A Pilot‘s Bewilderment

It was about 5:55 p.m., on January 2, 2015, when the Seneca hit trees and then struck the ground while the pilot was attempting a forced landing near Kuttawa, Kentucky. The commercial pilot and three passengers were killed. A fourth passenger received serious injuries. The flight had departed Tallahassee Regional Airport in Florida and was en route to Mount Vernon, Illinois. Kuttawa is about 96 miles from Mount Vernon. In his IFR flight plan, the pilot estimated the time en route would be three hours and 12 minutes. He reported that the airplane would have four hours and 50 minutes of fuel onboard.

That morning, the airplane had been flown from Key West, Florida, to Tallahassee. The pilot had ordered the tanks topped up at Key West and again at Tallahassee, where it took 67.7 gallons of 100LL according to a fuel receipt.

The airplane's fuel requirements, fuel system schematic and system operation were covered in documentation provided by the manufacturer. The Seneca II could carry 128 gallons, with five of those gallons being unusable. Each wing had two tanks. The outboard tank on each side had the fuel filler port. The fuel flowed to each engine from the inboard tank on the same side as the engine. As fuel was consumed from the inboard tank, it was replenished from the outboard tank.

In addition to an engine-driven fuel pump, each engine had an electric pump for use in starting, if there was a failure of the engine-driven pump and to keep fuel flowing if fuel vapors built up. The electric pumps were controlled by rocker switches on the electrical panel and had LO, HI and OFF positions. The manufacturer warned that excessive pressure and very rich mixtures would occur if an electric pump was turned to HI when the fuel injectors were functioning normally. The manufacturer said the electric pumps could be used on LO if the engines became unstable or fuel flow indications started fluctuating when the engines were at idle or when the airplane was cruising at high altitudes. A plumbing system returned unburned fuel and vapors to the tank from which they came.

The fuel selector control handles were on the console between the front seats and they were linked to valves in the fuel system plumbing. Each engine had a fuel selector that could be moved to ON, OFF or X FEED (crossfeed). The OFF position is obvious. With the selector ON, the engine drew fuel from the inboard tank on the same side. When X FEED was selected, the engine drew fuel from the inboard tank on the opposite side.

The airplane's operating handbook noted that when one engine is inoperative and the fuel selector for the operating engine is on X FEED, the selector for the inoperative engine must be in the OFF position. It also stated, "Do not operate with both selectors on ’X FEED.' Do not take off with a selector on ’X FEED.'"

At about 5:47, the pilot had been handed off from one controller to the next at Memphis Center and reported descending through 5,700 feet for 6,000 feet. He was given the altimeter setting for Paducah, Kentucky. About 30 seconds later, the pilot radioed, "Hello, this is two niner one, I've got problems, the, uh, I need to go to the nearest VFR airport both..." The rest was unintelligible.

The controller responded, "...Seneca two nine one, um, Memphis, I caught a little bit of that, you said you are having problems?"

The pilot replied, "Seems like I just lost both..."

At 5:51:12, the controller radioed, "Seneca two niner one, uh, you're cutting out, uh, mike thirty four airport, uh; Gilbertville, Kentucky, mike thirty four's eleven miles to the west." The airport was Kentucky Dam State Park Airport (M34). The pilot acknowledged and six seconds later radioed, "I see it, okay, I don't know what the hell is going on."

The controller then advised the pilot that the Gilbertville airport was now about 12 miles west of his current position.

At 5:52:01, the pilot radioed, "I don't know what's wrong, both my navs, both my mags went out, both engines are malfunctioning---everything's forward, it was running perfect, I have fuel, I just don't know, the right engine is out." The controller then asked for the number of people on board. The pilot replied, "I have five on board and I see the airport, I'm at, uh, again, my best glide speed here..." The rest of the transmission was unintelligible.

At 5:53:09, the controller asked, "...is your left engine still working?" There was no response. The controller tried to raise the pilot three more times and at 5:54:09, the pilot radioed, "Do you hear me, eight one two niner one." The controller replied, "Yes, sir, I got you, what's your status right now?" At 5:54:14, the pilot radioed, "Engines are not producing power. I don't know what's up. I do not see the airport, can you tell me the frequency of the airport?"

The controller told the pilot to stand by, and a few seconds later radioed that the airport was on frequency 123.0. At 5:54:43, the pilot asked, "Is there anything suitable; suitable landing area." The controller did not have good news: "No sir, that is the closest airport that is going to be to you---it's runway two seven---it's four thousand feet long, a hundred feet wide."

At 5:55:06, the controller advised the pilot that radar contact was lost. There was no response. The last radar returns showed the Seneca at about 2,700 feet and descending about 10 miles east of M34.

The controller tried three more times to raise the Seneca. When another aircraft came up on frequency, the controller told the pilot there was a Seneca about 20 miles ahead at twelve o'clock that had a problem and asked that pilot to try to raise the Seneca pilot on the center frequency. When that failed, the controller asked the pilot to try the M34 airport frequency. The Seneca wasn't on that frequency, either.

The other pilot reported seeing a flashing light about 15 or 20 miles ahead which appeared to be slightly higher, but could not be identified as the Seneca. At the request of the controller, the other pilot detoured to the vicinity of the Kentucky Dam State Park Airport, but was unable to see the Seneca or hear any radio transmissions or an ELT signal.

It was two days later when investigators got to the accident site, about eight miles east of M34. The airplane was upside down with the landing gear retracted. There was a strong odor of fuel. As the wreckage was moved, fuel flowed out of the left wing area. There were no pre-impact failures or malfunctions discovered with the engines or airplane systems.

The fuel selector for the left engine was in the X FEED position. The selector for the right engine was between the ON and OFF positions. Investigators believed the right selector's position was due to impact damage. When investigators looked at the positions of the fuel valves in the wings, they found that the left engine really was in X FEED and it had been using fuel from the right side. The right engine valve was in the ON position, also drawing fuel from the right tanks.

It was calculated that the usable fuel in the right tank would have been exhausted just about the time the pilot reported both engines had died. Investigators found further supporting evidence for the in-flight draining of fuel only from the right side. The yaw trim was positioned to the full nose-right position. It was likely the pilot turned it a bit at a time during the flight to compensate for the imbalance caused as the fuel was burned from only the right tank, making that side lighter.

Data recovered from two GPS receivers on the airplane pointed to a likely reason the left engine fuel selector valve was on X FEED. The operating handbook calls for the pilot to move each fuel selector to X FEED for a short time during taxi, then set each selector back to ON before takeoff. The GPS data indicated that the pilot taxied from the ramp at Tallahassee and onto the active runway in about three minutes. Investigators say that the short time span means he didn't stop to perform a complete run-up of both engines before takeoff. He would have had to skip a visual check of the fuel selector positions.

The GPS data showed that at 5:50:16 the airplane had slowed from 160 knots to 142 knots ground speed during its descent. The plane further slowed to a ground speed of 100 knots as the descent continued. At 5:51:25, the GPS data showed the airplane turning to a westerly heading while continuing to descend. The last GPS altitude recorded was 700 feet.

Using the operating handbook's performance charts, investigators figured that the airplane had been burning about 20 gallons per hour. The handbook says to allow 4.2 gallons for engine start, taxi and takeoff. They figured that during the two hours and 55 minutes from departure to the time of the accident, about 60 gallons of fuel would have been burned.

The NTSB determined that the probable cause of the accident was the pilot's failure to properly set the left engine fuel selector before takeoff and to recognize the incorrect setting during the flight, which resulted in fuel starvation and a loss of engine power on both engines.

With the fuel selector valves right there on the center console, you'd think the incorrect setting would have been obvious had the pilot only looked, even in dim night lighting. It's not as if he was a novice in airplanes equipped with fuel tank selectors. After all, the 48-year-old pilot held a commercial certificate and was a flight instructor. He was rated for single-engine and multi-engine airplanes, and was instrument rated. He had 2,300 hours, with just under 6 hours in the Seneca. A local mechanic told investigators the pilot had only flown the Seneca "a handful of times" before the accident flight. Still, even with low time in type, it's hard not to know where the fuel selectors are located when they're so close to your right hand.

A checklist was found in the wreckage. In looking at photos of what was found, I could see both sides had the Seneca's tail number at the top. There was no clue as to the source of the checklist. The size of the typeface appeared to be so small that I'm sure I would have trouble using it, especially at night. On one side was the "normal procedures" checklist. On the other side, with various sections outlined in red, were the "emergency procedures."

There were a dozen different categories of emergency procedures, none of which suggested the pilot try different positions of the fuel selectors to see if an engine would restart. It's speculation, but had the checklist contained a category headed "Engine Failure During Cruise" containing entries such as "check gauges for fuel quantity," "check fuel selector positions," and "set selectors to feed from fullest tank," the pilot might have begun to understand what was happening. He could have started burning some of that fuel in the left tank instead of remaining as bewildered as his radio transmissions suggest he was.

Peter Katz is editor and publisher of NTSB Reporter, an independent monthly update on aircraft accident investigations and other news concerning the National Transportation Safety Board. To subscribe, visit www.ntsbreporter.us or write to: NTSB Reporter, Subscription Dept., P.O. Box 831, White Plains, NY 10602-0831.