Over the past four-and-a-half decades, Robinson helicopters have periodically broken apart in flight due to what accident reports often describe as “undetermined reasons.” In most cases, investigators have established that these catastrophic break-ups were initiated by mast bumping: a phenomenon unique to helicopters with two-bladed main rotor systems in which the down-flapping side of the rotor hub contacts the rotor mast. In the absence of cockpit video or flight data recorders, however, investigators have generally been unable to determine the precise sequence of events that led to the mast bumps — until now.
On June 30, the Australian Transport Safety Bureau (ATSB) released its final report on the in-flight break-up of a Robinson R66 helicopter over the water near Hawks Nest, New South Wales, in October 2023. It is the first such accident for which investigators had access to cockpit video and flight data, thanks to a factory-installed recording system developed by Robinson Helicopter Company and a member of the public who discovered the dislocated recorder on a beach near the accident site and handed it over to authorities. Data from the recorder provided conclusive evidence of what happened in the final moments of the flight, which killed the pilot and a small dog that was riding on the back seat.
The uncertainty surrounding previous mast bumping accidents has created a sometimes-bitter divide in the helicopter industry: between those who blame the design of Robinson helicopters for these fatal crashes, and those who blame the pilots. The Hawks Nest accident does not neatly resolve this debate. Instead, it provides a disturbing example of how an unforgiving design, combined with pilot distraction and complacency, can lead to tragedy — and lessons that all pilots of Robinson helicopters should take to heart.
Flying on autopilot
According to the ATSB report, the accident flight commenced shortly before 9 a.m. on Oct. 26, 2023. Departing Cessnock Airport, the helicopter initially flew east toward Newcastle, which is north of Sydney on the New South Wales coast. The pilot, who had logged about 1,125 hours of flight experience, then requested and received clearance from nearby Williamtown air traffic control to follow a coastal visual flight rules (VFR) route north towards his destination of Wallis Island Airport.
The helicopter was equipped with a Genesys HeliSAS two-axis autopilot which maintained the aircraft’s attitude using pitch and roll control inputs only, leaving the pilot responsible for managing the collective and yaw pedals. The autopilot was engaged during the flight and maintained the pilot’s selected headings and altitudes.
Heading north along the coast, the autopilot held an altitude of around 500 feet above mean sea level (AMSL) until Anna Bay, when the pilot increased the target altitude to 900 feet. He also made several course corrections using the heading selection knob to maintain the aircraft close to the coastal VFR route. Indicated airspeed fluctuated between 110 and 120 knots (around 200 to 220 kilometers per hour).
The ATSB report states that “the pilot was occupied with non-flying related tasks for much of this time, specifically, mobile phone use and the consumption of food and beverages.” A snippet of video footage released by the ATSB shows the pilot holding a bowl containing a sandwich in his right hand, which would normally be used to manipulate the cyclic.
In an interview with Vertical, Robinson CEO David Smith contended that the ATSB’s dispassionate wording does not adequately convey the extent of the pilot’s distraction. “I sat and I watched the video and I’ve rewatched it a thousand times since then, and it’s troubling how much the pilot, in this case, was in their phone and was distracted by other things in the cockpit,” Smith said, suggesting that it was an instance of automation complacency.
While helicopter autopilots are often praised for their ability to enhance safety by reducing pilot workload, “as an engineer, that’s not what we intended when we designed these things, for people to be head down in an entirely different task,” he said.
Turbulent conditions
Meanwhile, the outside environment was sending warnings to the pilot that appear to have gone unrecognized. From the beginning of the flight, cockpit video footage shows small deviations in attitude and heading, consistent with the presence of light turbulence. Investigators estimated that by the time the helicopter reached the Newcastle area, it was flying in winds of around 25 to 30 knots (around 45 to 55 km/h), and scattered white caps were visible on the water.
The R66 pilot operating handbook (POH) warns pilots to not exceed 110 knots indicated airspeed (KIAS) except in smooth air and then only with caution. “In turbulence, use lower airspeed. If turbulence is significant or becomes uncomfortable for the pilot, use 60 to 70 KIAS,” the POH states. However, the pilot did not lower the collective to reduce power, which would have also reduced airspeed with altitude hold engaged on the autopilot.
As the helicopter passed Anna Bay it continued to follow the coast toward Yacaaba Headland, a mountainous formation connected to the mainland by a narrow isthmus. The headland rises steeply from the sea, forming a ridgeline from east to west with a peak elevation of 715 feet (218 meters).
Investigators estimated that the southerly wind at the time of the accident was blowing almost exactly perpendicular to the ridgeline. This created an area of orographic turbulence — turbulence created as air flows over and around terrain — behind and above the ridgeline. Theoretically, the pilot could have inferred this turbulence from the wind strength and direction, but it was also visible on the water north of the headland in the form of cat’s paws, dark patches on the sea surface that indicate the presence of localized gusts.
The helicopter passed over the southern shoreline of Yacaaba Headland with an indicated airspeed of 115 knots and the pilot still engaged in eating. Two seconds later, the aircraft encountered turbulent air and began to roll right, pitch nose down, and climb.
With his right hand holding the sandwich, the pilot placed his left hand on the cyclic control cross bar and began to make cyclic control inputs to counter the uncommanded aircraft attitude movements. Investigators noted that it was the only time during the flight that he used his left rather than his right hand to manipulate the cyclic, “an unconventional method of cyclic control” that “likely resulted from the pilot’s engagement with non-flight related tasks using their right hand immediately prior to encountering turbulence.”
After the initial roll to the right, the helicopter pitched nose up and rolled left, then right, then left again while continuing to climb. About 14 seconds after the initial encounter with turbulence, the helicopter pitched down to near level. Investigators determined that the turbulence directly contributed to a low-G condition that was exacerbated by a forward cyclic input by the pilot, which resulted in the weight of the helicopter being unloaded from the rotor disc.
As is typical for Robinson helicopters in low-G conditions, the aircraft began rolling to the right. Investigators explained the right roll as being primarily due to the downward force created by the helicopter’s asymmetrical horizontal stabilizer, which is greater at higher airspeeds — hence Robinson’s admonition to slow down in turbulent conditions.
By the time the roll began, “there was probably insufficient time to recognize and respond with application of aft cyclic, and the roll continued to develop,” the ATSB concluded. Within 2.5 seconds, the helicopter had reached 90 degrees of right roll, and it continued accelerating in the roll to about 285 degrees before breaking apart.
Investigators were able to establish that the pilot input full left cyclic as the final right roll developed, an instinctive action that pilots of two-bladed helicopters are warned against as it increases the risk of mast bumping. The investigators noted that the same outcome may have occurred without the application of left cyclic as the helicopter became inverted. Yet, they also cited a remarkable counterexample — an incident in 2013 in which a flight instructor and student in an R22 helicopter experienced a low-G condition and subsequent right roll while on a training flight in New Zealand.
In that instance, the instructor was “very aware” of the potential for left cyclic to cause mast bumping and attempted to keep the rotor disc perpendicular to the mast throughout the roll. The aircraft completed a full 360-degree roll, descending around 800 or 900 feet before the instructor was able to pull out of the dive and return to straight-and-level flight. While the helicopter showed evidence of mast bumping and was significantly damaged, the instructor and student landed safely.
Underestimated risk
To Robinson, the video evidence from the Hawks Nest flight depicts a pilot who wasn’t focused on the essential task of managing the aircraft and wasn’t paying attention to the outside environment. Slowing down to Robinson’s recommended airspeed for “significant turbulence” — 60 to 70 knots — and placing both hands and feet on the controls would have given the pilot more time to react correctly to a low-G condition with gentle aft cyclic to reload the rotor disc. He could also have easily altered his course to avoid the predictable area of turbulence behind Yacaaba Headland, just as he had previously tweaked his course to follow the coastal route.
“The pilot wasn’t engaged in flying the helicopter, meaning you could see from the video visual cues about the weather, you could see indications that should have informed the pilot, even if he had made choices to use his phone or whatever momentarily, as pilots do,” Robinson VP of engineering Sean Doyle told Vertical. “There were a lot of cues to that pilot that it was time to pay attention, time to fully fly the helicopter, and he was so engrossed in other activities that he wasn’t making the decisions that a pilot needs to make in order to remain safe.”
However, it is not clear that the pilot would have fully appreciated the risk posed by turbulence even if he hadn’t been distracted by other activities. Pilots who do not routinely fly in mountainous environments may be less attuned to patterns of orographic turbulence. And some pilots may not fully understand the connection between turbulence and mast bumping due to a widespread impression within the industry, reinforced by decades of official guidance material, that turbulence is always secondary to pilot control inputs as a cause of reduced G-loading.
In January 1996, the U.S. Federal Aviation Administration (FAA) implemented an airworthiness directive (AD) that required R22 flight manuals to be updated with information about main rotor stall and mast bumping. The information explained that mast bumping can result from the low-G condition created by an abrupt cyclic pushover in forward flight, and that “high forward airspeed, turbulence, and excessive sideslip can accentuate the adverse effects of these control movements.”
Pilots were urged to depart the area or land if they encountered moderate, severe, or extreme turbulence, but continued flight in these conditions was prohibited only for pilots with less than 200 hours of helicopter flight experience. When the AD was published in the Federal Register, the FAA explained that it chose to exempt “those pilots with sufficient training and experience from limitations that might in some cases substantially restrict their Model R22 flight operations” — reinforcing the erroneous impression that mast bumping was primarily a problem for low-time pilots.
In March 1998, Robinson issued Safety Notice (SN) 32, warning against flying in high winds or turbulence. The safety notice recommended reducing airspeed to between 60 and 70 knots if unexpected turbulence was encountered, and told pilots to avoid flying on the downwind side of hills, ridges, or tall buildings where turbulence will likely be most severe.
The original version of the safety notice did not mention mast bumping. Only in a later revision of SN-32 did Robinson explain that “a pilot’s improper application of control inputs in response to turbulence can increase the likelihood of a mast bumping accident.”
Over the past decade, Robinson has made further amendments to its POH, safety notices, and training material to emphasize the hazards of turbulence. However, the ATSB was not convinced it had done enough to warn of the potential for turbulence-induced low G by the time of the Hawks Nest accident in 2023.
“Robinson Helicopter Company operational guidance warned pilots of the risk of inappropriate forward cyclic inputs resulting in a low-G condition but did not identify the potential for a turbulence-induced low-G condition without forward cyclic input,” the ATSB states in its report, reiterating that its analysis of the recorded flight data indicated that turbulence directly contributed to the reduced G loading.
In a press release accompanying the report, ATSB Director Transport Safety Stuart Macleod emphasized: “In this case the pilot’s response to encountering the low‑G contributed to the development of the in‑flight break‑up, but we would caution all Robinson helicopter pilots that a low‑G condition can result from turbulence directly, and pilots must be ready to respond appropriately.”
Sharing lessons
Robinson CEO Smith said that since the accident, the company has ramped up its presence in Australia, including through engagement with the Australian Helicopter Industry Association. “We have a desire to work collaboratively with them to make sure that the training and lessons learned from this accident are addressed throughout the pilot preparation side of this,” said Smith.
The company has already rolled out a new avionics and autopilot training course at its headquarters in Torrance, California, which teaches best practices for operating these systems in various flight conditions. “For the purposes of a remote site like our Australian market, a very big market, we likely will end up using our approved instructors in the region,” Smith said.
Robinson is also revising several of its safety notices — which are used as the basis for much industry training material — to incorporate lessons learned from the Hawks Nest crash. SN-32 will contain new information about the causes and effects of turbulence and more explicit guidance on how to react to an unexpected turbulence encounter.
“We’ve really tightened it up with a series of bullet points, basically step by step: do this, do this, do this,” said Doyle. Additionally, he said, “we added a sentence to this that says some extreme conditions may be beyond the capability of the aircraft — do not fly if extreme conditions are anticipated.”
Robinson is also updating SN-41, which talks about pilot distractions. “This was originally written in 2013,” said Smith. “So if you think about this, in 2013 we did not have all of the digital devices we have today. It was the beginning of the world as we know it now. Now it is even more critical, and I really think we have an epidemic, because we see this in other cases . . . where people are not head-in-the-game, not really focused on the aspects that are required for the job.”
The company is also making revisions to one of its earliest safety notices, SN-11 on low-G pushovers, originally issued in 1982. According to Smith, this is being revised to reflect the company’s deeper understanding of the roll condition, based on studies that resulted in the replacement of Robinson’s original asymmetrical horizontal stabilizer with a new, symmetrical design.
The revised safety notice will clarify that cyclic pushovers are not the only way to enter a low G condition. It will describe what pilots can expect to experience under low G when flying with the different stabilizer designs, including a much more pronounced right roll with the asymmetrical stabilizer. Smith said that based on Robinson’s extensive flight testing, any roll rate that develops under low G with the symmetrical stabilizer is “extremely small,” and can be to either side depending on the airflow around the helicopter.
In its report, the ATSB found that the asymmetrical design “significantly contributed to the uncommanded right roll rate during low-G conditions and the risk of an in-flight break-up,” and “strongly encourages all operators of Robinson helicopters to fit the symmetrical horizontal stabilizer.”
The ATSB also called out the importance of the factory-installed video and flight data recorder, which it said was “invaluable” in understanding the factors that contributed to the accident. Accident investigation authorities have been calling for such recorders in small aircraft for years, and Robinson’s success in installing them on its light helicopter product line demonstrates that the technology is now viable for even the most weight-sensitive aircraft. According to Smith, Robinson believes that video and flight data from accidents like this one can play an important role not only in accident investigation, but in pilot education as well.
Smith said Robinson is currently in the process of selecting a simulation partner to develop a custom line of low-cost simulators, which it ultimately hopes to deploy with its dealers around the world. These will be used to provide pilots of Robinson helicopters with experiential, scenario-based training, informed by real-world accidents and incidents.
“The primary purpose is things like this, where when we see an accident like this, we can rapidly turn this into a scenario and push it to the fleet,” Smith said. “And if people are insured with people we know and work with for better rates, if they’re financed by Robinson, if they’re leased through a Robinson structure, we can enforce that these pilots should experience these scenarios in the device.”
