Not on the checklist

Visual and aural warning systems, especially in larger and more complex aircraft, can back up the checklists and alert crew members to concerns they’re not positioned to observe directly, such as a failed alternator or declining hydraulic pressure. Mechanized cross-checks of configuration settings for the phase of flight can be helpful, especially for takeoffs and landings.

But neither checklists nor annunciators are completely infallible, no matter how long they’ve been in service, and a pilot’s sixth sense that something is wrong is too easily ignored once the aircraft starts moving in earnest.

Two recent accidents bring this weakness into focus: On August 21, 2019, a Cessna Citation 560XL departing Oroville, California, was destroyed by fire following a rejected takeoff. Both pilots and all eight passengers were able to evacuate the aircraft without injury before the grass fire ignited by the runway overrun and subsequent gear collapse reached the aircraft. The Part 135 charter flight’s intended destination was Portland, Oregon.

Both pilots told investigators that they completed the before- and after-starting checklists prior to beginning their taxi to Runway 2 and completed most of the taxi checklist before reaching the threshold. The pilot stopped short of the runway so the co-pilot could contact NorCal Approach by cellphone to obtain their IFR clearance while he performed the rudder bias check, which involves setting the parking brake and then increasing power to 60 percent on each engine in turn to verify that the corresponding rudder pedal automatically moves to compensate for asymmetric thrust in the event of an engine failure. The cockpit voice recorder (CVR) confirmed that the pilot did not call out those steps while the co-pilot was on the phone.

After receiving their clearance, the flight crew entered it into the flight management system. Airport surveillance footage showed them waiting at the hold-short line for 3 minutes and 44 seconds. Before taking off, they spent another minute and 30 seconds resolving a No Takeoff annunciator on the instrument panel with indications that the autopilot was engaged and the trim setting incorrect. After determining that an iPad hosting the company electronic flight bag had slipped off the co-pilot’s lap and activated the autopilot, they disengaged it and reset the trim.

Eight seconds after the engines spooled up, the co-pilot confirmed that takeoff power had been set, then observed that the airplane was barely moving and said that something wasn’t right. The pilot replied that the airplane was moving and directed the co-pilot to call out speeds. Sixteen seconds later the pilot observed that the airplane was using more runway than expected; the co-pilot called out V1 and then rotation speed, and the pilot confirmed, “Rotate.” However, the airplane didn’t respond when he applied back-pressure. He told investigators afterward that “it was just a weird sensation” as he pulled the yoke all the way back to his chest without effect. At that point he applied full reverse thrust and maximum braking, but the Citation rolled off the departure end of the runway and struck a ditch, separating the landing gear and sparking a grass fire that eventually consumed the aircraft.

Investigators found the wreckage just short of 2,000 feet past the departure end of the runway. Most of the airframe had been consumed by fire, including almost all the right wing, most of the fuselage, and much of the cockpit. Heavy rubber skid marks ran from near the hold-short line down the entire length of the runway, across the overrun area, a perpendicular runway, and a taxiway, and into the grass. Examination of the partially melted parking brake valve, which was found loose in the wreckage, confirmed the cause of the skidding: The activating lever was found in the On position, and a CT scan showed that its internal valves were closed. While the flight data recorder didn’t monitor the parking brake, it had clearly been engaged throughout the attempted takeoff roll. An NTSB performance study found that the “unexpected retarding force” of the dragging brakes generated a nose-down pitching moment that exceeded the jet’s elevator authority, making it impossible to raise the nose.

The CE-560XL’s parking brake is operated by a pull knob at the bottom left edge of the instrument panel, not visible from the right seat. A cable connects the knob to the lever on the valve body. When the toe brakes are applied, pulling the knob closes check valves trapping whatever pressure is already held in the brake lines. The model received its type certificate in 1998; the certification standard governing the parking brake was established in 1965 and only required the parking brake to stop the aircraft when full brake pressure was set and takeoff power applied to one engine. Testing with both engines at takeoff power was neither required nor performed. That standard, defined in FAR 25.735, also did not require any cockpit indication that the parking brake had not been released, and parking brake status was not incorporated in the logic controlling the panel’s No Takeoff warning. Although that regulation was updated in 2002, the 2003 model airplane was only required to conform to its original type certificate.

Investigators also reviewed the original manufacturer’s checklists and the operator’s adaptation, which was approved by the FAA in May 2017. Neither contained any specific reference to the parking brake after the instruction to set it before starting the engines. The taxi checklists included a general item for “brakes-check” that doesn’t specify the parking brake, and the operator’s flow diagram only indicated the toe brakes. The pilot told investigators that he’d normally have called out “parking brake on, parking brake off” during the rudder bias check, but kept quiet while the co-pilot was on the phone with NorCal. The NTSB attributed the probable cause of the accident to “the pilot’s failure to release the parking brake before attempting to initiate the takeoff,” but also cited “lack of a checklist item to ensure the parking brake was fully released immediately before takeoff” as a contributing factor.

The results of an eerily similar accident two years later weren’t nearly as benign. All four occupants were killed and the airplane was destroyed when a CE-560XL crashed and caught fire attempting to take off from the Robertson Field Airport in Plainville, Connecticut, on September 2, 2021. Four other people on the ground were hurt, one seriously. Witnesses described the takeoff roll as “going slower” than others and passing the midpoint intersection with the gear still on the ground. One reported seeing “a puff of blue-colored smoke” at the rear of the airplane as it reached the two-thirds point of the 3,665-foot runway, which ends just before a steep 20-foot decline. Another stated that it went off the end of the runway in a nose-level attitude; when the ground dropped away, its nose pitched up without initiating any corresponding climb. The right wing struck a power pole, causing a shower of “softball-sized” sparks, and the engine noise changed to a “grinding” sound as the jet began a series of pitch and roll oscillations, striking the ground about 850 feet past the power pole and sliding into a building.

Fire consumed almost all the wreckage except the empennage. Once again, tire skid marks were found along the last two-thirds of the runway, and the parking brake knob and the valve it controlled “were found in the brake set position.” The flight data recorder showed that both engines were running at 91 percent N1 throughout, while a 16-degree increase in elevator deflection produced only a one-degree increase in pitch attitude. The weight-on-wheels indication showed “on the ground” until the ground dropped away, at which point the pitch attitude rapidly increased to 22 degrees nose-up, activating the stick shaker.

In its probable cause report on the Oroville accident, the NTSB pointed out not only “the lack of a No Takeoff annunciation warning that the parking brake was engaged, and lack of a checklist item to ensure the parking brake was fully released immediately before takeoff,” but also cited “the flight crew’s delayed decision to abort the takeoff.” When something seems wrong, it just might be—in which case it’s often better to figure out what before things start moving too fast for second thoughts.

David Jack Kenny is a former staff member of the AOPA Air Safety Institute.