LAX02FA097 · Grumman American AA1B

The engine lost power in the traffic pattern base to final turn during a post maintenance test flight and the airplane collided with a building short of the airport. The carburetor was overhauled about 11 hours prior to the accident. During post overhaul flights an intermittent carburetor rich fuel/air mixture was noted in the idle circuit, which prevented the engine from idling below about 1100 rpm without fouling and quitting from excessive fuel. The carburetor was removed and sent back to the overhaul shop who removed the installed needle valve assembly and replaced it with another identical new assembly. The carburetor was then installed on the engine. When the main fuel shutoff valve was turned on, fuel began to run out of the carburetor. It was again removed and returned to the overhaul shop. The shop owner suspected that a stuck or hanging up float may have been also been the problem. Since the carburetor was equipped with an Advanced Polymer float, which is larger than either the original brass or older style composite float, the shop owner carefully adjusted the float for lateral clearance between the float and the bowl wall, and, between the float clip and the needle valve. After several attempts on the test bench to achieve a stabilized fuel level, the carburetor finally passed. The pilot (an A&P) picked up the carburetor from the shop on the day of the accident, installed it, then flew the airplane for about 15 minutes. The accident happened on the second test flight. Post accident examination of the engine revealed sooted spark plugs typical of a rich fuel/air mixture. The carburetor was functionally tested by mounting it on a tilting test fixture. A trace amount of fuel was observed leaking from the discharge nozzle. When the fixture was rotated to a bank angle, fuel flowed freely from the nozzle. Tapping on the bowl stopped the flowage. Operational testing disclosed that the carburetor was operating at an excessively rich setting at idle speed. Disassembly disclosed that the Advanced Polymer float was clean and intact with no sign of damage. The float setting and bowl clearance was good. The needle valve seat assembly was inspected and measured and found to meet specifications. The float retractor clip and needle valve shoulder clearance was measured about 0.005-inch. The pivot pin/shaft that hinges the float assembly was found to be "tight" in the inside diameter of the Polymer Float hinge points; however, the float and shaft combination did rotate freely in the float bracket. The manufacturer's service manual (and incorporated overhaul instructions) were examined. There are three types of floats which can be installed. The originally designed floats are hollow brass chambers. The second type, no longer in production, consists of floats constructed out of composite materials. The third type is the Advanced Polymer floats, which are physically larger that either the brass or composite floats. With either the brass or composite floats, a typical 0.081-inch clearance exists between the float and the bowl chamber. The increased size of the polymer float reduces the float to bowl wall clearance to a typical 0.031 inches. The original sections of the service manual address the original brass floats and calls for a post reassembly minimum clearance of 0.005-inch between the float valve seat shoulder and the float valve retractor clip. Instructions E-955 (dated 03/18/99) have been incorporated into the manual and cover the installation of the polymer floats. This document requires the assembler only to "Insure that clearance exists between the float valve seat (shoulder) and the float valve retractor clip." There is no published minimum clearance limit. During final assembly, the carburetor fuel bowl and throttle body go together blind and without the ability to see the final internal clearances. Float clearance and height settings are critical to the proper metering of fuel proportional to airflow through the venturi of the carburetor. Any float drag against the wall of the bowl assembly could feasibly disturb the critical balance. The investigation measured several sources of free play in the carburetor removed from the accident airplane with a dial indicator. About 0.015-inch was measured rotating the bowl cover and fuel bowl halves with just snug bowl screws. A source of horizontal float centering free play not mentioned in E-955 is in the float hinge that is attached with screws to the bowl cover; the hinge can slide about 0.028-inch from either screw. Index marks were added by the investigator to each float tip, and the total sideways free play of the hinge measured about 0.229-inch at the float tips. The accident float flange was coated with a bead of black transfer material to test for rubbing on the bowl walls. The float was then reassembled into the accident carburetor bowl and throttle body assemblies following the service manual and kit instructions E-955. The carburetor was then rotated in various positions. Disassembly revealed black transfer markings on the inside of the bowl walls. The arc of float travel on the hinge point can bring the Polymer float against the bowl wall when the lateral movement is adjusted while installing the float bracket. An improperly centered float that rubs on the bowl wall may affect the float buoyancy and seating of the needle valve. The float retractor clip attaches to the needle valve. Without positive clearance from the needle valve seat shoulder a loss of float buoyancy or pressure to seat the valve can occur. Without proper seating of the needle valve due to float horn to bowl contact, positive fuel shutoff would be unlikely.

Verbatim NTSB analysis and probable cause from the NTSB dataset