China Eastern 737-800 MU5735 accident March 2022

[A0283]

The life feed shows video from this morning. The search at that time recovered what appears a big fragment of the top of one MLG, broken in half, the big connection rods broken in half (deformation - no plasticity it seems), trunnion on one side still shiny, the fragment required 5-6 man to carry, so say roughly at least 120kg weight.

Recovery status impression around 1050utc:

cockpit area - probably or only light material (documentation reported),
tail - maybe found today 25th,
horizontal stabiliser - big part suggested but not seen in photo or videostream yet,
winglet left hand - no sign,
winglet right hand, with part of the outer wing - right hand most likely,
wing itself - fragments of outer wing possible,
fuselage - from early on part of the top panel in front of the fwd escape hatch, probable,
fuselage - multiple large pieces of outer shell, all bent and torn and pushed flat,
engine - one certain, other no sign,
MLG/NLG - suggested is one of the main,
instruments - no sign or report of any till now,
instruments - wiring, no sign or report of any till now
total - in total about 183 parts, bigger fragments are collected in a row of tents on location,
total - a few hundred structure fragments and perhaps other types of fragments in between,
CVR - confirmed

Rough outline of the accident area:

The main area of interest is running N to S in what I call the gulley (low valley you might say). In the upper half of that NS-stretch is the main impact area which contains almost all (bigger) fragments). To the North of the gulley is a burned forrested area. The area with the 'possible burned landslide' (the other one of Lonewolf_50s two photos) is an estimated 300-400 m to the East and has no involvement with the accident.

[EDLB]

SCMP writes that they found potential parts 10km (6miles) from the crash site.
https://www.scmp.com/news/china/arti...mpaign=3171854

Which would mean that something departed long before impact form the plane.

[Mudman]

Originally Posted by EDLB

SCMP writes that they found potential parts 10km (6miles) from the crash site.
https://www.scmp.com/news/china/arti...mpaign=3171854

Which would mean that something departed long before impact form the plane.

10km radius around the crash site.

10km radius around crash site

[Teddy Robinson]

Originally Posted by EDLB

SCMP writes that they found potential parts 10km (6miles) from the crash site.
https://www.scmp.com/news/china/arti...mpaign=3171854

Which would mean that something departed long before impact form the plane.

from the article "Zheng Xi, head of the Guangxi regional fire and rescue office, said the piece of suspected wreckage was about 1.3 metres long and 10cm wide (4.3ft by 4 inches).
It was found 10km away and the search area was expected to be expanded by half, he told a press conference in Wuzhou, near the crash site."

Balance tab ?

[silverstrata]

.
When the 737 NG was first introduced it did lose a number of elevators in flight, due to excessive turbulence from the flight spoilers. So the max speed for spoiler deployment was reduced to 240 kt for a number of years - which was a pain in the arrse on the descent, because the spoilers are only really effective at high speeds (when the angle of attack is reduced).

To cure the problem, an extra hinge-point was added to the NG elevator, to stop them falling off.
Needless to say, losing elevators can spoil your day.

Silver

[nnc0]

Is it curious that the CVR module has been recovered but not the FDR?

Given the speed of Impact and trajectory I'd expect them to be found right beside on another. The crater isn't that big so where is it?

[EDLB]

The plane crashed with bullet speed into the ground leaving mostly small shrapnel sized parts. So the remains will be in the 30m radius by 20m depths shared with all other ingredients. The people working there have my deepest respect. The FDR memory will be found but it can take a while. The Chinese make a very professional job there so we will get to the true cause of this accident.

[DaveReidUK]

Originally Posted by nnc0

Is it curious that the CVR module has been recovered but not the FDR?

Given the speed of Impact and trajectory I'd expect them to be found right beside on another. The crater isn't that big so where is it?

I believe the recorders in one or both of the Max accidents were only found after much searching, so it's not altogether surprising that the Chinese FDR is still outstanding.

Both recorders will have been reduced to mangled fragments - only the stainless steel or titanium shell that protects the memory modules needs to survive - and the missing one could well be several feet down in the mud.

[grizzled]

Originally Posted by DaveReidUK

Both recorders will have been reduced to mangled fragments - only the stainless steel or titanium shell that protects the memory modules needs to survive - and the missing one could well be several feet down in the mud.

DRUK has it exactly right. Sadly, I have been to several accident sites involving high speed, near vertical impacts. You cannot imagine (don't want to imagine) such a scene. In addition to everything being reduced to simply fragments, those fragments are twisted, torn, dirt or mud-covered (sometimes burned) pieces of pieces of pieces. The orange boxes (exterior portions) are no exception, as they are usually the same dirty colour as everything else, and often in smaller pieces as well. Almost nothing looks like something you would recognize, because everything was crushed and conglomerated in an instant.

So, patience is called for, as is a large dose of respect and empathy for those doing the onsite work.

[Sailvi767]

Originally Posted by CW247

All 737-800s at MU are grounded
My hunch is a runaway trim.

Why would they not simply disconnect the trim or just grab the trim wheels? Both those fixes assume the system that automatically shuts down a runaway trim if the control column moves opposite the trim direction has failed. A runaway trim at cruise altitude should be a non event.

[Mookiesurfs]

Originally Posted by Sailvi767

Why would they not simply disconnect the trim or just grab the trim wheels? Both those fixes assume the system that automatically shuts down a runaway trim if the control column moves opposite the trim direction has failed. A runaway trim at cruise altitude should be a non event.

They could and they would. Probably not runaway trim, imo. A more plausible scenario for trim related loss of control at altitude is a gradual failure of a system that causes the autopilot to trim unremarkably, like a gradual fuel imbalance, rollback, or something else. The AP will trim to the limit of its authority and then kick off, handing you a plane that is well out of trim. This really should be manageable at 29,000’ but can be tricky the closer you get to 41,000’.

We can’t rule out a catastrophic event at cruise yet, bomb or otherwise.

[Sailvi767]

Originally Posted by Mookiesurfs

They could and they would. Probably not runaway trim, imo. A more plausible scenario for trim related loss of control at altitude is a gradual failure of a system that causes the autopilot to trim unremarkably, like a gradual fuel imbalance, rollback, or something else. The AP will trim to the limit of its authority and then kick off, handing you a plane that is well out of trim. This really should be manageable at 29,000’ but can be tricky the closer you get to 41,000’.

We can’t rule out a catastrophic event at cruise yet, bomb or otherwise.

The trim system in the 737 is very obvious if not annoying when running. That includes when the autopilot trims. If in fact they ignored the spinning trim wheels and noise as soon as the autopilot moved the elevator to counter the incorrect trim input the trim brake would engage and stop the runaway.

[PJ2]

FWIW, I don't see a longitudinal trim issue (runaway stab) as causal, particularly with fuel distribution which is primarily a lateral issue, (also, not sufficient fuel on board to cause a large imbalance issue), or an engine "roll-back", ("lateral roll followed by upset" scenario). Also, continuous stab trim operation would be immediately noticeable these days & handled as per (non-MAX) runaway drills.

So far, we haven't seen any thorough analyses of the vertical development of weather in the area. Tim Vasquez produced some great work, many years ago now, on the AF447 threads.

Whatever occurred, did so very swiftly.

Horizontal stabilizer jackscrew operation:

You Tube

[Mookiesurfs]

Originally Posted by Sailvi767

The trim system in the 737 is very obvious if not annoying when running. That includes when the autopilot trims. If in fact they ignored the spinning trim wheels and noise as soon as the autopilot moved the elevator to counter the incorrect trim input the trim brake would engage and stop the runaway.

It’s noticeable, but we’re talking intermittent inputs here, not a constant running. Anyone with time in type knows that the plane is constantly making intermittent trim inputs on its own, and that it is normal unremarkable background cockpit noise. A constantly running trim would be remarkable, but that’s not what we’re talking about.

[exosphere]

Regarding the wreckage piece allegedly found 10km away, it could well be that the initial upset caused it, but it’s not necessarly the reason for that upset. Entering such a dive would require a relatively quick pitch down which also means negative G’s, maybe exceeding final load factor.

[Al Symers]

Don’t rule anything out. If you saw some of the crazy sh!t I saw flying in China you’d agree that just about any scenario is possible.

59 y.o. Captain gets hard terrain warning. Must be punished. Downgraded to FO as he approaches retirement. Gets rostered with a leader captains (who downgraded him) son. Makes for an interesting cockpit dynamic. Who’s really in control in that situation? Anyone who’s flown in China knows it’s not the guy in the LHS. Up there it doesn’t matter a damn what’s right, it’s all about who’s right. Add in lots of hands on the controls, (a staple in Chinese aviation) and it doesn’t take much to create a disaster.

Might have just been the cranky old guy giving the yoke a shake to make a point, or stomping on the rudder (yup, seen that happen), with zero regard for consequences. Engine failure and stall also a possibility, keeping in mind that if you let a yaw develop in to a roll, it will eventually lead to aerobatics if left unattended.

Saying aviation in China is much safer than it used to be is like saying wars aren’t as messy as they used to be. The weapons may have changed but it’s the same result in the end.

[Turkey Brain]

Initial descent can be achieved with low positive G, no negative needed.

Originally Posted by exosphere

……. Entering such a dive would require a relatively quick pitch down which also means negative G’s, maybe exceeding final load factor.

Assuming the descent started at the time of the last cruise data point, a big assumption !
To lose 2,000 ft in 14 seconds is average about plus 0.3 g.

As long as the descent started at least 11 seconds of the first altitude readout of 27,025 ft then zero g would suffice.

Coincidently, if the descent was at an average of zero g for 11 seconds the rate of descent would be 21,737 ft/min at the next data point, 27,025 ft. using my simplified calcs. Assuming gravity, G = 10 m/s etc. So rough calcs.

The data point from FR24 gives a Rate of Descent at 21,696 ft/min.

So surprisingly similiar numbers. Maybe I made a mistake in my calcs.


But if correct, the aircraft didn’t need a high negative g bunt to achieve the initial descent, just a lack of lift.

All pretty horrible, but many scenarios achieve these numbers. 2 immediately spring to mind.

1: Wings level zero g profile.

2: a roll past or near 90 degrees.


1: The wings level zero g scenario seems highly unlikely, a jack fail or trim run away would in my guess be unlikely to achieve exactly zero g. I would assume some sort of dramatic manoeuvre but this is not an area where I have any expertise.

2: So very little lift component up. The problem with the roll case seems to be the lack of heading change.


The fidelity of the FR24 data may mask track changes. I note a 2 degree track change to the left passing 27,000 ft and 10 degrees left passing 25,000 ft. It’s back on track,( 100 degrees ) at 22,500 ft, but by 15,000 ft it’s now deviated to a track 40 degrees to the right.

A slow spiral ?

I can’t tell from the numbers. Maybe someone can.



equations used.

V^2 = U^2 + 1/2 a t ^2. V = velocity achieved, U initial velocity ( 0 in this case ) , a acceleration, t time

S = Ut + 1/2 a t ^ 2. S = distance

[fdr]

Originally Posted by PJ2

FWIW, I don't see a longitudinal trim issue (runaway stab) as causal, particularly with fuel distribution which is primarily a lateral issue, (also, not sufficient fuel on board to cause a large imbalance issue), or an engine "roll-back", ("lateral roll followed by upset" scenario). Also, continuous stab trim operation would be immediately noticeable these days & handled as per (non-MAX) runaway drills.

So far, we haven't seen any thorough analyses of the vertical development of weather in the area. Tim Vasquez produced some great work, many years ago now, on the AF447 threads.

Whatever occurred, did so very swiftly.

Horizontal stabilizer jackscrew operation:

https://www.youtube.com/watch?v=rxPa9A-k2xY

Nor do I;

reading the follow on posts, here is some physics for pondering for the other posts....

Planes fly nicely unless they are either -

1. commanded to do something else; or
2. upset by external factors; or
3. the fundamental stability of the toy gets upset.

A plane that is happily bumbling along at 1.0g doesn't suddenly decide to go to 0.3g. the trim solution is to an angle of attack, and it is not going to have a large excursion from that to some other value just because the autopilot decided to turn off. If the autopilot elected to enter a wild ride, then as long as the autotrim function is running the plane will happily retrim to the new angle of attack and do what the errant autopilot computer is commanding. The plane continues otherwise doing its short period or long period oscillations around the mean trim alpha. Yaay. has been like that since Wilbur and Orville decided that the world needed flight attendants to make it a more palatable life.

A pilot pushing on the prong will drop the nose, as long as he (or she, but most she's are not stupid enough to do what the boys do) is pushing. As soon as he (or she) lets go of the input, the plane is going to revert to the trimmed alpha, and it will do a recovery all on its lonesome. Not just nice to have, it is a mandatory design requirement, even for unstable designs, the 1's and 0's have to be able to guarantee that is always the case.

getting to a seriously nose low condition takes a lateral excursion in order to occur unless the tail has gone its own way, and as was previously mentioned, with Lauda 004, "Mozart?" when the inadvertent thrust occurred on an engine in the mid climb, the PW4060 issue resulted in a huge disruption of the flow over the wing section immediately behind the nacelle, and that caused a large lift loss as well as yaw. As the aircraft rolled from the loss of lift, and from the secondary effect of the yaw, eventually the structure got to a point of exceeding the strength of the tail and the tail failed, by memeory with an asymmetric failure of the horizontal stab/elevator and then complete failure of the vertical stab, whatever, shortly thereafter, from the loss of the horizontal stab, the Cm from the wing and the CG shift of losing a stack of weight at the tail resulted in a rapid pitch down, that ended up with an overload of the main wing from an excessive negative g loading, after which the wings departed, and what was left was not much at all. A full on failure of the tail usually ends with the sky filled with chaff, an inflight breakup is almost always the outcome. JAL103 was an exception to that.

An aircraft that has a lateral upset is the most common means of getting into a nose low attitude. This can be from curious causes, Silkair being one, or the Classic loss out of Jakarta last year with the autothrottle clutch pack issue, or from entry into turbulence, or from instrument failures, like Stansteds B747-200F loss, or Adam Air with the INS's attitude being reset in IMC... that'll do it every time. When the aircraft rolls to a high bank level or inverted, the longitudinal stability will tend to drop the nose as the plane enters a spiral at high bank angles or tries to do a split S if inverted. The planes stab trim doesnt need to be altered and the outcome is a degrading flight path. What doesn't happen is the plane going to 0.3 g or other, it is trying to get back to the trimmed alpha, and that means it will tighten at the high bank angles, so g loadings increase. Tracking data if it is reasonably high resolution may show that a track change has occurred, but getting the nose down without changing the trim needs the roll.

On the morning of this accident, I was planning a medevac into China for the following day, and the plane involved had a recent radar issue that had resulted in a radar swap. As part of the planning, I looked at the weather in southern and eastern China, about 4 hrs before the accident and there was convective weather forecast near Guangzhou and to the west, and otherwise the area was quite clear. tops above FL260 were not significant except in that area, and the lightning strikes were being displayed in that area, then next closest area was over Pusan/Pohang and south of Vladivostok.

The grounding of 200 plus B738s seems to be premature. Unless there is clear evidence of a stab tab or other item falling off around the top of the drop, there is not much likelihood int his case that a structural failure was involved. Parts that separated part way down were unlikely to be the cause, they are the consequence.

Having said all of that, occasionally stuff really is surprising. I recall one investigation where the crew got 12,500FPM upwards, busted the assigned altitude by 4000' topped out at 0.2g, and then got 12000FPM on the down side and busted the assigned altitude a second time for good measure. The pull up was 1.3g, the over the top was 0.2, and the pull out was 1.3g. That got a spot on the wall, for the rest of the day until the next incident surpassed it. During the rollercoaster, the Captain managed to complete the full multi language PA to the passengers welcoming them on board and thanking them for choosing to fly brand X. The FO has the misfortune to be in a spritely climb in a light plane and to encounter an entry into a jetstream that increased the headwind by well over 100kts over 4000' and he elected to stay in FCLH targetting Mach.... so the latter climb was all done with the thrust levers at idle, and one very confused FO. The cabin crew continued their excellent, long suffering service, and didn't leave any dents in the ceiling.

TBs maths is OK for a wings level entry, using a dFPV/dt, V input, to determine a g loading, but anything that achieves that in a wings level condition is a seriously catastrophic event, the fact that the aircraft had any reduction in the severity of the dive angle suggests it was still responsive to control un until the time it then wasn't. 73 drivers don't often do push overs, recall that they drivers are sitting a fair arm forward of the CG, and the rate of pitch needed to go to zero g gets the drivers up close and personal with the fuel panel and the pressurisation control panel unless well strapped in.

Pushing to zero g and holding it for any time is a good push on the prong, gonna get tired doing that for long, and to hold it, without running the stab trim, that is a big ask

[CodyBlade]

You Tube

[henra]

Originally Posted by fdr

TBs maths is OK for a wings level entry, using a dFPV/dt, V input, to determine a g loading, but anything that achieves that in a wings level condition is a seriously catastrophic event,

And a bizzare one for that matter. I do not see many mechanical ways to end up with a proper wings level 0g nose dive from cruise. Trim runaway with full ND deflection and then the AP letting go would rather end up with negative g and seems also extremely unlikely. Why would the trim want to do that? Mis- trim scenarios did happen in the past but that was typically at low speeds, e.g. during approach but at cruise speed???
Losing Vertical tail will involve yaw/roll (could potentially be the case but the trajectory speeds >M0,9 do not smell like any kind of tumbling). Losing horizontal tail at cruise speeds would lead to massive negative g and shedding wings downward. Terminal angle being 90°ish. Jackscrew nut leeting go. That might hypothetically be possible. But in a seven y/o aircraft?? And then, why exactly at TOD? why suddenly and fully in one go without any prior 'warning'? In the case of the Alaska MD the failure developed over some time. The flight profile here appeared absolutely clean until the drop.
I'm somewhat struggling to conceive a 'hard' mechanical failure leading to the profile we apparently have seen here.