Narrative:

Vertical deviation from assigned 5000 ft altitude departing ZZZ while dealing with a pitch failure. Situation: the departure was a flight into an unrestricted VFR sky; part 91; company owned; lightly loaded; 1979 mitsubishi mu-2 solitaire; IFR. The plane had not flown for 3 months; waiting for an engine overhaul. This was the return leg of the first cross country flight; with no passenger aboard. Relevant history: the aircraft has documented history of intermittent un-commanded pitch excursions from a 27-year-old bendix M4D autoplt. However; the FD/autoplt has been expensively maintained by the most respected M4D shop in the USA. And for several months (prior to being taken out of service for the left engine oh); the pitch excursions had stopped occurring. Incident: the day of this incident was the first flight after 3 months of downtime. Approaching 5000 ft in a fast 2500 FPM plus climb; the FD/ap stuck pitched up. Pitch sync became inoperative. Disconnect: I first attempted to stop the climb on autoplt; using the autoplt command control head. Unable to do so; I then hit the emergency autoplt disconnect while rapidly passing through 5600 ft altitude; while pushing the wheel (with no passenger); firmly forward. ATC caught this maneuver on radar as a deviation; but without loss of separation. Stuck v-bars: after the incident; the v-bars never returned to normal; remaining locked in an up position for the remainder of the flight; as well as after landing; as witnessed by two mechanics from two separate companies on the ground at ZZZ1. Pitch sync remained inoperative throughout the flight and on the ground after landing; which is the only primary leveling control for the v-bars. Additionally; after turning the FD completely off in-flight; the v-bars remained abnormally in view at the bottom of the ADI. Both inoperative pitch sync and abnormal stow were also witnessed on the ground. Hard failure. We further determined on the ground that de-powering and re-powering the aircraft did not clear any faults. It now appears we no longer have an intermittent situation -- but a solid failure that can now finally be traced and resolved.callback conversation with reporter revealed the following information: reporter stated this is a bendix M4D autopilot. He was not able to stop the uncommanded climb while using the autopilot command control head; requiring him to hit the emergency autopilot disconnect. This uncommanded vertical pitch lock problem has happened to his aircraft aprox 9 times. Eight of those incidents could not be duplicated by maintenance troubleshooting. After the ninth event of uncommanded automatic pitch-up and with the maintenance bench tests again not being able to duplicate the uncommanded pitch lock; reporter stated he happened to mention to the avionics techs; the pitch lock appears to happen when the aircraft is flying in cool to cold weather; especially fall and winter. This prompted the maintenance tech to place the autopilot in a very cool refrigerator and then subjected the unit to their bench tests. The autopilot unit immediately duplicated the uncommanded pitch lock in the pitch-up and pitch-down modes. The autopilot computer control cards are located in the tail cone and subject to whatever temperature is outside. This bendix M4D has a history of intermittent uncommanded pitch excursions. Honeywell had taken over repairs of the bendix M4D autopilot and the magnetic type clutch servos used for autopilot control of the rudder; elevators and spoilerons. That is; until recently; when honeywell announced they would no longer provide repair service on the bendix M4D autopilot or servos. Reporter further states his flight director ADI is a sperry HZ454 unit. The FAA has also issued sfar-2006-24981 #108; requiring a functional autopilot under IMC conditions.

Google
 

Original NASA ASRS Text

Title: A MITSUBISHI MU-2-SOLITAIRE PILOT EXPERIENCED AN UNCOMMANDED VERTICAL DEVIATION DUE TO UN-SELECTED PITCH LOCK OF HIS BENDIX M4D AUTOPILOT.

Narrative: VERTICAL DEVIATION FROM ASSIGNED 5000 FT ALT DEPARTING ZZZ WHILE DEALING WITH A PITCH FAILURE. SITUATION: THE DEP WAS A FLT INTO AN UNRESTRICTED VFR SKY; PART 91; COMPANY OWNED; LIGHTLY LOADED; 1979 MITSUBISHI MU-2 SOLITAIRE; IFR. THE PLANE HAD NOT FLOWN FOR 3 MONTHS; WAITING FOR AN ENG OVERHAUL. THIS WAS THE RETURN LEG OF THE FIRST CROSS COUNTRY FLT; WITH NO PAX ABOARD. RELEVANT HISTORY: THE ACFT HAS DOCUMENTED HISTORY OF INTERMITTENT UN-COMMANDED PITCH EXCURSIONS FROM A 27-YEAR-OLD BENDIX M4D AUTOPLT. HOWEVER; THE FD/AUTOPLT HAS BEEN EXPENSIVELY MAINTAINED BY THE MOST RESPECTED M4D SHOP IN THE USA. AND FOR SEVERAL MONTHS (PRIOR TO BEING TAKEN OUT OF SERVICE FOR THE LEFT ENG OH); THE PITCH EXCURSIONS HAD STOPPED OCCURRING. INCIDENT: THE DAY OF THIS INCIDENT WAS THE FIRST FLT AFTER 3 MONTHS OF DOWNTIME. APCHING 5000 FT IN A FAST 2500 FPM PLUS CLB; THE FD/AP STUCK PITCHED UP. PITCH SYNC BECAME INOPERATIVE. DISCONNECT: I FIRST ATTEMPTED TO STOP THE CLB ON AUTOPLT; USING THE AUTOPLT COMMAND CTL HEAD. UNABLE TO DO SO; I THEN HIT THE EMER AUTOPLT DISCONNECT WHILE RAPIDLY PASSING THROUGH 5600 FT ALT; WHILE PUSHING THE WHEEL (WITH NO PAX); FIRMLY FORWARD. ATC CAUGHT THIS MANEUVER ON RADAR AS A DEVIATION; BUT WITHOUT LOSS OF SEPARATION. STUCK V-BARS: AFTER THE INCIDENT; THE V-BARS NEVER RETURNED TO NORMAL; REMAINING LOCKED IN AN UP POSITION FOR THE REMAINDER OF THE FLT; AS WELL AS AFTER LNDG; AS WITNESSED BY TWO MECHANICS FROM TWO SEPARATE COMPANIES ON THE GND AT ZZZ1. PITCH SYNC REMAINED INOPERATIVE THROUGHOUT THE FLT AND ON THE GND AFTER LNDG; WHICH IS THE ONLY PRIMARY LEVELING CONTROL FOR THE V-BARS. ADDITIONALLY; AFTER TURNING THE FD COMPLETELY OFF INFLT; THE V-BARS REMAINED ABNORMALLY IN VIEW AT THE BOTTOM OF THE ADI. BOTH INOPERATIVE PITCH SYNC AND ABNORMAL STOW WERE ALSO WITNESSED ON THE GND. HARD FAILURE. WE FURTHER DETERMINED ON THE GND THAT DE-POWERING AND RE-POWERING THE ACFT DID NOT CLEAR ANY FAULTS. IT NOW APPEARS WE NO LONGER HAVE AN INTERMITTENT SITUATION -- BUT A SOLID FAILURE THAT CAN NOW FINALLY BE TRACED AND RESOLVED.CALLBACK CONVERSATION WITH RPTR REVEALED THE FOLLOWING INFO: REPORTER STATED THIS IS A BENDIX M4D AUTOPILOT. HE WAS NOT ABLE TO STOP THE UNCOMMANDED CLIMB WHILE USING THE AUTOPILOT COMMAND CONTROL HEAD; REQUIRING HIM TO HIT THE EMERGENCY AUTOPILOT DISCONNECT. THIS UNCOMMANDED VERTICAL PITCH LOCK PROBLEM HAS HAPPENED TO HIS ACFT APROX 9 TIMES. EIGHT OF THOSE INCIDENTS COULD NOT BE DUPLICATED BY MAINT TROUBLESHOOTING. AFTER THE NINTH EVENT OF UNCOMMANDED AUTO PITCH-UP AND WITH THE MAINT BENCH TESTS AGAIN NOT BEING ABLE TO DUPLICATE THE UNCOMMANDED PITCH LOCK; REPORTER STATED HE HAPPENED TO MENTION TO THE AVIONICS TECHS; THE PITCH LOCK APPEARS TO HAPPEN WHEN THE ACFT IS FLYING IN COOL TO COLD WEATHER; ESPECIALLY FALL AND WINTER. THIS PROMPTED THE MAINT TECH TO PLACE THE AUTOPILOT IN A VERY COOL REFRIGERATOR AND THEN SUBJECTED THE UNIT TO THEIR BENCH TESTS. THE AUTOPILOT UNIT IMMEDIATELY DUPLICATED THE UNCOMMANDED PITCH LOCK IN THE PITCH-UP AND PITCH-DOWN MODES. THE AUTOPILOT COMPUTER CONTROL CARDS ARE LOCATED IN THE TAIL CONE AND SUBJECT TO WHATEVER TEMPERATURE IS OUTSIDE. THIS BENDIX M4D HAS A HISTORY OF INTERMITTENT UNCOMMANDED PITCH EXCURSIONS. HONEYWELL HAD TAKEN OVER REPAIRS OF THE BENDIX M4D AUTOPILOT AND THE MAGNETIC TYPE CLUTCH SERVOS USED FOR AUTOPILOT CONTROL OF THE RUDDER; ELEVATORS AND SPOILERONS. THAT IS; UNTIL RECENTLY; WHEN HONEYWELL ANNOUNCED THEY WOULD NO LONGER PROVIDE REPAIR SERVICE ON THE BENDIX M4D AUTOPILOT OR SERVOS. REPORTER FURTHER STATES HIS FLIGHT DIRECTOR ADI IS A SPERRY HZ454 UNIT. THE FAA HAS ALSO ISSUED SFAR-2006-24981 #108; REQUIRING A FUNCTIONAL AUTOPILOT UNDER IMC CONDITIONS.

Data retrieved from NASA's ASRS site as of January 2009 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.