Failure to Enter

This story begins with a question: Why?

On Aug. 26, 2011, four people died in the crash of an Air Methods Eurocopter AS350 B2 helicopter in a cow field near Mosby, Mo. Why?

Here is what the United States National Transportation Safety Board (NTSB) determined: James Freudenberg, the pilot of the emergency medical services helicopter, was distracted by text messages during his shift and failed to properly preflight and fuel the aircraft. Along with flight nurse Randy Bever and paramedic Chris Frakes, he departed on a patient transfer mission believing he had two hours of fuel on board; in fact, he only had half that. Well before he arrived at Harrison Hospital to pick up the patient, Terry Tacoronte, Freudenberg had realized his mistake, and once at the hospital, he began searching for a place to refuel. He identified Midwest National Air Center, 58 nautical miles away, and chose to depart for fuel with Tacoronte and the medical crew on board.

Freudenberg reported having 45 minutes’ worth of fuel when he departed the hospital, but he actually had around 30 minutes of fuel — not enough to make his destination, and far short of the 20-minute reserve fuel requirement demanded by Federal Aviation Regulations (FAR). About a mile short of the airport, at an altitude of around 275 feet above ground level (AGL) and an airspeed of around 115 knots, the helicopter’s engine flamed out due to fuel exhaustion. Freudenberg failed to successfully enter autorotation, and the helicopter impacted terrain roughly 10 seconds later.

This narrative suggests a lot, but it doesn’t fully answer “why.” For most of us as pilots, it’s hard to relate to Freudenberg’s decision-making. He had multiple opportunities to abort the mission, or call for fuel. There is no evidence that he faced external pressure to push on; his coworkers indicated that he would have suffered no consequences worse than embarrassment for admitting his mistake. NTSB testimony indicated that Freudenberg might have been anxious to return to the base for a dinner date — hardly a valid reason for gambling with the lives of three passengers. 

Every flight involves some element of risk; some type of gamble. Freudenberg’s decision-making is horrifying to us mostly because the extremely high risk — of running out of fuel — so obviously outweighed the benefits of completing the flight. Freudenberg must have known that his chances of an engine failure due to fuel exhaustion were better than even. Did he believe he could make a successful autorotation in a worst-case scenario? If he did, why was he flying at such a low altitude when the engine failure occurred? Was he simply overconfident in his abilities, or did he not understand how to stack the autorotative odds in his favor?

Simply overconfident, or also ignorant: as I pored over the NTSB records for the crash, these struck me as the two possible answers for “why” Freudenberg took the risks that he did. But as I thought about the accident, I also started thinking about more everyday risk scenarios. Helicopters routinely operate low and close to the ground — that tends to be where the work is, and when we accept a low-level mission, we make a calculated decision to accept the risks associated with it. But what if we’re cruising from point A to point B at 300 feet above ground level, when there’s really no reason for us not to be at 1,000 feet AGL or higher? We’re putting ourselves in the same aerodynamic position that Freudenberg was in, and, consciously or not, we’re making the same evaluation of our autorotative abilities. We’re gambling that we’re good enough to make that safe forced landing if we need to — so good that we don’t need to take even the simplest measures to mitigate our risk.

Are we really as good as we think we are?

In its report on the Mosby accident, the NTSB defines autorotation as the state of helicopter flight in which the main rotor system is being turned by the action of air moving up through the rotor, rather than by engine power (something that is readily practiced in flight by reducing engine power and allowing the main rotor system to freewheel). This straightforward definition is notable for what it leaves out, namely, any mention of an emergency. An autorotation is what we perform on the way to the ground following an engine failure, or a drive system failure, or perhaps a tail rotor failure. But autorotation itself simply entails maintaining positive control of the helicopter while ensuring that the pitch of the main rotor blades — and therefore the drag acting upon them — remains low enough to allow their continued rotation.

“An autorotation is a maneuver. It’s not an emergency procedure,” American Eurocopter chief pilot Bruce Webb told me one day in May at the company’s headquarters in Grand Prairie, Texas. According to Webb, “The industry has done a good job of teaching autorotations. What it has done less well at is teaching forced landings.”

Grand Prairie was the first stop in my quest to understand the “why” behind the Mosby accident, because American Eurocopter, as the manufacturer of the accident helicopter, played a key role in the NTSB’s investigation of the crash. And Webb’s distinction — between an autorotation as a maneuver and as an element of an emergency procedure — is central to the Mosby case. As a helicopter pilot employed by a FAR Part 135 operator, Freudenberg had been required to demonstrate competency in autorotations during his initial and recurrent training. However, his failure to perform an autorotation in the context of a forced landing was fatal.

Without a flight recorder on board the accident helicopter, it is impossible to say exactly what happened during its final minute of flight. Inspection of the aircraft wreckage indicated that the helicopter impacted the ground in a 40-degree nose-low attitude, at a high rate of descent. There was minimal damage to the main and tail rotor blades, consistent with low rotor r.p.m. at the time of impact. Examination of the helicopter’s Turbomeca Arriel 1D1 engine confirmed that the engine had flamed out, but also revealed blade curling of the axial compressor, indicating that the gas generator was still turning at a high r.p.m. at the time of the crash. Based on known coast-down times for the Arriel 1 series, Turbomeca concluded that the helicopter must have struck the ground within the first 10 seconds after flameout.

All of this evidence points to main rotor blade stall, a topic that is covered exhaustively in virtually all primary helicopter flight training. When Freudenberg failed to enter autorotation, high pitch and drag on the main rotor blades reduced their r.p.m. to the point of aerodynamic stalling; suddenly robbed of lift, the helicopter plummeted to the ground. This much can be inferred from the wreckage. What is not clear is what actions Freudenberg did or did not take that led to this unhappy condition, and speculation as to this point became one of several focuses of the NTSB’s investigation.

Pete Gillies, the chief pilot of Western Helicopters in Rialto, Calif., is among those who believe that the Mosby crash can be explained by Freudenberg’s failure to input adequate aft cyclic in reaction to the sudden loss of power following engine flameout. “The crash was very simply the result of not bringing the cyclic back immediately — ‘instantly’ would be a better word — when the engine quit,” Gillies wrote to me in an email. At cruise airspeed, he explained, prompt aft cyclic would have tilted the main rotor disc back, allowing airflow from the helicopter’s significant forward airspeed to travel up through the rotor and provide an instant boost in r.p.m. Aft cyclic would also have been necessary to counter the severe nose-down pitching — and associated loss of r.p.m. — that would have occurred at that airspeed if Freudenberg lowered the collective to enter autorotation.

Gillies doesn’t discount the importance of lowering the collective to reduce drag on the main rotor blades and sustain autorotative flight: “If both hands are on both controls, great. Move them [the cyclic and the collective] at the same time,” he told me. But he is adamant about the importance of aft cyclic, emphasizing, “Don’t bring the cyclic back in time and pay a huge price. A stalled rotor system in an engine-out autorotation marks the beginning of the end. There is no recovery from this situation.”

As part of its investigation into the Mosby crash, the NTSB performed tests with American Eurocopter pilots in the company’s AS350 full-flight simulator. I recreated these tests during my own visit to Grand Prairie. I had flown the simulator previously as part of American Eurocopter’s Inadvertent IMC (Instrument Meteorological Conditions) training course (see p.64, Vertical, April-May ’12), so after a brief reintroduction to the simulator, I established myself in cruise flight at around 115 knots and 275 feet AGL. I was then given a series of simulated power failures from this airspeed and altitude — the same configuration that Freudenberg had put himself in at the time of the crash.

When I did nothing in response to the power failure, rotor r.p.m. decayed instantly and I impacted the ground within about five seconds. When I responded with aft cyclic only, the aircraft climbed briefly before rotor r.p.m. decayed with the same result. When I responded only by lowering the collective, the helicopter nosed over and we plunged to the ground like a lawn dart. It was only after the coordinated use of significant aft cyclic and down collective in response to the power failure that I was able to perform a successful autorotative landing (and then only after some practice; 275 feet is not a lot of altitude to work with).

As the simulator tests showed, there are a few different ways in which Freudenberg could have failed to enter an autorotation, and it’s quite possible that he simply froze on the controls, unable to believe he had run out of fuel just a mile short of his destination. In its report on the accident, however, the NTSB gives particular consideration to the possibility that Freudenberg failed to input sufficient aft cyclic, as Gillies suggests. The report observes that helicopter pilot training in general tends to emphasize the importance of lowering the collective to enter autorotation, often to the exclusion of other control inputs. For example, it points out that the United States Federal Aviation Administration’s (FAA’s) Helicopter Flying Handbook — the FAA’s primary resource for helicopter flight training — has very little information about the entry stage of an autorotation, and does not mention the importance of aft cyclic when entering an autorotation from high-speed cruise flight.

Of course, no pilot learns how to perform an autorotation simply by studying the Helicopter Flying Handbook. Aft cyclic was not absent from my own primary training in autorotations: “down, right, aft” (for “down collective, right pedal, aft cyclic”) was the mantra drilled into my head by my instructors when I was learning to perform autorotations in the Robinson R22. This type of deliberate, coordinated entry, however, was very much in the “autorotation as maneuver” category. And although I did go on to gain practice in responding to engine failures simulated by unexpected throttle reductions (as well as practice in simulating them, when I was working as a flight instructor), it was the flight school’s policy that none of these take place at airspeeds above 80 knots. When you’re entering an autorotation at 75 knots, it simply doesn’t take a great deal of aft cyclic to maintain a level pitch attitude and arrive at the R22’s recommended glide speed of 60 to 70 knots.

Autorotation entry at 75 or 80 knots is probably a reasonable thing to practice in an R22, which is often hard-pressed to cruise much faster. But what about in larger aircraft with higher cruise airspeeds, such as the AS350? Freudenberg came to Air Methods a year before the accident at an autorotational disadvantage: as a 2,000-hour former U.S. Army pilot, all of his recent flight experience had been in Boeing AH-64 Apaches, twin-engine helicopters in which the prospect of total power loss is more remote than it is in single-engine aircraft. So it is unlikely that Freudenberg would have been particularly sharp on autorotations during the final stages of his Army career. As part of his new-hire training with Air Methods, he received 4.2 hours of flight instruction in the AS350 B2, including training in autorotations. Based on interviews with other Air Methods pilots, however, the NTSB concluded that most of these practice autorotations were probably entered at speeds of around 80 knots, consistent with traditional flight training.

Moreover, it appears that Freudenberg never experienced an actual reduction in power during these practice autorotations. Like many operators who don’t want to risk damaging revenue-generating aircraft, Air Methods generally does not practice full touchdown autorotations (that is, autorotations that end in a landing on the ground). Because the B2 model has a floor-mounted fuel control lever without an idle detent, Air Methods conducted autorotation practice in these aircraft without retarding the throttle, lest it accidentally go fully closed and prevent a power recovery. 

Because I was having a hard time understanding what a practice autorotation without a power reduction could accomplish, I asked American Eurocopter assistant chief pilot Mike Newell to demonstrate one in an actual AS350 B2. We did a lap around the traffic pattern at Grand Prairie, then Newell entered a smooth, fast descent toward the runway from an altitude of about 700 feet AGL. The descent did indeed have the “feel” of an autorotative glide, and Newell said that power-on autorotations are a good way to warm up pilots who might be anxious about the maneuver. However, the entry certainly bore little resemblance to the reduced-power autorotations I also practiced during the flight, which, especially at higher airspeeds, required me to react to a significant drop in r.p.m. and a noticeable right yaw.

This difference also stood out to the NTSB, which noted that Freudenberg did not have practice in entering autorotations from cruise airspeeds, or the opportunity to experience symptoms of a true power loss. In its report on the Mosby accident, it concludes, “the autorotation training that the pilot received in the Eurocopter AS350 B2 was not representative of an actual engine failure at cruise airspeed and likely contributed to the pilot’s failure to execute a successful autorotation.” This assessment goes a bit further toward explaining “why” Freudenberg made the gamble he did. Overconfident he certainly was, but it also appears that he didn’t have the experience necessary to make a realistic assessment of his own abilities.

Pete Gillies believes that all pilots should be taught “cyclic back” as the initial response to an engine failure. “In no way should the movement of the cyclic be tied into the pilot’s opinion of airspeed, power being pulled, inertia of the rotor system, etc.,” he told me, arguing that pilots don’t have time to make fine judgments about control inputs in an emergency situation. “The movement aft of the cyclic should be as automatic as lowering the collective. Just do it!”

Others in the helicopter industry, including Bruce Webb, are not prepared to go this far, fearing that an emphasis on aft cyclic could cause pilots to neglect the importance of down collective in responding to a power failure. Instead, Webb believes the emphasis should be on a coordinated application of flight controls appropriate to the flight regime — which is more or less the position taken by the NTSB. In its report on the Mosby crash, the agency highlights the necessity of “immediate and simultaneous control inputs” in responding to a loss of power, but limits its discussion of aft cyclic to situations involving high-speed cruise flight. The report doesn’t address power failures in, for example, out-of-ground-effect hovers, which require a very different control response.

What Gillies, Webb, and the NTSB agree on is that many helicopter pilots are poorly prepared for the reality of a power failure, just as Freudenberg was. Which naturally raises the question: what can they and their employers do about it? More realistic training is the obvious answer, but in an actual helicopter, this demands two things. One is a skilled and confident instructor (certainly more skilled and confident than I was when I began teaching autorotations, with just 250 hours of total flight time). The other is an organization willing to accept the risk of aircraft damage that comes with practicing reduced-power autorotations, particularly touchdown ones.

Not all helicopter operators can meet these requirements, but manufacturers’ schools and some specialty training organizations can. “Because we train all the time, we’re able to perform maneuvers to a level of proficiency that most pilots don’t normally have the opportunity to achieve,” said Newell, describing American Eurocopter’s flight instructors and their comfort with touchdown autorotations in particular. “If you could get someone here once a quarter, they would have all the confidence in the world.” For many commercial operators, quarterly training at that level simply isn’t practical, but annual training should be. For example, I later spoke with Greg Hildenbrand, executive director of Life Star of Kansas, an air medical operator that uses two AS350 B2s and one AgustaWestland AW119 Koala. Because insuring LifeStar helicopters for touchdown autorotation practice is prohibitively expensive, Hildenbrand sends his pilots to the Era Training Center in Lake Charles, La., once a year instead. Such training isn’t cheap, he admitted, but “one of the first things I heard when I entered the industry was that the best safety program is a well-trained pilot,” he said.

Yet even the most skilled instructor faces certain practical limitations when simulating engine failures in a real helicopter. For this reason, the NTSB is recommending that all air medical pilots also receive scenario-based training in simulators or flight training devices (FTDs), which can more accurately replicate the symptoms of an actual engine flameout. In its report on the Mosby crash, the agency concludes that if Freudenberg had received autorotation training in a simulator rather than a helicopter, “he would have been better prepared and might have more effectively responded to the engine failure during the accident flight.” In fact, after the accident occurred, Air Methods implemented simulator training for all of its AS350 pilots, in addition to revising its autorotation training guidance to emphasize the importance of aft cyclic.

Like many pilots, my first experiences in simulators and FTDs left me disoriented, queasy, and unconvinced of their value. But I got over my discomfort — and became a huge fan of synthetic training devices — after completing a full transition course in the CAE 3000 AS350 FTD in 2011 (see p.78, Vertical, Oct-Nov ’11). By the time I completed American Eurocopter’s Inadvertent IMC training course later that year, I was ready to recommend simulators to everyone. Nevertheless, I decided to pay a visit to the experts at FlightSafety International’s Dallas Learning Center to get their take on the use of simulators in autorotation training.

Greg McGowan, FlightSafety senior vice president of operations, admitted that simulator training can be “a very humbling experience” for anyone unaccustomed to it. Although FlightSafety has been training helicopter pilots for 30 years, pilots with no previous exposure to simulators are often as wary about them as their predecessors were when they stepped into the first Bell 222 simulators three decades ago. Part of this is simply lack of familiarity and comfort with the machines. But as Pat Leone, FlightSafety director of standards, pointed out, “a lot of helicopter pilots do not want to display to an instructor an incapacity or a weakness” — which is often what comes out during emergency procedures training that is much more realistic than what a pilot is used to. Yet such training can also be a confidence-builder in the long run, Leone said: “It’s an ego booster when people find they can manage it.”

I spent about an hour with Leone and assistant director of training John Healey in FlightSafety’s Eurocopter EC135 full-flight simulator, which was developed in partnership with Metro Aviation (see p.92, Vertical, Oct-Nov ’12). During the flight, we practiced complete power losses (as twin-engine helicopters like the EC135 are, sadly, just as susceptible as the AS350 to running out of fuel). Healy also demonstrated an autorotative landing in the context of a fenestron failure, which brought up considerations unique to twin-engine aircraft. Finally, Leone, who was at the instructor’s station in the back of the simulator, gave Healey a surprise power failure as he was flying out of LaGuardia Airport to give me a virtual tour of New York City. Although his left hand was off the collective and gesturing toward the skyline at the time, Healey managed to enter autorotation and land us safely in an empty lot.

As Vertical columnist Shawn Coyle points out in The Little Book of Autorotations, it is difficult to realistically duplicate the flare and touchdown phases of an autorotation in a simulator. “Unfortunately,” he writes, “this is interpreted by experienced pilots to mean that the whole experience of autorotations in a simulator is not worthwhile training, when there really is a lot that can be learned with good economy of both time and money as well as improved safety.” My experience at FlightSafety demonstrated that simulators are outstanding for practicing the quick reactions and decision-making required in forced landing scenarios — for practicing, in other words, autorotations as emergency procedures. As McGowan put it, “I can’t imagine having a flight operation where there was a flight simulator available, and not using it.”

No story about autorotation training would be complete without a visit to the Bell Helicopter Training Academy (BTA) in Fort Worth, Texas, which for many years has been the industry’s mecca for power-off flight. The BTA’s experienced instructors don’t simply teach autorotations as maneuvers; they teach them as art. Under their guidance, students learn how the many variables in an autorotation — airspeed, altitude, rotor r.p.m., trim — can be manipulated to arrive safely at a designated landing spot, every time.

The day after my visit with FlightSafety, I made the pilgrimage to Bell Helicopter’s training facility at Alliance Airport, where winds were gusting to 30 knots. Senior flight instructor specialist Larry Sommers took me for a flight in a Bell 407 to the school’s practice area, a few miles north. There, he showed me how he introduces his students to full-down autorotations through a sequence of maneuvers — steep approaches, quick stops, and hovering autorotations — and let me practice a few straight-ins. Then, he demonstrated how he trains students in forced landing scenarios from more challenging flight regimes. For example, he had me set up an orbit over the practice area to mimic what a law enforcement helicopter might fly over the scene of an incident. After identifying the only “safe” landing spot, he rolled off the throttle mid-way through the orbit, entered autorotation, and showed me how to manipulate the flight controls to land where we needed to. “That’s what we try to do, is show [students] how to work the aircraft to their advantage,” he said. “That’s one thing we’re really good at, is expanding that envelope.”

This was my second visit to the Bell Helicopter Training Academy. My first had been on a Whirly-Girls scholarship in 2007, when, as a 1,000-hour pilot fresh out of flight instructing, I had been blown away by the knowledge, skill, and professionalism of the instructors. It immediately became my goal to someday achieve the same level of mastery in a helicopter — which, unfortunately, hasn’t happened. In my case, that’s largely due to the fact that I stopped flying full-time to become a magazine editor, but the Mosby accident demonstrates that many higher-time, working line pilots are also failing to meet that bar. A pilot with several thousand flight hours may become very good at the type of flying that he or she does day in and day out. When it comes to essential maneuvers like autorotations, however, it’s not uncommon for high-time pilots to be worse at them than they were in flight school.

It’s also not uncommon for them to be more complacent. It’s telling that people like Sommers — who are highly competent to make a safe forced landing from 275 feet and 115 knots — would avoid flying at that altitude to begin with. “I try to remind guys, bring them back to the basics,” Sommers said. “If I see someone flying too low, I ask them, ‘What would happen right now if your engine quit?’ ”

Pilots like Sommers appreciate how much the outcome of a forced landing owes to luck, in a way that Freudenberg did not. For every Michael Campbell — the New York Helicopter pilot who was celebrated as the next Capt. “Sully” Sullenberger after making a safe forced landing on the Hudson River on June 30 — there’s also an Alex Kelley, the pilot who, along with two others, died in the crash of an Air Methods AS350 B3 in Tucson, Ariz., in 2010. According to the NTSB report for that accident, Kelley apparently successfully entered autorotation from 800 feet AGL following engine flameout due to an improperly installed fuel inlet union. But he was forced to maneuver over a set of power lines on the way to a clear landing spot, and the helicopter impacted a concrete wall instead. Although the NTSB considered the recency of his training as a factor in the accident, it concluded, “Because the engine failed suddenly at a low altitude over a congested area, more recent training may not have changed the outcome.”

When reading NTSB reports, it’s tempting to look for reasons why such accidents could never happen to us. That is especially true in the Mosby case, and I think it’s fair to say that most pilots — no matter how inadequately trained, how fatigued, or how anxious to return to base — would never push a low-fuel situation to the extent that Freudenberg did. Yet even the most blatant cases of pilot error hold an uncomfortable mirror up to us as individuals, and as an industry. Simply put, we’re not as good as we think we are — but we have the tools we need to get better.