Narrative:

We were cruising at FL240 on the fossl 5 arrival to phx. We were running late so I had engaged rta mode to make up time and we were doing ok as we were projecting an arrival within 15 mins of schedule. Shortly after passing sacho intersection the first officer went off frequency to get ATIS and call company. Shortly after that ATC gave us direct tonto. I entered this into the computer and the aircraft turned towards tonto. Suddenly, to my surprise, I observed the autothrottles decreasing power and the speed bug come back approximately 20+ KTS. I checked the rta page and noticed that an extra 2 mins had opened up in the rta window and since our previous ETA was still entered, the computer was slowing down to compensate for the new routing. Wanting to take full advantage of the shortcut I deselected VNAV and increased speed via the mode control panel. I observed the autothrottles increase power and turned my attention to reprogramming the computer. In the middle of my programming efforts ATC cleared us to cross tonto at 12000 ft and 250 KTS. I reset the altitude on the mpc and went back to programming the computer which didn't like the parameter that I was entering so I had to choose another. That done I noticed that I had a flashing yellow autoplt warning light and that for some reason we were in control wheel steering pitch. By this time I was a bit fed up with the computer so I elected to start my descent in vertical speed mode even though we were well outside our top of descent point. I engaged vertical speed and dialed in a 1000 FPM descent, whereupon the autoplt defaulted to level change. Thinking this was the normal 'glitch' that some of our aircraft have, in that vertical speed will default to level change if you spin the vertical speed wheel too fast, I reengaged vertical speed and observed 1000 ft in the window. However, the autothrottles did not come back very far and the aircraft did not seem to be responding to the commanded descent. I checked the altimeter and noticed that we were 300 ft high and climbing. By this time I am getting confused. During the resulting rechk of my instruments I notice to my horror that the altimeter is reading 25300 ft not 24300 ft as I had originally perceived. We had gained over 1000 ft in less than 1 min. In retrospect the 1000 ft that I saw in the vertical speed window after the second engagement must have been +1000 ft not -1000 ft. I immediately retarded the thrust levers manually and pitched over manually just as the first officer came back after getting ATIS, etc. During this entire incident ATC didn't say a word to us, there were no TA's or RA's from the TCASII and we continued the remainder of the arrival normally. I can only surmise that when I deselected VNAV that the autoplt for some reason defaulted to control wheel steering pitch and not altitude hold as it should and with the addition of power to increase speed the aircraft started to climb. With ATC giving us the crossing restr immediately thereafter and my resetting the altitude alerter, this removed altitude protection as the alerter would then only warn of deviations from 12000 ft. This was an interesting combination of events. One would think that after 34 yrs of flying, of which almost 14 yrs are flying jet airliners, that I wouldn't be surprised at how fast a situation can go to hell in a hand basket. My only conclusion is that the time interval for disaster is getting shorter. Thinking back I now remember a similar 'happening' about a month ago when an autoplt, for no apparent reason, defaulted to control wheel steering pitch. At the time I was not distracted by any duties and caught the 'glitch' immediately. As usual it happened in the approach phase and we had no time to troubleshoot the problem to discern whether it was human or software error. Whenever things like this happen, my initial reaction is to assume that I did something wrong in programming or mode selection. I now believe that any error on my part has been over use of the computer to begin with. In the future I will only use it for rudimentary LNAV, VNAV and as a backup if feasible. I will no longer engage in 'advanced FMC' as my current first officer refers toit. The computer requires entirely too much attention to be paid to it considering the supersaturated ATC environment that we operate in. Please note that this incident occurred 120 mi from destination, in cruise and only 1 amended clearance. Not to blame ATC but I was surprised that some alarm did not alert the controller to this. My guess is that he too was distracted by some other duty that required his attention to be diverted momentarily from his screen.

Related reports:

Google
 

Original NASA ASRS Text

Title: B737 CREW HAD ALTDEV IN ZAB CLASS A.

Narrative: WE WERE CRUISING AT FL240 ON THE FOSSL 5 ARR TO PHX. WE WERE RUNNING LATE SO I HAD ENGAGED RTA MODE TO MAKE UP TIME AND WE WERE DOING OK AS WE WERE PROJECTING AN ARR WITHIN 15 MINS OF SCHEDULE. SHORTLY AFTER PASSING SACHO INTXN THE FO WENT OFF FREQ TO GET ATIS AND CALL COMPANY. SHORTLY AFTER THAT ATC GAVE US DIRECT TONTO. I ENTERED THIS INTO THE COMPUTER AND THE ACFT TURNED TOWARDS TONTO. SUDDENLY, TO MY SURPRISE, I OBSERVED THE AUTOTHROTTLES DECREASING PWR AND THE SPD BUG COME BACK APPROX 20+ KTS. I CHKED THE RTA PAGE AND NOTICED THAT AN EXTRA 2 MINS HAD OPENED UP IN THE RTA WINDOW AND SINCE OUR PREVIOUS ETA WAS STILL ENTERED, THE COMPUTER WAS SLOWING DOWN TO COMPENSATE FOR THE NEW ROUTING. WANTING TO TAKE FULL ADVANTAGE OF THE SHORTCUT I DESELECTED VNAV AND INCREASED SPD VIA THE MODE CTL PANEL. I OBSERVED THE AUTOTHROTTLES INCREASE PWR AND TURNED MY ATTN TO REPROGRAMMING THE COMPUTER. IN THE MIDDLE OF MY PROGRAMMING EFFORTS ATC CLRED US TO CROSS TONTO AT 12000 FT AND 250 KTS. I RESET THE ALT ON THE MPC AND WENT BACK TO PROGRAMMING THE COMPUTER WHICH DIDN'T LIKE THE PARAMETER THAT I WAS ENTERING SO I HAD TO CHOOSE ANOTHER. THAT DONE I NOTICED THAT I HAD A FLASHING YELLOW AUTOPLT WARNING LIGHT AND THAT FOR SOME REASON WE WERE IN CTL WHEEL STEERING PITCH. BY THIS TIME I WAS A BIT FED UP WITH THE COMPUTER SO I ELECTED TO START MY DSCNT IN VERT SPD MODE EVEN THOUGH WE WERE WELL OUTSIDE OUR TOP OF DSCNT POINT. I ENGAGED VERT SPD AND DIALED IN A 1000 FPM DSCNT, WHEREUPON THE AUTOPLT DEFAULTED TO LEVEL CHANGE. THINKING THIS WAS THE NORMAL 'GLITCH' THAT SOME OF OUR ACFT HAVE, IN THAT VERT SPD WILL DEFAULT TO LEVEL CHANGE IF YOU SPIN THE VERT SPD WHEEL TOO FAST, I REENGAGED VERT SPD AND OBSERVED 1000 FT IN THE WINDOW. HOWEVER, THE AUTOTHROTTLES DID NOT COME BACK VERY FAR AND THE ACFT DID NOT SEEM TO BE RESPONDING TO THE COMMANDED DSCNT. I CHKED THE ALTIMETER AND NOTICED THAT WE WERE 300 FT HIGH AND CLBING. BY THIS TIME I AM GETTING CONFUSED. DURING THE RESULTING RECHK OF MY INSTS I NOTICE TO MY HORROR THAT THE ALTIMETER IS READING 25300 FT NOT 24300 FT AS I HAD ORIGINALLY PERCEIVED. WE HAD GAINED OVER 1000 FT IN LESS THAN 1 MIN. IN RETROSPECT THE 1000 FT THAT I SAW IN THE VERT SPD WINDOW AFTER THE SECOND ENGAGEMENT MUST HAVE BEEN +1000 FT NOT -1000 FT. I IMMEDIATELY RETARDED THE THRUST LEVERS MANUALLY AND PITCHED OVER MANUALLY JUST AS THE FO CAME BACK AFTER GETTING ATIS, ETC. DURING THIS ENTIRE INCIDENT ATC DIDN'T SAY A WORD TO US, THERE WERE NO TA'S OR RA'S FROM THE TCASII AND WE CONTINUED THE REMAINDER OF THE ARR NORMALLY. I CAN ONLY SURMISE THAT WHEN I DESELECTED VNAV THAT THE AUTOPLT FOR SOME REASON DEFAULTED TO CTL WHEEL STEERING PITCH AND NOT ALT HOLD AS IT SHOULD AND WITH THE ADDITION OF PWR TO INCREASE SPD THE ACFT STARTED TO CLB. WITH ATC GIVING US THE XING RESTR IMMEDIATELY THEREAFTER AND MY RESETTING THE ALT ALERTER, THIS REMOVED ALT PROTECTION AS THE ALERTER WOULD THEN ONLY WARN OF DEVS FROM 12000 FT. THIS WAS AN INTERESTING COMBINATION OF EVENTS. ONE WOULD THINK THAT AFTER 34 YRS OF FLYING, OF WHICH ALMOST 14 YRS ARE FLYING JET AIRLINERS, THAT I WOULDN'T BE SURPRISED AT HOW FAST A SIT CAN GO TO HELL IN A HAND BASKET. MY ONLY CONCLUSION IS THAT THE TIME INTERVAL FOR DISASTER IS GETTING SHORTER. THINKING BACK I NOW REMEMBER A SIMILAR 'HAPPENING' ABOUT A MONTH AGO WHEN AN AUTOPLT, FOR NO APPARENT REASON, DEFAULTED TO CTL WHEEL STEERING PITCH. AT THE TIME I WAS NOT DISTRACTED BY ANY DUTIES AND CAUGHT THE 'GLITCH' IMMEDIATELY. AS USUAL IT HAPPENED IN THE APCH PHASE AND WE HAD NO TIME TO TROUBLESHOOT THE PROB TO DISCERN WHETHER IT WAS HUMAN OR SOFTWARE ERROR. WHENEVER THINGS LIKE THIS HAPPEN, MY INITIAL REACTION IS TO ASSUME THAT I DID SOMETHING WRONG IN PROGRAMMING OR MODE SELECTION. I NOW BELIEVE THAT ANY ERROR ON MY PART HAS BEEN OVER USE OF THE COMPUTER TO BEGIN WITH. IN THE FUTURE I WILL ONLY USE IT FOR RUDIMENTARY LNAV, VNAV AND AS A BACKUP IF FEASIBLE. I WILL NO LONGER ENGAGE IN 'ADVANCED FMC' AS MY CURRENT FO REFERS TOIT. THE COMPUTER REQUIRES ENTIRELY TOO MUCH ATTN TO BE PAID TO IT CONSIDERING THE SUPERSATURATED ATC ENVIRONMENT THAT WE OPERATE IN. PLEASE NOTE THAT THIS INCIDENT OCCURRED 120 MI FROM DEST, IN CRUISE AND ONLY 1 AMENDED CLRNC. NOT TO BLAME ATC BUT I WAS SURPRISED THAT SOME ALARM DID NOT ALERT THE CTLR TO THIS. MY GUESS IS THAT HE TOO WAS DISTRACTED BY SOME OTHER DUTY THAT REQUIRED HIS ATTN TO BE DIVERTED MOMENTARILY FROM HIS SCREEN.

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.