Narrative:

The aircraft involved was an advanced cockpit aircraft, fully automated, and certified for CAT III approachs. The widely scattered thunderstorms were no factor. We had the FMS programmed for the scurry arrival into dfw with a planned landing on R17L, autoplt on, captain's leg (mine). FMS and autoplt performed beautifully, hitting all fixes at assigned altitudes and airspds. When arriving at the refil intersection, the aircraft started a turn to the right to the appropriate heading of 350 degrees. The aircraft had been level at 11000' since seago intersection with 11000' in the altitude select widow, and was at the appropriate speed of 210 KTS at refil. After confirming that the aircraft was indeed turning to the correct heading at the correct place, I glanced down at the commercial plate for 17L to verify the initial approach altitude and G/south intercept altitude. The timing was perfect; during the second it took for me to glance at the approach plate, the FMS sent a spurious signal to the autoplt to descend, even though 11000' was in the altitude window. I heard the altitude warning go off, looked up and saw the altimeter unwinding and a descent rate approaching 1500 FPM. I immediately turned off the autoplt, stopped the descent, and began a climb back to 11000'. The altimeter was reading about 10600' when the descent rate was stopped. The copilot and I discussed the problem at length, after landing, and could not determine why the aft ever left 11000' with no command input from the MCP, or why the FMS signaled a descent. After the excursion, all functions of the FMS and MCP appeared to work normally. I have no idea what caused the problem. As long as we have computers, we will continue to have anomalies making it mandatory for crews to remain alert and vigilant.

Related reports:

Google
 

Original NASA ASRS Text

Title: ADVANCED COCKPIT ACR LGT BEGINS UNANTICIPATED DESCENT AT REFIL ON THE SCURRY 9 STAR TO DFW.

Narrative: THE ACFT INVOLVED WAS AN ADVANCED COCKPIT ACFT, FULLY AUTOMATED, AND CERTIFIED FOR CAT III APCHS. THE WIDELY SCATTERED TSTMS WERE NO FACTOR. WE HAD THE FMS PROGRAMMED FOR THE SCURRY ARR INTO DFW WITH A PLANNED LNDG ON R17L, AUTOPLT ON, CAPT'S LEG (MINE). FMS AND AUTOPLT PERFORMED BEAUTIFULLY, HITTING ALL FIXES AT ASSIGNED ALTS AND AIRSPDS. WHEN ARRIVING AT THE REFIL INTXN, THE ACFT STARTED A TURN TO THE RIGHT TO THE APPROPRIATE HDG OF 350 DEGS. THE ACFT HAD BEEN LEVEL AT 11000' SINCE SEAGO INTXN WITH 11000' IN THE ALT SELECT WIDOW, AND WAS AT THE APPROPRIATE SPD OF 210 KTS AT REFIL. AFTER CONFIRMING THAT THE ACFT WAS INDEED TURNING TO THE CORRECT HDG AT THE CORRECT PLACE, I GLANCED DOWN AT THE COMMERCIAL PLATE FOR 17L TO VERIFY THE INITIAL APCH ALT AND G/S INTERCEPT ALT. THE TIMING WAS PERFECT; DURING THE SECOND IT TOOK FOR ME TO GLANCE AT THE APCH PLATE, THE FMS SENT A SPURIOUS SIGNAL TO THE AUTOPLT TO DSND, EVEN THOUGH 11000' WAS IN THE ALT WINDOW. I HEARD THE ALT WARNING GO OFF, LOOKED UP AND SAW THE ALTIMETER UNWINDING AND A DSCNT RATE APCHING 1500 FPM. I IMMEDIATELY TURNED OFF THE AUTOPLT, STOPPED THE DSCNT, AND BEGAN A CLB BACK TO 11000'. THE ALTIMETER WAS READING ABOUT 10600' WHEN THE DSCNT RATE WAS STOPPED. THE COPLT AND I DISCUSSED THE PROB AT LENGTH, AFTER LNDG, AND COULD NOT DETERMINE WHY THE AFT EVER LEFT 11000' WITH NO COMMAND INPUT FROM THE MCP, OR WHY THE FMS SIGNALED A DSCNT. AFTER THE EXCURSION, ALL FUNCTIONS OF THE FMS AND MCP APPEARED TO WORK NORMALLY. I HAVE NO IDEA WHAT CAUSED THE PROB. AS LONG AS WE HAVE COMPUTERS, WE WILL CONTINUE TO HAVE ANOMALIES MAKING IT MANDATORY FOR CREWS TO REMAIN ALERT AND VIGILANT.

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.