NTSB Factual Report

HISTORY OF FLIGHT

On January 19, 2004, at 1334 central standard time, a Piper PA-46-500TP, N1968W, operated by a private pilot, made a gear up forced landing following a loss of engine power on takeoff. The takeoff was being made on runway 28R (3,599 feet by 75 feet, asphalt) at the Flying Cloud Airport (FCM), Minneapolis, Minnesota. The pilot was not injured. The airplane was substantially damaged. The 14 CFR Part 91 personal flight was operating in visual meteorological conditions on an instrument flight rules flight plan. The intended destination for the flight was Sioux Falls, South Dakota.

The pilot reported that after preflighting the airplane he used a ground power unit to start the engine. He reported the outside temperature was 7 degrees Fahrenheit and the engine start was normal. He stated he received his clearance and taxied to runway 28R for takeoff. The pilot continued to report, "I did a lengthy run up including checking all gauges (at 1,900 RPM's), testing the overspeed governor and full deice test of everything but prop heat. I also tested the throttle lockout at engine idle as well as reviewing all engine gauges again. All systems appeared normal, and all operating temps and pressures were as they have been in the past." The pilot reported that the run up was longer then normal due to the overspeed problem he had in the past. The pilot reported he was cleared into position and hold at which time he "set the transponder to alt, placed both fuel pump and ignition to manual, turned on pitot heat and landing lights, and turned on window defog and taxi pulse."

The pilot reported that after being cleared for takeoff, he advanced the throttle to 1,270 pound of torque before releasing the brakes and everything was normal. He stated that upon reaching 85 knots he rotated, achieved a positive rate of climb, and retracted the landing gear. The pilot continued, "Just after the wheels were up the engine shuddered with a very substantial loss of power. It felt like a car that was hitting on half its cylinders." He stated he brought the power back to idle and focused on getting the airplane back on the runway. He stated he knew he was going to go off the end of the runway so he elected to leave the landing gear retracted. The pilot reported that after the airplane touched down he "…pulled the emergency fuel cutoff." He reported the airplane traveled straight ahead, crossed a taxiway, and came to rest about 100 yards past the taxiway. He stated that as soon as the airplane stopped, he shut off the battery and exited the airplane via the rear door.

During a subsequent interview the pilot reported that he did not have time during the accident sequence to look at the engine instruments to see what they were indicating. He also stated that the airplane felt like it was slowing down during the takeoff and there was a different pitch or tone from the airplane. He stated it sounded as if the engine was straining to produce power. The pilot stated that prior to takeoff the engine instruments were indicating 2,000 revolutions per minute (rpm) and 1,270 pounds of torque.

AIRCRAFT INFORMATION

The pilot reported he purchased the airplane in August 2003 and he had the aircraft registration number changed from N715MA to N1968W on October 6, 2003. The airplane total time at the time of the accident was 183 hours. The airplane was equipped with a Pratt & Whitney PT6A-42A, s/n PCE-RM0128 engine. The last annual inspection on both the airframe and engine was performed on January 8, 2003, at an aircraft total time of 41 hours.

The pilot reported that after he purchased the airplane he almost immediately noticed the engine was running hotter than the airplanes belonging to other pilots that he talked to. The pilot subscribed to a service that monitors information acquired by a data acquisition unit (DAU). A representative from the service stated the pilot contacted them every couple of weeks to see how the engine on N1968W was performing compared to engines in other PA-46-500Ts that they monitored. This representative stated that he informed the pilot that his engine typically ran 10 to 15 degrees hotter than other engines, but that the temperatures were still a "fair margin" below red line on the inlet turbine temperature (ITT) gauge. The pilot stated he had discussions with Piper regarding the engine temperatures and they requested he have the ITT probe and torque indicating system calibrated.

Logbook records dated September 16, 2003, state "Performed calibration test on the torque and T5 systems. Found both systems within limits at this time. Adjusted the fuel control to prevent engine from starting warm and fast." The work order for this inspection states that the ITT system was indicating 13-14 degrees higher than the input and the torque indicating system was reading 10 ft/lbs above input. Both indications were considered within limits and neither indicator was changed.

The pilot reported that during a flight on December 31, 2003, while descending from 26,000 feet to 24,000 feet, he received a propeller rpm overspeed warning. The pilot stated he began the descent at 900 feet per minute (fpm) at 800 pounds of torque. The pilot reported that during the descent the rpm overspeed warning sounded and the indication was fluctuating between 2,000 and 2,200 rpm. He stated he reduced power and raised the nose of the airplane as soon as the alarm sounded. He stated the fluctuation continued for approximately 40 seconds prior to stabilizing back at 2,000 rpm. The pilot reported the alarm stopped once the airspeed began to decrease. The pilot stated that after the alarm stopped he increased the power and continued the flight without incident.

The pilot communicated with Piper, Pratt & Whitney, and a local repair station following the indication of a propeller overspeed. Following an inspection of the airplane and engine it was determined that the problem was most likely the tachometer generator. The tachometer generator was replaced. A Product Condition/Compliance Summary of the work performed stated "Per Piper's request, rung wiring from tach gen to DAU. No faults found, also replaced tach gen with new." The Piper Specification sheet for the Propeller Tachometer states "The DU [DAU] shall drive DISCRETE OUTPUT 3 to GND when the reported Np value is greater than or equal to 2040 RPM for greater than five seconds." A representative from Piper stated the tachometer generator that was removed from N1968W was tested and no discrepancies were found. The accident occurred on the next flight.

WEATHER

The weather at FCM 19 minutes after the accident was sky condition 10,000 feet broken, visibility 10 statue miles, wind variable at 3 knots, temperature minus 13 degrees Celsius, dewpoint minus 22 degrees Celsius, altimeter 30.29 inches of mercury.

WRECKAGE AND IMPACT

The airplane traveled off the end of runway 28R, sliding across the terrain and a taxiway before coming to rest approximately 1,500 from the end of the runway.

Post accident inspection of the airplane revealed the belly skin, lower engine cowling, and radome were damaged by ground contact. There was a hole in the aft pressure bulkhead and, both the left and right wings were punctured at the landing gear.

The power lever was found in the idle position and the fuel control was in the cutoff position. The manual override lever for the fuel control was found in the unlatched position. The pilot stated he did not use the override lever and had no idea how it got in the unlatched position, unless he inadvertently kicked it with his foot as he was leaving the airplane. He stated that had he suspected a fuel control problem he would have used it, but he didn't have enough time to do anything except shut the engine down and land.

TESTS AND RESEARCH

Propeller Governors

During the manufacturing process of new PA-46 airplanes, the primary propeller governors are shipped to Pratt & Whitney where they are installed on the engine prior to the engine being shipped to Piper. The overspeed governors are shipped to Piper where they are installed on the engine prior to the engine being installed in the airplane.

Both the primary and overspeed propeller governors were removed from the engine and tested at Woodward in Rockford, Illinois, on February 5, 2004.

Primary Governor

A visual inspection of the primary governor, Woodward part number 8210-007, Pratt & Whitney part number 3035930, s/n 130806980, was conducted. The lockwire found on the airbleed system reset-eccentric screw was non-standard for Woodward and the torque sealant had been removed. Bench testing of the unit revealed the maximum speed setting was 3,856 rpm. This setting is 11 rpm above the maximum test limit of 3,840 +/- 5 rpm. The as-received setting of 3,856 rpm equated to 2,008 or 101% propeller rpm.

Documents provided by Woodward indicate the primary governor was set at a maximum speed setting of 3,842 rpm when it was last tested at their facility on July 11, 2001.

Overspeed Governor

A visual inspection of the overspeed governor, Woodward part number 210638, Piper part number 463-264, s/n 13093188, revealed the inside of the speed setting hex screw did not contain any torque sealant. The sealant on the outside threads of the screw was cracked near the base of the screw and the housing.

The specified maximum speed setting is 4,047 +/- 10 rpm. This value corresponds to 2,120 propeller rpm or 106 percent of propeller rpm. Bench testing of the unit revealed the maximum speed setting was 3,810 rpm. This setting is 227 rpm below the maximum specified in the test protocol. The as-received setting of 3,810 rpm equates to 1,996 or 99.8% propeller rpm.

As tested the Minimum Speed - Reset setting was 3,289 rpm which is 241 rpm below the minimum limit of 3,570 +/- 40 rpm.

Documents provided by Woodward indicate the Maximum Speed setting was 4,047 rpm and the Minimum Speed - Reset setting was 3,580 rpm when the overspeed governor last tested at their facility on August 2, 2001.

Torque Sealant

According to Woodward, their standard practice is to fill the head of the adjustment hex screw with torque sealant to prevent unauthorized field adjustment of the governors. During a tour of the Piper facility in Vero Beach, Florida, it was noted that the torque sealant on an overspeed governor that was installed on an engine in the plant did not contain torque sealant inside the hex screw. Others viewed in the plant did contain the sealant inside the screw.

A representative from Piper reported that if the maximum speed setting on the overspeed governor was adjusted during production, it would not have been shown in the aircraft paperwork. However, if a problem with the propeller rpm was written up during a test flight the adjustment would have been indicated in the aircraft paperwork. The same representative stated that in the past Piper had made adjustments to the overspeed governor, but that practice was no longer continued and was most likely stopped prior to the production of the accident airplane. Piper would not reapply torque sealant if they had adjusted the governor.

A review of the flight test documents revealed the following four entries regarding the adjustment of the maximum reverse rpm.

11/19/01 Discrepancy: Max Rev RPM low, 1800 s/b 1820-1880 Q=320.

Action: ADJ Reverse 1820/360

11/20/01 Discrepancy: Max Reverse Q=380 Np=1820 See #12

Action: Adjust Rev. TQ=340 Np=1840

11/21/01 Discrepancy: Repeat #10 Max Rev Torque High 370Q @ 1850 RPM

Action: Adjust Rev

12/10/01 Discrepancy: Repeat #10 - Reverse 340/1810

Action: Adjust Rev

ENGINE

The engine was shipped to Pratt & Whitney in Montreal, Canada, where it was torn down on May 4 and 5, 2004, under the supervision of an investigator from the Canadian Transportation Safety Board.

The compressor discharge pneumatic lines were intact from the gas generator fitting to the FCU. The power turbine control pneumatic lines were intact at the fuel control unit, but had been disconnected at the propeller governor for its removal. All chip detectors and filters were clean of debris. Circumferential rubbing was visible in the compressor section and the bearings were intact. All of the turbine blades were intact and rubbing was evident on the blade tips and on the shroud. Debris (dirt and grass) was found in the intake screen, the combustion can, and in the compressor. The fuel control unit, fuel pump, flow divider, fuel nozzles, fuel heater, compressor bleed valves and the torque limiting unit were all tested. No anomalies that would have precluded normal operation of the engine were noted during these tests.

FLIGHT TESTS

Flight testing was performed at Piper in Vero Beach, Florida, on April 28 - 29, 2004. The flight tests were performed using one of Piper's PA-46-500TP test airplanes, N402MM. Both the primary and overspeed propeller governors were removed from the airplane and sent to Woodward for calibration. The governors were calibrated to the settings found on the governors which were removed from N1968W after the accident. The governors were then shipped back to Piper and installed on N402MM just prior to the flight tests.

A test plan was developed by Piper and reviewed by NTSB, Woodward, and Pratt & Whitney prior to the testing. The test plan stated, "For the purpose of this testing, an overspeed governor will be adjusted by the supplier to a setting of 99.8% engine RPM (1,995 engine RPM, 3,810 governor RPM) in the "reset" mode and installed on an aircraft for testing. Setting the overspeed governor to this value in the "reset" mode allows the pilot to operate the aircraft with this RPM setting while in a test mode and, with the flip of a switch, the pilot will be able to return to the normal mode of governor operation which is at an RPM setting of 106%. This setup will help minimize the safety risks during the flight test portion of the test plan."

The test plan called for engine ground tests, high speed taxi tests, and flight tests. All of the tests were completed with the exception of one set of high speed taxi tests which was not completed based on the results of the other tests and the increased risk due to the gusty wind conditions. The results of the tests were recorded on the DAU.

On the first day of testing the outside temperature ranged from 23 - 27 degrees Celsius with wind velocities varying from 13 to 21 knots. On the second day the outside temperature ranged from 24 - 26 degrees Celsius with wind velocities varying from 12 to 15 knots.

During the ground tests the propeller RPM (Np) dropped from 100% in normal mode to 96% (1,920 RPM) in reset mode. This RPM change stabilized within 3 seconds following each mode change. The second set of tests were to check the response of the airplane with rapid throttle movements. During each of the rapid throttle movement tests, the system stabilized within 5 seconds of the throttle movement. The third set of tests were the high speed taxi tests. The data indicated the RPM was limited to 96% during these tests. The flight test crew reported no "abnormal oscillations or RPM hunting" during the high speed taxi tests. The fourth set of tests was to simulate a takeoff and climb. These tests were done at an altitude of 5,500 feet mean sea level. The data indicated the RPM was limited to 97% with "no abnormal oscillations, engine shudder, or RPM hunting." The last set of tests consisted of a normal takeoff and flight around the traffic pattern. The data showed the "RPM limited to 96% NP with no abnormal oscillations, engine shudder, or RPM hunting."

The Piper Specification Sheet for the Primary Engine Display Page Layout states, "The Propeller Tachometer (Np) digital counter shall display 2,000 RPM when sensed Np is greater than or equal to 1,960 RPM and less than or equal to 2,040 RPM."

ADDITIONAL INFORMATION

Parties to the investigation were Piper Aircraft, Pratt & Whitney, and Woodward.

NTSB Probable Cause

A loss of engine power for undetermined reasons.