Narrative:

Unexpected disengagement of autopilot in turbulence & IMC. While in cruise at 8;000 ft MSL through moderately turbulent clouds I turned my head to the right & down while refolding L24 chart to see next navaid. I don't recall if my glance back to the instruments was routine scan or in response to physical sensation of displacement. Simultaneously I noticed that the attitude indicator showed a steep bank (more than 45 deg.) and the altimeter's hundreds needle was rapidly moving down. Apparently extensive training in unusual attitude recovery from many years ago took over. I noticed that my eyes went to vsi & airspeed indicators; faster than my conscious thinking; & confirmed attitude & altitude; I was in a diving turn. Fortunately; habits embedded from long ago training continued to prevail. I simultaneously leveled the wings & reduced power. As the wings neared level I then eased the pitch back to recover the dive. By this time conscious thinking was caught up with reflexive behavior and I was sensitive to not over stress the airframe. Just a second after I had established a climb attitude the controller called my north number & stated that I was off assigned altitude. As I was replying to him that I was already recovering on the altitude I also noticed that the dg was at least 40 degrees off heading. As I banked for the heading correction I then noticed that the autopilot was in ready mode with the lighted annunciators for heading and altitude both off. The autopilot disengagement had happened earlier in the day on the first leg of my trip. However; the previous situation was a non-event because I was in smooth air & the inherent stability of the aircraft held heading and altitude fairly close. I probably noticed the first incident more quickly since my distraction had only been the length of time necessary to set the next frequency to the stand by window of the radio. I am particularly aggravated by this situation since in recent months I have spent about $2;000 trying to resolve a different problem with the altitude hold feature. This work included sending the autopilot computer unit back to the manufacturer (s-tec) for trouble shooting and overhaul. Until the recent altitude hold problem and today's disengagement surprises this has been an amazingly stable and completely reliable autopilot. I would have expected overhaul to maintain this level of reliability.

Google
 

Original NASA ASRS Text

Title: BE35 pilot at 8000 FT in IMC experienced an autopilot disengagement that was undetected until the aircraft entered a diving right turn with more than 45 degrees bank angle. As recovery began ATC questioned the reporters altitude.

Narrative: Unexpected disengagement of autopilot in turbulence & IMC. While in cruise at 8;000 FT MSL through moderately turbulent clouds I turned my head to the right & down while refolding L24 chart to see next navaid. I don't recall if my glance back to the instruments was routine scan or in response to physical sensation of displacement. Simultaneously I noticed that the attitude indicator showed a steep bank (more than 45 deg.) and the altimeter's hundreds needle was rapidly moving down. Apparently extensive training in unusual attitude recovery from many years ago took over. I noticed that my eyes went to VSI & airspeed indicators; faster than my conscious thinking; & confirmed attitude & altitude; I was in a diving turn. Fortunately; habits embedded from long ago training continued to prevail. I simultaneously leveled the wings & reduced power. As the wings neared level I then eased the pitch back to recover the dive. By this time conscious thinking was caught up with reflexive behavior and I was sensitive to not over stress the airframe. Just a second after I had established a climb attitude the controller called my N number & stated that I was off assigned altitude. As I was replying to him that I was already recovering on the altitude I also noticed that the DG was at least 40 degrees off heading. As I banked for the heading correction I then noticed that the autopilot was in ready mode with the lighted annunciators for heading and altitude both off. The autopilot disengagement had happened earlier in the day on the first leg of my trip. However; the previous situation was a non-event because I was in smooth air & the inherent stability of the aircraft held heading and altitude fairly close. I probably noticed the first incident more quickly since my distraction had only been the length of time necessary to set the next frequency to the stand by window of the radio. I am particularly aggravated by this situation since in recent months I have spent about $2;000 trying to resolve a different problem with the altitude hold feature. This work included sending the autopilot computer unit back to the manufacturer (S-Tec) for trouble shooting and overhaul. Until the recent altitude hold problem and today's disengagement surprises this has been an amazingly stable and completely reliable autopilot. I would have expected overhaul to maintain this level of reliability.

Data retrieved from NASA's ASRS site as of April 2012 and automatically converted to unabbreviated mixed upper/lowercase text. This report is for informational purposes with no guarantee of accuracy. See NASA's ASRS site for official report.