NTSB Factual Report

On June 19, 2001, about 1045 eastern daylight time, a Grumman American AA-1B, N8957L, registered to a private individual, operating as a 14 CFR Part 91 personal flight, crashed while attempting a forced landing in the vicinity of Oak Hill, Florida. Visual meteorological conditions prevailed and no flight plan was filed. The airplane received substantial damage and the private-rated pilot and a passenger were not injured. The flight originated from Space Coast Regional Airport, Titusville, Florida, about 15 minutes before the accident.

According to the pilot, he had routinely switched his fuel source from the left tank to the right tank about 8 minutes before he started losing rpm and fuel pressure. He switched the fuel selector back to the left tank, but engine speed continued to decay to under 500 rpm, although it did not quit completely. He was at 1,200 feet msl, and did not have time to accomplish the entire "engine failure in flight" checklist. He did not attempt using carburetor heat. The aircraft was properly approved for the FAA supplemental type certificate for use of automotive fuel and 87 octane automotive fuel was being used at the time.

According to an FAA inspector, the pilot chose an unpaved road about the same width as the wing span, and on landing rollout, the right wing collided with mailboxes and a palm tree that separated the wing, (and landing gear) near the wing root. The right wing tank was compromised during the crash sequence, and evidence of fuel spillage was present. The right wing contained about 1 gallon of auto fuel and the left wing contained about 6 or 7 gallons of auto fuel.

During post crash examination of the Lycoming O-235-C2C engine at an aircraft maintenance facility with FAA oversight, all fuel and ignition components were inspected for possible causes of the symptoms reported by the pilot during the engine loss of power. A small amount of fine particulate matter was found in the 4 or 5 oz. of auto fuel drained from the carburetor bowl, although all filter screens in the fuel system were clean and unobstructed. The inspector characterized the contamination as, "...light to moderate muddiness, such as silt in suspension." The carburetor bowl assembly was removed from the carburetor, and an additional small amount of encrusted residue was visible. It was decided to continue the operational checks without removing the carburetor bowl contaminate. Two attempts were made to test run the engine, but after satisfactory starts, the engine would abruptly shut down of its own accord. On the third attempt, using throttle manipulation to prevent another shutdown, run time was adequate to operationally test the magnetos, and both tested normal. The indicated fuel pressure, with and without the boost pump on, was 4 to 5 psi, which is in the normal range. Removal and reinstallation of the carburetor to another engine for operational testing revealed normal performance of the carburetor. All components of the fuel and ignition systems were again re-examined, including the boost pump, the engine driven pump, the fuel lines, the fuel selector valve, both magnetos, the magneto p-leads, and the magneto switch. All components again tested normally.

The carburetor, a Marvel Schebler, model MA-3A, PN 10-3103-1, Serial No. Y-29-5001, was subjected to FAA repair station disassembly examination, with NTSB oversight. The carburetor bowl contained an amount of encrusted, tan-colored residue within its angular recesses, that when scraped and deposited on a napkin, amounted to about 1/4 gram. Testing of the carburetor on a fuel flow tester revealed the float level was properly adjusted. The main nozzle was the correct part number for its application, and the fuel flow rate was to specification. The venturi assembly was found misaligned with respect to its relationship with the main nozzle, but it was seated squarely within the carburetor throat into its machined mating surface and would not have been a factor as to correct fuel/air flow or engine performance.

Although the atmospheric conditions existing at the time of the original engine power loss put the flight in the regime of, "serious icing at glide power" on the icing probability charts, this line of investigation was not pursued since the engine exhibited the same power instability during postcrash testing at the maintenance facility.

NTSB Probable Cause

A partial loss of engine power while at a low altitude cruise due to fuel contamination, and the subsequent emergency descent and landing to a rural road resulting in collision with tree and mailbox post.