What happened
The aircraft was in cruise flight at 9,000 feet when the pilot performed a fuel switch from the right tip tank to the right main tank. Approximately two to five minutes after this transition, the engine stopped and the propeller began windmilling. The pilot attempted to restart the engine but was unsuccessful. While being vectored toward an airport, the pilot determined that altitude could not be maintained and elected to perform an emergency landing in an open field. During the landing, one of the aircraft's wings separated.
The investigation
An engine test run was conducted on March 19, 1999, under FAA supervision. Prior to the start, the fuel screen was removed from the fuel servo; investigators found fuel with minimal sediment and no water present. Fuel was also noted in the flow divider. While the mixture control showed free travel from stop to stop, the throttle travel was limited from idle to half throttle because the cockpit control had been bent.
During the engine run, the engine exhibited a rough running condition and emitted black smoke upon startup. The engine ran smoothly and the black smoke ceased only when the mixture control was leaned out to approximately half travel. At idle, the oil pressure gauge indicated minimum oil pressure. Further tests showed that advancing the throttle without leaning the mixture caused the engine to quit, and moving the mixture from lean to full rich caused the engine to run rough, emit black smoke, and die within seconds.
On June 14, 1999, a flow test was performed on the Precision Fuel Servo Injector (model RSA-10ED1, S/N 8791). The test revealed that the servo was running rich at the idle and "off" idle settings, though it remained within specifications at higher RPM settings.
Investigators also examined the condition of the component upon its arrival at the NTSB office in Miami. The unit had been shipped in an unprotected FedEx envelope with only plastic wrapping and no padding. Although the mixture control lever was found bent and jammed in the full rich position, it was determined that this damage occurred during transit from the FAA office in Atlanta to the NTSB office and did not affect the results of the fuel servo flow test.