Search crews in the remote southern Indian Ocean have completed a task so vast and technically ambitious that it once seemed impossible: to scan a three-mile-deep, 120,000 sq km swathe of seabed using a side-scan sonar “towfish” in hopes of finding the wreckage of missing Malayia Airlines 777 MH370. After considerable delay due to mechanical problems and bad weather, the final square miles were scanned on October 11 by the research vessel Fugro Equator. The $180 million project turned up no trace of the missing plane, though searchers did find several long-sunken sailing ships.
The Fugro Equator will next use an AUV, or autonomous sub, to scan selected areas where the rugged seabed topography was too rough for adequate imaging by the towfish. “The total combined area of the spots that will be surveyed with the AUV is very limited, but still required to ensure that no area has been missed,” says Fugro spokesman Rob Luijneburg.
The Australian National Transport Board (ATSB), which is overseeing the search, expects this fill-in work to be completed by the end of February.
The fact that that the Pennsylvania-sized towfish scan had been completed was first noticed by Richard Cole, a space scientist at University College London who has been meticulously logging the search ships’ movements via online tracking services and then posting charts of their progress on Twitter. “At the completion of Equator’s last swing in mid-October the target of 120,000 square kilometers had been achieved, at least as far as my calculations show,” Cole wrote me last week. Both Fugro and the ATSB subsequently confirmed Cole’s observation.
The 120,000 sq km area has special significance in the effort to find MH370, because ministers from the four countries responsible for the search have made it clear that if nothing turns up within it, the search will be suspended. Unless new evidence emerges, the mystery will be left unsolved.
Plans to search the seabed were first mooted during the summer of 2014, after officials realized that metadata recorded by satellite-communications provider Inmarsat contained clues indicating roughly where the plane had gone. At first, investigators were confident that the wreckage would be found within a 60,000 sq km area stretching along the 7th ping arc from which the plane is known to have sent its final automatic transmission. When nothing was found, ministers from the four governments responsible for the search declared that the search zone would be doubled in size.
In December, 2015, officials declared that the search would be completed by June, 2016. In July of 2016, Malaysia’s transport minister indicated that it would be finished by October; weeks later, a meeting of the four ministers pushed the completion back to December. Last week, the Australian Safety Transport Board announded that “searching the entire 120,000 square kilometre search area will be completed by around January/February 2017.”
In an email to me, ATSB communications officer Dan O’Malley said his organization will issue a report on the seabed search once the full scan is completed. Under ICAO guidelines, Malaysia will only be obligated to release a comprehensive final report on the investigation once it has been formally terminated; so far, Malaysia has only talked of suspending the search, not ending it.
The bulk of the work has been carried out by ships pulling a sidescan sonar device on a long cable. This so-called “towfish” uses reflected sound waves to create an image of the sea floor. By sweeping up and down the search zone in much the same way that a lawnmower goes back and forth across a lawn, searchers have been able to build up a comprehensive image of the search area’s bottom.
But, just as a landscaper might have to use a weedwhacker to clean up areas around rocks or stumps, searchers will have to fill in gaps in the scan where underwater mountains, volanoes and escarpments have prevented the towfish from getting a close enough look.
“A total area for search by the AUV is difficult to give because it concerns a number of relatively small spots that all are relatively difficult to reach and in difficult terrain,” Luijnenburg says.
The fill-in work will be carried out by an Autonomous Underwater Vehicle deployed from the Fugro Equator. The Kongsberg Hugin 100 is capable of diving to depths of up to 15,000 feet and can maintain a speed of 4 knots for up to 24 hours before being retrieved by the mothership. Whereas the side-scan sonar of the towfish has a resolution of 70 cm, the AUV’s sonar has a resolution of 10 cm, and so can image the seabed in much greater detail, as well as taking photographs when necessary.
Meanwhile, as the AUV work progresses, a Chinese vessel will deploy an Remotely Operated Vehicle (ROV) to take photographs of targets previously identified as being of interest. The ATSB has stated that none of these targets are “category one” targets, namely those likely to have come from MH370, however. Says Cole, “In the absence of category one targets there must be a list of targets from the sonar search that look the most interesting, so the question is how far down that list they are going to go.”
While the fill-in work must be carried out in order for the work to be declared 100 percent done, little prospect remains that the missing plane will be found in the southern Indian Ocean.
NOTE: This story was updated 10/26/2016 to include comments from Fugro spokesman Rob Luijnenburg.
@sinux
..related to that idea, some observers feel that if the plane was ice cold due to depressuring, that might explain why the INMARSAT pings might be wrong.
@JeffW
…yes and we might add to that that Mike Chillit is one of the most vocal believers in off-shore Exmouth (orig locale) as the best site. But I think that is closer to 22-25S
@sinux, I don’t know what your source of information for these assertions is, but based on what I’ve seen in ATSB and DSTG reports, I don’t think that the manufacturers of the avionics aboard MH370 would agree with you. There seems to be very little grounds for doubting the integrity and accuracy of the ping-rings.
@TBill:
That will be yet another guy with nothing to do and too much spare time on his hands to plan grusome acts…. Well, I think it is a good idea, and perhaps crewmembers could be elected to be educated to manage that. Or others esp. chosen/fitting for that within the organisation. I see some obstacles though but obviously two pilots behind a locked door is not the way to go. Now, there are an awful lot of flights, and really few incidents, but one incident is one too many.
I am almost longing for the man-size postal pneumatic dispatch system I once saw an outline for in the press. Hamburg–Bangock in just over an hour. By way of parts of an old natural gas pipeline through Central Asia, perhaps. What could possibly go wrong with that?
@Sinux
The BTO values depend primarily on the speed of light and the path length. There may be a small clock related variation in the fixed delay associated with processing in the AES.
We have good reason to believe that the speed of light has been constant for at least the last century.
@Jeff
For example the SDU manual states for AC Voltage at SDU terminals:
“100 to 122 V ac, 360 to 800 Hz (normal operation)
92 V ac, 320 Hz minimum; 134 V ac, 800 Hz
maximum;”
So the SDU can work between 92V and 134V that’s quite a big margin.
I don’t know if it will be able to establish a link at those extreme values.
But it will between 100V and 122V.
As the nominal power from a IDG is 115V you’ll note it’s more sensitive to higher voltage spikes than to voltage drops.
But of course that’s the view from outside of the SDU “black box” if you will. We don’t quite know how the voltage is rectified inside.
From the manual we know it can throw the following errors :
08 O 63 REF OSC FAILURE/VOLTAGE TOO LOW
08 O 64 REF OSC FAILURE/VOLTAGE TOO HIGH
My hunch is that as the SDU was never meant to be used as a tracking device, the fact that processor over/under clocking (up to a certain amount at which damage can occur to electronics) was not compensated for as the link can still be established.
Who cares if the SDU answers a tad too early or too late (as long as it’s within it’s allocated time slot)?
So I wouldn’t be so sure about the “accuracy” of the BTO. ATSB and DSTG assume a perfectly functioning aircraft under autopilot control…
They don’t explain why the SDU rebooted.
If we assume the electrical was in a non standard configuration (due to X problem or Y person’s actions) then SDU could have been switched on/off due to automatic load shedding. How stable is the voltage in a situation like that, when ELMS tries to maintain a balance?
PS if you haven’t already, see : https://core.ac.uk/download/pdf/140642.pdf
@DennisW
“There may be a small clock related variation in the fixed delay associated with processing in the AES.”
Exactly my point. And some of the BTO values need to be compensated for that. We know because we can calibrate at the gate for example.
Would you please be able to quantify the impact of a +/- 10% clock speed variation on the arc radius?
Even if it’s just a back of the envelope calculation…
@Sinux
The BTO bias is on the order of 500,000 microseconds. If that is mostly related to processing delays in the AES, then a 10% variation in clock speed would have an enormous effect some +/- 50000 microseconds. The BTO values themselves are only on the order of 12,000 – 15,000 microseconds.
However, clock speed variation due to both large temperature and large voltage variations is not going to exceed a part in 100,000 or some 5 microseconds maximum which would amount to path variations of less than 2 kilometers. Noise in the BTO measurements is an order of magnitude larger than that.
In the 20 previous flights of 9M-MRO examined by the DSTG, the BTO information was very consistent with the actual aircraft position for all the flights. Basically I would not hang your hat on this one.
@Keffertje:
120,000 sq.km, that is three Netherlands.
I was afraid it would be even more (but it is, in the sense that is goes up and down (I think)).
@DennisW
Thanks for the back of envelope calculation!
Please consider this :
In Underwater Water Search Area 3Dec 2015 page 20.
They recommend using a correction factor of 7820us from calibration (when?). They state that the ref clock was 128Hz (ie 7812.5us). That would mean the actual clock was running at 127.8Hz at calibration. That’s already a 7.5us difference.
Applied to the 18:25 BTO, (where N=5), that’s a 37.5us difference.
Applied to the 00:19 BTO, (where N=4), that’s a 30us difference.
You stated above :
“clock speed variation due to both large temperature and large voltage variations is not going to exceed a part in 100,000 or some 5 microseconds maximum”
Please explain where you got that 5us figure?
From my ruff calculations, 37.5us corresponds to around 10km of arc displacement…
That’s for a 0.2Hz clock speed variation
But it gets worse, if you do the calculation for 18:25 BTO, you’ll notice that the actual clock was at 127.6Hz.
00:19 BTO the clock is around 127.95Hz
We know that because we have a delta to a base value.
But what about the base value itself? There seems to be at least 1 N in that value.
So for a 0.4Hz clock difference at 18:25 the BTO difference is 15us already.
That’s a 0.3% variation in processor clock speed.
Just for the fun of it, I had a look at my computer’s clock speed (I know it’s not the same type of processor). It varies from 0.9Ghz to 1.6 Ghz within seconds…
@DennisW
“In the 20 previous flights of 9M-MRO examined by the DSTG, the BTO information was very consistent with the actual aircraft position for all the flights.”
In those flights the plane always landed where it was supposed to.
There was no malfunction of any sort.
I’m positing a scenario where the power supply is not stable.
That could explain the clock variations, and why they haven’t found the plane where the maths said it should be 😉
@JeffWise
“…….. I think that many people on this forum are attracted to the idea that the plane might have crashed significantly further to the northeast, at or beyond Broken Ridge, but attempts to explain how this might have happened run up hard against the constraints of the BTO analysis.”
Jeff, there are at least two tracks well to the northeast that fit the BTO’s accurately and both pass through the Cocos Islands waypoint.
1. The target waypoint of UXORA gives a possible final location around 20S to 21S with just three waypoints within the five BTO’s from 19:41 through to 00:11 and a constant ground speed through to 22:41 while fitting all the BTO’s without changing the true track at any of the BTO’s.
2. The target waypoint of IPKON gives a possible final location around 8S to 9S with just three waypoints within the five BTO’s from 19:41 through to 00:11 being required to give a constant ground speed and fit all the BTO’s like a glove without changing the true track at any of the BTO’s.
@Sinux
“clock speed variation due to both large temperature and large voltage variations is not going to exceed a part in 100,000 or some 5 microseconds maximum”
“Please explain where you got that 5us figure?”
The 5us comes from the part in 100,000 when applied to the BFO bias of 500,000 microseconds.
Your assumption is that correction factors are due to clock frequency changes. My assumption is that the correction factors are due to operations performed in the AES that were different at those times i.e. the software was executing a different sequence of steps.
A crystal oscillator simply cannot change its frequency by values as large as you are suggesting. It is a very high Q device, and changes on the order of parts in 100,000 for even very cheap devices like the oscillator in your PC would be a very generous upper bound.
Cheap digital watches are better than half a second per day, and there are 86,400 seconds in a day, i.e. better than a part in 100,000.
@Freddie
Interesting theory about Path to IPKON do you have a link for a report on this path. Approx Constant speeds do seem to be most likely, from peoples interpretations of the data. But headings/tracks for most peoples interpretations of the data of recent (as nothing found in the current search area so far) seem to vary up and down all the Arcs from 19:41 onwards.
@DennisW – but didn’t you discuss oscillator drift when temperatures drop significantly ? If power is unstable for the crystal oven then this might throw off calculations….
@MH
Yes, I did with respect to BFO values where there is much more sensitivity to oscillator frequency. Remember that L-band, the frequency the AES uses to communicate to the satellite is about 1.6GHz. That would be 1,600,000,000Hz. The Inmarsat claim of 7Hz error is about 1 part in two billion. 1Hz in 2,000,000,000Hz. Quite different than the 1 part in 100,000 I applied to the BTO discussion above.
@MH
Oops! Dropped a decimal place above. 7Hz is about a part in 200,000,000. The rest of the comments still apply.
@Buyerninety. Engine volume lessened too by any buried and not sonically reflective. The engines might well be separated from items which glide underwater and separated from themselves, in that depth some way. I would have expected the undercarriage, attached to other reflective structure, to be a prime tell tale.
@Jeff Wise. “Basically, they’re very confident that the plane isn’t anywhere they’ve scanned.” I think you are alluding to Fugro. I have trouble tying that in with the Gordon expectation of a low-to-mid 70s chance of spotting it, though I know he was talking about probabilities. My question is as to the overall possibility that the wreckage could be missed in an area nominally covering a 95% probability, the remaining 5% being a widespread low probability area. Can you help with that?
@Johan. “…the Dutchmen involved were inexperienced” Oh they were clever indeed if that is how they were assessed, taking advantage of the King’s absence.
They probably had a hand in the FI released during the investigation.
The seabed search. Adding to Jeff. https://en.wikipedia.org/wiki/Side-scan_sonar.The Chinese ROV has a video capability and I suppose the AUV does also. Something took the pics of the lost towfish on the bottom and the old wreck found. Magnetometers are used in submarine search and geology but are limited as to the type of metal they can detect, such as steel, quite limited detection range and discrimination from ferrous rocks.
@DennisW – its not the calculations done by those who like that punishment but more or less if the temperature drops significantly to affect the avionics equipment that the error compounds by factors on how each error is applied in a data chain to the final BFO. So the aircraft may have flown in a different flightpath but the BFO shows to the SIO.
@David, I’m afraid you’re conflating two separate issues. What Luijnenburg is saying is that they’ve scanned every square inch of the seabed (or will have once the fill-in work is done) and if something was there, they would have seen it. It won’t have been hiding under a bush or something.
What Gordon was talking about was the sum of the probability densities within the search box. So, before the search started, they drew a box and said, “It will definitely be within this box.” Then after they’d finished the first 60K and started on the second, they said, “Well, there was always a chance that the plane flew the way we think it did, but it wound up outside the search box–let’s say we’re 90 percent.” And then by the time I talked to Gordon, when it was pretty clear the plane wasn’t in the search zone, he retroactively lowered the probability to 70 percent or whatever.
I really think you have to take these after-the-face estimates of probabilities with a grain of salt. They’re a way of saying, in effect, “Our analysis was correct, we just got unlucky.” As I wrote in an earlier post, I just don’t think there’s any way a plane descending at 18,000 feet per minute is going to wind up 40 nm away.
@David:
So you suspect foul play? May very well be. They are tricky them Batavians. It is not so reassuring that the truth has been hidden for 400 years. We should perhaps slow down on our pace?
Thanks for the additional info on the seabed search.
If you wonder why the Russians carry missiles with them while passing through Öresund, it is because we’ve taken over the Dane’s habit of pouring bricks over passing ships. And now we even have a bridge to through from. Poor Russians.
@MH
Not sure if your last post is a question or a statement, but the errors associated with reasonable assumptions relative to AES oscillator drift are too small to allow a Northern flight path.
The errors are large enough to expand the possible range Southern paths, however
@Jeff:
“…descending at 18,000 feet per minute is going to wind up 40 nm away.”
It will only be the engines down there, or right below, or? There must be a considerable offdrift for the remaining splinters through 5000 m water.
@Jeff. Thanks. ‘I just don’t think there’s any way a plane descending at 18,000 feet per minute is going to wind up 40 nm away’.
No indeed, not if not manned unless it pulled out (possible)in which case 40km would surely be a stretch too far anyway. Or those BFOs were fallacious: unlikely, given the Boeing opinion. If manned there could have been a dive at the top to get away from decompression on engine failure. Oxygen available but bends and altitude sickness also, after a while. An APU start would take a while and anyway above 22000 it would not provide pressurisation I think.
@DennisW – should have used punctuation for that of a question.
@Johan. The rocks. Nice that NATO would have help.
@sinux, re your post October 27, 2016 at 5:32 PM
Are you suggesting, that the calibration of the BTO’s “at the gate” is effectively worthless for later analysis of the pings in flight, because “at the gate” everything was “normal” (voltages and temperatures), but later, in flight, (due to probable temp or voltage swings, or a combination of both), the AES clock speed was most likely “slowed”, and thus the “recorded BTO” under those conditions was larger than what would have been recorded if everything was operating “normally” ?
If that is so, then the “actual” ping arcs radius later in the flight would be less than we have calculated, ie, a shorter radius from the SSP.
The question then becomes, how much shorter, ie, what additional “correction” do we have to apply to the recorded BTO’s ?
The only way I can think of to try to quantify that, would be to work out what the limits of the clock speed variations were before the SDU failed to work at all.
Another point, is that it seems to have been assumed by all and sundry, that the the SDU reboot was a “restoration of power” (deliberate or by hapenstance is not relevant here) and that it was operating “in a stable state” therafter, BUT, that “new” stabele state may not have been the “normal” stable state (say due to temp).
But what if the situation is more complicated than that ?
If the aircraft has ongoing severe electrical system issues, what if the voltage swings were effectively “continuous”, but not too large ?
Is possible (probable) that the 3 “boxes” of the Satcom may be effected differently ?
Is it possible that whichever of those 3 boxes maintains “the logon” information, only rebooted “the once”, then “remained powered”, but one or more of the other two boxes, rebooted multiple times during the flight, because they were “more sensitive” to those voltage swings than box one ?
If that were the case, I presume that when boxes 2 or 3 rebooted, they would run their firmware codes in ROM, to get back “on line”.
Would box one, “know” that either box 2 or 3 had been “out to lunch”, (perhaps multiple times) between pings ?
Presumably not, because if it did, you would expect it to have initiated a new log on, and it didn’t.
So, is it possible that the end unit in the chain, the actual antenna unit itself, and the electronic system within it, could be the source of an apparent additional “clocking” delay ?
@David:
I thought NATO was helping us…
We’ve been messing with the Russians since the 8th century. According to our history Russia is nothing but the Vikings who didn’t want to come back.
@Johan. Not enough sun. They went down south I gather. Ottomans, Sicily. Not forgetting Normandy.
@David:
And Crimea, possibly. Lot’s of sun there. And water with some temperature. To sothen the arms after rowing down the Dnepr.
@David: still off-topic:
Á propos rowing, the Vasa’s demise, the longships’ success and the archipelagic province of Roslagen, look up Wikipedia’s article for “Rus’ (name)”. I dare not give the link.
@Johan. “.. the Vikings who didn’t want to come back.”
Puts me in mind of the Weslh as being the Irish who couldn’t swim.
for Ottomans I meant Byzantium
Will look up Rus thanks, and for the exchange.
@David:
Ha ha. It took a while for the penny to drop down.
It was probably a more multi-faceted thing, but with your permission I’ll change it to “the Vikings who couldn’t find their way back”.
And here’s another from a Norwegian example: When Russia annexed Crimea the national average intelligence quotient rose in both countries.
@Johan. ““the Vikings who couldn’t find their way back”. Ah yes. Well adapted.
I read about the Rus’ and the rowers. Your history lessons in school must occupy a fiar chunk of the curriculum.
I suggest reverting to topic or The Wise might become The Terrible.
PS The New Zealand Prime Minister reckons that their emigrants to Australia likewise ‘luft the untellugence’ of both countries. It might have been a bad line and to be fair he would have had to wait for the ‘phone.
@DennisW
“Your assumption is that correction factors are due to clock frequency changes.”
No, that’s not my assumption
“My assumption is that the correction factors are due to operations performed in the AES that were different at those times i.e. the software was executing a different sequence of steps.”
Yes, that’s also my assumption!
At 18:25 and 00:19, the AES was clearly doing something extra. What’s odd is why did it take 39000us at 18:25 vs 31000us at 00:19 ?
My hypothesis is that at 00:19 the microprocessor in SDU was receiving more voltage, and thus was calculating faster than at 18:25.
Do you have an other explanation?
I agree with your remarks regarding the OCXO drift / temp variations, but we’re looking at an other part of the electronics now.
We don’t know how much delay is added by the electronics processing to the BTO during normal operations. It’s canceled out by the BTO offset.
And this method works well when the SDU is operating normally.
But to me there is reasons for concern that the electrical supply wasn’t normal after Igari. Thus, we can’t be sure that the microprocessor was calculating at the same speed during the whole flight.
Honeywell would be able to test that. But so far I haven’t seen indications that they did.
@Ventus45
Yes that could be the implications. Bear in mind, I’m not talking about OCXO frequency variations. That’s an added effect. DennisW explained that part very well.
Regarding your boxes 2 (HSU) and 3 (HPA) being out of commission but SDU still functioning :
I doubt that’s possible. SDU voltage tolerance is tighter than HSU and HPA.
-SDU : 100 to 122 V ac (in normal mode)
-HSU : 96 to 122V
-HPA : 96 to 122V
SDU will shutdown first in case of under voltage.
I still think that if there was an electrical problem, the AES was load shed.
(Imagine for example, left IDG lost/disabled, passenger O2 generation starts (high priority), AES is shed (low priority). Later O2 generation stops, more power is available, AES is re-powered.)
Speaking of Russia I note Ukrainians hackers have accessed Kremlin email accounts;
Payback? Russia Gets Hacked, Revealing Putin Aide’s Secrets
http://www.nbcnews.com/storyline/ukraine-crisis/payback-russia-gets-hacked-revealing-putin-aide-s-secrets-n673956
Also @Jeff indicated it was a couple of Ukrainians, not Russians, who boarded MH370 on its final flight. It would be strange for Russia to get involved directly in MH370 if this was the case. Finally 169 of the passengers were Chinese and Russia and China are allies.
@David: I have feared that for a while already. It is the way of the sword.
@SteveB: The Ukrainians may perhaps want to think twice about who would come instead of Putin if he falls. And what would then happen in/to eastern Ukraine. As an American artist once (not once really) so wisely put it: “Be careful what you wish for”.
I wouldn’t use the word “ally” for China and Russia, but I might be wrong.
@David: history lesson:
Let’s say we are speeding through some of the millenia… And staying with nice bits in some centuries. And there is Vasa, which took a while (333 years) to decide upon. But the wreck-finder was smart: he probably told royalists it was a great symbol of royal power and the social-democratic political majority that it was a fine example of a royal f**k-up. So everyone was finally happy to pull the corpse up by the hair and exhibit it.
@SteveB:
Thanks for the article! There was also more to it than I first thought.
In trying to estimate the electronics delay in BTO, I stumbled upon a stone : “the BTO definition” !
Many people have asked what is the BTO exactly.
It’s not the round trip time of data from GES to AES back to GES.
That round trip time is in the order of more than 500’000us.
That’s NOT what Inmarsat recorded.
On page 23, the bayesian report states :
“the expected time of arrival of each communications burst is compared with the actual time of arrival and the difference between the two is referred to as the Burst Timing Offset (BTO).”
Which seems to be corroborated by Inmarsat’s journal of navigation report pg 2 :
“[the BTO] is essentially the delay between when the transmission was expected (given a nominal position of the aircraft) and when it actually arrives, and is a measure of twice the distance of the aircraft from the satellite.”
Please note the use of “essentially”!
Further the bayesian goes on to say :
“The nominal aircraft position is at zero altitude directly
below the satellite’s nominal orbital position of 64.5°E longitude, zero latitude and an altitude of 35788.122 km.”
And they give the value for BTO bias : 499962us
On the surface, voilà! All explained! But is it really?
Just before saying “essentially” Inmarsat said :
“The BTO is a measure of how long from the start of that time slot the transmission is received.”
So now is it linked to the frame time slot, or to a nominal position? Why would Inmarsat use a dummy satellite position when they know exactly where their satellites are? On top of that, they would have to program one nominal position per satellite, a lot of unnecessary work.
Back to bayesian report page 22 :
“The duration of each time slot is sufficient to account for all possible positions of the aircraft with respect to the satellite.”
Wait a minute… if an aircraft were to fly right underneath the sub-satellite point, at an altitude higher than 0, the BTO would be negative… Is that possible?
At the GES the burst would have seemed to arrive before the nominal time slot began…
That sounds fishy, I doubt Inmarsat’s engineers would have left a slight chance of recording negative BTOs.
The bayesian explanation is then “essentially” correct.
It works for the purpose of determining where the aircraft is within error margins.
How about an other explanation.
This is mine, please correct me if I’m wrong :
From : http://www.paul.sladen.org/download/aaib/sladen-20140703-briefing-note.pdf
“One SU arrived every ~160 milliseconds, with
them being logged to the Signalling Unit logfiles as they went.”
If I were to design a way to record the BTO in a database, and if I were told not to record the round trip time (to save space?), here is how I would do it :
(round_trip_time) % SU_time_slot
(The modulo of round trip time by signaling unit time slot)
This way, the data retained is minimized and we don’t have negative BTOs!
Does it work?
Indeed!
And the advantage is that we can now back calculate the other delays due to everything but speed of light. For example, for the 16:00 BTO, using SU_time_slot=160000us (this might be off, I don’t know, only using the value from Paul Sladen) :
round trip time : 160000 x 4 + 14820 = 654820us
speed of light delay : 510478us
extra delay : 144’342us
@ 16:29 extra delay : 144’340us
@ 16:55 extra delay : 144’300us
@ 17:07 extra delay : 144’279us
An other interesting point is that with this explanation, the error column in tab3 pg 6 of journal of navigation becomes irrelevant. You don’t need to use an average Bias value anymore.
The BTO is what it is :
A measure of how long from the start of the time slot the transmission is received.
@TBill, I read up today in MY elections and PKR (head is spinning). What is interesting is that BN (MY’s reigning party) had the biggest loss against opposition (incl. PKR) in 2008. It was a victory election for PKR despite a lot of corruption and bribery going on by BN. Guess what? Elections were on March 8th, 2008. The choice of date received huge criticism from PKR (Anwar Ibrahim). He accused the PM of it being a dirty trick because changing the electoral rules so short before elections would prevent certain people not being able to vote. March 8 was a landslide victory for Anwar Ibrahim, more so than the 2013 general elections. So for PKR, March 8 is a stand out date, a date of celebration and victory over the sitting government for the first time since 1956. That’s a big deal. Shortly after this victory Anwar gets arrested on the same sodomy charges. On March 8, 2011 a judge hands Anwar a major victory in his sodomy trial. What are the chances on the date? Then MH370 goes missing on March 8, 2014. Coincidence? I don’t believe in coincidences.
@Sinux
Interesting post. Frankly, I skimmed over the calculation of BTO bias in the “Bayesian Methods…” book. I did not think they would screw that up, so I paid no attention to it.
At first blush (and first cup of morning coffee) it would seem that there might be a bit of a disconnect here. I’ll take a hard look, and get back to you.
@Sinux
So I did the “Bayesian Methods…” calculation and came up with a value of 500040us. So we have for BTO bias:
Gordon: 499962us
Inmars: 495679us
Dennis: 500040us
The difference between my value and Gordon’s value is trivial. Amounting to less than 24km in range. All this says is that Gordon did the math correctly.
The difference between the Inmarsat value and Gordon’s value is non-trivial. Amounting to almost 1300km in range.
From page 40 of “Bayesian Methods…”
“The majority of the messages available from the accident flight are R1200 messages for which Tchannel = −4,283 µs.”
When this value is used to adjust the Gordon value of 499962 the Inmarsat value of 495679 results. This is the value everyone, myself included, has been using for BTO bias.
So, in my view, the world is a happy place. I am not sure what your exact issue is.
You said to Sinux, “there seems to be very little grounds for doubting the integrity and accuracy of the ping rings”
Hmmm. I admire your certitude. Can you be absolutely confident that some brain box smarter than either of us, is not having a good laugh at our expense?
But seriously, I agree with you 100% on that one. The chances that the electrical system malfunctioning during this flight are vanishingly small. The people who pose this possibility in order to account for why the plane hasn’t been found yet, are clutching at straws.
@Dennis
Nice to know the world is a happy place, at least for now :).
@ROB
Yes, I am asymptotically approaching my problem solving limit for the day and it is not even 9AM yet in Cali. I may be able to knock off before noon, drink heavily, and take a long nap.
@Keffertje
Interesting…I was initially confused with all the 8’s but I now see you are going back in history to 2008 and 2011 and 2013. At least that confirms my hypothesis that somehow Z requested the March 8 date, even if I turn out to be wrong about the moonrise. Sort of like the Ides of March, but a week early.
I am trying to find a comment by someone else, basically saying that the airline industry is strongly suspecting some kind of foul play probably factored into the loss of MH370. In any case, I still feel that is true.
@Keffertje:
The last Court of Appeal’s verdict against Anwar, tearing up the previous aquittal and sending him back to jail again came the same evening as the flight. You wonder if the Malaysian authorities work on any other day of the year. To all appearances, MY is saying either of two things: Anwar is on life-time salary from us to divert the masses attention from the ways of the political system, it is all theartre; or, we won’t accept you gaining any position in our political life — period; say what you like, but we are gonna make life very hard for you and your followers until you stop. It is not necessarily undemocratic formally, but it is cynical and it is evident for anyone to see. “Don’t go bathing in that lake”. One of the catches here is that what you refer to as the party, is in reality a coalition of 19 parties, in power since 1956. It comes close to what is referred to as majority oppression, and it is on the other hand a message to the people that what is good for those 19 parties is good for anyone, and that you, if you oppose their politics, is placing yourself on a political continuum that lies outside of (negating) that vast (repesentative) majority’s responsible opinions. Socially that can be very unwise, and politically it c a n be a bit strange to question authority for the wrong or not good enough reasons. They are saying that voting for Anwar is not the responsible thing to do. It is against society. Maybe this is only “party politics” Malaysian style, maybe it is a bit more.
@Keffertje
Nice catch on the dates.