NTSB Factual Report

On July 5, 2001, at 1701 Pacific daylight time, a Bell 47D-1, N64580, experienced a rough running engine on short final approach and entered an autorotation to the airport at Riverside, California. The helicopter made a hard landing and sustained substantial damage. Zemlock Helicopters was operating the helicopter under the provisions of 14 CFR Part 91. The commercial pilot and one passenger were not injured. The personal flight departed Chino, California, about 1630. Visual meteorological conditions prevailed, and no flight plan had been filed.

The pilot stated that he and another pilot departed Chino and flew to Riverside to practice landings. The second pilot flew to Riverside, made the first landing, and turned control over to the first pilot for the next landing. They each took a turn flying the helicopter to landing.

As the first pilot descended through 175 feet above ground level (agl) on his second attempt, he noticed the tachometer bouncing back and forth between the yellow arc and the green arc. He could hear the engine running and continued his approach. He saw a needle split with the engine rpm (revolutions per minute) needle falling below the rotor rpm needle. He felt a loss of power and the engine began to run rough. He transitioned to an autorotation and initiated a flare at 50 feet agl. He pulled up on the collective, and hit the stop while several feet in the air. The helicopter dropped to the ground, bounced once, and then came to rest.

The main rotor continued to turn; but the engine continued to run rough and the pilot shut it down. He was unaware of any damage to the helicopter until he exited.

The helicopter and engine were examined. Both main rotor blades exhibited leading edge damage to their tips. The tail rotor drive short shaft was on the ground beneath the fuselage. Scuff marks were observed on the short shaft and the structure below the short shaft.

The helicopter was recovered and the engine was test run twice, but it continued to run rough. The upper crankcase cover was removed and the crankshaft was found cracked.

The Safety Board Material's Laboratory examined the crankshaft and prepared a factual report. Pertinent facts from the Laboratory's narrative report follows; the full report with photographs is attached to this report.

The crankshaft fractured through the cheek on the gear side of the third main bearing journal from the drive flange. The main and rod bearing journals were individually numbered beginning at the drive flange. The fracture was then located between the No. 3 main and the No. 5 rod journals. The only markings on the crankshaft were raised numerals iu07806 $ Bl0 on the side of one crank cheek.

Optical examinations of the fracture faces found beach marks and crack arrest lines indicative of fatigue cracking though the majority of the crank cheek. Further examinations established that the fatigue initiated at two origins on the aft radius surface of the No. 3 main bearing journal about halfway between the journal and cheek surfaces.

The fatigue propagated in two stages. The initial stage was on a plane at approximately 90 degrees to the radius at the origin (45 degrees to the longitudinal axis of the journal) and had penetrated about 0.3 inches into the cheek. The area of the initial fatigue was very smooth and flat with subtle arrest lines.

In the second stage of propagation the fatigue reinitiated at multiple origins all along the terminus of the first stage and propagated on a plane parallel to the journal axis. Prominent arrest lines, multiple ratchet marks, and a coarser surface characterized the second stage area. The overall extent of fatigue and the terminus of the second stage were obscured by mechanical damage to both fracture faces. However, it was estimated that more than 80 percent of the cheeks cross section was penetrated by fatigue.

An oil stain covered the aft radius but no damage was obvious. However, removal of the stain by vigorous wiping with a cotton tipped swab revealed wide spread surface damage to the radius. The damage extended from the journal surface up to the level of the origin area and completely around the journal. The surface damage consisted of circumferential scoring, intermittent tearing, and material flow. The radius also had a slight bronze hue. Further cleaning and inspections found similar damage and coloration on the forward radius of the No. 3 main bearing but no damage to other journal radii. A photograph showing the typical surface on other radii is included for comparison.

When viewed in profile using a back light projector the aft radius appeared uneven and slightly disrupted. Away from the fracture, the radius measured 0.136 inches and appeared typical of other measured main bearing radii. The other radii, however, displayed smooth continuous profiles without noticeable disruptions.

A longitudinal metallurgical cross section was cut through the aft half of the fracture slightly away from the fatigue origins. The removed section was electroless nickel plated, polished, and etched with 2 percent Nital (2 percent concentrated nitric acid in ethanol). The etchant revealed a case hardened microstructure consisting of tempered martensite. The visual depth of the case was about 0.025 inch. High magnification examination found a very thin (~0.0004 inch) white layer at the surface of the radius and along the manufactured surfaces of the adjacent cheek. The white layer appeared consistent with an iron nitride layer formed during nitriding (a case hardening process for steels where nitrogen is diffused into the steel to create a hard case). Microhardness surveys adjacent to the fracture and perpendicular to the radius found an average hardness of 42.8 HRC (Knoop) between 0.003 and 0.20 inches. At a location immediately adjacent to the fracture, the case area between the white layer and 0.003 inches deep was softer, averaging only 36.5 HRC (Knoop). Core hardness, measured well away from the surface, averaged 33.2 HRC.

A second longitudinal metallurgical section was cut through the aft radius of the No. 3 main journal at a location away from the fracture. The section showed the entire radius and a portion of the adjacent bearing surface. Etching the specimen with 2 percent Nital revealed a layer of disturbed material in the radius at the location corresponding to the fatigue origins. The layer measured about 0.0003-inch deep and appeared consistent with localized overheating and smearing of the surface. Further examinations found a large area of overheated material on the bearing surface of the journal away from the radius. This area measured 0.3 inches wide and 0.03 inch at its deepest, and contained multiple transverse intergranular cracks. Close examination of the microstructure within the overheated region found untempered martensite at the surface.

The direct visual examination of the No. 3 main bearing journal surface showed heavy circumferential scoring and roughening but no visual indications of overheating. The majority of the other journals, both rod and main, also displayed similar scoring but to lesser degrees.

Measurements established that the diameters of all main bearing journals were between 2.238 and 2.240 inches. All rod journals measured 1.927 inches in diameter. Conversations with a distributor for new Franklin engines indicated that standard bearing sizes for 6A-335 engines are 2.250 to 2.249 inches for mains and 1.9375 to 1.9365 inches for rods. Bearings for 0.010- and 0.020-inch undersize journals are available for both the main and rod journals. The measurements on the fractured crankshaft are consistent with 0.010 inch under size journals for both the mains and rods.

NTSB Probable Cause

The engine crankshaft's failure as the result of fatigue cracking initiated by surface damage and localized overheating due to the installation at overhaul of incorrectly sized bearings.