Boeing 737 Max8 Aircraft Crashes and Investigations [Part 1] 20

[itsmoked]

Indonesian 737 Max 8 - Splashed

Most reports I've seen mention the aircraft had an 'undisclosed issue' before take-off. Does anyone know what that was?

Surely the crew was in contact and reporting something during the relatively long disaster.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[LittleInch]

My point to a certain extent though is that maybe it wasn't the actual sensor ( the bit of metal in the photo) that was faulty, but the analogue input card or perhaps the determination of what the analogue input was.

For control systems I use you can get say a varying voltage of 1 to 5V. That goes into the analogue input card which generates a number between say 1000 and 5000. How that number is then interpreted is down to the setting in the PLC so that in this instance 1000 means -30 degrees and 5000 means + 30 degrees. The actual input into the controller then uses these interpreted values to do the actual control.

If those setting somehow got screwed up you have your discrepancy. It's pretty clear from the trace that both AoA indicators were working as they more or less mirror each other, but were about 20 degrees apart.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Alistair_Heaton]

A mechanical throwing a thought out.....

Would a bad earth screw it all up?

 

[IRstuff]

Since the AoA sensors were disagreeing, that should have been a failure condition for safety of flight during the design phase to determine what the system should do, and unless the system has enough smarts otherwise to get past that, it should not have been attempting to control flight with conflicting sensor data. That seems to be a serious design fault at Boeing, if that's what happened. Is there even a flight condition that can cause such readings that would be backed up by the AHRS data. Moreover, it looks like the AoA sensors were disagreeing, even before takeoff.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Alistair_Heaton]

to be honest if they are so mission critical on these FBW machines there should be three of them.

 

[zeusfaber]

Resolvers are traditionally very good at coping with frequency- and voltage- wild supplies, which is one of the reasons they became so popular for safety-critical signalling (they're also less inclined to wear a hole in the track about their habitual null point before flicking their output wildly from one extreme to another than potentiometers).

The interesting thing is that the traditional way of reading a resolver signal has long since changed from being an electromechanical receiver (perhaps with its own servo loop) to being some contraption built from phase sensitive rectifiers and comparators - which may (if the designer has thought about it very much) provide adequate power supply quality and EMI resistance.

A.

 

[waross]

you can see the circuit diagram on the side of it. Its just a potentiometer.

No, not potentiometers. They look like some sort of Selsyn transmitters. (Self synchronous transmitters)
I saw tens of thousands of dollars damage done in an instant due to a reversed connection to a selsyn circuit.
The Selsyns matched and zeroed. But as soon as an error was detected the correction signal was such as to increase the error rather than to reduce the error.
I'm not saying that this was the issue here, but there are a lot of ways that a Selsyn can be improperly wired to give a faulty signal.
Some miswiring may cause a fixed offset in the indicated position.
Some errors may cause a signal reversal so that the worse the error gets, the harder it tries to make it even worse.
If the Selsyns are mounted so that they face opposite ways, eg, both facing towards the end of the wing, then either different models of Selsyn must be used on each side, or the aircraft wiring connections must be reversed on one Selsyn.
If the Selsyns are Left and Right, then installing the wrong "hand" Selsyn would most likely cause a reversed error signal.
I am assuming that the AOA sensors are plug connected. If they must be connected wire by wire then the possibility of a mistake is much multiplied.
I still can't believe that a system that is showing a 20 degree error between sensors dose not turn itself off instead of fighting the pilots into the ground.



Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[waross]

Hello Alistair;
Thank you for your toleration of my comments.
In regards to a "Wild Frequency".
There are several ways to cope with a varying speed of an input shaft.
First, the input speed and the frequency of a synchronous generator are considered to be "locked" together.
But there are exceptions.
A generator is normally excited with DC.
If, however the generator is excited with AC then the frequency may be changed relative to the normal synchronous frequency.
Packages are available to vary the frequency of the excitation as the speed varies so as to provide a stable output frequency despite variations of the input speed.
If the frequency does change with speed changes then there are two main solutions.
1> The generator output may be rectified and then inverted back to AC at a stable frequency.
This system has been used successfully in small, permanent magnet wind generators, often with a battery bank on the DC link to provide backup.

2> All equipment may be designed to accept a varying frequency.
This is not a problem with electronics which typically rectifies the input to DC and may easily be made both voltage and frequency agile.
Large AC motors and transformers do not respond well to wide frequency variations.

You asked about a bad ground connection. That depends on the purpose of the ground connection.
In power circuits we often ground circuits to limit the voltage in the event of some other failure.
We may also ground a circuit to avoid damage due to a condition known as a "Discontinuous ground fault", a fancy name for an arcing ground fault. I doubt that this is the case in aircraft wiring but I don't know.
If a ground return is used to save a wire, then a bad ground may be the same as a bad connection in any circuit conductor. If this is common practice in aircraft then yes, but probably not for a signal circuit. A wire may be used rather than a grond return so that all circuit conductors have similar impedance.



Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Alistair_Heaton]

No problems at all sir...

Its quiet nice to be honest talking about these thigs with people that are more into the design and systems than I am, I was mostly structural when I was working as an engineer,.
If the techs want to know about a fatigue fracture I can help out but the systems stuff I am just an educated pilot.


If its saves weight then they won' use a wire. Its the same with the hydralics (which i have a bit more clue about than electrics) They will vent to a nearby low pressure line rather than run another pipe back to the reserve. The classic case which used to happen was the none retrun valve would get stuck and then the hydralic pump would explode as it was using the internal pressure behind the plate as its reference pressure. The NRV gets blocked and then the oil seapage into the pump for lubrication would gradually pressurise the internals which couldn't vent. This then meant the output was raised to 2000 psi above the internal pressure of the pump. Eventually the ouput pipe would blow off.. then they would replace the pipe... then the pump would explode.. In my experence usually over wales in in the UK.

the reason why i said a grounding fault is the fun and games i have had over the years with car electrics

 

[Sparweb]

Still looking at those function plots from the preliminary report.
Some things to bear in mind (I didn't realized until now):
Auto Pilot was off for the entire accident flight.
In the previous flight A/P was attempted briefly then shut off for the rest.
The automatic trim is "mostly" overridden by full flaps down but not all the time (time stamp 23:23:30),
Pitch trim position varied in the top half of its range for most of the flight,
Pitch trim position became more aggressively "down" in the final 2 minutes, but crew reduced their counteracting "ups",
Airspeed was fairly constant, but seems to be slow.

I can't form a clear picture without additional information:
[ul]
[li]Units of scale on the trace plots[/li]
[li]Position of primary flight controls (stick and rudder) for both pilot and copilot.[/li]
[li]Engine power settings and RPM[/li]
[/ul]

In fact the ABSENCE of these plots is worrisome in itself. I am concocting in my head some crazy scenes on that flight deck.

No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
STF

 

[SwinnyGG]

I'm curious about the nature of how exactly different AOA readings from each side would be accounted for under all possible real world conditions...

The airflow over any airliner's body is going to be highly variable. Apparent wind in flight can come from any vector which has a strong enough axial component to keep the airframe aloft.. surely there must be some situations where the two AOA sensors will not exactly agree because of variable airflow effects at different points on the aiframe surface.

So even with both sensors functioning, there must be some situations where perfect function provides mismatched AOA signals; that means that detecting faults is not as simple as 'AOA signals do not agree'.

Of course it's possible I've got something wrong in there but it seems logical in my mind.

 

[itsmoked]

jgKRI; I'd agree with that but there are other sensors who's inputs can corral or limit the allowed discrepancies in those areas of expected divergence. Certainly if they are disagreeing while the craft is sitting on the ground the cockpit should be alerted.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Alistair_Heaton]

Its now being said that only one sim in the world has this MCAS system actually coded in it.

So its no wonder nobody has been able to reproduce the flight in the sim apart from boeing who has the only sim in Seattle.

It appears as such the vast majority of pilots flying the machine have had no training with the system and most have no knowedge of it or how it works never mind the failures.

Boeing is sying its covered in the QRH with excisting procedures.

 

[human909]

Its now being said that only one sim in the world has this MCAS system actually coded in it.
So its no wonder nobody has been able to reproduce the flight in the sim apart from boeing who has the only sim in Seattle.

It appears as such the vast majority of pilots flying the machine have had no training with the system and most have no knowedge of it or how it works never mind the failures.

Wow. That just makes Boeing even more screwed on this one IMO. From what I can see a faulty sensor and run away MCAS was almost an inevitability with the design as it was. It was just a matter of when, and how it would be dealt with by the pilots.

Boeing is sying its covered in the QRH with excisting procedures.

The 'existing procedures' do cover it. But the symptoms that are faced are different from most pilots would expect. Even the ones who did get it right on the previous flight still were still confused about the symptoms. They never reported run away trim.

Boeing is ass covering on this one. (Of course like most accidents there are multiple points of failure. If communication and maintenance attention on the issue had been better, there would have been a much better chance of resolving the issue and hopefully providing the feedback to prevent further incidents for all other operators.)


QF 72 had a similar incident to this. Again a imperfect design, by seemly not as bad a design as MCAS. Either way there were significant injuries and the issue was addressed by Airbus.

https://en.wikipedia.org/wiki/Qantas_Flight_72

 

[waross]

Despite the large number of people involved in signing off on this design I suggest that a relatively small number were actually familiar with the programing and details of the system.
I suggest that most of the people were basing their approval on descriptions supplied by the first small group.
I suspect and hope that Boeing will revise their internal procedures:
How a new design is checked:
Possible component failure.
Unintended consequences of failure.
How the above mentioned checks are verified.
How information concerning changes is supplied to pilots.
I suggest that Boeing may be in need of a lot of overall procedure changes that are not directly related to this specific failure.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[berkshire]

A note in AVWEB this morning says that Lion air is cancelling further orders of the 737 Max 8.
B.E.

You are judged not by what you know, but by what you can do.

 

[Alistair_Heaton]

the AD is now out

http://rgl.faa.gov/Regulatory_and_G...5835b004fc051/$FILE/2018-23-51_Correction.pdf

quiet why the AoA disagree remains an option I really don't know.

 

[itsmoked]

Is Manual Stabilizer Trim completely manual? (non-electric)

What does the FEEL DIFF PRESS light indicate?

Wow, So STAB TRIM CUTOUT electrically disengages the trim?
All those poor blokes needed to do was disengage? Something seems wrong about this, surely the flight crew knew they where having a trim battle, wouldn't they head for STAB TRIM CUTOUT pretty quickly? Or do you think they didn't actually realize it was a trim issue?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Alistair_Heaton]

The feel system is in there because of powered controls.


When we fly manual we fly in a trimmed condition. Which means all it needs is fingers to manipulate the controls and if you took your hand off then the plane would continue without deviation. When the aircraft is away from the trimmed position for a speed and power setting with powered controls without feel you would never know the plane was out of trim and when you released the controls the plane would then try and restore itself to the speed that it was trimmed for. So the fee system gives control force feedback which increases the further away from the trimmed position you are at.


I belive the Stabilzer trim is a metal cable to the trim tab on the elevator. On the classic you know when the aircraft is trimming because the wheel moves at high speed and is noisy and is right next to your knee. (to note I have only ever been in a 737 classic sim and on a jumpseat)

There are other things which can cause a nose down pitch one of which being an extremely forward CoG On a pax aircraft though its extremelly unusual for the CofG to change suddenly during flight. In frieghters more than a few aircraft have been lost just after departure when the load has shifted rearwards and taken the aircraft into the unstabilised area of the CoG flight envelope.

I wouldn;t say it disengages the trim more it cuts the power to it. There are situations where it jams and is in a locked position which could be due to icing on the tail plane. I have had this situation after accepting an aircraft which had be deiced with type IV fluid a couple of days earlier and they hadn;t cleaned out the trim tab joints down the back and it had turned into a gel and had frozen at altitude where the temp was -60 deg C... which is a different problem. But that also gives changing control forces as the speed changes and you alter configuration.


I don't have a clue how much trim change occurs when you go from takeoff flap to 0 flap on the 737. There will have been loads going on and alerts triggering in that cockpit. 90% of the problem is usually identifying the issue and running the correct QRH page. If you misidentify and run the wrong page it can make matters much worse. Actually i suspect if they had shut an engine down it would have likely kicked it into a none normal mode and killed this MCAS system.

 

[itsmoked]

Thanks for that great info Alistair!

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Alistair_Heaton]

https://youtu.be/A5POuch59YU

That's a 737-800 sim run for an engine fire from a couple of low hours guys I think who are doing some multi crew course It gives you an idea how the wheel moves normally.


Engine failures and fires like this example are relatively easy to sort out because the symptoms are pretty easy to solve as to the reason for the issue and the solution is obvious. ie secure the engine and fire the fire bottles