At a press conference in Canberra today, Australian Deputy Prime Minister Warren Truss state that “further refinement of satellite data” indicated that missing Malaysia Airlines Flight MH370 had turned south earlier than investigators had originally thought. This implied, he said, that the plane had most likely would up further to the south than previously estimated.
The previous assumption was laid out in a report released in June by the Australian Transport Safety Board (ATSB), which included the map shown here. The document described a methodology for determining the search area which suggested that the plane did not make a single turn to the south and then fly on a straight-ahead course into the southern ocean, but rather lingered near western Sumatra for the better part of an hour.
After the report was issued, a loose coalition of experts from around the world called the Independent Group (of which I’m a part) released a statement which questioned the ATSB’s methodology, and in particular pointed out that signal data related to an attempted satellite phone call at 18.40 UTC indicated that the plane was already established on a course to the south. This fact allows the range of possible flight paths to be narrowed considerably.
As fate would have it, Truss’ announcement came just one day after the Independent Group issued a follow-up statement reminding the authorities that its own analysis suggested a search area futher to the south.
“The data is nothing new, but the fact that the Australian government has chosen to issue this statement is very interesting,” says Independent Group member Victor Iannello.
Truss was vague as to where the priority search area had shifted, saying it remained “within the search area” previously laid out in the southern Indian Ocean. This area, however, is more than 1,500 miles long.
I will in this post outline the evidence showing the pings were transmitted after the plane had crashed.
From the NYT article on March 22nd:
“Chris McLaughlin, a vice president at Inmarsat, the satellite communications firm, said technicians pulled all the logs of all transmissions from the plane within four hours of its disappearance. Then, after a day without sign of the plane, they began scouring the company’s databases for any trace of Flight MH370. ‘We decided to go have another look at our network to see if there was any data that we had missed,’ McLaughlin said. It turned out there was, Inmarsat technicians identified what appeared to be a series of fleeting ‘pings’ between Flight MH370, a satellite over the Indian Ocean and a ground station in Perth, Australia…”.
This was how the Washington Post on March 30th reported on this belated discovery of the pings:
“…That same Sunday, sitting at his kitchen table in Potomac, Md, John Mackey started to realise the enormousness of the unfolding mystery. He runs network operations for Inmarsat and shuttles between its offices in London and Washington. As the 47 year old engineer drank his coffee and ate a bagel, he scrolled through the emails on his smartphone. One was from the company’s ground station in Australia. A technician had captured some data from a Malaysian flight. Then Mackey read the news about Flight MH370…….”.
Thus, the post 1707 UTC transmissions, the pings, did not even appear in Inmarsat’s data logs in the first place. What could possibly be the reason for this?
The signals, as I understand it, are burst of radio waves. The only possible reason why the pings failed to register on Inmarsat’s data log when the logs were first pulled, would be because those signals were very weak in signal strength. Only faint traces were detected, too faint to register on the logs, which is exactly as reported in the NYT.
How is it possible that the signal strength of the pings was so weak, if they were transmitted when the plane was still in the air?
If the plane was still flying during those 6 or 7 hours, the engines would have been running and the pings would have been transmitted on the plane’s regular AC or DC power which would mean the signal strength of the pings would be more or less the same as the prior transmissions.
One or two weak pings may be explained away- the plane was turning or banking at that moment in time etc, but these pings were transmitted over 6 or 7 hours and they all failed to register on the logs.
In my humble view, there could only be one explanation, the pings were transmitted over that period of time, on battery power and not on the plane’s regular AC or DC power. Which would lead to the conclusion that the plane’s engines were not running during such period of time.
Incidentally Inmarsat have refused to disclose the data on the signal strength (Rx DBm) and the related information C/No (noise) and BER (bit error rate). These fields have been excluded from the 47 page data log released on May 27th.
In my next post, I will set out several “mathematical observations” about the BFO and BTO data for the pings showing that such data merely reflect the movement of the satellite against MH370 which had already crashed by then and that such data were transmitted by an SDU which was no longer getting the necessary input from the plane’s AIMS for proper Doppler correction after the catastrophic event at 1.21am at IGARI.
There is a typo in the quote from NYT. “…pulled all the logs of all transmissions…” should read “….pulled the logs of all transmissions…”
Yes, I may be a bit hard at it.
Things do have a way of evolving and cliques often expand. That’s usually a good thing for accurate assessments.
I’ve been at it since March also, so don’t feel lonely.
Anytime you want to discuss any of my opinions, I welcome and encourage that.
Just bring your facts and we’ll have at it.
Some have floated the idea of a massive mechanical failure to explain the MH370 event.
I don’t subscribe to any such notion. Just for information purposes I have posted a partial description from Boeing re: one particular failure scenario:
“If both IDG’s and the APU generator are inoperative, a backup generator powers essential airplane equipment. To reduce electrical loading on the backup generator, the following systems are inoperative:
TCAS, SATCOM, Right HF radio, ……
Source: Boeing B777 – Systems Summary (Electrical)
Alex/John – Bloody hell, It would mean we have to account for the radar data but I remember speculating a while ago that there may be any number of planes in the air in SE Asia without transponders at any time. That may help explain the lack of a reaction from Malaysian Air Force? The floating ping theory was put up by someone a while back but somewhere west of Malacca Straits.
It’s the first time I have thought about the pings going unnoticed. Where are the batteries? Would the satcom antennae have it’s own? Could it have drifted as a piece of floating wreckage faintly bouncing the signal?
Matty – Perth
Where’s the conclusive evidence re: primary radar data for MH370?……..There isn’t any.
I hope Alex will continue……
Pretty much my point. For a while I’ve had a hunch that the SIO search could be wildly wrong. Then there is the old rule – plane’s are almost always found pretty close to where you lose contact.
@Matty, thanks for defending our tone of cordial exchange and discussion here.
@Matty – Perth
I agree.
Lots to consider or reconsider.
Alex said, “The only possible reason why the pings failed to register on Inmarsat’s data log when the logs were first pulled, would be because those signals were very weak in signal strength.”
I think a more likely reason is they were looking for an exchange of ACARS data, not login/handshake activity.
VictorI,
The transmissions for the first period (1707 UTC and prior) also included ‘log in/handshake activity’. See the entries at the beginning of the 47 page data log, referred to in my previous comments. There is no indication these transmissions did not show up on the logs when they were first pulled, in contrast to the post 1707 UTC ‘log in/handshake activity’. The transmissions for the prior period also included a failed uplink message at 1803 UTC. The NYT article mentioned only the post 1707 UTC pings as the signals which did not register on the logs. In addition to not being found on the original data logs, the pings were also described as ‘fleeting’ and indirectly as ‘traces’.
@Victori
Not sure I can necessarily agree with you……
“The search for MH370 would have been hopeless had it not been fitted with a system called Classic Aero, a type of Aircraft Communications Addressing and Reporting System (ACARS) which transmits data on location, altitude, heading and speed.
ACARS, as we now know, can be turned off manually, via a switch on the ceiling of the cockpit or behind the throttles between the pilot and co-pilot. MH370’s ACARS system was switched off at 1.21am on March 8, two minutes after the pilots’ last verbal communication with the ground. It is this deliberate act that has convinced investigators the pilots were on some form of suicide mission.
But Classic Aero also has a second terminal that operates independently of ACARS and cannot be switched off while the aircraft still has power.
Once every hour the system sends out a “ping” to satellites operated by Inmarsat. The pings play no part in ACARS, and merely serve to synchronise timing information and keep the connection to the satellite network alive.”
I think they knew what to look for.
A log is a log. A transmission is either entered into the log or it is’nt. There are no ‘weak’ log entries. Most probably Inmarsat was’nt looking for any log entries hours after the airplane disappeared from ATC radar in their first ‘pull’ of the logs.
@John: There is no “second terminal”. There are just layers to the communication protocol. The ACARS is an application layer that rides on the SATCOM protocol when the ACARS is routed via satellite. When the ACARS was disabled, the lower level communication continued, as evidenced by the handshakes.
Most likely, the logs are quite lengthy as they have ALL the activity for all planes. A filter in the form of a database query would be applied to extract the relevant records, and the filter might have been based on ACARS criteria. Later, the criteria might have been based on the SDU ID, which revealed the lower level handshakes. I don’t see this as mysterious at all.
@Richard: the topic Matty & I were discussing was the feasibility of the path the ATSB used to defend the site they chose to listen for pings. They published this path on May 1; it had LISTED speeds of 323KTAS in its “southern arc” phase. If the Inmarsat pings are to be relied upon, then those speeds are FORCED by that location.
Your model would help defend a decision to put the listening devices in at YOUR result (s33?); but instead, they chose s20 – 1,700km NE of s33.
The ATSB moves its detectors to a place the plane could not possibly have reached (if the Inmarsat pings were authentic). They then detect sounds the plane’s black box could not possibly have made. Yet the Australian PM loudly proclaims them authentic, and an eight week search ensues. And you find nothing suspicious about this?
John, Matty,
1. There is no evidence of any radar having picked up MH370 after 1.30am.
2. The Malaysians claim that their Butterworth primary radar detected a plane which made a turn back at around the area where MH370 went missing and that this aircraft then crossed back over Peninsular Malaysia.
3. On closer scrutiny, all that was detected by the Butterworth radar was a blip coming from the other direction heading to KL, which blip had emerged at the edge of that radar’s range subsequent to MH370 disappearance. From the NYT article again:
“…As air traffic controllers struggled to re-establish contact with Flight MH370, military radar at the Butterworth air force base on Malaysia’s west coast picked up an unidentified aircraft near where the plane disappeared. But the watch team, normally an officer and three enlisted personnel, either failed to notice the signal or decided not to designate and track it as a ‘zombie’, which would have pushed the information up the chain of command and possibly alerted air command. At a briefing on the base the next night, about 80 air force personnel were told there was ‘no proof’ the unidentified signal showed the missing plane making a sharp turn, flying back across Peninsular Malaysia and then turning again and heading northwest over the Straits of Malacca, a person familiar with the situation said. But the investigators now believe that is exactly what happened….”.
4. When the story about this pickup by the Butterworth radar first surfaced, the Malaysian air force chief Daud said there was a possibility of a turn back on a reciprocal heading ie back to KL and that the blip was later spotted over Pulau Perak at 2.40am.
5. Butterworth was not the closest radar to IGARI/BITOD. A much closer radar and the radar tracking MH370 from 1.21am to 1.30am from IGARI to BITOD, was the military radar at the Gong Kedak airbase at the northeast of Peninsular Malaysia.
6. The Gong kedak radar’s range extends all the way to the coast of southern Vietnam whereas the Butterworth radar’s range ends at around IGARI (see the recent post by Duncan on the various radars).
7. The ATC at Subang (KL) would have the returns of all civilian and military radars from around the country ie from both Butterworth and Gong Kedak in addition to the civilian primary TRACON radars and secondary radars co mounted at the various airports including the airport at Kota Bahru.
8. Neither the Gong Kedak radar station nor the ATC at Subang reported seeing any aircraft making a turn back or any unidentified blip that could have been MH370.
9. What was spotted on the Butterworth radar in all likelihood was CES539 a flight from Shanghai to KL which was coming from the other direction and which was only around 100 NM east of MH370 when MH370 went off secondary radar at 1.21am at IGARI.
10. The Butterworth radar tracked CES 539 heading to KL and from KL another flight most probably KLM836, which departed from KL around the time CES arrived at KL and which flew over Pulau Perak at around 2.40am.
11. To the people manning the Butterworth radar which was a primary radar, a blip (MH370) had disappeared at the edge of the radar range, and subsequently a blip had emerged from around the same area (CES539) and this blip was tracked heading back to KL and from KL a blip (KLM836) was tracked heading to Pulau Perak. So the deduction that this series of blips could have been MH370 heading back to KL and then later turning northwest to Pulau Perak.
12. For the people at the Gong Kedak radar station this blip (CES539) was tracked all the way from the southern coast of Vietnam so there is no issue of any turn back and for the ATC at Subang which would have the SSR signal from CES539, this aircraft was CES539.
@Victor1
My references came from here:
http://www.telegraph.co.uk/news/worldnews/asia/malaysia/10720009/MH370-Britain-finds-itself-at-centre-of-blame-game-over-crucial-delays.html
While I’m not sure Alex is on to something my only point was I believe Inmarsat knew what to look for.
Your interpretation is certainly plausible though.
Weak or not, the BTO data which are well understood, seem to indicate, that the plane moved further and further away from the satellite from 19:41 onwards. At each ping it was further away from the satellite. A stationary or floating plane cannot have increased these distances by that much, but a flying plane can do that. That the satellite moved first north, then south, thus mimicking to some degree the alledged movements of the plane, is an interesting observation. But the satellite’s movements have nothing to do with the simple fact, that the reaction time of the pings got longer and longer from 19:41 onwards, thus indicating an object moving as fast as a flying plane.
If we take that observation at face value, we simply have to conclude that the plane kept flying for hours. Of course the SDU could’ve been compromised somehow, thus skewing the reaction time systematically. But isn’t it highly unlikely, that the SDU kept pinging for hours, consistently indicating a speed, which can only be explained by a flying object?
I think your observations are interesting, Alex, but IMHO they can’t be explained by a SDU which somehow remained functional and afloat for several hours after a crash of mh 370 in the South China Sea.
VictorI,
I posted a comment to u earlier in reply but it seems it did not go through, so i will try again here.
1. The transmissions for the first period (1707 UTC and prior) also included ‘log in/handshake activity’, see for eg the entries at the beginning of the 47 page data log as well as the failed uplink message attempt at 1803 UTC.
2. There is no indication those above referred earlier transmissions were also missing from the data log when first pulled. The NYT said to the effect only the post 1707 UTC pings were not in the original log and that these pings were ‘fleeting’.
Alex: No doubt Malaysia has made a series of inaccurate and misleading comments about the primary radar and other matters. One could speculate as to whether it was a deliberate attempt to mislead or whether it was to hide incompetence. There is some question as to whether ANY of the radar data published was from Malaysian radar.
On the other hand, Thailand has supplied radar data, and has even questioned why their data was not requested until days after the disappearance. Many of us believe that the published primary radar data came mostly if not completely from Thailand.
See this link:
http://www.dailymail.co.uk/news/article-2583553/Thai-military-says-missing-flight-MH370-followed-twisting-path-Strait-Malacca.html
John, these newspaper articles are interesting, but there were so many stories floating around in the early days, so many contradictions, that we could fill several books with them. Even quality news sources were engaged in that game. And all this confusion and misinformation floating around is certainly unique in aviation history and definitely worth of more scrutiny. But we can’t build a whole theory on one or two of those pieces.
@Alex
Still digesting your entire concept, but agree completely about the radar data.
I don’t believe much of it can be trusted at all. (Perhaps more importantly, the interpretation thereof)
@ littlefoot
Not trying to build any theory from newspaper articles.
Not especially with my experience with the media.
Just posting a source (the second terminal is really irrelevant). What can be gleaned I believe is that Inmarsat knew what data to look for. Simple as that.
In fact Alex’s idea conflicts with my own personal theory as to what happened.
@Alex, don’t get hung up on the word ‘fleeting’. That’s journo speak and not tech language. I think they wanted simply to illustrate the fleeting nature of immaterial pings compared to something material like debris. I certainly wouldn’t read into it a comparison between different sorts of pings. And again, it’s only a newspaper article.
Alex said, “There is no indication those above referred earlier transmissions were also missing from the data log when first pulled. The NYT said to the effect only the post 1707 UTC pings were not in the original log and that these pings were ‘fleeting’”
We have no idea exactly what data was extracted in the first query. I believe the NYT statement is consistent with a second query that included an SDU ID filter as opposed to an ACARS filter.
@ John, point taken. As to theory building, that was meant more for Alex than you. Especially since you have not disclosed your favored scenario so far 😉
@littlefoot
I wish I could disclose it, but that would be cruel I believe since relatives of my suspect actor still await news about their loved one.
I found it rather repugnant when some ran around pointing the finger at the Captain or Co-pilot. I mean, how would you feel?
I’ll only say no one has publicly stated they have developed the same theory as I have. I’ll also say just this………I believe the plane was headed to Perth.
To understand how the pings could have been transmitted, one needs to know the SATCOM architecture on board MH370.
1. The SATCOM system was fitted by Honeywell. It would be from the MCS series developed jointly by Honeywell and Racal, I do not know the exact model but if i have to guess i think it was the MCS 4000/7000 series. The manual for the MCS 4200/7200 series described as an ‘augmented’ version of the former, can be found online.
2. The SATCOM configuration on board MH370 was a dual SDU dual antenna configuration. The primary or master SDU was linked to the High Gain Antenna via the High Power Amplifier and the Beam Steering Units. The back up or slave SDU was linked directly to the Low Gain Antenna.
3. This dual configuration is provided under ARINC 741 the governing standard for SATCOM systems. The slave SDU/LGA subsystem acts as a back up to the main SDU/HGA subsystem and would only come on when the main system has failed. The back up system can only transmit ‘low rate packet data’.
4. The SDUs were located at the rear of the plane, above the overhead luggage compartment. The LGA shaped like a shark fin was located just above or near this part of the plane, on top of the fuselage. The HGA was a dual side mounted antenna from Ball Aerospace, located above the door to the rear (shaped like a plate).
5. The MCS series SDU has an internal battery. Its main purpose would appear to be to power memory. I have not been able to get the specifications for this battery.
6. Primary power for the SDU would be the plane’s regular AC and/or DC power. The SDU also has inputs for secondary power sources although it is not known to me whether the SDUs on MH370 were hooked up to a secondary source of power or a separate stand alone battery in the form of a UPS, whether by itself or as part of a larger group of related avionics.
7. A lightning strike that managed to penetrate into the plane’s systems would have fried all the wires/cables and knocked out the primary systems functioning at the time of the strike.
8. This would appear to be what happened to the plane as none of the systems ever revived.
9. The SDU that transmitted the pings had to be the back up SDU transmitting directly through the low gain antenna, as the back up system was designed to so do.
10. The main system was out already, ACARS never revived.
11. Transmission of basic level signals through the back up system would require a lot less power than the transmission of signals through the main system which would involve the High Power Amplifier and Beam Steering Units.
12. The back up SDU on MH370 after 1.21am would not be any different from a handphone that had lost all its main apps but was still powered by a battery.
13. For the handphone, as long as there is battery power, the handphone can still respond to the pinging by a cell tower by emitting the standard reply signal to the cell tower.
14. I would contend that the back up SDU was in the same situation, it would have been able to respond to the satellite hourly pinging and transmit the “I am here” signals or pings, as long as there was battery power. Essentially that was what the back up SDU did, after 1.21am.
@John, thanks for at least outlining the nature of your theory. So, you do believe that the plane kept flying and was heading southwards?
Discussing certain scenarios almost always involves finger pointing – at the pilots, the crew, at some of the passengers, at the ineptness of the authorities – it depends on what scenario we favor. I see your moral dilemma. I personally choose to discuss all possible scenarios, no matter what, because I believe that by discussing them with others, plausibilities can be checked. I think a lot can be said in favor of ‘The-Captain-did-it’ theory. It was my favored scenario for a long time. But because of discussions at this blog I changed my mind. While I still don’t entirely exclude the possibility, that the captain was involved to some degree, my theories certainly have changed and I see this baffling case in a different light.
The only scenario I personally exclude is the random-disaster theory. I think this was premeditated by one ore more perpetrators.
@richard
A couple more points that I’m wondering if you could shed light on:
1. We have recently been told that data traffic from a call made to the flight deck satellite phone (which was reportedly not picked up) has provided additional position/heading information, leading investigators to tweak the inferred flight path. What is the nature of this data? And, why did it take almost six months to integrate this data into the analysis? Wouldn’t digital data from all post-diversion communications with MH70 be one of the first things investigators would look at to figure out the trajectory of the plane?
2. Have investigators been given full, unrestricted access to all digital communication traffic from the plane post diversion, including IP traffic (if any) and, if so, have they fully analyzed both the metadata and the content (if any)?
@littlefoot
Yes, I believe the plane kept flying, with the ultimate destination being Perth.
I agree about discussions, but pointing to the ineptness of the authorities doesn’t equate to pointing a finger at an individual on the plane (Admittedly with no firm evidence) whose relatives still pray for their safe return.
VictorI,
The Thais made the same comedy of errors as the Malaysians did, in terms of the purported radar pick up.
I have in comments made on the TMF blog, most of which addressed to Bruce Lamon, set out the what the Thais thought they saw on their Surat Thani radar.
1. When the plane disappeared, Thai ATC (and all other ATCs in the region) said they did not pick up anything that could have been MH370.
2. Then on March 15th as reported in Thailand’s main newspaper Bangkok Post, Thailand’s military said their main radar system at Haadyai only picked up MH370 departing from KL and that was the ‘first and last time’ they saw MH370.
3. It was only a few days later on March 18th, 10 days after the plane’s disappearance that another military officer came forward and said their Surat Thani radar picked up a blip first heading to KL and then ending up northwest, the route first mentioned by the Malaysians.
4. The same points made in my previous comments regarding the purported pickup by the Butterworth radar would apply equally to this alleged pickup by the Surat Thani radar.
5. The Surat Thaini radar was the furthest away from IGARI/BITOD with both the radars at Haadyai and Phuket closer.
6. Its range, like the Butterworth radar, also ends before IGARI.
7. So the same questions would arise- how come the radars closer to the plane did not pick up any unidentified blip and how come Thai ATC which would have the returns from all radars both civilian and military also did not see anything.
8. Again it looks like a case of the Surat Thani radar having picked up CES539 coming from the other direction heading to KL.
9. Unlike the Butterworth radar, the Surat Thani radar could not have tracked CES539 all the way to KL and from KL KLM836, so presumably they only saw a blip going over Pulau Perak at the time indicated initially by the Malaysians of 2.40am, when that blip came into range close to that island.
10. At the end of the day, those that claimed they saw something that could have been MH370 bear the burden of proving that the blip was indeed MH370.
11. To begin with, these people have to produce the radar track of those purported blips.
12. The fact that until today, no recording of any radar track or any other data has been produced, by the Malaysians or for that matter by any country, speaks volumes.
@Alex
“At the end of the day, those that claimed they saw something that could have been MH370 bear the burden of proving that the blip was indeed MH370.”
I COULDN’T AGREE MORE!
@Luigi, since you’re around: thanks for the info about the Curtin University ‘boom’, which was registered at the time mh 370 might’ve run out of fuel. So, there’s now a geological explanation. And we can probably safely disembark from this train of thoughts.
@Brock
>the topic Matty & I were discussing was the feasibility of the path the ATSB
>used to defend the site they chose to listen for pings. They published this path on May 1;
>it had LISTED speeds of 323KTAS in its “southern arc” phase. If the Inmarsat pings are to
>be relied upon, then those speeds are FORCED by that location.
By the 1st May report I assume you mean the report to the ICAO, dated 9th April, that was released on 1st May. This was from the Malaysian authorities; the role of ATSB in producing that report is unclear (to me at least). The factors used in determining the search areas are described in appendix A of the later ATSB report. In the era of late March/early April this includes the following:
‘The area that was crossed by air route M641 was classified as red (Priority 1), the next two priorities, yellow and green, were then defined moving south along the arc from this position’
As you will recall, possible paths of the aircraft determined by waypoints were much discussed early in the incident, and clearly this was still a driver for the investigation when the search pattern for Ocean Shield was decided. It was in this red area where the acoustic ‘pings’ were detected. The report implies that the corrections to the BFO data due to the satellite temperature and other issues was not folded in at that time – those turned out to be very large (up to 9Hz). Without those corrections good BTO/BFO fits on curved paths to the red-zone (at lower speeds) were possible – I described some in a note written in early June. It seems perfectly reasonable that those later adjustments were not available, or at least not fully accepted, when the search pattern for Ocean Shield was set. The ATSB report appendix A makes it clear that by June the BTO/BFO was the main determinant of the search area, but it wasn’t earlier, and it wasn’t the same data earlier.
We can’t assume that all of what is known now was clear and accepted at the end of March.
To try and not appear a complete patsy for the authorities, I have serious concerns over the analysis of the data presented in the ATSB report, particularly:
1. there is no description of the statistical techniques used to determine the width of the search area. This is absolutely key. All the statistical terms used are qualitative rather than quantitative.
2. the random error on the BFO is defined under one model at 5Hz (standard deviation = 1 sigma). However, examination of the raw BFO data (there are several runs of BFO values early in the flight) suggest the random error is much smaller, around 1.6Hz. I added a contribution due to unknown small vertical speeds, to make a total random error of 2.1Hz. This figure gives the expected minimum errors in the fitting using a particular statistical technique. A figure of 5Hz (apart from being a round number) just doesn’t match to the fitting achieved.
3. the presentation of the validation data is naïve. There must be a number of candidate paths that match the BTO/BFO data for the validation flights to within the errors. The diagrams only show one candidate path per flight. This might be the best fit to the data, but that is irrelevant, it is the width of the sheath of statistical significant path matches that is important and that is not presented.
4. Although improved, appendix G of the report is still disappointing. For example, whoever wrote it used an STK table of satellite positions/velocities with a cadence of 5 minutes, then kludged it for aircraft data two minutes different in time when parameters calculated are changing faster than that. We are supposed to be impressed by the error of 0.3Hz in the match to the modelled course (table 5), but the error added by the kludge is twice that, and what happened to the noise? A noise of 5Hz (or whatever is correct) should be apparent in any attempt to do a precise point-by-point comparison – so it would seem the bias term was calibrated from that measurement or the calculation was just lucky. Either way, not comforting.
It may be that appendix G was added later to guide the interested and the actual modelling by the separate teams that went into the definition of the final search areas was more rigorous, but I can only go by the report as published.
@Richard, thanks for your detailed contributions!
@luigi
>We have recently been told that data traffic from a call made to the flight deck satellite
>phone (which was reportedly not picked up) has provided additional position/heading
>information, leading investigators to tweak the inferred flight path.
This is based on a short section of the August 28th press conference so there is not much to go on. The BFO data corresponding to the data packets sent from the aircraft as part of that call attempt were always in the raw data file from Inmarsat, but it would seem were not used in the official modelling. This is presumably because the analysts did not believe their track models (speed/heading etc) were applicable for the aircraft’s motion at the time of the call so the BFO data would not help in the fitting. If adjustments to the data calibration mean the track models might now match the call time, then it could be used in the fit. It doesn’t fit any of the statistically significant tracks in my modelling, but that doesn’t prove anything.
On your second point on IP traffic, I don’t know the answer. The report says that the functions that worked over the satellite link were audio communication, Aircraft Communication Addressing and Reporting System (ACARS) and the In-Flight Entertainment equipment. It was stated in the preliminary report that ACARS transmissions stopped early in the flight. There were very few data packets sent during the flight, at least as recorded in the log, so I would doubt that there was any IP traffic, at least after ACARS stopped. Others might know more.
Thank you. Not sure why the investigators would have incorrectly (as it turned out) discarded the phone call-related data points from their modeling, if that’s the reason. What is your take on the satcom reboot? At what point did that become known to investigators? You say that there were very few data packets sent during the flight according to the log. Does that mean that there was hardly any use of the internet on the plane even before the diversion? Was internet service available to passengers? Does the log enumerate data traffic both pre-diversion and after the reboot? Does the log come from Immarsat’s own database, or was it filtered by the authorities in any way? Presumably, if there was internet service available, someone must have gone over all the Internet traffic from the plane for the whole flight with a fine-tooth comb looking for clues. Would that have been done by Malaysian authorities, or would Immarsat have done that?
Okay, enough questions for now…
@Richard,
You wrote:
“I will presume you refer to the data when the aircraft was stationary at KL airport. The trend of the BTO data in that period corresponds to what would be expected according to the BTO algorithm in appendix G of the report, that is the BTO should increase by 100us per 30mins due to motion of the satellite. See the graph below (R-channel BTO values only).
https://www.dropbox.com/s/hm6lbgtopdhjged/BTO_at_KL.jpg?dl=0”
I agree with your graph, but not your assumption nor your conclusions.
First, while the line you’ve fitted may be the least squares line, the correlation is so poor that you cannot properly call it a “trend.” My calculation results in an R_squared of something like .0000001. You cannot conclude that there is a trend from this. The mere fact that the stationary BTOs are within 100us of each other is irrelevant. The point of correlation is to show that the “BTO” variable us dependent on the “distance” variable, and not some other unknown variable. With weak correlation, we can’t be sure that the relatively narrow range of BTOs isn’t caused by temperature, for example, which would also be relatively constant at the terminal.
But then, even if you see a trend, you’re seeing a trend that contradicts the hypothesis. The satellite, thoughout the time frame of your graph, is moving TOWARDS the airport, not away. So the trend should be DECREASING BTOs, not increasing ones. The satellite started heading south at about 19:36. It was never as far north as KL, so from a KL reference point, the satellite is approaching until 19:36, then receeding (I believe for another 12 hours, but don’t quote me there.)
While I realize the “100us every 30mins” came from ATSB, it doesn’t make any sense that a linear rule would apply to a satellite in an elliptical orbit.
@Richard: Yes, I refer to charts in the Preliminary Report released May 1. WHOEVER cares to take responsibility for those charts chose a “best estimate” flight path whose “red zone” terminus MH370 LACKED THE FUEL TO REACH.
For a waypoint to drag the ATSB past its own performance limit defies logic.
Richard, I’m afraid this looks very much to me like an entity keen to search much closer to port for a while, with subsequent rationalizations of gradually increasing plausibility and/or complexity, to convince and/or confuse. It simply won’t wash. Hence the petition.
Since you conceded concerns re: disclosure, I will concede something in your direction: I believe the PEOPLE of the ATSB – right on up to their leader, Martin Dolan – are people of high ability and character, who’ve simply and sadly been ensnared in someone else’s web. I suspect key facts were “managed” at levels higher than Martin.
@Richard
Let’s see if we can tone it down a little.
Answer me this……
Why does ATSB refuse to even release the recordings of the putative pings?
Give some id’s for the nameless and faceless “group of international experts.”
Why were we told that the Bangladesh Navy was going to investigate the claims of GeoResonance. (FYI I personally place no credence in those claims)
Who EXACTLY *controls* and make these statements and decisions?
That’s enough for now.
Thanks
@Richard – I’m several posts behind you here. I still take issue with some of your BTO conclusions, detailed above, but I also wholeheartedly agree with this statement of yours:
“4. Although improved, appendix G of the report is still disappointing. For example, whoever wrote it used an STK table of satellite positions/velocities with a cadence of 5 minutes, then kludged it for aircraft data two minutes different in time when parameters calculated are changing faster than that.”
My suspicion, actually, is that the satellite positions at the nearest 5-minute interval produced better results than the correct positions would have. The opposite is possible, too, for sure.
But we’re back to the same place – the ATSB either used the wrong data deliberately to add support for their conclusion, or they used the wrong data out of incompetence or laziness. Neither is reassuring.
@Brock, who exactly has ensnared whom in your opinion? And what could be the motives for this obfuscation? I’m not asking in a sarcastic way, I’m simply interested in your opinion. I see plenty of motives for the Malaysian authorities to be less than forthcoming and cooperative, some of these motives being actually pretty mundane. I also see plenty of public relations mismanagement, especially from Inmarsat. I see errors, which is only to be expected in such a baffling case. Many of us shake their heads, when we look at some early decisions. But machinations from factions above the ATSB would take this conundrum to an all together higher conspiracy level.
So, would you care to elaborate?
@richard
>> This is based on a short section of the August 28th press conference
>> so there is not much to go on. The BFO data corresponding to the data
>> packets sent from the aircraft as part of that call attempt were always
>> in the raw data file from Inmarsat, but it would seem were not used
>> in the official modelling.
One possible explanation is that the scientists within Immarsat and the ATSB have never been given unfettered access to the totality of the data traffic that went through the satellite, and their analysis has therefore has been limited by political constraints (under the usual rubric of “ongoing criminal investigation,” “national security,” “need to know,” and “state secret”). As far as I can recall, the attempted calls to the flight deck satphone were not publicly acknowledged until months after the aircraft disappeared. Perhaps it was only then that the analysts became aware of the missed traffic and had the leverage to get a peek at the data.
@JS
>But then, even if you see a trend, you’re seeing a trend that contradicts the
>hypothesis. The satellite, thoughout the time frame of your graph, is moving
>TOWARDS the airport, not away.
The BTO is a function of the total distance from the ground station (via the satellite) to the aircraft (and back). The distance from the ground station to the satellite is increasing in this period as the satellite moves North so the BTO increases. The distance from the satellite to the aircraft is also increasing, if modestly, due to the increase in the altitude of the satellite. Anyway, in total the round-trip distance increases, so the BTO increases. We are not trying to prove the model here – the model is a given. If you want _this_ data to prove the model, there aren’t enough data points; if there is proof of the model it is in the validation data on the other flights, which we don’t have.
>While I realize the “100us every 30mins” came from ATSB, it doesn’t make any sense
>that a linear rule would apply to a satellite in an elliptical orbit.
I have worked from the raw data and the satellite elements, not from the ATSB report. The trend is pretty close to linear over such a limited time.
@littlefoot: thanks for your interest.
From a very long list – each far-fetched – I select the US as the least-UNlikely culprit. It seems plausible they were responsible – by accident or design – for MH370’s fate, want the world not to know this, and work hard in the background to put bad info out, and keep good info in. The US is on a very short list of entities capable of both deed and coverup, and I doubt the Aussies would work this hard to protect any other culprit on that list.
If accidental (war games gone wrong) or reactive (to a real/perceived hijacking), motive would be simply to conceal embarassment. If intentional, motive could be anything from espionage (offense or defense) to terrorism (economic, directed at Malaysia, or false-flag, directed at whoever blocks their path to discount oil, these days). I have listed these in what I feel are DECLINING relative probabilities.
But honestly: who cares what MY pet theory is? That we are FORCED to speculate is a symptom of the disease our petition means to cure. We all need to stop skeet-shooting each other’s theories, and start drilling investigators for bedrock facts.
Thanks, Brock.
Who cares what anyone of us here thinks? I like your statement, that the fact we are forced to speculate (about things, which shouldn’t need speculation I might add) is a symptom of the disease.
Littlefoot – I’ll repost this old link of yours in case some of the new guys haven’t seen it.
http://www.bloomberg.com/apps/news?pid=newsarchive&sid=aK4daf8MD.Bw
And another –
http://www.techrepublic.com/article/satcom-terminals-ripe-for-malware-exploitation/
I have held onto the Curtin boom for a while as I assumed it would be a pretty simple matter to corroborate/discount with other seismic data sources. Maybe they don’t share? Tonnes of seismic sensors out there. Any case a red herring bites the dust.
VictorI,
Inmarsat was just the SATCOM link provider, SITA was the service provider to Malaysia Airlines (MAS).
After MAS announced at 7.24am that the flight was missing, SITA would have asked Inmarsat (the SATCOM link provider) and whoever was the VHF link provider (ARINC etc) to forward MH370 data to SITA. Inmarsat said they forwarded their data to SITA that Saturday itself, within 4 hours. The data from Inmarsat would comprise all transmissions from MH370 transmitted through the SATCOM system. The aircraft had an special ID number and it would have been a matter of inserting this ID number into the system or database and then printing all transmissions from such ID from time of original log in at KL onwards, into readable data logs. My layman’s understanding anyway.
On the point whether the SDU could have been in a compromised or degraded state, transmitting ‘distorted’ BFO and BTO data as a result, during the period of the pings.
1. First of all I would like to pose this question to all the scientists/engineers/mathematicians out there: Is it not a ‘tale of 2 halves’ in respect of all the graphs of the data from MH370 whereby the second half comprising of the period from 18:25 UTC onwards shows a markedly different pattern as compared to the earlier part of the flight?
2. For this second half, the respective graphs for (a) the satellite velocity (b) the satellite movement on the Z axis (c) the BTO and (d) the BFO, all have the same shape, like a ‘tick’ mark, with a dip at the time of the 2nd ping at 1941 UTC and a ‘linear’ path from then on. The satellite happened to reach its northern apex, ie turned, at around this time, at 1936 UTC. There was no such pattern for the first half.
3. For the second half, BFO = satellite velocity in knots plus an offset of around 90 (the same offset when the plane was stationary in KL incidentally). There was no such pattern for the first half.
4. For the second half, for the pings not affected by the eclipse, 1 km of movement by the satellite on the Z axis = 8 plus us in BTO. There was no such pattern for the first half.
5. For the second half, the alleged flight path of the plane constructed from the BFO and BTO data, borrowing a word from @littlefoot, mimics the movement of the satellite for the corresponding period, firstly northwards at 360 degree north (see the other Richard’s comment on Duncan’s blog) and then southwards not later than the time of the second ping at 1941 UTC at a 180 degree heading/track (see VictorI, Denver and many others) . There was no such pattern for the first half.
6. The BFO values for the second half of the graph and in particular for the later pings, reversed engineered work out to speeds for the plane exceeding the maximum speed of a B777. There is no such issue for the first half.
7. The BTO for certain transmissions for the second half reflect a ‘fixed offset’ of as much as 4600us and others have ‘variable delay’ (according to Inmarsat). No such fixed offset or variable delay can be discerned for transmissions for the first half.
8. The ratio of BTO to BFO for the second half for pings not affected by the eclipse, fixed offset or variable delay, works out to around 71. There is no such pattern for the first half.