On October 8, 1996, at 1647 hours Pacific daylight time, a Richard's Lancair IVP, N477RM, collided with trees in a lemon orchard after a loss of power near Camarillo, California. The airplane was destroyed and the airline transport pilot and passenger received serious injuries. The airplane was being operated by the pilot as a personal flight. The flight originated from the Camarillo airport about 1615. Visual meteorological conditions prevailed. The experimental airplane was built from a kit by the pilot and was completed on September 10, 1996. The airplane had been flown about 12 hours since completion and was completing Phase I of its flight testing. The pilot indicated that there were 32 gallons of fuel onboard at the last takeoff, all of which was in the left tank. The airplane's fuel selector valve had been modified from the original design. The fuel selector valve's stem was machined to incorporate a both position. The kit manufacturer does not recommend a both position for its low wing designed aircraft. The pilot indicated he lost engine power while cruising at 4,500 feet above sea level with the fuel selector valve in the left position. He was unable to restart the engine and landed the airplane off the airport. The pilot stated, "I had no recourse but to land the aircraft off the airport in the most suitable area observed in the glide time remaining." The engine and fuel system components were removed from the airplane and sent to manufacturer's facilities in Mobile, Alabama, for operational tests. The engine oil sump, pickup tube, starter adapter, and engine mounts were damaged as a result of impact and were replaced. The engine was installed in a production test cell and serviced with oil. The engine started immediately and was run to full throttle after warm up. An oil leak was observed at the number 5 exhaust push rod tube. According to the manufacturer, the bent tube was a result of being hit by the engine mount that broke during impact. The push rod and tube were replaced and the engine was started again. There was no evidence of any further oil leak. The engine was run to full throttle and did not develop power. The engine was shut down again and the paint on the number 6 cylinder was found scorched. Review of photographs taken before the engine was shipped to the manufacturer revealed evidence of a scorched area on the cylinder. The number 6 cylinder was removed and replaced along with the piston and rings. The replaced cylinder and piston exhibited heavy scoring. The engine was started again and achieved full rpm and power. The engine was at full power to establish a base line for planned tests of the fuel system. After the engine stabilized, the crankshaft flange broke, allowing the club propeller to separate. According to the manufacturer, a fatigue crack initiated from a prior crack on the crankshaft outer diameter. This was generated when the crankshaft flange was bent during the accident impact sequence. Since the engine was no longer available to test the aircraft fuel system, the fuel selector valve and attached plumbing was taken to a flow bench at the engine manufacturer's facilities. The system was installed on the flow bench to record inlet and discharge pressures at various fuel flow rates. The test was performed to determine if air would leak past the fuel selector valve if one of the inlets was dry, simulating the fuel configuration of the accident airplane. Air bubbles were observed through the vapor ejector line of the flow bench when line opposite to the selected fuel tank was uncapped simulating the empty tank. It was noted as the air bubbles increased the fuel flow would drop off. The fuel selector valve was disassembled. The valve had three O-rings that separated the fuel feed ports from the fuel return ports. The color of the center spindle was gray and there was no seal separating the right, left, and both ports of the selector valve from each other. The fuel selector valve center spindle was measured and the outside diameter was 1.0001 inches. The center bore was 1.0053 inches, leaving 0.0052 inches clearance between the spindle and the bore. Two other fuel selector valves from the same manufacturer that were modified to incorporate a both position were disassembled and measured. The clearances measured 0.0046 and 0.0043 inches. According to the engine manufacturer ,the scoring of the number 6 cylinder and piston was a result of high cylinder temperature from a lean fuel-air mixture. The lean mixture developed from aeration of the fuel through air bypassing the fuel valve from the empty tank. According to the person who modified the fuel selector valves, there was no machining done to the outside diameter of the spindle or the center bore. The person who modified the valves also performed tests to see if air or fuel would bypass the spindle using modified and unmodified fuel valves. The tests incorporated one modified valve and two unmodified valves. According to the modifier, the results revealed all three valves leaked and the dimensional clearance between the spindle and the bore was different in all three. Lancair publishes a newsletter for the builder of their kit airplanes. The Lancair News Letter of the first quarter of the year 1996, contained an article on the fuel selector valves. According to the article, ". . .early units had more leak by than deemed acceptable." The revised spindle was identified with a gray finish versus the black finish of the originals. In another Lancair News Letter for the fourth quarter of 1996, the fuel selector valve spindle is addressed again. The news letter announces, "There will be a 3rd generation spindle for the valve. The issue has been one of rotational friction getting too high for comfort. The version should eliminate the high friction situation. It does require some lathe work on the spindle." According to the news letter, "The significant difference between the spindles are: Increased O-ring glands cut into the spindle side to put less pressure on the O-rings."