On April 30, 2005, about 1715 Alaska daylight time, a tundra tire equipped F-19 Taylorcraft, N3691T, sustained substantial damage following a loss of engine power and loss of control during the takeoff/initial climb from runway 36 at the Chena Marina Airpark, Fairbanks, Alaska. The commercial certificated pilot was fatally injured, and the sole, pilot-rated passenger, received serious injuries. Visual meteorological conditions prevailed for the 14 CFR Part 91 personal cross-country flight. The flight was operated by the pilot/airplane owner and his wife, who was the pilot-rated passenger. The intended destination was Bettles, Alaska, and a VFR flight plan was filed.

The NTSB investigator-in-charge (IIC) interviewed several witnesses at the accident site and via telephone. They all related essentially the same information. All the witnesses were either at the Chena Marina airstrip, or along the perimeter road system north of the airstrip. The witnesses stated the airplane made a normal takeoff run from runway 36, and had reached an altitude of approximately 300-500 feet above the ground, when the engine abruptly lost all power. The airplane made an almost immediate, steeply banked turn to the left, pitched nose down, and entered a left turn spin, of one to two turns duration. The spin stopped, and the witnesses described a partial recovery, with the airplane's nose attitude starting to return to level prior to impact. One of the witnesses, who was closest to the airplane when it lost power, stated that she was unsure if the engine actually stopped, as it sounded as if the engine had continued to run, or idle, at a very low power, and may have had a momentary burst of power prior to impact. The airplane crashed in a boggy area of low brush and small, under 20-foot high, trees. It came to rest upright, with the nose of the airplane in a pond, and the engine under water.

The airplane's co-owner/passenger was initially interviewed by the IIC via telephone while she was hospitalized, on May 9, and in person at the hospital, on May 17. The IIC also had additional telephone conversations with her after her discharge. She stated that she has a private pilot's license, and is qualified to operate the accident airplane. She indicated that on the day of the accident flight, she preflighted the airplane, and found no indications of any mechanical problems or fuel contamination. She said the airplane was last fueled on April 27, with 100 low lead fuel from a fuel vendor, Alaska Air Fuel, at the nearby Fairbanks International Airport. Shortly before the accident flight, she drained fuel samples from both wing auxiliary tanks' quick drain sumps and the main header tank. She recalled that the auxiliary tanks were full, and the main tank had 3/4 or more. Her husband was the pilot-in-command, and prior to takeoff, he performed a pretakeoff engine run-up and systems check with no apparent mechanical problems. He then back-taxied about 2/3 of the way toward the takeoff end of runway 36 (4,700 feet long), and began the takeoff roll. The takeoff and initial climb were normal, until they had passed the departure end of the runway, and were about 300-500 feet above the ground, when the engine suddenly lost power, as if the engine throttle had been abruptly pulled to the idle. She said her husband immediately made a hard, steep, left turn to the south, towards the departure end of 36, and as the airplane rolled into the turn, it entered an inadvertent stall, the nose pitched nearly straight down, and the airplane started a spin to the left. She was unsure how far the airplane rotated to the left, but stated that the rotation stopped, and a partial recovery to a nearly level attitude was accomplished prior to crashing into a swamp. She also said that the engine never completely quit running, and it may have regained some power just prior to impact.

The passenger stated that they had purchased the accident airplane in October, 2003, and that it had an infrequent yet recurring problem of power losses during takeoff/initial climb. She reported that it had occurred once while she was flying the airplane, on or about October 2, 2004, while taking off from Fairbanks International Airport, with her husband as a passenger. She said the loss of power scenario was almost exactly the same as the accident flight, and about 300 feet above the ground, the engine suddenly lost power, but never completely stopped. Her husband immediately took the flight controls, and made an abrupt, hard turn to return to the airport. He was able to make a successful landing on a taxiway without damaging the airplane. After landing, engine power returned, and the airplane was able to taxi under its own power to the ramp area, where it was run-up to full power, without any apparent mechanical problems. They left the airplane at the airport, and it was recovered later by her husband. The passenger stated that she was aware of possibly two other loss of engine power events that happened while her husband was flying the airplane. On those occasions, like the incident at Fairbanks International Airport, he was able to return and land at the departure airstrip.


The commercial pilot was fatally injured. His shoulder harness failed at the point where the single anchor, or tail strap, attached to the V-shaped shoulder restraint straps. The passenger received serious injuries, and was hospitalized for several weeks. Her shoulder restraint system did not fail. Additional information regarding the restraint systems for the pilot and passenger are contained in the Tests and Research portion of this report, and in the Public Docket for this report.


The pilot held commercial privileges for airplane, single engine land and instrument ratings, and private privileges for airplane, single-engine sea. According to his personal flight log, he had accumulated approximately 1068 flight hours at the time of the accident. He had logged approximately 36 hours in the accident airplane. His second class FAA medical was issued October 19, 2004, with the limitation that he must wear correctives lenses for distant vision.

The pilot was employed by Colorado State University as a biologist, and was working under contract to the United States Army at Fort Wainwright, Alaska. As part of his duties as a biologist, he was authorized to fly certain government-owned airplanes, and on March 24, 2005, had successfully completed a recurrent Pilot Evaluation/Qualification Check written and check flight in a Aviat Husky A-1, administered by a Department of the Interior (DOI) check pilot and FAA certificated flight instructor. As part of the flight check, according to the evaluation form and an interview with the check pilot/instructor, the accident pilot successfully performed stalls, and one low level simulated loss of engine power and forced landing approach to a meadow. The check pilot/instructor indicated that the accident pilot did not attempt to complete a 180-degree turn during the maneuver, and appropriately selected an emergency landing site ahead of the airplane.

On September 12, 2004, the pilot was awarded a certificate from the FAA for completing a safety seminar, "Practical Risk Management For Pilots" FAA-Wings-024. The pilot had also completed a flight course in an aerobatic airplane, titled: "Emergency Maneuvering-Modules 1, 2, 3" from CP Aviation, Santa Paula, California, on November 24, 2002. According to the CP web site, modules 1, 2, and 3, cover stall spin awareness, in-flight emergencies, and basic aerobatics, including unusual attitude recoveries.


The airplane was a single-engine, 1977 model year Taylorcraft F19 tailwheel airplane, with two seats in a side-by-side configuration. It was equipped with a Continental O-200, 100 horsepower-rated engine. At the time of the accident, available logbook information, tachometer and Hobbs meter readings indicated that the total service hours on the engine was 1,545, and the service hours since major overhaul were 561. The airplane's engine maintenance log book covers the period from April 15, 1989, to the last entry, the date of the annual inspection, February 15, 2005. The first entry notes: "Previous Eng[ine] logs lost. Engine removed for major overhaul. Engine time is same as airframe tach time at time of removal."

The airplane had been modified with a climb pitch propeller, and retro-fitted shoulder harness and seat belt combinations at both the pilot and passenger seats. The airplane did not originally come equipped with shoulder harnesses, and there was no logbook entry regarding their installation. Additional shoulder harness information is contained in the Tests and Research section of this report.

The pilot's logbook indicates two prior power loss events, on August 7, 2004, when he was flying the airplane, and on or about October 2, 2004, when his wife was flying and he took the controls and landed at Fairbanks International Airport. His logbook entry for the August 7 loss of power is followed by the phrase "carb ice?". A witness to the August 7 event at the Chena Marina airstrip said that he was on the road to the south of the airstrip, and the airplane passed nearly overhead, when he heard the engine cut out, and the airplane banked very steeply, like a "wing-over", and turned towards the runway. The witness, a student pilot, said he didn't think the airplane would be able to make it back to the runway, but it did. There's also an entry dated October 8, 2004, that alludes to a loss of engine power at Fairbanks International the preceding weekend, (the flight with his wife) and that a magneto was replaced. The pilot's wife believes there may have been one more loss of engine power that occurred while her husband was out hunting in the Fall of 2004, during takeoff from a remote airstrip in the Alaska Range. The pilot's logbook reflects that he went hunting with the airplane in September 2004, but it does not indicate that there was a loss of engine power during that period.

The airplane's last annual inspection was completed at the Chena Marina Airpark by Apex Aviation on March 15, 2005. During that inspection, the carburetor was opened, the float assembly adjusted as needed, reassembled, and the fuel system flow tested with no observed anomalies. The aviation mechanic who completed the work and certified the airplane as airworthy, indicated that the engine was run-up, and no discrepancies noted.

A review of the airplane's engine logbook entries and invoices from Apex Aviation, disclosed that the airplane's engine had repairs accomplished on September 13, 2004, to resolve a rough running engine. The discrepancy, or "squawk sheet" that accompanied the invoice noted, in part: "Will not run-up over 1500 rpm. Trouble shoot eng[ine]. Run up, found cold cyl[linder] number 1...found stuck valve... ." According to the work order, the valves for that cylinder were removed and repaired, and the post-repair engine run was satisfactory. Another invoice from Apex Aviation, dated November 30, 2004, for work performed October 8, noted, in part: "Cruise to FAI [Fairbanks Airport] and troubleshoot rough eng[ine]. R&R R/H mag [remove and replace right-hand magneto] due inop, time to engine, ops chk."

On May 12, 2005, the NTSB IIC had a telephone interview with a friend of the accident pilot. During the interview, the friend related that during a conversation with the accident pilot on April 28 or 29, 2005, the accident pilot indicated that he was still concerned about the loss of engine power events, and thought they might be related to fuel flow problems.


Visual meteorological conditions prevailed. Fairbanks International Airport, about a mile southeast of the accident site, was reporting at 1739, about 19 minutes after the accident: clouds, few at 2,000 feet; 10 statute miles visibility; temperature 71 F, Dew point 23 F; wind, 110 degrees at 12 knots, with gusts to 17 knots.


The Chena Marina Airpark runway has a gravel surface, and is approximately 4,700 feet long, and 50 feet wide. It is located to the west of and immediately adjacent to the Chena Marina Seaplane waterway, at an elevation of 427 feet msl. The Chena Marina complex is privately owned and maintained. The area to the north of the airport, in the general vicinity of the accident site, is essentially flat, with minor relief, a marshland swamp of approximately 2 acres, a gravel roadway with lightly traveled, smaller connecting roads, and scattered small trees and brush. The area north of the accident site had been mostly cleared in preparation for proposed housing sites.


An on-site inspection by the NTSB IIC, with FAA aviation safety inspectors from the Fairbanks Flight Standard District Office assisting, commenced on May 1, at 0800, and was completed on May 2.

The airplane came to rest about 175 yards north of, and about 75 yards west of, the north end of runway 36. It was upright, at the north edge of a shallow pond located in a marsh, with the engine and propeller mostly underwater. The pond was approximately 3 feet deep at its deepest, and about 30 feet in diameter. The longitudinal axis of the airplane was on an approximate 180 to 360 magnetic degree orientation, with the nose of the airplane pointing to 180 degrees, towards the departure runway. There were a few small trees, approximately 2 to 3 inches in diameter, just aft of the wreckage, that were bent over, and may have been struck by the airplane prior to impact. There was no significant soil disruption at the airplane's point of rest.

All major components of the airplane were accounted for at the immediate crash site. The portion of the fuselage aft of the wings' trailing edge was intact with little to no deformation. The wings were bent down, nearly uniformly, resembling a partially inverted "V", and the trailing edges of both wings were detached from the fuselage. Both wing spars were broken near their respective lift struts. The left aileron control actuator rod had fractured. The lift struts for both wings were bent and compressed. The main landing gear, which was equipped with over-sized "Tundra" tires, was collapsed. The top of the cockpit section had been modified with a skylight, and a portion of the top of the cabin had been removed by rescue personnel. The cockpit bench seat frames for the pilot and passenger were distorted. The pilot's seat frame was pushed upwards, and the passenger's was compressed downwards. Both flight control wheels were extended and bent towards the floorboards. The engine throttle was pulled out to approximately mid-span, and captured by impact damage. The mixture control was in, and the fuel cut-off was out. The carburetor heat was in (OFF). The throttle and mixture cables were attached at their respective ends. Flight control continuity was established from the ailerons to the control wheel, and from the rudder to the rudder pedals. The elevator cables were followed to the cabin/cockpit area, but could not be moved due to cockpit deformation/crushing. The two-bladed propeller was nearly straight, with only slight forward bending at one tip. The leading edges of the propeller did not display any significant gouges, chord-wise scratches, or other impact damage.

The airplane was equipped with one, 12-gallon main header tank located aft of the engine firewall, and two, 6-gallon wing tanks, one mounted in each wing. The fuel system only allows fuel to flow from the main header tank to the engine; the wing tanks are used to replenish the header tank via gravity flow as needed. The wing tanks' respective cockpit valves to the header tank were found in the "off", or closed, position. The right wing tank tested positive for water contamination at the lowest portion of the tank, when a paste that reacts to water was inserted into the tank on a stick. The left tank did not react to the water paste. Both tanks appeared to be full, or nearly full, with what resembled and smelled like 100 octane low-lead aviation fuel. The main header tank had been breached during the accident, and the breach was under water, and thus the tank could not be tested for contamination. According to the passenger, the main tank was about 3/4 full at the time of departure. The fuel vent tube to the main header tank was intact, and appeared clear. When the IIC blew into the vent tube, an odor of gasoline could be smelled, and bubbles were seen and heard in the header tank.

The wreckage was removed from the site on May 1 via helicopter, and taken to a nearby maintenance facility for further inspection and storage. The engine was removed, and taken to another maintenance facility for inspection and partial disassembly.


An autopsy on the pilot was performed by the State of Alaska, Office of the State Medical Examiner, 4500 S. Boniface Parkway, Anchorage, Alaska, on May 3, 2004. The autopsy final diagnosis lists the cause of death as multiple blunt force/impact injuries due to an airplane crash. No evidence of any significant disease was discovered.

Toxicological tests on specimens taken from the pilot were completed by the FAA's Civil Aeromedical Institute laboratory in Oklahoma City, Oklahoma. The tests were negative for all drugs tested except for Ephedrine and Phenylpropanolamine, which were found in the pilot's urine. Ephedrine and Phenylpropanolamine are drugs commonly found in over the counter cold medications.


The cockpit/cabin structure was principally intact, with minor lateral and vertical compression. As noted, the bench seat frames were distorted, and the left, pilot's seat, was pushed slightly upwards, and the right, passenger's seat, was compressed downwards. The pilot's control wheel was trapped by impact with the wheel extended, and the control wheel handles were distorted and pushed forward. As previously addressed, the pilot's shoulder harness failed at impact. A review of the pilot's autopsy report by the NTSB's Medical Officer physician, disclosed that the pilot had more than one fatal injury, and at least one of the fatal injuries was consistent with high vertical deceleration forces.


Engine and related components:

On May 1 and 2, a partial engine disassembly and inspection was conducted at Airframe Alterations, a maintenance facility at the Chena Marina Airpark, under the direction of the NTSB IIC. Present and participating were three FAA airworthiness inspectors from the Fairbanks Flight Standards District Office.

The engine, a four cylinder Continental Motors O-200A, serial number 255707, was hung from a chain hoist for initial inspection. There were no indications of any preimpact catastrophic engine failure. The carburetor had been knocked off the intake manifold by impact, but it appeared intact and undamaged. The intake bosses for the number 1 and 3 cylinders were broken off/cracked. The spark plugs were removed, and were observed to have normal wear and deposits when compared to a Champion Check-A-Plug chart. The engine rotated freely, and was developing compression at each cylinder (low, 64/80, high, 74/80). The accessories drive gears rotated when the crankshaft rotated, and all valves and rocker arms moved sympathetically with the crankshaft. All fuel supply lines and screens/filters were checked and found serviceable. Magneto timing was checked and found correct. The spark plug leads sparked when the magneto was operated by hand cranking the propeller. The exhaust mufflers were examined to ensure that the flame spreader cones were intact, and not blocking the exhausts. They were intact, and did not block the exhaust stacks.

On May 26, the engine was completely disassembled under the direction of two FAA Fairbanks Flight Standards airworthiness inspectors. No evidence of any preimpact mechanical malfunction was discovered.

The carburetor, a Precision MA-3SPA, serial BE2215512, was disassembled. Swamp water and associated debris was in the bowl. The white polymer Delrin float, part number 30-804, code 0274, tolerances were checked, and were within factory specifications. The float moved freely, was not cracked, and did not contact the sides of the float bowl. Further disassembly of the carburetor disclosed no evidence of any preimpact anomaly, other than the float's needle valve retraction forks, which were bent up, away from the body of the float. A new, exemplar float was examined, and the retraction forks were noted to be parallel to the body of the float. The accident float's retraction forks are stainless steel, and rigid. They were bent away approximately 1/32 of an inch from the parallel position (see photographs in report docket). The FAA airworthiness inspectors, all certified aviation mechanics with inspector authorizations, noted that they were concerned that the bent retractor forks may not have exerted sufficient force on the needle valve when the float dropped to unseat the needle valve. They noted that if the needle valve was not unseated, the lack of fuel in the carburetor bowl could result in possible fuel starvation to the engine, and a loss of engine power.

The carburetor was cleaned of the swamp debris and water, partially reassembled, and a non-permeable dye was placed along the edges of the polymer float. The carburetor was shaken fore and aft, up and down, to see if any of the dye transferred from the carburetor float, indicative of the float touching the bowl sides. No dye was seen on the bowl sides.

The carburetor was reassembled, and sent to the manufacturer, Precision Airmotive Product Corporation in Everett, Washington, for further testing. On July 12, 2005, under the direction of a Seattle, Washington-based NTSB air safety investigator and an FAA aviation safety inspector, the carburetor was flow tested, disassembled, and inspected. The inspection disclosed the bent retractor forks, and noted: " This caused the needle to not make contact with the adjustment tab, but instead be controlled by the retractor clip." According to the FAA Inspector present, the technician performing the inspection noted that the bent retractor forks were abnormal. However, three more similar carburetor assemblies were opened by Precision staff at the request of the NTSB investigator, and according to the FAA inspector, all three had the retractor forks bent up slightly from the float body. The FAA airworthiness inspector who participated in the on-site field investigation at Fairbanks, has issued a Malfunction Or Defect Report (MDR) regarding the bent retractor forks. A copy of that MDR is included in the report docket.

A flow check of the carburetor was performed, and was within the manufacturer's acceptable parameters, with the exception of an excessive flow at the idle cut-off position.

Shoulder Harness Assemblies:

The airplane did not come from the factory equipped with shoulder harnesses, and there was no log book entry indicating when they were installed, or who the manufacturer was (as with the engine logbooks, the airframe logs prior to 1989 were reportedly lost). There were no manufacturer's markings or labels on either of the harnesses, nor the corresponding seat belts. The surviving passenger/owner said the belts and harnesses were the same as when they purchased the airplane. During an interview with one of the previous owners, it was disclosed to the IIC that he thought the harnesses were installed about 1982. He was uncertain, but thought they came from a national aviation supplier, Wag Aero. Wag Aero was unable to identify the harness or the harness material as being one they had distributed.

According to the FAA, there is no inspection criteria specifically for shoulder harness assemblies installed or manufactured prior to 1980. Given the age of the airplane (1977), without markings on the harnesses, or a logbook entry delineating the date of installation/manufacturer, inspection of the harness during maintenance of the airplane would not be required, other than for a generic, general condition assessment and security of attachment.

The pilot's shoulder harness webbing was shaped like an inverted "Y", with the bottom, or tail of the "Y" acting as an anchor strap that was threaded through and around a steel bar that ran across the top of the cabin, perpendicular to the long axis of the airplane. The anchor bar was welded into the top of the airframe tubular structure, and was approximately 5 degrees aft of a vertical line drawn from the top of the pilot and passenger's seat frames. The two arms of the "Y" webbing draped over the pilot's shoulders, and attached to the seatbelt coupler. The "Y" portion of the harness was comprised of a single strap, that was folded over on itself in midspan to create an angle that would allow the harness to drape over both shoulders. The folded segment was then attached to the tail anchor strap by covering the fold with a fold of the anchor webbing, and securing it all with thread in a box stitch pattern. A visual inspection of both the pilot's and passenger's harnesses did not reveal any signs of excessive wear, although both appeared lightly soiled and faded.

The pilot's and passenger's shoulder harnesses were sent to the NTSB Materials Laboratory for testing and failure analysis. The report notes, in part, that the pilot's harness failed at the stitching where the anchor strap was sewn to the junction of the shoulder restraints. The webbing material was tensile tested to failure, albeit without the ability to test the box stitching at the attachment point. The material was found to fail at 3,840 pounds, exceeding the minimum of 2,250 pounds as specified in Technical Service Order C22f for webbing material, which was applicable when the harnesses were reportedly purchased. A test was also performed on the passenger's harness. The webbing material again failed at a strength exceeding the minimum requirements, but the box stitching joint failed at 995 pounds, significantly below the minimum of 1,500 pounds.

The report also notes that that pilot's anchor strap was 1.75 inches shorter than the passengers, and while the end of the passenger's anchor strap was cut with an instrument consistent with a hot knife frequently used to cut webbing, the end of the pilot's anchor strap displayed indications of a previous overload tensile failure. A copy of the engineer's Materials Laboratory Factual Report is included in the docket of this report.

Ignition Switch

The ignition switch, a Bendix five-position switch (OFF/RIGHT MAGNETO/LEFT MAGNETO/BOTH MAGNETOS/START), part number 10-357200-1, was removed by the IIC and the circuitry tested with a multi-meter for continuity and proper grounding. The switch was also disassembled and inspected for any signs of arcing or failure. The switch operated properly at all positions, and no evidence of any internal mechanical problem was discovered.


The wreckage was released to Alaska Air Service, an agent for the owner's insurance company, on May 2, 2005. The NTSB IIC retained the Precision carburetor, Bendix starter switch, and shoulder harnesses from the pilot and passenger stations for additional testing. Those items were released to Alaska Air Service on March 16, 2006.

Use your browsers 'back' function to return to synopsis
Return to Query Page