Narrative:

Departing oakland, a normal climb profile was used. After leaving 10000 ft, I accelerated from 250 to 320 KTS. Reaching 320 KTS, I engaged the autoplt. It is important to note that the autoplt on this aircraft has a pitch trim wheel which is used to set vertical speed. When the autoplt is not in use, this pitch trim wheel is supposed to follow the vertical speed of the aircraft, so that when the autoplt is engaged, a pitch change won't result. Anyway, when I engaged the autoplt, a pitch change did result. The aircraft pitched down. I quickly checked the position of the autoplt pitch trim wheel and noticed that it was positioned to command a 1500 FPM descent. I then repositioned the pitch trim wheel manually to about 2000 FPM climb and a positive rate of climb was maintained at all times. From this point, I maintained 320 KTS with the pitch trim wheel until reaching about FL180. I then engaged the airspeed hold feature of the autoplt, and I also engaged the altitude select and heading feature. By this time we had been cleared to FL290 and FL290 was set into the altitude alerter. I and the captain fly this route often. The aircraft differences involved with the aircraft we flew this night created a little extra workload. In and of itself, this wouldn't have been a problem. We had briefed the differences prior to leaving the ramp, but we (or at least I) found that the differences were enough to interrupt some fairly ingrained flow patterns. Also, add to that the fact that the upcoming deviation occurred at xaam in the morning, normal working hours for us, but still your brain doesn't exactly leap into gear at that hour. Now, back to what happened. As we were approaching the palmdale VOR, we were given a 'descend at pilot's discretion' clearance to FL240. We were given a frequency change and a new clearance for the ziggy 3 arrival, cross majek at 14000 ft, descend to 13000 ft. The captain noticed on his altimeter that we were at FL298 and called it to my attention. I checked my altimeter and it said FL299. I also looked at the altitude alerter which had our 'pilot's discretion' altitude of FL24O dialed in. I immediately disconnected the autoplt and began a descent to our assigned cruise altitude of FL290. We then dialed in 14000 ft which reflected our new clearance to cross majek at that altitude. With the captain's concurrence, we conducted a test. Sure enough, when we leveled off at 14000 ft, and later at 13000 ft, the altitude select captured the altitude. But, the pitch/trim wheel on the autoplt stayed in a pitch down condition. Furthermore, when we selected a new altitude on the altitude alerter, the altitude capture green light changed from green to amber, and the stuck pitch trim wheel then caused the aircraft to descend. The opposite would have happened after leveling off from a climb. The altitude select would have leveled the aircraft at FL290, but the malfunctioning pitch trim wheel would have remained in a climb condition. When the new lower altitude of FL240 was selected into the altitude alerter, the altitude select light would have gone from green to amber, and the stuck trim wheel, commanding a climb, would have caused the aircraft to climb. We did get a green altitude select light on the autoplt console, however. And we monitored the aircraft's altitude and noticed that it maintained FL290 normally. In fact, we later found that the altitude hold light will not come on unless the autoplt thumb wheel is detented in the neutral 0 rate of climb position, and because it wasn't, it didn't. Quite simply, the altitude select was maintaining our altitude, not the altitude hold. To sum up, an altitude deviation occurred because of several events: 1) mechanical failure of the automatic vertical speed following system in the autoplt thumb wheel. 2) interrupted flow patterns caused by a nonstandard cockpit confign. 3) several unrelated electrical glitches, like repeated false autoplt disconnect warnings. 4) the time of day, around xyam in the morning, and the normal degradation in performance associated with it.

Related reports:

Google
 

Original NASA ASRS Text

Title: ALTDEV ALT EXCURSION.

Narrative: DEPARTING OAKLAND, A NORMAL CLB PROFILE WAS USED. AFTER LEAVING 10000 FT, I ACCELERATED FROM 250 TO 320 KTS. REACHING 320 KTS, I ENGAGED THE AUTOPLT. IT IS IMPORTANT TO NOTE THAT THE AUTOPLT ON THIS ACFT HAS A PITCH TRIM WHEEL WHICH IS USED TO SET VERT SPD. WHEN THE AUTOPLT IS NOT IN USE, THIS PITCH TRIM WHEEL IS SUPPOSED TO FOLLOW THE VERT SPD OF THE ACFT, SO THAT WHEN THE AUTOPLT IS ENGAGED, A PITCH CHANGE WON'T RESULT. ANYWAY, WHEN I ENGAGED THE AUTOPLT, A PITCH CHANGE DID RESULT. THE ACFT PITCHED DOWN. I QUICKLY CHKED THE POS OF THE AUTOPLT PITCH TRIM WHEEL AND NOTICED THAT IT WAS POSITIONED TO COMMAND A 1500 FPM DSCNT. I THEN REPOSITIONED THE PITCH TRIM WHEEL MANUALLY TO ABOUT 2000 FPM CLB AND A POSITIVE RATE OF CLB WAS MAINTAINED AT ALL TIMES. FROM THIS POINT, I MAINTAINED 320 KTS WITH THE PITCH TRIM WHEEL UNTIL REACHING ABOUT FL180. I THEN ENGAGED THE AIRSPD HOLD FEATURE OF THE AUTOPLT, AND I ALSO ENGAGED THE ALT SELECT AND HDG FEATURE. BY THIS TIME WE HAD BEEN CLRED TO FL290 AND FL290 WAS SET INTO THE ALT ALERTER. I AND THE CAPT FLY THIS RTE OFTEN. THE ACFT DIFFERENCES INVOLVED WITH THE ACFT WE FLEW THIS NIGHT CREATED A LITTLE EXTRA WORKLOAD. IN AND OF ITSELF, THIS WOULDN'T HAVE BEEN A PROB. WE HAD BRIEFED THE DIFFERENCES PRIOR TO LEAVING THE RAMP, BUT WE (OR AT LEAST I) FOUND THAT THE DIFFERENCES WERE ENOUGH TO INTERRUPT SOME FAIRLY INGRAINED FLOW PATTERNS. ALSO, ADD TO THAT THE FACT THAT THE UPCOMING DEV OCCURRED AT XAAM IN THE MORNING, NORMAL WORKING HRS FOR US, BUT STILL YOUR BRAIN DOESN'T EXACTLY LEAP INTO GEAR AT THAT HR. NOW, BACK TO WHAT HAPPENED. AS WE WERE APCHING THE PALMDALE VOR, WE WERE GIVEN A 'DSND AT PLT'S DISCRETION' CLRNC TO FL240. WE WERE GIVEN A FREQ CHANGE AND A NEW CLRNC FOR THE ZIGGY 3 ARR, CROSS MAJEK AT 14000 FT, DSND TO 13000 FT. THE CAPT NOTICED ON HIS ALTIMETER THAT WE WERE AT FL298 AND CALLED IT TO MY ATTN. I CHKED MY ALTIMETER AND IT SAID FL299. I ALSO LOOKED AT THE ALT ALERTER WHICH HAD OUR 'PLT'S DISCRETION' ALT OF FL24O DIALED IN. I IMMEDIATELY DISCONNECTED THE AUTOPLT AND BEGAN A DSCNT TO OUR ASSIGNED CRUISE ALT OF FL290. WE THEN DIALED IN 14000 FT WHICH REFLECTED OUR NEW CLRNC TO CROSS MAJEK AT THAT ALT. WITH THE CAPT'S CONCURRENCE, WE CONDUCTED A TEST. SURE ENOUGH, WHEN WE LEVELED OFF AT 14000 FT, AND LATER AT 13000 FT, THE ALT SELECT CAPTURED THE ALT. BUT, THE PITCH/TRIM WHEEL ON THE AUTOPLT STAYED IN A PITCH DOWN CONDITION. FURTHERMORE, WHEN WE SELECTED A NEW ALT ON THE ALT ALERTER, THE ALT CAPTURE GREEN LIGHT CHANGED FROM GREEN TO AMBER, AND THE STUCK PITCH TRIM WHEEL THEN CAUSED THE ACFT TO DSND. THE OPPOSITE WOULD HAVE HAPPENED AFTER LEVELING OFF FROM A CLB. THE ALT SELECT WOULD HAVE LEVELED THE ACFT AT FL290, BUT THE MALFUNCTIONING PITCH TRIM WHEEL WOULD HAVE REMAINED IN A CLB CONDITION. WHEN THE NEW LOWER ALT OF FL240 WAS SELECTED INTO THE ALT ALERTER, THE ALT SELECT LIGHT WOULD HAVE GONE FROM GREEN TO AMBER, AND THE STUCK TRIM WHEEL, COMMANDING A CLB, WOULD HAVE CAUSED THE ACFT TO CLB. WE DID GET A GREEN ALT SELECT LIGHT ON THE AUTOPLT CONSOLE, HOWEVER. AND WE MONITORED THE ACFT'S ALT AND NOTICED THAT IT MAINTAINED FL290 NORMALLY. IN FACT, WE LATER FOUND THAT THE ALT HOLD LIGHT WILL NOT COME ON UNLESS THE AUTOPLT THUMB WHEEL IS DETENTED IN THE NEUTRAL 0 RATE OF CLB POS, AND BECAUSE IT WASN'T, IT DIDN'T. QUITE SIMPLY, THE ALT SELECT WAS MAINTAINING OUR ALT, NOT THE ALT HOLD. TO SUM UP, AN ALTDEV OCCURRED BECAUSE OF SEVERAL EVENTS: 1) MECHANICAL FAILURE OF THE AUTOMATIC VERT SPD FOLLOWING SYS IN THE AUTOPLT THUMB WHEEL. 2) INTERRUPTED FLOW PATTERNS CAUSED BY A NONSTANDARD COCKPIT CONFIGN. 3) SEVERAL UNRELATED ELECTRICAL GLITCHES, LIKE REPEATED FALSE AUTOPLT DISCONNECT WARNINGS. 4) THE TIME OF DAY, AROUND XYAM IN THE MORNING, AND THE NORMAL DEGRADATION IN PERFORMANCE ASSOCIATED WITH IT.

Data retrieved from NASA's ASRS site as of July 2007 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.