On September 14, 2007, an Air Tractor AT 602, N8522P, a single-engine agricultural airplane, sustained substantial damage when it made a forced landing to a field after a partial loss of engine power near Quitague, Texas. The commercial pilot/owner was not injured. No flight plan was filed and visual meteorological conditions prevailed for the local aerial application flight conducted under 14 Code of Federal Regulations Part 137.

In a written statement, the pilot said that he was traveling to a cotton field southeast of Quitague, Texas, when the engine started to lose power. He checked the power settings, turned on the fuel boost pump, and looked for a place to land. The pilot then dumped his chemical load and made a forced landing to a cotton field. In a subsequent interview with the FAA, the pilot reported that after the airplane came to rest, he removed his helmet and realized that the engine was still operating. He then placed the condition lever in the cut-off position and exited the wreckage. The pilot described the power loss as a "min-flow" condition.

A Federal Aviation Administration (FAA) safety inspector performed an on-scene investigation. According to the inspector, the airplane sustained damage to the propeller blades, the landing gear, both wings, the fuselage, and the vertical stabilizer. The airplane and engine were taken to a facility for further examination.

The engine (model PT6A-65AG, serial number PCE-PN0060)) was examined by the Safety Board on September 27, 2007. The engine cowling was still on the airplane and there was no external damage. The engine cowling was removed. The engine modules rotated freely and there was no binding. The fuel control unit (FCU) linkage was intact and positive movement of the linkage was established when the power quadrant was moved inside the cockpit. The safety that protected the P3 filter housing was cut and the proper installation of the filter was verified. This airplane had an Supplemental Type Certificate (STC) for the application of a Py flexible line that was installed from primary governor to FCU. Examination of this line revealed no damage. All associated Px lines were examined and no anomalies were noted. All B-nuts were properly installed and safetied were applicable. The FCU filter and inlet screen were removed and examined. The filter bowl was full of fuel and the screen was absent of debris. The airframe fuel filter was removed and was also full of fuel and absent of debris. The flow divider input was removed and there was some residual fuel noted; however, there was no fuel found in the flow divider drain.

The FCU was removed and sent to Pratt and Whitney Canada where it was examined on November 7, 2008. under the supervision of the Transportation Safety Board of Canada. The examination revealed that the CDP bellows was .013 inch longer than the dimension marked on the bellows at the time of manufacture. The bellows and the rod guide were sent to Woodward Governor Company, Rockford, Illinois, for a more detailed investigation.

The rod guide was examined on January 8, 2008, under the supervision of the Safety Board. According to the Woodward Engineering and Materials Laboratory Investigative reports, the rod guide was installed in a test fixture to check for leakage from the seal between CDP and overboard drain. The fixture with the rod guide installed was submerged in a fluid and the air pressure was applied to the fixture. A steady stream of bubbles was seen exiting the fixture, indicating that the seal was leaking. In the fuel control, an orifice is installed upstream of the bellows so that when pressure is bled off the Py system, the pressure in the bellows cavity is reduced, which reduces fuel flow to the engine and reduces engine speed and power. A fixture was made up to hold a Py orifice. This was plumbed into the line to the rod guide test fixture with pressure taps upstream and downstream of the orifice. Up to 100 psig was applied to the test setup. The leakage through the rod guide assembly was only enough to create a pressure drop across the orifice of 0.2 psid. According to the Woodward, this was not enough to cause a significant flow decrease from the fuel control. Therefore, it was concluded that the rod guide assembly probably did not contribute to the decrease in engine power.

The bellows, which was comprised of a two ply Be-Cu hydroformed bellows that was soldered to a 410 stainless steel end and a 410 stainless steel seat, was configured so it could be pressurized and find a leak site. The bellows was placed in a clamp (to avoid over extending the bellows), immersed in Stoddard solvent and pressurized to approximately 50 psi with helium. The approximate leak area was then determined by watching for escaping bubbles and marked. The bellows was placed in the scanning electron microscope (SEM) so the marked area could be examined in more detail. Unfortunately, the bellows surface was dirty, which obscured the surface in the leak area. When the surface contamination was analyzed using energy dispersive spectroscopy (EDS) it was found to be composed of copper (Cu), carbon (C), oxygen (O) and sulfur. In an attempt to remove this dirt, the bellows was ultrasonically cleaned first in acetone than in ethanol. Neither of these solvent cleaning steps was successful. Next the bellows was ultrasonically cleaned using a mixture of Alconox cleaner in water. This process succeeded in removing the surface contamination. The bellows was again place in the SEM and the marked area examined to try to find the cause of the leak. Two areas of interest were found in the marked area. The first was a small pinhole approximately 55 micrometers (um) in diameter. The second area was located approximately 2 mm from the pinhole and appeared to be a series of small holes. While both areas of the bellows had a corroded appearance, the cleaning required to removed the surface contamination in order to find the leak made it impossible to determine what was attacking the bellows. Once the leak was determined to be caused by holes in the bellows, the next step was to section the bellows to remove the area containing the holes. This was accomplished using a dremel tool to cut through the bellows. Once this section was removed, the two ply’s were separated and individually inspected. First, the outer ply was placed in the SEM and the interior surface examined. This revealed that while there appeared to be numerous holes on the outer side of the ply, only one was found on the inner side of the ply. The inner ply was also examined in the SEM, but no defects were found that perforated that ply.

The measurable change in size of the bellows assembly was due to a leak in the bellows itself. The leak was caused by a hole in the bellows outer ply. The hole appears to be due to pitting corrosion, but since the corrosive agent could not be determined, it was not possible to determine at what point the corrosion began. There was a deposit that was present on the bellows when received. It was composed of copper, carbon, oxygen and sulfur, and was likely copper sulfate. This is a general corrosion product that is commonly found on Be-Cu bellows and typically does not lead to pitting.

On July 3, 2006, Pratt and Whitney Canada issued Service Bulletin 13408, titled TURBOPROP ENGINE FUEL CONTROL UNIT - REPLACEMENT OF, which applied to the PT6A-65AG and PT6A-65B engines. The purpose of the Service Bulletin was to reduced the chance of a power rollback due to reduced fuel flow as a result of a leak in the Fuel Control Unit (FCU) bellows from material irregularities. The Service Bulletin told the operator to replace the hydromechanical FCU with a new one which incorporated a more robust P3 bellows design and an air flow deflector to prevent P3 air impingement on the bellows. Pratt and Whitney recommended that the FCU be replaced the next time the engine was disassembled and access was available to the necessary subassembly (i.e. module, accessories, components, or build groups).

A review of the aircraft logbooks revealed that this Service Bulletin (which was not mandatory) was applicable to this engine, but had not been complied with at the time of the accident.

The engine had last undergone a 100-hour inspection on September 12, 2007, just two hours before the accident. At that time of the accident, the engine had accrued a total of 2,071.5 hours.

Use your browsers 'back' function to return to synopsis
Return to Query Page