Narrative:

We had just completed a STAR and were being vectored by ATC for a visual approach. WX was VMC with moderate to light turbulence, plus windshear reported from other aircraft on final approach. Our assigned altitude was 5000 ft with altitude and heading modes selected on the flight director panel. Autoplt was engaged. I was the PF on this leg. The controller issued us a new heading, which was entered with the heading bug. I also selected bank limit to keep the bank angle shallow, in the interest of passenger comfort. In the turn, the captain asked me if I had selected the bank limit feature. I replied yes, but the indicator light on the bank angle button was very faint, so we both had to look closely to see if it was on. It was. As I returned to my scan, I noticed the aircraft had left its assigned altitude and was slowly descending in the turn. For just a moment, I thought we had received a new altitude assignment with the last vector, but the altitude on the screen (not the altimeter) was still 5000 ft. No new altitude was assigned. I noted this to the captain, who suggested to disconnect the autoplt. This I did, and began correcting the approximately 200 ft altitude deviation. I also glanced at the flight director panel and noticed that no vertical modes were engaged. At no time did either of us cancel any vertical mode during the turn. About this time (5-10 seconds after I began to correct the descent and disconnected the autoplt) the controller called asking us to verify our altitude. We then completed the visual approach. This was the first time I had seen the altitude hold mode on the flight director disconnect without the pilot's input. Callback conversation with reporter revealed the following information: reporter indicated at the time the turn was initiated, both the captain and first officer were scanning for reported traffic. The first officer remembered the first indication of no altitude hold was that all vertical mode lights were off. He also recalled the heading mode lights were indicating properly. Another point made was that the shape coding and location of the vertical and heading control knobs were significantly different so that the possibility of the error of using one for the other was remote. It was possible that the excursion of 200 ft was not great enough to trigger the warning system of an altitude deviation. Also the parameters, which in one case energize the warning system in the aircraft and in the other alert the ground controller, should be adjusted to allow the crew warning to precede the controller warning with sufficient time for the crew to react. The crew was at a loss as to an explanation for the apparent loss of altitude hold and as a result did not write up the malfunction.

Google
 

Original NASA ASRS Text

Title: THE ACFT WAS IN STRAIGHT AND LEVEL FLT WITH THE AUTOPLT ENGAGED AND ALT AND HDG MODES SELECTED. WHEN A NEW HDG WAS SELECTED THE ACFT BEGAN A SHALLOW DSCNT IN THE TURN.

Narrative: WE HAD JUST COMPLETED A STAR AND WERE BEING VECTORED BY ATC FOR A VISUAL APCH. WX WAS VMC WITH MODERATE TO LIGHT TURB, PLUS WINDSHEAR RPTED FROM OTHER ACFT ON FINAL APCH. OUR ASSIGNED ALT WAS 5000 FT WITH ALT AND HDG MODES SELECTED ON THE FLT DIRECTOR PANEL. AUTOPLT WAS ENGAGED. I WAS THE PF ON THIS LEG. THE CTLR ISSUED US A NEW HDG, WHICH WAS ENTERED WITH THE HDG BUG. I ALSO SELECTED BANK LIMIT TO KEEP THE BANK ANGLE SHALLOW, IN THE INTEREST OF PAX COMFORT. IN THE TURN, THE CAPT ASKED ME IF I HAD SELECTED THE BANK LIMIT FEATURE. I REPLIED YES, BUT THE INDICATOR LIGHT ON THE BANK ANGLE BUTTON WAS VERY FAINT, SO WE BOTH HAD TO LOOK CLOSELY TO SEE IF IT WAS ON. IT WAS. AS I RETURNED TO MY SCAN, I NOTICED THE ACFT HAD LEFT ITS ASSIGNED ALT AND WAS SLOWLY DSNDING IN THE TURN. FOR JUST A MOMENT, I THOUGHT WE HAD RECEIVED A NEW ALT ASSIGNMENT WITH THE LAST VECTOR, BUT THE ALT ON THE SCREEN (NOT THE ALTIMETER) WAS STILL 5000 FT. NO NEW ALT WAS ASSIGNED. I NOTED THIS TO THE CAPT, WHO SUGGESTED TO DISCONNECT THE AUTOPLT. THIS I DID, AND BEGAN CORRECTING THE APPROX 200 FT ALTDEV. I ALSO GLANCED AT THE FLT DIRECTOR PANEL AND NOTICED THAT NO VERT MODES WERE ENGAGED. AT NO TIME DID EITHER OF US CANCEL ANY VERT MODE DURING THE TURN. ABOUT THIS TIME (5-10 SECONDS AFTER I BEGAN TO CORRECT THE DSCNT AND DISCONNECTED THE AUTOPLT) THE CTLR CALLED ASKING US TO VERIFY OUR ALT. WE THEN COMPLETED THE VISUAL APCH. THIS WAS THE FIRST TIME I HAD SEEN THE ALT HOLD MODE ON THE FLT DIRECTOR DISCONNECT WITHOUT THE PLT'S INPUT. CALLBACK CONVERSATION WITH RPTR REVEALED THE FOLLOWING INFO: RPTR INDICATED AT THE TIME THE TURN WAS INITIATED, BOTH THE CAPT AND FO WERE SCANNING FOR RPTED TFC. THE FO REMEMBERED THE FIRST INDICATION OF NO ALT HOLD WAS THAT ALL VERT MODE LIGHTS WERE OFF. HE ALSO RECALLED THE HDG MODE LIGHTS WERE INDICATING PROPERLY. ANOTHER POINT MADE WAS THAT THE SHAPE CODING AND LOCATION OF THE VERT AND HDG CTL KNOBS WERE SIGNIFICANTLY DIFFERENT SO THAT THE POSSIBILITY OF THE ERROR OF USING ONE FOR THE OTHER WAS REMOTE. IT WAS POSSIBLE THAT THE EXCURSION OF 200 FT WAS NOT GREAT ENOUGH TO TRIGGER THE WARNING SYS OF AN ALTDEV. ALSO THE PARAMETERS, WHICH IN ONE CASE ENERGIZE THE WARNING SYS IN THE ACFT AND IN THE OTHER ALERT THE GND CTLR, SHOULD BE ADJUSTED TO ALLOW THE CREW WARNING TO PRECEDE THE CTLR WARNING WITH SUFFICIENT TIME FOR THE CREW TO REACT. THE CREW WAS AT A LOSS AS TO AN EXPLANATION FOR THE APPARENT LOSS OF ALT HOLD AND AS A RESULT DID NOT WRITE UP THE MALFUNCTION.

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.