The July 5 issue of Flight School Business discussed the accident that killed four members of the Florida Institute of Technology (FIT) aviation department on November 11, 2010 (“Multiplying Risk”). Their Piper PA-44-180 Seminole was so completely consumed by the post-impact fire that it took the National Transportation Safety Board more than a year to issue its factual report, which came out on December 1. The formal determination of probable cause has yet to be published, but in the meantime, the factual report confirms much of what was reported immediately after the accident and adds some new details.
The four were on their way home from the Bahamas to FIT’s base in Melbourne and had stopped at Palm Beach International (KPBI) to clear customs. The left-seat pilot was a current FIT student with a commercial certificate and about 300 hours of flight time; almost 50 hours were multi-engine, but he hadn’t flown a twin in nearly a year. A staff instructor was in the right seat; another staff instructor and the student’s brother (who worked for the program) were in back. They took off from PBI about five minutes after 6 p.m.—10 minutes after the end of civil evening twilight—on a VFR flight plan. Thirteen seconds after they received their takeoff clearance, the CFI radioed the tower reporting that they’d suffered an engine failure and that “we would like to turn around and land, please.” The tower controller cleared them to land on any runway. Footage from a surveillance camera showed the airplane’s beacon beginning a slow left turn that continued until the explosion.
The factual report confirms four crucial points reported earlier:
- The left engine’s fuel selector was found in the Off position, and there was no fuel in the carburetor or lines. There was no evidence of any other failure in the left engine.
- Both throttles were close to full forward.
- Neither propeller was feathered.
- The gear selector switch was down, and the landing gear was down and locked.
The report also contains the results of tests conducted with another FIT Seminole.
Investigators found that with the fuel selector turned off, the engine ran for five minutes and 50 seconds at taxi speed; for 36 seconds at full takeoff power; or for 10 seconds at takeoff power after five minutes at taxi speed. Transcripts show that the CFI reported the power loss not quite six minutes after the flight was cleared to taxi and seven seconds after they read back their takeoff clearance.
The flash memory chips from the airplane’s avionics suite somehow survived the fire. Data recorded during that final flight show the airplane reaching its maximum altitude of 170 feet 20 seconds after the left engine’s rpm began to decay, and 18 seconds after the airplane’s first hard roll to the left. At this point its airspeed was 73 knots, well below the blue-line speed of 88 but above the 56-knot VMC. The last data point was recorded only 14 seconds after that, by which time the airplane had slowed to 68 knots while banking left more than 20 degrees. Its heading had changed 90 degrees, from east to north. The right engine stayed at 2,650 rpm throughout.
We will probably never know who was flying the airplane during those last 30 seconds, but these facts give some hints about his or her intentions. With very little altitude to work with and the gear still down, the preferred alternative would have been to close both throttles and land straight ahead on grass, taxiways, or whatever surface was available. A sidestep to the parallel runway also might have been possible—about 3,000 feet of pavement remained between the point of impact and the departure end. But it doesn’t appear that the pilot was trying to do either. The fact that the right engine stayed at full throttle, together with the CFI’s reference to turning around to land, suggests that they were still trying to fly when they lost control.
When a light twin loses an engine, maintaining airspeed and directional control are crucial. Especially in a low-powered twin like a Seminole operating close to maximum gross weight, achieving any positive rate of climb depends on cleaning up the airplane and pitching for blue-line speed. As long as the gear stayed down and the left propeller was windmilling, this one had no chance of reaching a safe altitude.
There’s no way to guess whether the pilot consciously ignored those steps, assuming they’d be landing again in a few minutes anyway, or whether the cockpit simply became too chaotic too quickly. The engine failure occurred at just about the worst possible time, requiring the pilot in command to take the correct action almost instantaneously to avoid disaster. Yes, 30 seconds is long enough to retract the gear and feather the prop, provided you act immediately, decisively, and correctly. It’s a good bet that no one on board had ever experienced an actual engine failure just after takeoff, especially one at high weight, low airspeed, and low altitude at night. Confusion would be understandable.
Genuine emergencies do occur on training flights, and when they do they’re no less dangerous than on any other flight. The chance of an emergency doesn’t change. An instructor can’t afford to relax his or her guard too much, even on a pleasant flight with friends; neither can a student, for that matter, or any other pilot. Every takeoff demands equally close attention. When an emergency does arise, there are no short cuts. All the steps in the prescribed response are there for a reason, and a pilot skips any one of them at his peril.
In this case, it seems likely that a mis-set fuel selector precipitated the entire disaster. Aside from the fact that rigorous adherence to the pre-takeoff checklist should have caught that, there’s the question of why it was turned off in the first place. One possibility lies in the Seminole’s taxi checklist, which calls for checking the crossfeed settings during taxi. Doing this passes each selector through the Off position twice. While it’s certainly good to verify that crossfeed works, this may be one of the cases where an operator can justify establishing procedures that diverge from aspects of the published checklists.
Related Articles
