Introduction

On 27 November 1987 South African Airways flight SA295 departed from Tapei, Taiwan. The flight was due to stop in Mauritius to drop off and pick up passengers before flying on to Johannesburg. There were 159 passengers and crew on board. The pilot in command was Captain Dawie Uys, 49, a veteran with more than 13 843 hours flying experience, 3 884 of those on the Boeing 747 aircraft type.

The plane he was flying was a seven-year-old Boeing 747, named the Helderberg, of the type called a Combi. SAA operated two of these aircraft in its fleet in the 1980s. The idea behind the design was that on low density routes, an airline could remove some of the seats from the rear of the passenger deck and create a cargo bay in the freed up space at the back.

About nine-and-a-half hours after take-off the crew reported smoke aboard the aircraft to the control tower at Plaisance Airport in Mauritius. 15 minutes later all radio contact was lost. It was subsequently established that a fire had started on board causing the aircraft to crash in the Indian Ocean. There were no survivors.

With 159 lives lost the Helderberg disaster remains the worst accident on record for SAA. Although the crash was investigated by a commission of inquiry headed by Judge Cecil Margo - and including a number of international experts from the UK, USA, Mauritius, Taiwan and South Africa - a number of questions around the crash have never been answered.

This explanatory void has been filled by a number of unsubstantiated theories, relating mostly to alleged sanctions busting by the then National Party government.

However it may be that the answers to these questions may have been hiding in plain sight. Recent research into the history of in-flight airliner fires may provide the vital clues about the most pressing question of all, namely how the fire started. Before going on to discuss this matter it is important to first set out the established facts of the case.

The flight

On the afternoon of 27 November 1987, flight SA295, departed Taipei's airport at 14h23 UTC. This was about 85 minutes later than planned. The delay was due to the presence of heavy thunderstorms over the airport. These storms delayed take-off as well as the arrival of inbound flights with passengers booked on the SAA flight to Johannesburg.

During the delay, however, a random check was conducted on pallets of cargo in the rear of the main deck by Taiwanese customs officials. They did not record anything untoward and found the manifest to agree with the actual cargo they inspected. Tellingly, according to the record, among the pallets they inspected was pallet PR, situated at the right front of the cargo area.

The aircraft took off without incident and maintained regular contact with all scheduled control stations along its path until it headed out over the Indian Ocean en-route to a stop-over in Mauritius ahead of its early Saturday morning arrival in Johannesburg. All indications are that the flight passed without incident for the greatest part of that ocean crossing.

About 300km north east of the island, however, the crew of SA295 radioed the control tower at Plaisance Airport in Mauritius to report that they had encountered a smoke problem aboard the aircraft. The captain stated that they had already set in motion the required emergency procedures - among which was a descent to 14 000ft.

The last radio call from the airliner to the tower at Plaisance appeared to indicate that the Captain felt sufficiently in control of the aircraft to make a 180 degree turn and conduct a normal landing for the prevailing conditions.

Sadly, no further contact was made with the airliner nor did it land in Mauritius within the time planned on the schedule.

After the passage of some hours it became apparent the aircraft was in distress. A search and rescue effort was launched by the Mauritian authorities near the island using a helicopter but no sign of the airliner or its occupants could immediately be found.

At daybreak ships and aircraft from many nations joined the search -many of their own accord. Later in the day the crew of a private aircraft located some items of floating debris and two oil slicks on the sea some 240km from Mauritius. Once surface vessels had retrieved the debris examination of it showed that the impact forces with the sea had been catastrophic and that there was little hope of finding survivors.

The investigation

The South African Civil Aviation Act then in force provided for the establishment of an official board of inquiry into any airline accident - which was mandated to establish the cause and culpability for the accident. Thereafter, the board was to make recommendations which would prevent a similar accident in the future. All hearings in such investigations were conducted, as per the provisions of the applicable legislation, in public and any interested parties could attend and/or submit testimony.

Accordingly, an official board of inquiry was set up headed by an internationally respected expert in the field of aviation, Justice Cecil Margo. Among the board members were Mr Tompkins of the USA, Mr Wilkinson of the UK, Mr Justice Goburdhun of Mauritius, Dr Funatso of Japan, Mr Lung from Taiwan and Messrs' Gilliland and Germishuys of South Africa.

The first task was to examine all available evidence, the collection of which presented formidable challenges. The ocean floor in the region of the accident lay some 4 400m below the surface. Operation Resolve was launched at huge cost to locate and then recover the wreckage. Given the great depths involved, special equipment had to be manufactured for this purpose. It was a pioneering effort of deep sea salvage that remains unsurpassed to this day.

The search and recovery process took more than 2 years. All the wreckage on the sea floor - laying in two debris fields some 600m apart - was mapped and photographed. It was apparent from the wreckage that a fire had been raging in the rear cargo bay. The investigators decided to try and retrieve as many items showing evidence of fire as possible.

Given the complications presented by the depth of the site, the weather and time on hand, however, not all of these items could be retrieved. A vital clue in any accident is the data captured on the Cockpit Voice Recorder (CVR) and the Digital Flight Data Recorder (DFDR). These recorders were not found attached to the rack that normally held them in place at the rear of the aircraft.

An SAA engineer Dr. Theuns Kruger applied considerable mathematical skill and ingenuity to the task of locating the CVR. This item was eventually found almost exactly within the area he had predicted, somewhere between the main debris fields. The DFDR, however, was not recovered.

In addition to the CVR, priority targets such as particular portions of the rear cargo bay, items of cargo and other wreckage, were retrieved and re-assembled on a frame representing the aircraft in Hangar 1 West at Jan Smuts Airport (later renamed OR Tambo International Airport.)

The CVR was carefully preserved and sent to the USA to be analysed. It was found to contain only a short period of relevant conversation - spanning the period from shortly before the initial fire alarm to about 80 seconds thereafter when the recording stopped.

The exact time when these discussions took place could not be pin-pointed by the board but, taken on usual crew reaction and procedures in such an emergency, the relevant recording could not have been made more than 5 minutes or so before the first radio call to Mauritius. Analysis of the recorded events, however, indicate that the crew followed their emergency checklist procedures promptly.

The official Board of Inquiry found that, on the available evidence, it was apparent that the fire had taken place in the cargo bay between 300 and 250km from Mauritius.

The Board visited a mock-up of the rear cargo bay constructed by SAA technical where a fire was simulated using cardboard and other packaging materials determined to have been in the cargo bay. The mock-up did however not fully represent the aircraft in every detail as the wiring looms and other structural components used in the actual aircraft were not included. The test confirmed the difficulties involved in attempting to contain and extinguish a fire in the cargo area of a 747 Combi.

Conclusions of the Board

After hearing evidence from various parties - ranging from the supplier of the airliner, its operator and various civil aviation agencies involved with the certification and testing of the Combi design, the Margo Board found that the firefighting regulations and safety equipment regulations applicable to the aircraft were woefully inadequate. The procedures and equipment standards governing the cargo bay had first been set up in the 1940s for far smaller aircraft. Shockingly, they had never been updated to keep pace with the growth in size of airliners in the interim.

The board made many suggestions as a result - one of which involved the banning of the Combi concept as posing an unacceptable risk to life and property. They also suggested that automated fire-fighting facilities be installed in similar cargo bays in future. Another suggestion was that provision be made for self-contained, isolated CVR and Data Recorder power supplies with a time-code being added to the recordings as a matter of course.

The board was, however, unable to determine the exact cause of the fire, nor exactly how the aircraft came to crash in the sea. While some evidence suggested that controls were compromised and the fuselage most likely broke in two at about 14 000 feet, the manufacturer vigorously contested such scenarios.

The Board, therefore, restricted itself to making findings on the most important steps required to prevent a recurrence of a similar accident in future - which, in terms of the law, was its primary purpose. In any event, a definitive finding on the cause and consequences of the fire would not have altered the fact that any fire needed to be detected and extinguished promptly. As a consequence of the board's recommendations, the fire-fighting regulations and equipment specifications involved with similar cargo bays (known as Class B bays) were overhauled throughout the world to prevent any recurrence of the tragedy.

The unanswered questions

That the board did not pinpoint the source of the fire has, however, remained a point of anguish for family members of the victims of the crash, who are still seeking some kind of closure. For those alleging foul-play, the added refusal of the authorities to re-open the investigation has provided fertile ground for speculation in the intervening 27 years.

Much has been made by various parties over the years of allegations that the National Party government had been using the aircraft of the national airline to smuggle weapons. The principal claim is that a type of oxidant used in rocket fuel called Ammonium Perchlorate (APC) was aboard the aircraft and that this had ignited. This theory was put to the original Board of Inquiry by an expert testifying on behalf of the manufacturer of the aircraft and it has been regularly hauled out before the court of public opinion ever since.

Why the then government would have seen fit to be importing this substance is a mystery. It was manufactured in the Republic and had been available in the country for more than a decade before the crash occurred. A simple faxed order to Sonchem would have procured as much as was required without the need to airfreight it aboard a passenger airliner.

Nevertheless, if it were true that the airline was carrying dangerous items, one would have expected the powers-that-be to have heaved a collective sigh of relief at the facts of the accident location - conveniently far out in the ocean at nearly its deepest point. They could then simply have issued a statement that said "Sorry about the crash. That part of the sea is too deep to find anything." and left matters at that. They did nothing of the sort.

Operation Resolve was a hugely expensive undertaking which swallowed up much-needed foreign exchange. It involved volunteers, aircraft, ships and expertise from many countries - including the former Soviet bloc. Many SAA and Civil Aviation Department staff spent more than 2 years in Mauritius, far from home. Endless hours were dedicated to the investigation.

Had the then government had had something to hide it seems unlikely that they would have gone to such great lengths to shed light on the cause of the accident. It would also have been very risky to allow international involvement at the heart of the investigation if there was any chance of any dirty secret coming to light.

New information on in-flight fires

I have been researching and writing in the field of aviation safety for more than 23 years. Historically, the majority of airline accidents on record have been caused by a tragic confluence of seemingly unrelated circumstances that, on a given day and with specific parameters, lead to a fatal result.

In hindsight, therefore, most airliner accidents are fully explainable through logical assessment of the known facts. Examination of the record of the particular type of aircraft and the evidence on hand from the actual accident being examined - more often than not - lead investigators to the cause of the tragedy.

Research for a new book project begun in 2010 led me to in-depth detail concerning numerous in-flight fires aboard other airliners which have taken place since the loss of the Helderberg.

A large proportion of the fires on record have been ignited by a specific type of polyimide coated wire, known as MIL-W-81381 or Kapton after the trade name of the insulation material (see video here). This type of wiring was fitted throughout the Helderberg.

This wiring - thinner than a few strands of human hair - was at first believed to be extremely durable. It has, however, been demonstrated that the insulation of the wire developed cracks after just a few years of service. Additionally, it has been proven that in older passenger jets such wiring with cracked insulation is prone to a phenomenon known as wet-arc tracking in humid environments. Such humid conditions tend to develop inside passenger aircraft after a number of hours of flight due to the condensation of moisture from the exhaled breath of passengers. Dry arcing is also a problem where the insulation has abraded and is exposed against the aircraft structure.

I believe, however, that the Helderberg may well have succumbed to the explosive, damaging effects of wet arc tracking.

The extremely high heat given off during this arcing has repeatedly set fire to thermal insulation blankets installed against airliner fuselages - most notably AN26 or Metalised Mylar insulation blankets.

These blankets - which were also fitted to the Helderberg and exposed in the cargo bay area - were, in the 1970s and 1980s, erroneously believed to be non-flammable. They have latterly been removed from airliners after it came to light that they are highly flammable when directly exposed to a heat source (see here). Once they have ignited they burn rapidly and release large amounts of toxic smoke.

The combination of aging and cracked Kapton wiring looms clamped directly against or near these dangerous insulation blankets has, however, ignited many in-flight fires since 1975. The problem was, however, never seen for what it was until more and more aircraft fell victim.

A long list of fires

The current body of technical knowledge about Kapton and insulation blankets was, however, never on tap to the investigators of the Helderberg tragedy. This is due to the methods used to record the incidents and a reluctance by interested parties to face up to the seriousness of the problem.

Even now the information about these fires and the involvement of aged, cracked Kapton wiring and flammable insulation blankets lies scattered in the archives of various airlines and international aviation bodies. Where arcing might have caused the loss of instruments, the matter has often been listed only as an "instrument failure".

Thus a concerted search was required to find the events caused by these two materials.

The first serious alarm was raised in the 1970s when TWA alerted Lockheed to the problems their wiring experts had discovered with Kapton after performing some simple tests to inspect the durability of the insulation materials. As a consequence they asked that their Tristar airliners not be fitted with it. The Lockheed executives met with the TWA executives. A lot of negotiating took place. The orders proceeded.

In the interim, the US military grounded and re-fitted - where possible - thousands of aircraft due to accidents blamed directly on arcing wires. New build aircraft were not permitted to be wired with Kapton.

Early in 1988, a month or so after the loss of SA295, articles on hand record that United Airlines wrote to Boeing stating its refusal to accept any further aircraft wired with Kapton wiring. In a repeat of the TWA concerns, the manufacturer managed to persuade the airline to accept the airframes - against the wishes of its technical specialists who were over-ruled.

The matter raised itself again in 1993 after a McDonnell Douglas MD82 aircraft experienced an in-flight fire while on approach to land in Copenhagen. The Scandinavian investigators found that wet arc tracking in the Kapton wiring started the fire which then burned through the insulation materials. The energy released was massive, with more than 140 Amps of electrical energy release generating temperatures as high as 2000 deg C.

The arcing melted fuselage parts and ignited the insulation material against the inner skin of the aircraft which subsequently burnt at temperatures exceeding 1000 degrees C. Of interest is that in this, and many subsequent events, the circuit breakers of the aircraft had not been tripped until the fires were well established. This allowed the arcing to continue for lengthy periods and generate more than sufficient heat to set fire to the thermal insulation materials.

Another example of the tragic consequences of this combination of factors was the accident to Swissair flight 111 which caught fire near Nova Scotia, Canada in 1998. That fire was also found to have been caused by explosive arcing of the aircraft wiring which, in turn, set fire to the insulation blankets. The resultant blaze quickly brought down the aircraft. Swissair 111 finally brought this dangerous combination of materials into the spotlight and led to action by regulatory bodies to prevent a recurrence (see National Geographic programme here - especially from 37 minute mark).

Kapton wiring has now been superseded in all newly designed Boeing aircraft since the late 1990s. It has been replaced by a Teflon coated version called TKT that has, apparently, resolved the arcing issue. Metallised Mylar and AN26 - the names of the principle insulation blanket materials used in airliners in the 1970s and 1980s - have also subsequently been banned and removed from all airliners.

A new look at old evidence

So, did the Margo Board miss all this? Not really.

There are disparate references to arcing wires, globules of molten copper from the wiring, and detached and molten insulation materials, scattered about the evidence brought to the board as well as in the final, official report. At the time, however, the combination of Kapton wiring and the insulation materials were thought to be benign and they were, therefore, probably never suspected as possibly lying at the root of the blaze.

Of interest is one particular source - independent of the airline or state machinery - that references the arcing of wiring, molten copper globules, extreme temperatures and detached insulation in the Helderberg accident.

This is the Burgoyne Report submitted to the original Margo inquiry. It was, interestingly enough, commissioned by the manufacturer of the aircraft. This report stated that the fire had burned fairly rapidly and with high heat in its initial stages.

It furthermore details evidence of arcing in the wiring recovered from the region of the cargo bay and states that the retaining clips of the insulation blankets in the area of the fire's origin had melted, permitting the insulation blankets to fall on to the cargo of pallet PR and others nearby.

A V-shaped pattern of fire damage on the fuselage skin was described in the Burgoyne report as the region of the highest heat. It placed this area between fuselage formers 1800 and 1820 in the region of stringers 14 and 15 (see illustrations). Of note is that this area is well above the cargo pallets where most experts of the day, ignorant of the twin dangers of Kapton and the flammable insulation blankets, surmised - or stated - the fire must have started.

Cargo bay photograph:

A photograph of a Boeing 747-Combi cargo bay showing the wiring loom (within the white frame) up against the insulation material. The position of pallet PR, long suspected as being somehow involved in the SA295 fire is ghosted over the image in red. The relationship between the area highlighted in the Burgoyne report (Pink circle), the wiring loom, the silver/white insulation material and the pallet is clearly evident. (Credit: Original Photograph © Walter Van Bel, Belgium used with permission. Annotations by the author.)

Putting the puzzle together

After reading through yet another modern report detailing the combustion sequence caused by Kapton wiring clamped against the insulation blankets, I decided to plot the region of the highest heat detailed in the Burgoyne report on diagrams of the recovered wreckage from the Helderberg.

I also discovered that a major Kapton wiring loom ran above pallet PR in the huge cargo bay. (See photograph) I then plotted this loom's position on copies of the same drawings.

Superimposing these two plots reveals that they coincide at precisely the point of highest heat mentioned in the Burgoyne report and where droplets of condensation could have run down from the crown of the fuselage - possibly pooling around the top wires in the bundle.

Given the age of the wiring it is very likely that hair-line cracks had developed in the insulation at this time (a known problem on that wiring type after a few years) and the entry of droplets of moisture could have started high temperature arcing of more than 2000 degrees C - the usual temperature recorded in these types of incident.

Wreckage plot 1:

The area of highest heat plotted against a drawing of the Helderberg to show its relative position together with the location of the wiring raceway against the inner fuselage. The numbers refer to debris item numbers recovered by the investigators. (Credit: Drawing from the Margo Report. Annotation by the author.)

Arcing wiring in the loom would then easily have ignited the insulation blankets. Once alight these would then have dropped onto the cargo below after having melted the nylon retaining clips holding them against the fuselage.

The findings in the Margo and Burgoyne reports, in my view, support each other when read in light of the current knowledge about these materials and this ignition scenario.

Further credence to this theory is lent by the recovered CVR tape.

The background noise on the tape is fairly regular until shortly prior to the sounding of the smoke alarm. An increasing amount of loud pops and clicks can then be heard on the recording leading up to the triggering of the smoke alarm and the discussion of the matter by the crew. The CVR records the engineer stating at the onset of the smoke alarm that circuit breakers were popping out.

Wreckage plot 2:

The area detailed in the Burgoyne report as the location of the greatest heat in theHelderberg fire plotted on a drawing of the fuselage. The recovered wreckage is traced in black on the fuselage sketch. (Credit: Drawing from the Margo Report. Annotation by the author.)

Later during communication of the flight crew to Mauritius they said they had "...lost a lot of electrics. We don't have anything on the aircraft now." The tape also has further loud pops on it as well as changes in pitch and amplitude which could have arisen as the voltages in the wiring loom fluctuated and the tape speed varied.

Thus the CVR tape may well have recorded the origin of the fire and its exact timing.

On the basis of our current knowledge regarding in-flight fires, I would argue that the Margo Board of Inquiry could not have looked into all the possibilities regarding ignition of the fire. I do not believe that this was a deliberate oversight. It was, more than likely, caused by its reliance on submissions by the respective manufacturers and specialists as to the alleged fire safety of the wiring and insulation material, and by the state of knowledge at the time.

Answers to many questions

My view is that there is more than enough information available worldwide today regarding the danger posed by these materials, present on the Heidelberg, to warrant revisiting the matter. This hypothesis would, moreover, explain a number of unresolved issues around the accident. These include:

• The late onset of the fire condition after many hours - the condensation would have taken time to form from exhaled breath and steam from galley ovens and kettles.
• The clicks and pops picked up on the CVR, starting shortly before the activation of the smoke alarm.
• The point made by the Burgoyne report that the initial combustion was a high-energy source and situated against the fuselage above the cargo pallet and not within pallet PR which had been checked by the customs officials in Taiwan.
• The Burgoyne report stated that the fire was later of a small area and burned at the top of the cargo pallets - this would be consistent with burning insulation blankets dropping onto the cargo.
• The melting of aluminium fuselage components seen on recovered wreckage fits with the known temperatures and energy release reached in Kapton arcing events - many have been recorded as being in excess of 140 Amperes.
• The large volume of smoke involved as testified to at the Margo hearings - this is consistent with the outcomes of tests involving burning AN26 and metallised Mylar insulation blankets.
• The change in pitch on the CVR recording and the loss of power to the unit - this would all be expected if the wiring loom underwent an arcing event or fire.
• It would also explain the evidence recorded in the evidence of recovered wiring with traces of arcing events and the tell-tale molten copper globules.
• The total lack of physical evidence of explosive substances or any other likely sources of ignition on the recovered wreckage or in the declared cargo.
• The inability of the crew to contain the fire.
• The damage to sections of the crown of the fuselage, high up and far away from the small sofa sized fire on top of the cargo described in the Burgoyne report - this is consistent with the fire spreading rapidly through the insulation material to the top of the fuselage.

The case for a new investigation

In keeping with the government's open offer to view and consider new, credible and verifiable evidence, I provided the Hon. Minister of Transport a briefing document on my findings earlier in the year. She was most gracious in receipt of this and informed me that the matter had been passed to the CAA for follow-up.

I suggest that the thousands of photographs and the many hours of video recordings of the wreckage be re-examined as a first resort. In addition, a thorough, faithful reconstruction of the cargo bay, complete with all wiring looms, sensors, switches, insulation materials, mock-up cargo pallets, ventilation systems and other detail could then be constructed. Wet arcing could be induced in the region of highest heat recorded in the Burgoyne report and the results observed.

If the information collated can provide a possible resolution of an, as yet, unexplained airframe loss then all the effort and expense in following up on the leads by all concerned will be more than worthwhile.

Such an exercise might complete the record and permit the goal of aviation safety investigation to be met. All relevant information from accidents should always be available in order to alert aircrew of the dangers exposed by investigators so as to prevent a similar tragedy.

Such a resolution - if it does indeed transpire - would also, I trust, bring closure to the family members of victims of this terrible accident.

The core issue for me is maintaining the integrity of the Air Accident Investigation process in South Africa. We owe it to history to properly investigate this matter and, if possible, to finally put it to rest.

Note: Numerous e-mails were sent to the manufacturer of the Helderberg for comment on the matters raised herein. However, no response from Boeing had been forthcoming by the time of publication.  

Mark D Young has been researching and writing about civil aviation accidents for more than 23 years. He published a book on the history of SAA accidents in 2007 titled "A Firm Resolve". His book on the new evidence found in regard to SA295 is called "Resolved - The story of a Boeing called Helderberg." Full details of the research are available here.  

Click here to sign up to receive our free daily headline email newsletter