The discussion prompted by last week’s blog post raised some interesting issues that I think are worth discussing in further detail.
First, I wrote last week that “At 18:22, MH370 vanished from primary radar coverage over the Malacca Strait. Three minutes later—about the amount of time it takes the Satellite Data Unit (SDU) to reboot—the satcom system connected with Inmarsat satellite 3F-1 over the Indian Ocean and inititated a logon at 18:25:27.”
Commenter LouVilla earlier today laid out the issue with more clarity, writing:
MH370 flew out of radar range @18:22.12 UTC. All of a sudden @18:25.27 UTC, the AES sent an Login-Request to the satellite. This are 03:15 Minutes between this two events. When the AES is without power supply for a while and reboots after power is available again the AES needs approximately 02:40 Minutes to sent an Login Request (ATSB Report Page 33). 03:15 minus ~ 02:40 = ~ 35 seconds. So, the perpetrator must activated the left bus again at around 18:22.47 UTC, 35 seconds after MH370 flew out of radar range.
The close sequence of these events does, in my mind, raise the possiblity that they are connected. How would a perpetrator know that he has left radar coverage? Among the possibilites would be a) some kind of radar-energy detector (like that used by automobile speed-trap radar detectors) brought on board by the perpetrators, or b) prior scouting by allied agents. This latter idea would be far fetched for a suicidal pilot but quite feasible for, say, Russia, which spends quite a lot of time probing the radar coverage of its NATO neighbors.
Of course the timing might just be a coincidence.
A second point I’d like to address is the idea that Zaharie or Fariq might have de-powered the satcom by isolating the left AC bus. One problem with this scenario, as I’ve previously mentioned, is that it would be difficult for a pilot to know just what else they would be taking off line in isolating the left AC bus. I later realized that I had underestimated the problem.
In a fascinating blog post on Flight.org an airline pilot who goes by the handle “Ken” describes going through a simulated left AC bus failure in the course of a training session. He notes that among the systems lost were Window Heat (Left) and a Primary Hydraulic Pump (Left). “No biggie,” he writes, but adds that in addition:
…there are a whole host of ancillary services lost. Many of these are reflected by the amber lights on the overhead panel. Having looked at the roof – you later discover even then that it’s not the whole story. In this particular scenario we decided to return to KLAX. Part of the return process was fuel jettison down to maximum landing weight. Guess what? Without the Left Bus – the main tank jettison pumps are failed. You’ll be advised of this… when you start the fuel jettison. I didn’t give this a second thought… but the discussion we had afterwards that included a talk about this little quirk of the Boeing EICAS/ECL was interesting. There are no EICAS/STATUS messages to advise you of everything you’ve lost, and in many cases, until you attempt to use something that’s failed – you won’t know about it. Older aircraft used to publish a Bus Distribution List (Electrical and Hydraulic) so that you’d know exactly what you’d lost with a particular electrical bus failure – but not on the 777. My fellow pilots were vaguely disturbed by the lack of information.
It’s not impossible to imagine that one of the pilots cooked up a plan that involved switching off the satcom by isolating the left AC bus, but to do so they would have had to do intensive research into the issue. And even then, they would have to have grappled with the fact that in doing so they might disable other systems that they weren’t aware of. All told, this would be a complicated and risky strategy. And to what end? If the satcom was deselected for ACARS and the IFE was switched off (both of which are easily accomplished from the cockpit) then there would be no reason for a pilot to fear that the satellite would give away his position.
Another suggestion that has been made is the idea that the co-pilot, having been locked out of the cockpit, went down into the E/E bay and started pulling circuit breakers at random, hoping that in so doing he would succeed in de-powering the flight deck door lock, and instead power cycled the satcom by mistake. I don’t think this makes much sense, since a) this would require to know that such a circuit breaker exists in the E/E bay, but not know where it is, and b) I just can’t imagine a trained airplane pilot pulling circuit breakers at random.
In general I think we should resist any explanations that require complicated series of actions to take place as a result of a random series of happenstances. Boeing 777s are not Rube Goldberg contraptions; they are multiply redundant and extremely robust. Neither a fast-moving fire nor a panicked copilot are likely to remove the exact components at exactly the right time (and then replace them at exactly the right time!) by chance alone.
Finally, I think it’s time to raise a very important issue regarding the search of the southern Indian Ocean. Last week, I wrote that the search had failed. Some people took umbrage at this suggestion, pointing out that the original 60,000 sq km area has not yet been searched. To that, I say fair enough. Perhaps I jumped the gun. I’m willing to go along with those who say that we need to wait until the entire 120,000 sq km are searched. But then what?
For many, the matter will have been laid to rest: if the plane is not there, then it did not go there. It will be time to scratch the “ghost plane” hypothesis off the list and move on to see what other options are on the table. Well and good–this is how scientific investigation moves forward.
However, I am concerned that some people might refuse to come along. Already some commenters have pointed out that there may be crevasses into which the debris could have sunk, or underwater hills in whose sonar shadow the wreckage may be lurking. Or maybe there was a gap between the search swathes. These are all valid points, but they are also points that the Fugro searchers are certainly well aware of. They know exactly what part of the seabed each sonar image covers. They can tell where the gaps are, and they can send UAVs to probe the shadows and the gullies. Their entire mission hangs on them covering every square inch of the designated area, precisely so that that when they’re finished no one can say, “well, you only covered 99.99 percent, therefore we don’t know it isn’t down there.”
We all have to be open-minded about the data, no matter how fervently we may believe that our personal hypothesis is correct. It’s unsportsmanlike to call on the ATSB to search a particular ocean, at great expense and effort, and then when they’ve spent the money and time say, “Well, I don’t believe in your result, you probably screwed it up.”
We can be skeptical about the authorities’ handling of the investigation–I’m sure none of us would be here if we weren’t–but at the end of the day we have to have some basic faith in the honor and competence of the investigators. Otherwise, we just have to throw our hands in the air and declare that nobody knows anything.
Just a short question :
We have an full handshake @00:10.59 UTC (BTO/BFO). Why ? The last activity before this handshake was @23:15.02 UTC (MAS call).
~54 minutes, not the usual 60 minutes. How can we explain this 6 minute difference ?
@Gysbreght,
I think I have a fair idea. I own 2 aircraft and fly them regularly. Of course they are much lighter than a B777, and don’t fly nearly as fast. But I have experienced a descent rate of perhaps 12,000 ft/min (only in a spin in light aircraft) and I can assure you that recovery from that takes a deal more than 650 ft.
Consider the fact that at altitude in a heavy aircraft the elevator is much less effective, and it is most unlikely that there is sufficient elevator to achieve a pull up of 2G.
So tell me about your explanation.
@LouVilla,
Lou, not sure about your link. When I look at earth.nullschool, I get a 1024 where you have 1208 in the link. That is not surface. Also your link’s coords are not in the current search area.
I just looked again, I come up with 26kph at 37S, 89.5E (using my fat finger on the iPad). That is Beaufort 4, 1-2m waves, according to Mr WikiPedia, at 0000Z on 2015/03/08.
I’d like to point out again, that for the prevailing sea state at that time, the conditions in the hours before that time are more important and that nullschool is a predictive model, rather than an observation database. It can be taken as aguide only, not as factual information.
Cheers
Will
@Flitzer_Flyer:
Your quote is not quite accurate. From the official MAA report on the incident:
The DFDR traces in the report show that the recovery from the dive was accomplished in approximately 1000 ft height.
@MuOne
I’d like to point out that at the same time wind for south and east of Christmas Island shows about 5-15 km/h which is a lot more favorable for ditching.
@gysbreght and @Flitzer_Flyer: The recovery time (and drop) is a strong function of the initial conditions. Consider a plane in a vertical dive. If too much elevator is applied quickly, the angle of attack (AoA)will be high, and the wing will stall, and might even reach a flat spin. Let’s say the plane is able to pitch up by 15 deg from straight down and the wing is not stalled at this AoA. Most of the lift force from the wing contributes to increasing the horizontal speed, not decreasing the vertical speed, as the lift force is perpendicular to the airflow over the wings. As the horizontal speed increases, the plane can pitch up further (again, limited by the AoA for stall), and the plane progressively recovers from the steep descent. (The vertical component of the drag vector, parallel to the airflow vector over the wings, will also slow the vertical descent.) The entire lift from the wing cannot be applied in the vertical direction immediately without inducing a stall.
@MuOne :
I checked this. You´re absolutely right.
This link shows me on 03/08/14 @01:00 UTC, Data : Wind@Surface, Mode : Air, Height : Sfc
26 kph @37S,89.5E (using my fat finger,too).
This is Beaufort 4, correct. Thank you for your clarification on this issue.
Since it may be of wider interest, here is a screengrab of a picture in the report:
http://i.imgur.com/tBDqilR.png
@Gysbreght: Yes, it took 1000′ to recover because it had substantial horizontal speed during the descent and it’s velocity vector relative to the air was far from vertical. Most of the lift could be applied in the vertical direction without stall. The recovery time is limited by the AoA of stall and the structural limits of the wings.
@VictorI:
The initial condition was given as 15,000 ft/min rate of descent. In a vertical dive at FL350/ISA that is 148 kTAS or 83 kIAS, much too slow for a B777. At that speed there wouldn’t be much drag or lift or control authority, and the airplane would therefore accelerate downwards at approximately the acceleration of gravity which is 19,05 kts/second. To achieve a normal acceleration of 2.5 g it would have to accelerate to at least 1.6 times the stall speed, which is of the order of 170 kIAS in clean configuration.
@Gysbreght: My only point is that in the discussions between you, @airlandseaman, and @Flying_Fitzer, there was disagreement on performance based on differences in initial conditions, i.e., the horizontal and vertical components of velocity.
You have to be careful about stated stall speeds. I suspect that 170 KIAS refers to speed at which lift equals weight, i.e., the velocity vector is steady (but not necessarily horizontal), not the minimum speed at which lift is generated. As long as there is flow over the wing and the AoA is less than the stall angle, I would expect that some lift should be generated down to zero speed.
@Gysbreght: OK. Re-reading your comments, I see that you are equating stall speed with the speed for 1-g lift. All’s good.
P.S.
In order to pull the wings off it would have to exceed a loadfactor of 1.5 * 2.5 = 3.75, and to achieve that it needs two times the stall speed.
P.S.2
Just to keep it simple; to kill 15,000 ft/min rate of descent in 650 ft of height requires a vertical deceleration of 1.5 g. At any reasonably probable airspeed that corresponds to a ‘sensed’ normal acceleration of approximately 2.5 g.
@Gysbreght: But again, if the plane is in a vertical dive, the 2.5g of vertical force can only be applied progressively and the plane will fall much more than 650 ft. If the plane is closer to horizontal flight, the lift force is better aligned with the vertical, and the descent is more easily arrested. The ability to arrest a descent is a function of the initial velocity vector.
@VictorI:
Please explain how a B777 gets into a vertical dive at an indicated airspeed of 83 knots.
@VictorI, Gysbreght, For present purposes I think we all can agree that, without having any further information, it’s plausible that a plane flying at cruise altitude would have enough room to recover from a 15,000 fpm descent.
For this to occur, we would have to posit that the plane ran out of fuel and began plunging toward the ocean before a pilot thought better of it and pulled the nose up in order to effect a lengthy glide far from the 7th arc. Whether or not this eventuality would seem likely, the scenario would allow for a union of Mike Exner’s interpretation (that the 0:19 BFO implies a steep descent) with that recently proposed on DuncanSteel.com by Hyman, Martin, and Bennett (that the absence of seabottom wreckage means a pilotless final spiral did not occur).
@Victor and Gysbrebht,
I think both are contributing parts of the solution, Victor more so.
There are couple of important points . . Any lift generated is perpendicular to the wing, and not vertical, and indeed contributes to increasing the horizontal speed (and not necessarily arresting the vertical descent) which is why it is absolutely necessary to manage the speed while pulling out. It is also critical not to exceed the critical AOA and induce a stall (which is not speed dependent).
The fact is a big aircraft at say 35,000ft does not have sufficient elevator authority to generate even 2G, but it will continue to accelerate in a dive unless the speed is managed with speed brakes, and only gradually does the elevator become sufficiently effective to pull up as the AOA is managed.
P.S.
The argument started with ALSM’s assertion that a sinkrate of 15,000 ft/min cannot be arrested without pulling the wings off. I countered that with a simple back-of-the-envelope calculation that showed it is entirely possible. I have not done the detailed sums that would show that even a vertical dive at 83 kts IAS could be recovered without pulling the wings off.
@Gysbreght,
I don’t think anyone was talking about a “vertical” dive, at any speed.
The discussion is about a B777 descending at a rate of 15,000 ft/min, but with a horizontal speed of something approaching 500 knots. The vertical component of velocity can be calculated, but the aircraft attitude is certainly not “vertical”.
It seems to me you are contradicting your own quote.
You think I completely misunderstood VictorI’s comments?
@Jeffwise,
please feel free to stop or erase this conversation as much as you like. As long as it is allowed to continue, I will not have my inputs qualified as nonsense by anybody.
I think you should exchange e-mails and continue there
When a plane runs out of fuel it doesn’t necessarily start ‘plunging toward the ocean’. Normally it just continues flying, but it needs to descend at a shallow angle of about 3 degrees to maintain airspeed. However, if the pilot (or autopilot) instead of descending, maintains altitude, the airspeed will reduce until the airplane eventually stalls. What may have occurred is that the pilot initially tried to maintain altitude, allowing the speed to decay until he was warned of the approaching stall by the stickshaker. His logical response would then be to move the control column forward, the nose would drop, the flight path would become increasingly steep downwards, and airspeed would increase until the pilot pulled the nose up and continued to descend at the steady glide angle of about 3 degrees.
@Gysbreght and @Flying_Fitzer: To be honest, I believe we are all in agreement, and always have been. If the problem is properly formulated as to the initial condition of horizontal and vertical speed, we all arrive at the same conclusions about recovery. That was my main point.
@jeff: re: the “however likely” part of [pilot change of heart] as means, however likely, of reconciling [long glide theory] with [BFO-indicated descent rate @ 00:19]:
– pilot sits around until fuel is spent
– pilot actively pushes into 20-degree dive
– …to explain 15k’/min descent rate at arc7
– pilot reconsiders immediately after arc7, pulls out into long glide
– …to explain zero seabed wreckage under arc7
– in (by my calc) 3-3.5M waves, plane crashes violently anyways, leaving floating debris in SIO
– this debris evades ALL detection and down-current shorelines for 16+ months
Even with this “magic bullet” flight path (arguably implausible to begin with), we still haven’t explained nil surface debris.
Anyone trying to clear a path for this scenario by blowing down the avionics house of sticks needs to save enough breath for the surface debris house of bricks.
@All who still consider the ISAT data reliable: as usual when discussing BFO’s, I’m forgetting key elements of past discussions. Are we to respect the JoN directive to disregard 18:27 & 00:19 BFOs? If not, did we EVER respect it? If so, when & why did we STOP? (I’m guessing expert views are a direct function of how cold they think the mechanism was at each time, but would appreciate clarification.)
@spencer
”The ghost flight scenario makes no sense unless IMO you believe Z took himself out at a moment coinciding/just after the FMT…and this seems HIGHLY unlikely.”
Exactly. Apparently the person flying the aircraft survived till 18:25 but suddenly dropped dead right after the turn… how can this be explained ? If there was a aircraft failure (any type) then how could the person have survived for more than an hour up to 18:25 ? I think it is safe to assume that there was a conscious person in the cockpit beyond 18:25 and most likely saw a incoming sattelite phone call.
@Brock
I do think the ISAT data is reliable subject to the physical constraints on oscillator accuracy which has been over-estimated IMO. The BTO accuracy I think is fairly estimated.
I am amused by the new IG glide path scenario. First we have the zombie hypothesis which is what lead to the IG/ATSB consensus search area. Not we have the zombie awakening hypothesis to explain why the plane was not found at the zombie terminus. As usual the analytics are well done based on the assumptions. Despite my lame attempts at humor here, I think it was good to take a look at what a glide path might imply as opposed to the spiral descent.
I am inclined to wait a bit longer to see what the search reveals. There is still a lot of area left. Back to my initial point, the terminus is very sensitive to BFO regardless of what ancillary assumptions are made. I am far from ready, despite my long standing public reservations relative to the assumptions, to abandon the search in that general area.
Also, there is a finite probability (I would estimate 10% or so based on the AF447 reports) that the wreckage was missed by the sonar search.
Brock,
I’m quite sure that both BFOs are valid, and most likely they represent uncorrected values (see my earlier discussions with Victor and Niels). No steep dive. Moreover, I already saw it is possible to account for 00:19 BFO as uncorrected value in the optimization functional and obtain reasonable results. The difficulty is to account for the sequential flame out of the engines, particularly the time.
@Brock
“– in (by my calc) 3-3.5M waves, plane crashes violently anyways, leaving floating debris in SIO”
I don’t think you can pinpoint the wave height with 0.5m error margin because we don’t even know the exact location in southern SIO but it’s safe to estimate it at 2-5m(for the whole area considered by ATSB) judging by wind speed on that link posted today.
However even at the minimum wave height it would still leave debris, remember that Airbus that landed perfectly in Hudson river…it left some debris too and it was a river, landed down the stream.
After 16 months the chance of no debris surfacing anywhere is very very small.
That landing was far from perfect and do you have a reference for floating debris?
@Brock and StevanG
I think the effects of hurricane Gillian (around 20 March +/-) which passed over CI and into the SIO near the current search area has been under-appreciated relative to scattering debris. Also, I have not seen much of an attempt to model currents anywhere in that region. I’ve seen promises, but no results. Maybe I missed them.
@Gysbreght
I can’t find it now but I remember well there was a guy on reddit who claimed there was small amount of debris scattered around.
OK not exactly perfect landing but certainly the one would you expect from experienced pilot.
@DennisW
http://earth.nullschool.net/#2014/03/08/0000Z/wind/surface/level/orthographic=-263.68,-46.42,1208/grid=on
have you seen this link? the area around CI is very blue especially east of it, this was around the time of the final ping
problem is, WHY after 7 hours to carefully land in SIO empty area and not contact anybody during flight even with aparently working satvoice
in case of hurricane, there will be probably far better chance that we will already have many souvenirs with MAS logo everywhere, NOTHING found yet
Ah I see now you are referring to debris flow after the crash, well yeah it might be underestimated but out of so many pieces (assuming it crased like AF447) there is a small chance nothing would get found.
@oleksandr
please, can you remember where you discussed the “uncorrected BFO values” after restarts?
@airlandseaman
please how good is OCXO oven thermal insulation? tea in thermal bottle stays hot for a hour without problem…
Why BTOs between 00:19:29 and :37 (so 9 secods) are so different? (23000 and 49660)
For me, it looks more as if there was some UFO skipping ultra fast – I am simply confused
@Falken
Yes, your argument has merit. I am merely making excuses. The debris issue is truly the elephant in the room.
If the flight ended in a glide it is unlikely to have crashed like AF447 which was very deeply stalled at impact and in the entire descent.
@Gysbreght If I got it right the glide would require a sane person behind the stick? Why would anyone sane want to go there at the first place then?
@Dennis if CI theory is indeed true the sea was so calm there it’s possible the amount of debris would be negligible in case of controlled ditch
@StevanG: If you had followed my ‘yarn’ you wouldn’t ask.
Falken,
I think it was back in May. I realized that both BFOs are likely correct when I wrote the 18:25 and 00:19 sequences of logs into two columns. See my post May 11, 7:58pm for example. In my opinion both the anamalous values of 273 and -2 Hz are results of misfed data from INS. Zeroes are the most natural choice.
In support of the CI theory I would seriously consider the possibility that use of the autopilot was denied to the pilot as I suggested earlier. In that case he would have to fly manually and that is much easier at a lower altitiude, say FL250, and at a lower speed.
@StevanG
If the PAX were alive, which is embedded in my hypothesis, some would have exited the aircraft. Does not make sense (to me) that a controlled ditch occurred without some evidence of it.
Maybe I should help Professor Chen with a smooth ditch followed by a tip-over scenario. Just kidding. I am as lost as everyone else. The debris issue is killing my enthusiasm relative to any ocean terminus.
@Gysbreght
Certainly lower altitude. I’ve been using 8000 meters, and lower speed to satisfy BTO. I had not considered the possibility of AP denial, but rather deliberate AP tweaking or elective manual control.
Exiting the aircraft some 40-50 miles off the coast gives one very low chance for survival.
But yes then some human bodies would have probably washed ashore by now.
@StevanG
Yes. Plus life jackets, carry-ons, various service items, …
I wonder if northern parts of the arc could fall under thermohaline circulation which would take debris to who knows where.
OT? interesting link about AF447 case – good to stay calm ahead of streams of alerts, if possible
http://es.slideshare.net/auxesis/escalating-complexity-devops-learnings-from-air-france-447
@StevanG
Yes, they would, and that was an early postulate I had relative to no debris. Further investigation was not able to characterize the magnitude of the THC relative to surface winds and other factors, so I abandoned the idea.
Maybe it went down some where in Malaysia into the jungle or a larger lake (such as Kenyir Lake)..