AUTOPILOT PROS & CONS

Doesn't anybody do proper modelling of the aircraft behaviour (such as
is done for airliner flight simulators, in order to derive the
equations which are then solved in the sim) and check that the control
loop parameters of the AP provide a sufficient margin?

No. It's not done.

There have been too many cases of APs going unstable under certain
conditions.

I have a Twin Comanche with a Piper Autocontrol autopilot (single
axis), and I can esily force it into a divergent pitch oscillation
mode. All I have to do is let go of the controls in a trimmed
airplane, let it fall off into a spiral past about 30 degrees
(eventually it will, no matter how well trimmed and how smooth the
air), and then just turn on the autopilot in wing level mode. You
know, like pilots who lose it in IMC are told to do. The autopilot
will keep the wings level, but the plane will go into a divergent pitch
oscillation.

Gotta love it.

Michael

I think when this happens in some very old AP, which used to work
previously, it is probably because it used electrolytic caps for
various time constants, and over the decades these dry out, so the
derivative term goes up, the integral term goes down, and the
stability goes out of the window.

That sounds great, but I had the Century test set available to me when
I installed the autopilot, and I used it (the Autocontrol is actually a
Century II private-labeled, which is obvious when you read the service
manual) and everything tested within limits. The autopilot actually
works fine - it does a good job of holding wings level or heading, and
a fair job of tracking the LORAN. Roll stability is fine regardless of
airplane attitude in which any roll mode is engaged. The problem
occurs in pitch.

Anything modern should not suffer from this, and current stuff should
do the control loop in software anyway.

I can make my friend's STEC do the same thing in his Bonanza. An
identical STEC in a Warrior won't do it. Neither one will do it if the
pitch mode is engaged, but in roll mode only it will happen in the Bo.

Michael

Pitch stability is much harder in comparison and I wouldn't even know
where to start. One would need to be a control systems expert.

Pitch stability isn't very difficult if you have pitch sensing and
pitch control. I guess I'm a little bit of a control systems expert
(I've taken control theory courses at the graduate level and designed
and built some oddball control systems) and I could do a pitch control
with no problem. So can the A/P manufacturers. The problem I
described doesn't occur with any 2-axis A/P that I know of, in any
airplane.

The problem is that roll and pitch are dynamically coupled (prove it to
yourself - apply roll input and the nose drops) and the coupling is
different between airplanes, and for a given airplane is likely a
function of mass, cg, inertial moment (controlled mostly by how far
outboard the fuel is), airspeed, angle of bank, etc. Without knowing
any of those things, it's pretty much a given that if there is a
divergent pitch mode (and many planes have one) the A/P will find it
one way or another. What bothers me is not that a given
airplane/autopilot combination can be forced into such a mode, but how
trivially easy it is. No effort has been made to suppress this
undesirable behavior.

If I were going to fix this problem, I would start by having the unit
read the bank at startup, and go into a recovery mode designed for the
airplane, probably involving stabilizing the angle of bank and then
slowly reducing it. Not saying that's a guaranteed fix, but it's
simple and IMO would probably work. BTW, this is simple enough to do
in an analog design, and pretty trivial in software. But there's no
incentive to do it - causing a divergent pitch oscillation by
overcontrol of the roll axis doesn't make an autopilot unairworthy - no
rules cover this.

It may make it unsafe if used in the manner often advocated, but unsafe
does not mean unairworthy (and vice versa).

Michael

There are secondary effects but these are second order. If you apply
roll, the main result will be roll. The pitch change should be small.

Think about this carefully. Imagine you have the plane trimmed for
straight and level flight. Now let go of the controls and allow the
plane to enter a spiral. Will the pitch change be small? Somehow I
doubt it. When an airplane rolls, the vertical component of lift is
reduced. This is a first order effect on pitch. The second order
effects modify it.

I think that either there is more to this, or the manufacturers are
not doing the most basic analysis of the stability margin.

Your latter conjecture is most likely correct. I once interviewed an
electrical engineer who worked for Meggitt. If he is at all typical,
they are simply not capable of doing the most basic analysis of the
stability margin (of course I did not hire him).

The thing which suprised me, on the KFC225, was that the software does
not detect an increasing error. So, if say a roll servo goes dead, it
never realises it.

S-TEC's are the same. A friend of mine had a servo go intermittent,
and finally die. Never got any diagnostics.

I even have a couple of videos of this. The unit
even passes its power-up tests, including servo tests, with a totally
dead servo.

I believe the only failure mode the KFC is designed to detect is a dead
open circuit. Some A/P's won't even do that.

I realise this isn't on the same topic as control
stability but if they don't get this right then .... ?

Your implication is that their design is not up to the standards you've
come to expect from a professional operation. And it's not.

I often wonder just what "certification" really means in this
business.

It means the paper is worth five times as much as the actual product.
It's obvious to you how poorly autopilots are designed because you have
a good understanding of the technology. If you believe that the
situation is different for other aircraft components, I have a bridge I
think you might be interested in purchasing.

Michael