Probable Cause #26: Four Minutes
Pilots have many different phobias when it comes to flying personal airplanes. Some cringe at the thought of flying IFR over mountains, whether in a single or a twin. Others refuse to fly over large bodies of water, choosing to skirt Lake Michigan via the Gary/Chicago/Milwaukee route to Oshkosh each year. No matter which operation curdles your blood, few pilots disagree that an in-flight fire is the most dangerous and feared emergency they can face.
The reasons include pride -- no pilot likes to think he or she can get into a situation from which proper application of technique or a checklist cannot extricate them -- and a visceral reaction to the idea of being burned. The thought of trying to control an airplane while flames erupt around you is not pleasant.
This kind of emergency, especially when fed by the combustibles in an engine compartment, can very quickly have dire consequences. With a 100-plus-knot slipstream helping it, an in-flight fire can quickly burn through a wing spar and fail a wing, or penetrate the cabin firewall and enter the aircraft’s cabin.
Fortunately, in-flight fires are very rare occurrences. However, on December 20, 2002, just such an event brought down a turbocharged Piper Saratoga II into a residential area near Woodbury, Conn. The 1200-hour Commercial pilot and his passenger were fatally injured and the airplane was destroyed. The accident occurred at night in visual conditions during a planned flight from the Rutland State Airport in Rutland, Vt., to the Republic Airport in Farmingdale, N.Y.
History Of The Flight
The accident airplane had been airborne for about 50 minutes when the pilot contacted the New York TRACON, reporting level at 6500 feet. Less than a minute later, the pilot declared an emergency; there was heavy smoke in the cockpit and an engine fire. The controller immediately directed the airplane to Waterbury-Oxford Airport in Waterbury, Conn. The approach controller then coordinated with the tower controller at Waterbury-Oxford, and reported that the airplane was five miles northwest of the airport, descending out of 5500 feet.
The tower controller stated that he turned up the runway lights and called airport emergency personnel before making visual contact with the airplane, scanning it with binoculars and noting nothing unusual. After other duties drew away his attention from the accident airplane, he again made visual contact with the airplane, which by then appeared to be entering a close-in left downwind for Runway 36 at Woodbury. At this time, he noticed that the airplane was now engulfed in yellow flames originating from around the engine area. The flames dissipated slightly, then the airplane began a steep, final descent. The tower controller never received any radio transmissions from the airplane.
A ground-based witness saw the airplane fly past him from left to right. The airplane was about 10 feet away, and 20-25 feet above him. The front of the airplane was on fire, and he observed a “red glow.” He did not see smoke and did not hear the engine running, hearing only the sound of rushing air as the airplane passed by. He lost sight of the airplane behind trees, then heard the sound of the impact.
The wreckage was scattered over two residential backyards; all primary components of the airplane were located in this area. According to the NTSB, the majority of in-flight fire damage was located on the right side of the airplane, from the engine cowling back to the cabin door, which displayed extensive impact and fire damage. Soot and aluminum splatter was found on the outer surface of the door in a pattern consistent with airflow. The exterior lower portion of the right-side forward cargo door (the Saratoga has a baggage compartment between the engine and the cabin) revealed soot and aluminum splatter in the direction of the slip stream. The right wing’s leading edge also exhibited soot and aluminum splatter, consistent with the airflow over the wing during flight. Soot and aluminum splatter were also noted on the right horizontal stabilizer.
The splatter of aluminum on various parts of the forward fuselage, cabin door, right wing and tail testify to the intensity of the fire.
Investigators focused on various engine components in trying to find the source of the fire. The engine had been overhauled with new cylinders and accessories on September 16, 2002. The engine was reinstalled on October 2, 2002, and an annual inspection was performed. The NTSB singled out two components for special attention: the fuel system and the turbocharger.
Examination of some of the fuel line fittings on the left side of the engine revealed they were loose. While the fuel pump remained attached to the engine, its top adjustment housing was separated from impact. The fuel pump screws were determined to have fractured in a manner consistent with overload. Visual examination of the fracture surface on the fuel pump top adjustment housing revealed features consistent with overstress.
At the time of the accident, an airworthiness directive (AD) was in effect for the installed fuel pump. It addressed the torque of the screws on the plate of the fuel pump, and required initial and repetitive inspections of the relief valve attaching screws. The actions specified in the AD were intended to prevent fuel pump leaks, which could result in an engine fire. The airplane’s logbooks indicated compliance with this AD.
The outlet side of the turbocharger oil supply line, with a partially intact check-valve and B-nut, were closely examined. The B-nut was finger loose and appeared to be fully engaged with the check valve. Examination of the check-valve’s external threads revealed they were cross-threaded, and only the last few threads would fully engage. The interfacing threads of the B-nut did not reveal corresponding damage. Examination of the surface of the oil line ferrule that contacted the tip of the check valve showed a non-uniform spiral pattern that was non-concentric. The oil supply line had been replaced during the engine overhaul, but the check-valve was not. There was no record of it ever having been replaced.
The NTSB report does not mention the airplane’s fuel selector -- whether it was even found in the wreckage and, if so, how it was positioned. If it had been in the “off” position and the fire continued to burn, that would be a good clue that the fire was oil-fed. Since we don’t know what steps the pilot took to extinguish the fire or prepare for the emergency landing, we can’t determine if the engine stopped on its own or whether the pilot’s remedial actions were responsible.
Unfortunately, the available evidence did not allow the NTSB to determine the fire’s origin. However, the FAA issued a subsequent AD against the installed fuel pump type, requiring replacement of certain components with new ones of a different design. (You can read the NTSB Probable Cause report here.)
By definition, there are many potential fire sources in the average general aviation airplane’s engine compartment. A relatively complex installation like the engine in the turbocharged accident airplane increases the chances for a small leak to turn into a conflagration under the right (wrong) circumstances. Given both the catastrophic nature of such a fire and its inability to be easily extinguished, operators should pay close attention to engine-compartment leaks and the tightness of various fittings.
At the end of the day, stuff happens. When a fire erupts in an airplane’s engine compartment, the time available to get it safely on the ground can be measured in seconds, not minutes. In this instance, less than four minutes elapsed between the pilot’s initial report of an engine fire and the crash.
That’s barely enough time to read this article.