I have seen that predictive capability many times over when the target A/C gets out of coverage.

That is a feature of the animation onscreen. The raw dataset available for download, which I hope anyone talking above had been using, is just plain record of the collected ADS-B transmissions.

FR24 seem to truncate some of it (sub zero FL readings) but otherwise it is what it is.

-- x --

Not sure if the TOD discussion is relevant, i.e. whether the events unfolded after the descent was instructed by ATC.

Observation. Take it for what you will: The final three ADSB speed outputs show the groundspeed decreasing; which is not indicated by any noticeable pitchup in the two videos, coupled with an increase in the rate of descent.

Possible speedbrake deployment?

A true pearl of wisdom - always understand the instruments and sampling process behind the data.

Back in the late 1990's, I worked 737-3/4/500. We had repeated issues with poor maintenance with Chinese operators - things like throttle cables breaking due to excessive wear less than 100 hours after they were supposedly inspected. When I worked at a car dealership while in college, we called that a 'wall job' - customer brings in the car, it gets parked up against a wall for a few days - customer comes back and picks up the car with a sizable repair bill when no actual work had been done (didn't happen at my dealership, but I knew of dealerships where it did).
I'd rather assumed that sort of thing was a thing of the past as the Chinese market had matured - maybe not...

A runaway trim scenario in which the pilots were surprised would seem unlikely given the 737 manual Trim wheel would have been spinning and clicking away a few inches from each pilot. Hard to not notice that.

Don't take for granted it was in maintenance, China has strict Covid policy at the moment, domestic flying is way down with 60% of flights canceled. Could have just been grounded due to canceled flights.

737 Throttle cable?

However this began & finished, it wasn’t “throttle cables”.

What is irritating is how thinly veiled the racism is and how quickly it surfaces on a thread like this.
Contrary to popular belief you don’t need to be able to speak English to fly a jet.
Last time I checked there are 6 official ICAO languages, Chinese being one of them.
Chinese system is different, ATC stands for Air Traffic Control.
This is where you fly, this is your offset, this is where you descent. The pilot operates the airplane and ATC controls the airplane.
Different doesn’t mean it’s wrong or somehow inferior to what “we” do.
So typically Pprune-ish, "well 30 years ago I had this happen or 30 years ago that was the case so it must be the same now blah blah blah"
NO.
That’s exactly what Boeing's response was after not one but two fatal accidents.
’Well it wasn’t a western company so they have done it wrong’.
STOP doing that.

I flew in China for ten years
 

A lot of people had posted their opinions of what could have happened.

I flew in China for 10 years, being last flight in 2020. I'm not racist, just pragmatic.

The ATC there is crap. Not only the control itself, but also the attitude from ATC to the to pilots. Chinese pilots accept all kind of stupid requests from the ATC without a word. It is almost impossible a Chinese pilot refuse a stupid ATC instruction. Specially if the crew had a 2700 hrs Cpt.
It is well possible that they had an engine failure , request to descend and got a "stand by" as an answer. And they lost speed until the stall. It really would not surprise me at all.
I really think that China lasted a lot without a fatal accident. They had imminent accidents though out this years.
I remember foreign pilots had years of discussions in the cockpit because Chinese pilots flew at night with their radar off, and worst, with cockpit windows covered with newspapers to avoid radiation. Its happened until the day a wide body China Eastern flew into an isolated CB and lost 10000 ft. And so on. Many, but many stories. If you want to know more download flying upside down.pdf from the Internet.

So I really believe this accident had a partnership between ATC and crew. But.... as everything in China we will never know. BTW , probably the boxes were destroyed, but the communication between the plane and ATC wasn't revealed yet. Probably will never be.

Fortunately for all foreign pilots flying in China , it happened with a local crew.

My condolences to all families involved.

If that were the case, we would be seeing a decrease in ground speed before the start of any descent. As it is, the aircraft appeared to be flying straight and level, then there's 15 seconds where no ADS-B data is available and in that 15 seconds, the aircraft has dropped 50 knots GROUND Speed but had already lost 2000 feet of altitude and was now descending at 21.000fpm. Correct me if I'm wrong, but first reaction to engine failure in cruise on a 737 is disconnect autothrottles, select MCT on the operative engine, perform any recall items and then maintain cruise flight level until reaching single engine drift down speed which can take a couple of minutes.

Here we have 15 seconds between cruising along steadily and suddenly descending rapidly. My initial thoughts were a runaway trim scenario - maybe not the sort which has the trim wheels clanking around - but slow and insidious over a period of minutes which isn't noticed - because the coincidence of an upset happening where they would usually be at TOD is just too much to overlook. So, they set the altitude to, say, FL 100, the AP commanded nose down, but plane was horribly out of trim due some defect. Crew were preoccupied with briefing the approach and before they realised what was happening, they were already in the merde...

The Chinese media are already reporting that they entered the range of Guangzhou ATC at 14:17, started the rapid descent at 14:20 and the ATC repeatedly tried to reach them without any response until the end. Take that however you will.

Also, can’t help but agree that the last part you wrote seems completely unnecessary.

One of the recorders has been found.

https://www.scmp.com/news/china/poli...ox-found-plane

If this was a jet upset and the crew was trying to address the ROD with a standard recovery, I am wondering where the lateral changes in track come into play?

36 seconds after the stable flight path was lost, the aircraft has completed a 40 degree turn to the left.

36 seconds after that is has turned approximately 65 degrees to the right, all the time descending.

This of course is not how you would expect a UA recovery from a deep nose down attitude to be carried out.

Something else is going on here.

From press release: "The captain had accumulated 6,709 flight hours total, the first officer 31,769 flight hours, the second officer 556 hours, all of them in good family relations."
So somewhat more than 2700.

There is some issue in the translation, the 556-hour pilot was the first officer. The first officer listed as having 31,769 hours was a check and training captain, however, in China, they are listed as a F/O on crew manifests.

Data recorder found
 

Actually 30,976 feet per minute is about 566 km/h, not 600 km/h. But you are right to be suspicious about that number, because it's quite close to the maximum vertical speed that can be encoded by ADS-B: 32,640 feet per minute (597 km/hour). So it wouldn't surprise me if that 30,976 fpm value is the maximum that can be reported by this particular type of aircraft due to some additional limitation, and the actual vertical speed probably reached even higher values.

ADS-B encodes vertical speed as a 9 bit integer (there are also two additional bits for direction - up/down - and source - GNSS or barometric), and each increment of that integer represents 64 feet/minute. So there are only 512 possible values for vertical speed, between 0 and 511. And because 0 is regarded as "no data", you have to substract one from that encoded value before multiplying by 64. That gives a maximum of 510 x 64 = 32,640 feet / minute. You can't get more than that encoded in an ADS-B data packet.

Second CAAC press conference today. Key takeaways:

- one black box found, severely damaged
- ATC communications "normal"
- weather not a factor
- Captain 6709h, 1st F/O 31769h (double confirmed), 2nd F/O 556h, all have good family relationships


China changes incredibly fast. I think flight crews who worked there twenty years ago experienced a vastly different place. ATC is an issue as they are military and speak to you in that tone. regulators generally shrug complaints off and ask "why change it when it works?" ... CAAC meanwhile has used a hard enforcement approach to radically change a formerly unsafe aviation market into one that hasn't had a crash like this one in ten years in spite of its massive size. Generally they have adopted an approach of producing excellent investigative material and publicizing it widely in Chinese to get maximum trust from the flying public. They don't care at all about writing anything in english, which I presume is why a lot of folks outside of China haven't noticed their changed practices vs ten years ago.

Often, the SIMPLEST explanation is the correct one….

A NEAR VERTICAL dive in a B737 can only be caused by SUSTAINED PILOT INPUT.

After the SILKAIR MI 185 disaster in 1997, several scenarios including RUDDER HARDOVER, JET UPSET, ENGINE FAILURE, STALLS, DEPRESSURISATION and SUSTAINED PILOT INPUT were simulated by the human factors group, which were part of the accident investigation team. (see final report).

SUSTAINED PILOT INPUT produced a nearly identical rate of descent and just under 3nm lateral distance travelled.

The other scenarios didn’t get anywhere near the same descent or distance travelled.Also, the aircraft exceeded the local speed of sound, causing the ‘boom’ sounds locals heard in the final stages of descent.

I hope the recorders can be recovered in a usable condition. But like the SilkAir accident, I’m sure they would have been disabled before the dive.

Has the tail section been found ?

​​​​​​
How do US vs Chinese fatal crashes match up over the last decade?

your maths is about right, and the aircraft has a resultant TAS down the flight path of 484 kts, At the lower levels that is an IAS of around 460-475 roughly, which is about 100kts over the Vne of the plane. At very high mach numbers, above the height where a change in the descent occurred to a momentary apparent recovery, the b737 wing reduces lift for a given AOA from the development of strong normal shock on the underside of the wing. a normal shock at high mach is expected, the 737 has a visible shock at around 0.78M, the Classic had one quite observable at at around 0.735. On the top, at speeds over MMo, the shock on the bottom of the wind results in a rapid loss of lift for a given AOA, which results in a decaying flight path, or a pretty hefty increase in AOA needed to maintain level flight.

Adam Air DHI 574, and Silk Air reached speeds higher than this aircraft, but both of those had very odd reasons for the upset. Adam Air was not far removed from Air Asias bad day out for causation. The usual testing limit for Mdive is around 0.06M above the proposed MMo, roughly. The B737 is not much fun at taht but it will do it, but 100kts over Vmo is a tall ask for the aircraft. The increased AOA needed to effect a recovery results in very high buffet and that is one of the items that impacts flutter boundaries.

The simulators that are certified for flight training do not properly represent the aircraft at speeds well in excess of Vmo/MMo, they give a sense of comfort that the MACH/CL, and MACH CM data does not support, the trim change is observable in the MMo-Md dive data. Back in the dark ages this was described as Mach tuck, and the explanation out of the AP3456A was related to the shock foot movement rearwards.... yes, that does occur, but the tuck comes from the development of the shock on the bottom of the wing, and that has a very abrupt loss of CL for the trimmed AOA, and that results in a stunning Cm change. The above just suggests that getting nose low in a jet transport goes bad pretty quickly, and the simulator does not have great fidelity for training in that area. For the occasions that I did tests out beyond MMo to Md the recovery was of enough concern that the flights were planned to have a high thrust level in the descent, so that if the tuck was excessive, a speed reduction from reducing thrust could be made. Some planes are just beautiful, like the Falcon, it is impeccable to very high speeds, with full authority without buffet. The B737 Classic has buffet that was odd given the rule, and the NG has a normal envelope that essentially alters to achieve compliance of the rule, through the little orange line that starts to drop down below the MMo indication, same for a few others, and that gives the compliance to the rule, albeit at a mach a little lower than the nominal MMo. In upsets, where the ATR is engaged, the thrust comes back rapidly to idle. If that hasnt occurred, then more has happened than can be explained by a simple upset. That info resides in the engine EEC memory chips, which have some possibility of having survived. The CVR and DFDR of this aircraft probably survived the average deceleration but any heavy structure that impacts them in the deceleration would have way above the 3400g test levels. The average deceleration would be around 2000-2100g, its about 50/50 the FDR or CVR will have taken higher impact loads.

All in all, would be surprised that the plane in this case was the trigger of the event, and an upset can occur without substantial crew error, or any action by the crew. An upset in severe weather can put the plane into a position that care is needed. The odd thing is that the most significant factor in the recovery of a high speed jet upset is the detection-recognition time of the crew before an effective recovery action is commenced. The only exception that comes to mind on that is the stunning survival of the B747 of CAL on its way to the USA that held together well beyond where it should have failed, it was pretty daggy by teh time it got to California, but it didn't splash down.

10 years without a fatal incident in China. There were two fatalities in the same period in the US.

https://news.cgtn.com/news/2022-02-2...Pra/index.html

https://www.airlines.org/dataset/saf...-air-carriers/

China and the US are very close in number of system seats. China briefly was ranked first during Covid but that has since changed in favor of the US again.
Change could be temporary though as China is building a ton of new mega airports and is set to add 8,000 new 100+ seat aircraft in the next 20 years.

https://centreforaviation.com/analys...e-world-521779

That is true for the MI185 case, the flight path was very well defined and it was difficult to replicate it. Other steep descents in the 30-50 degree range have occurred for other causes, the trim runaway being a low likelihood cause (except when someone does something pretty dumb with a system architecture unwittingly).

The data gives enough information to exclude most stalls as a cause, and a number of items can give a roll excursion that will drop the nose to below 30 degrees nose down in short order. The contenders for that sort of thing is a simple TS entry and disorientation, a slat element extension at cruise, a TR deployment, a revisit of the yaw damper/rudder reversal issue etc. Most of these are not likely but they get to the same point in time and space without being MI185 or EGYPT 990, or really odd events like Adam Air 574.

Flash 604 had some other controversial system possibilities raised that were discounted yet had occurred to the type previously, and in fact was photographed in a SAF B737 which got the crews attention. 574, the autopilot disconnection wasn't the basic problem, the turning off of reference systems for the APLT and the ADIs of the crew pretty much made that a most unfortunate event.

Maybe. Or maybe not. I teach people to fly these things. Done simulator with the China Eastern guys, this aircraft. Nice guys, I like them.

It would not surprise me at all if the report says they just kept doing “X“ until the plane crashed; never figured out what was going on. Human beings are fragile things. I would trust the machine over the man. I have seen people lock right up. And the other one just sits there. Totally possible. It is just how we are, we are fallible.

484 TAS is closer to 440 CAS.

G-EZJK made it to 429 and already the AAIB bulletin noted the crew's actions on the yoke alone would not have stopped the descent.

Was right seat in a Viscount simulator (white on black AH) when the left seat got it upside down at 16,000 and just sat there with me screaming in his ear we were upside down until we hit the ground.

Would be a good simulator scenario to have in the curriculum, but don't prebrief the guy expected to take over.

@ Krautland : That was a long time ago, they are civil now. The problem still is that the airspace around the civil routes is mostly military airspace and cannot be penetrated, even for weather.

I have to disagree here. It was an ex NASA astronaut, the Minister of Transport for Canada that called the FAA to tell them that his experts had reviewed the data from both Max 8 crashes and had recommended (and accepted) to the PM that the Max be grounded in Canada. At the time, Trump viewed the Chinese grounding (1st group to do so) as 'political' and anti American. It was the final straw when Transport Canada ordered the grounding.

Check the Washington Post article dated 2019/03/13/

[QUOTE=A-3TWENTY;11204388]A lot of people had posted their opinions of what could have happened.


I flew in China the last 15 years and I still fly into China. I don't know if I should call it racist but I sense a lot of expats here seem to have the superiority complex.
Admittedly I have not operated with a mainland Chinese in the seat but I had a few in the jumpseat as "navigator" into non open airports. They seem reasonable when we had discussion about work.
I agree ATC is crap but it is not true to say the locals comply without a word. I can speak the language so I understand everything on the frequency. Often I heard pilots argued with ATC and some got very fiesty. And so you think a 2700 hrs captain from elsewhere will always speak up??? So there are no near mishaps in other parts of the world?? I do not agree with the punitive culture but it has been 10 years without a fatality. It is quite scary if you look at the safety stats of a large European/"Asian" airline but nobody talks about them even when they had a crash. When there is any incident Chinese related everybody jumped on it. When something happened and a local pilot saved the day nobody talks about it.
Having worked in a few companies with diverse workforce and also as a checker, I have seen people with substandard performance from every continet; so is good performance.

CVR was found yesterday, the base of the unit had significate damage, the cylindrical memory unit which contains the data has the appearance of being in good condition was sent to Beijing last night. The NTSB has also been formally invited to participate in the investigation. https://www.reuters.com/article/chin...on-redirect=in

https://cimg7.ibsrv.net/gimg/pprune....c471c82204.png
Source: Aviation Accident Investigation Center of Civil Aviation Administration

They are indeed. But this is when the training should kick in hard and achieve;

• Establish who has Controls. State it,
• Memory Items,
• Inputs to attempt to establish Straight And Level,
• QRH whilst the other is PIC.
• Work QRH Items for stated Outcomes.

I think they were doing more than just sitting there. I think they were very very busy. I think it's far easier to "Lose it and Lock Up"' in the sim (knowing it's only your job that's over and not your life) - a relatively consequence-free environment than when you're actually there in the air.

Sim work - it's hard to say - there are differences in the state of the mind - in some ways you're harder focused in the sim than real World and in other ways you're softer focused. What you cannot do in the sim is convince someone they're about to die if they get it wrong.
Knowing you're going to die will mean you will do ANYTHING you can to avoid that.
Looking at the somewhat spurious data so far - it looks very likely that the pilots were severely deprived of flight controls from on or soon after the initial upset occurred all the way to the impact on the ground. That would suggest a loss of one or more control surfaces.

A simpler, and more likely, explanation for ADS-B blips would be that there weren't any - the aircraft could well have continued to transmit normally, but for some reason those transmissions weren't picked up by FR24's crowd-sourced receivers, possibly due to an unusual aircraft attitude where the transponder antenna was blanked.

Unfortunately we are living in a COVID world where there are a lot of rusty people in the aviation industry. I have seen many maintenance errors as a result, and have been expecting to see many more. It would not surprise me if this was somehow related to an error. There are a lot of mechanisms in place to get Flt crew back up to standard after considerable time not flying, but in the maintenance field it is very different. And when workplaces are restructured to save costs during this mess, things are more likely to go wrong.

One example of how things can go wrong from a simple error. An aircraft is on the ground for an extended layover and all drains, static ports & pitot probes are covered up (etc, etc). All covers are removed at the return to service, however the pitot covers on the vertical stab are missed. It does not get picked up on the walk around. Aircraft goes flying and everything seems fine until the pilot disconnects the autopilot and tries to fly it manually with high airspeed. The elevator feel is subsequently compromised and the elevator controls are over sensitive.

In no way am I suggesting this as a cause, just an example of the unexpected. I have personally seen aircraft go flying with gear pins left in, drain covers still fitted, and maintenance paperwork not completed. The Aviation industry is different to what it was 2 years ago. Everyone needs to take an extra minute to double check what they, and others, have or have not done.

As many who have worked with or associated with Chinese aviation over the last 30 years would know the countries attitude to aviation is indeed changing to one of trying to maintain reasonable standards. This also involved change of attitudes towards the ex military pilots who were unsuited to airline style work and move to more western style management of training/checking processes for civil aviation. As with many other areas in China the industry is very punitive if you mess up, you don't get many chances if you stuff up, but that's not just China. Considering the challenging flying in China, with mountainous terrain and terrible weather and smog conditions they do a pretty good job, considering the amount of flying done these days. If you try and pin some cultural BS on their pilots it would just be from a prejudice mind.

Thank you for bringing this up.
Yes, investigation should include records of maintenance performed in this CoViD era.
And check all possible tasks that may have been missed (or exceeded if time-controlled).

Fortunately the throttles and flight controls are not interlinked.
With over 10K hours on the 737 including the 800 and Max, and some training in accident investigation, I'd like to offer a few observations.
First off, I don't believe it is possible to reach the ground in two minutes without the throttles pushed up. No one keeps power in, in a dive. Its instinctive for anyone that's been flying for any period of time to reduce power to idle.
Now with power at idle I don't think even vertical you can descend 29K feet in two minutes.
Now to get vertical, you would have to have a condition where not only are the flight controls unresponsive but they have had to go into a full nose down position by themselves, and be resisting input to reduce the angle of descent, which without some kind of assistance would not normally occur on their own.
OR you would have have to a failure of the flight control system as well.
So we are talking two separate events, by systems not related, except by the person flying.
I could be grossly mistaken, but I don't believe you will find much data coming out of China that isn't manufactured. The airplane has been in service 7 years, so bugs have been worked out. This leaves one of several possibilities.
Crew action, maintenance or some external factor.