Descending Into The Fog

The accident took place on April 7, 2015, in Bloomington, Illinois. The airplane was a Cessna 414A Chancellor on a Part 91 personal flight. The airplane crashed about 2.2 nautical miles (nm) east-northeast of the Runway 20 threshold at Central Illinois Regional Airport (KBMI). It was about 1.75 miles east of the localizer centerline. The GPS altitude of the accident site was 854 feet MSL. The airplane impacted upright, in a nose low attitude. The Safety Board said there wasn't much of a horizontal debris path, which indicated the airplane had stalled and spun in.

The Probable Cause statement says the accident was a result of "the pilot's failure to maintain control of the airplane during the instrument approach in night instrument meteorological conditions, which resulted in the airplane exceeding its critical angle of attack and an aerodynamic stall/spin." It adds a couple of contributing factors: pilot fatigue; increased workload because the airplane's glideslope equipment had a problem; and the airplane being loaded outside of its aft balance limit.

The 414A was powered by two turbocharged Continental TSIO-520-NB engines, which had been upgraded to 325 horsepower through modifications under an STC. The propellers were three-blade models, constant speed, full feathering. The airplane was pressurized and seated seven people. Winglets, vortex generators and wing spoilers had been installed under STCs. Fuel capacity was just over 213 gallons, with 206 gallons usable.

The airplane originally had been registered in Canada. In 1986, it was brought back into the U.S. and given a standard airworthiness certificate and new registration number by the FAA.

The airplane had been flown to Indianapolis International Airport (KIND) in Indiana on April 6.

Bloomington, where the airplane was based, is on central time. Indianapolis is in the eastern time zone. At about 12:07 a.m. eastern time on April 7, the airplane took off from KIND on an IFR flight back to KBMI. The pilot had selected Lambert-St. Louis International Airport as his alternate and had received at least four weather briefings that day, including one fairly close to departure time.

The flight was cleared to climb to 8,000 feet and proceed direct to KBMI. Just after 11:44 p.m. central time on April 6, the pilot was cleared to descend to 4,000 feet. They were about 42 nm south-southeast of KBMI at the time. At 11:52:06, the pilot contacted Peoria TRACON, reported level at 4,000, and asked for the ILS for Runway 20 at KBMI. The controller told him to expect radar vectors for the ILS to Runway 2, and the pilot radioed, "If we could we'd like to do, uh, two zero." The controller said, "Sorry about that, uh, vectors ILS two zero" and then told the pilot to turn right heading 330 degrees. The controller told the pilot that "Information India now current," and the pilot acknowledged and added, "We'll take lower when able."

At 11:59:16, the controller cleared the flight to descend to 2,500 feet and the pilot acknowledged.

At one second after midnight central time on April 7, the controller told the pilot to turn left heading 290 degrees. The pilot acknowledged. Thirty-eight seconds later, the controller radioed that the flight was, "Five miles from EGROW (the outer marker) turn left heading two three zero, maintain two thousand five hundred until established on the localizer, cleared ILS runway two zero approach." The pilot's readback was correct.

According to the flight path as reconstructed by investigators, at 12:01:26, the 414A crossed through the final approach course while still on a 230 degrees heading. It then turned to a south heading and made course corrections on both sides of the localizer. At 12:01:47, the controller told the pilot to cancel "on this frequency or on the ground, radar services terminated and frequency change approved." The flight was 3.4 nm from EGROW, was on the localizer, and was holding 2,400 feet MSL.
According to a recording that investigators obtained from a private source, at 12:02:00, the pilot radioed on KBMI's common traffic advisory frequency (CTAF) that he was "coming up on EGROW, ILS Runway 20 full stop." There were no further transmissions from the pilot recorded on the CTAF or FAA controllers' frequencies.

Reconstruction of the flight path indicated that, at 12:03:12, the airplane was to the right of the localizer centerline when it crossed the EGROW at 2,100 feet. Radar coverage in the area was somewhat spotty, and at 12:03:46 the airplane descended below 1,500 feet and was lost from radar. At 12:04:34, it became visible on radar again about 1.7 nm north of the runway threshold at 1,400 feet. By 12:05:08, it had climbed to 2,000 feet MSL and maintained a south course. Then it began a left turn to the east and descended to 1,500 feet. Then it started climbing and, at 12:05:42, was again at 2,000 feet at a position three-quarters of a nautical mile west of the localizer centerline. Five seconds later, it was lost from radar at 1,800 feet. At 12:06:11, it reappeared on radar at 1,600 feet and climbed to 1,900 feet in nine seconds, a rate of about 1,800 feet per minute. It had descended to 1,600 feet when the final radar return was recorded at 12:06:25. That last radar return was positioned just about over the crash site.

The published elevation of KBMI is 871 feet MSL, and the elevation of the Runway 20 touchdown zone is the same. The runway is 8,000 feet long by 150 feet wide. On an ILS, the minimum altitude would be 1,071 feet MSL, 200 feet above the surface. For a localizer-only approach, the pilot should not descend below 1,260 feet, 389 feet above the surface.

Weather observed at 12:05 a.m. by the Automated Surface Observation System (ASOS) at KBMI included an overcast ceiling at 200 feet AGL; visibility one-half mile in light rain and fog; temperature and dew point both at 13°C; altimeter setting 29.98; and runway visibility range for Runway 29 variable from 4,000 to 6,000 feet. Both of the airplane's altimeters were found in the wreckage with Kollsman settings between 29.98 and 29.99.

Investigators found that the coaxial cable that feeds from the glideslope antenna to a splitter, which then sends the glideslope signal to two independent NAV/COM/GPS radios in the cockpit, was disconnected. The twist-lock BNC connector was not secured to the antenna. Investigators believe it was that way before the accident, and they point to other connectors in the glideslope antenna circuit being firmly secured. Without the cable firmly connected to the antenna, no usable glideslope signal would be available.

The airplane was equipped with a primary navigation display, which presented compass, map, flight plan and RMI data on a 4-inch screen. When they checked data for the accident flight, which had been recorded by the system, they found that both of the glideslope displays would have shown a large "X" with no glideslope pointer for the pilot to see. The investigators reviewed maintenance records for any signs of work on the glideslope cable or antenna, and found nothing. They could not determine a reason for the glideslope antenna disconnect.

Investigators did a half dozen telephone interviews with residents who lived near the approach path to Runway 20. One said the airplane "sounded like [it was] 20 feet above [the] house. Sounded louder than a jet does. Shook my house." Another said the aircraft flew over and "got extremely loud," and he could hear it climbing. He said he looked outside and "couldn't see 15 feet" due to the fog.

Another witness who lives about a mile and a half from the end of Runway 20 said the airplane was so loud it sounded like "it was in the room with us." He offered his opinion that the pilot was going to set the airplane down short of the runway, then throttled up and climbed.

Yet another witness said the plane was so loud it woke up her 5-year-old son. She said the airplane made two passes, and she thought, "Why are they crop dusting at night?"

The NTSB declared that fatigue was a factor in the accident, pointing to the pilot likely having been awake for about 18 hours at the time of the accident and that the accident took place about two hours after the pilot's normal bedtime. However, that's downplaying what investigators learned from another pilot who reported having a lengthy conservation with the accident pilot at an FBO in Indianapolis while waiting for their passengers. The pilot was very relaxed, he said, and did not appear to be fatigued or ill.

The pilot was 51 years old, a flight istructor, held an FAA second class medical certificate, had about 12,100 hours with 1,150 in the 414A, and was type-rated for Cessna Citation, Learjet 35, Sabreliner, Falcon 10 and Embraer Phenom jets.

Three investigators had a 45-minute interview with the accident pilot's wife. An NTSB medical officer took part. The wife painted a picture of a man who was not under stress, was consistently well-rested, and had no financial or other issues. She said he had not been tired or fatigued. He didn't smoke, and only occasionally had a glass of wine.

The pilot's wife said on April 6, he got up at around 6:00 a.m. and followed his usual routine. She said he had to be at work at 7:30. He told her that he would be working late and that she shouldn't wait up for him. The pilot knew it would be a long day.

My anecdotal experience is that when pilots know in advance to expect a long day, they usually prepare for it and incorporate rest periods into the day's routine. Although the NTSB examined his 72-hour history, medical background and experiences as a runner, I didn't see a chronology of his ground activities leading up to the accident flight. The NTSB did state that "there was no evidence of cumulative sleep loss, acute sleep loss, or medical conditions that indicated poor sleep quality for the pilot."

The recorded radio conversations did not show that the pilot was slurring his speech nor forgetting things. In fact, he corrected the controller who made a mistake in the requested runway and consistently gave correct readbacks. He had no trouble switching from frequency to frequency and made an accurate announcement of what he was planning to do when he came up on the Bloomington CTAF.

Calculations by investigators found that the airplane had taken off on its flight from KBMI to KIND about 160 lbs. overweight, with an aft CG outside of the normal operating envelope, and landed 287 lbs. over the maximum landing weight and with an aft CG outside of the limits. They calculated that the accident flight took off from KIND about 271 lbs. over the maximum takeoff weight, with a CG about 4.37 inches aft of the limit. They said at the time of the accident it was 366 lbs. over the maximum landing weight, and the CG was about 3.71 inches aft of the limit. The NTSB said this contributed to the accident, and it very well may have by making the airplane a bit less stable. However, the pilot got the plane safely to KIND, and who knows whether operating the plane outside of limits had happened before with no consequences. Even if it hadn't, the pilot would have experienced how the airplane handles with an aft CG during the flight to KIND, if the NTSB's calculations are correct.

The NTSB report points out that whatever the pilot was doing in the way of maneuvering in the final two or three minutes of the flight, he wasn't executing the published missed approach procedure. At the missed approach point for a localizer approach, 0.9 nm from the threshold, it's a climb straight ahead to 1,500 feet, then a right turn to 270 degrees and continued climb to 3,000 feet while en route to a holding fix. The accident pilot was turning to the east instead of the west.

No problems were detected with the engines, propellers, vacuum pumps or flight controls. The landing gear was retracted. Both throttles were found at idle, both prop levers were found full forward, and the mixture levers were found at intermediate settings. All control levers were found bent to the right.

The Safety Board used available data to create a simulation study for the accident approach. It found that when tracking inbound from EGROW, the airplane was one or more dots to the right of the localizer centerline, not a big deal. It was consistently one or more dots below the glideslope, which could be a big deal but not in this case. At almost the missed approach point for a localizer approach, the airplane crossed through the localizer from right to left while on a 90-degree turn to the east. The airplane's calibrated speed was 150 knots at EGROW, and it subsequently slowed to 80 knots a couple of times. The simulation study found that it stalled at least once while deviating to the east.

With no glideslope signal, and the ASOS reporting ILS minimums, a 12,000-hour ATP-rated pilot surely would know that descending to 389 feet on a localizer-only approach likely would not bring the runway environment into sight. Did the pilot deliberately go lower, thinking he'd be safe taking a look at the ILS minimum of 200 feet? Was the maneuvering off the localizer a deliberate attempt to feel out whether he was dealing with a ragged ceiling, which might afford him an opening somewhere? Was he relying on a panel full of sophisticated avionics to help him know where he was at all times?

Knowing that the NTSB concludes the accident was caused by "the pilot's failure to maintain control," and that fatigue, loading and the lack of a glideslope were factors, surely is useful information. But it doesn't explain the most important thing: why a healthy and serious ATP-rated pilot with 12,100 hours got into this situation in the first place.

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.