Aftermath: Interfaces

How a GA pilot's confusion and a controller's indifference led to fatal consequences.

(April 2011) If the souls of the pilots whose untimely ends are chronicled in NTSB accident reports could be assembled for a focus group on accident prevention, some would say of their final flights, "Yes, I was really asking for it. I should have seen that accident coming from a mile away." Others would still be scratching their heads and wondering "What went wrong? It should have been so simple!"

In the latter category, I suspect, would be the pilot of a Van’s RV-10 who crashed in April 2008 on the way from Lebanon, Tennessee, to Lakeland, Florida, for the Sun ’n Fun fly-in.

There were two private pilots in the airplane. One, 63 years old, had aircraft single-engine land and instrument ratings. His total time was 1,770 hours, including almost 360 hours in actual instrument conditions. He was the owner of a 1976 Cardinal in which he had logged 130 hours of actual instrument flying in the past seven years and 6.4 hours in the preceding six months.

The other pilot, the builder of the airplane, was 64. His total time was 526 hours, of which 68 were in the RV-10, 5.2 in the preceding 90 days. He did not have an instrument rating.

The airplane had logged 75 hours since receiving its special airworthiness certificate 15 months earlier. It was equipped with a glass panel consisting of a Dynon EFIS and dual Grand Rapids MFDs but had no vacuum — or electrically driven — attitude instruments.

The RV-10 left Lebanon in midmorning. Southern Georgia and eastern Alabama lay under the influence of a stationary front; stratus and stratocumulus clouds covered the region, with tops generally around 4,500 feet and ceilings of 1,000 feet or better. Visibilities below the clouds were eight miles or more.

The NTSB could not establish which pilot was in which seat, but since an instrument flight plan had been filed, the instrument-rated pilot was assumed to be the pilot flying.

The distance from Lebanon to Lakeland is about 550 nm. Near the halfway point, cruising in the clear at 9,300 feet, the pilot checked in with Atlanta Approach and requested a VOR approach to Weedon Field at Eufaula, Alabama, on the Georgia border. Thirteen minutes later, the controller cleared the flight to descend to 4,000, and when the RV-10 was 20 miles from the airport he cleared it for the approach. The pilot responded, “We may, I think we are going to fly the whole approach here, I think we’ll let you know when ... ” The rest of the transmission was unreadable.

The intent of this somewhat garbled statement was probably that, rather than come straight in, the pilot would fly to the VOR, which is on the airport, and make the procedure turn, and that he would advise the controller when turning inbound on the final approach. But instead he began a descending right turn to a heading of 090 and then transmitted, without explanation, “We’re going to have to come up with an alternate plan here.”

He then requested and received vectors to Auburn-Opelika airport, 30-some miles behind him; but while on the way back northward toward Auburn, he changed his mind again and requested vectors to Columbus, Georgia, explaining that he wanted an airport with an ILS. Actually, Auburn has an ILS; but the controller vectored the flight as requested and cleared it for the ILS Runway 6 approach at Columbus.

Shortly thereafter, the airplane flew across the ILS and also triggered a low altitude alert. The controller relayed the altitude warning to the pilot, who acknowledged both errors. The controller then told the pilot to climb to 3,000 feet and turn right to a westerly heading.

The pilot said, “Do you have any location that the weather is 2,000 feet or better?”

The controller replied that the pilot could consider checking the weather at Auburn, and the pilot replied, “OK, we’ll do that.”

That was his last transmission. Two minutes later, the RV-10 crashed in wooded terrain near Seale, Alabama, 17 nm from the Columbus airport. The final moments of the flight, recorded on a handheld GPS in the airplane, told a tale of pilot disorientation. The airplane first descended in a right turn from 2,700 feet to 1,300, then zoomed back upward to 2,770 before again starting to descend. It then apparently entered a spin, finally crashing only 400 feet horizontally from the last GPS fix, recorded when it was still almost 2,000 feet above the ground.

During the 14 minutes that it was in the clouds, the RV-10 had deviated as much as 400 feet above and 1,200 feet below its assigned altitude. The controller had relayed two low-altitude alerts to the pilot and told him on five different occasions that he was not on his assigned heading.

The pilot never declared an emergency, but at some point it might have begun to strike the controller that all was not well aboard the RV-10. As the course and altitude deviations multiply, it becomes increasingly apparent that this flight is a special case. Either a noninstrument pilot who is faking it has gotten in over his head, or a pilot who is qualified and equipped as required is somehow hampered in his execution of what should be simple maneuvers.

The pilot’s indecision, his requests first for an airport with an ILS and then for one where the ceiling was 2,000 feet or better, and his failure to explain to the controller what was going on all suggested a potentially precarious situation. The pilot is having trouble controlling the airplane, and he is grasping at anything that might simplify an instrument approach — which, with a 1,000-foot ceiling, should not be a challenging approach to begin with. He needs help.

In its analysis of the causes of the accident, the NTSB identified as a contributing factor the pilot’s lack of familiarity with the airplane’s electronic instrumentation. Although his log reflected many hours of instrument flying, most of it had been done in his Cardinal, which had conventional instruments.

Glass panels are easier to use than “steam gauges,” in part because they collect lots of information into one place; but they are easier only once a pilot has become used to them. Programming any electronic display to depict an ILS approach is a complex task; it’s unlikely that a pilot could figure it out amid the distractions of an IFR flight. The noninstrument-rated owner, on the other hand, had had no reason to learn how to use his flat panel for instrument approaches.

Any confusion created by the instrument display might have been exacerbated by the pilot’s lack of experience in the airplane, which had stability and control responses quite different from the Cardinal’s. There was no indication in his log that he had flown the RV-10 before. And there could have been another element: anxiety.

I wonder whether pilots who have never felt acute fear in an airplane in instrument conditions realize how subversive it can be. A pilot I knew once spoke of an increase in “mental viscosity” that can occur — thoughts that normally flit through the brain like fire slow to a slushy ooze. I experienced this once after picking up a load of ice in a Beech Sierra — not a powerful performer to begin with — at night over the mountains of western Arizona. I remember staring at the DG, strangely unable to figure out what direction I needed to be going. Fortunately, there was a hole over Prescott; I don’t know how competent an instrument approach I would have flown that night. I was right to be scared: After we landed, slabs of ice fell from the wings onto the taxiway.

This pilot’s indecision and his errors of heading and track can be seen as the results of discovering, when he entered IMC, that the digital display was more difficult to interpret than he had expected. Increasing anxiety made the difficulty greater and greater.

It seems that a tipping point came when the pilot asked the controller for an airport with a 2,000-foot ceiling. The request may have been unreasonable — weather cannot be made to order — but rather than try to get the information for the pilot the controller unhelpfully suggested that he "consider checking the weather" at Auburn. Certainly that suggestion cannot have been made with the sarcasm that its reproduction on the printed page suggests. But the pilot had already given indications of being in over his head; he didn't need one more task. Not only the instrument interface failed this pilot; the ATC interface did as well.
This article is based on the NTSB's report of the accident and is intended to bring the issues raised to our readers' attention. It is not intended to judge or to reach any definitive conclusions about the ability or capacity of any person, living or dead, or any aircraft or accessory._