I was cruising in LNAV/VNAV PTH on autopilot A. As an experiment I disconnected the ap by pushing hard forward on my yoke. However, it appears that I broke something since I am now unable to reconnect ap A. I can disconnect and reconnect autopilot B all day long by pushing the button on the MCP, but I cannot find any way of reconnecting autopilot A. Plane is in trim and flying level. I'm assuming it is operator error - does anyone have an idea why autopilot A appears to have stopped working?

To partly answer my own question (for the benefit of others who run into the same), it appears that pushing on the yoke caused an "FCC_A INOP" failure. Clearing that failure restored autopilot A to service. Does anyone know why pushing on the yoke would case that failure?

You can disco the A/P by manual override, but indeed there's a force around 110 lbs (if memory serves me) that will indead break something, rendering the FCC INOP for the remainder of flight. Did I get that from the intro manual? I believe so, would have to check. EDIT Yeah it's on 116 intro manual. To save you ten seconds:

[...] While any autopilot is engaged in CMD mode, yanking at thecontrols really hard will result in "breaking" the controller shearrivets (real value 110lbs of force) and destroy the onside FCC(flight control center). The associated autopilot cannot be re-engaged. [...]

Edited August 7, 201114 yr by badderjet

EDIT Yeah it's on 116 intro manual. To save you ten seconds:

Thanks, Etienne. Learning to fly the NGX is heaps of fun / Tom

Hi,This might be a stupid question, but I have the same problem. When I push my yoke, autopilot A is switched off, and won't be available for the rest of the flight.You wrote that it can be fixed by clearing the failure. Where can I clear this? Many thanks in advance,Frans Plugge

Hi Frans, It is in the failures section of the CDU. If you are having difficulty finding the right pages, there is a description in the Introduction document starting around page 80.

Read closely above, as it certainly is a factor HOW you push your yoke. A gentle push (or pull, roll, whatever) should disconnect only - a hard yank will break the system for the remainder of the flight though.

I wonder how the system knows the difference between breaking the connections and simply wanting to go into CWS.

What, the NGX or the real thing? On the real thing something physically breaks (that's why it's inop for the remainder of the flight) at more than about 110 lbs control force. The NGX has "fly-by-software", as they call it. As you will probably physically break your hardware as well when applying 110 lbs to it they probably just take deflection or also rate-of-deflection in order to make a decision.

The AP controls the yoke through some servos. By measuring how much power those servos draw, the system can work out how much power is being exerted on the yoke to override the AP commands. Given a small rate of deflection, the system has time to react and disconnect the servos, but if a large enough force is applied, those shear rivets give way. Obviously, this is designed this way so that in case of an AP error and disconnect failure, you can still overpower it and disconnect it the hard way. I somehow don't think that's an option on an Airbus.

I somehow don't think that's an option on an Airbus.

Weird, I kinda have the same feeling... LOL

Read closely above, as it certainly is a factor HOW you push your yoke. A gentle push (or pull, roll, whatever) should disconnect only - a hard yank will break the system for the remainder of the flight though.

Etienne, Even when I push the yolk gently, the yolk still has to travel a significant distance before I see either CWS-P or CWS-R on the annunciator. After this point, I cannot re-engage CMD A. Even though I have not checked for any failures to clear, the movement is gentle enough such that I should not have any failures but I will check the next time and will also try re-engaging CMD mode with CMD B. Also, when the A/P disconnects completely (manual flight), I cannot get CMD A to re-engage. Would you know of any conditions that must exist prior to re-engaging CMD A, i.e., what will stop it from re-engaging in level, non-turbulent flight (not on the ground) ? Thanks,Zach

Yeah, actually you might be right. I know what you're talking about, and I've had similar problems. Especially because in the pitch channel, if you stay within 250' (IIRC) of your previous altitude while going into CWS-P, then ALT HOLD will reengage. This can only be overcome by applying significant control force so as to not leave CWS, and this usually results in violent pitch rates and vertical speeds until you leave that 250' zone. The CWS joystick sensitivities could really need some adjustment, but then again the values have to work on a variety of hardware. Also, I've had the issue that CMD would not engage at all. Actual engagement conditions are not rocket science, but there are certainly some weird things happening that make CMD stop working for no sane reason. This behavior is often preceded by the FD bars not appearing with active FMA modes, at least when they definitely should. I posted a thread about this some time ago. BTW I think the failure you're looking for when the shear rivets break is "FCC A/B FAIL".

I wonder if there is an audible "snap" or "pop" as this happens.

Negative. The rivets break in the gear bay (IIRC), so I doubt you could hear anything.