On Sept. 30, 2020, the U.S. Air Force released the report of the Accident Investigation Board about the mishap of an F-35A at Eglin Air Force base.
According to the report, the accident occurred at 21.26LT on May 19, 2020, when F-35A aircraft tail number 12-005053, operated by the 58th Fighter Squadron, 33rd Operations Group, assigned to the 33rd Fighter Wing, crashed 4,600 feet down the runway 30 (slightly left of the centerline) at Eglin Air Force Base (AFB), Florida. The pilot ejected safely from the aircraft (sustaining non-life threatening injuries) while the aircraft, valued at $175,983,949, rolled, caught fire, and was completely destroyed.
The F-35 set and held 202 knots calibrated airspeed (KCAS) throughout the approach and landing; as a consequence, the aircraft touched down approximately 50 KCAS faster than normal, and approximately 8 degrees more shallow than desired for landing, at a 5.2 degree Angle of Attack. After landing, the pilot attempted a recovery maneuver (go-around) and remained aboard the aircraft about 5 seconds before ejecting.
Here’s an excerpt from the AIB report:
The nose of the aircraft drove down at a high rate of speed and the nose gear contacted the runway immediately after the main landing gear. Next, the MA [Mishap Aircraft] experienced a significant nose-high bounce. After the initial bounce, the MP [Mishap Pilot] made stick inputs consistent with an attempt to recover and set a landing attitude. However, the MP’s stick inputs quickly fell out of synch with the aircraft pitch oscillations and aircraft control cycles. Two seconds after touch down, the MP set and held aft stick, which would normally bring the nose of the aircraft up. Approximately one second after commanding aft stick the pilot also commanded full afterburner on his throttle. Both of those actions are consistent with an attempt to establish an attitude that would have allowed the aircraft to take off and go-around for another landing attempt. The horizontal stabilizers remained in full deflection down, which would tend to keep the nose of the aircraft down, despite the pilot holding aft stick for three seconds. After being unsuccessful in the attempt to go-around after multiple and progressively worsening bounces, the MP released the stick to eject.
The AIB President found, by a preponderance of the evidence, that the mishap was caused first, by the MA touching down at 202 KCAS, and second, by the MA flight control surfaces, namely the tail of the aircraft, conflicting with the MP inputs upon landing, resulting in the MP’s inability to recover from the aircraft oscillation. The AIB President also found by a preponderance of the evidence that four additional factors substantially contributed to the mishap. The substantially contributing factors are: the MP landed with Speed Hold engaged and using an alternate crosscheck method, the MP Helmet Mounted Display misalignment distracted the MP during a critical phase of flight, MP experienced cognitive degradation due to fatigue, and the MP lacked systems knowledge on flight control logic.
Therefore, the root cause for the crash was the pilot setting a “speed hold” of 202 knots, about 50 knots in excess, the F-35 approach AOA was also 8 degrees shallower than the normal parameters for the gross weight at the time of landing. This caused the aircraft to bounce, with the nose of the aircraft rising “rapidly and excessively”. The pilot attempted to stop this attitude with a forward stick input followed by aft stick but “this began a series of multiple and increasingly violent pitch oscillations on combinations of main landing gear only, nose landing gear only, and all three landing gear”. In other words, this series of forward and aft stick inputs during the first two seconds after touch down created what is called PIO (Pilot Induced Oscillations) “quickly fell out of synch with commands from the Flight Control System (FCS) and, combined with nose gear bounce, resulted in the FCS becoming over-saturated.”
In the presence of large and aggressive stick inputs the flight control system, based on its Control Law (CLAW) logic, became saturated and unresponsive, and ultimately biased the flight control surfaces toward nose down. At the two second mark, the MP set and held aft stick to try to reestablish a landing attitude. The horizontal stabilizers trailing edges remained full deflection toward nose down despite the pilot holding aft stick. The MP was unable to overcome the nose-down bias in order to reestablish a landing attitude or execute a go-around. The MP reported feeling confused, helpless, and ignored. Three seconds of pilot flight control input was not enough time to overcome the saturation caused by two seconds of prior inputs. The flight control system failed to orient the aircraft to the appropriate attitude for a go-around, and thus avoid catastrophic loss of the aircraft. After three seconds of attempting to go-around, and after multiple and progressively worsening bounces, the MP released the stick to eject.
Another contributing factor was that the HMD misalignment distracted the pilot during a critical phase of flight.
The HMD misalignment, and subsequent conflict in flight data, consumed the MP’s attention and was a source of distraction from the final descent through flare and touch down. The MP had never experienced HMD misalignment at night. Worse yet, the HMD was misaligned low as opposed to high. The low alignment resulted in the aircraft coming in too high for landing, which conflicted with ILS and visual data. The MP was relegated to using the HMD data to set his aimpoint as the F-35 lacks an alternate instrument such as the HUD found in legacy aircraft. The MP fought his own instincts to push further into the darkness short of the runway to correct his trajectory. This contributed to the steep approach. F-35A simulator profiles train F-35 crews to fly HMD-out approaches, but do not emphasize HMD-misaligned approaches, which aggravated the impact of the MP’s HMD misalignment on the night of the mishap. The focus required to mentally filter the degraded symbology, green glow of the HMD projector, visually acquire nighttime runway cues, correct and then set the aimpoint, fight the “hand of God” effect of the darkness short of the runway, and monitor glide path trends, distracted the MP from engaging APC or slowing to final approach speed.
Various things also contributed to what the AIB report calls a cognitive degradation: mental fatigue due to the retasking ahead of the sortie; interrupted sleep; concern about the possible exposure to Covid-19 notified the day prior to the mission. Moreover, the HMD’s projection brightness, also known as “Green Glow,” became increasingly distracting as the approach continued.
The MP attempted to dim the brightness but green glow feedback intensified as the MA approached the runway, peaking at approximately 300 feet AGL. The MP reported having to squint through the green glow in order to pick up on the environmental cues in the runway environment. This further contributed to the MP’s level of distraction on final.
Breathing through the on-demand oxygen system , something that is called “work of breathing” also had an impact on the mental fatigue of the F-35 pilot.
Furthermore, lack of knowledge of the multiple flight control logic modes also played a role in the mishap. Despite the complexity, there is minimal discussion about these in the F-35 manual and relevant academics. Moreover, F-35A simulator models do not accurately represent the aircraft flight dynamics meaning that the simulator allows for high-speed landings and recovery after bouncing:
The MP reported having been able to land the aircraft at the same parameters as seen during the mishap event. Two members of the AIB team were also able to successfully land the simulator at the MA’s speed and attitude. The LM mishap technical report verified the disjoint between actual MA performance and the simulator model and stated, “the pitch rate sensitivity evident in flight was not observed in piloted simulation or initial F-35A, T/N 12-005053, 19 May 2020 24 attempts to match the maneuver with offline simulation.” If the MP had a better understanding of the CLAW logic and how the aircraft would respond to inputs, and did not have the negative learning from the simulator, he might have been able to recover the aircraft despite the high speed landing, which is why this is a contributing factor to the mishap.
The AIB report does not specify any restriction following the mishap, but for sure there are some lessons learned that will have to be translated into corrective actions.