Narrative:

On the westminster 3 arrival into bwi, we had been cleared by ZDC to cross lizio intersection at FL250. We were descending with the center autoplt engaged in the flight level change mode, 290 KTS speed selected, in cloud and in moderate chop. Our descent profile would have us level at FL250 several mi west of lizio. As we passed FL253 in the descent, the autoplt started a smooth, rapid pitch down, and the amber autoplt light on the center panel illuminated. Both flight directors initially commanded a significant pitch down, with the autoplt decreasing pitch towards the command bars. The autoplt was immediately disengaged the command bars then reversed to show a pitch up command as we passed the target altitude. Pitch altitude reached about 10 degrees nose down, and vertical velocity approached -5000 FPM as a manual pull-up was initiated. An estimated 1.5 G pull-up was accomplished to minimize the amount of altitude deviation. Minimum altitude during the pull-up was FL245, and maximum altitude during the recovery was FL252. I told center that we were unable to maintain FL250 due to an autoplt malfunction, and the controller cleared us to continue our descent to cross mumsy intersection at 15000 ft. After I read back the clearance, the controller inquired about our climb, and I explained that we were recovering from an autoplt malfunction. The controller then cleared us to descend and maintain FL240 and switched us to a different sector. The descent to FL240 was hand flown with the flight directors still engaged, and at FL243 both flight directors again commanded a rapid pitch down. Both flight directors were selected off and then back on, and the right autoplt was engaged. All flight director indications and autoplt actions were normal for the remainder of the flight. After the pull-up, I checked with the first flight attendant to determine if anyone had been standing or injured, and was told that everyone had been seated and there were no problems. As a precaution, I had asked the cabin crew to prepare the cabin for landing and remain seated, starting just prior to top of descent and continuing for the remainder of the flight. As far as I know, no faults were found in the flight directors. I wonder how far things would have deteriorated, and how much altitude would have been lost if we hadn't perceived a problem and disconnected the autoplt as soon as we did. I think our reaction time may have been slower than optimum due to fatigue (on duty for 8+ hours, domicile time, time since awakening approximately 18 hours for me). I think we were both task saturated during the pull-up. I remember coaching the first officer to increase the pull-up rate, being concerned about possible traffic at FL240, and at the same time not wanting to hurt anyone in the cabin with an abrupt pull-up. I found it difficult to make a coherent readback of ATC's new clearance while monitoring the instruments, resetting the MCP, and communicating with the first officer, in moderate chop and IMC. If I had that to do over, I would not have acknowledged a new altitude clearance or tried to set it in the MCP until we were stabilized (and calmed down) at the last assigned altitude. 'Standby,' would have been a more useful response from me, and probably would have unloaded both us and the controller. I was very relieved to learn that the flight attendants and passenger had followed my earlier directions to situation down and buckle up for turbulence. From a system standpoint (if I understand it correctly, and if the 'book' describes it correctly) this couldn't have happened, unless all 3 flight control computers simultaneously sent an incorrect command (twice!). Apparently, the left flight director (source: left FCC), right flight director (source: right FCC), and center autoplt (source: center FCC) all commanded a pitch down at the time they should have commanded a pitch up (at altitude capture). The amber autoplt light illuminating indicated ' a degraded operating condition exists in engaged autoplt.' further, 'autoplt engagement requires at least 2 operable FCC's. Commands from the 2 FCC's are compared to prevent 1 FCC from commanding an autoplt hardover...during single autoplt operation, failures affecting the engaged mode are annunciated on the ADI.' we saw no such annunciation. As we were both watching the flight director indications closely during the second leveloff at FL240, we were quite interested to see both flight directors 'jump' to a pitch down command, at about the time we expected them to command a pitch up for altitude capture. One hypothesis is that the FCC's were experiencing interference from a portable electronic device in use aboard the aircraft. Unfortunately, this didn't occur to us at the time in-flight, or during the logbook write-up activities afterwards. It would be interesting to learn if there are any similar autoplt/flight director malfunctions in the ASRS database that have been linked to electronic interference.

Google
 

Original NASA ASRS Text

Title: A B757-200 DEVIATED FROM ASSIGNED ALT DUE TO CTR AUTOPLT MALFUNCTION.

Narrative: ON THE WESTMINSTER 3 ARR INTO BWI, WE HAD BEEN CLRED BY ZDC TO CROSS LIZIO INTXN AT FL250. WE WERE DSNDING WITH THE CTR AUTOPLT ENGAGED IN THE FLT LEVEL CHANGE MODE, 290 KTS SPD SELECTED, IN CLOUD AND IN MODERATE CHOP. OUR DSCNT PROFILE WOULD HAVE US LEVEL AT FL250 SEVERAL MI W OF LIZIO. AS WE PASSED FL253 IN THE DSCNT, THE AUTOPLT STARTED A SMOOTH, RAPID PITCH DOWN, AND THE AMBER AUTOPLT LIGHT ON THE CTR PANEL ILLUMINATED. BOTH FLT DIRECTORS INITIALLY COMMANDED A SIGNIFICANT PITCH DOWN, WITH THE AUTOPLT DECREASING PITCH TOWARDS THE COMMAND BARS. THE AUTOPLT WAS IMMEDIATELY DISENGAGED THE COMMAND BARS THEN REVERSED TO SHOW A PITCH UP COMMAND AS WE PASSED THE TARGET ALT. PITCH ALT REACHED ABOUT 10 DEGS NOSE DOWN, AND VERT VELOCITY APCHED -5000 FPM AS A MANUAL PULL-UP WAS INITIATED. AN ESTIMATED 1.5 G PULL-UP WAS ACCOMPLISHED TO MINIMIZE THE AMOUNT OF ALTDEV. MINIMUM ALT DURING THE PULL-UP WAS FL245, AND MAX ALT DURING THE RECOVERY WAS FL252. I TOLD CTR THAT WE WERE UNABLE TO MAINTAIN FL250 DUE TO AN AUTOPLT MALFUNCTION, AND THE CTLR CLRED US TO CONTINUE OUR DSCNT TO CROSS MUMSY INTXN AT 15000 FT. AFTER I READ BACK THE CLRNC, THE CTLR INQUIRED ABOUT OUR CLB, AND I EXPLAINED THAT WE WERE RECOVERING FROM AN AUTOPLT MALFUNCTION. THE CTLR THEN CLRED US TO DSND AND MAINTAIN FL240 AND SWITCHED US TO A DIFFERENT SECTOR. THE DSCNT TO FL240 WAS HAND FLOWN WITH THE FLT DIRECTORS STILL ENGAGED, AND AT FL243 BOTH FLT DIRECTORS AGAIN COMMANDED A RAPID PITCH DOWN. BOTH FLT DIRECTORS WERE SELECTED OFF AND THEN BACK ON, AND THE R AUTOPLT WAS ENGAGED. ALL FLT DIRECTOR INDICATIONS AND AUTOPLT ACTIONS WERE NORMAL FOR THE REMAINDER OF THE FLT. AFTER THE PULL-UP, I CHKED WITH THE FIRST FLT ATTENDANT TO DETERMINE IF ANYONE HAD BEEN STANDING OR INJURED, AND WAS TOLD THAT EVERYONE HAD BEEN SEATED AND THERE WERE NO PROBS. AS A PRECAUTION, I HAD ASKED THE CABIN CREW TO PREPARE THE CABIN FOR LNDG AND REMAIN SEATED, STARTING JUST PRIOR TO TOP OF DSCNT AND CONTINUING FOR THE REMAINDER OF THE FLT. AS FAR AS I KNOW, NO FAULTS WERE FOUND IN THE FLT DIRECTORS. I WONDER HOW FAR THINGS WOULD HAVE DETERIORATED, AND HOW MUCH ALT WOULD HAVE BEEN LOST IF WE HADN'T PERCEIVED A PROB AND DISCONNECTED THE AUTOPLT AS SOON AS WE DID. I THINK OUR REACTION TIME MAY HAVE BEEN SLOWER THAN OPTIMUM DUE TO FATIGUE (ON DUTY FOR 8+ HRS, DOMICILE TIME, TIME SINCE AWAKENING APPROX 18 HRS FOR ME). I THINK WE WERE BOTH TASK SATURATED DURING THE PULL-UP. I REMEMBER COACHING THE FO TO INCREASE THE PULL-UP RATE, BEING CONCERNED ABOUT POSSIBLE TFC AT FL240, AND AT THE SAME TIME NOT WANTING TO HURT ANYONE IN THE CABIN WITH AN ABRUPT PULL-UP. I FOUND IT DIFFICULT TO MAKE A COHERENT READBACK OF ATC'S NEW CLRNC WHILE MONITORING THE INSTS, RESETTING THE MCP, AND COMMUNICATING WITH THE FO, IN MODERATE CHOP AND IMC. IF I HAD THAT TO DO OVER, I WOULD NOT HAVE ACKNOWLEDGED A NEW ALT CLRNC OR TRIED TO SET IT IN THE MCP UNTIL WE WERE STABILIZED (AND CALMED DOWN) AT THE LAST ASSIGNED ALT. 'STANDBY,' WOULD HAVE BEEN A MORE USEFUL RESPONSE FROM ME, AND PROBABLY WOULD HAVE UNLOADED BOTH US AND THE CTLR. I WAS VERY RELIEVED TO LEARN THAT THE FLT ATTENDANTS AND PAX HAD FOLLOWED MY EARLIER DIRECTIONS TO SIT DOWN AND BUCKLE UP FOR TURB. FROM A SYS STANDPOINT (IF I UNDERSTAND IT CORRECTLY, AND IF THE 'BOOK' DESCRIBES IT CORRECTLY) THIS COULDN'T HAVE HAPPENED, UNLESS ALL 3 FLT CTL COMPUTERS SIMULTANEOUSLY SENT AN INCORRECT COMMAND (TWICE!). APPARENTLY, THE L FLT DIRECTOR (SOURCE: L FCC), R FLT DIRECTOR (SOURCE: R FCC), AND CTR AUTOPLT (SOURCE: CTR FCC) ALL COMMANDED A PITCH DOWN AT THE TIME THEY SHOULD HAVE COMMANDED A PITCH UP (AT ALT CAPTURE). THE AMBER AUTOPLT LIGHT ILLUMINATING INDICATED ' A DEGRADED OPERATING CONDITION EXISTS IN ENGAGED AUTOPLT.' FURTHER, 'AUTOPLT ENGAGEMENT REQUIRES AT LEAST 2 OPERABLE FCC'S. COMMANDS FROM THE 2 FCC'S ARE COMPARED TO PREVENT 1 FCC FROM COMMANDING AN AUTOPLT HARDOVER...DURING SINGLE AUTOPLT OP, FAILURES AFFECTING THE ENGAGED MODE ARE ANNUNCIATED ON THE ADI.' WE SAW NO SUCH ANNUNCIATION. AS WE WERE BOTH WATCHING THE FLT DIRECTOR INDICATIONS CLOSELY DURING THE SECOND LEVELOFF AT FL240, WE WERE QUITE INTERESTED TO SEE BOTH FLT DIRECTORS 'JUMP' TO A PITCH DOWN COMMAND, AT ABOUT THE TIME WE EXPECTED THEM TO COMMAND A PITCH UP FOR ALT CAPTURE. ONE HYPOTHESIS IS THAT THE FCC'S WERE EXPERIENCING INTERFERENCE FROM A PORTABLE ELECTRONIC DEVICE IN USE ABOARD THE ACFT. UNFORTUNATELY, THIS DIDN'T OCCUR TO US AT THE TIME INFLT, OR DURING THE LOGBOOK WRITE-UP ACTIVITIES AFTERWARDS. IT WOULD BE INTERESTING TO LEARN IF THERE ARE ANY SIMILAR AUTOPLT/FLT DIRECTOR MALFUNCTIONS IN THE ASRS DATABASE THAT HAVE BEEN LINKED TO ELECTRONIC INTERFERENCE.

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.