SEA97FA001
SEA97FA001

HISTORY OF FLIGHT

On October 4, 1996, approximately 1215 hours Pacific daylight time, a Boeing Vertol BV-107 II, 196CH, registered to and operated by Columbia Helicopters, Inc., was destroyed when it collided with terrain following a loss of control in flight during cruise. The crash site was three miles east of the southern boundary of the Aurora airport (refer to CHART I). A post crash fire confined to the engine area was extinguished following the crash. Both pilots and the onboard mechanic were fatally injured. Visual meteorological conditions existed and no flight plan had been filed. The flight, which was a conformity maintenance check flight, was to have been operated under 14CFR91, and originated from the Aurora airport, Aurora, Oregon, at 1138.

Witnesses, many of whom were initially attracted by the unusual sounds from the rotorcraft, reported observing it maneuvering erratically in the vicinity of the accident site, and then tumbling out of control to ground impact. Specifically, one witness reported observing the rotorcraft's rotor blades impact one another. Another witness described the sound as like "metal hitting" and described the maneuvers as "flipping." Another witness reported seeing the rotorcraft flying "straight and level for three or four seconds before it went vertical." Several other witnesses observed the rotorcraft flying away from the Aurora airport approximately 30 minutes before the accident and then return during which they observed it "tumble" (refer to witness statements and attached FAA witness statement transcriptions).

A flight instructor on an instructional flight, who was taxiing out to runway 17 at the Aurora airport, reported that he "heard a helicopter make a position report" (this radio transmission occurred approximately 12:10). He could not recall what was said during the radio transmission but reported that "about 10-15 seconds later (he) heard a stuck mike on the radio with the same helicopter noise. This lasted about 10-15 seconds. Then the mike was un-keyed" (refer to attached statement).

PERSONNEL INFORMATION

PILOT-IN-COMMAND:

The pilot-in-command, who occupied the right seat in the cockpit, held an airline transport pilot certificate as well as a flight instructor's certificate. According to the operator, he had accrued a total of 14,778 hours of flight experience of which 11,841 were as pilot-in-command (PIC), and 14,668 hours were logged in rotorcraft. Additionally, he was reported to have logged 8,880 hours in the Boeing Vertol BV-107 rotorcraft of which 8,269 hours were as PIC. The PIC held a type rating in both the BV-107 and the BV-234 rotorcraft.

CO-PILOT:

The co-pilot, who occupied the left seat in the cockpit, held a commercial pilot certificate. According to the operator, he had accrued a total of 4,112 hours of flight experience of which more than 2,500 hours were as pilot-in-command (PIC) and 4,036 hours was logged in rotorcraft. Additionally, he was reported to have logged 2,449 hours in the Boeing Vertol BV-107 rotorcraft of which 1,809 hours were PIC. The co-pilot held a type rating in the BV-107 rotorcraft.

CREWMAN:

The crewman, whose location in the rotorcraft could not be determined, held an FAA airframe and powerplant mechanic certificate. According to the operator, he had been engaged in maintenance on the rotorcraft during its preparation for flight testing and, as was customary for the operator, was assigned to assist during the accident test flight.

AIRCRAFT INFORMATION

N196CH, serial number 407, was a Boeing manufactured derivative of the model 107 rotorcraft built for Sweden as a model HKP-4, and which had been acquired by Columbia Helicopters, Inc., to be converted to a civil model BV-107-II in accordance with FAA Project Number TDO639NY-R. The rotorcraft had a total of 7,073.0 hours of airframe time at the time its "experimental" certification was approved on August 5, 1996. And, on October 2, 1996, the rotorcraft was issued a maintenance release for its first conformity test flight.

On October 3, 1996, N196CH, was flown for 1.4 hours, including four landing, from the operator's base at the Aurora airport.

On October 4, 1996, N196CH, departed the Aurora airport at 1138 hours on its second test flight (refer to photograph 1 which shows the accident aircraft departing on the accident flight). The aircraft had departed with 1,800 pounds of Jet A fuel.

WRECKAGE AND IMPACT INFORMATION

The aircraft crashed in an open, plowed, agricultural field. The latitude and longitude of the crash site (point A on DIAGRAM I) was 45 degrees 13.52 minutes North and 122 degrees 42.73 minutes West, respectively. The elevation of the site was approximately 175 feet above mean sea level (MSL) (refer to CHART II). The majority of the airframe came to rest in four major sections (refer to SCHEMATIC I, DIAGRAM I and photograph 2). The forward fuselage (including the cockpit) and forward pylon/rotor head assembly (section A) was observed to be furthest west. This section came to rest with its longitudinal axis oriented along a 223/043 degree magnetic bearing line (nose towards the southwest)(refer to photograph 3). The center cabin area (section B) was located slightly east and adjacent to the aft lower fuselage section containing both engines (section C)(refer to photograph 4). The aft pylon and rotor head assembly (section D), which came to rest furthest to the east lay approximately 75 feet from the forward cabin area (refer to photograph 5). These four major sections of the rotorcraft lay along an approximate 270/090 degree magnetic bearing line.

The forward rotor head assembly remained attached to the forward airframe (section A) at its pylon. All three fiberglass rotor blades (red, yellow and green) remained attached to the rotor hub (refer to photographs 6 and 7). However, the blades displayed shattering damage towards their outboard sections and tips. The aft rotor head assembly remained attached to the aft pylon (section D). The aft pylon separated from the fuselage. Again, all three fiberglass rotor blades (red, yellow and green) remained attached to the rotor hub (refer to photograph 5). Again, the blades displayed shattering damage towards their outboard sections and tips.

The synchronization drive shaft, which consists of five successive tubes connecting the forward and aft transmission units, was examined at the site. Shaft numbers four and five (aluminum and steel respectively) were found connected together with the aft end of shaft five attached to the aft transmission unit (refer to photograph 8). Shaft numbers one and two (aluminum) were found connected together with the forward end of shaft one attached to the forward transmission unit. The number two synchronization shaft was observed to be broken at its midpoint and the aft end of this shaft as well as the entire shaft number three were not found within the main wreckage (refer to photograph 9). The entire number three synchronization shaft (aluminum) was located lying in the field approximately 90 feet and 159 degrees magnetic from the forward cabin (section A)(refer to DIAGRAM I and photograph 4). The coupling at each end was absent and the rivets, some of which remained in the shaft, exhibited flush smearing consistent with rotational or longitudinal overload. Additionally, a diagonal impact near the forward end of shaft number three was observed. The impact was consistent with a rotor blade outboard leading edge impact (refer to photograph 10). The aft 40% of the number two shaft was located lying in an adjacent field (refer to photograph 11) bearing approximately 133 feet and 87 degrees magnetic from the number three shaft (refer to DIAGRAM I). A number of smaller aluminum fragments of drive shaft were recovered from the site and these, along with the aft number two shaft section and number three shaft were reassembled at the reconstruction site. The separation at the approximate midpoint of the number two drive shaft was consistent with a rotor blade strike and there was no evidence of any disconnect of the drive shaft prior to the blade strike.

Numerous small fragments of rotor blades, fuselage skin, and fiberglass were observed to be distributed over an area extending 1,400 feet. The general distribution (magnetic track) of the fragments was found to lie along an approximate 004 degree bearing line with many of the smaller fragments having fallen into a filbert orchard north and east of the crash site (refer to CHART II). The size and weight of fragments gradually increased approaching the crash site, with the lightest fragments most distant.

Both the forward and aft pylon and rotor head controls, as well as the control cables and rods within the tunnel connecting the rotor heads, were examined at the site. No evidence of any pre-impact disconnect was found. The engines were observed to have remained within the aft fuselage (section C) which had sustained a post crash fire.

The wreckage was recovered and transported to an indoor facility several miles away for partial reconstruction. During the recovery process it was noted that the right side of the forward cabin/cockpit area, including the main cabin entry and the flight control closet area, which houses much of the rotorcraft's control linkages, was substantially crushed inward (refer to photograph 12).

MEDICAL AND PATHOLOGICAL INFORMATION

Post mortem examination of the pilot-in-command, co-pilot, and crewman, was conducted by Clifford C. Nelson, M.D., at the Offices of the Oregon State Medical Examiner, 301 NE Knot Street, Portland, Oregon, on October 5, 1996. Toxicological evaluation of samples from all three crewmen was conducted by the FAA's Toxicology and Accident Research Laboratory, Mike Monroney Aeronautical Center, P.O. Box 25082, Oklahoma City, Oklahoma. The resultant tests were found to be negative in all three crewmen (refer to attached Toxicology reports).

OFF SITE EXAMINATION AND RECONSTRUCTION

During the off-site wreckage reconstruction phase, the rotorcraft's two General Electric CT58-140-1 turboshaft engines were examined. Examination of the power turbine rotor blades of both left and right engines revealed uniform tip curl opposite to the direction of rotation. Additionally, there was no evidence of any uncontained ejection of engine components from either engine casing.

The forward and aft rotor blades, which had been removed from their respective rotor heads at the site, were reassembled with their associated fragments at the reconstruction site. There was no evidence of any pre-accident inflight loss of components/sections of any of the six rotor blades.

The flight control continuity check of the forward cabin/cockpit area, including the flight control closet area, was continued at the reconstruction site. It was necessary to cut away the external airframe skin in order to access this area. Once accomplished, many of the flight control rods were observed to display evidence of bending deformation, separations characteristic of impact overload, and scratching and paint abrasions (refer to photograph 13).

During the examination and disassembly of the flight control closet, a disconnect was noted at the point where the lower bearing end of the "aft directional and lateral" output pushrod connects to the inboard clevis of the bellcrank within the forward section of the mixing unit (refer DIAGRAM II and photographs 14 and 15). The bolt specified for this installation, a AN (Air Force - Navy aeronautical standard) 464, was absent, as was the nut, washers and cotter key.

The mixing unit was removed from the flight control closet as was the disconnected "aft directional and lateral" output pushrod, and both components were examined more closely (refer to photographs 16 and 17). The pushrod displayed some minor bending deformation and longitudinal scratches of its painted surface. Additionally, there was no evidence of any significant mechanical damage in the pushrod's inner bushing end characteristic of impact deformation against a threaded bolt. However, the opposing "forward directional and lateral" output pushrod, as well as the "aft collective pitch and longitudinal" output pushrod, both of whose lower bearing ends remained attached to the mixing unit, were broken (refer to photographs 16 and 17 and DIAGRAM II). The "forward collective pitch and longitudinal" output pushrod, which remained attached at both ends, to both the mixing unit and the collective pitch and longitudinal input bellcrank, was unbroken but exhibited extensive scratching and bending deformation (refer to photographs 16 and 17 and DIAGRAM VI).

The forward mixing unit section, including the disconnected bellcrank was separated from the entire mixing unit assembly, cleaned with solvent, and examined, as was the pushrod, (refer to photograph 18). It was determined that this bellcrank, P/N 107C2606-8 (refer to photograph 19), was in fact, a collective bellcrank which had been installed in the forward (lateral portion) of the mixing unit, and not the lateral bellcrank, P/N 107C2606-9, called for in Boeing Drawing 107C2606 (refer to DIAGRAMS III, IV, V). A correctly installed collective bellcrank, P/N 107C2606-8, was found to be installed in the aft (collective portion) of the mixing unit, as called for in Boeing Drawing. Note: P/N 107C2606-8 (aluminum) is equivalent dimensionally to P/N 107C2606-1 (magnesium) as detailed in both the Illustrated Parts Catalogue and Boeing Drawing 107C2606. The same applies to P/N 107C2606-9 (aluminum) and P/N 107C2606-2 (magnesium).

The forward cabin/cockpit area, including the flight control closet, was subsequently re-examined for any loose hardware (bolts or nuts) which might have been the disconnected AN464 bolt. A bolt matching the type used to attach the remaining three output pushrods was discovered loose in the flight control closet area. No matching nut was found.

Discussions with the operator and Boeing Vertol indicated that a disconnect of the aft directional and lateral pushrod at the mixing unit would render the aft rotor head incapable of receiving cockpit issued lateral and directional control inputs. The forward rotor head would continue to receive such control inputs thereby creating a control force differential between the two rotor heads (refer to DIAGRAM VI).

TESTS AND RESEARCH

The bellcrank (P/N 107C2606-8) removed from the forward mixing unit, along with both the connected forward and disconnected aft lateral pushrods, and the loose bolt, were hand carried to the Safety Board's Materials Laboratory Division for further examination (refer to photograph 19). Examination of the components was conducted on February 27, 1997 (refer to attached Metallurgist's factual report).

The lower ends of both the (disconnected) aft directional and lateral pushrod, P/N 107C2551-13, and the (connected) forward directional and lateral pushrod, P/N 107C2551-11, control rods during normal assembly are fastened to the forward mixing unit bellcrank, P/N 107C2606-9 by bolts through the clevis tangs and the rod end bearings on each control rod. These bolts are shown in the illustrated parts catalog and assembly drawing as being inserted from the forward (cockpit) side of the bellcrank and secured with a castellated nut and cotter pin on the aft (tail) side. The required fastening hardware includes a NAS 464-4-17 bolt, three AN1 960 washers, (two thick -416, one thin -416L), an AN 320-4 or MS2 17826-4 castellated nut and an AN 381 cotter pin. The bolt passes through a NAS 75-4-010 sliding bushing installed in the forward tang of the clevis and an NAS 77-4-23 flanged bushing inserted into the aft tang. The flanged bushing is shown installed with the flange on the inside of the clevis. The required buildup of exemplar fastening components is shown in figure 2 (metallurgist's factual report). The bushings were found in place in the bellcrank at the accident reconstruction site (refer to photographs 20 and 21) but removed prior to the metallurgy examination.

The bolt suspected of having come from the left clevis connection had head markings identifying it as a NAS 464 close tolerance shank bolt. It had a shank diameter of approximately 0.25 inches, a grip length of 1.06 inches and an overall shank length of 1.41 inches. The bolt was plated with what appeared to be conversion coated cadmium except on the shank and washer flat surface. The dimensions and surface finish were consistent with a NAS 464-4-17 bolt. The bolt had a nearly identical appearance and dimensions to the NAS 464 bolt removed from the right clevis of the bellcrank. The bolt suspected of coming from the left clevis connection along with the right clevis bolt and the exemplar buildup are displayed in figure 3 (metallurgist's factual report).

The right clevis bolt was received with two thin "-416L" washers installed (see arrow, figure 3 metallurgist's factual report). The measured overall width of the clevis (P/N 107C2606-8) between the outer surfaces of the tangs (including the paint layer) was 1.075 inches. The specified overall width of this examined clevis (P/N 107C2606-8) between the outer surfaces of the tangs (excluding the paint layer) was 1.062 inches whereas the specified overall width of the clevis called for in the illustrated parts catalogue (P/N 107C2606-9) between the outer surfaces of the tangs (excluding the paint layer) was 0.960 inches (refer to DIAGRAMS II, IV, and V).

Based on calculations using measurements from the exemplar parts, the bolt suspected of coming from the left clevis would not be long enough to allow the cotter pin to be properly inserted through the nut and bolt using the required arrangement of washers (as shown in figure 3, metallurgist's factual report). However, when only two thin washers are used, like that found for the right clevis assembly, the cotter pin can be inserted. To verify the calculations, trial assemblies were performed using the required arrangement of washers and one using only two thin washers. Figure 4 (metallurgist's factual report) shows the two assembly configurations assembled on the right clevis. As can be seen, the cotter pin hole is only exposed when two thin washers were used. With the thick washer assembly the cotter pin hole was not exposed (hidden by the unslotted portion of the nut) and a cotter pin could not be inserted to safety the nut.

The shank of the bolt suspected of coming from the left clevis connection showed a fine circumferential grinding pattern typical of original manufacture. In addition, three narrow circumferential contact marks were noted on one side of the shank at positions approximately 0.1, 0.69 and 0.88 inches from the underside of the head. The opposite side of the shank was marked by a small band of shallow longitudinal scratches. Both the contact patterns and longitudinal scratches were characteristic of a bolt which had been used in an assembly.

Examination of the bolt threads with a stereo microscope found light circumferential scoring on both the pressure and non pressure flanks of the threads in the area of bracket "A", figure 5 (metallurgist's factual report). A few areas of intermittent light scratching were noted on the threads between the cotter pin hole and the shank in the area of bracket "B" in figure 5 (metallurgist's factual report), but none extended completely around the bolt. A few thread crowns adjacent to the shank were mechanically damaged and deformed on the pressure flanks. The cotter pin hole had an as manufactured appearance with no scratches, gouges or scoring characteristic of contact with a cotter pin.

In comparison, the bolt removed from the right bellcrank clevis showed a continuous scoring pattern of both the thread flanks between the cotter pin hole and the bolt end, in the area of bracket "C", figure 5 (metallurgist's factual report). The threads between the cotter pin hole and the shank, bracket "D" in figure 5 (metallurgist's factual report), were heavily marked on the pressure flanks consistent with contact by nut threads. The cotter pin hole for this bolt showed two prominent score marks for the full length of the bore surface. The unmarked arrow in figure 5 (metallurgist's factual report) denotes the location of one of the score marks. The scores were at diametrically opposed locations in the bore, aligned perpendicular with the longitudinal axis of the bolt and consistent with insertion or removal of a cotter pin.

The surfaces of the tangs for both the right and left clevises were optically compared in the area of the bushings and holes. Figures 6 and 7 (metallurgist's factual report) show the forward and aft faces of the bellcrank, respectively, with the bolts removed. The paint around the bushing on the forward face of the right clevis was cracked and disturbed in a circular pattern consistent with contact by a circular object, see arrow "A" in figure 6 (metallurgist's factual report). The circle of disturbed paint was about 0.5 inches in diameter. The AN 960 washers used in the assembly have an approximately 0.5 inch outer diameter. On the forward face of the left clevis the paint was cracked and appeared to have been lifted from the surface around the hole (bushing had been removed) in a circular area, see arrow "B" in figure 6 (metallurgist's factual report). The damaged paint was not tightly attached to the bellcrank surface and could be easily removed.

Circular impression ridges of paint were visible encircling both clevis holes on the aft surfaces of the bellcrank. Light scratch patterns in the paint within each impression were consistent with contact with a circular object. The aft face of the bellcrank is shown in figure 7 (metallurgist's factual report) with arrows denoting the circular paint impressions.

The inside faces of the left clevis were mostly devoid of surface marking except for a small raised paint ridge on the face of the flanged bushing. In contrast, the inside faces of the right clevis showed a prominent circular contact area on the face of the flanged bushing and cracked paint on the surface of the sliding bushing. A comparison of the markings on the flanged side of the bushings from the left and right clevis is shown in figure 8 (metallurgist's factual report).

During examinations it was noted that the inner diameter of the sliding bushing from the left clevis was greater than the sliding bushing in the right clevis and the exemplar bushing, see figure 9 (metallurgist's factual report). NAS 75-4 bushings have a 0.25 inch nominal inner diameter (ID). The left sliding bushing ID measured 0.27 inches. All other bushings had a nominal 0.25 inch ID. The right clevis and exemplar bushings were also chamfered at the ID, as indicated by arrow in figure 9 (metallurgist's factual report), and the left bushing was not.

Optical examination of the left control arm lower rod end bearing uncovered a dent in the bearing shield. The dent, shown in figure 10 (metallurgist's factual report), was consistent with over travel contact with the bearing ball.

ADDITIONAL INFORMATION

On-site examination and investigation commenced on the evening of October 4 and continued through October 12,1996, after which the wreckage was released to the owner/operator. A number of components were retained for further metallurgical examination and returned June 26, 1997, as documented on the attached "receipt of aircraft parts" (NTSB Form 6120.15).

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