On April 28, 2004, at 1503 eastern daylight time, a Beech A36, N3243P, was substantially damaged during a forced landing at Bedford/Laurence G. Hanscom Field (BED), Bedford Massachusetts. The certificated commercial pilot was not injured, and the passenger received minor injuries. Visual meteorological conditions prevailed, and no flight plan had been filed for the local post-maintenance test flight, which was conducted under 14 CFR Part 91. The flight was the first flight after the installation of an overhauled engine. The pilot was the owner of the airplane, and the passenger, who was the president of the maintenance company that installed the engine, Nagle Aircraft Incorporated, was onboard to monitor its performance. According to the pilot, he started the engine about 1445, and subsequently taxied the airplane to runway 29. Engine run-up was normal, and the takeoff was "normal with [the] engine running smooth." After rotating the nose at 78 knots, the pilot crabbed into a 30-degree crosswind, then accelerated to 110 knots. About 200 feet above the ground, the pilot raised the landing gear, and about 400 feet, he noticed that the engine rpm had dropped to 2,550. "Moments later," the engine failed and the tachometer went to zero. The pilot lowered the airplane's nose to maintain 90 knots, and determined that there wouldn't be enough runway remaining to land straight ahead. He turned the airplane to the left, to aim for taxiway W, and unsuccessfully attempted an engine restart with the "electric fuel pump." The pilot then turned the airplane further to the left, to cross runway 05 perpendicularly, and aimed for a gap within a line of parked airplanes. The airplane touched down just after crossing runway 05, all three landing gear collapsed, and the airplane came to a stop in the midst of the parked airplanes without striking them. Two Federal Aviation Administration (FAA) inspectors initially examined the airplane. According to the president of Nagle Aircraft, "the FAA, immediately after the accident, clearly noted that there was no fuel staining identified within the engine compartment and no fuel odor present in and around the engine compartment." Both inspectors subsequently replied that they did not recall the statement. However, they did note that they could not inspect the engine compartment closely because the left cowling was jammed and could not be opened. On May 17, 2004, the airplane was examined under Safety Board direction. Among the items examined, the cockpit fuel boost pump lever was found in the "high" position, and once the cowling was removed, there was no evidence of fuel spillage or spray within the engine compartment. In addition, torque paste on the engine-driven fuel pump's return line B-nut was found to be about 1/4 turn counter-clockwise (loosened) in relation to the torque paste on the receiving end of a 90-degree elbow fitting. When the B-nut was hand-tightened, the torque paste on both fittings were aligned. When the B-nut was further tightened with a wrench, the paste was again misaligned, but about 1/4 turn further clockwise. Multiple engine runs were subsequently conducted. When the B-nut was hand-tightened to where its torque paste was aligned with the 90-degree fitting torque paste, and also when it was fully tightened with a wrench, the engine ran smoothly. When the B-nut was loosened to nearly where it was found during the initial examination, the engine quit. During further testing, while the engine was running, the B-nut was loosened to a point where the engine started to quit. The auxiliary fuel pump was then turned on, to the "high" position, and the engine ran normally, but fuel was sprayed toward the left side of the engine compartment from the fitting. When the fitting was further loosened, the engine quit and could not be restarted, even with the auxiliary fuel pump operating. The testing was repeated several times with the same results. The return fuel line, which was new, was subsequently removed and visually examined, with no anomalies noted to either the hose or the B-nut threads. There were also no anomalies noted to the receiving, 90-degree elbow fitting threads, which were also new. Subsequent to the examination and testing on May 17, 2004, personnel from Nagle Aircraft conducted another series of tests on a different airplane, but with a similar engine and fuel supply system. According to the company president, the fuel return B-nut fitting was loosened to the point where it was found on the accident airplane, and the engine ran numerous times without quitting. When the fitting was loosened sufficiently to allow for fuel leakage (fuel was then being sprayed in the compartment), the engine quit. During all of the company's subsequent tests, when the engine began to quit, the auxiliary fuel pump was placed in the "high" position, and the engine continued to run.