HondaJet Runway Excursions Raise Questions

Honda Aircraft’s HondaJet has experienced eight runway excursions in the last 22 months at the time of this writing. Are these merely statistical coincidences attributable to individual pilot errors, as opined by supporters of the jet, or do these reflect a design with challenging handling characteristics, as suggested by others? These are important questions best answered by solid data, objective investigations and the application of human factors engineering seeking to optimize the pilot-vehicle-interface.

The most recent runway excursion occurred on Dec. 6, 2024, during landing in Muskegon, Michigan. Information from the Aviation Safety Network indicated the Honda HA-420 HondaJet experienced an excursion when landing on Runway 24. The aircraft struck the 3,000 ft. remaining runway sign. The latest METAR reported winds were light at 3 kt. from 010, 10 sm visibility, an overcast ceiling at 4,500 ft., and a remark that a band of snow commenced at approximately 30 min. before the official METAR’s time of publication and lasted 11 min. The airport manager stated to local media that the aircraft sustained minor damage.

Unfortunately, a high-speed runway excursion after a takeoff attempt at Mesa-Falcon Field (KFFZ) on Nov. 5, 2024, resulted in five fatalities. The pilot and three passengers were fatally injured, along with the single occupant of an automobile that was struck by the aircraft. One passenger was seriously injured.

The NTSB’s preliminary investigation used security video to capture images of the aircraft accelerating on Runway 22L for about 3,000 ft. With 2,100 ft. of runway remaining, the aircraft began to decelerate. The publicly available ADS-B data suggests that the aircraft lost an insufficient amount of velocity before going off the end of the runway.

It overran the departure end of the runway, struck the airport perimeter fence and continued across a roadway, whereupon it struck a single vehicle.

The Secure Data card showed that the aircraft accelerated to about 130 kt. before it began to decelerate. Why was the takeoff rejected at an abnormally high speed? The preliminary report gave no indication. Flight control continuity was verified by the investigation team on site. The control column gust lock was located uninstalled. The aircraft brake and anti-skid system were examined, with no anomalies noted. No engine anomalies were noted. Clearly there are many questions for the investigation team to analyze.

Data Reveal

One of the advantages of business aircraft is the ability to efficiently fly to and from airports closer to the business’ operating locations. Often these airports have runways that are shorter and narrower than those served by air carriers. Additionally, many of the runways used by business aircraft lack the surface features (grooves, crowning, PFC) that lessen the effects of runway contamination. Business aircraft frequently operate into airports and runways with limited snow removal capabilities and unreliable runway condition reporting. This operating environment creates a higher risk for runway excursions, and it places a premium on a pilot’s landing performance assessment, a pilot’s ability to accurately fly a stabilized approach at the recommended speeds, as well as the jet’s ability to decelerate smoothly.

The Flight Safety Foundation’s Aviation Safety Network contains 26 runway excursion events involving the HondaJet that have occurred worldwide, including in the nations of Saudi Arabia, Brazil, Canada, Japan and the U.S. Fourteen have taken place since March 2022. Investigations are underway on six.

Is this number out of the ordinary? That is hard to tell. A true apples-to-apples comparison would be needed with other aircraft sharing similar weight and operating environment (runway lengths, widths, surfaces and pilot crewing, for example). Furthermore, that comparison would need to compare “x number of incidents per 100 landings.” Obtaining the number of landings for several other types of comparable aircraft is not possible.

One of the limitations with the accident and incident data is that we simply do not know how many “close calls” have occurred. Flight Data Analysis programs would yield insight into whether the accidents represent a few limited events, or more realistically, if these are the proverbial tip of the iceberg. This type of data has been utilized to track the effectiveness of changes in procedures and training, as well as detect if a specific pilot needed additional individualized training.

Unfortunately, business aviation is far behind the air carrier industry in the utilization of this important safety data collection and analysis system, even though it would benefit greatly. Regrettably, FOQA data was not available for this analysis, which will become obvious several times during the following discussion.

Controllability on the Ground

A common problem noted in the incident and ASRS reports was difficulty with controllability on the ground.

For example, NASA ASRS report #1966378 (dated January 2023) described losing directional control during landing rollout when attempting to correct back to the centerline. The PIC’s narrative reported a smooth landing by the SIC just slightly right of center line approximately 1,000 ft. from the threshold with good initial deceleration. The pilot in command (PIC) announced and deployed the speed brake for the second in command (SIC). The aircraft initially tracked straight, just right of center line. SIC applied braking after approximately 3 sec., whereupon the aircraft developed a strong drift to the right. PIC waited for correction, and then announced and took controls as the aircraft approached the side of the runway. SIC released and announced, "You have controls."

The PIC applied initial and increasing control and brake input to arrest the drift. However, the aircraft response to the PIC’s input was delayed, leading to more control and brake input. A swerve then developed to the left. The PIC released some brake and attempted control with rudder/nose wheel steering. A pilot-induced oscillation (PIO) developed toward the left side of the runway. The PIC then applied maximum braking. The aircraft began to skid to the right. The aircraft came to a stop on the edge of the runway. Ultimately, the aircraft stopped on the runway with only a left blown tire.

A formal investigation of a similar event was conducted by the Japan Transport Safety Board. The accident occurred on March 13, 2021, at the Kohnan Aerodrome in Japan.

The trainee, who was training to upgrade to captain level, started to correct for the aircraft’s deviation from the centerline while starting to depress the brake pedals to decelerate. The change in heading became greater than 15 deg./sec. due to repeated corrections. In addition, the aircraft tilted, resulting in a change of the weight-on-wheels (WOW) mode of the left landing gear into the air mode. The aircraft deviated from the runway and came to a stop in the grass. The captain and the trainee disembarked from the aircraft by themselves, and no one was injured. The accident investigation report concluded that the tires skidded and the travel direction could not be controlled because the aircraft was tilted and large lateral acceleration was generated due to the excessive corrections on the travel direction, resulting in the reduced capability of steering control and the main landing gear braking control, and leading to the lost control of the travel direction.

The FAA’s Airplane Upset Recovery Training Aid points out that: “aerodynamic principles do not change, but aircraft design creates different flight characteristics. Therefore, training and experience gained in one model or type of aircraft may or may not be transferable to another.”    

A pilot’s control input is based in part on their expectation of what the response will be in the aircraft.  Pilots are not adept at quickly and expertly knowing the predictable response characteristics of an aircraft that handles differently from their past experience.  The same would apply during other pronounced perturbations in an aircraft’s trajectory caused by brake abnormalities, tire skidding or crosswinds.

Dr. Sheldon Baron worked on aircraft control for decades at the famed NASA Langley Research Center and was subsequently appointed as a member of the prestigious National Academy of Sciences and National Research Council Committee on Human Factors. In the book, Human Factors in Aviation, Dr. Baron provided important insight and historical perspective into why innovative aircraft configurations with different dynamic characteristics have experienced persistent difficulties with PIOs.

Each of our limbs has specific properties in terms of its strength, response rate and ability to make large versus small movements. This is important when considering the use of the legs for steering control. The legs are large and capable of powerful movements. The legs are less adept at making the fine motions required for a jet with sensitive steering qualities. This is further complicated due to varying lengths of legs and seat adjustment, both of which will modify the legs’ reactions and place feet at potential angles which can inadvertently result in forward flexion (called plantar flexion).

Depending on foot placement, this could apply pressure to the upper portion of a rudder pedal, resulting in inadvertent brake pressure. Those with taildragger experience can quickly understand the absolute importance of foot placement to prevent the unwanted consequences of foot flexion. Thus, the ergonomics of the seat position, rudder pedal position, varying leg lengths and foot size are variables that can complicate the dynamics of an aircraft’s handling on the ground. This is further complicated because of the wide range of pilot experience levels flying this jet.

Can enhanced training on ground handling successfully “reprogram” a pilot’s control inputs, as some have mused? Probably not as much as proponents would like. Dr. Baron noted that unanticipated pilot-vehicle-interface problems were often accompanied by training difficulties. Important points in Dr. Baron’s chapter included the importance of designing the dynamic handling aspects of the aircraft with consideration of the human capabilities rather than vice versa.

Unfortunately, there are many aviation examples demonstrating the difficulty in retraining ingrained habits. Professional flight crewmembers with thousands of hours of flight experience and rigorous repetitious training in routine maneuvers have deeply ingrained motions that will be difficult to change, especially when time pressure is short, such as a crosswind landing or unexpected event during the high-speed phase of a takeoff, when a pilot is most likely to revert to previously ingrained habit patterns.   

When previously learned habit patterns are inappropriate for the current aircraft, environment or procedures, it is called “negative habit transfer.” The role of negative habit transfer cannot be lightly disregarded. The work of the late Dr. James Reason, former Emeritus Professor of Psychology at the University of Manchester, found that negative habit transfer increases the chances of human error by five times.

Stop Ability

Most business jets are equipped with ground spoilers, thrust reversers and anti-skid braking systems. The HondaJet lacks thrust reversers, leaving the aircraft reliant on a braking system to decelerate. Furthermore, the relatively small tires have only a small contact area with the runway surface. Earlier versions of the HondaJet did not have spoilers to reduce the lift after landing. This was corrected on the Elite II version.

The second notable commonality in the accidents and incidents involved problems with the jet’s stop ability. Without Flight Data Analysis it is difficult to accurately determine how often other pilots have experienced challenges with the stop ability of the HondaJet. A search of the NASA ASRS database found five reports revealing pilot difficulties with braking since April 2019.

The question of stop ability is illustrated by an accident on May 18, 2023, in Summerville, South Carolina. The NTSB’s preliminary report reveals that the pilot delayed departure with his five passengers to let rainstorms pass through the Summerville area. The pilot said he had landed at Summerville “hundreds of times” in both dry and wet conditions and never had an issue with lack of sufficient runway to land. Runway 24 at Summerville is 5,000 ft. long and 75 ft. wide with an ungrooved, asphalt runway.

The pilot said he flew the approach at a Vref of 120 kt. with full flaps and immediately applied full brake pressure upon touchdown. He said, "the brakes immediately began to pulsate in anti-skid mode and because of that, very little braking effort was being done to slow the plane despite full pedal pressure." He described the response of the anti-skid system as having a slower on/off cycle than he had experienced on previous wet runway landings where the system had a more rapid on/off response.  The left brake "grabbed,” and the aircraft nose suddenly yawed to left. This began a sequence of left and right skids.

The pilot said he used the rudder pedals to keep the aircraft on the runway and was able to straighten the nose out right before they went off the runway into the grass. The aircraft slid down a series of embankments before coming to a stop.  All six occupants were able to safely exit through the main cabin door. A post-crash fire destroyed the aircraft.

NTSB investigators noted tire marks on the runway that appeared chalky and extended for 1,550 ft. The tire marks began to make an “S” shape along the runway, and approximately 275 ft. before the departure end, the tracks entered the grass on the left side of the runway and subsequently returned.

A barely noticeable detail was contained at the end of the NTSB’s preliminary report of the Summerville accident that is worth noting: An investigator from the FAA’s Aircraft Certification Office (now called Certification Branches) with responsibilities for type certification, continued operational safety questions, investigation and service difficulties, was present.

Crosswind Handling Challenges

Stop ability and controllability on the ground are more challenging when crosswinds are present. The HondaJet’s unique design includes short landing gear, placing the wing very low to the ground. A slight bank on touchdown can result in a wing strike. While other business jets have a published “maximum demonstrated crosswind,” the HondaJet has a definitive crosswind limitation of 20 kt.

A warning in the aircraft flight manual advises: “For landings with a crosswind greater than 15 kt., the approach must be made using a crab technique. A landing in these conditions using a wing low method could result in a wingtip strike.”

The NTSB’s investigation of another recent runway excursion involving a HondaJet at Houston’s William P. Hobby Airport on Feb. 17, 2023, illustrates the convergence of these design features to make crosswind landings especially challenging. Data indicated that the aircraft was 14 kt. or more above the published landing reference speed when it crossed the runway threshold, and it touched down about 2,000 ft. from the threshold. The pilot applied aileron controls for the crosswind and applied the brakes. However, no braking action was noted, whereupon the aircraft drifted left and departed the runway’s surface. A portion of the ring wing separated.

The latest wind report prior to the landing indicated a 70-deg. crosswind at 15 kt., gusting to 25 kt. The corresponding crosswind gust component was about 24 kt., which exceeded the airframe’s crosswind limitation. It should be noted that the pilot made two requests with the approach controller to land on a different runway, but those requests were denied.

The brake system touchdown protection is designed to prevent brake application until wheel spin-up occurs to avoid the possibility of inadvertently landing with a locked wheel due to brake application. After WOW has been true for three seconds, power braking is enabled. The NTSB’s investigation found that the left and right WOW parameters transitioned from air to ground consistent with initial touchdown; however, the left WOW parameter transitioned back to air about 2 sec. later. The right WOW parameter remained on ground until the aircraft departed the runway pavement. It is likely that the lack of positive WOW parameters inhibited brake application due to the touchdown protection function and resulted in the pilot not observing any braking action.   

The excess airspeed, extended touchdown and transient WOW parameters were consistent with the aircraft floating during the landing flare and with the application of aileron controls for the crosswind conditions. The aircraft was not equipped with wing-mounted speed brakes, which would have assisted in maintaining WOW during the initial portion of the landing.

The NTSB determined the probable cause to be the pilot's loss of directional control during landing. Contributing to the accident was the pilot's decision to land with a crosswind that exceeded the limitation for the aircraft.

Crosswinds were noted as contributing factors in five of the incidents and accidents listed in the Aviation Safety Networks database, although once again I must stress that we do not know if these are rare events or if these instead reflect a systemic trend that would be revealed in a good Flight Data Analysis program. Nor would we know if special-focus training efforts have produced measurable positive outcomes.

Waiting for Investigations

The FAA recently released Information for Operators (InFO 24014) encouraging all organizations to incorporate safety management systems (SMS) as a standard business practice. The SMS should include a process to identify safety hazards and assess the associated risks, a process to implement remedial action when necessary, and provisions for continuous monitoring and regular assessment of safety management activities.

Flight Data Analysis, which forms an important component of many SMS in the industry, would have provided engineering-quality data for in-depth analysis to the questions evident in these incidents and accidents. This oversight needs to be corrected, not just for this particular jet. Its usefulness would be invaluable to all of business aviation.

This author submitted a series of questions to the Honda Aircraft. These questions centered upon whether the manufacturer was considering any design changes or enhancements to improve the stability, handling characteristics and stopping ability of the HondaJet.

Honda Aircraft responded quickly. “Honda Aircraft Co. is always seeking continuous improvement and looking for additional opportunities to enhance the customer experience. As a company, safety is our highest priority, and every model of the HondaJet has been engineered to a high and exacting standard. It is our practice to assist NTSB investigations where a HondaJet is involved and are sure to take appropriate action after a careful review of any recommendations that are made.”

The recurring pattern certainly begs the question if fundamental design and/or training deficiencies need to be corrected. Are the expectations and control techniques common in other business jets not necessarily transferable to flying the HondaJet? Will any of these investigations connect the dots between the other incidents? Does the inclusion of an investigator from the FAA’s Certification Branch in the Summerville investigation portend a possible recommendation for a reexamination of the pilot-vehicle interface?

We will have to patiently wait for the final reports of the investigations underway for answers to these many questions.