Autopilot uncommanded disconnect - certification requirements.
 

Hope someone can help. I was flying my 737 NG in max rate mode (trying to meet an ATC restriction) ...was at least 6,000 feet below max cruise alt wings level and still above the amber band (start of min manoeuvre speed) when the autopilot disconnected and the aircraft pitched significantly nose down. What are the certification requirements of an autopilot in this regime....the engineers are just saying tested OK nothing wrong with it.

NFF - The line engineers favourite abbreviation. The fact is that in an NG below optimum you can actually climb in the amber band....just don't go doing 30 degree banked turns!!! The autopilot should not disconnect in the circumstances you describe.

FAR 25.672 has the basic airworthiness certification requirements. They include an appropriate warning that must alert the pilots to failures.

AFAIK, in practice that extends to the autopilot disconnect alarm sounding EVERY time the autopilot is disconnected, intentionally or otherwise. That is why the second press of the disconnect button is required to silence the alarm. if the alarm did NOT sound in your case, I believe that would be a grounding discrepancy.

Thanks guys....anyone like to comment on whether you would expect the autopilot to be disconnecting 10 knots above manoeuvring speed (top of amber band in the NG case).

Greetings,

737 and 727 200 , if , using a spoon you hit (not hard of course) the control colum the auto pilot would trip, this was more than 20 years ago and we still dont know why :confused:
Aircraft pitch down and autopilot trip could be a transcient from ADRU altitude increase to which the autopilot will respond to a pitch down to maintain the demand, and then disconnect, investigation will; as one colleague wrote close the matter with NFF.
Auto pilots are certified using DO 178B level A, this Design Order is highly constraining for hardware and software, specifically when it comes to CAT III.
Level A means NO FAILURES hardware, or software within MTBUR, to the exception of fail passive and fail operational for CAT II/III B.
The component of the board processing the signal are different from the board monitoring, the power supply is different, and the software is different on the boards, and all boards are physically segregated from one another:ok:

Autopilot doesn't disconnect. Just above 1.3 Vs for the current config, it enters Command Speed Limiting mode, shows the underspeed symbol in the speed window (A for Alpha - symbol for angle of attack) and uses either pitch or thrust (whichever was engaged) to maintain that speed (as best it can). A speed 15 knots greater than than the limiting speed needs to be selected to remove the symbol.

Kijan pointed out that there was a history of AP disconnects for something other than what it says in the book...if I hadn't known that it wasn't one of the 12 things an autopilot should disconnect for then I wouldn't have snagged it!

I read the question. You asked:
I answered.

Just a wild guess - it may have reached max elevator to maintain the attitude for the speed at that height? If it then had insufficient nose-up available....? Was it a steady speed climb or decaying?

BOAC...it was in max rate climb but was essentially maintaining mach 0.70 for the minute or so prior to disconnect. What do you think?

This can be a problem with older autopilot types, even the NG aircraft have a rather basic autopilot, carried essentially over from the earlier models, with minor improvements...at least according to a good friend of mine who was an instructor on the NG series.
In fact, I asked him specifically about this.
Also, he mentioned that there was just one aircraft that he had flown (and is now back flying) which had the very best autoflight system he had come across, one that does not trip either off nor to CWS in the event that the elevator runs out of authority (all-flying stab, elevator mechanically connected thereto), so these aforementioned nuisance trip problems simply do not occur....out of trim annunciated for the pilots, yes, well in advance.
The design is over thirty five years old, and still works faithfully.

TriStar.

Quite likely not.;)

Actually, ASFKAP, several still operate from Portugal, and the last time I has there, it was still a member of the EU.
Several still in the UK, as well...RAF.

Dragging this back on topic from the 'Aviation History and Nostalgia' gang, I don't know, RMC. Either you had a duff A/P or it reached some limit and appears to have been out-of-trim (nosedown) at disconnect? We need some avionics input here from an NG qual techie.

.. come on, guys .. if you don't cut 411A some nostalgia slack .. I'll start reminiscing about the halcyon Electra days (another fine Lockheed product) ... and probably even get a tad maudlin that I didn't get an opportunity to fly (although I had a few pax rides in) the larger Lockheed product ...

So, what's it to be ... one bloke reminiscing (and invariably politely) ? ... or two ?

so great that the crews would be still talking about it ten years after they'd been withdrawn from service....

touché, good sir .... brilliant ...

Even the beloved machine got its quirks sometime:

http://www.tristar500.net/articles/cstmp_stansted.pdf

Did the pilot in LH got command ranks after all ?

Yes, it does, however, if flown properly, these problems don't develop.
Case in point.
This nonsense of having the First Officer fly the approach, then hand over for landing to the Captain, can provide difficulties, and I expect this was the outcome, as a First Officer (upgrade to Captain) occupied the LHS.
Also, once no dual is annunciated at a low altitude, the affected tripped off autopilot (if it does trip off) should not be re-engaged, as otherwise problems can occur, especially if autothrust is engaged at the same time.

When in doubt, follow the Lockheed supplied Aircraft Flight Manual, which clearly indicates the way the airplane should be operated for automatic approach/land maneuvers.

RTF (Lockheed)B...for best results, as otherwise pilots are just asking for trouble.

And, this does not just apply to one type either.
The aircraft manufacturer goes to rather great expense to indicate how the airplane is to be flown and the procedures to be used...problems arise when airlines 'think' they know better.:rolleyes::rolleyes:

Hi guys

I must say that the civilized fencing going on here is providing us upstarts with some great technical and operational knowledege in the process

ASFKAP...411A....please do go on:ok:

.. one of the inevitable consequences of being olde pharts .. is that we pick up these and those bits of information and knowledge along the way ...... a strength of PPRuNe I might suggest ...

- no objections, John, on the proper thread? While you are cutting posts, could you delete RMC's double post since he seems unable/unwilling/(unaware?) to do so?

May I respectfully suggest, as an ex-mod (and an olde phart), a thread on how wonderful the 1011 was (correctly in AH&N, of course), and this one on A/P disconnect in the climb, since we do not yet appear to have an answer apart from ASF's?

BOAC,

(a) thanks for the observation re double post .. I'd missed it altogether .. now merged and tidied up

(b) suggest you shoot a suggestion to 411A re a nostalgia thread on the Type .. he'd have to be the best man to start such a thread ... I think that we'd have Buckley's of getting him not to put the occasion L1011 plug into threads ....