NTSB Factual Report

HISTORY OF FLIGHT

On August 4, 1997, about 1605 eastern daylight time, a Piper J3C-65, N24565, was destroyed when it impacted the ground during a forced landing, after takeoff from the Danbury Municipal Airport (DXR), Danbury, Connecticut. The certificated private pilot received serious injuries, and the pilot rated passenger was fatally injured. No flight plan was filed for the local flight that departed DXR. Visual meteorological conditions prevailed for the personal flight conducted under 14 CFR Part 91.

In the pilot's written statement, he stated the airplane had been flown for about 10 hours in the weeks preceding the accident, and it "performed flawlessly." On the day of the accident, the airplane was "topped off" with 10 gallons of aviation fuel, and the pilot preflighted the airplane and seated a passenger in the front seat. He then went through the engine starting procedures, where "the engine started and ran smoothly."

The pilot departed DXR with the passenger and flew a local flight, where he performed steep turns, slow flight, and approach and departure stalls. After a 1 hour flight, the pilot returned to DXR and experienced radio problems. After landing, the pilot had a maintenance facility examine the radio, where it was determined that the problem was due to the antenna ground connection.

The pilot further stated:

"...The engine was restarted to check for ignition interference and the engine started with one pull. The radio checked good on the ground, and the engine was shut down. With the objective of air-testing the radio...a second flight was initiated. Once again the engine started with the first pull. The aircraft was cleared to runway 17, a routine run-up was performed and takeoff clearance received. The aircraft had been on the ground less than 30 minutes. Takeoff power was correct and 565 [N24565] accelerated, took off and climbed normally. At about 200 feet AGL, the engine began to sputter and the nose was lowered. As the nose lowered, the engine surged to full power. When normal climb attitude was restored, the engine resumed sputtering. With partial power apparently available, and a forbidding wall of hills ahead, I initiated a relatively steep turn to the left to return to the field. On entering the turn the engine failed completely, the aircraft entered an accelerated stall and crashed to the ground."

Several witnesses reported that the airplane departed runway 17. During the initial takeoff climb, the witnesses heard the airplane's engine intermittently lose power. The airplane was then observed in a left turn when it descended and impacted the ground. One witness stated, "...engine started to miss again, pilot then tried to turn around. He made a 180 degree nose high turn when aircraft stalled and made a 1/2 turn spin into ground." After impact, the airplane was consumed by fire.

The accident occurred during the hours of daylight approximately 41 degrees, 23 minutes north latitude, and 73 degrees, 29 minutes west longitude.

PILOT INFORMATION

The pilot held a private pilot certificate with ratings for airplane single and multiengine land, rotorcraft helicopter, and instrument airplane.

His most recent Federal Aviation Administration (FAA) Third Class Medical Certificate was issued on January 10, 1997.

The pilot's estimate of flight time in the Pilot Operator Report stated that he had accumulated about 4,350 total flight hours. He also listed 1,590 hours in single engine airplanes, of which 58 hours were in make and model. He further listed 7 hours in make and model within the previous 30 days.

AIRCRAFT INFORMATION

During telephone interviews with the pilot, and in his written statement, he stated that the airplane had been manufactured in 1938; however, it had been "meticulously overhauled with a new fuselage and engine." He stated that the airplane had been flown roundtrip from DXR to California in 1995, and that during the 5 years since it had been purchased, "...565 has never exhibited any operating abnormalities." The pilot also stated that in addition to the fuselage tank, the airplane had an auxiliary fuel tank with a capacity of 10 gallons.

AIRPORT INFORMATION

The DXR runway 17 was 3,135 feet long, with taxiways on both sides of the runway. The distance from the accident site to the departure end of runway 17, was about 1,000 feet. From intersection of runway 17 and runway 26, the departure end of runway 26 was about 2,500 feet.

WRECKAGE INFORMATION

The airplane wreckage was examined at the accident site on August 4 and 5, 1997. The examination revealed that the airplane was consumed by a post crash fire, and all major components were accounted for at the scene. The airplane came to rest at DXR, south of runway 8-26, and east of runway 17-35, on an approximate magnetic heading of 175 degrees.

Flight control continuity was established from the cockpit controls to all primary control surfaces.

The engine crankshaft was rotated by hand through the propeller flange bolts. Valve train continuity and compression were obtained on all cylinders, except for the number 4 cylinder, which had impact damage. All top spark plugs were removed, and the spark plug electrodes were light tan and gray. The right and left magnetos were removed and bench tested. During the magneto bench test, the left magneto produced spark at three of the four towers, and the right produced none.

Examination also revealed that the fuel selector control in the cockpit, and the fuel selector valve assembly, were in the off position. The fuel selector valve was removed from the wreckage, and there was no sign of impact damage to the linkage. When placed to the on position, sediment was observed partially obstructing the fuel valve opening. When air pressure was applied to the fuel selector valve in the off position, a flow of air was detected at the opposite end.

COMMUNICATION

A review of the DXR air traffic control communication transcripts revealed that at 1600:19, the pilot of N24565 stated to ground control, "five six five at BAC with papa ready to taxi for takeoff." The airplane was cleared to taxi to runway 17, and at 16003:42, the pilot contacted DXR tower and advised that he was ready to takeoff. N24565 was cleared to takeoff, and acknowledged, "cleared to go five six five," at 2003:52. At 2004:42, the pilot radioed, "and tower we got to turn out right here we just had our engine sputter." That was the last recorded transmission from N24565.

TESTS AND RESEARCH

Carburetor

The main body of the carburetor assembly was bench tested at Mattituck Aviation, Mattituck, New York, on October 2, 1997, under the supervision of an NTSB Investigator. The Investigator's report stated that during the examination, it was noted that the main body of the carburetor had sustained some impact damage in the form of a dent in the vicinity of the float. When fuel was pumped into the float chamber, it was noted that the float rose until its travel was impeded by the dent. Furthermore, fuel was observed to be leaking from the area of the dent. The float was then removed and the dent straightened. When the float was reinstalled, the float and needle valve assembly functioned, and the float appeared within limits.

Fuel Selector

The fuel selector and the material observed in the selector were sent to the National Transportation Safety Board Materials Laboratory in Washington, D. C.

In the Material Laboratory factual report, it stated the following:

"...The fuel line fitting was removed from the valve and when the selector was placed in the 'off' position, the fuel selector valve was not sealed...The valve ball was found to be loose in its fitting...The condition of the valve ball seal prior to the post-crash fire could not be determined."

The report further stated:

"An energy dispersive X-ray spectrum (EDS) analysis was performed on material taken from the inside of the [fuel] selector valve, the inside of the fuel line fitting and the material submitted. This material was determined to be flourine-rich, consistent with fluoroplastics (Teflon).

J3 Fuel System Examination and Test

On September 19, 1998, a Safety Board Investigator examined the fuel system of a Piper J3-65, converted with an 85 horsepower engine. The examination was conducted at the Aeroflex-Andover Airport, Andover, New Jersey. The on/off fuel selector valve was located on the center of the fuselage, below the forward instrument panel. It was positioned about mid way between the instrument panel and the firewall. Fuel from the tanks traveled down to the valve, where a single fuel line entered the valve on the side closest to the pilot seat, towards the rear of the fuselage. Fuel then exited the valve on the forward side, closest to the engine, down a flexible fuel line, through the firewall to the engine carburetor.

The fuel valve consisted of a single shut off lever mounted on top of the valve. The lever was mounted with the control arm pointed forward toward the engine. Connected to the end of the lever arm was a cable and housing. The cable ran to the left side of the fuselage, where it entered the left fuselage side wall. The cable was connected to a push pull knob mounted on the left fuselage wall. The knob was recessed into the left side wall and was located about mid-arm to shoulder high level, next to the forward seat.

When the knob was all the way forward, it pushed on the cable, and pushed the fuel valve arm towards the right side of the fuselage, to the ON position. When the valve was ON, the arm was to the right of the valve center, and pointed at the 1:00 to 2:00 o'clock position. When the knob was pulled aft, it pulled the control cable, and the valve arm was pulled left, to the OFF position. The valve arm then pointed at the 10:00 o'clock position. When pulled by hand, the knob and cable did not move easily, but required several pounds of force to move.

A test was then conducted about 30 minutes after the engine had been shut down. The fuel valve was placed in the OFF position, and the Safety Board Investigator sat in the back seat and held the brakes. The engine was then hand propped and started. It was run at 1,100 to 1,200 RPM for about 15 seconds, and 700 to 800 RPM for the remaining time. The engine ran for 1 minute 46 seconds before it ran out of fuel.

ADDITIONAL INFORMATION

Fuel Selector Information

The fuel selector shut off valve in N24565 was located in the forward cockpit. The knob could be moved forward and aft. The "ON" position was forward, and the "OFF" position was aft.

During a telephone interview with the pilot he stated that he did not recall which fuel selector valve position was "ON" and "OFF". He further stated that "as I remember, aft is open and forward is closed." Also, that he would not turn the fuel selector valve off unless he was putting the airplane away.

The pilot also stated that he did not turn the fuel shut off valve "OFF" at any time on the day of the accident, and that the impact may have moved the valve to the "OFF" position. He also stated that it took "a lot of hand pressure" to move the selector.

Pilot's Observations

During telephone interviews with the pilot and in his written statements, he stated that fuel exhaustion, water in the fuel, the magnetos, and internal engine damage could not or were not the cause of the loss of power. He further stated that "The cause HAD TO BE fuel supply."

Although the pilot stated several times that the "engine was impeccable," and had run flawlessly for the past 5 years with no carburetor malfunctions, he stated, "...it would appear most likely that carburetor malfunction was the cause of the accident."

Fueling and Maintenance

A local fixed base operator's records revealed that the airplane was refueled with 11.8 gallons of fuel earlier on the day of the accident. Also according to the fuel records, the last refueling prior to August 4, 1997, was October 7, 1996.

The refueler reported that the airplane had two fuel tanks, and that both tanks were "topped off." The refueler also stated that he had observed the airplane flying for about 45 minutes the week before the accident, and that before that, "I have not seen the plane fly in months."

Fuel samples were taken from the fuel truck August 5, 1997, and observed to be absent of contamination. Another fixed base operator reported that they performed maintenance on the airplane's radios earlier that day, and the purpose of the flight was to test the radios.

FAA Flight Training Handbook - AC 61-21A

According to the FAA Handbook:

"If an actual engine failure should occur immediately after takeoff, and before a safe maneuvering altitude is attained, it is usually inadvisable to attempt to turn back to the field from which the takeoff was made...it is generally safer to immediately establish the proper glide attitude, and select a field directly ahead or slightly to either side of the takeoff path."

The airplane wreckage was released on August 5, 1997, to the DXR Airport Administrator, and receipt of the wreckage and all retained parts was acknowledged by Allen Ryan, on November 5, 1997, a representative of the owner's insurance company.

NTSB Probable Cause

The pilot's failure to takeoff with the fuel selector in the full on position, which resulted in fuel starvation. Also causal was the pilot's improper in-flight decision to attempt a 180 degree turn back to the runway, which resulted in the inadvertent stall/spin.