The Danger of Repeated Omissions

About nine minutes later, the pilot asked Atlanta Center for a descent to 10,000, which was granted. After another minute, he reported an "engine fluctuation" and requested a further descent and deviation to Selma (KSEM), which was close by. Center cleared him to 5,000. Shortly afterward, the pilot, his voice tense but calm, declared an emergency. 

The center controller offered the pilot his choice of Runway 15 or 33; he was currently lined up with 15. The pilot reported two souls on board and said that the prop was turning but he could not tell whether the engine was producing power. "I may be dead-stickin' it," he said. 

Three minutes after the pilot declared an emergency, the center controller advised him that Selma was at 12 o'clock, 9 miles. The pilot said that he could not make a straight in—presumably he was still too high—and would fly a pattern. Then he added, "I've got a fire, looks like as well, sir."

Half a minute later, the pilot requested the frequency for Selma. The controller gave him the CTAF frequency, 122.7. 

That was the pilot's last communication. The Lance had not yet turned toward the runway, but it remained in radar contact a little while longer. The airport was now at 9 o'clock, 4 miles, the controller said.  Then, "Do you copy? Radar contact lost, respond if you can."

Now then the wheels of accident response began to turn. Another controller provided the phone numbers of the police and the airport at Selma. A cropduster spotted the heavily fragmented wreckage of the Lance in a field. It had struck the ground at high speed a few miles from the runway; both pilots had died in the crash.

The National Transportation Safety Board's accident investigation naturally focused on the engine. It had failed catastrophically, the No. 6 (rearmost) connecting rod breaking, smashing a hole in the top of the crankcase and disintegrating into pieces, most of which could not be found. The No. 5 rod had also separated from the crankshaft. The camshaft was broken in two. Almost all of the connecting rod bearings were deeply scored—the crankshaft journals not so badly—and the aft end of the crankshaft was discolored by intense frictional heating. There were also score marks in the oil pump.

The key finding was that the perforated cylindrical screen through which the lubrication system takes in oil from the sump was 60-percent filled with a mixture of carbon flakes and metallic material. The accident report conflates "metallic" and "magnetic," leaving some confusion about whether the "metallic" debris, which accounted for three-quarters of what was found, was entirely ferrous or not, and therefore about how much of the non-magnetic material may have been non-ferrous metal as well. In any case, however, the verdict of the NTSB was that the obstruction of the screen had led to oil starvation and to the eventual—and inevitably—failures of the rods.

The investigation therefore focused its attention upon the mechanic who had maintained the airplane.

It emerged that the owners had persistently reported low oil pressure. The mechanic had removed and replaced the oil pressure relief valve, removed and flushed the oil cooler, and at least twice raised the oil pressure using the adjustment screw. In response to a direct question as to whether he had cleaned the suction screen at the last inspection, the mechanic replied that the oil cooler had been flushed. The NTSB seemed to consider this reply evasive.

Normally, every step an A&P takes during an inspection is recorded. The oil filter had been cut open and examined at every annual, and nothing unusual reported. Recently, a spectroscopic oil analysis had been performed, but the mechanic did not know the result. Significantly, however, his notes on the most recent annual inspection, less than 90 days before the accident, made no mention of removing and cleaning the suction screen, which is recommended to be done every 50 or 100 hours, depending which guidance you consult. In fact, it turned out that the most recent record of cleaning the screen was dated 11 years before the accident.

It was impossible to know how long the accumulation of debris had been there. The NTSB assumed that it had been there a long time, and was the reason for the repeated reports of low oil pressure. The investigation concluded that the engine failed because of the blockage in the screen, and identified as the probable cause of the accident "the mechanic's failure to clean the oil suction screen during the most recent maintenance..."

All pilots imagine being in the position in which the two pilots of the Lance found themselves, and wonder how they would react. Certainly, the actions of the Lance pilots were appropriate and, as people in the business like to say, "professional." They must have heard a great deal of alarming noise as the No. 6 connecting rod broke loose from the crankshaft and punched a big hole in the crankcase, but they remained calm, they made the correct decisions, they flew the airplane. They were lucky that there was a good airport in easy gliding range and they were already talking to a controller. 

They believed that their problem was simply a power loss. A dead-stick landing, while challenging, was very likely to succeed. The reason they came down so fast is not clear—rather than buy themselves as much time aloft as possible, they descended from 15,000 feet to near sea level in nine minutes—but it turned out to have been the right thing to do when fire broke out in the cowling.

We cannot know what happened in the final minute of the flight, when the two pilots lost control of the airplane. Probably they were blinded or overcome by smoke. The NTSB's description of the wreckage mentions "thermal damage"—that is, the effects of fire—both ahead of the firewall and behind it. Wiring and thermal insulation under the glareshield were burned. 

Frustratingly—because any inflight fire is a potential teaching tool for other pilots— the NTSB does not explain, or even speculate upon, how the fire got past the firewall or what the original source of ignition may have been. The gear and flaps were up and the mag switch turned off, but the fuel selector was on the left tank. It is possible that the continuing flow of fuel to the engine fed the fire. 

The fatal crash was a sad and terrible consequence of what may have seemed, before, like a minor omission. But it was a minor omission many times repeated. 

"For want of a nail..."