In August 2015, a Florence, South Carolina, doctor, 51, left his home field early in the morning on a vacation trip whose eventual destination was the Grand Canyon. He was flying his 1963 Beech B33 Debonair, which he had owned since 2008. His first stop was to be Hot Springs, Arkansas, 689 nm distant, but shortly after taking off he amended his IFR clearance to swing southward around a line of weather. He later changed his destination to El Dorado, Arkansas.
Arriving at El Dorado, he was cleared for the GPS approach to Runway 4, but on breaking out he canceled IFR and circled to land on Runway 31. The wind was 17 gusting to 25 from 010 degrees and the field was VFR with scattered clouds at 1,900 feet and an overcast layer at 3,600. The last recorded radar return came at a location southwest of the field and about 2½ miles out. At that point, the airplane had been airborne for five hours and 36 minutes, and had traveled 722 nm.
The Debonair crashed 2,200 feet short of the runway, diving vertically into an area of 100-foot trees. The pilot was killed. Fire broke out and almost entirely consumed the cabin, destroying, among other things that would have been helpful to accident investigators, the fuel-tank selector.
One of the first things an investigator looks at after a crash is the propeller blades. When an airplane hits an obstacle or the ground under power, the blades are twisted in the plane of rotation and exhibit chordwise scoring. If the engine is not producing power, the blades instead fold backward and the telltale chordwise marks are absent. The soft aluminum spinner may also provide evidence of rotation or lack of it. In this case, it appeared from both the propeller blades and the spinner that the engine had not been developing power on impact.
Sometimes, after an accident apparently related to an engine failure, superficial repairs can be made to the engine and it can be run on a test stand to determine why it lost power in flight. In this case, the engine could not be run, but the crankshaft could be turned, compressions measured and the valves and spark plugs examined. No mechanical reason for the loss of power was found. Significantly, however, the fuel distributor — the little “spider” on top of a fuel-injected engine — had no fuel in it.
Accident investigators reconstructed the flight from ATC radar records. Correlating forecast winds aloft with observed groundspeeds, they concluded that the Debonair’s average true airspeed in cruise was 142 knots. That speed corresponded, according to the manufacturer’s performance data, to a power setting of about 55 percent, and should have provided an endurance of six and a half hours, plus 45 minutes reserve at reduced power.
The B33 has one fuel tank in each wing, and a total usable fuel capacity of 74 gallons. Like those of most low-wing airplanes, its fuel-tank selector does not have a “both” setting. Its Continental IO-470 engine, rated at 225 hp, would most likely be operated on the rich side of peak EGT, with a specific fuel consumption of .45 to .5 pounds of fuel per horsepower per hour. Some pilots, however, believe that they do their engines a favor by running them rich, so it is possible that the fuel consumption was greater than book. For that matter, the engine was getting close to its TBO; if it was down on compression, its fuel consumption may have been a little higher than expected.
The most recent fuel purchase investigators were able to document was of 7.6 gallons, four days before the accident. Such a small purchase suggests a pilot topping his tanks for an anticipated long flight. Investigators could not determine whether the plane had been flown since.
What was apparent, however, was that the near-vertical path of the airplane through the tall trees into which it fell strongly implied a stall-spin. Absent witnesses, and the accident having occurred after the airplane ceased to be visible to ATC radar, it was impossible to know whether the loss of control took place while the Debonair was already on final approach or as it was turning from base to final. The evidence of loss of engine power, however, presumably before the loss of control, was very strong. But the airplane had evidently not run out of fuel; if it had, it would not have burned.
The National Transportation Safety Board gave an uncharacteristically cautious statement of the probable cause, blaming the crash on: “The loss of engine power for reasons that could not be determined during post-accident examination of the accident or based on the available information, and the pilot’s failure to maintain adequate airspeed and his exceedance of the airplane’s critical angle of attack after the engine power loss, which led to an aerodynamic stall and loss of control.”
In its synoptic discussion of the accident, however, the NTSB suggested a more detailed theory. A convective system, with distant lightning, was approaching the airport from the north as the Debonair circled to land. “It is possible that the pilot became distracted … by the approaching weather system and may not have selected the fullest fuel tank for the approach.” Then, going farther out on a limb, the author speculates that “the pilot likely attempted to extend the airplane’s glide to reach the runway and exceeded its critical angle of attack.”
Maybe so. The evidence of fuel starvation is pretty strong, and if the pilot instinctively reacted to the sight of approaching treetops by pulling back on the yoke, he wasn’t the first to do so.
It’s not only possible but likely, when tanks are nearing empty, that one tank is significantly emptier than the other, particularly when there is only one person in the airplane. The pilot sits about a foot to the left of centerline; if he weighed 160 pounds, the plane would be out of equilibrium by 160 foot-pounds. If you assume that the fuel is 4 feet from the centerline, the right tank would need to have seven gallons more in it than the left just to balance the airplane. If you had 13 gallons remaining, you could have three gallons on the left and 10 on the right, and the airplane would feel balanced. In gusty conditions, the possibility of unporting in the nearly empty tank would be increased.
Putting together the possibility that a pilot might use more fuel than expected, that “full” tanks might not be perfectly full, that there may be significantly more fuel on one side than the other even when the plane seems to be in equilibrium, and that fuel gauges may be inaccurate, you have the makings of an occasional perfect storm of fuel-quantity surprises. The handbook predicts speeds and fuel consumptions with great confidence, but your results may vary.