NTSB Factual Report

On October 5, 2001, about 1600 Pacific standard time, an experimental Ellsberg One Design, N55TE, nosed over during a forced landing about eight miles north of Bend, Oregon. The pilot, who was the sole occupant, received minor injuries, and the aircraft, which is owned and operated by the pilot, sustained substantial damage. The 14 CFR Part 91 aerobatic practice flight, which departed Bend Municipal Airport about 1545, was being operated in visual meteorological conditions. No flight plan had been filed. There was no report of an ELT activation.

According to the pilot, just after he entered the aerobatic box, and while still at cruise, the engine began to violently shake. He therefore shut the engine off, and began a glide toward a nearby road in order to attempt an emergency landing. During the descent, the entire aircraft continued to vibrate to such a degree that is was very difficult for the pilot to see, and he became concerned that the engine might separate from the airframe. When he reached the ground, the pilot attempted to land on a narrow two-lane road, but during the landing roll, the windmilling propeller continued to shake the aircraft to an extent that made it hard for the pilot to see the restricted landing area. During the rollout, one of the aircraft's main gear departed the side of the road and encountered soft terrain. The tire then began to sink into the soft terrain, and the aircraft nosed over onto its back. According to the pilot, a quick check of the engine area revealed that a propeller counterweight attach bracket (including the counterweight) had separated from its mounting collar.

A post-accident inspection revealed that the subject propeller was a non-certified 200 series Whirlwind composite. The composite blades were used with a conventional McCauley propeller hub, internal components, and counterweighs. Whirlwind Propellers had connected 500 series McCauley counterweights and their respective attach brackets to the propeller blade collar with two AN-6 bolts and two AN364 elastic stop nuts. According to the associated McCauley propeller assembly torque table and the manufacturer of the blades (Whirlwind Propellers), these bolts should be torqued to 27 to 30 foot-pounds (324-360 inch-pounds). Although the separated counterweight was never located, during the inspection, a calibrated torque wrench was used to see how much force was needed to loosen one of the AN-6 bolts that held the other counterweight attach bracket to the apposing propeller blade collar. The bolt broke loose at a torque of 210 inch-pounds (17 foot-pounds), which is approximately 115 inch-pounds less than the required torque. The torque on the second bolt could not accurately be determined as it would have changed when the first bolt was loosened.

According to the pilot, this propeller had been supplied to him by Whirlwind as a temporary replacement, while his propeller was undergoing a periodic inspection at the Whirlwind facility. According to Whirlwind, after the accident, they sent out a notification to all Whirlwind Propeller owners, asking them to check the subject bolts for proper torque. Reportedly, there have been no under-torqued bolts discovered, and no history of nuts backing off their installed torque specification.

NTSB Probable Cause

the separation of a propeller counterweight mounting bracket due to incorrect mounting bolt torque. Factors include significant powerplant vibration during the forced landing, no suitable landing area, and soft terrain next to the narrow road on which the pilot attempted his emergency landing.