737-500 crashes in Kazan, Russia

[Sidebar]

Is there the proper name available (a correct terminology) for this variety of accident ?

Loss of control?

[pdw]

Right, that's the MAKE ... sort of like C.F.I.T. is a certain make, or type.

MODEL would be the "variety", like 'a load shift' .. or 'the inadvertant/unexpected PITCH UP into precarious attitude' (as here) .. or ' windshear stall ' .. or ' wing separation '. As accident starters these are eligible underneath a " loss of control " sub-heading, but these aircraft are already doomed at the point when their respective accidental scenarios began if an unrecoverable accident sequence is already in progress well before actual "loss of control" has been achieved or recognized.

[Gogona]

What seems interesting to me too is that the Co-Pilot had exactly 150 hours before moving onto the 737. I don't know the russian aviation regulations at all but I would imagine that those are the bare minimums to get your CPL. So at some point they must have had 737s flying around with a captain with, hopefully, +2.000 hours and a Co-Pilot with less than 200 hours...

Believe it or not, but it's a common practice in Russia. There's a tremendous deficiency in the numbers of commercial pilots, and that makes airlines forced to hire the new crew members right out of college. So these guys jump into the big jets directly from their SEP trainee aircraft.

[swordfish]

What am I missing in flying an aeroplane? You pour the piss to the engines, the nose pitches up, you push forward on the (power-assisted) control yoke, retrim, and fly out of the situation.

Non-event.

Eric...please help out! You're the voice of credibility here..

[cncpc]

Is there the proper name available (a correct terminology) for this variety of accident ?

The airliner noses up inadvertantly too much here ... while (understood to be) in a process where a steady approach turned into an overshoot procedure. Not like the one in Iraq on youtube where the cargo airliner stalls due to the load shift ... but looks like a similar result

Whip stall would be in the running as a proper name, I suppose.

[pdw]

In practice, did it only once or twice, the nose had to be gradually pulled up to at least 45 degrees above the horizon so the full power stall could be made, where I recall the advice to watch for when the "whip stall" tries to turn it (172) in the direction opposite to its single engine rotation for the tendency of a quick spin entry.

Here two engines would give it a better chance to stay straight at TOGA power, ... just so tragic not having the necessary height to recover.

If whip stall is primarily about the suddenness into a stall, like how it must have happened here (where the turning aspect would only be the secondary concern on singles) ... it's probably close enough ... to call it that.

[cncpc]

In practice, did it only once or twice, the nose had to be gradually pulled up to at least 45 degrees above the horizon so the full power stall could be made, where I recall the advice to watch for when the "whip stall" tries to turn it (172) in the direction opposite to its single engine rotation for the tendency of a quick spin entry.

Here two engines would give it a better chance to stay straight at TOGA power, ... just so tragic not having the necessary height to recover.

If whip stall is primarily about the suddenness into a stall, like how it must have happened here (where the turning aspect would only be the secondary concern on singles) ... it's probably close enough ... to call it that.

I accidentally did a whip stall once when I was an instructor in a 172. Full power, full flaps, student pulled back sharply just before the break, we rotated to the nose down vertical almost instantly. An oil can went out the back window over the baggage area, and another one came forward at quite a clip and hit me in the side of the head and put a sizeable dent in the glareshield. No spin.

That was a 172. This is a 737. I only suggest this because of what I saw in the video, and that is as vertical as we were and more or less instantly. I have no idea whether a 737 even does a whip stall.

[pdw]

Speaking of those loose payload items, anything that could move a bit would do so relatively to the rest of the airframe when multiple extra tons of thrust happen all-at-once from beneath each wing of a 737.

"Pitching up too steep" is the result of the delay to provide immediate forward yoke at overshoot decision-time which quickly brings on the need for greater sudden forward-yoke force to counter the quickly-building TOGA thrust where then much stronger down elevator is required, with the added risk that more tail-up is required than is possible with that control input.

In this kind of control problem, is the upward-steering control input for the horizontal tail requiring only a maximum reduction of HS down-force ... or does it turn into an up-lifting tail as well ?

The resulting steep angle also tests any weaknesses in aft constraints of luggage or other payload, allowing any loose payload items to weigh more aft.

When nearing the full power stall it likely also yaws to the side with the least thrust (not necessarily equal) in the fight to recover the decaying go-around airspeed. Probably wouldn't look like how a whip stall feels ... when seeing the larger airliner do that in slow motion.

(EDIT: The grammar needed to be corrected in the second paragraph.)

[YQkC]

To me this speaks to the importance of upset recovery training. The FAA has been implementing rules stipulating required aircraft upset simulator training. The point was really driven home in the case of the Colgan crash in Buffalo where the mishandling of the crew not only caused a jet upset but prevented them from recovering from a very recoverable event.

[lownslow]

Is this the upward-steering of a horizontal tail here requiring only the maximum reduction of HS down-force ... or lift as well ?

What?

Whatever it is you asked, I doubt the issue here was the design or construction of the airplane.

LnS.

[pdw]

Pitch got out of control, but appears to have taken a while. Almost a minute to impact and still over the airport.

This sounds like at overshoot was tracking slow, when it got away. Lots of thrust, but perhaps not enough speed to correct pitch ?

[GyvAir]

Looks like there is a fair bit lost in translation regarding descent and negative vertical acceleration, but it sounds like there was a considerable delay between applying go around thrust and any attempt by the pilots to control the pitch attitude of the aircraft.
Would the trim work automatically with the autopilots disengaged on a 737-500?

From AvHerald:

Information released by Rosaviatsia on Nov 28th 2013 based on preliminary investigation results by MAK states, that the captain of the flight (47, ATPL, 2736 hours total, 2509 hours on type, 528 hours in command) was rated for CAT I ILS approaches only, the first officer (47, no type of pilot rating provided, 2093 hours total, 1943 hours on type) was rated for CAT II ILS approaches. According to preliminary information the crew was significantly (4km) off the approach track prompting ATC to query the crew. Corrections were made, the aircraft remained significantly right of the extended runway centerline however until the crew selected heading 250 into autopilot (heading select mode) and the aircraft intercepted the localizer automatically about 2nm short of the runway threshold at about 1000 feet AGL, the glideslope did not capture due to height however. After passing the missed approach point the crew discussed a go-around and communicated with dispatch, then disengaged both autopilots engaged in the automatic approach and continued manually on flight director. Engines accelerated to 83% N1 (near Go-Around Thrust) and continued at that speed until almost impact. The aircraft began to pitch up under the influence of engine acceleration and flaps retraction reaching 25 degrees nose up, the stabilizer trim system wound - most likely automatically - nose down commanding the aircraft into a dive. In the meantime the crew retracted the gear, there had been no input on the yoke since deactivation of autopilot until that time, the airspeed had decayed from 150 to 125 KIAS. The crew now applied full forward pressure, the aircraft began to accelerate again after reaching a minimum speed of 117 KIAS at 700 meters/2300 feet above the runway, and began to rapidly descent, EGPWS alerts "SINK RATE" and "PULL UP" sounded, there was no reaction to the extreme nose down attitude however and the vertical acceleration became negative. The aircraft impacted ground at 75 degrees nose down at about 450 kph at coordinates N55.608818 E49.276852, the impact occurred 45 seconds after initiating the go-around and 20 seconds after reaching the maximum height.

[pdw]

....a considerable delay between applying go around thrust and any attempt by the pilots to control the pitch attitude of the aircraft ....

From AvHerald:

... After passing the missed approach point the crew discussed a go-around and communicated with dispatch, then disengaged both autopilots engaged in the automatic approach and continued manually on flight director. ......

Go around being automatic after the MAP when too high, and when the pilots are 'discussing' there at any length with ATC (along with focus on disengaging flight equipment to go manual for the rare event), ... does that divert too much attention from the required routine ?

EDIT: to refrase the question

[Eric Janson]

What am I missing in flying an aeroplane? You pour the piss to the engines, the nose pitches up, you push forward on the (power-assisted) control yoke, retrim, and fly out of the situation.

Non-event.

That's what should have happened - the investigation will reveal more about why it didn't.

This aircraft did not stall - the speed was very low but still above the stall speed. The trajectory flown appears to be almost parabolic.

The 737 has a speed trim system that will trim nose down under the conditions of this go-around.

When I flew the 737 my company had 20 degrees nose up as an absolute limit - anything above this was an unusual attitude. Nose high with a decreasing speed results in a loss of elevator effectiveness and an increasing difficulty getting the nose down.

My company had a special Simulator session dealing with nothing but unusual attitudes and recovery from them. We used to get a few unusual attitudes on the re-current checks as well.

Not clear why they were off track initially - maybe they had a map shift and were trying to intercept the ILS in LNAV (At my company we had to intercept in heading mode never in LNAV).

On the older 737 models the only time a dual autopilot approach is flown is during an autoland. In this case a single autopilot go-around can be made.

In all other cases the approach is flown with a single autopilot engaged. A go-around will need to be flown manually. In some models of 737 the autopilot will not engage until the Flight Director bars are almost centered.

It looks like the crew just sat there and let it all get out of hand. I'd be very interested to see what kind of training they had.

[burhead1]

....a considerable delay between applying go around thrust and any attempt by the pilots to control the pitch attitude of the aircraft ....

From AvHerald:

... After passing the missed approach point the crew discussed a go-around and communicated with dispatch, then disengaged both autopilots engaged in the automatic approach and continued manually on flight director. ......

Go around being automatic after the MAP when too high, and when the pilots are 'discussing' there at any length with ATC (along with focus on disengaging flight equipment to go manual for the rare event), ... does that divert too much attention from the required routine ?

EDIT: to refrase the question

No winds in this one?

[pdw]

"Winds" were 230T @ 17G22kts (8G11MPS).

Approaching Kazan from the west the aircraft was 4km to the north/northeast of rwy 29 centre line of 304T/292M, a deviation corrected by adjusting course to 42degrees left of 304T with the heading 250M/262T likely maintained for about a minute where the "150kts" airspeed is around 138kts ground speed in the 30 degrees left crosswind for a reduced left crab.

Next, the rwy-heading intercept-turn is made "automatically" approx 30-32 degrees to the right, to a 70-79degree left crosswind. Here at 2km-final/1000AGL the ground speed stays steady even if airspeed reduces 10kts during or right after the turning. There the Glide Slope is likely missed by only a slim margin with help of the high/breakeven ground speed as the strong on-final left crosswind crosses the 304T-track at 40 degrees closer to perpendicular, the increasing left crab angle is at the expense of the reducing headwind/ lowering airspeed.

Was then airspeed getting a touch low for this go-around initiation ?
The pitch up stems from a combination of control changes : reducing the flap (1st noseup-tendency) with a low thrust line for the high power just applied (2nd noseup-tendency), that then required forward elevator attention (both nose up tendencies left unchecked) but were not yet sensing a threat from steeper climb-angle while in the 'low airspeed but high speed ground speed' environment.

Q: Is 'nose up tendency' allowed to proceed here intentionally ? If there's less elevator effectiveness at the pre-TOGA airspeed ... leaves the undetected 'reduction' in the elevator's nose-down effectiveness to progress because it is just not seen or felt. What pitchup degrees did the pilot see ... same as we're told from the FDR ? The excess pitchup degrees are probably more like 5-10 ... where the lateral left and right turns just before where quite steep at 30 left and 30 right ...

[pdw]

(30 degree heading changes at 15 degrees of bank)

Burhead1, the autopilot performed both 30deg heading changes ... and the bankangle used there was much steeper than any excess increase in pitch-up angle that got them seconds later.

The "winds" here make lateral adjustments to the course a bit more rocky too, so the sensation of pitch-change for the nose-high exceedance left so little to feel in comparison.

Would it not be better to leave the whole go-around to the autopilot in a situation like this ... after so much erratic low-level turning ?

[Diadem]

pdw, you do know that wind has absolutely no bearing whatsoever on airspeed, right? It might produce some illusions due to drift, and windshear can temporarily cause a change in airflow over the wings, but otherwise the aircraft moves through the air regardless of the direction in which that air is moving. Imagine a balloon filled with air, with a tiny airplane moving around within it. If you move the balloon, the air inside it moves in exactly the same direction and at the same speed. Likewise, the airplane continues to move through the air within the balloon at the same speed and in the same direction. It's only when comparing the movement of the aircraft to the ground that winds become relevant, and then only for things like groundspeed, crab angle, and touchdown speed. Wind has absolutely nothing - zero, nada, zilch - to do with the aircraft's performance while it's in the air unless there's windshear. The plane moves through the air in exactly the same way regardless of whether the airmass is moving at fifty knots or zero. Aerodynamics isn't affected by groundspeed, and a crosswind wouldn't affect the aircraft's pitch, or airspeed, or whatever you're trying to say.

[pdw]

The plane moves through the air in exactly the same way regardless of whether the airmass is moving at fifty knots or zero.

That's exactly the point to make here, IMHO.

The turn onto runway heading at 2km from the threshold at 1000AGL was not the right estimate to catch the glide slope. On another day with just 10kts on the beak it might have worked fine, for 500ft per km. There the ratio of ground speed to airspeed increases in the turn which helped being high ... contrary to the goal.

[burhead1]

The plane moves through the air in exactly the same way regardless of whether the airmass is moving at fifty knots or zero.

That's exactly the point to make here, IMHO.

The turn onto runway heading at 2km from the threshold at 1000AGL was not the right estimate to catch the glide slope. On another day with just 10kts on the beak it might have worked fine, for 500ft per km. There the ratio of ground speed to airspeed increases in the turn which helped being high ... contrary to the goal.

Makes perfect sense!

[AJV]

It looks like the crew just sat there and let it all get out of hand. I'd be very interested to see what kind of training they had.[

[/quote]/quote] me too

[pdw]

Doesn't look like there was all THAT much time there, since at first were traveling at a fairly good ground-related clip up until just before disengaging the autopilot.

From there is only " 25 seconds " until the steep portion of the climb is already over (at 117 kts) and the first 10 seconds are consumed in the spooling of engines. With the higher ground speed, to see a satisfactory initial climb angle when coming directly out of their lower descent pitch there would need a significantly larger positive pitch change at first (so pitching it up more there, or too much, wasn't hard to do).

It reads like the gear is up by the "117kts" and also that manual trim is not ruled out (ie which means PIC focus / time is also spent on that).

Then there's the seconds of stabilizing (Yoke Back Pressure Required / YBPR-1), arresting their descent as power comes in (YBPR-2), while retracting 15 degrees of flaps (YBPR-3), then raising/steering the pitch to the best angle (YBPR-4); so up to the "117kts" where the accident sequence is in progress, there was just 15 seconds of time.

In addition, 'gear down' holds wheel drag well below the thrust-line and the pilot flying it manually will be already holding the YBPR(5) to counter its nose-down drag right at the beginning of the retraction cycle. Gear is also selected up during this short climb portion, which momentarily adds increase to positive pitch until their travel passes the zero thrust-line with the engines. Then in addition to these 5, as the wheels are briefly still protruding from the wells now just ABOVE the thrust-line of the engines, a stiff REDUCTION of Yoke Back Pressure is needed there only briefly for that extra surge of pitch-up moment that is not coming from Yoke Back Pressure.

[lownslow]

I started flying professionally with the goal of one day getting into the big airlines but now, I just don't know man, you make it sound really difficult. Borderline scary, even.

LnS.

[pdw]

... you make it sound really difficult.

O.k. I'll admit it IS difficult to calculate how pitch-up could have been so excessive to set up for this accident in just a few seconds.

The highlighted safety statement on the report suggests it's being seen as a matter of practicing going missed manually ...

"They continued on in 'Fight director mode' after the MAP", and now "Rosaviatsa re-iterated" (Aviation Herald report/ Kazan Nov 2013) 'for pilots to do SIM practice for missed approach in 'Flight director mode' '.