were “refined” to include a kick-over into this second mode? If so, I don’t know how they are getting enough speed to traverse the arcs.

    I did write to the ATSB over a year ago to request clarity on their bizarre “up to 15 minutes” characterization of time spent on one engine, given what outside experts said this did to a/c speed, and its devastating implications for the BTO fit. While they have responded to most of my requests, they took ath 5th on this one.

  143. Richard says:


    1) I use the word ‘probable’ in my last paper because I am not certain of what happened around Car Nicobar in respect of timing and fuel usage. Considering a range of 29S to 31S – I believe the numerical probability is > 50%. The floating debris simulator only uses forward drift and not reverse drift calculations.

    2) Re: the para: “From a starting point on the 7th Arc at 30°S, the simulation shows debris arriving south west of Reunion after 487 days. This fits the timeframe of the Flaperon find after 508 days. The possible tracks show significant dispersion as the efficiency factor is varied, hitting Reunion and Mauritius but now including South Africa and Tanzania.” When the efficiency factor is varied sufficient to place ‘Roy’ in Klein Brak (not just ‘SA’ – Klein Brak, 1,455km around the horn from the NE start of the ‘SA’ shoreline) by day 655, on what day does debris THEN reach Réunion?

    The simulator shows 161 days from Reunion to Klein Brak River. Working forwards 487 days + 161 days = 648 days at Klein Brak River. Working backwards 655 days -161 days gives 494 days to Reunion.

    Here is a link to the simulator output based on actual GDP Buoy data:


    3) Do you have path model versions (à la Version 13.x, 16.x, etc.) calibrated for either or both of your last two papers: “McMurdo=27s” and “Wilkins=30s”? If so, would you be so kind as to make them available to this forum?

    Here are link to both my worksheets exported from Apple Numbers to Microsoft Excel. The export said not all formulae could be transferred and last values were used instead. If you have a problem please let me know:



  144. Gysbreght says:

    @Ge Rijn: You’re right about the RAT powering right flaperon and left elevator. Since the flaperon is for roll control and the elevator for pitch control I don’t think one compensates the other. The RAT also powers the ailerons and some of the spoilers. I believe there is adequate control for steady flight and normal maneuvering.

    The point I was making about AoA is that at a certain weight and speed it varies with normal acceleration (“g”-level). When the system maintains a speed close to the 1g stick shaker speed, small variations in AoA due to control inputs or turbulence may cause intermittent exceedance of the AoA stall warning threshold and intermittent activation of the stickshaker.

  145. DrB says:


    My offset turn method #3 above (with two 25 degree bank turns) has two things going for it:

    (1) You can start at the 10 NM past MEKAR fix at 18:22:12. You don’t need extra tailwind or several miles of 18:22 position error to the west.

    (2) The 18:40 phone call occurs after the first turn abeam of IGOGU but before the second turn just north of ANOKO + OFFSET. The 18:40 BFOs are very well matched by a track at ~202 degrees true in between these two turns.

    Here is a plot of this type of turn along with BTO and BFO plots and notes.

  146. TBill says:

    What I seem to see (in FS9 PSS-777) is you can do up to 90 deg turn with almost no overshoot between waypoints (without offset). Once you come to a sharper angle, there is overshoot. FS9 starts the turn at approx. 10 nM before the waypoint, which for a sharper angle, would have to start the turn (for example) at 16 nM before IGOGU to make it look like your curve down to ANOKO.

    Basically my turns look a lot like Victor’s turns in FSX. How well that matches real 777 is another question.

  147. Brock McEwen says:

    @Richard: thanks much. Will pore over each item.

    And thank you for the clarification, and the additional data. It could be inferred that you are making the following leap:

    1) 487 from 30s to SW of Réunion
    2) 161 from Réunion to Klein Brak
    3) therefore 548 from 30s to Klein Brak

    But I would rather verify this than assume this; perhaps debris after 487 days is sufficiently south of Réunion to MISS Klein Brak altogether. I think it would improve the paper’s credibility if a path all the way from 30s to Klein Brak is modeled directly – feel free to jack up the efficiency factor as high as you need to to get it there – and then we can assess the plausibility of all the other arrival times.

    I’m disputing neither your result nor the possibility that Roy travelled that far, that fast. It’s just that adrift.org had a very difficult time getting debris from 30s to Klein Brak in only 655 days. And the recently published leeway data on Roy (LESS than for the flaperon) would seem to further counter-indicate hypothesized impacts as far south as 30s. So it would be prudent for drift studies from here on in to address head-on this challenging piece of debris.

    (It is also not clear to me from your reply that the 487 and the 161 each used the same efficiency factor – would appreciate explicit confirmation. But this is a small point, as I expect they did.)

  148. buyerninety says:

    I would caution that there may be other factors impinging upon the behaviour seen
    in VictorI’s png picture,

    Due to that flightpath being programmed as having ANOKO as the HOLDING FIX, with a
    specific programmed INBD CRS/DIR (180°/R TURN probably) and other complementary
    inputs (as detailed here;)
    the flightpath behaviour may be different to that that would be seen if the programming
    were simply ‘to offset a route with IGOGU and ANOKO as simple fly-by waypoints’.

    Additionally, another factor is that I believe the FMC have been programmed such that
    WITT is the ARRIVAL airport (@VictorI can comment?).
    This may be affecting that displayed flightpath in the following manner:
    If WITT is the ARRIVAL airport, the FMC may regard IGOGU and ANOKO as STAR waypoints
    (Standard Terminal Arrival Route) instead of as simple fly-by waypoints. If the FMC is
    treating IGOGU as a STAR waypoint, it may be assigning different characteristics to it
    than if it were treating it as a fly-by waypoint, for example it may treat IGOGU as a
    fly-over waypoint (which could explain why that flightpath is represented as passing
    through IGOGU rather than being represented as merely flying to IGOGU and then directly
    on to the next waypoint). (I’m referring to the non-offset path there, of course.)
    That a waypoint could be either a fly-by waypoint, or a fly-over waypoint when part of
    a STAR procedure, is suggested to me by the example of this document;
    wherein KENDI is designated a fly-over waypoint only in certain specific approaches.

    A check of the Indonesian AIP for WITT or e.g. the navigraph charts for RNAV approaches
    to WITT, may be called for.

    One other minor point – I note that although IGOGU and ANOKO appear as STAR waypoints
    for, I assume, RNAV approaches to WITT, we do not see any magenta paths to WITT.
    I can only assume that ANOKO may be a STAR waypoint in more than one RNAV approach to
    WITT, and because no one specific RNAV approach has been programmed into the FMC, or
    no specifc approach runway has been programmed, the FMC therefore does not have
    sufficient information to represent any particular path to WITT.

    (I provide the above thoughts only as an aid to other posters – readers may know that
    ANOKO has no place in any theory of mine… .)

  149. Victor Iannello says:

    @buyerninety: No, WITT was not the arrival airport. The overshoot is a function of the turn angle and perhaps the turn radius (determined by the speed and the bank angle). And I only considered fly-by waypoints (with and without hold patterns). As @TBill said, for turns greater than 90 deg, there seems to be some overshoot for the case considered. I suspect that the FMC attempts to keep the track error from the route below some threshold. For larger turn angles, this comes at the expense of some overshoot. Overshoot adds to the length of the turn, but there might be other advantages relative to traffic for keeping the track error small.

  150. DrB says:


    Thanks for your inputs on the question of how a B777 would turn from IGOGU to ANOKO with a 15 NM right offset. Hopefully Andrew or another B777 pilot will weigh in. I wouldn’t be shocked if the flight simulators were different from the real aircraft for this somewhat unusual turn situation.

    It would be interesting to see what the flight simulators do for a 15 NM offset and waypoints NILAM IGOGU NILAM. That is, what if you do a 180 degree turn at IGOGU? Does the route go 15 NM past IGOGU (in a semicircle), or does it follow some other pattern?

  151. Andrew says:

    @Ge Rijn:
    Thank you for posting the article about the TAC function. Most of the information appears to be taken from the Boeing technical training notes used for B777 engineer training, of which I have a copy. The one thing the notes don’t mention is the removal of the TAC rudder trim input when the TAC fails. Nevertheless that does make sense, as a TAC failure might be caused by erroneous inputs that produce incorrect rudder movements. I’ll happily retract my previous thought bubble!

    Regarding the vexatious issue of turn tracking, the short answer is I can’t say for sure what happens in the case of a large offset combined with a track change of more than 90° – it’s not something we normally do in the aircraft. With zero offset, I would expect the aircraft to turn early to smoothly intercept the outbound track without overshoot, unless the waypoint is a fly-over type, where the aircraft must over fly the waypoint before turning. Fly-over waypoints are sometimes depicted on SIDs or STARs, but are rarely seen on air routes. However, I’m not sure how the FMC treats a turn where there is a large offset.

    What I can say is that the picture that is drawn on the pilot’s ND does not always reflect the path the aircraft will actually fly. It’s quite common to see a picture similar to that posted by Victor, where the track drawn on the ND shows an overshoot. However, the FMC seems to recalculate the turn as the aircraft approaches the waypoint and then turns early to smoothly intercept the outbound track. Sometimes the recalculated path is displayed on the ND, sometimes not!

    May I ask how you entered the flight plan into the FMC legs page? Was each waypoint entered separately, or were the waypoints NILAM and IGOGU entered as part of the route N571 as far as IGOGU, with ANOKO inserted after IGOGU? I’m not familiar with FSX or the PMDG model, but is it possible the simulator treats the waypoints as fly-over waypoints if they are entered separately, but fly-by if entered as part of a route?

    I hope to have some ‘play’ time in a Level D simulator late next week and, depending on the time available, will look at this and some of the other issues that have been mentioned.

  152. Ge Rijn says:


    No thanks to me. It was @buyerninety who posted an article on TAC..

  153. Andrew says:

    Ah, my apologies. Thanks to @buyeminety!!

  154. Andrew says:


    Regarding the LNAV heading vs track after a route discontinuity issue, according to Boeing:

    “LNAV remains engaged after the airplane passes the last flight plan waypoint of the first waypoint preceding a flight plan discontinuity. Upon passing this waypoint, a CDU message is displayed, LNAV guidance outputs a heading command and maintains heading until LNAV is disengaged through the MCP. If, after transitioning to this heading mode, a flight plan path is intercepted, LNAV guidance is again provided to the defined flight path.”

    That still doesn’t answer the magnetic vs true question – still working on that!

  155. Andrew says:

    Correction: First sentence in the above quote, “of” should read “or”.

  156. Richard says:


    “It could be inferred that you are making the following leap:

    1) 487 from 30s to SW of Réunion
    2) 161 from Réunion to Klein Brak
    3) therefore 548 from 30s to Klein Brak”

    No, I said 487 days to Reunion plus 161 days from Reunion to Klein Brak River = 648 days (not 548 days) from 30°S on the 7th Arc to Klein Brak River. This aligns with “Roy” being found after 655 days.

    Most Floating Debris from 30°S hits a central point in Madagascar, where it either goes North around the top of Madagascar or South around the bottom of Madagascar. Both routes experience very fast currents, which are frequently used by sailors to make extremely fast progress.

    The next area of extremely fast currents is near the coast of South Africa, known as the wild coast, where the Aguhlas current is also very strong reaching up to 5 knots. The Aguhlas current is very narrow and follows the continental shelf. Again sailors use this current to great effect when heading South by sticking to the line of the continental shelf and avoid it, by heading further out to sea, when travelling North.

    The path shown from Reunion to Klein Brak River is based on “actual” paths of drifter buoys from the GDP program. There is, of course, a difference between a drifter buoy and a flat panel such as “Roy”. The predominant wind at MossI el Bay is from the SSE at an average speed of 12.3 mph and maximum speed of 16.3 mph in December 2015, which would help “Roy” to beach.

    I did not use the Adrift data as this has been aggregated and averaged. I also noted the anomaly in the Adrift data around Mossel Bay.

    I use a database built out of the GDP raw data.

    My simulation does not use one efficiency factor for the whole trajectory across the Indian Ocean. The efficiency factor varies considerably depending on position and time of the year. I use different efficiency factors for each segment of the trajectory and each month of the year.

  157. Ge Rijn says:

    @Gysbreght @others

    A to you and others perhaps interesting PDF about i.e. actuation of elevator, rudder, flaperon, aileron etc. and failure modes starting at chapter 11.5.


    I found the ‘damped’ and ‘blocked’ modes especially interesting relating to the elevator, rudder and flaperon. F.i. flutter protection is mentioned.

  158. Victor Iannello says:

    @Andrew: The waypoints I entered were all “fly-by”. For small course changes, the turn began before the waypoint and there is no overshoot. For larger course changes, there is overshoot.

    From Ch 15, “Flight Management Systems”, in the Avionics Handbook:
    Transitions that are flyby but require a large course change (>135°) typically are constructed for a planned overshoot because of airspace considerations. Turn initiation and waypoint sequence follow the same algorithms except the course change utilized in the above equations is reduced from the actual course change to delay the leg transition and create the overshoot. The amount of course change reduction is determined by a balance in the airspace utilized to perform the overall maneuver.

  159. Brock McEwen says:

    @Richard: Thanks for the info on local currents.

    Re: your first para: by “No”, are you correcting my typo only (yes, 648 was my intent, quite clearly -apologies), or was there something else you disputed about how I summarized your logic?

    Other than to correct my typo, the only thing you seem to be disputing – seeming to correct my “SW of Réunion” to your “Réunion” – is confusing, because “my” SW of Réunion was taken directly from your report:

    “From a starting point on the 7th Arc at 30°S, the simulation shows debris arriving south west of
    Reunion after 487 days.”

    I think that the distinction between a location and SW of it is potentially profound. If we are ever to join you in setting a probability in excess of 50% on a particular impact point, it would be comforting to see strong probability density surrounding Réunion – not just SW of it – and for the Roy find to be likewise shown to be plausible via a single path projection – not two paths stitched together (with a disconnect in the middle).

  160. Gysbreght says:

    @Victor: “Transitions that are flyby but require a large course change (>135°) typically are constructed for a planned overshoot ”

    Is the course change at IGOGU >135°?

  161. Richard says:


    My apologies, I did not realise it was a typo, it was what you wrote.

    You appear to ask questions, that are not even there to be asked.

    If you do not accept my results that is fine, I am not asking you to accept my results.

    You asked for the information and I provided it.

    Can we just agree to disagree please!

  162. Victor Iannello says:

    @Gysbreght: I don’t think the 135° is a hard limit. There might also be some variation in behavior between FMS suppliers. The point is that overshoot will occur for large turns contrary to opinions expressed here.

  163. DennisW says:

    Yesterday, on the eve of the third anniversary of its disappearance on March 8, 2014, with 239 people aboard, Professor Charitha Pattiaratchi said that UWA’s reverse-drift modelling put MH370 “at Longitude 96.5 E Latitude 32.5 S with a 40km radius”.


  164. Brock McEwen says:

    @Richard: believe me: I am doing my best to understand your work. I am also doing my very best to ask questions clearly and politely. Please bear with me on both fronts.

    Are you able to get your model to drift debris directly from “Arc 7 @ 30s” to Klein Brak in 655 days?

    A simple “no” doesn’t kill your theory. Models can be conservative, storms can introduce paradigm shifts, and debris discovery dates and locations can sometimes be off. When I was putting together my aggregate drift study, I didn’t require a location to explain every single piece.

    All I seek is clarity.

  165. Brock McEwen says:

    @ALSM: Re: Jitter:

    My read of the BTO fact sheet – Figure 2 & its accompanying commentary in particular – is that items 2 (sawtooth) and 4 (rounding) are the dominant determinants of jitter error. I say “dominant” not because the other noise items aren’t present, but because they seem to be EITHER not present OR simply overridden by the rounding.

    Is this a reasonable interpretation?

    If so, is it fair to say that two BTO values logged 1 second apart are not, strictly speaking, independent observations, since their errors are statistically correlated via both the rounding and sawtooth functions?

  166. Brock McEwen says:

    @All re: 4600 offset:

    (all values in microseconds, unless otherwise specified)

    I actually prefer the nomenclature @ALSM uses in his paper: there are really 3 bias terms – one for each Channel Type. ALSM gives these as:

    T1200: 500,659
    R1200: 495,679
    R600: 491,079

    With the magical 4600 being merely the difference between the last two.

    While the first two were calibrated by the IG to the ISAT pdf’s early BTO values themselves, the R600 number was derived as “R1200 number minus 4600”. This is because 4600 is the number Inmarsat supplied separately for the difference between the two.

    Correctly, the IG did not just dutifully plug this number in, and carry on – they attempted to validate the 4600 against whatever was in the May 27, 2015 ISAT PDF for R600 signals at known locations.

    Sadly, the tarmac logon had been redacted from the file.

    Luckily, by December 23, 2014, these records were supplied. So now we had R600 records at known locations. As @ALSM’s paper correctly shows, the 4600 value seemed to be a near-perfect fit, given differences between parallel R600 and R1200 values at 16:00 UTC.


    1) I don’t think it was optimal to validate the 4600 by comparing adjacent R600 and R1200 values. By doing this, we have two instances of jitter error, one being subtracted from the other. Better might have been to solve for the R600 such that the adjusted BTO matched the known round trip gate-sat-Perth-sat-gate time at the speed of light (14,801 us, I believe). This method back-solves to 491100, not 491079. This equates to a differential vs R1200 of 4579, not 4600. Not a big deal, but not negligible.

    2) There were only two such BTO values in the “ground truth” portion of the data log – and these were only 1 second apart. I suspect that the reason these were both the exact same value (19380) was largely because these had the same jitter contribution. So we really just have a single point estimate for this increment – 4579 – and we know that this is +/- whatever jitter value happened to be buried within the 19380. If the DSTG paper is to be trusted, this embedded error has a sigma of 62. So our “ISAT log-estimated” 95% confidence interval for the 4600 is actually [4457, 4701].

    3) While it is fine to step back and say, “well, sure, if that’s all we had, we’d need wide error bars – but we already have Inmarsat’s averaging of this increment over “many” historical records. How many is “many”? I have not seen the raw data behind the 4600 – have any of you?

    4) The December, 2014 ISAT pdf contains 519 BTO values, broken down thus:
    Known locations (pre-17:30): 499
    T1200: 419
    R1200: 78
    R600: 2 (spaced 1 second apart, so likely not independent)

    Unknown locations (post 18:22): 20
    T1200: 6
    R1200: 12 (incl Arcs 2-6)
    R600: 2 (18:25~Arc1 and 00:19=Arc7)

    So, while the opportunity to independently validate R1200 was ample, the opportunity to independently validate R600 has been extremely limited.

    Inasmuch as the search box centred on Arc7 – and inasmuch as the 18:25 R600 BTO seems low by ~100 after reducing it by the full 4600, I think we should press for the raw data behind Inmarsat’s derivation of its 4600.

    If actual was closer to 4500, the search box was centred 20km too far NW.

    (This is why I am so interested in getting precision on deceleration dynamics while on one engine, Brian.)

  167. DrB says:


    The turn from IGOGU to ANOKO is 116 degrees, which is smaller than the 135 degrees in the information supplied by Victor.

    Maybe Andrew will be able to tell us what the Level D Simulator does.

  168. Victor Iannello says:

    Comments here are closed. Please continue the discussion under the new post.