SPECIAL INVESTIGATION REPORT 72-2

ACCIDENT INVESTIGATION REPORT

BUREAU OF AIR SAFETY INVESTIGATIONLIBiiTtii/ COPY

Beech 65-80 Queenair AircraftVH-CMI near Alice Springs on20 January, 1972.

Air Safety Investigation Branch

SPECIAL INVESTIGATION REPORT 72-2

GIDENT INVESTIGATION REPORT

Department of Civil Aviation

Australia

Connair Pty. Ltd. Beech 65-80Queenair Aircraft VH-CMI nearAlice Springs on 20 January, 1972,

The investigation of this aircraft accidentwas authorised by the Director-General ofCivil Aviation pursuant to the powersconferred by Air Navigation Regulation 278.

Prepared by:Air Safety Investigation Branch Melbourne

September, 1972

©Commonwealth of Australia 1972: The contents ofthis publication may not be reproduced in whole or inpart without the written authority of the Departmentof Civil Aviation. Enquiries should be addressed tothe Air Safety Investigation Branch, Box 1839 Q,G.P.O., Melbourne, 3001.

CONTENTS

Section 1

1.11.21.31.41.51.6

1.71.81.91.101.111.12

1.131.141.151.16

Section 2

Section 3

INVESTIGATION Page 1

History of the Flight 1Injuries to Persons 2Damage to Aircraft 2Other Damage . 2Crew Information 2Aircraft Information 2

History 2Loading 2

Meteorological Information 3Aids to Navigation 3Communications 3Aerodrome and Ground Facilities 3Flight Recorders 3Wreckage 3

Wreckage Trail 3Powerplant Layout 4Powerplant Examination 4System Examination 5Structural Examination 5

Medical and Pathological Information 6Fire 6Survival Aspects 7Tests and Research 7

ANALYSIS 7

CONCLUSIONS 12

APPENDICES

List of Occupants andProbable Seating Arrangement Appendix A

Transcript of Tape RecordingAlice Springs Tower 20 January 1972 B

THE ACCIDENT

At approximately 0745 hours Central Stand-ard Time on 20 January 1972, there was anin-flight fire in a Beech 65-80 Queenair aircraft,registered VH-CMI, which resulted in the separ-ation of the starboard engine and the starboardouter wing. The aircraft subsequently struck theground some seven miles south-west of AliceSprings Airport in the Northern Territory. At thetime of the accident, the aircraft was engaged inoperating a charter flight for the purpose of carry-ing passengers, mail and freight from Alice Springsto Ayers Rock. The aircraft was destroyed by fireand impact forces and the pilot and the sixpassengers were killed.

1- Investigation

1.1 HISTORY OF THE FLIGHT

Beech 65-80 aircraft, VH-CMI, was assignedto operate a charter flight departing Alice Springsfor Ayers Rock at 0730 hours on 20 January 1972.The aircraft was owned by Connellan AirwaysPty. Ltd. and operated by Connair Pty. Ltd. whohold an appropriate charter licence issued by theDirector-General of Civil Aviation. The aircraftwas reg is te red for r egu la r pub l i c transportoperations.

Before boarding the aircraft the pilot attend-ed at the self briefing facility at Alice SpringsAirport and compiled a visual flight rules (VFR)flight plan covering two return flights to AyersRock. This flight plan, which was lodged with theflight service.officer on duty, indicated that theaircraft would proceed to Ayers Rock on thedirect track of 235 degrees magnetic at a cruisingaltitude of 6,000 feet and that the estimatedflight time was 64 minutes.

At 0733:06 hours, the pilot called AliceSprings Tower by radio and requested a clearanceto taxi. The aircraft was cleared to taxi for take-off on Runway 12, (i.e. the runway which isaligned 116 degrees magnetic). While taxying thepilot received and acknowledged an airways clear-ance which was in accord with his flight plan and,at 0735:23 hours, the aircraft was cleared to takeoff and make a right turn for the purpose ofjoining the departure track. The aircraft carried

out an apparently normal take-off which waswatched by the aerodrome controller on duty inthe control tower until such time as the aircrafthad commenced its turn to the right.

At 0740:06 hours the pilot reported thatthe aircraft's departure time was 0739. He wasgiven the area altimeter setting and instructed tochange from the tower frequency to the flightservice frequency. Note : Aeronautical Inform-ation Publication requires that a pilot shall reporthis departure time when established on the de-parture track at no further distance from theaerodrome than five miles. The departure timeshall be the current time less an adjustment forthe distance from the aerodrome.

At 0740:25 hours the pilot made a normalcontact with Flight Service and this call wasacknowledged. The next radio transmission fromthe aircraft was recorded at 0741:40 hours andconsisted of a one-second transmission on theflight service frequency. No message was passed,the only modulation being a background of enginenoise with a "beep-beep" sound superimposedupon it. At 0742:20 hours, 40 seconds after theprevious transmission, the pilot reported on towerfrequency : "we are returning to the field I have,er, a asymmetric condition, er, a fire in the star-board engine." Throughout this transmission therewas a continual "beep-beep" sound in the back-ground. The tower controller immediately clearedthe aircraft to return for landing on Runway 12and requested it to report on final approach. Thisinstruction was not acknowledged from the air-craft but a muffled clicking noise, similar to amicrophone switch being operated or a suddenbreak in an unmodulated radio transmission wasdetectable at this time on the record of radiocommunications. There was no further radio con-tact with the aircraft and at 0742:52 hours theaerodrome controller observed a puff of blacksmoke in the sky to the south-west of theaerodrome.

The only person who saw the aircraft afterits departure from the immediate vicinity of theaerodrome was Mr. B. L. Parsons who, at the time,was located some 5 miles south-west of the airportdriving an item of road construction equipment ina south-westerly direction. He reports that he firstobserved the aircraft flying diagonally across theroad ahead of him from left to right and that itappeared and sounded normal at that time. Thewitness, some time latej, saw a cloud of blacksmoke to his left and at the same time saw anaircraft, beneath the cloud, diving steeply towardsthe .ground. It appeared to make one complete

revolution about its longitudinal axis before dis-appearing behind trees and no smoke arose fromthe position where the aircraft disappeared. Thewitness was unable to recall, with any degree ofaccuracy, the time interval between the twosightings.

1.2 INJURIES TO PERSONS

Injuries Crew Passengers OthersFatalNon-fatalNone

1

1.3 DAMAGE TO THE AIRCRAFT

The starboard outer wing and the starboardengine with its associated cowls, mounting struc-ture and propeller became detached from theaircraft in flight and were substantially damagedwhen they fell to the ground. The remainder ofthe aircraft was destroyed by impact forces whenit subsequently dived to the ground.

1.4 OTHER DAMAGE

There was no damage to property exceptthat which was carried in the aircraft.

1.5 CREW INFORMATION

The pilot-in-command of the aircraft, NigelClifford Halsey, aged 25 years, was the sole crewmember and held a valid commercial pilot licenceendorsed for the aircraft type. His total flyingexperience amounted to 1,558 hours of which 481hours had been gained in multi-engined aircraft,i n c l u d i n g 55 hours in Beech 65-80 aircraft.Mr. Halsey had passed the necessary theoreticalexamination for a senior commercial pilot licencebut he had not accumulated the flying experiencenecessary to qualify for the issue of this class oflicence.

On the day prior to the accident the SeniorRegional Pilot of the company conducted a pro-gress and base check on Mr. Halsey in conjunctionwith a flight to Mt. Isa and return in VH-CM1. Thebase check, which was carried out at Mt. Isa,included a simulated failure of the starboardengine followed by an asymmetric go-around andlanding. The progress and base checks were bothcarried out satisfactorily.

Mr. Halsey's most recent medical examin-ation was carried out on 23 September 1971 whenhe met the medical requirements for a commercial

licence. There was no evidence from his previousmedical history, his activities on the evening pre-ceding the accident, or from the post-mortemexamination that pilot incapacitation contributedin any way to the accident.

1.6 AIRCRAFT INFORMATION

History Beech 65-80 aircraft, Serial No. LD-12,was constructed in the United States of Americaand imported to Australia in 1962 as a used air-craft. At that time its total time in serviceamounted to 310 hours. The aircraft passedthrough the hands of a number of owners beforebeing acquired by Connellan Airways Pty. Ltd.The certificate of registration was transferred tothat company on 20 May 1970 and was valid until19 May 1979.

There was a current certificate of airworthi-ness for the aircraft which was valid until 10December 1973 and the records indicate that,prior to this accident, the aircraft had flown atotal of 4,017 hours since new and 555 hours sinceits last major inspection. The port engine hadcompleted 2,138 hours in service, including 168hours since its last overhaul, whilst the starboardengine had completed 2,299 hours in service in-cluding 716 hours since its last overhaul. All ofthese operating periods were within the maximumhours between overhauls specified in the mainten-ance manual and approved by the Director-General.

A Check 3 inspection was carried out on theaircraft on 13 January 1972 and, with the ex-ception of a minor discrepancy in respect of radio,the worksheets were appropriately certified bylicensed engineers. Maintenance Release No. 1142was issued on the morning of 20 January 1972, onthe completion of the daily inspection and nodefects were recorded.

In the three-week period preceding theaccident a recurring defect had been reported by anumber of pilots. This defect consisted of an inter-mittent fluctuation of 10 to 20 Ib per hour in thestarboard flowmeter indications during cruisingflight. In an effort to rectify this defect, filters andlines were cleaned and all of the major fuel systemcomponents were replaced. Although this actioneffected some improvement it seems likely thatthis defect was present to a reduced extent at thetime of the accident.

Loading Prior to the aircraft's departure fromAlice Springs, a load statement was prepared and acopy of this statement was handed to the pilot.

The passenger list and seating plan are shownat Appendix A. Although the two seats used byMr. and Mrs. Kneipp are known it has not beenpossible to establish the particular seat which eachoccupied. In addition to the six passengers whowere carried, 27 Ib of baggage, 60 Ib of freightand 27 Ib of mail were stowed in the rear locker.

The gross weight of the aircraft at the timeof take-off has been calculated to have been 7,342Ib which is less than the maximum permissiblegross take-off weight of 8,000 Ib. The load dis-tribution was such that the centre-of-gravity ofthe aircraft was within designated limits.

1.7 METEOROLOGICAL INFORMATION

Prior to completing the flight plan for theflight the pilot was provided with the terminalforecast for Ayers Rock and details of the ap-propriate area forecast. These forecasts indicatedthat the weather would be fine with a visibility of15 to 20 miles, there would be 1/8 of cloud at theaircraft's operating altitude of 6,000 feet, and thewind at that altitude was forecast to be variable at8 knots. The surface wind at Alice Springs, givento the aircraft by the tower shortly before take-off,was light and variable tending north-easterly andthe surface temperature at the time was 27°C. Anupper wind observation, made at 0900 hours atAlice Springs Airport, recorded the 3,000 feetwind as 055 degrees 10 knots and the 5,000 feetwind as 320 degrees 8 knots.

1.8 AIDS TO NAVIGATION

The availability or serviceability of radionavigation aids was not a factor in this accident.

1.9 COMMUNICATIONS

During the flight all communications be-tween the aircraft and ground were carried out onVery High Frequency radio channels. These com-munications were recorded on magnetic tape anda transcript of the relevant portion of this tapeappears at Appendix B. There is no evidence ofany deficiency in the provision of communicationfacilities.

1.10 AERODROME ANDGROUND FACILITIES

The availability of these facilities was not afactor in this accident.

1.11 FLIGHT RECORDERSThe aircraft was not fitted with either cock-

pit audio or flight data recorders and there is norequirement for Australian registered aircraft ofthis category to be so equipped.

1.12 WRECKAGE

Wreckage Trail The accident site was locatedin relatively flat featureless country some sevenmiles south-west of Alice Springs Airport. The areais sparsely covered by stunted eucalypt scrub andthere is no nearby habitation.

The wreckage trail extended for a distanceof some 5,000 feet. Some 4,650 feet of the trailpreceded the main impact point whilst debrisresulting from the main impact was thrown for-ward a further 350 feet.

The following major components of theaircraft were located along the wreckage trail atthe indicated distances preceding the main impactpoint:

Starboard main landing gearStarboard engine and propellerStarboard engine augmenter

(inboard)Starboard outer wingStarboard engine augmenter

(outboard)

500 feet640 feet

1,100 feet1,200 feet

1,700 feet

The nature of these components and theirdistances from the main impact point clearly in-dicated that the aircraft had sustained a majorstructural failure in flight.

The earliest part of the wreckage trail con-sisted of flakes of heat-blistered paint and dropletsof molten metal. Closer to the main impact pointand through the area of the major componentslisted above, a large number of lesser componentsand pieces of the aircraft structure were located.

The main wreckage, comprising the completefuselage, the empennage, the port wing, the portengine and the inboard section of the starboardwing, struck the ground at high speed in a steepnose-down attitude at an angle of about 70 degreesto the horizontal. The impact resulted in disin-tegration of most of the structure.

There was abundant evidence from the star-board cen t re w ing , the inboard end of thestarboard outer wing and the structure and com-ponents at the rear of the starboard engine firewall,that an intense fire had occurred in the nacellearea behind the firewall. It was also apparent thatthe wing spearation had occurred because the heat

to which the wing spars had been subjected hadsubstantially reduced their load-carrying capacity.

During the on-site investigation, all of thesignificant items of wreckage were identified,photographed, and initially examined. A surveyplan (see Fig. 1) was also made of the locations ofall components, after which they were collectedand transported to a vacant hangar at Alice SpringsAirport. Here a detailed examination of thewreckage and a re-assembly of the starboardnacelle area was carried out. Subsequently anumber of components were transported toMelbourne for further specialist examination.

Powerplant Layout Each engine nacelle in thisaircraft is divided into two compartments by astainless steel firewall. The engine and its ancil-laries are mounted forward of the firewall andenclosed in cowlings which include side panelswhich may be opened to provide access to theengine for maintenance purposes. The rear nacellecompartment behind the firewall contains themain landing gear wheel-well, the oil tank (whichis mounted on the upper rear side of the firewall),numerous fuel and oil lines and a number ofelectrical cables. The wing spars pass through thiscompartment.

A separate exhaust system is provided forthe three cylinders on each side of each engine.Three short exhaust pipes, one from each cylinder,lead rearwards and are clamped together near theiraft extremities where they are directed into a shortlength of stainless steel ducting called the first-stage augmenter. The output of each first-stageaugmenter is then directed into a main augmenter.The main augmenters are larger insulated ductsclamped to cut-outs on each side of the firewalland exhausting rearwards from each side of therear nacelle; they vent the engine cooling airflowwith the entrained exhaust gases to atmosphere.The main augmenters are positioned close to thewheel-well walls and the nacelle external skinincludes a fairing covering each augmenter exceptfor the discharge orifices.

Powerplant Examination Both propellers werestripped and examined. The examination showedthat the starboard propeller had been featheredand was not rotating at the time of ground impact.The port propeller was found to have been op-erating within the constant-speeding range at thetime that the main wreckage struck the groundand the damage to the propeller blades, clampsand locating tubes was consistent with the propel-ler having been abruptly stopped from a rotatingcondition.

The strip examination of the port enginerevealed no evidence to suggest that it was notcapable of normal operation prior to being severelydamaged by ground impact.

The starboard powerplant, with the enginecowlings and about 18 inches of nacelle structureaft of the firewall, struck the ground heavily in aninverted attitude. The impact resulted in fracturingof the crankcase, and the engine-driven accessorieswere broken from their mounting pads. Thecylinders, exhaust pipes and the first-stage aug-menters remained in approximately their correctrelative positions, but the main augmenters fellseparately. All engine components were retainedwithin the cowls although impact distortion causedthe side cowl panels to burst open.

The strip examination of the starboardengine revealed that the shells of No. 2 mainbearing had sheared the dowels which locate themin their housing and that they had rotated withinthe housing and had also been displaced axially.The loss of restraint of this bearing had been of aprogressive nature, involving fretting away of thetangs on each bearing half, followed by progressivefailure of the stepped dowels eventually leading tocomplete freedom of the bearing to rotate andmove axially on the journal. The spin and axialdisplacement would have had the effect of de-priving the No. 2 main bearing and the Nos. 3 and4 big-end bearings of lubrication. The No. 2 mainbearing had lost all bearing metal and its bronzebacking, leaving only the steel shell, and the No. 3big-end bearing shells had.completely disintegrated,leaving the connecting rod and the crankshaftjournal in a scorched and blackened condition.The No. 4 big-end bearing also showed the effectsof lack of lubrication.

The disintegration of the No. 3 big-endbearing had resulted in a gross increase in bearingclearance with a consequent over-travel of thepiston. As the piston passed bottom dead centreits skirt was pounded by the crankshaft counter-weights and ultimately the No. 3 connecting rodfailed at about its mid-point. The piston and theupper half of the connecting rod lodged in theupper part of the cylinder, but the lower half ofthe connecting rod remained attached to thecrankshaft. The flailing action of this part of theconnecting rod, whilst the engine continued torotate, punched a hole some eight inches square inthe top of the crankcase, penetrated the inductionmanifold at the bottom of the crankcase, breachedthe oil gallery, and removed a three-inch section ofthe camshaft, thus stopping valve operation onNos. 1 , 2 , 3 and 4 cylinders. Both the supply and

scavenge sides of the oil system for this enginewere found to be heavily contaminated with metalparticles, and all bearings showed evidence ofhaving been lubricated with metal-contaminatedoil.

The engine was found to have had AvcoLycoming Service Instruction 1112C incorporatedat the crankcase bearings. This repair schemeinvolved the removal of fretted material from thearea of the through-studs at Nos. 2 and 3 mainbearings and the insertion of selected spacers toretain the correct bearing housing diameter whenthe crankcase halves were bolted together. Evi-dence was found of fretting of the mating crank-case surfaces at Nos. 1, 3 and 4 bearing housings,but not at the No. 2 location.

System examination The aircraft's electrical,fuel, oil, instrument, flight control and enginecontrol systems were subjected to the mostdetailed examination possible having regard to thesevere fire and impact damage which had occurred.

The insulation had been almost completelyburnt away from the electrical looms in thestarboard nacelle area and a number of areas ofelectrical arcing were present on some of the wires.The grouping and location of these arcing in-dications, and the fact that they were associatedwith mechanical damage to the cables, indicatedthat the arcing had occurred at the time of wingseparation. The fire damage to the main electricalloom was only moderate at the connector on theright hand side of the firewall, being most intenseat a point some 13 inches aft of the firewall.

The starboard wing fuel system was re-assembled and it was found that several sections offuel line were missing. The missing sections wereall on or near the inboard side of the landing gearbay and comprised:

(a) virtually the whole of the vapourreturn line;

(b) those sections of the main tankvent line within the landing gearbay; and

(c) sections of the main tank supplyline, the engine supply line, andthe crossfeed line.

The remaining portions of fuel line weresubjected to metallurgical examination as de-scribed in Section 1.16 of this report. The fueltank selector in the starboard nacelle was foundselected to the main tank. The cables whichoperate the fuel tank selector were heat-affectedand broken. The evidence suggests that breakageoccurred at the time of wing separation.

The four-gallon capacity oil tank for thestarboard engine, which is mounted on the upperrear face of the firewall, was partially burned awayand fire had been present inside the tank.

The examination of the aircraft systemsrevealed no evidence of any defect or malfunctionhaving been present prior to the engine failureand fire.

The aircraft was equipped with an enginefire detection system and a one-shot fire sup-pression system covering the area forward of eachengine firewall. It was not possible to determinefrom the wreckage examination whether the firewarning system had operated. The fire extinguisherbott le from the starboard nacelle had beendamaged by the nacelle fire and its contents hadbeen discharged. It was established that the ex-plosive charge in this extinguisher had been firedby electrical detonation rather than by the effectsof heat.

Structural Examination The examination ofthe aircraft structure revealed that, with the ex-ception of the starboard engine nacelle and itsadjacent structure, all of the structural damagewhich occurred was consistent with impact withthe ground. There were large areas of severe paintblistering on the rear fuselage and the empennageas a result of the nacelle fire, but no structuraldamage occurred in these areas from this source.

All identifiable parts of the nacelle structure,the adjacent centre wing and outer wing structure,the firewall and the engine cowlings, landing gearcomponents and doors were re-assembled to moreclearly assess the fire path and structural damage.

An intense fire had been present in the rearnacelle and the resultant structural damage hadbeen largely confined between the main wing ribswhich form the side walls of the landing gear bay.This bay is 20 inches wide.

The f ron t wing spar failed in upwardbending, the lower spar cap failing some six inchesfrom the inboard rib and the upper cap at twopoints, two inches and eleven inches respectivelyfrom the inboard rib. The outboard failure of theupper cap was consistent with an in-flight bendingfailure, while the inboard failure appeared to haveresulted from ground impact. A section of sparweb in the region of the failure was missing. Therear wing spar, which forms the rear wall of thelanding gear bay, also failed in upward bending,the upper spar cap failure being adjacent to theinboard rib and the lower cap failure some fourinches further ouboard. Most of the inboardsection of spar web was missing and the remainder

was severely heat affected and partly burnt away.The spar caps were subjected to a metallurgicalexamination as described in Section 1.16 of thisreport.

The remainder of the aircraft structure inthe landing gear bay area was affected by fire tovarying degrees ranging from slight staining tocomplete consumption. The general pattern of firedamage forward of the wing front spar was fairlyuniform over the width of the wheel bay. Aft ofthe front spar the fire damage was concentratedmore on the inboard side of the landing gear bayas evidenced by the destruction of the lowerportion of the inboard side rib, the concentrationof fire damage on the rear spar close to that rib,and the pattern of burning of the upper wing skin.

The outboard side rib was virtually un-damaged by fire aft of the front spar, but forwardof the spar there was evidence of a severe fire zonein the lower half of the wheel bay on the outboardside. A large section of the rib had been consumedby fire in the region adjacent to the main aug-menter. The maximum intensity of fire damagehad occurred about 15 inches aft of the firewall,where the entire lower panel of the rib had beenburnt away.

The landing gear actuator and its supportstructure, on the forward face of the front spar,were very severely fire-damaged. All other com-ponents of the landing gear were also fire-damagedand the nature of the damage showed that theright main gear had extended during the course ofthe fire, this having occurred because of the failureof the actuator support structure. That portion ofthe tyre which is normally contained within thenacelle was severely scorched by fire, and the tyrewas found to be deflated. The landing gear doorswere both damaged by the fire within the landinggear bay.

The starboard wing flap, which was brokeninto two pieces, was severely damaged by fire onits lower surface in line with the inboard half ofthe landing gear bay. The nacelle fairing which isattached to the lower surface was almost com-pletely burnt away. The nature of the damageshowed that the flap had been in the retractedposition during the course of the fire and at thetime of wing failure.

The outboard main augmenter fairing wasseverely affected by fire and nearly all paint wasburnt from its lower half. A section of the fairingwas burnt away at its lower forward corner whereit attaches to the firewall and an adjoining sectionof the lower edge of the engine side cowl panel

was missing. The paint on this cowl panel wasblistered over its lower half by heat of internalorigin, and the forward part of the blisteringpattern corresponded in shape to the exhaust pipesof cylinders Nos. 3 and 5. The paint was com-pletely burnt away in the area adjacent to theou tboa rd f i rs t -s tage augmenter, which itselfshowed indications of abnormally high tempera-tures. There were no similar heat effects on theinboard first-stage augmenter or on the inboardside cowl panel.

The firewall was extensively crushed anddistorted by impact of the power-plant with theground in an inverted attitude. In addition, thelower section of the firewall was bent rearwardsalong a line passing through the two augmentercut-outs and the pattern of soot deposits in thebuckled areas showed that fire had been presentafter this rearward deformation had taken place.The most severe area of fire damage on the rearface of the firewall was at its mid-depth betweenthe wheel bay side ribs. The forward face of thefirewall showed no indications of fire, but therewas a pattern of smoke staining through theoutboard augmenter cut-out.

The rear clamps on the main augmentersremained attached to them, but the clamp bracketshad been burnt away from their supportingstructure. The forward clamps remained in pos-ition around the firewall cut-outs after the mainaugmenters became detached. In both cases theouter bolt through the clamp halves had beentightened such that the faces were tightly clampedtogether whereas there was a considerable gapbetween the clamping faces at the inner side ofeach clamp. There was evidence of good clampingaction in the form of numerous areas of contactbetween the clamps and the lips on the forwardend of each main augmenter. In the case of theoutboard augmenter, however, these areas hadbeen heavily, sooted over in a manner which in-dicated that fire had been present after, theaugmenter had been displaced from its forwardclamp.

1.13 MEDICAL AND PATHOLOGICALINFORMATION

Post mortem examinations were carried outon all of the victims and no condition was foundwhich could have contributed to the accident.

1.14 FIRE

The details of the in-flight fire have beencovered in the preceding section of this report.

There was no significant ground fire. When theaircraft had not been sighted eight minutes afterreceipt of the message that it was returning, fire-fighting vehicles were despatched in the directionfrom which the aircraft had been expected toarrive. Due to difficulty in locating the position ofthe wreckage, they did not reach the accident siteuntil 58 minutes later but this was of no con-sequence in the particular circumstances.

1.15 SURVIVAL ASPECTS

This was not a survivable accident.

1.16 TESTS AND RESEARCH

A metallurgical examination was made of anumber of structural components and fuel linesfrom the starboard nacelle area to determine themode of failure, the maximum temperature, theduration of heating and the presence of any pre-existing damage.

All of the pieces of fuel line which wererecovered were examined, and some of them werefound to contain holes or splits similar in appear-ance to pressure bursts. It was found that in eachcase of fracture or bursting the failure had oc-curred while the particular section of pipe hadbeen at an elevated temperature in the range400-600°C, the bursts all showing evidence oftemperatures close to 600°C. In some cases thefracture surface was bright, which was indicativeof the fracture occurring at wing separation withrapid removal of the fracture surface from the firezone, whereas in others the fracture was sootedand heat stained which indicated that the fracturedpipe had remained in the fire zone for some periodafter failure. In all cases the bursts occurred due tointernal pressure whilst the pipe was weakened byheat, and fuel was probably added to the fire as aresult of these breaches. There was no evidence ofany pre-existing defect, in any of the fuel linesrecovered, which could have initiated or con-tributed to the nacelle fire.

The examination of the spar caps close tothe positions of failure indicated that a gradient ofhardness (and maximum temperature reached)existed through the cap sections. Experimentswere carried out on unaffected samples of theactual front spar lower cap in an air-recirculationfurnace in an endeavour to establish the durationof heating of the failed spar caps. Equipmentlimitations precluded the development of tempera-tures beyond 900°C and the observed internalh a r d n e s s / t e m p e r a t u r e gradients could not be

achieved at this temperature. From a plot offurnace temperature against the time of heatingrequired for the centre of the cap to reach theappropriate temperature it was clear that theactual flame temperatures must have been inexcess of 900°C and that the duration of heatingmust have been of the order of one minute.

From the observed hardness figures it wasalso estimated that the load carrying capacity ofthe spar caps was reduced by more than 80 percent in the period of maximum heating.

A chemical analysis was made of sootdeposits on the internal structure of the starboardwing in an effort to establish whether they hadresulted from the combustion of Avgas or engineoil. The results were inconclusive.

2 - Analysis

The analysis of this accident is primarilyconcerned with the origin and progress of theintense fire in the starboard engine nacelle whichresulted in the separation of the starboard outerwing. Beyond the point in time at which wingseparation occurred no action was available to thepilot which could possibly have influenced thesubsequent course of events and consequently theanalysis of the operational aspects of the accidentis confined to the period prior to wing separation.

There was no evidence that the start-up, taxiand take-off were other than normal and at0740:25 hours, some five minutes after beingcleared to take off, the pilot made a normal radiocontact on flight service frequency. The contentsof this transmission, which terminated at 0740:31hours (see Appendix B), clearly indicate that hewas not aware at that time of any abnormaloperation. The next transmission from the aircraft,69 seconds later at 0741:40 hours, obviously wasmade after an engine had failed since it consists ofa recognisable landing gear warning horn signal.This horn would only be operating if one or boththrottles were retarded whilst the landing gear wasnot locked down. Since there is evidence that,prior to the accident, the aircraft sustained afailure of the starboard engine, it is reasonable toassume that this event took place during the 69second period between these transmissions andthat, at the time of the second transmission, thepilot had probably completed and certainly com-menced initial feathering action. In the nexttransmission, a further 40 seconds later, the soundof the landing gear warning horn was still present

and the pilot reported that he had an asymmetriccondition and a fire in the starboard engine. Theevidence that the starboard engine fire extinguisherwas discharged electrically indicates that, on be-coming aware of the fire, the pilot carried out thenecessary emergency procedures.

There is little doubt that the black smokepuff observed by the tower controller was theresult of the separation of the starboard wing andengine. It is known to have been visible at 0742:52hours, some 22 seconds after the last transmissionfrom the aircraft was completed, and it obviouslyappeared some time prior to this. The recording ofthe ground/air communications provides valuableevidence of the time scale of the events andindicates that the maximum period between thepilot becoming aware of an abnormal operatingcondition and the in-flight break-up of the aircraftwas something less than 2 minutes and 21 seconds.

Based on the aircraft's performance and theelapsed time from take-off it is probable that aheight of between 2,500 and 3,000 feet aboveground level had been reached at the time that theemergency occurred. At the time of wing failurethe aircraft had completed a turn on to a reciprocalheading and was returning to the airport. Cal-culations made from witness observations of theblack smoke puff, together with a trajectory plotof selected items of wreckage, indicate that theaircraft broke up at a height of about 2,000 feetabove ground level.

In assessing the engineering evidence sum-marised in Sections 1.12 and 1.16 of this report,consideration must first be given to the sequenceof the two major events — the failure of the engineand the occurrence of a fire.

Had a fire existed behind the firewall forlong enough to burn through the oil tank andconsume its contents or cause them to be lost, it ispossible that an engine failure could have occurredas a result of loss of oil supply.

There are several sources of combustiblematerial aft of the firewall, but no evidence of aleakage of any sort was found. However, severalsections of fuel line were not recovered, includingthe whole of the vapour return system aft of thefirewall, and an undetected rupture or leak in oneof these areas must be considered a potentialsource of fuel. In addition, the starboard enginehad a history of fluctuations of fuel flow, exhaustgas temperature and. cylinder head temperatureduring cruising flight. Such fluctuations couldpossibly be attributed to air leaking into the fuelsupply line under the suction head produced by

the engine-driven fuel pump when operating withthe electrically-driven boost pumps off. A cor-ollary to this proposition is that a fuel leak couldhave existed with boost pumps on, during thetake-off and climb segments of the flight. Themain evidence against the existence of such a fuelleak is that, during the course of the attempts totrace the cause of the flowmeter fluctuations aprolonged ground inspection of the starboardnacelle had been carried out with the main boostpump running to pressurise the engine supply linesand no leaks were found.

Even if it is assumed that some leakage ofcombustible material had occurred behind thefirewall there is other evidence that a fire in thisarea did not precede and cause the engine failure.

, Firstly, the No. 2 main bearing failure, which isknown to have been the first step in the enginefailure sequence, was of a progressive nature. Itsdeterioration probably extended over some hoursof engine operation, and bearing material waswidely distributed throughout the engine oilsystem. This indicates that an oil supply had beenavailable to and was circulating through the engineduring the bearing failure process. Secondly, if afire behind the firewall had preceded and causedthe engine failure the time involved from the startof the fire to failure of the wing would be in-consistent with the evidence regarding the time ofheat exposure of the spar caps. Section 1.16 ofthis report shows that the spar had been subjectedto high temperatures for about one minute. It isknown from the communications record that theengine had failed at some time prior to 0741:40hours and that the aircraft was still intact duringthe last transmission which terminated at 0742:30hours. Thus the shortest possible time from enginefailure to wing failure is 50 seconds and this, initself, is close to the one-minute spar exposuretime. If a substantial fire had existed in the rearnacelle for long enough to burn through the oiltank, consume its contents, and finally cause abearing failure, the total time of spar cap heatexposure would have been for too long to beconsistent with the metallurgical evidence.

It can thus be concluded that the enginefailure preceded the development of a fire aft ofthe firewall and it appears certain that the enginefailure was the direct cause of the fire. Theproblem which remains is to determine where thefire started and to trace the course of its develop-ment to the point of causing a catastrophicstructural failure.

Looking firstly at the engine failure, it isevident that the failure resulted in a large hole in

the crankcase, a breaching of the oil gallery,penetration of the induction manifold, and a cam-shaft failure which stopped valve operation ofcylinders 1, 2, 3 and 4. As long as the enginecontinued to rotate there would have been arelease of combustible material into the enginecompartment by virtue of oil being sprayed fromthe crankcase failure and an Avgas/air mixturebeing forced out of the broken induction manifoldby supercharger action. It is not known how longthe engine continued to rotate after failure of theconnecting rod, but the nature of the damage tothe various components and the presence of a largequantity of oil in the engine compartment showthat it did so for a significant time interval.

Several ignition sources would have beenavailable inside the engine compartment; Themixture of oil, Avgas and air entrained by theaugmenter system could have been.ignited by theexhaust gases from Nos. 5 and 6 cylinders whichwould still have been functioning-.'prior to engineshutdown. Other possibilities are,, arcing from adamaged high-tension lead, ignition,of the fuel/airmixture in the induction manifold if the forwardend of the camshaft had stopped in a positionwhich left one of the inlet valves open, or auto-ignition of the engine oil on a still-hot exhaustpipe. The latter appears to be a very likely possibil-ity because the paint blistering on the outside ofthe outboard side cowl shows a distinct patternsuggesting that a fire was at some stage burning onthe outside of Nos. 3 and 5 exhaust pipes.

The fire around the forward sections of theseexhaust pipes was apparently not very intense.There was no observable fire damage to the engineor its mountings nor to any of the components orsystems forward of the firewall. The cowl panelsshowed no internal evidence of heat except in thearea immediately adjacent to the first-stage aug-menter. Here there were very pronounced heateffects and the augmenter itself showed evidenceof abnormally high temperatures, indicating thatan intense fire had been burning in and around thefirst-stage augmenter.

It therefore appears reasonable to concludethat the fire resulted from the release of a con-siderable volume of combustible material followingthe crankcase failure, this being ignited either bycontact with hot exhaust pipes or with exhaustgases, and being almost completely entrainedwithin the outboard augmenter system.

It now becomes necessary to consider themeans by which this fire, apparently containedwithin the normal exhaust system, was able topenetrate into the nacelle area behind the firewall.

A number of possibilities exist. One is thatthe engine compartment fire burned through theside cowl and the flames passed externally rear-wards and toward the bottom of the nacelle,entering the wheel bay through the wheel cut-outin the landing gear doors. However, the hole in thecowl panels at the rear lower corner of the sidepanel, adjacent to the firewall, was a fairly smallone and showed little evidence of local meltingaround its edges, suggesting that the missing pieceshad been broken away rather than burnt away.Nor was there any evidence of a streaming patternin the paint blistering aft of this hole - the patternof paint blistering in this area was in fact muchmore suggestive of having been caused by firewithin the nacelle. Finally, the forward extremitiesof the landing gear doors were not significantlyblistered on the external surfaces as would havebeen the case if flames had entered the wheelcut-out from such an external source.

Another possibility arises from the fact thatthe burning which occurred in the augmentersystem would probably have produced a consider-able torching effect aft of the main augmenteroutlet. The possibility was considered that such aflame had entered'the nacelle by burning throughthe lower nacelle skin. There was certainly evi-dence of much exterior paint blistering aft of theaugmenter, and the rear nacelle skin had beencompletely burnt away on the lower surface aftof the landing gear doors. However, the paintdamage showed no clear streaming pattern aft ofthe augmenter outlet, and at the rear of the nacellethe structural fire damage was strongly indicativeof having been of internal origin. The main pointerto this was the fact that the destruction of thelower nacelle structure extended directly rearwardsin line with the landing gear bay side rib, suggestingthat it had been caused by a fire burning withinthe confines of the bay after the landing gear doorshad opened. In addition, a flame torching rear-wards from the augmenter would have beendirected along a path lying just outboard of theside rib at Wing Station 88 and there was no signof any penetration of the skin in this area.

Another possible mode of entry of fire intothe landing gear bay area would be the leakage of asubstantial quantity of avgas or oil behind thefirewall at the time of engine failure, this beingignited externally by the torching with a sub-sequent flash back into the nacelle. A coincidentlong-term leakage is not considered likely forreasons discussed earlier and there is no evidenceof a defect in the fuel system which could have

produced a fuel leak in the early stages of thisparticular flight. However, the severe vibrationwhich would have occurred whilst the enginecontinued to rotate after failure could conceivablyhave caused some disruption of the fuel or oilsystems aft of the firewall. All of the fuel line nutsappeared to have been tight at the time of the fire,with no evidence of straining, nor was there anyevidence of a pre-existing fuel line defect whichcould have been aggravated by vibration to pro-duce a sizeable leak. However, a number ofsections of fuel line were consumed in the fire andit is not possible to exclude any of these sectionsas the source of such a fuel leak. The only relevantevidence in this respect is that the fractures re-maining were all high-temperature overloadfailures, indicating that at least the ends of themissing sections had failed as a result of the fireand the subsequent structural break-up.

Consideration has also been given to thepossibility that the engine vibration induced apartial collapse of the nacelle structure, whichcould very readily have produced a fuel or oil linefailure aft of the firewall. There is certainlyevidence that the nacelle had drooped to someextent prior to the final structural failure, butthere was no evidence that this was induced byvibration. In fact, the compression failure of thelower nacelle skin panel aft of the firewall showedevidence of failure at high temperature, indicatingthat it had occurred well after the development ofthe internal fire and probably as a result of theweakening of the panel by that fire.

The final possible entry path to be con-sidered is via the wheel bay side rib adjacent to theoutboard augmenter. Both main augmenters werefound in the wreckage trail and they had both lefttheir forward clamps still attached to the firewall.Although the forward lips of both augmenters andtheir respective clamps showed evidence of goodclamping contact, the presence of substantial sootdeposits over these clamping areas in the case ofthe outboard augmenter showed that fire had beenpresent subsequent to its detachment.

The indications were quite consistent withthe displacement of the outboard main augmenteras a result of severe engine vibration and the con-sequent entry of the intense first-stage augmenterfire into the space between the main augmenterand the wheel bay side wall. In the absence of anysignificant cooling airflow on the inboard side ofthis rib, penetration could be expected to occur ina very short time. Once having entered the wheelbay area, several sources of additional fuel for thefire would become available by way of the fuel

lines and the oil tank, which would probably stillhave contained a quantity of oil after the engineceased rotation. The vapour return line and thosefuel lines in which fuel was not flowing would havebeen particularly susceptible to rapid breaching bythe fire. It should also be noted that the heat-induced bursting failures of the main and auxiliarytank supply lines were at comparatively low pointsin the system and would have allowed a substantialdischarge of fuel even with the boost pumps notrunning.

The overall pattern of fire damage is con-sistent with the above hypothesis; particular pointsbeing the almost complete destruction by fire ofthe side rib in the area immediately adjacent to theoutboard main augmenter and the evidence of highinternal temperatures surrounding the main aug-menter as shown by the complete removal of paintfrom the lower nacelle panel outboard of the siderib.

The adequacy of the clamping action on theforward end of the outboard augmenter could notbe accurately assessed. However, it was noted thatthere was a gap between the clamp halves of theinner bolt, whereas the outer clamps faces weretightly bolted together. To ensure maximumclamping efficiency a gap should be present be-tween the clamp halves at both the inner andouter faces with both bolts correctly torqued.

Although other possibilities cannot be com-pletely excluded, the available evidence supports aprobability that the last-mentioned hypothesisrepresents the actual mode of entry of the fire intothe landing gear bay area.

Thus the probable sequence of events maybe summarised as follows:

(1) An engine failure which involvedrupture of the crankcase, the in-duction manifold and the oil galleryoccurred as a result of a progressiveloss of restraint of the Number 2main bearing shell.

(2) The engine failure allowed therelease of a large quantity offlammable fluids into the enginecompartment.

(3) The flammable mixture, consistingmainly of engine oil, was ignited bycontact with exhaust gases or withthe hot exhaust pipes and resultedin an intense fire burning in anda r o u n d the outboard first-stageaugmenter.

(4) Engine vibration subsequent to thefailure and prior to the cessation ofengine rotation resulted in the dis-placement of the forward end ofthe outboard augmenter from itsfirewall clamp and allowed the fireto enter the space adjacent to thelanding gear bay side rib at WingStation 88.

(5) The fire penetrated the rib andentered the landing gear bay.

(6) The resultant internal fire forwardof the wing front spar destroyedthe landing gear actuator supportstructure, allowing the landing gearto extend.

(7) Additional fuel was added to thefire by the breaching of fuel linesand the burning away of the rearface of the oil tank.

(8) The fire spread rearwards throughthe landing gear bay area, rapidlyweakening the wing spars to theextent that wing failure occurred,this being preceded by a downwardcol lapse of the engine supportstructure.

The final aspect to be considered is thereason for the loss of bearing restraint whichinitiated the above sequence of events. There arethree design features which can contribute to therestraint of the main bearing shells. The primarymeans of restraint against the tendency of thebearing shells to rotate is by the clamping actionor "nip" of the bearing housings on the shellswhen the two crankcase halves are bolted together.Additional restraint against rotation and alsoagainst any axial movement of the shells is pro-vided by the small tangs on the outer faces of theshells, which engage in appropriate recesses in thecrankcase housings, and by the dowels which arefitted in the housing and engage in slots in thebearing shells.

It is known that fretting of the matingcrankcase surfaces at the Number 2 and 3 mainbearing saddles can occur in service, and that theresulting loss of metal can change the size of thebearing saddles sufficiently to cause loose through-bolts or excessively tight crankshaft bearing fits.Avco Lycoming Service Instruction No. 1112Cwas introduced to deal with this problem byremoval of fretted material and the installation ofselectively fitted spacers to maintain the correct

housing diameter when the crankcase halves arebolted together.

The failed engine was found to have had thisrepair procedure incorporated, but it was notpossible to determine when or where it was carriedout. There was no relevant Log Book entry and theoverhaul agency had no record of such workhaving been done.

There is evidence that, subsequent to therepair being incorporated, further fretting hadoccurred in the bearing saddle areas, but it is clearthat such fretting played no part in the bearingfailure process. In fact, fretting was apparent atmain bearing positions 1, 3 and 4, but not at thefailed Number 2 bearing location. This suggeststhat the Number 2 bearing saddle was less suscep-tible to fretting by virtue of having been moretightly clamped together than the others, and thisin turn implies that there may have been signifi-cantly less "nip" on this particular main bearingshell. The existence of such a condition couldexplain why the Number 2 bearing sustained thetype of failure observed, but it must be remem-bered that the engine had achieved a total of 716hours operation since the last overhaul and itappears most unlikely that any gross error inassembly tolerances could have existed for such along time before resulting in a failure. The crank-case failure was so extensive around the Number 2bearing area that it was not possible to make anymeasurement of the housing diameter.

In considering the causal areas for thisaccident it seems clear that the sequence of eventscommenced when loss of restraint of the Number 2bearing shell resulted in an extensive failure of theengine structure. Engine failures of this magnitude,however, have occurred previously on frequentoccasions and the purpose of the firewall is tocontain, within the engine compartment, any firewhich may result from such occurrences. Thea v a i l a b i l i t y of combustible fluids within theengine compartment is limited by design require-ments and this factor provides a further restrictionon the intensity and duration of any enginecompartment fire. On this occasion, probably as aresult of the displacement of the forward end ofthe outboard augmenter from its clamp, the fire-wall did not fulfil its intended function and theengine compartment fire entered an area wherecombustible fluids were readily available.

There is evidence from the contact marks onthe clamp and on the lips at the forward end of theaugmenter that clamping, adequate for any normalsituation, had been achieved. However, the severityof the vibration, which would certainly have been

present in an engine failure of this nature, is notknown, and may have exceeded the limits of afitting of this design. Alternatively, the clampingaction, while being adequate for normal situations,may have been less than adequate for abnormalsituations by virtue of the uneven distribution ofthe gap on each side of the clamp halves. Thereason for a loss of clamping effectiveness in thecircumstances of this particular engine failurecannot be positively determined.

3-Conclusions

1. The pilot was properly licenced and experi-enced for the flight and there was no evidence ofany pilot incapacitation which may have contrib-uted to the accident.

2. There was a current certificate of airworthi-ness for the aircraft and, although it is probablethat the starboard engine had developed an internaldefect at the time the flight was commenced, thereis no evidence to show that this should have beendetectable in the normal maintenance procedures.

3. Shortly after take-off, while the aircraft wasclimbing to cruising height, the starboard enginesustained a massive internal failure which resultedin a large quantity of oil and fuel vapour beingreleased into the engine compartment. At the sametime, probably as a result of severe vibration, theforward end of the outboard main augmenterbecame displaced from the clamp which attachesit to the firewall cut-out.

4. The combustible fluids in the engine com-partment became ignited, probably by contactwith outboard exhaust system components, andthe resulting fire was carried by the. airflowthrough the outboard firewall cut-out. Due to thedisplacement of the main augmenter the fireburned through the side wall of the wheel bay andentered the rear nacelle area.

5. Intensified by a further supply of fuel whichwas made available by heat breaching of the oiltank and fuel lines, the fire impinged on andweakened both the main and rear spars and thisled to separation of the starboard outer wing andthe starboard power plant.

6. The time interval between the engine failureand the separation of the starboard wing andpower plant was less than 2 minutes and 21seconds.

7. There is evidence that the pilot acted ex-peditiously to shut-down the starboard engine andthat he activated the engine fire extinguishingsystem.

CdllSG The probable cause of the accidentwas that, following an engine failure which re-sulted in severe vibration and a fire in the outboardrear section of the engine compartment, theintegrity of the firewall and its attached exhaustducting was lost and structure at the rear of theengine nacelle was thereby exposed to fire.

Appendix A

LIST OF OCCUPANTS AND PROBABLE SEATING ARRANGEMENT

SEAT

1

2

3

4

5

6

7

8 or 9

8 or 9

10

OCCUPANT

Nigel Clifford HALSEY

Douglas George FR1EDRICHS

Gregory Brian CAIRNS

Michael Scott CAIRNS

Laura Anne CAIRNS

Unoccupied

Unoccupied

Eleanor Short KNE1PP

Robert Frederich KNEIPP

Unoccupied

APPROX.AGE

25

39

18

m

17

51

60

ADDRESS

Alice Springs

. Ayers Rock

Alice Springs

Alice Springs

Alice Springs

U.S.A.

U.S.A.

INJURY

Fatal

Fatal

Fatal

Fatal

Fatal

Fatal

Fatal

Appendix B

TRANSCRIPT OF TAPE RECORDINGALICE SPRINGS TOWER 20 JANUARY 1972

LEGEND

TWR Alice Springs TowerFS Alice Springs Flight ServiceFSI Flight Service position one* This transmission was recorded on the tower tape through the flight

service monitor facilityNOTE: Times of transmissions are Central Standard Time (G.M.T. + 91A hours).

Times included in transmissions are minutes past the appropriate G.M.T. hour.

TIME

0733.06

.11

.31

TO

TWR

CMI

FROM

CMI

TWR

.200735.06

.08 .

.13

.21

.23

.260740.06

.10

.16

.18

.20

.22

.25

.28

.310741.40*0742.20

TWRTWRCMITWRTWRCMITWRTWRCMITWRTWRFSTWRFSCMIFS

TWR

CMICMITWRCMICMITWRCMICMITWRCMIFSTWRFSCMIFSCMI

CMI

0742.50

0742.52

CMI

TWR

FIRE

TWR

FIRE

TWR

FIRE

TWR

FIRE

TWR

Alice Tower good morning CMI for Ayers Rock, request taxiclearance.

CMI good morning this is Alice Tower runway 12 wind light andvariable, tending nor' easterly QNH1014 time 02 !£ line up.

CMI

CMI request airways clearance.

CMI clearance track 235 cruise 6000.

CMI 6000.

CMI is ready ."•

CMI, clear for take off, make right turn.

CMI right turn.

CMI departure 09.

CMI area QNH 1014 call Alice Springs 122.1

CMI

(answering intercom) FSI.

Departure CMI 09.

CMI

Alice Springs good morning CMI listening 122.1

Good morning CMI readability 5.

CMI

- BEEP - BEEP (on 122.1) with V* second engine noise.

Alice Tower CMI, er we are returning to the field I have er, aasymmetric condition er a fire in the starboard engine. (ContinuedBeep Beep in the background.)

CMI Roger right base runway 12 report on final (Noacknowledgement is received to this transmission.) There is aclick — phut (microphone noise?)

Yes we copied that mate

Roger can you turn out

How far out is he

You can see him there, see that smoke

72-2