An Aircraft Built for Speed Loses Power

The original airplane, now in the EAA museum, was built in the late 1930s. It was stored, incomplete, when World War II loomed, and it never flew. Intended to compete in the Coupe Deutsch de la Meurthe races, it had a drastically tapered, forward-swept wing (with an aspect ratio of 3.3), a minimal empennage, and contra-rotating tractor propellers driven by two Bugatti supercharged straight-eight engines of 450 hp each, mounted amidships, one behind the other. Racing speeds in those days were around 300 mph—the dominant competitors in Europe were carefully streamlined, but conventional Caudron monoplanes with six-cylinder inverted, inline engines of around 300 hp. Clearly, the Bugatti, if it worked, would be faster. Besides, some people thought it was the most beautiful airplane ever.

Bugatti’s aeronautical engineer, Louis de Monge, packaged the pilot, the engines and their cooling radiators ingeniously, wasting nary a cubic centimeter of the fuselage’s slender, perfectly streamlined spindle. The modern replica used two Suzuki Hayabusa motorcycle engines, 1.3-liter straight-fours, nominally rated at 175 hp at 9,500 rpm. Their integral gearboxes were retained, set in sixth gear. Hydraulic clutches, intended to protect against harmonic resonance during start-up, connected the engines to slender drive shafts that ran forward on either side of the pilot to a speed reduction unit driving the two fixed-pitch props. Thus, the two engines and propellers were almost entirely independent of one another. The only potential point of failure common to both was the lubrication system for the nose gearbox.

Unusual for an event involving a single fatality and a unique airplane, the National Transportation Safety Board (NTSB) produced a detailed analysis. The circumstance was clear: the aircraft stalled during initial climb from the 13,500-foot runway at Clinton-Sherman Airport (KCSM) in Oklahoma. The airplane was equipped with half a dozen GoPro cameras that recorded its demise in granular detail. Using those recordings, the NTSB could dissect the accident/flight second-by-second, and reconstruct a test flight from a year earlier.

The first flight of Le Rêve Bleu—the name, conferred by the pilot, means “The Blue Dream”—took place in August 2015. The runway hop, intended to check stability and control, was successful, although during the rollout a brake pedal failed and the elegant airplane ended up nosed over in rain-soaked ground beside the runway. The second flight, in October, was a single circuit of the airfield. The airplane lifted off at 80 knots with both engines turning at about 6,000 to 6,500 rpm—presumably propeller-limited and well below their speed for maximum power. At 500 feet agl, it leveled out and slowly accelerated to about 110 knots. As it turned final, the engines continued to turn at around 6,000 rpm, although the throttle levers had been backed off considerably—behavior consistent with a propeller pitched for climb. 

Although the second flight was without incident, it was not the flight expected of a 2,800-pound, 350-hp airplane in the hands of a pilot fully confident in its performance. It suggested, in fact, an airplane with barely sufficient power. In any case, the pilot’s impressions were not made public. I have not found among the reports on various aspects of the project that are posted online any discussion of what took place or what changes, if any, were made during the 10 months that elapsed between the second and third flights.

The third flight was announced to be the last. The airplane was destined for a museum in the U.K. This must have been a disappointment to the many people who had contributed money, time, and effort to the project, hoping some day to see it—on YouTube, at the very least—roar past them at speed, a blue streak out of the past.

The pilot accelerated gradually, as if feeling out the airplane anew. He rotated at a little above 80 kias after having rolled nearly 8,000 feet. He retracted the gear immediately. The airplane climbed, but its attitude seemed somewhat nose-high. Then the angle of climb seemed to diminish, the right wing dropped, then came back up. A moment later the other wing dropped and the nose sliced to the left. The airplane was only 100 feet in the air when the stall occurred, and it was impossible to recover.

In-cockpit video revealed that about 30 seconds after the airplane became airborne the left engine’s rpm began to drift upward, approaching the redline. (“Left engine,” here, means the engine controlled by the left throttle lever.) The pilot pulled the left throttle back to idle and that engine rolled back to 7,000 rpm. He pushed the right lever fully forward but the rpm of the right engine did not change, but that of the left engine, surprisingly, surged upward again, briefly reaching 9,500 rpm.

During this sequence, which lasted 20 seconds, the indicated airspeed, which had never exceeded 85 knots, gradually bled off, and the angle of attack—displayed on a conspicuous digital indicator at the top center of the instrument panel—steadily increased, eventually reaching 18 degrees. The NTSB found that the probable cause of the crash was “the pilot’s failure to maintain airspeed during an engine anomaly...the reason for which could not be identified during post-accident examination.” Although the NTSB declined to say so, the “engine anomaly” was certainly a slipping clutch on the left engine.

Aviation usually avoids complex drivetrains. Each non-rigid transition from one component to another involves losses, and no matter how well engineered the separate components may be, their potential interactions are difficult to foresee. There was always a possibility, even a probability, that Le Rêve Bleu might suffer a failure of one of its engines or drives. With an empty weight of over 2,500 pounds, it was extremely heavy for its size. Its large wing area, 223 square feet, gave it a comparatively low wing loading, but its 27-foot span was a liability at low airspeed. So was the extreme taper, which would predispose it to a wing drop at the stall. Running at 7,000 rather than 9,500 rpm reduced the raw horsepower available from the engines by between 15 and 20 percent. The two-stage reduction gearing took away a few more percentage points, as did the inevitable inefficiency of a fixed-pitch propeller at an off-design airspeed. According to a 1945 NACA report on the interactions of contra-rotating propellers, the windmilling front prop could have peeled off another 10 percent of the thrust available from the other one.

It was foreseeable that the airplane would have trouble climbing at low speed; its drag was at a minimum at around 110 kias, which was the “blue line” speed marked on the airspeed indicator for single-engine flight, and so at 85 kias, it was “behind the power curve.” The single most important action the pilot could take at the first sign of engine trouble would be to get the nose down, even if this meant belly landing in the rough. He certainly knew this and most probably rehearsed the proper response in his mind.

So why, when he lost power on one engine, did he fail to maintain airspeed? Nobody knows. But the fact a 10,000-hour former fighter pilot, intimately familiar with his airplane and able to plan for the scenario, failed to execute the indispensable response is a caution to us all. We may not react as well as we imagine.

This article is based on the National Transportation Safety Board’s report of the accident and is intended to bring the issues raised to our readers’ attention. It is not intended to judge or to reach any definitive conclusions about the ability or capacity of any person, living or dead, or any aircraft or accessory.