Roll goes past roll limit

[paulrkuntz]

I have observed what appears to be a bug in the autopilot software. (I'm running the 6.0 version). When the autopilot initiates a turn to a new heading, if the turn is to the left, it tends to roll past the roll limit and keep rolling. I have left it alone up to 70 degrees of roll angle, then intervened when it showed no sign of reversing the roll. this is very repeatable.

My roll limit is set to 20 degrees. When rolling right, it seems to behave as expected. I've tried adjusting roll sensitivity, but the behavior continues.

Last weekend I noticed that I am nearly always getting a roll slip alert when rolling to the left, but not to the right. I don't see a mechanical explanation for this -- the servo and roll torque tube arms are both at 90 degrees to the linking arm at neutral and there is no obvious friction in the system. There may be an aerodynamic difference between left and right roll, and I have noted that the aileron forces are relatively high and cruise speeds, although easily controllable when hand-flying. The roll force is currently set to 100% and it still happens.

So I tried applying some manual left aileron when turning left to assist the roll servo, and when I do that, everything works as expected. The plane rolls smoothly up to the roll limit, maintains that bank angle through the turn, then smoothly rolls level to intercept the new heading and course line.

So, here's what I am calling the software bug:
The roll servo is still applying 100% force to the ailerons, even when it's slipping, but it appears that when the autopilot senses servo slip, it stops monitoring the bank angle when reaching the bank angle limit and continues to apply roll input. At least, that's how it's behaving in my airplane.

Suggestions?

 

[led]

Paul,

This can sometimes happen if the airplane has a heavy wing (for example, due to fuel burn.)  The AP can no longer hold up the heavy wing, slips, has insufficient torque to recover, and continues slipping.  Meanwhile the heavy wing continues the roll.

If 100% torque is insufficient for your installation, you may need to use a larger servo size.   If you are able to feel a significant difference in the force needed to override the servo on the ground in one direction or the other, that may indicate trouble with the servo.  Contact our support folks at support@dynonavionics.com and they can help you sort that out.

 

[paulrkuntz]

You explain possible reasons for the servo to slip, but the slipping is not what I am concerned about.

Even if the servo is slipping at max torque, it's still applying plenty of aileron force to roll the aircraft. The problem is that the autopilot is rolling the aircraft past the roll limit. The symptoms indicate that the autopilot software ignores the roll limit if the servo is slipping. If that is what's happening, then I think this is an oversight in the software design. It's that suspected software error that I'm trying to highlight here.

 

[dynonsupport]

Paul,
That's an interesting experience that we're not familiar with. Any chance you have a Dynon Diagnostic log of that flight? If not, if you can get it to happen again a log would be very interesting to us.

 

[paulrkuntz]

Yes, I do have data logs from the flight (I'll have to download them, but they're still there). Is the Dynon Diagnostic log the same thing as the user data log?

 

[dynonsupport]

The diagnostic log is not the same and is only one hour long, so if you have flown more than that since the event, it's gone.

It's in the same menu as export user logs but named differently.

 

[paulrkuntz]

I did some more trouble shooting last weekend and found that I actually had some aileron binding between the aileron and the wing. It wasn't immediately apparent when hand flying, and it didn't normally appear during preflight, but I did find some slight binding during a preflight on Sunday. So I think that could explain the behavior. I opened up the clearance between the end of the aileron and the wing where the binding was occurring, and I think that cleared up the strange behavior. I am still getting some aileron slippage in flight, however, and concluded I need to upgrade the servo for more torque. I've already arranged that with Dynon Customer Service, so I'll see if the larger servo gets rid of the aileron slip. I believe that will get me to normal autopilot operation.