PPRuNe Forums - 737 AOA failure - System implications

PPRuNe Forums (https://www.pprune.org/)

- Tech Log (https://www.pprune.org/tech-log-15/)

- - 737 AOA failure - System implications (https://www.pprune.org/tech-log/620455-737-aoa-failure-system-implications.html)

737 AOA failure - System implications

Trying to noodle through all the system effects from an AOA failure that generates a false stall signal from the SMYD computer to the active FCC (some of this could also result from a bad airspeed input/blocked pitot, but want to focus on AOA for now). A stall signal from an SMYD will activate certain system response. Good in an actual stall, but unhelpful if result of bad AOA input. From what I have read so far, there is no "compare" function between AOA's that would inhibit these system responses.

Here's a list of things that may happen:

Stick shaker activation (failed side).
Elevator Feel Shift Module (EFSM) activation. Independent of active FCC. Heavier elevator control forces.
Stall ID function of Speed Trim System (STS) trims nose down at .27 units/sec. Can be stopped by aft column cutout switch
Autoslat activation. Independent of active FCC. Full extend slats for any flap setting except 0. Full extend placard limit 230 knots, and will effect LE slat position lights on forward console. Amber "LE FLAPS TRANSIT" inhibited. Green "LE FLAPS EXT" would only illuminate for TE position that would command full extend (varies by specific tail number). Not a huge problem except that it potentially creates another distraction.

This is in addition to the other system implications of unreliable airspeed and altitude (no A/T, no FPV, possibly no A/P).

Informed comments from those with 737 experience appreciated.

737driver
I assume you are talking about the MAX.

Quote:

From what I have read so far, there is no "compare" function between AOA's that would inhibit these system responses.

There is no compare function because the MCAS was shockingly designed with zero redundancy. The two sensors are independent for alternate use. Unlike In Airbus where there is triple redundancy and they automatically compared and single erring one is self detected. There also if two erring ones have common error they can reject the good one. In Max MCAS too much authority was invested without any redundancy.

Yes, I think we are all aware of the MCAS deficiencies. However, there are other systems misbehaving due to a false stall signal, on BOTH the NG and the MAX. I am trying to make sure I’m covering them all.

737 Driver, an interesting viewpoint.

No FD, or none relevant to the situation.

The list should be restated with the appropriate level of alerting to add emphasis - tactile most salient, stick shake and control force, then flashing visual alerts, visual with audio repeat, visual or audio only.

The first to be discarded would probably be audio, then visuals as judged by relevance to the situation. Over time, perhaps stick shake in proportion to the increasing stick force, ultimately a single focus of attention - pull.

Subsequent activity probably subconscious, revert to basics - auto throttle and AP engage, FD(?), undo last switch action, backtrack seeking to identify a mistake, cause.

None of the the alerts actually identify the exact problem and action, nor might they post mod as described so far.

The two AoA vanes can be compared to provide a display of ‘AoA Disagree’ - currently an optional fit, but it would be of no value in resolving a vane AoA input problem, nor would it be of any benefit as a mandated post mod addition - similarly with AoA EFIS dial display.

Mind, memory, seven bits of information limit; little spare mental capacity due to surprise and continuing startle - because the corrective activity does not resolve the problem.

I understand from the D-AXLA report that the A320 compares the AOA readings during level flight to Attitude, but it does nothing with the data except alert maintenance, not the crew.

Quote:

section 1.6.11:
ADIRU
In straight and level flight, when the Mach greater than 0.75, a comparison between the attitude and the aeroplane angle of attack is made by each ADIRU. A class 3 maintenance message(10) is generated if the difference between these two parameters exceeds 0.6 of a degree.

Would it be possible to generate a dynamic sanity check figure to compare against the AOA sensor readings to determine if an AOA sensor should be ignored?

I'm thinking using Ground Speed and vertical speed to generate a velocity vector, and then comparing that to attitude to generate a ballpark AOA with a tolerance - if the measured AOA is outside of tolerance, disregard the sensor reading as spurious.

Quote:

Originally Posted by Blythy (Post 10448309)

I understand from the D-AXLA report that the A320 compares the AOA readings during level flight to Attitude, but it does nothing with the data except alert maintenance, not the crew.

Would it be possible to generate a dynamic sanity check figure to compare against the AOA sensor readings to determine if an AOA sensor should be ignored?

Entirely possible - I believe 787 (and 777?) does exactly that, since only has 2 physical AOA sensors.

Retrofitting 787 ADIRU (if that is where that function is, I think probably) into a 737 is left as an exercise for the reader.

Deciding whether it would be cheaper to do that than have two crashes and a long grounding is left as an exercise for the accountant, with the note that the exercise is purely academic since crossing fingers behind back and saying "it'll never happen" is always going to be the cheaper choice.