Narrative:

The first officer was the PF and the captain was the PNF for this leg. The captain was recently upgraded and still on 'high minimums.' the first officer had approximately 250 hours in the aircraft. The aircraft was in a descent passing through approximately FL210. I believe at the time of the incident the descent clearance was to cross brave intersection at 10000 ft. The aircraft was 'in range' for landing at milwaukee (mke). The PNF was just entering a higher workload phase of the flight due to getting ATIS, the in range call to company, and setting up for an instrument approach into mke. In addition, the automatic pressurization control was MEL'ed which resulted in the PNF having to manually manage the cabin altitude during the descent. As the PNF was beginning to take care of the in range tasks, the right dc generator failed. This resulted in the PNF losing the flight director, autoplt, and also the FMS as his primary navigation source. The PNF called the failure/cancelled the master caution. Per procedures, the right dc generator was reset and began producing power again. The PNF indicated that he would bring up the flight director for the PF who was now hand flying 'raw data' and asked him if he wanted the vertical mode restored which was in use previously (which was vflch). The PF gave a positive response that he did want vflch restored. Note: flch stands for flight level change which is used when changing altitudes at a constant airspeed. This mode adjusts pitch to seek the programmed airspeed in a descent/climb. The vflch is a 'VNAV' mode that uses the vertical navigation programmed into the FMS (in this case the airspeed profile), while normal flch uses the airspeed that is 'bugged' manually. The PNF armed the 'navigation' mode for lateral navigation which will use the navigation source selected on the side that it is coupled to (in this case the right) and then cycled through flch and VNAV expecting to see LNAV (indicating lateral navigation off of the FMS) and vflch to arm/engage. However, the lateral mode armed but did not engage and the vertical mode stayed in regular flch. This caused a moment of confusion on the part of the PNF who was sure the proper modes had been selected. While thinking about why the desired modes had not engaged, the PNF realized that the aircraft had stopped descending and was starting to pitch up and indicate a climb. He realized that autoplt had already been engaged by the PF and that the PF had left the manual speed selector bug at a much lower airspeed then what he was previously flying from the FMS. The aircraft had started to pitch up to seek this lower airspeed which not only killed the descent but was initiating a climb. At this point, the PNF called the deviation and the PF switched to vertical speed mode and set about a 1500 FPM descent which again initiated a descent as assigned by ATC. The PNF then looked over at the primary flight display (pfd) of the PF and realized that although the PF had coupled the flight guidance control computer back to the right side, the PF was still in 'green needles' which meant that he had not reselected the FMS as his primary navigation source after the dc generator failure. This meant that both LNAV and VNAV modes were not available. The PNF announced that the PF was still in 'green needles,' selected the FMS for the PF and rearmed the desired lateral and vertical modes which now engaged properly. ATC did not comment or inquire about the momentary arrest of the descent and climb and no other traffic was in the area. Contributing factors: the PF had no experience with this type of failure in flight and how it would effect the flight guidance he was receiving on the right pfd. It was a mistake to engage the autoplt as soon as the flight director bars came up before verifying that desired flight attitude/path would be maintained. Furthermore, it should have been announced to the PNF that the autoplt had been engaged. Also, the failure to reselect the FMS as the primary navigation source resulted in the requested lateral mode not being available. The workload of the PNF was higher then usual due to the fact that the cabin altitude had to constantly be managed/monitored due to the automatic mode being MEL'ed. In the immediate aftermath of thedc generator failure, he was too focused on this and the desire to get the APU started up and the APU dc generator on line so that dc power redundancy would be available in case the right dc generator failed again while in the terminal environment for an instrument approach at night (which it did). The PNF should have been more focused on the inexperience of the PF in this type of a situation. Although it was noticed that the flight guidance computer was re-coupled to the first officer's side, it was a crucial mistake not to verify that the first officer was also back on the FMS before trying to engage a VNAV mode. The captain had noticed earlier that the first officer was not in the habit of setting the manual speed bug to match what he was flying from the FMS. The captain considered this 'bad form' but not against company policy and had decided not to 'badger' his first officer over a 'style issue.' however, had the speed bug been set consistently with the descent profile, the flight path deviation would not have occurred in-spite of the VNAV mode (vflch) not engaging and the autoplt being engaged before verifying it would maintain the desired flight path.

Google
 

Original NASA ASRS Text

Title: A DO328 PIC'S RPT ON THE INTRICACIES OF THE FMS IN DSCNT MODE WHEN A PRIMARY ELECTRICAL SUPPLY PROB INTERRUPTS THE DSCNT PROC ON FREQ WITH ZAU.

Narrative: THE FO WAS THE PF AND THE CAPT WAS THE PNF FOR THIS LEG. THE CAPT WAS RECENTLY UPGRADED AND STILL ON 'HIGH MINIMUMS.' THE FO HAD APPROX 250 HRS IN THE ACFT. THE ACFT WAS IN A DSCNT PASSING THROUGH APPROX FL210. I BELIEVE AT THE TIME OF THE INCIDENT THE DSCNT CLRNC WAS TO CROSS BRAVE INTXN AT 10000 FT. THE ACFT WAS 'IN RANGE' FOR LNDG AT MILWAUKEE (MKE). THE PNF WAS JUST ENTERING A HIGHER WORKLOAD PHASE OF THE FLT DUE TO GETTING ATIS, THE IN RANGE CALL TO COMPANY, AND SETTING UP FOR AN INST APCH INTO MKE. IN ADDITION, THE AUTOMATIC PRESSURIZATION CTL WAS MEL'ED WHICH RESULTED IN THE PNF HAVING TO MANUALLY MANAGE THE CABIN ALT DURING THE DSCNT. AS THE PNF WAS BEGINNING TO TAKE CARE OF THE IN RANGE TASKS, THE R DC GENERATOR FAILED. THIS RESULTED IN THE PNF LOSING THE FLT DIRECTOR, AUTOPLT, AND ALSO THE FMS AS HIS PRIMARY NAV SOURCE. THE PNF CALLED THE FAILURE/CANCELLED THE MASTER CAUTION. PER PROCS, THE R DC GENERATOR WAS RESET AND BEGAN PRODUCING PWR AGAIN. THE PNF INDICATED THAT HE WOULD BRING UP THE FLT DIRECTOR FOR THE PF WHO WAS NOW HAND FLYING 'RAW DATA' AND ASKED HIM IF HE WANTED THE VERT MODE RESTORED WHICH WAS IN USE PREVIOUSLY (WHICH WAS VFLCH). THE PF GAVE A POSITIVE RESPONSE THAT HE DID WANT VFLCH RESTORED. NOTE: FLCH STANDS FOR FLT LEVEL CHANGE WHICH IS USED WHEN CHANGING ALTS AT A CONSTANT AIRSPD. THIS MODE ADJUSTS PITCH TO SEEK THE PROGRAMMED AIRSPD IN A DSCNT/CLB. THE VFLCH IS A 'VNAV' MODE THAT USES THE VERT NAV PROGRAMMED INTO THE FMS (IN THIS CASE THE AIRSPD PROFILE), WHILE NORMAL FLCH USES THE AIRSPD THAT IS 'BUGGED' MANUALLY. THE PNF ARMED THE 'NAV' MODE FOR LATERAL NAV WHICH WILL USE THE NAV SOURCE SELECTED ON THE SIDE THAT IT IS COUPLED TO (IN THIS CASE THE R) AND THEN CYCLED THROUGH FLCH AND VNAV EXPECTING TO SEE LNAV (INDICATING LATERAL NAV OFF OF THE FMS) AND VFLCH TO ARM/ENGAGE. HOWEVER, THE LATERAL MODE ARMED BUT DID NOT ENGAGE AND THE VERT MODE STAYED IN REGULAR FLCH. THIS CAUSED A MOMENT OF CONFUSION ON THE PART OF THE PNF WHO WAS SURE THE PROPER MODES HAD BEEN SELECTED. WHILE THINKING ABOUT WHY THE DESIRED MODES HAD NOT ENGAGED, THE PNF REALIZED THAT THE ACFT HAD STOPPED DSNDING AND WAS STARTING TO PITCH UP AND INDICATE A CLB. HE REALIZED THAT AUTOPLT HAD ALREADY BEEN ENGAGED BY THE PF AND THAT THE PF HAD LEFT THE MANUAL SPD SELECTOR BUG AT A MUCH LOWER AIRSPD THEN WHAT HE WAS PREVIOUSLY FLYING FROM THE FMS. THE ACFT HAD STARTED TO PITCH UP TO SEEK THIS LOWER AIRSPD WHICH NOT ONLY KILLED THE DSCNT BUT WAS INITIATING A CLB. AT THIS POINT, THE PNF CALLED THE DEV AND THE PF SWITCHED TO VERT SPD MODE AND SET ABOUT A 1500 FPM DSCNT WHICH AGAIN INITIATED A DSCNT AS ASSIGNED BY ATC. THE PNF THEN LOOKED OVER AT THE PRIMARY FLT DISPLAY (PFD) OF THE PF AND REALIZED THAT ALTHOUGH THE PF HAD COUPLED THE FLT GUIDANCE CTL COMPUTER BACK TO THE R SIDE, THE PF WAS STILL IN 'GREEN NEEDLES' WHICH MEANT THAT HE HAD NOT RESELECTED THE FMS AS HIS PRIMARY NAV SOURCE AFTER THE DC GENERATOR FAILURE. THIS MEANT THAT BOTH LNAV AND VNAV MODES WERE NOT AVAILABLE. THE PNF ANNOUNCED THAT THE PF WAS STILL IN 'GREEN NEEDLES,' SELECTED THE FMS FOR THE PF AND REARMED THE DESIRED LATERAL AND VERT MODES WHICH NOW ENGAGED PROPERLY. ATC DID NOT COMMENT OR INQUIRE ABOUT THE MOMENTARY ARREST OF THE DSCNT AND CLB AND NO OTHER TFC WAS IN THE AREA. CONTRIBUTING FACTORS: THE PF HAD NO EXPERIENCE WITH THIS TYPE OF FAILURE IN FLT AND HOW IT WOULD EFFECT THE FLT GUIDANCE HE WAS RECEIVING ON THE R PFD. IT WAS A MISTAKE TO ENGAGE THE AUTOPLT AS SOON AS THE FLT DIRECTOR BARS CAME UP BEFORE VERIFYING THAT DESIRED FLT ATTITUDE/PATH WOULD BE MAINTAINED. FURTHERMORE, IT SHOULD HAVE BEEN ANNOUNCED TO THE PNF THAT THE AUTOPLT HAD BEEN ENGAGED. ALSO, THE FAILURE TO RESELECT THE FMS AS THE PRIMARY NAV SOURCE RESULTED IN THE REQUESTED LATERAL MODE NOT BEING AVAILABLE. THE WORKLOAD OF THE PNF WAS HIGHER THEN USUAL DUE TO THE FACT THAT THE CABIN ALT HAD TO CONSTANTLY BE MANAGED/MONITORED DUE TO THE AUTOMATIC MODE BEING MEL'ED. IN THE IMMEDIATE AFTERMATH OF THEDC GENERATOR FAILURE, HE WAS TOO FOCUSED ON THIS AND THE DESIRE TO GET THE APU STARTED UP AND THE APU DC GENERATOR ON LINE SO THAT DC PWR REDUNDANCY WOULD BE AVAILABLE IN CASE THE R DC GENERATOR FAILED AGAIN WHILE IN THE TERMINAL ENVIRONMENT FOR AN INST APCH AT NIGHT (WHICH IT DID). THE PNF SHOULD HAVE BEEN MORE FOCUSED ON THE INEXPERIENCE OF THE PF IN THIS TYPE OF A SIT. ALTHOUGH IT WAS NOTICED THAT THE FLT GUIDANCE COMPUTER WAS RE-COUPLED TO THE FO'S SIDE, IT WAS A CRUCIAL MISTAKE NOT TO VERIFY THAT THE FO WAS ALSO BACK ON THE FMS BEFORE TRYING TO ENGAGE A VNAV MODE. THE CAPT HAD NOTICED EARLIER THAT THE FO WAS NOT IN THE HABIT OF SETTING THE MANUAL SPD BUG TO MATCH WHAT HE WAS FLYING FROM THE FMS. THE CAPT CONSIDERED THIS 'BAD FORM' BUT NOT AGAINST COMPANY POLICY AND HAD DECIDED NOT TO 'BADGER' HIS FO OVER A 'STYLE ISSUE.' HOWEVER, HAD THE SPD BUG BEEN SET CONSISTENTLY WITH THE DSCNT PROFILE, THE FLT PATH DEV WOULD NOT HAVE OCCURRED IN-SPITE OF THE VNAV MODE (VFLCH) NOT ENGAGING AND THE AUTOPLT BEING ENGAGED BEFORE VERIFYING IT WOULD MAINTAIN THE DESIRED FLT PATH.

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.