History of the flight
The Robinson Helicopter Company model 22 helicopter (R22)
registered VH-UXF was engaged in aerial mustering operations with
another R22 helicopter registered VH-AOP. The pilot of UXF had
returned from a refuelling stop and had recommenced operations in
the mustering area. When no radio transmissions were heard from UXF
for approximately 15 minutes, the pilot of AOP commenced a search
for UXF. After a short search, the wreckage of UXF was located at
the edge of a claypan. The helicopter had impacted the ground
heavily in an upright attitude and both the occupants survived the
impact with what appeared to be severe injuries. After making them
comfortable the pilot of AOP departed to seek medical assistance.
He returned approximately 80 minutes later with a doctor. The
doctor reported that in the intervening period of time, both
occupants of UXF had died as a result of their injuries.
Wreckage findings
On-site examination of the helicopter revealed that it had
impacted heavily with little forward speed. Both main rotor blades
exhibited evidence of low speed rotation at the time of impact and
also exhibited minor damage from small tree strikes. The tail rotor
exhibited evidence of high speed rotation at the time of impact and
was destroyed by ground contact. The helicopter remained upright
and onsite examination revealed a failure in the clutch shaft which
on initial inspection appeared to be from impact damage. The
helicopter was recovered to an engineering facility for a more
detailed examination.
The examination at the engineering facility revealed what appeared
to be a pre-impact failure of the clutch shaft. The shaft was taken
to the ATSB laboratory in Canberra for further examination. The
remainder of the examination of the recovered helicopter revealed
that the engine was operating at high power when the helicopter
impacted the ground. In addition, no pre-impact damage to other
helicopter systems was found that would have contributed to the
accident.
Failed Part examination
The clutch shaft had been installed in the helicopter on 30
October 2002 and had 886.2 hours time-in-service since new.
The clutch shaft fitted to UXF (P/N A166-1, S/N 5570) had
fractured at the point of connection to the main rotor gearbox
flex-plate yoke (P/N A907). The fracture displayed clear evidence
of torsional fatigue cracking, following a spiral path from within
the yoke connection that extended around the shaft for
approximately 340 degrees over an axial length of around 25
millimetres. The area of final overload failure represented
approximately thirty percent of the overall shaft cross-section and
showed mostly ductile shear features.
Disassembly of the shaft-yoke connection found low levels of
break-out torque required for both assembly bolts. The inner and
outer bolts loosened at torque values of 7 and 4 foot-pounds
respectively; markedly below the 20 foot pounds (240 in-lb's)
tightening torque that the aircraft maintenance manual specified
for the connection. Removal of the bolts and bearing blocks
subsequently revealed that the blocks had been installed directly
over the painted yoke surface. Inspections found no evidence of any
attempt to remove the paint from the block seating locations prior
to assembly.
Both connection bolt shanks showed fretting and other prominent
evidence of loaded contact against the yoke and shaft hole sides.
The orientation of the bolt damage indicated that the looseness of
the connection had allowed the transmission of shaft torque loads
via shear loading of the bolts. Extensive fretting and corrosion of
the mating shaft and yoke surfaces, found after removing the
fractured stub provided further evidence of looseness and movement
within the connection. Both mating surfaces showed no evidence of
the primer coating specified by the maintenance manual. An oily
residue remained in some areas, as did an accumulation of a soft
yellow coloured material, typical of a jointing or similar
non-drying compound.
Initiation of the torsional fatigue cracking had occurred from the
bore of the inner shaft bolt hole, at an area of fretting damage
similar to the damage on the bolt shank surfaces. The fretting
damage was confined to one side of the hole, and similar damage was
also found within the other shaft bolt holes and the aligning yoke
holes.
Metallurgically, the clutch shaft material and heat-treatment were
sound and complied with the manufacturer's specification for the
component. No evidence was found suggesting any contributory
deficiencies within the design or manufacture of the clutch
shaft.
Examination of Serviceable Assembly
A comparable clutch shaft & yoke unit was obtained from the
same aircraft operator and disassembled at the ATSB laboratory.
While the yoke surfaces in that case were unpainted and hence the
bolt tensions remained higher (evidenced by break-out torques of 12
and 16 foot-pounds), the connection had still been unstable and
showed appreciable fretting of the mating surfaces and on the bolt
shanks. The similar, non-hardening jointing compound had been used
within the connection in lieu of the prescribed primer coating.
Cracking was not found within any of the shaft bolt holes.
Previous Failure
In 1992, a R22 helicopter sustained a torsional fatigue failure of
the A166 clutch shaft, with the aspects of the failure appearing
very similar to the failed component described above. A review of
the technical report provided by the then Civil Aviation Authority
concerning that failure confirmed the prior instability and
looseness of the shaft-yoke connection as being a factor
contributing to the failure. Similar fretting was found over the
mating surfaces of the shaft and within the bolt holes. The
torsional fatigue cracking had initiated from the same location and
propagated to a comparable extent before failure.
A subsequent inspection of the shaft that had been retained for
exemplar purposes also found no evidence of the use of a primer or
similar coating over the shaft connection.
Use of non-approved primer
The R22 maintenance manual states that a zinc chromate primer or
epoxy resin be sprayed onto the mating portions of the A907 yoke
and the end of the A166-1 shaft. These two components are then
"wet" assembled (joined together while the primer or epoxy is still
wet) and the bolts torqued to the pre-determined value prescribed
in the manual. The company maintaining the helicopter had
substituted a compound called "Mastinox" at some point early in
their operation and maintenance of the type. The chief engineer
stated that the use of Mastinox was instigated well before he had
joined the company and that it had become an accepted maintenance
practice. The use of "Mastinox" was not in accordance with the R22
maintenance manual, and communication with the manufacturer has
confirmed that the use of Mastinox is "not approved".
R44 Clutch shaft-to-yoke assembly
The R44 helicopter has an identical assembly of the clutch shaft
to yoke, only the part sizes are larger than those of the R22.
There have been no documented cases of failure of the R44 clutch
shaft in Australia.
ANALYSIS
The A166-1 clutch shaft fitted to UXF had failed as a result of
torsional fatigue cracking, originating from the bore of the inner
bolt hole used for securing the adjoining drive yoke. The
initiation of fatigue cracking was directly attributed to the
looseness of the shaft-yoke connection, which allowed the
transmission of rotational loads through the connection by shear
forces acting on the bolt shanks and transmitted via the bolt
holes. The point loading about the holes produced by that behaviour
produced a significant stress-raising effect on the localised
material structure. In the presence of the associated fretting and
corrosion damage that was sufficient to initiate fatigue cracking
under normal shaft loads.
Under normal intended security of the connection, rotational
forces are transmitted uniformly via the friction between the shaft
and yoke surfaces. Security of the connection is established by
adequate bolt tension and the assembly of the connection with a
curing or drying primer. Corrosion protection is also assured with
the use of the manufacturer approved primer(s).
The assembly of the yoke-shaft connection without first cleaning
away the paint from underneath the block seating locations was
considered a major factor in the loss of sufficient bolt tension
and hence the looseness of the connection. The
compressibility/conformability of most paint coatings allows
applied bolt tension to be progressively lost and hence renders
such coatings unsuitable for use within stable, load bearing bolted
connections.
The assembly of the yoke-shaft connection using a jointing or
similar compound in lieu of the specified primer(s) is also
considered to be a factor contributing to the looseness of the
connection. The oily, anti-friction properties shown by the
compound would have acted to reduce the surface friction within the
connection thus increasing the loads transmitted through the bolts
and bolt holes. The non-drying properties of the compound also
prevented the effective "lock-up" of the connection when assembled,
allowing the surface movement, fretting and bolt
interference.
The potential exists for the similar factors described above that
led to the failure of the examined R22 shaft to exist within the
R44 fleet.
The yoke assembly covers the initiation and propagation site of
the failure and therefore it cannot be easily observed in its early
stages. Inspection of the yoke-to-shaft mating area during a daily
inspection, by either a pilot or a maintenance engineer, is
unlikely to discover evidence of the crack. By the time the crack
has propagated to a length where it is visible outside the yoke,
complete failure is imminent.
