Maintenance matters

Flying Safety, Nov, 2004

1. The B-nut securing the PS3 sensor to the top of the ME was found backed-off.

2. The canon plug connection between the T5 amp and the main fuel control (MFC) was discovered to be cross-threaded and apparently tightened with tools (against common maintenance practice of hand-tightening).

3. The MFC serial number did not match the serial number documented as being installed.

These three items, although separate in nature, proved to be directly related to the cause of this mishap. Interior engine inspection revealed significant damage to the engine aft of the combustion chamber. The HPT and LPT were completely destroyed along with the engine bearings and associated hardware parts as a result of severe overtemp conditions. Portions of the HPT had melted into molten metal and passed to the LPT and into the tailpipe causing extensive interior object damage (IOD). In addition, the outer seals on the LPT jammed as a result of the IOD and subsequently seized the engine.

TEMS data revealed the aircraft never flew in the engine disturbance envelope and the throttle was properly rigged. Investigation revealed no FOD or hardware failures occurred in the engine prior to the mishap sequence. However, lack of fuel was a factor. The fuel supply tested good post-mishap. The MFC was suspected as a cause during the compressor stall portion of the mishap investigation, and was PQDR'd. Analysis showed it was functioning as designed during the mishap sequence. In addition, the canon plug connects important signals from the T5 Amp to the MFC. However, post-mishap analysis proved the canon plug functioned despite its substandard connection. However, the loose PS3 B-nut found on the engine during disassembly contributes signals to the MFC, affecting fuel scheduling. Investigation revealed that a loose PS3 would have provided erroneous signals to the MFC, which would have caused a reduction in fuel flow to the engine below a level required for engine operation.

The investigation turned next to the cause of the overtemp sequence. The TF-34 maintains logic which, when the throttle is set at idle and the engine reduces core RPM below 56 percent, will automatically cause an auto-start attempt. This auto-start procedure would ignore the erroneous signals which caused the flameout and subsequently flood the combustion chamber with fuel to re-light the engine. This excessive flow of fuel caused the overtemp condition, which then caused the Class B category damage.

The key to the mishap remains highlighted by the undocumented MFC change. An MFC change would affect both the PS3 sensor B-nut and the T5 Amp canon plug connection. As mentioned earlier, both of these items were found and criticized for substandard maintenance practices (especially the canon plug). An MFC change would have forced the loosening (and hopefully, the re-tightening) of the PS3 B-nut. Also, the connection between the T5 amp and the MFC would have been loosened and re-tightened. However, it is impossible to investigate this change, since it was undocumented and could have occurred at anytime between the overhaul in 2000 and the mishap flight. There are no indications as to who, when, or why the MFC were switched between the ME and a second engine. Does this sound like quality maintenance? Not to me. We must document everything we do and follow the rules established to prevent mishaps.