Ron Natalie wrote:
Yes, that was the cause of the AP disconnect. The 15 yo overpowered
the autopilot causing it to silently disconnect the aileron servos.
Neither the right seat pilot nor the captain standing behind noticed
this. When the airplane started roll subsequently they assumed they
had somehow commanded an autopilot-controlled hold entry. They
then allowed the bank to progress to 50 degrees.

This leads to an incipient problem you can have in a private aircraft
as well. Without sufficient power, the autopilot trying to maintain
altitude can drive the aircraft into a stall. It was finally at the
onset of the prestall buffeting that the copilot started to try to
recover, unfortunately while you can overpower an autopilot easily,
overpowering a 15yo holding the other yoke is not as easy.

Its hard to compare an transport aircraft quality autopilot with an air
mixer. The specific problems brought up in the Russian accident were 1)
The pilots were never trained that overpowering the autopilot would
result in a *partial* disconnect. The roll disconnected but not the
pitch and 2) When such an event happens the autopilot disengage horn
does not sound. Clearly a misstep in design.

Now compare that to the Cessna. If you tell the Cessna to hold altitude
and pull power all the way back to idle it will get pretty slow, but it
will not stall, the nose will drop and the computer is smart enough to
give up altitude to prevent stall (I've done it). There is no partial
disconnect in the Cessna. The autopilot is either engaged or not. If
you press the red button the entire thing goes off line. In the above
accident a non-rated person was banking the aircraft in excess of 60
degrees of roll when the problem happened. If you use more than 60
degrees of roll in your procedure turn, you probably have bigger
problems in your Cessna than the autopilot.

-Robert