Bravo J-Class
 

As someone whose professional life is all about researching cognitive bias, I find few phenomena as disconcerting as pilots turning a blind eye.

Airbubba
 

I'm not an engineer either but certainly on the 777 GPS timing is used by the airplane information management system (AIMS). Even the flight deck clocks/chronos are effectively "slaved" to the GPS timing signal via AIMS.

( Source is the 777 FCOM 2)

Gysbreght
Just to keep this pot boiling, some comments to try and back the position as stated in post #73! I will assume that the main issue is whether an exact match of a model to the data is good or bad.



I can only speak from my experience and training in analysing physical data. All data is noisy and dealing with the errors is a big part of the data interpretation. There is always the risk of developing a complex physical model with lots of variables that the analyst is trying to get out of the data, e.g. in the case of a plasma, the temperature (sometimes several), density and chemical composition. If there are only a few data points, the model can be so complex that almost any data, including the noise, can be fitted by changing the variables to some set of values. If this happens, then the model is too complex and very little can be extracted from the result. So if any model can exactly match the data values as measured (assuming there is more than one data value), the analyst should be suspicious. One of the major reason for having a model is to be able to recognise the errors in the data and deal with them.


So the predictions of the model, when compared to the data, must still show a disparity/difference that is consistent with the known errors in the data.


In my experience it is usual to keep the comparison of the data and the model in the frame of the data where the errors are well understood. If others wish to do it another way, that’s fine, but the errors have to be transferred in some fashion because they have not gone away.

Hyperveloce
That group will have to speak about the BTO/BFO comparison themselves. It doesn’t match the model I have proposed (it’s miles off the BFO model), otherwise it would have appeared in my paper as a potential solution consistent with the data.

The radar tracks

Someone asked a while ago about the consistency of what I have been proposing with the radar track along the Straits of Malacca. I have investigated this by fixing the latitude of the 18:29UT ping to be North of Sumatra. The two speeds in the track model set are then the speed of the second leg from 18:29UT to 19:41UT, and then the speed of the subsequent legs (all the same). Again, I compare the BFO predictions of each track with the data. Statistically, the solutions are the same as those in my original paper. However, in this case the required speeds of the second leg to be consistent with the data are an output of the modelling. I have plotted the speed on the second leg against latitude of the final point at 00:19UT, i.e. position along the 00:19UT ping ring, for the statistically acceptable solutions. The result is shown in the graph below.

final_search_areas_zps2591eb0d.jpg Photo by RichardC10 | Photobucket

This is a complex graph, so some notes:

a. All of this is constrained by the need to be consistent with the radar data.

b. The blue points are solutions to the 3 ping fit (20:41 to 22:41UT) in the paper. With the reported search areas shift to the South these are less important.

c. the brown points are solutions to the 4 ping fit (19:41 to 22:41UT). This corresponds to section 7.7 of my report.

d. the green points are solutions to the both the 3 and 4 ping fits.

e. the red line I have drawn in shows the minimum 2nd leg speed to get to a particular latitude. The aim is to show just how fast the aircraft has to travel on the 18:29UT to 19:41UT leg to get it South enough for the rest of the BFO values to match.

f. the aircraft has to be at the speed for the later legs by 19:41UT, that is whatever the average speed between 18:29 and 19:41UT, the aircraft has to have slowed to the later leg speed by 19:41UT. If it is still at the higher speed the fit at 19:41UT is very bad.

g. If the aircraft flew North of Sumatra, the minimum 2nd leg speed to get to the Northern red-zone search area at ~21degS is 410kt, so lower than the speed at loss of contact, so not a constraint.

h. However, to get to the Southern areas now being mooted (say 28S), the minimum average speed on the 2nd leg is 540kt (given that the speed of the aircraft has to be at much lower value of <330kt by 19:41UT, the maximum has to be greater than 540kt). The minimum average speed on the 2nd leg to get to the Southern edge of the ICAO green zone is 500kt. This speed may be a limit set by the aircraft analysis

If the constraint to be consistent with the radar data is dropped, the course can be much more Southerly, as indicated in the maps of the original report. Using the 4-point BFO fit, the goodness of fit of the possible tracks is shown in the graph below. A speed limit of 500kt on the first leg is used (the figure used above). The 90% confidence limits give a range of 23S and 30S along the 00:19UT ping arc for the final destination.

confidence_limits_zpsa555d4ad.jpg Photo by RichardC10 | Photobucket

So here is a risky prediction, based on a model I have not been able to validate. So what can go wrong?

1. The Southern edge of the ICAO green-zone was set by a maximum 2nd leg speed of 500kt, with the possible tracks set by the need to be consistent with the radar data.

2. when the new search area is declared, it will be between 23S and 30S, which corresponds to the 4-point ping fit.

To be clear, this comes with a big health warning as I think I am making the model too complex for the available data.

Also, again to be clear, the model proposed in the paper is a hypothesis of what the Inmarsat analysis process might have been. Apart from the published MH370 data log, I have no other data to validate the model, so cannot claim it is verified and hence that any prediction from it is correct.

Three more questions...

- is it known how the BTO data published ("raw data" PDF) were measured? It is evident that the digits are not the whole go and back signal time. Regardless what is substracted to the whole trip of the signal, it have to be at least two different very accurate time stamps: the first when the signal is sent by the GES, the second when it comes back. So, how could be computed the BTO at, for example, 18:25:27,421 (value is 17120 usec)? It is "the end of the link lost period that began at sometime between 17:07:48 and 18:03:41". It is a RX from the aircraft, without any TX from the GES before which could have initiated the RX (and timestamped). 1.431 second later, the GES answers (Log-on confirm), followed by a sequence of 4 TX. When was measured the "17120 usecs"? And between which times?

- is there any possibility that the "activity" of the satellite unit (when it logs on or when it answers a "ping" from the GES) could be "viewed" from the cockpit, even on a sub sub screen of a normally never used display?

- has it been published something accepted or "credible" about the reason of the "re"log at 18:25?

There is a post on the DuncanSteel blog where some infos are written about the communication link, particularly this:
Is this seems correct? Could it be applied to a passenger door blown (or opened) in flight?

That is five questions, actually...

My question is why didn't they send long range drones to inspect the floating debris sites more closely? Or did they? It seems to me that this whole search seems full of loose ends and lacks thorough investigation. eg. RQ-4A Global Hawk with 3,000 mile range. Apparently it can survey as much as 40,000 square miles (100,000 km2) of terrain a day.

The false pings still don't appear to be clear as to their cause. Yet have they repeated the tests to verify the source of the false pings? i.e. ships bar fridge, other part of ship or the TPL-25 itself?

Could there be a public checklist reporting web site that site that highlights every key piece of evidence and states whether the evidence has been ruled in or out and provide the reasoning as to why. Seems the data needs to be open to public scrutiny.

Seems to me that exhaustive and complete verification is required, if they are to eliminate the noise and really focus on the most probable likelihoods. Their current approach seems quite spurious.

Hey, discussed, explained, and dismissed months ago.

Cloudless, ground stations, analysis, etc.

Why should there be a public website? I am sure the coordination centre has it all recorded on as dag packet. It is for the official report to cross I's and dot t's.

Well the latest is they are now saying the hunt could take decades. source:Telegraph.co.uk

The mind boggles that they could not find any surface debris or clues in the initial period after the plane was reported missing. Now they say decades? Seems to me that maximum effort would have been best spent early and not over decades.

Initially no one had any idea it had flown for hours to nowhere; it's pretty difficult to see how anyone could have done more sooner.

Admittedly too much time was wasted in the wrong area because the Malaysians didn't believe it hadn't just crashed on route to China.

It's clearly a slow process now to map the sea floor and look systematically in areas thought promising. They might get lucky, they might not.

You seem to be criticising the search because the outcome wasn't to your liking. That doesn't mean it was wrong.

I don't believe the ship borne false ping scenario... This would have been spotted at a very early stage as they were lowering the detector into the water....


Just dropped it in to the water, Pings strength 9++.
Lower the cable, Pings strength 5.
Even lower still, Pings strength 2.


Someone must have spotted that.

Officials say MH370 was in controlled flight
 

New York Times article, June 23rd.“MH 370 Was in Controlled Flight After Contact Was Lost, Officials Suspect”
Link > http://www.nytimes.com/2014/06/24/wo...=29625191&_r=0
SOME HIGHLIGHTS;
“Was probably not seriously damaged in the air and remained in controlled flight for hours after contact with it was lost, until it ran out of fuel over the southern Indian Ocean.”
“Their conclusion, reached in the past few weeks, helped prompt the decision to move the focus of the search hundreds of miles to the southwest.”
“ main evidence for the conclusion lies in a re-examination of Malaysian military radar data and in a more detailed analysis of electronic “handshakes,” or pings, that the aircraft exchanged with an Inmarsat satellite over the Equator, senior officials involved in the investigation said.”
“The altitude readings from the radar now appear to have been inaccurate, officials said.”
“Malaysian radar equipment had not been calibrated with enough precision to draw any conclusions about the aircraft’s true altitude. “The primary radar data pertaining to altitude is regarded as unreliable,” said Angus Houston.”
“Other officials involved in the crash investigation have suggested that either of the plane’s pilots might have commandeered the aircraft in order to commit suicide, or that a smoke from a fire in the fuselage might have overcome the pilots and passengers but left the engines and autopilot working normally.”
Data Guy Notes; Boeing service letter cautions to fuselage fires may be seen as Appendix 1 (Chinese) Aviation Safety Council Final Report, fire on EVA Airways, 2/23/08, Flight BR67. Report # ASC-AOR-11-02-001.Link > http://www.asc.gov.tw/downfile/ASC-A...1(English).pdf

As the aircraft is now assumed to follow designated waypoints as it flew NW it is likely that it was in LNAV.

If the assumption now is that the aircraft flew at cruise altitude and normal speed it follows that there could be 2 endurance times and distances
1 Pack on
2 Pack off. With the bleeds or packs off the endurance would be extended due to better fuel economy.

How does this tie in with the final ping timewise (which equates to empty tanks). On that basis it should be possible to draw a series of circles from the last probable waypoints; MEKAR IGREX or the FIR at 92E into the southern Indian Ocean.

Search to take decades?

On whose dime? :confused:

This question is based on the practical matter of effort versus benefit. If the estimation is that the aircraft was in control and that a human agent was involved, is it really necessary to find the aircraft to arrive at some means of preventing such incidents in the future?

At some point, when dealing with the human mind, you can't engineer a "solution" to how people think and behave.

Looking for various loopholes that may have been exploited does not IMO require finding the aircraft on the ocean floor.

There is a major issue here, that professional pilots need to take note of. That is the level of trust that people feel that they can put in the flight crew that have locked themselves away in the cockpit. This is the real reason for the need to find the airframe.

As we see here immediate offense taken at suggestions that the pilot may be the one that 'hijacked' the aircraft (and for the same reason this may be modded out). This immediate reaction is the reason that no FOQA or DFDR is streamed and that no CVR data is streamed from the cockpit. It is also the reason that video recordings are not made and streamed. IFF this was a pilot initiated incident then I suspect that legislative and insurance company action will be taken to enforce such data streaming. IFF it was not pilot initiated then perhaps trust can remain in pilots.

Taking offense at the suggestions and 'turning a blind eye' to the problem will not make it go away. The current mood is that it was pilot initiated. Without finding the wreckage and the recorders and numerous personal cell phones with recordings that mood will steadily swing toward less trust being given to pilots, and eventual enforcement of streamed recordings.

It is in everyone's interests that the wreckage is found and the reason for what happened is discovered.

Thanks for looking this up, I appreciate it. :ok: I've never flown the Triple but have operated several other Airbus and Boeing glass products over the years.

I agree, not sure why these arcane technical posts about GPS timing survive but any discussion of possible human factors seems to be taboo in the modern commercial incarnation of PPRuNe.

I remember reading of the likely cause of the Silkair 185 crash here on PPRuNe a few days after the mishap from a poster with good sources and, sadly, family members on the plane. I also remember the healthy debate on the Egyptair 990 mishap.

Rumor in U.S. federal law enforcement circles has it that an FBI report on MH 370 is in limited distribution with analysis from the BRIU and computer forensics folks at Quantico. Perhaps some of the recent news reports stem from this alleged document which may be very speculative.

BFO model
 

Something I don't understand about the D3 compensation: what is the purpose to send a pilot frequency from Italia or the Netherlands, so that the 3F1 satellite measures/observes the doppler shift versus a point in Europe
1) when a shift versus Perth has to be corrected ?
2) when the actual compensation seems to be implemented, not onboard the sat, but on the ground in Perth ? (why send a pilot freq. to the sat ?).

The south trajectories outputed by the MC sim at 470 kts trying to fit (best fit in green) all the BTO/BFO values (red plots) excepted for 16:55 and 18:27): they remain compatible with the radar track, cross only the northern tip of Indonesia (80 km east of Banda Aceh).
https://drive.google.com/file/d/0B3s...it?usp=sharing
https://drive.google.com/file/d/0B3s...it?usp=sharing

Ian W, well answered on why there is global aviation industry value in finding out what physical evidence may tell about what happened, and a few hints at who and how.

Whose dime? My offered suggestion is Malaysia, and the Flag Carrier Airline of that country. Put the burden of cost on those with the formal oversight for safe and proper operations of their various aircraft and crews.

This may sound harsh, but I think is realistic and just.

"BFO model":
What I understand is that there are two ways for the Doppler compensation from the aircraft unit:
- "hearing" a [constinuously?] transmitted signal from the satellite which frequency is known: from the difference between the supposed frequency and the received one, the shift of frequency to emit is computed,
or
- computing the necessary shift from the flight speed and direction, knowing the theoretical position of the satellite. From what the radial speed between the aircraft and the theoretical position of the satellite is computed has not been, IMHO, stated (GPS and/or ADIRU). What is admitted about this is that this compensation is faulty because the satellite is not in this position (north of it all the time of the flight) and because it moves.

It has been said that the first way was using a whole canal to do that, and so is not "economical" and MH370 used the second way which is not asking for a pilot frequency.
Without correction, the radio signal uses a larger bandwidth, so this correction allows much more canals inside the same band width.


Another thought: IMHO, people using precise BFOs values to guess a flightpath are surevaluating these data: or you take the option of a single direction and a single speed value after the [supposed] turn to south (that is, aircraft on AP), and you don't need BFOs to find a flightpath, or you presume changes in direction, speed and possibly flight level, and you can find such a large bunch of possible flightpathes that you can cross the 7th arc from Sumatra to something like 40S 70E, and flying a hundred miles more after fuel exhaustion (IFF the last ping was generated by fuel exhaustion, APU starting, ..., and not from the "same something" which made the system reboot at 18:25:xx).

A communications engineer would be better placed to comment, but a couple of points:

a. it seems that the system design wanted the pilot frequency generators (for correction of the uplink and downlink Dopplers) on the ground rather than on the satellite. Perhaps this was to give flexibility in changing the frequency if needed, possibly due to interference, or to avoid having to provide on-board redundancy for that function.

b. the pilot frequency is received on-board at a shifted frequency, but this shift will be measured by the L-band receiver, so can be removed from the total Doppler on the pilot frequency when it is received at Perth.

This system concept seems to provide the required output parameter of the downlink Doppler.

Purpose of pilot signal
 

My understanding is that the pilot is used primarily to measure frequency errors induced by the satellite when up or down converting between L and C bands. Doppler-induced errors are a lesser problem. Pilots won't originate on board the satellite since the idea is to mimic the end-end communications link.