The following is the full text of a letter sent by the author to the FAA.
This letter is to inform you of an incident that occurred during a routine flight from Austin, Texas (AUS) to Addison Airport, Dallas, Texas (ADS).
This incident in no way deals with any violation of current FAR’s or use of ATC in and emergency fashion. The flight in question took place on March 10, 1997 and the incident at approximately 19:36 CDT.
I have written this letter to bring to light a possible connection to the earlier crashes involving Piper Malibus. The facts are as follows:
Piper Malibu N9298Y operated by National Institute for Community Banking, owned by D&L Flyers, Inc. Dallas, Texas. PIC was myself, Jason E. Palisin holder of ATP certificate 291621500. Aircraft was purchased in October of 1996 from Columbia Aircraft Sales, Groton, Connecticut. Total time of aircraft is 2400, 900 on the engine. I have 315 time in type.
The problem with the aircraft is mostly noted on accent or decent while using the Altitude Preselect/Alerter. The problem is that of the Encoding Altimeter, not the Autopilot. Earlier it was suggested that the culprit of the late Eighties in-air break-ups was caused by the KFC-150 Autopilot system. My accounts may suggest the encoding altimeter may have played a role into the mishaps.
N9298Y is equipped with a KFC-150 Flight director/autopilot, a King transponder/with flight level readout, a KX-297B Preselect/Altitude Alerter, and a KLN-90A IFR GPS. All of these items are tied into electrical data lines tapped off the encoding altimeter. The problem of pitch-over was first recognized the day of purchase on the pre-sale inspection test flight. During a climb, the digital readout of the transponder, started to blink, first showing the correct altitude we were at, e.g. (095) then it would blink and during the blinks, the readout would change: (095)-(414)-(530)-(358)-(095) then would go blank (—-) at that time, the altitude alerter would sound a warning horn and finally the GPS Message light would illuminate and blink that an altitude fail had occurred. When all of this happened, the aircraft went from a climb attitude, approximately 3.5 degrees at 650 FPM to a pitch-over approximately 2-3 degrees before we had the chance to disconnect the autopilot. The problem was remedied by recycling the Encoder circuit breaker and resetting it, restoring the functions of the autopilot and Mode C of the transponder. The problem resurfaced one or two times and after reset, did not show up again.
The incident that occurred on March 10 could be categorized as a severe pitch-over.
Descending into the DFW Class B, out of 9500, VFR, and using the Preselectors’ rate of decent/accent controller, I set up a descent of 500 FPM, indicated airspeed was 175 kts, 5 knots below the yellow arc on the airspeed indicator. Upon passing 9000, the transponder started to blink, before I knew it, the warning horn went off, and the aircraft pitched-over so hard I was off my seat for a second. Hitting the autopilot interrupt for control wheel steering, I stopped the pitch-over re-established the descent rate I wanted, and released the CWS. At that moment, the aural warning went off again, and the aircraft pitched over again, this time more violently. When this occurred, I saw the descent rate controller go from my pre-selected rate of 500 FPM down to 1,100 FPM up them immediately to 3000 FPM down, and the elevator trim wheel started to race to the nose down trim location. At this second time, I disconnected the autopilot and re-established the descent profile by hand with no other incident.
Once on the ground I started to think about the earlier crashes and started to piece together my own puzzle, and what I saw. These were my conclusions:
The autopilot is a simple computer that responds to the inputs of what the pilot wants it to do, to interrupt it does not reset it, and when re-activated will return to its last instructions or commands, therefore positive understanding and knowledge of limitations are imperative for the operation of such systems.
The Malibu, although a single engine piston, is a very complex aircraft, pilots transitioning from an A36 Bonanza or Cessna T-210 are in for a big surprise. This airplane demands constant training, recurrent training and requires respect.
The autopilot/flight director carries out the last instructions given to it, therefore if the encoder begins to give conflicting information, the airplane may not know where it is at, and therefore tries to compensate with the last data received. If the encoder, as shown by the transponder readout, kicks offline and the last info sent was that the aircraft is at 41,400 feet it tries to return the aircraft to where it originally was told to go, i.e., the altitude pre-selected.
The pitch-over was only encountered in a climb or a descent, not in level flight, and only when the KX-297B was used for pitch rate control. Basically when the preselector is dependent on information sent by the encoder.
I was aware when the phenomenon was about to occur, and therefore prepared for the pitch deviations. Pilots unaware have no prior warning.
Upon having an avionics shop look at the possible problem, they reported that the encoder may have a problem running four lines off of it, possibly that the lines have crossed or that there maybe a voltage problem. This was evident when they said that the problem may have occurred after the installation of the KLN-90B GPS.
Pilots unfamiliar with the Malibu, its slick and speedy design can cause the aircraft to accelerate quickly into the yellow arc or even red line, especially in a descent, therefore if this pitch-over had occurred, with the same problem that our aircraft had, in a descent, with the airspeed very high, the possibility of over-stress is very possible.
If the pitch-over occurred in turbulent or IFR conditions, or both, the pilot may not have noticed until the airspeed was observed. In a split-second decision, if the pilot tried to pull out of the dive without disconnecting the autopilot, the autopilot would try to compensate for the pilots input and put it into a steeper dive.
The fact that the rate controller feature of the KX-297B automatically defaults to the steep descent rate profile may also be suspect, I feel that if the Preselector would have leveled off the aircraft or defaulted to zero in case of an uninstructed disturbance, the autopilot would not try to chase it.
The fact that the crashes could have been linked to pilot error, or improper decision making may be relevant, but this airplane requires pilots to understand super critical airfoils, the understanding of high-altitude operations and autopilot theory and application. In no way do I consider myself an expert, and I certainly do not know everything about flying, but I hope this report of events helps in the determination of cause. It is a tragedy that these crashes occurred, but not knowing why is tough also. I hope this helps.
In conclusion to this report, I will be available for discussion or further investigation should the need arise.
Thank You,
(signed)
Jason E. Palisin
ATP, CFI, CFII, MEI
