On May 3, 2004, at 2345 Pacific daylight time, a Bell OH-58A+, N397E, struck a power line during an autorotation to a residential street near Bakersfield Junior College, Bakersfield, California. The loss of engine power preceded the autorotation. The Kern County Sheriff's Department (KCSD) operated the helicopter as a routine patrol mission, as a public-use flight under the provisions of 14 CFR Part 91. The flying pilot (FP), a commercial helicopter rated pilot, and the commercial helicopter rated tactical flight officer (TFO) were not injured. Visual meteorological conditions prevailed for the local area surveillance flight, and a company visual flight rules (VFR) flight plan had been filed. The flight departed Meadow's Field Airport (BFL), Bakersfield, about 2226. The flight was scheduled to terminate at BFL.

The National Transportation Safety Board investigator-in-charge (IIC) interviewed both pilots. According to the pilots, the first flight of the night lasted 2 hours, with no discrepancies encountered. They indicated that after the first flight, they refueled the helicopter; however, during the refueling the fuel truck ran dry. The pilots indicated that 20 gallons of fuel had been pumped onboard the helicopter. While the deputies were on break, the operator refilled the fuel truck, and prior to the accident flight an additional 24 gallons of fuel was pumped onboard the helicopter.

The accident occurred about 1 hour 15 minutes into the second flight. The pilots received a call regarding a stolen car about 2 miles from their current position. The FP made a 180-degree turn to respond to the call. Halfway through the turn, both pilots heard a "big boom" and the helicopter shuddered. The FP leveled the helicopter and assessed the situation. He made a decision to land in a nearby parking lot. About 30 seconds later while they were en route to the parking lot, the pilots heard another loud boom followed by the engine out audio and warning lights.


Pilot-in-command/Flying Pilot (FP)

According to the FP's written statement to the Safety Board, he reported that they departed BFL about 2226 as Air One. About 2345, Air One was in the vicinity of Alfred Harrell Highway, west of Fairfax when they were requested to respond to a possible carjacking in the Oildale area. He noted their altitude at 1,100 feet mean sea level (msl), with an airspeed of 60 knots, prior to adding power. The pilot added power (about 80 percent torque) and initiated a right turn towards the west while accelerating to about 80 knots.

During the turn both he and the tactical flight officer (TFO) heard a loud boom and felt a vibration in the helicopter. The FP immediately reduced the power by lowering the collective, and reduced the airspeed to about 60 knots. They discussed the cause of the noise and were in the process of deciding to land or return to BFL when they heard a second loud boom followed immediately by a loss of engine power. The FP lowered the collective and entered into an autorotation.

The southwest parking lot of Bakersfield College was in front of them, and had been chosen as their primary emergency landing zone prior to the loss of engine power. While maneuvering to land, the FP recognized that he was going to overshoot the landing zone (LZ). He initiated a flare, but decided to continue over flying the LZ, while maintaining his airspeed. The TFO suggested Haley Street for landing. The FP made a turn northbound and lined up with Haley Street. About 60 to 75 feet above the ground the FP initiated the flare for landing. During the flare, they felt a "strong jolt" and an increased rate of descent. The FP reported noticing a loss of rotor rpm (revolutions per minute).

The FP maintained a level attitude, and about 10 feet above the ground he applied full collective pitch, and the helicopter landed hard with very little forward airspeed. After they secured and exited the helicopter, they realized that the helicopter had clipped power lines.

Tactical Flight Officer (TFO)

According to the TFO's written statement to the Safety Board, he reported that they received a call to respond to a possible carjacking about 2 miles to the west of their location. The TFO noted that during the turn to respond to the call, the pilot added power. While still in the turn he heard a loud boom and felt a vibration in the helicopter. At first he thought they had hit a bird and they were discussing the situation and their options when the heard a second loud boom, followed by a loss of engine power.

The TFO reported that during the first loud boom he noted the engine oil pressure gage and N1 gage were in the normal range. He further noted that the time between the two booms was between 30 seconds to a minute. When the engine quit, the TFO heard the engine out audio warning through his helmet and saw both the red and yellow caution warning lights illuminate on the instrument panel. At that point he turned his attention outside the helicopter to aid the FP in finding a suitable landing spot.


Pilot in Command/Flying Pilot

According to KCSD, the pilot was hired on August 22, 1977, and assigned to the Aviation Division on July 16, 1988. The pilots' most recent quarterly check ride, conducted by KCSD, was completed on March 10, 2004.

A review of Federal Aviation Administration (FAA) airman records revealed the FP held a commercial helicopter certificate.

The pilot held a second-class medical certificate issued on June 26, 2003. It contained the limitation that the pilot must wear corrective lenses.

According to NTSB Form 6120.1/2, Pilot/Operator Aircraft Accident Report, the pilot reported accumulating a total flight time of 5,773.4 hours, with a total flight time of 548 hours in the accident make and model helicopter. He logged 91.3 hours in the last 90 days, with 36.6 hours in the accident make and model. He logged 34.9 hours in the last 30 days, of which 28.3 hours were in the accident make and model. The FP logged 3.6 hours in the accident make and model in the last 24 hours.

Tactical Flight Officer

According to KCSD, the pilot was hired on November 28, 1984, and assigned to the Aviation Division on April 1, 1995. The pilot's most recent quarterly check ride, conducted by KCSD, was completed on March 9, 2004.

A review of Federal Aviation Administration (FAA) airman records revealed the TFO held a commercial helicopter certificate.

The pilot held a second-class medical certificate issued on August 8, 2003. It contained no limitations or waivers.

According to NTSB Form 6120.1/2 Pilot/Operator Aircraft Accident Report, the pilot reported accumulating a total flight time of 2,821.2 hours, with a total time of 765 hours in the accident make and model helicopter. He logged 99.2 hours in the last 90 days, and 46 hours in the accident make and model. He logged 30.5 hours in the last 30 days, of which 8.7 hours were in the accident make and model. The TFO logged 4.5 hours in the accident make and model in the past 24 hours.



The helicopter was a Bell OH-58A+, serial number 70-15442. A review of the helicopter's logbooks revealed a total airframe time of 6,667.1 hours at the last 100-hour inspection. The last 100-hour inspection was completed on April 13, 2004. An annual (Phase) inspection was completed on February 25, 2004, at an aircraft total time of 6,499.4 hours.


A Rolls-Royce/Allison, model 250-C20B, serial number CAE 835723 engine, was installed on the helicopter. Total time on the engine at the last 100-hour inspection on April 13, 2004 was 8,787.2 hours.

Acro Aerospace, Inc., Richmond, British Columbia, Canada, conducted a 1,750-hour inspection of the engine, which was completed on December 30, 2003. As a result of the inspection the second stage gas producer turbine nozzle diaphragm was replaced by machining out the old diaphragm and brazing in a new diaphragm. The engine then ran for an additional 161.7 hours before the accident.

According to the flight crew, at the time of departure (2230), the helicopter had a total of 88 gallons of Jet A fuel prior to the accident flight. Examination of the maintenance and flight department records revealed no unresolved maintenance discrepancies against the helicopter prior to departure.

The Rolls-Royce/Allison 250-C20B is a two-shaft turboshaft engine with a combination compressor, which consists of a six-stage axial compressor attached to a one-stage centrifugal compressor. The engine incorporates a reverse-flow annual combustor, a two-stage high-pressure turbine (also referred to as the N1, gas producer turbine), and a two-stage low-pressure turbine (also referred to as the N2 power turbine). The gas path along the Rolls-Royce/Allison 250 engine flows into the inlet, through the compressor's axial and centrifugal stages, into two external air transfer tubes and to the combustor, which is located at the very rear of the engine. The gases then turn 180 degrees toward the front of the engine and proceed through the two-stage gas producer turbine (N1) and a two-stage power turbine (N2). Finally, the gases are directed out of the exhaust collector and upward through two exhaust outlets.

The gas producer (GP) turbine, consisting of turbine wheels and nozzles #1 and #2, drives the compressor section of the engine through an inner shaft, while the power turbine (PT), consisting of turbine wheels and nozzles #3 and #4, drives the power output gear (to the main rotor transmission) and the accessory gearbox through an outer shaft. The inner shaft rotates independently within the outer shaft.



The Safety Board investigator and representatives from Rolls-Royce/Allison engines and Kern County Sheriff's Department, who were parties to the investigation, examined the helicopter and engine at the Kern County Sheriff's Department hangar on May 13, 2004. There were no discrepancies noted with the airframe.

Investigators noted that the visual examination of the engine revealed no external damage or damage to the compressor inlet or first stage compressor blades. Kern County maintenance personnel removed the engine for further inspection. Maintenance personnel removed the engine's magnetic chip detector plugs. Investigators noted the plugs were free of debris. Investigators were unable to manually rotate the compressor through the inlet guide vanes. However, the power turbine output shaft (N2) rotated freely, thus establishing drive continuity of N2.

The engine was shipped to the Rolls-Royce Engine Services (RRESO) facility in Oakland, California, for further examination. The investigative parties, along with inspectors from the Federal Aviation Administration (FAA) reassembled at RRESO on May 19, 2004, to examine the engine.

Investigators and RRESO personnel noted the following items: the number 1 turbine wheel had excessive heat damage, including partial loss of all airfoils; however, the snap ring remained intact and installed. Carbon deposits were found on the pinion to turbine coupling. When the power and gas turbine sections (PT and GP) were separated, the gas producer turbine section could not be manually rotated and was noted to be in a bound condition. The third-stage nozzle from the PT contained metal splatter. The number 8 bearing, located between the combustor and the GP, was intact, but not lubricated. The 19-24 seal displayed heat discoloration, but remained in the turbine support. The GP tiebolt connecting the GP turbine to the compressor coupling was removed intact; however, it displayed heat discoloration associated with high temperatures.

The number 2 turbine wheel was locked in place. Investigators noted that the second-stage turbine nozzle diaphragm fractured into several pieces and portions of the diaphragm were missing.

The following components were sent to the Safety Board's metallurgical facility in Washington, D.C., for a more detailed metallurgical examination:
The number 1 turbine wheel, p/n 23073853, s/n X527485
The number 2 turbine wheel, p/n 23073854, s/n X510254
The second-stage turbine nozzle, p/n 23031938, s/n RT015
Portions (fractured pieces) of the diaphragm, marked Acro 23-31248

The metallurgist made the following visual observations of the number 1 turbine wheel, number 2 turbine wheel, and the second-stage turbine nozzle.

The blades of the stage 1 and 2 turbine wheels showed significant damage from heat and deterioration. The stage 1 blades were solutioned beyond 0.1-inch of the rim and became molten at the fractured tips. The stage 2 blades were solutioned beyond 0.4-inch of the rim. Significant metal spray consistent with this deterioration was found in various locations.

The stage 1 blade track contained significant metal spray without rub marks. The stage 2 blade track also contained metal spray to a lesser extent. However, the stage 2 blade showed circumferential scoring/rub marks along the middle of the track. Axial scoring and rub marks with a separation consistent with blade spacing were noted. X-ray energy spectroscopic (EDS) analysis of the second-stage nozzle revealed that the stage 1 track had metal spray transfer consistent with Mar-M-246, and the stage 2 track contained metal spray transfer consistent with Hastelloy X. Additional metal spray layers were found on the stage 1 and stage 2 tracks. Both stage 1 and stage 2 tracks contained spray consistent with Metco 443NS. Review of Acro Aerospace repair records indicated that both tracks were sprayed with Metco 443NS.

Second Stage Turbine Nozzle and Diaphragm

The metallurgist noted that the diaphragm was etched with "Acro 23-31248." He reported that several of the nozzle airfoils exhibited tears along the trailing edge. Two sections of sheet metal diaphragm material were recovered from the second stage turbine nozzle. One section was a ring (0.34 - 0.45 inch wide) of the outboard portion of the diaphragm at the braze attachment area. Even though the ring of material was fully detached from the two brazing locations, it was still retained in the nozzle casting by mechanical deformation. The second piece of the diaphragm was the interior seal area. Examination of the fracture surfaces on the recovered portions of the diaphragm indicated that there were no pre-existing cracks, and the damage noted was consistent with tensile overstress (rub damage and rough features).

The metallurgist cut the ring of the second-stage nozzle diaphragm material in order to remove it from the nozzle casting. He noted that the diaphragm portion of the nozzle is secured to the cast rim portion by brazing between the outer ring of the diaphragm and the nozzle in two locations, one at the leading edge side of the nozzle and one at the trailing edge side of the nozzle. Once the ring was removed, the two braze surfaces were examined. On the leading edge side, the braze alloy filled between 10 - 50 percent of the joint width, with an estimated average fill of about 30 percent. On the trailing edge side, braze alloy filled about 50-100 percent of the joint width, with an estimated average fill of about 75 percent. The metallurgist also noted that in addition to the incomplete penetration of the brazing material, significant rubbing and impact damage was noted on the casting braze fracture surfaces.

According to Rolls-Royces' Authorized Maintenance Center Overhaul Information Letter 4 (AMC-OIL 4), for a class 1 joint, such as the brazed joint between the diaphragm and rim portion of the second-stage nozzle, the braze alloy should fill the exposed extremities in a continuous line for the first 20 percent of the joint width, or 1.5 times the thickness of the thinnest member being joined, and have a minimum total coverage of 80 percent of the joint. The metallurgist noted that the new-make requirements from Rolls-Royce were similar and covered by Rolls-Royce specification EIS (Engineering Inspection Specification) 1201.

According to the metallurgist, the leading edge braze fill of 10 - 50 percent of the joint width, and the average fill of 30 percent did not meet Rolls-Royce AMC-OIL 4 specifications requirements. The trailing edge braze fill of 50 - 100 percent of the joint width met the specifications of the AMC-OIL 4 specification; however, the average fill of 75 percent did not meet the specification requirements.

Two metallographic cross sections were made through a portion of the second-stage nozzle casting at the diaphragm braze joints. At the cross section locations, the braze on the leading edge joint only covered 10 - 20 percent of the joint width, and the braze on the trailing edge joint covered 60-100 percent of the joint width. At the trailing edge joint, the maximum thickness of the braze alloy was measured as 0.009-inch. According to Rolls-Royce Parts Repair Procedures Letter (PRPL) PRPL-2-D004 titled "Second Stage Turbine Nozzle Assembly - Repair Salvage" the braze gap prior to brazing should be 0.002 - 0.004 inch. PRPL-2-D004 replaced AMC-OIL 12.

An EDS analysis indicated that the base material of the second stage nozzle casting and the second stage nozzle diaphragm were consistent with the respective engineering drawing requirements.


A review of KCSD's records indicated that the engine was removed and sent to Acro Aerospace for the 1,750-hour inspection on October 17, 2003. During the inspection, the second-stage nozzle diaphragm was repaired per AMC-OIL 12 and AMC-OIL 4.

During the course of the investigation, Acro Aerospace (AcroHelipro) was added as a party to the investigation.

Rolls-Royce indicated that the company had been aware of the nozzle separations from past events. These previous events had been attributed to either improper brazing (lack of penetration), or fracturing of the outer portion of the diaphragm at the trailing edge bend due to thermal fatigue cracking. According to the Safety Board metallurgist, there was no evidence of thermal fatigue cracking in the second-stage nozzle diaphragm.

Rolls-Royce AMC-OIL 12 (PRPL-2-D004) specifies that the thickness of the braze gap filler alloy should be between 0.002-0.004 inch. However, the Acro Aerospace router (paperwork that accompanies the part as it goes through the repair process) also included information from AMC-OIL 4 for an alternate braze filler alloy if the braze gap was between 0.004 - 0.008 inch. According to Rolls-Royce, the AMC-OIL 12 procedure is specifically for repair of the second-stage nozzle diaphragm and takes precedence over AMC-OIL 4. AMC-OIL 4 is for general field braze repairs. It should be noted that Acro Aerospace was provided the PRPL from Rolls-Royce and was required to draft the router as a method of compliance with the PRPL's, prior to conducting the repair work. The above router was presented for approval to Rolls-Royce, who accepted Acro Aerospace's method for repair of the second-stage nozzle diaphragm, including the improper information regarding the braze alloy to be used for a gap of 0.004 - 0.008 inch. Inclusion of the reference to a gap of 0.004-0.008 inch reportedly created confusion at the overhaul shop (Acro Aerospace) on what braze gap was allowed. After the accident, Acro Aerospace removed reference to AMC-OIL 4 from the router for repair of the second stage nozzle diaphragm.

The Safety board metallurgist indicated that the thickness of the braze alloy diaphragm repair was measured in one area as at least 0.009-inch. According to the Safety Board metallurgist, this exceeded the 0.002 - 0.004 inch requirement specified in AMC-OIL 12. This thickness is also above the improperly referenced gap of 0.004 - 0.008 inch specified in AMC-OIL 4.

The Acro Aerospace router for the repair of the second-stage nozzle also included requirements for visual and fluorescent penetrant inspection of the brazed joints after completion of the repair. These methods are not capable of detecting internal voids unless the voids are connected to the surface. When the part is initially manufactured, an immersion ultrasonic inspection is required in addition to the visual and fluorescent penetrant inspections. The immersion ultrasonic inspection is capable of detecting internal voids in the braze joints even if they are not connected to the surface.

Rolls-Royce and Acro Aerospace reviewed the history and processing of the second stage nozzles repaired by Acro Aerospace. It was noted that Acro Aerospace used the hydrogen fluoride cleaning method to clean the large majority of the nozzles prior to brazing. However, the accident nozzle and seven others were cleaned using vapor degreasing. The other nozzles that were cleaned with vapor degreasing were retrieved and examined destructively, and most were found to contain cracking in the brazed joints. Destructive evaluation of representative nozzles that had been cleaned using the hydrogen fluoride cleaning method showed that none contained cracks, including some that had braze thicknesses as large as 0.020-inch.


The IIC released the helicopter to the operator on May 14, 2004, and the engine on May 20, 2004. Components retained for additional testing (the Number 1 Turbine Wheel p/n 23073853, s/n X527485, the Number 2 Turbine Wheel p/n 23073854, s/n X510254, and the Second-Stage Turbine and Nozzle Diaphragm p/n 23031938, s/n RT015) were released to the operator on October 12, 2004.

As a result of the accident and through numerous meetings, Rolls-Royce issued a new PRPL (PRPL 2-D004) to include hydrogen fluoride cleaning of the second stage turbine nozzle diaphragm, as well as, immersion ultrasonic inspection and ultrasonic inspection requirements during and after the brazing process. Acro Aerospace rewrote their router procedures to reflect the changes to the manufacturers revised procedures.

Use your browsers 'back' function to return to synopsis
Return to Query Page