Accident Overview
The accident occurred during a post-maintenance test flight when the engine lost power while the aircraft was in the traffic pattern, transitioning from base to final turn. The airplane subsequently collided with a building before reaching the airport.
Pre-Accident Maintenance History
Approximately 11 flight hours prior to the accident, the carburetor had been overhauled. During subsequent post-overhaul flights, an intermittent rich fuel/air mixture in the idle circuit was noted, preventing the engine from idling below about 1100 rpm without fouling and quitting due to excessive fuel. The carburetor was removed and returned to the overhaul shop, where the installed needle valve assembly was replaced with another identical new assembly. After reinstallation, fuel leaked from the carburetor when the main fuel shutoff valve was turned on, prompting another removal and return to the shop.
At the shop, the owner suspected a stuck or hanging float. Because the carburetor was equipped with an Advanced Polymer float—larger than the original brass or older composite floats—the owner carefully adjusted the float for lateral clearance between the float and the bowl wall, as well as between the float clip and the needle valve. After several test bench attempts to achieve a stabilized fuel level, the carburetor passed inspection.
On the day of the accident, the pilot—an Airframe and Powerplant (A&P) mechanic—retrieved the carburetor from the shop, installed it, and flew the aircraft for about 15 minutes. The accident occurred on the second test flight.
Post-Accident Examination
Examination of the engine after the accident revealed sooted spark plugs, consistent with a rich fuel/air mixture. The carburetor was functionally tested on a tilting fixture. A trace amount of fuel leaked from the discharge nozzle, and when the fixture was rotated to simulate a bank angle, fuel flowed freely from the nozzle. Tapping on the bowl stopped the flow. Operational testing disclosed that the carburetor was operating at an excessively rich setting at idle speed.
Disassembly showed the Advanced Polymer float was clean and intact with no damage. Float setting and bowl clearance were within specifications. The needle valve seat assembly met specifications upon inspection and measurement. The clearance between the float retractor clip and needle valve shoulder measured about 0.005-inch.
Carburetor Analysis
The pivot pin/shaft that hinges the float assembly was found to be tight within the inside diameter of the polymer float hinge points; however, the float and shaft combination rotated freely in the float bracket. Examination of the manufacturer's service manual revealed three types of floats: original hollow brass chambers, discontinued composite floats, and the larger Advanced Polymer floats. With brass or composite floats, typical clearance between float and bowl chamber was 0.081 inches; the polymer float reduced this clearance to about 0.031 inches. The original manual specified a minimum clearance of 0.005-inch between the float valve seat shoulder and the float valve retractor clip for brass floats, but a later instruction (E-955, dated 03/18/99) for polymer floats only required ensuring that clearance exists, with no published minimum.
During final assembly, the carburetor fuel bowl and throttle body are joined without visibility of internal clearances. The investigation measured free play in the accident carburetor: about 0.015-inch from rotating the bowl cover and fuel bowl halves with snug screws, and about 0.028-inch horizontal float centering free play from the hinge screws. Total sideways free play at the float tips measured about 0.229-inch. A bead of black transfer material was applied to the accident float flange to test for rubbing; after reassembly and rotation, black transfer markings appeared on the inside of the bowl walls, indicating contact. An improperly centered float that rubs on the bowl wall may affect float buoyancy and needle valve seating. Without proper clearance from the needle valve seat shoulder, positive fuel shutoff would be unlikely.