Buttonology

The flight

Just before 11 p.m. on December 29, 2016, a Cessna 525C Citation CJ4 vanished less than a minute and a half after taking off from Cleveland’s Burke Lakefront Airport. A week later the first pieces of the wreckage were retrieved from the surface of Lake Erie. Major portions of the aircraft—including both wings, most of the empennage, the cockpit’s center pedestal assembly, and the right engine—were located under 40 feet of water and recovered during the two weeks that followed.

At 10:55 p.m., the pilot received clearance to take off from Runway 24R with instructions to turn right to a heading of 330 degrees and maintain 2,000 feet. The airplane lifted off just before 10:57 p.m. ADS-B position data showed that it began turning immediately beyond the departure end of the runway while climbing through 550 feet above the ground. The right turn continued as the jet climbed to 2,450 feet agl before nosing down into the lake on an approximately 030-degree heading. During that time, its cockpit voice recorder (CVR) recorded nine “altitude” or “Pull up!” warnings in addition to one for bank angle, plus an overspeed warning that continued through the last 15 seconds of the tape. The CVR also recorded two instances in which the pilot replied to ATC transmissions without pushing the push-to-talk switch, suggesting he might have been suffering overload.

The pilot

The airplane was flown by its owner, a 45-year-old private pilot rated in single-engine airplanes, multiengine airplanes, and helicopters. He’d received his CJ4 single-pilot type rating just three weeks before the accident, then completed a simulator-based recurrent training course nine days later. Of his 56.5 hours in type, only 8.7 were logged as pilot in command, which included his checkride. He was not, however, a new jet pilot. Before buying the CJ4, he’d owned a Cessna 510 Mustang that he’d flown for nearly 375 hours during the previous two years—almost one-third of his 1,205 hours of career flight experience.

The instructor who provided his CJ4 transition training described him as a “very sharp guy” who “came to the lessons prepared.” The instructor’s preferred procedure was to activate the flight director “at or prior to” the runway threshold, center the heading bug once lined up on the runway, and activate the autopilot after climbing to at least 300 feet. The instructor also observed that for pilots transitioning from the Mustang to the CJ4, “things happen fast,” especially “down low, things happen really quick.”

Before the accident, the pilot had flown his five passengers from Columbus to Cleveland to attend an NBA game. (The Cavaliers beat the Boston Celtics, 124-118.) At the time of the accident, he’d been awake for about 17 hours.

The Weather

While not technically instrument meteorological conditions, the practical effect was the same: a moonless night over a dark, featureless lake. Burke Lakefront reported a broken layer at 2,300 feet that the airplane presumably entered just before it began to descend. Visibility was nine miles underneath, with west winds of 22 knots gusting to 31.

The Airplane(s)

For all their family resemblance, distinctions between the Cessna 510 and 525 models aren’t limited to performance. The NTSB made particular note of several small but potentially significant differences in panel layout. In the Mustang, the autopilot engagement button is located near the center of the lower row of the flight guidance panel; an indicator light next to it illuminates when the autopilot is engaged. The autopilot engagement button of the CJ4 is second from the right in the upper row of the flight guidance panel, with no associated indicator. In both models, the primary indicator of autopilot engagement—the only one in the 525—is an icon in the upper row of the primary flight display (PFD).

And the PFDs themselves are configured differently, potentially complicating a tired pilot’s ability to monitor the flight. The CJ4’s PFD uses the familiar “inside-out” display, where the simulated horizon moves around a fixed symbol representing the airplane. The Mustang’s PFD uses the opposite arrangement: The horizon remains fixed, while the simulated airplane moves up, down, and sideways.

The investigation

Damage to the memory chip prevented recovery of any useful information from the airplane’s AReS II flight data recorder. Much of the instrument panel was never recovered, so the actual status of aircraft systems couldn’t be definitively determined. Taken in combination with the ADS-B-derived flight track, the series of altitude, bank, and airspeed warnings recorded on the CVR led NTSB investigators to conclude that the pilot probably tried to engage the autopilot as he’d been trained, then either failed to notice that it hadn’t happened or succumbed to spatial disorientation while attempting to hand-fly the jet at low altitude with negligible visual references at night. The NTSB finding of probable cause specifically cited “negative learning transfer”—CFI-school jargon for old habits carrying over—as having contributed to the crash.

But like many aviation disasters, this one was the product of a collection of factors that conspired to create a singularly unforgiving situation. In a more familiar aircraft, autopilot engagement wouldn’t have been in doubt. In good visibility in daylight or over well-lit terrain, the airplane’s trajectory would have been obvious and readily corrected. And a better-rested pilot probably would have been more attentive in cross-checking systems before takeoff and monitoring the aircraft afterward. Circumstances that leave little margin for error can exact a heavy price for mistakes that might otherwise seem trivial.