Narrative:

I was delivering my plane, a raytheon/beech A36, from pae (paine field, everett, wa) to the shop at tiw (tacoma narrows, gig harbor, wa, 40 NM south) for its annual inspection. It was a dark and stormy night, raining, with a strong (20-40 KTS) headwind, light to moderate turbulence, 5-7 mi visibility, 700-1500 ft scattered to broken, 3000-4000 ft overcast. Although short, this IFR flight goes into, then below, the seattle class B airspace and is quite busy with frequency and altitude assignments. Under these conditions, especially the turbulence, the workload was high for a single pilot. At one point, I felt myself getting behind the airplane so I reduced power and slowed the plane down from 160 KTS to 120 KTS to give myself more time. This worked, it caught me up with the airplane, but I feel it contributed to what came next. Sea approach vectored me to intercept the tiw localizer for runway 17, then cleared me to descend to and maintain 2000 ft MSL until GS intercept at the FAF, 11 mi ahead. Both navigation #1 and #2 were dialed in to the ILS frequency, I was hand flying on the flight director. After a few mins, the GS indicator of the #1 navigation appeared at the top of the HSI, then smoothly, gradually descended to 1/2 DOT above the center, where I lowered the gear, began my descent, and contacted tiw tower. The tower said that his radar showed no target at the FAF, only one still 7 mi out from the FAF. At that point, I saw that the #2 GS still showed the GS full scale above me. I had not reached the FAF after all, and had begun my descent (and switched frequencys) too soon. I was at about 1500 ft MSL, 500 ft below my last assigned altitude and only about 100 ft AGL. The HSI GS indicator had failed, responding to any GS signal by always indicating 1/2 DOT below the GS. The primary cause of this altitude deviation was the HSI failure, but I had enough information available to know that I was not at the FAF, had I looked. Distance to the airport was displayed on the GPS and the approach plate gave the distance from the FAF to the threshold, not exactly the same measure, but close enough to show that I was almost 3 times farther away than I thought. I could have used the xradial to identify the FAF, but I had tuned both navs to the localizer frequency. Another factor was the unusually low ground speed, due to the reduced power and strong headwind, so the time elapsed to what I thought was the FAF didn't seem too short. The HSI had exhibited this GS failure mode twice before in-flight, but always performed flawlessly in the shop, so the problem had not been found or fixed. In addition, both of the previous failures had been sudden and obvious, snapping quickly to the 1/2 DOT low indication, leading me to the belief that I could identify when it had failed by watching it, a belief that was obviously wrong. Upon returning to my assigned altitude of 2000 ft, I turned on the autoplt and set it to hold altitude while I reconfigured the radios. When I looked up again, the autoplt had deviated above my assigned altitude, climbing to 2300 ft. The approach mode had superseded the altitude hold and since the HSI was indicating 1/2 DOT below the GS, the autoplt began climbing in an attempt to center it. At that point, I turned off the autoplt/flight director altogether, returned to my assigned altitude, and hand flew the remainder of the approach on the #2 navigation/GS. Since I was delivering the plane for its annual inspection anyway, I added this HSI problem to the list of things to be checked. Until the problem with the HSI is found and fixed, ILS approachs will be flown only with the #2 navigation/GS, without the autoplt, flight director or HSI. Finally, I learned that any component that is known to fail intermittently can not be relied upon even when it appears to be behaving. Anything that is working, sometimes must be considered broken, always.

Related reports:

Google
 

Original NASA ASRS Text

Title: A VERY LOW TIME PLT, FLYING A VERY COMPLEX ACFT, RPTED A SERIES OF ERRORS DURING A SHORT FLT FROM PAE TO TIW.

Narrative: I WAS DELIVERING MY PLANE, A RAYTHEON/BEECH A36, FROM PAE (PAINE FIELD, EVERETT, WA) TO THE SHOP AT TIW (TACOMA NARROWS, GIG HARBOR, WA, 40 NM S) FOR ITS ANNUAL INSPECTION. IT WAS A DARK AND STORMY NIGHT, RAINING, WITH A STRONG (20-40 KTS) HEADWIND, LIGHT TO MODERATE TURB, 5-7 MI VISIBILITY, 700-1500 FT SCATTERED TO BROKEN, 3000-4000 FT OVCST. ALTHOUGH SHORT, THIS IFR FLT GOES INTO, THEN BELOW, THE SEATTLE CLASS B AIRSPACE AND IS QUITE BUSY WITH FREQ AND ALT ASSIGNMENTS. UNDER THESE CONDITIONS, ESPECIALLY THE TURB, THE WORKLOAD WAS HIGH FOR A SINGLE PLT. AT ONE POINT, I FELT MYSELF GETTING BEHIND THE AIRPLANE SO I REDUCED PWR AND SLOWED THE PLANE DOWN FROM 160 KTS TO 120 KTS TO GIVE MYSELF MORE TIME. THIS WORKED, IT CAUGHT ME UP WITH THE AIRPLANE, BUT I FEEL IT CONTRIBUTED TO WHAT CAME NEXT. SEA APCH VECTORED ME TO INTERCEPT THE TIW LOC FOR RWY 17, THEN CLRED ME TO DSND TO AND MAINTAIN 2000 FT MSL UNTIL GS INTERCEPT AT THE FAF, 11 MI AHEAD. BOTH NAV #1 AND #2 WERE DIALED IN TO THE ILS FREQ, I WAS HAND FLYING ON THE FLT DIRECTOR. AFTER A FEW MINS, THE GS INDICATOR OF THE #1 NAV APPEARED AT THE TOP OF THE HSI, THEN SMOOTHLY, GRADUALLY DSNDED TO 1/2 DOT ABOVE THE CTR, WHERE I LOWERED THE GEAR, BEGAN MY DSCNT, AND CONTACTED TIW TWR. THE TWR SAID THAT HIS RADAR SHOWED NO TARGET AT THE FAF, ONLY ONE STILL 7 MI OUT FROM THE FAF. AT THAT POINT, I SAW THAT THE #2 GS STILL SHOWED THE GS FULL SCALE ABOVE ME. I HAD NOT REACHED THE FAF AFTER ALL, AND HAD BEGUN MY DSCNT (AND SWITCHED FREQS) TOO SOON. I WAS AT ABOUT 1500 FT MSL, 500 FT BELOW MY LAST ASSIGNED ALT AND ONLY ABOUT 100 FT AGL. THE HSI GS INDICATOR HAD FAILED, RESPONDING TO ANY GS SIGNAL BY ALWAYS INDICATING 1/2 DOT BELOW THE GS. THE PRIMARY CAUSE OF THIS ALTDEV WAS THE HSI FAILURE, BUT I HAD ENOUGH INFO AVAILABLE TO KNOW THAT I WAS NOT AT THE FAF, HAD I LOOKED. DISTANCE TO THE ARPT WAS DISPLAYED ON THE GPS AND THE APCH PLATE GAVE THE DISTANCE FROM THE FAF TO THE THRESHOLD, NOT EXACTLY THE SAME MEASURE, BUT CLOSE ENOUGH TO SHOW THAT I WAS ALMOST 3 TIMES FARTHER AWAY THAN I THOUGHT. I COULD HAVE USED THE XRADIAL TO IDENT THE FAF, BUT I HAD TUNED BOTH NAVS TO THE LOC FREQ. ANOTHER FACTOR WAS THE UNUSUALLY LOW GND SPD, DUE TO THE REDUCED PWR AND STRONG HEADWIND, SO THE TIME ELAPSED TO WHAT I THOUGHT WAS THE FAF DIDN'T SEEM TOO SHORT. THE HSI HAD EXHIBITED THIS GS FAILURE MODE TWICE BEFORE INFLT, BUT ALWAYS PERFORMED FLAWLESSLY IN THE SHOP, SO THE PROB HAD NOT BEEN FOUND OR FIXED. IN ADDITION, BOTH OF THE PREVIOUS FAILURES HAD BEEN SUDDEN AND OBVIOUS, SNAPPING QUICKLY TO THE 1/2 DOT LOW INDICATION, LEADING ME TO THE BELIEF THAT I COULD IDENT WHEN IT HAD FAILED BY WATCHING IT, A BELIEF THAT WAS OBVIOUSLY WRONG. UPON RETURNING TO MY ASSIGNED ALT OF 2000 FT, I TURNED ON THE AUTOPLT AND SET IT TO HOLD ALT WHILE I RECONFIGURED THE RADIOS. WHEN I LOOKED UP AGAIN, THE AUTOPLT HAD DEVIATED ABOVE MY ASSIGNED ALT, CLBING TO 2300 FT. THE APCH MODE HAD SUPERSEDED THE ALT HOLD AND SINCE THE HSI WAS INDICATING 1/2 DOT BELOW THE GS, THE AUTOPLT BEGAN CLBING IN AN ATTEMPT TO CTR IT. AT THAT POINT, I TURNED OFF THE AUTOPLT/FLT DIRECTOR ALTOGETHER, RETURNED TO MY ASSIGNED ALT, AND HAND FLEW THE REMAINDER OF THE APCH ON THE #2 NAV/GS. SINCE I WAS DELIVERING THE PLANE FOR ITS ANNUAL INSPECTION ANYWAY, I ADDED THIS HSI PROB TO THE LIST OF THINGS TO BE CHKED. UNTIL THE PROB WITH THE HSI IS FOUND AND FIXED, ILS APCHS WILL BE FLOWN ONLY WITH THE #2 NAV/GS, WITHOUT THE AUTOPLT, FLT DIRECTOR OR HSI. FINALLY, I LEARNED THAT ANY COMPONENT THAT IS KNOWN TO FAIL INTERMITTENTLY CAN NOT BE RELIED UPON EVEN WHEN IT APPEARS TO BE BEHAVING. ANYTHING THAT IS WORKING, SOMETIMES MUST BE CONSIDERED BROKEN, ALWAYS.

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.