Plane with 71 on board goes missing after taking off from Moscow

[vaztr]

WOW. amazing to see that the FDR data has been 'processed' so quickly

[Gabriel]

We need more sticky notes pasted to the airspeed indicator of pilots learning to fly in IMC.

http://avherald.com/h?article=4b4cb236&opt=0

In the afternoon of Feb 13th 2018 the MAK reported, that decoding of the FDR data has been completed. Preliminary analysis shows, that the pitot heatings for all three pitot probes were off while the pitot heatings had been turned on prior to departure on the previous 15 flights. About 2:30 minutes after becoming airborne a special situation developed at about 1300 meters of height and a speed of 465-470 kph (250 KIAS) when a disagreement between the speed readings #1 and #3 developed with the speed reading #2 not registering, #1 was showing about 30kph (15 knots) more than #3, an according message was issued. No significant altitude deviations between the pitot systems were noticed. At about 2000 meters height speed reading #1 began to reduce while #3 increased, another speed disagree message was issued. The crew disconnected the autopilot and continued in manual control. Speed readings from #3 reached 540-560 kph (290-300 KIAS), #1 speed readings continued to decrease. 50 seconds after the autopilot was disconnected the aircraft experienced vertical loads between 0.5 and 1.5G, the #1 speed reading reached 0, the #3 began to decrease reaching 200 kph (108 KIAS), the aircraft pitched down to about 30-35 degrees below horizont, the vertical load was 0G. Before collision with the ground #3 speed readings began to rapidly increase reaching 800 kph (432 KIAS) just before impact, #1 speed readings remained at 0. The pitch angle remained at 30 degrees below horizont until impact, 5 seconds prior to impact a right bank of 25 degrees developed. The MAK wrote: "A preliminary analysis of the recorded information, as well as an analysis of similar cases that occurred in the past, suggest that the development of a special situation in the flight could be caused by incorrect data on the flight speed on the pilots indicators, which in turn was apparently due to icing of the pitot probes when the heating systems are off."

Strange no mention of any GPWS warning. While this comes from the FDR, not CVR, I believe that the GPWS status is being recorded in modern FDRs.

[Evan]

the #3 began to decrease reaching 200 kph (108 KIAS), the aircraft pitched down to about 30-35 degrees below horizon

St_ll

I'd like to buy a vowel...

If this is accurate, the #2 and #3 pitots must have been unaffected. Two must be in agreement or the autopilot would have self-disengaged. That 108kts speed is probably reliable,

But why is the pitot-heat not automatic? For instance, on the A320, you have a choice between AUTO and ON. Both are on in the air. There is no OFF in the air (unless a CB is pulled or the air/ground sensing system fails...)

[3WE]

Familiar

Familiar...like something they do multiple times during each flight...select power and attitudes to attain climbs, descents and level offs...multiple times each flight. Did I say every day, multiple times per flight...familiar???

...memory items...memorized training...memory items...memory items....

Please, encourage guys to memorize your numerous "items" (that's plural, you know)...I know, "familiar" just isn't sexy enough.

However, I'd ask if memorizing that pitot heat should generally be turned on when flying in cold clouds, might be emphasized a bit more than memorizing the exact power setting and attitude one should use for UAS.

[Gabriel]

Two must be in agreement or the autopilot would have self-disengaged. That 108kts speed is probably reliable,

I am not so sure... Probably the 800 kph (432 KIAS) of #1 just before the impact was accurate. There is no way for an airplane to accelerate from 100 kts to 430 kts in a few seconds, not even in free fall. There was an accident in Argentina in 1997 (the last pax jet fatal accident in Argentina) where the pilots thought they were too slow and about to stall. The maxed the thrust and started a descent but the IAS was still going down. They extended the slats and of course the oversped and overstressed them, because you know, Newton's opinion is that an airplane descending with max thrust and nose-down pitch is not going slow regardless of what the IAS indicates. Ok, the thing is that they totally lost control and started to fall almost vertically, the IAS still recorded low, until at some point it suddenly increased to almost Match 1. Of course that was not the plane accelerating faster than the gravity, but the ice in the pitot melting. I am thinking that something similar may have happened here (I am talking of the airspeed indication, not the slats). The magnitude of the impact seen in the video and the size of the debris seem to confirm a very very high energy impact, more like 400 kts than 100 kts.

[Evan]

I am not so sure... Probably the 800 kph (432 KIAS) of #1 just before the impact was accurate. There is no way for an airplane to accelerate from 100 kts to 430 kts in a few seconds, not even in free fall. There was an accident in Argentina in 1997 (the last pax jet fatal accident in Argentina) where the pilots thought they were too slow and about to stall. The maxed the thrust and started a descent but the IAS was still going down. They extended the slats and of course the oversped and overstressed them, because you know, Newton's opinion is that an airplane descending with max thrust and nose-down pitch is not going slow regardless of what the IAS indicates. Ok, the thing is that they totally lost control and started to fall almost vertically, the IAS still recorded low, until at some point it suddenly increased to almost Match 1. Of course that was not the plane accelerating faster than the gravity, but the ice in the pitot melting. I am thinking that something similar may have happened here (I am talking of the airspeed indication, not the slats). The magnitude of the impact seen in the video and the size of the debris seem to confirm a very very high energy impact, more like 400 kts than 100 kts.

What about 432 KIAS in a dive with full thrust applied? We have seen how licensed and experienced pilots sometimes get themselves into a stall and then firewall the throttles without doing too much else.

Stall warning, full thrust, pull-up, stall, Newton, uncommanded dive, overspeed, loss-of-control, fireball?

[Evan]

Familiar...like something they do multiple times during each flight...select power and attitudes to attain climbs, descents and level offs...multiple times each flight. Did I say every day, multiple times per flight...familiar???

OK, let's start over with your first assumption... Gather one-hundred commercial transport pilots very familiar with flying in RVSM airspace and ask them when was the last time they manually set pitch and power for level flight at cruise altitude.

[ATLcrew]

OK, let's start over with your first assumption... Gather one-hundred commercial transport pilots very familiar with flying in RVSM airspace and ask them when was the last time they manually set pitch and power for level flight at cruise altitude.

This commercial pilot had to do it about two years ago. Still alive.

[Evan]

This commercial pilot had to do it about two years ago. Still alive.

Every day, multiple times per flight?

Anyway, 3WE is making the same moot point as always... it's not about an inability to fly using the most basic airmanship, it's about human factors getting in the way (and perhaps a paucity of upset recovery training, or just learning it wrong). The reason memory procedures and checklists exist is to prevent those human weaknesses from interfering with that airmanship, something that seems to happen every time there is a crash like this... Something to which every commercial pilot better know they are susceptible. Memory procedures are there to quickly stabilize things so that you may think clearly and methodically apply your airmanship.

[3WE]

Every day, multiple times per flight?

I was wrong about the "set" part.

What scares me is that you might be correct that they don't pay one single damn bit of attention to what the power setting is at all whatsoever, since they are so busy memorizing all of your memory items... you know that acronym PM and then the duty that the PF has to do whatever the hell "M" stands for while the auto-everything flies the plane.

That being said, I see that someone once set the power (and I also recall that this person once operated a crappy regional jet, while naked, and having no autothrottles)...so a question back to you, is if I might gather 100 ATP's and the odds that one of those might just operate a CRJ, or any number of turpoprops, or the Cape Airways boys...by the way, where's my black and white voo doo doll that all aircraft and all pilots use auto throttles all the time?

By the way...did you throw out a big acronym?: RVSM? See, I thought that the autopilot maintained the altitude using elevator inputs, so that you didn't need autothrottles to set exactly the right power anyway...so RVSM or not, as Hillary once said, "what difference does it make".

[3WE]

Anyway, 3WE is making the same moot point as always...

Indeed. Knowing that it's a good idea to turn on the pitot heat when flying in cold clouds and how that might be prioritized versus how exactly to maintain altitude is moot point.

[Quench]

Does pitot heat ON apply heat to the static ports also ?
If the static ports iced up too that may have added to the confusion.

[Gabriel]

What about 432 KIAS in a dive with full thrust applied? We have seen how licensed and experienced pilots sometimes get themselves into a stall and then firewall the throttles without doing too much else.

Stall warning, full thrust, pull-up, stall, Newton, uncommanded dive, overspeed, loss-of-control, fireball?

The problem is not the speed but the rate at which the speed changed, aka acceleration (tangential acceleration to be more accurate). The gravity is about 19 knots per second. That means that in a vertical fall with no lift, no drag, no thrust, the speed increases 19 knots every second. The plane went from 108 to 432 knots. I the above condition that would take 17 seconds. And the plane was not falling vertically. And 30/35 degrees nose down will give you like 0.5G of tangential acceleration (without drag). And the engines will give you like 0.2G (without drag). So if the speed increased that much in one or 2 dozen seconds, it's unlikely that it was accurate in both extremes, and chances are that they will miss-indicate less, not more. That would make the 430 knots the accurate one. And probably the pilots had been acceleration since many seconds earlier, when the indicated airspeed was going down.

[Evan]

I was wrong about the "set" part.

What scares me is that you might be correct that they don't pay one single damn bit of attention to what the power setting is at all whatsoever, since they are so busy memorizing all of your memory item

No, busy managing the autopilot, because...

By the way...did you throw out a big acronym?: RVSM?

It is not kosher for pilots to manually fly under normal conditions at flight level in RVSM airspace (and why would they?). RVSM was made possible by the height-keeping abilities of modern autopilots. So, as most flights in the developed world operate in this space, autopilot is usually engaged at 400ft and remains engaged until final. I'm serious. How many pilots could tell you what a "familiar" pitch and power setting is for an A330 at 250,000lbs and FL360? I think that (combined with human factors) is why many of the unreliable airspeed incidents revealed by the AF447 investigation—none using the memory item values, all improvised—resulted in upsets that were reported to be difficult to manage. Think about it.

If I recall correctly, AF447 was tooling along at somewere between 1° and 0° pitch just prior to the autopilot disengagement. And that was at something like 75% N1 (slowing for turbulence penetration speed)! Is that what would pop into your head as familiar pitch and power settings?

At full CL thrust and 5°, you are going to depart flight level a bit, but you are safe from stall and overspeed and you can fine tune that with the FCOM in a minute or two. That's the idea.

[Evan]

The problem is not the speed but the rate at which the speed changed, aka acceleration (tangential acceleration to be more accurate). The gravity is about 19 knots per second. That means that in a vertical fall with no lift, no drag, no thrust, the speed increases 19 knots every second. The plane went from 108 to 432 knots. I the above condition that would take 17 seconds. And the plane was not falling vertically. And 30/35 degrees nose down will give you like 0.5G of tangential acceleration (without drag). And the engines will give you like 0.2G (without drag). So if the speed increased that much in one or 2 dozen seconds, it's unlikely that it was accurate in both extremes, and chances are that they will miss-indicate less, not more. That would make the 430 knots the accurate one. And probably the pilots had been acceleration since many seconds earlier, when the indicated airspeed was going down.

I don't want to quarrel with physics, and you know your physics, but XL888 went from 99kts at 3000ft to 263kts at 340 ft (where the recording stopped) in a relatively shallow -14° final recorded pitch. And Swiftair AH5017 went from about 190kts at around 9000ft to 380kts at 1600ft (where the recording stopped) with pitch varying from 80° to 58° where it ends. Based on that precedent, does it seem plausible that you could go from 108kts at 6000ft to 432kts before the recording ends in a nearly vertical dive with TOGA thrust? Or does physics not allow for that?

EDIT: ah. forget it. I just saw that the pitch was limited to around -30° throught the descent.