N279WP – Beechcraft (Textron Aviation) A36TC – 2003-07-04 NTSB Accident Report

HISTORY OF FLIGHT

On July 4, 2003, approximately 1330 Pacific daylight time, a turbine-powered Beech A36TC, N279WP, impacted the terrain during a power-out forced landing in a wheat field about three miles northwest of Pangborn Memorial Airport, Wenatchee, Washington. The commercial pilot/owner in the right seat, who was acting as pilot-in-command of this flight, received fatal injuries, the private pilot in the left seat, who was flying the aircraft at the time of the power loss, received serious injuries, and a student pilot passenger in a back seat received minor injuries. The aircraft, which was owned and operated by the pilot-in-command, sustained substantial damage. The 14 CFR Part 91 personal pleasure flight, which departed Dryden Airport, Cashmere, Washington, about 5 minutes earlier, was being operated in visual meteorological conditions. No flight plan had been filed. There was no report of an ELT activation.

According to the two surviving occupants, the purpose of the flight to Pangborn Airport in Wenatchee was to refuel the aircraft prior to heading back to their home on Vashon Island, which is located in the Puget Sound near Seattle, Washington. Prior to departing Cashmere, the pilot-in-command of the upcoming flight performed a preflight inspection of the aircraft. Neither of the other two occupants remembered seeing the pilot-in-command drain fuel from the fuel system sumps, but they did not feel it was factually accurate to say that he did not. After completion of the preflight inspection, the pilot-in-command boarded the aircraft, and the engine was started in preparation for taxi. According to the pilot in the left seat, after the engine was started, the panel-mounted Shadin Miniflo fuel management system indicated a "Fuel Remaining" of 235 pounds, and at that same point in time, the left main fuel tank gauge was at the top of the yellow arc (approximately 15 gallons), and the right main fuel tank gauge needle was at the bottom of the yellow arc (approximately empty). The aircraft was then taxied to the end of the runway, and the private pilot in the left seat made the takeoff using runway 25. After takeoff, he executed a right climbing turn, and then headed toward the northern end of the town of Wenatchee. He then leveled off in smooth air about 2,000 MSL (Mean Sea Level), and established a cruise speed of about 130 knots. Sometime during the first two minutes after level-off, the right main tank "Fuel Low" light came on and stayed illuminated. Within one minute of the illumination of the Fuel Low light, the engine lost power and began to spool down. Soon after the engine started to lose power, the pilot-in-command activated the fuel pump switches and moved the engine start switch to the down position (starter motor and igniter activation). He also rocked the wings in what the other occupants thought was an attempt to free up any fuel possibly trapped in the fuel tanks. The restart attempt was not successful, and the engine continued to spool down at a rate that was slower and smoother than what would be experienced during a ground shutdown. Approximately 10 seconds after the first indication of a loss of power the propeller came to a full stop, and within 25 seconds it appeared to be in a fully feathered position. At that point the throw-over yoke was moved from the pilot in the left seat to the pilot-in-command in the right seat, so that he could take control of the aircraft for the anticipated power-off forced landing. Initially the pilot-in-command headed for Fancher Field, a previously active airfield that is now closed, but because the aircraft was not going to be able to glide that far, he made the decision to put the aircraft down on the northwest-to-southeast axis of a wheat field located about one mile northeast of the airstrip (see satellite photograph #1). As he neared the field at an airspeed of approximately 110 knots, the pilot-in-command saw that there were power lines just off the northern (approach) end of the field, so he would be unable to touch down near the edge of the field, and instead had to clear the wires prior to descending to a touchdown point. According to the other two occupants, after crossing the wires and descending to the ground, the pilot essentially flew the aircraft onto the terrain without any significant glide or float. According to the pilot in the left seat, at the point of touchdown, the airspeed was approximately 100 knots. After the touchdown, the occupants felt a deceleration rate that indicated to them that significant braking was taking place, but, as the aircraft reached the southeast end of the field, it exited the wheat, crossed about 150 feet of grassy terrain, and then slid over the edge of a steep-sided narrow ravine. After crossing the northern edge of the ravine, the aircraft became momentarily "airborne," and then descended to a point of impact near the bottom of the ravine's southern wall.

According to the back seat occupant, who had been sitting in one of the front seats during the previous flight, he recalled that the Miniflo had indicated between 240 and 250 pounds of fuel remaining while the aircraft had been in the pattern for its last landing, and the fuel tanks gauges were at or near the aforementioned indications during the approach for landing. According to this occupant, at some time during the approach for the previous landing, the pilot-in-command stated that in the past, with the fuel gauges in that same configuration, and the Shadin indicating between 240 and 250 pounds, he had flown for an additional hour. This occupant further stated that although his memory was fuzzy, he thought he had seen the right main fuel tank "Fuel Low" light blink during the descent for landing, but it did not stay on.

WRECKAGE AND IMPACT INFORMATION

The aircraft initially touched down just over half way through the approximately 2,100 foot long wheat field, and then rolled/slid about 1,200 feet before going off the edge of the ravine. The ravine itself was about 50 feet deep, and had walls inclined at an angle of about 40 degrees. The aircraft came to rest upright on an easterly heading, with the tip of the left wing sitting at the very bottom of the ravine, and the right wing pointing uphill along the southern wall. Except for the most outboard section of both wings (containing the internal tip tanks), the entire wreckage was intact at this one location. The right wing tip, which had separated along the first full rivet line inboard of the tank, was laying on the ground under the aircraft's empennage (which was sitting elevated across the very bottom of the ravine). The left wing tip, which separated along the same rivet line, was found about 20 feet further up the ravine, and also at the ravine’s very bottom. The primary impact force had been attenuated through the lower right portion of the aircraft nose, with the forward baggage area being pushed backwards into the area directly forward of the instrument panel. In addition to that damage, the right wing, outboard of the flap-aileron junction, was crushed aft, with that portion of the wing bent upward at about 15 degrees. In addition, the left horizontal stabilizer and elevator where bent upward about 90 degrees at a point located at half of their span. The remainder of the airframe sustained mostly minor damage.

There were two areas of earth at the wreckage site that had been soaked with a liquid that had a smell and feel consistent with aviation Jet-A fuel. The first was below the point where the fuel line that feeds fuel from the fuel selector valve to the engine exited the lower left side of the cabin. This portion of the aircraft structure had been crushed up and aft, and fuel was dripping slowly from this area when investigators arrived at the scene about five hours after the accident. The fuel entering the sandy soil at this location had penetrated an area about three feet wide and one foot deep. The second location was at the point where the left wing tip had separated from its inboard structure. The investigative team dug a one foot wide by one and one-half foot deep hole at this location, which was considerably more rocky, and found that fuel had penetrated beyond the limits of that hole. The soil removed from this hole was considerably more saturated with fuel than that in the aforementioned area. There were no areas under the right wing where fuel had leaked into the earth. A further inspection of the airframe did not reveal any areas of surface staining as would be consistent with leaking fuel streaming aft over the aircraft skin while in flight, nor where there any stains on the bottom of the wings consistent with fuel leaking while the aircraft was parked on an airport ramp. The fuel selector, which is located near the knee of the left seat pilot, was found in the "ON" position.

After removal from the accident site, the aircraft was placed in a private hangar at Pangborn Field, where the investigative team performed a further inspection/teardown of the airframe and engine systems. During that process no anomalies were detected in the airframe or flight controls, and the occupant restraint systems locked properly when subjected to hand-applied acceleration loads.

As part of the inspection process, both main fuel tank lines and both header tank excess-fuel return lines were capped, and the system was pressurized with 30 pounds of air by attaching a fitting at the engine-driven fuel pump. During the two minutes that this test was run, no leaks were detected anywhere in the system, and there was no drop in pressure until one of the fittings was backed off from the header tank to ensure that this part of the system was being pressurized. After the system was depressurized, the portion of the system downstream of the main tank feed lines was disassembled in order to check for blockages or restrictions in any of the lines, and to test the fuel selector valve, the airframe fuel filter, all of the one-way check valves, and the return line orifice screens. In addition, both the primary and standby fuel pumps were tested to confirm rotation of their motors and the production of positive pressure at their outflow fittings when electrical power was applied. No anomalies were found anywhere in the system.

After completion of the field inspections, the engine was shipped to the custody of the Federal Aviation Administration Flight Standards District Office in Indianapolis, Indiana, where an NTSB-supervised partial teardown inspection was completed at the Rolls-Royce engine testing facility. The partial inspection included evaluation of the engine driven fuel pump, the engine driven fuel pump filter element, the oil filter element, the upper and lower chip detectors, and a 50 PSI pressure test of the compressor discharge pressure system (Pc). There were no anomalies or signs of contamination detected in any of the subcomponents, and the fuel pump filter bowl contained 10 ml of a straw colored fluid that had a smell and texture consistent with Jet A fuel. At the conclusion of this inspection, the impact-damaged propeller reduction gearbox was replaced by a test unit, and then, after replacement of any impact-damaged oil, fuel, or air line, the engine was test run, under the direct supervision of the FAA, to an abbreviated Production Test Specification (PTS). During the test, the engine started normally, and responded normally to all throttle inputs from idle to a full takeoff N1 of 51,181 rpm. No anomalies, malfunctions, or overheat conditions were detected during the test, and all turbine, compressor, and gearbox vibration levels were normal.

AIRCRAFT INFORMATION

According to the aircraft records, an Allison 250-B17C turbine engine was installed on this aircraft in October of 1986, in accordance with Supplemental Type Certificate (STC) AS3532NM. As part of the Soloy Conversion, the fuel system was altered, and a pair of 20 gallon Osborne tip tanks were added. Then in September 1989, the aircraft was damaged by the collapse of a hangar during conditions created by Hurricane Hugo. In 1990, the aircraft was purchased by the current owner, and soon thereafter Mace Aviation (Inspection Authorization 2164638) began "major repair" activity to repair the damage caused by the hangar collapse. In addition, Mace Aviation initiated "major alteration" activity to the wings and fuel system, with the intent of acquiring a repeatable STC for an A36TC with extended wings and a three internal tanks per wing configuration. According to documentation provided to the FAA as part of this investigation, the outboard leading edge and tank wells were parts normally used for the 110 gallon fuel system of the Beech B36TC. The inboard 40-gallon tanks were the same as the fuel cells approved for a B36TC, the outboard 15-gallon tanks were the same as those approved for a Beech B-60 Duke, and the 12-gallon integral wing tip tanks were the same as those approved for a Beech B-60 Duke. According to Mace Aviation, this alteration created a configuration similar to that of a Beech B36TC, but with the addition of an integral wing tip tank, or a configuration similar to a Beech B-60 Duke 200-gallon system, but without the box section fuel cell. In addition to the structural, system and aerodynamic changes resulting from this alteration, this configuration changed the routing of the excess fuel return line from the 40 gallon inboard tank (as in both the B36TC, and the A36TC turbine conversion) to the newly-installed integral tip tank and the outboard portion of the newly-installed outboard leading edge cell. This significantly lengthened the route the excess header tank fuel had to take in order to flow through the system and return to the fuel pump. In addition, the main fuel system vent line no longer was routed directly to the 40 gallon inboard tank (as in both the B36TC, and the A36TC turbine conversion), but instead to the integral outboard tank (with a vent standpipe interconnect to the outboard leading edge fuel cell).

Mace Aviation submitted a packet documenting this major alteration to the FAA for approval (including a FAA Form 337, dated April 2, 1993, and a Structural Substantiation Report, dated December 10, 1994), but the structural report based its conclusions only upon a discussion of the commonalities of the parts listed in the Illustrated Parts Catalogues of the A36TC, the B36TC, and the B-60. Therefore, since no acceptable approved engineering data, and no flight test data of any type, was ever submitted to the Aircraft Certification Office, the FAA canceled the STC approval project on October 23, 1997, and returned the entire documentation packet to Mace Aviation.

An NTSB review of this aircraft's airframe logbook revealed that an entry had been made on August 8, 1996, briefly describing the repairs to the hangar-damaged structure, followed by the statement, "Also mods listed on 337's, outboard leading edges installed with 15 gallon fuel bladders, and Baron B-58 wingtips with 12 gal of fuel. Fuel system mods are all Beechcraft manufacture." A series of NTSB discussions with the FAA determined that because there were no further 337's related to the fuel system and wing major alteration submitted after October 23, 1997, and because there was no FAA Field Approval entry in Block #3 of the Form 337 submitted as part of the original STC packet, there was no associated FAA Form 337 on file with the FAA at the time of the accident.

In addition, an NTSB review of the data provided to the aircraft owner by Mace Aviation revealed that there was no FAA Approved Airplane Flight Manual Supplement or Pilot Operating Handbook Supplement related to the changes created by this major alteration (although there was a detailed supplement for the previously altered fuel system related to the turbine conversion). There was also no data provided to indicate that flight tests had determined that this altered fuel system was able to provide adequate fuel to the engine at all power settings, flight attitudes, and fuel loads. It was also noted that the plaque/sticker located near the fuel ON/OFF selector valve still indicated that there were 74 gallons of useable fuel in this system, instead of the correct figure of approximately 130 gallons. Also, a further review revealed that although Mace Aviation stated that the fuel gauges had been overhauled and calibrated at the time of the repair/alteration, there was no record of the low fuel quantity light actuation system being calibrated since the current owner acquired the aircraft (although in October of 2002, Cutter Aviation, in Phoenix, Arizona, determined that at some point prior to their respective tanks becoming fully empty, the subject lights would illuminate).

MEDICAL AND PATHOLOGICAL INFORMATION

According to Douglas County Autopsy Report D-2003-65, the pilot-in-command died as a result of multiple skull fractures and intracerebral hemorrhages due to blunt forces impact to the head. The manner of death was classified as accidental.

A forensic toxicology examination was performed by FAA's Bioaeronautical Research Laboratory, and the pilot's blood indicated negative for carbon monoxide, cyanide, and screened drugs. No ethanol was detected in his vitreous fluid.

TESTS AND RESEARCH

As part of the investigative process, the Shadin Miniflo Fuel Management System panel-mount fuel flow indicator (P/N 912045T-D, S/N 9201), and the fuel line-mounted Shadin fuel flow transducer (P/N 660526, S/N 013) were removed from the aircraft and shipped to the custody of the FAA Principal Avionics Inspector (PAI) responsible for the oversight of the certificate held by Shadin Company Incorporated. The PAI then took the components to the Shadin Company test facility where the 42.8 kPPG (42,800 Pulses Per Gallon) transducer was tested on a fuel flow test bench for a duration of 141.7 seconds. During that period 1.21842 gallons was flowed past the transducer impeller, which recorded 51,957 pulses. These results yielded a calculated K-factor or 42.643 kPPG, or a difference of 0.4% from 42,800 kPPG, which is within the allowed plus or minus 2% deviation. In addition, the retrieval of the data in the non-volatile memory of the panel-mounted unit, which had a 46.8 kPPG sticker on its top surface, revealed a "Fuel Used" of 392 pounds, a "Fuel Remaining" of 220 pounds, a "Full Fuel" load number, which is normally entered into the unit soon after the unit's initial installation in the aircraft, of 900 pounds, and a K-factor correlation number of 46.8 kPPG. During a directed "self-test" the unit indicated "Good." This was followed by a series of function switch tests, during which no anomalies were found, and where it was confirmed that the P-factor in the unit was 6.7 (pounds per gallon of jet fuel). After these indications were read, in order to test the unit's current ability to calculate "Fuel Remaining," the unit's parameters where changed to a fuel load of 600 pounds, and then it was sent an indication of a fuel flow of 300 pounds per hour (30 pounds for every 6 minutes/one tenth of an hour). At five minutes, the "Fuel Remaining" indication was 574 pounds (a one pound deviation). At 10 minutes the deviation remained at one pound (549 verses 550), and at the 12 minute test termination point, the deviation still remained at one pound (539 verses 540).

Although these tests did not find any evidence of an operational malfunction or anomaly in the panel-mount unit or the transducer itself, it did reveal that the 46.8 K-factor manually entered into the panel unit and recorded on the aforementioned sticker (an action normally taken by the unit installer) did not match the 42.8 K-factor of the transducer (a requirement for accurate function of the system). It was further determined that this discrepancy would give a "Fuel Used" indication on the panel unit that was approximately nine percent (9%) less than the actual amount of fuel being used by the engine. In addition, a review of the records previously collected from Cutter Aviation, revealed that their avionics shop had installed the panel mount unit in this aircraft on December 11, 2002. The Cuter records show that at the time that the Miniflo panel unit was installed, an older version Shadin panel-mount unit was removed, and the new unit was connected to the 42.8 kPPG transducer of the older system. Although this is an acceptable practice, the installer would then have to make a correct manual entry of the 42.8 K-factor into the panel unit. Instead, a K-factor of 46.8 had been entered. Based upon this information, the IIC contacted the Cutter Aviation avionics shop, and after explaining the findings, asked if there was any documentation of the K-factor entry, or any explanation of why there was a discrepancy between the transducer and the panel unit K-factors. As of the date of this report, Cutter Aviation has not been able to supply any documentation of the K-factor entry, or any fact-based explanation of exactly why there was a discrepancy.

A review of the Shadin Miniflo Digital Fuel Management System Operating Manual, and in-depth discussions with Shadin technical personnel, revealed that the system is a fuel flow measuring system, and not a fuel quantity sensing system. It is not capable of measuring and subsequently displaying the quantity of fuel in any tank/s. It instead measures total fuel flow to the engine, and then electronically subtracts the amount of fuel used from the total fuel load number that was manually entered into the system by the pilot. The system then displays a calculated "Fuel Remaining" amount based upon the subtraction the Miniflo has performed. The total fuel load amount entered into the system by the pilot can be either a pre-selected full-fuel load amount (900 pounds for this aircraft), or any amount the pilot chooses to enter that is less than the full-fuel load. The system itself cannot account for fuel leakage or fuel theft, and manually entering an increase or decrease in the fuel remaining, whether intentional or not, will reset the fuel used value to zero. According to the Preflight Procedures section of the Miniflo Operating Manual, "A visual inspection and positive determination of the usable fuel in the fuel tanks is a necessity."

ADDITIONAL DATA AND INFORMATION

According to records obtained from several fixed-base operators (FBO's), the aircraft took on 115.91 gallons of Jet A fuel on May 17, 2003. Then on June 13, while in Portland, Oregon, the aircraft was fueled with exactly 60.00 gallons of Jet A fuel. During its last refueling, which took place on June 15, in Olympia, Washington, once again exactly 60.00 gallons of Jet A fuel was added to the tanks. According to the private pilot who was in the left seat during the accident flight, and who was also present during the subject last refueling, the aircraft's tanks were topped off during the refueling at Olympia. Reportedly, after that refueling, the aircraft was flown directly back to Vashon Island (approximately 0.3 of an hour). It then flew 1.4 hours on June 16, and 1.3 hours on July 1. On the flight prior to the accident flight, the aircraft flew from Vashon Island to Cashmere (approximately 0.7 of an hour), for an approximate total time since the last refueling of 3.7 hours, and a minimum of six takeoffs. Due to the nature of the flights on June 16 and July 1, an accurate average fuel flow could not be determined, but according to the subject private pilot, who had accompanied the owner in this aircraft a number of times before, the normal cruise fuel flow was around 170 pounds per hour, and the takeoff fuel flow was about 200 pounds per hour.

Based upon the aforementioned fuel flow figures, and the fact that after the major alteration was completed there was no amendment to the Pilot's Operating Handbook that would have indicated to the pilot that the amount of unusable fuel in the system had changed from the previously published amount of seven gallons, the NTSB calculated the approximate fuel remaining the pilot should have expected at his destination. In the two-minute climb, the aircraft should have consumed approximately one gallon of fuel (consumption at .4975 gallons per minute for two minutes). For the remaining distance of about 11 miles after level off, the aircraft should have consumed approximately two gallons of fuel (consumption at .4179 gallons per minute for five minutes). With the additional consumption of approximately one-half gallon of fuel from start up to takeoff, the total expected fuel used for this flight would be approximately three and one-half gallons. Therefore, based upon the left fuel tank gauge indication at start up (15 gallons), the pilot should have expected to arrive at his destination with approximately four and one-half gallons remaining (15 total, minus 7 unusable, minus 3.5 en route). This would have given him approximately 11 minutes of fuel remaining at cruise power.

Although with an unaltered fuel system, the pilot should have expected approximately four and one-half gallons of fuel remaining after landing, he had no way of determining the extent to which the unapproved fuel system major alteration, especially the routing of the excess fuel returned to the wing tanks system, could have altered the total amount of unusable fuel.

On January 5, 2005, the aircraft wreckage was released to Mr. Mark H. Smith, a representative of the family.