I think the answer to the OP is “almost anything”. Going by fatal accident reports over the years many people get killed by totally trivial problems causing distraction. However if you substitute “would” for “could” then the answer is really only major structural failure causing complete loss of control, unless you had a parachute of course.
Peter wrote:
That is the key Q: can you engage the AP on the ground (whether moving or not) in a way which makes it drive the trim?
Yes, you can. There was a fatal accident in the US where this happened. I read the report but it was some time ago and I don’t know where to find it now. It might even have been discussed on EuroGA.
A SEP Cessna (I believe it was a 172 or possibly 182) with a KAP140 pitched up and stalled immediately after takeoff. The trim was found in the fully nose-up position.
The investigation concluded that the pilot had accidentally engaged the A/P while doing other checks — but after setting take-off trim and checking controls free. The A/P box was close to some of the things to be checked. On the ground, the elevator on a 172 rests in a nose-down position. When the KAP140 was engaged, it would sense tension in the elevator cables and drive the trim fully nose-up in an attempt to reduce the tension.
The solution was to introduce a delay so that the A/P will not engage from a momentary push on engage button — you have to hold it for a short while.
Yes – prob99 that’s the one I mentioned.
I think you could do that with almost any autopilot. On my KFC225, a touch of the AP button
sets the autopilot into a capture of the current pitch and (if the wings are within IIRC 7 degrees of level) a capture of the current roll. If you press AP while on the ground, the captured pitch is presumably the, ahem, current pitch, so in theory the pitch axis will not be driven (no need to) but in reality there is a small drift in the system so the yoke will be driven in both pitch and roll, very slowly.
@Airborne-Again
That’s actually similar to the reason why I never use the VS+ALT feature of the autopilot after takeoff but only IAS. The reasoning behind that is: If you have accidentally set an altitude in the autopilot altitude window lower than the elevation of the airport, the autopilot will fly you straight into the ground if you press VS+ALT. Doing that on a IMC departure (probabaly) killed at least 2 Cirrus pilots.
If you press only IAS and have set the IAS wrong the DFC90 autopilot will not stall the airplane but first warn you (“envelope protection active”, “too slow”) and then even lower the nose. Not nice, but will not kill you.
Edit:
Of course you have to check the set altitude aswell, but manipulating the IAS is less risky than the Altitude. Best is to simply climb out with IAS without altitude capture, which can be added at a safe altitude.
That’s actually similar to the reason why I never use the VS+ALT feature of the autopilot after takeoff but only ALT+IAS. The reasoning behind that is: If you have accidentally set an altitude in the autopilot altitude window lower than the elevation of the airport, the autopilot will fly you straight into the ground if you press VS+ALT. Doing that on a IMC departure (probabaly) killed at least 2 Cirrus pilots.
Are you saying that the DFC90 will automatically calculate the polarity of the VS according to whether the preselect altitude is above or below the current altitude, at the time ALT is pressed? That would be highly unusual.
How does it automatically calculate the value of the VS?
Also very very few GA autopilots have an IAS mode.
If you set an altitude lower than the elevation of the airport and press VS+ALT after takeoff, the airplane will descend to that altitude at -500 ft/min. And therefore it is safer to use IAS only (not IAS+ALT).
Peter wrote:
But also the autopilot installation, at the bottom of the stack, was thought to be unusual.
Not unusual at all in the 182. Although the pilot in question must also have been very unlucky as he must have had some climb mode selected in the AP before engaging it. You do need quite a lot of force to overpower the AP in a C182.
That said, an uncommanded trim can kill you, or least try to kill you. Happened to me last year in a C172. It’s an airplane I very rarely fly and am not overly familiar with the AP. Made sure (I thought) that the AP was in ‘heading mode’ only, no VS selected. While reading up on the approach procedures at my destination – there’s a rather complicated noise-abatement procedure – I noticed something wasn’t right. Sure enough, this a/c has auto-trim which had decided to wind all the way nose-up and I was approaching the stall. Push nose over, AP OFF! The weird thing is, after re-engaging the AP – now obviously carefully monitored! – it worked as intended. After returning to base the next day discussed this with a couple of CFIs who are familiar with the installation and nobody could figure out what had gone wrong….
There are two very different issues:
)@Peter
The AP in the C182 I normally fly is an S-Tec and takes quite some force. The one in the event I wrote about above was a King, IIRC. No need to overpower, as disengaging and re-trimming solved the issue, so cannot comment on the forces required to overpower the runaway trim. It felt pretty normal during the ground checks.
The human reaction of trying to pick up a wing drop with aileron results in the wing suffering the wing drop, to move further into the stall regime as the aileron deflected down increases the camber of the outer wing, and lowers the critical AoA resulting in a deeper stall and consequently entering an incipient spin/auto rotation. Quote
I don’t rush to agree with this for the vast majority of certified GA aircraft. I have found following much inquiry that this is mostly historical for former military types (Harvard has been presented as vulnerable to this), and thus the wives tail lingers on.
recall that the prevailing design requirement for spin handling and tolerance reads (my bold):
(ii) No control forces or characteristic encountered during the spin or recovery may adversely affect prompt recovery;
(iii) It must be impossible to obtain unrecoverable spins with any use of the flight or engine power controls either at the entry into or during the spin; and
……..
Then, using ailerons and rudders in the proper direction, it must be possible to maintain wings-level flight within 15 degrees of bank and to roll the airplane from a 30 degree bank in one direction to a 30 degree bank in the other direction
…..
the airplane must respond immediately and normally to primary flight controls applied to regain coordinated, unstalled flight without reversal of control effect
;Quote
Uncoordinated flailing around of any flight control certainly will not promote recovery, but it’s not going to prevent it either, just delay. I have seen “rudder only” spin recovery pilots actually enter a spin from an intended stall, because of their over use of the rudder, resulting in very uncoodinated flight at the stall (which is how one enters a spin). The only flight manual I have read which discourages aileron use is that of the Cessna Caravan, which simply says something like over use of aileron during spin recovery may prolong the recovery because of the spoiler which extends with the upgoing aileron. I can say that the Caravan recovers from spins very well.
If you’re worried about how your aircraft will handle during a spin recovery, refresh your spin training, and then avoid spin conditions, as a delayed recovery from a spin will very likely not be the fault of the characteristics of a GA plane, compared to the pilot skill.
I do very much agree that the first pilot action at approach to a stall should be to crisply lower the AoA. That, combined with keeping the plane coordinated will prevent spin entry.
