DFC90 and Reported Pitch Porpoising

Like similar reports of roll oscillations in aircraft that install the DFC90, there are some reports of "pitch porpoising" in DFC90 equipped aircraft when the autopilot is engaged.  The typical report notes that in cruise flight, sometimes not noticed until about 30-60 minutes into flight, the aircraft will start a gentle, slow period (10-30 sec from peak to peak) oscillation in pitch.   There is usually a resultant 20-80 foot altitude deviation in the process.   Disconnecting and then reconnecting the autopilot often clears the issue for a while or the rest of the duration of the flight.

And, like in the case of the reported roll oscillations, these reports have all been traced to an out-of-spec pitch trim motor.   Repairing (by cleaning the built up carbon deposits on the internal brushes) or replacing the pitch trim motor has solved the reported behavior in all cases that Avidyne is aware of.

As noted in another thread in this forum, the DFC90 (and DFC100) system will actively use both the pitch trim and pitch servos to fly the airplane.  This results in a very responsive autopilot and solves the "deadband" challenges that would faced otherwise.  This has the affect of relying a lot on the pitch trim motor for behind-the-scenes fine movements to precisely track the autopilot loops.   So, in aircraft with "dirty" pitch trim motors, a form of slow motion "AIO" (autopilot induced oscillation) occurs because it slowly ramps up the input voltage required to start the pitch trim motor moving.  The "out-of-spec" pitch trim motor results in higher than spec startup voltages required to get the motor turning.  That means that instead of the expected 1.0 or 1.5 volts required to get the motor turning, it may now need 3 or 4 or 5 volts to get it moving.  The autopilot will slowly ramp up the voltage until it senses the motor beginning to turn.  This lag in motor turning, in turn results in the "AIO" and hence the observed "pitch porpoising".

We do not view this as a "broken" autopilot but instead, the DFC system is less tolerant of sloppy trim motors than predecessor autopilots may have been, hence the much improved performance with "in spec" aircraft.

What are you recomending to "clean" the servos with?