Misdiagnosis: Convair Crash at Wonderboom (part 2)
Last week, we went through the flight and crash of a Convair 340/440 at Wonderboom Airport in Pretoria. As a quick recap, as the Convair rotated the left engine caught fire. The flight crew declared an emergency but they lost control of the aircraft while still turning back as the fire melted through the aileron cables. The aircraft crashed through a factory building and impacted the ground on the other side.
The engineer responsible for the maintenance of the aircraft, who was sitting on the jump seat with the flight crew, was killed in the crash. The flight crew, two of the passengers and four people on the ground were seriously injured.
Last week we looked at the issues on the flight deck and that the flight crew did not appropriately react to the fire that they knew was burning in one of their engines. It is easy to wonder why they disregarded the checklists and did not make any attempt to extinguish the fire, which rapidly burned through control cables making it impossible for them to fly the aircraft. However, we also need to remember that the flight crew is often blamed for their errors because they are the last line of defence when things go wrong. Sure, we expect competence from commercial pilots but also they are in the one position where simple mistakes can’t be forgiven. A simple error on the apron or in the hangar may not feel like it has such overwhelming consequences but the truth is, it is just as much a part of the chain of events leading to the crash as the flight crew’s mistakes.
So although there are a lot of questions that can be asked about the flight crew’s response, there’s a key issue that we have to not lose sight of: an engine should not burst into flames on take-off.
When the investigators arrived on the scene, they first determined that fuel and oil were all as expected. The right engine had been operating normally up to the moment of impact. They focused on the left engine as the passengers clearly reported that they’d seen flames as the aircraft rotated.
The left-hand engine was located just behind the fuselage at grid station J25. The engine exhibited fire damage signatures near the carburettor. The accessory gearbox had broken off from the housing assembly. The left three-bladed propeller had broken off from the gearbox and was located at grid station K25. The first blade was missing, the second blade had disintegrated from the mid-section,while the third blade was fairly intact. The damage observed on the propeller blades was indicative of damage caused during the running of a low-powered engine. The exhaust tubes were still secured on the centre section of the wing. The exhaust exhibited damage caused by an in-flight fire and had signatures of overheating and discolouration.
That doesn’t really tell us a lot other than confirm what the passengers had already reported: the engine was on fire but still running when the aircraft crashed.
But when the investigators disassembled the engine, they found something much more damning: the pistons in cylinder 7 and 13 were damaged and in a way that had nothing to do with the impact.
Now I’m not going to pretend I know the intricacies of how an engine works (although I know there’s a few people in the comments who could break this down!) but the report makes it clear that for efficient operation, there are three requirements.
- The piston rings must be in good condition in order to provide maximum sealing during the stroke of the piston. There shouldn’t be any leakage between the piston and the walls of the combustion chamber.
- The intake and exhaust valves need to close tightly so that there is no loss of compression.
- The opening and closing of the valves must be such that the highest efficiency is obtained when the engine is operating its normal-rated RPM.
The compression rings and oil ring packs of piston No 13 were clearly damaged, with hard carbon deposits which had built up over time on the piston crown. This damage showed that cylinder No 13 had not been operating efficiently for some time, which had reduced the power available to the left engine.
More importantly, though, the exhaust valve head of piston no 7 had fractured before the impact. This must have happened during the take-off roll. The fracture caused a backfire and led to the left engine manifold pressure dropping, which the Pilot Monitoring had specifically called out shortly before he called V1.
This backfire would have caused the initial flames that the passengers in the cabin reported at the engine cowl flaps above the cowling.
Having learned what they could from the engine, the investigation’s attention shifted to the Convair’s maintenance, which had been done by a local Aircraft Maintenance Organisation (AMO). And there, they found a rather convoluted story.
The Aircraft Maintenance Organisation had been audited by the South African Civil Aviation Authority (SACAA) the year before the accident, in September 2017, specifically to determine whether they could be authorised to offer full maintenance to Convair 340/440 aircraft.
The SACAA did not authorise the maintenance company for full maintenance for the aircraft, as the company did not have adequate facilities for the maintenance activities and it did not have a training programme for certifying personnel on the Convair 340/440, which was required. The maintenance company was rated for category B (structural repairs and spray painting) only.
The maintenance company then submitted paperwork to show that a hangar agreement was in place which would extend their facilities. They also submitted paperwork to show that they had a licensed maintenance engineer and that he was the only maintenance engineer in South Africa licensed for the Convair 340/440 aircraft. There was literally no one else in the country who could do it.
The SACAA accepted this new submission and issued the maintenance company with A and C ratings, which allowed them to offer line maintenance on the Convair 340/440 aircraft.
On the 18th of June, less than a month before the accident, the maintenance company sent another letter to the SACAA requesting approval to be upgraded to full maintenance on the Convair 340/440. They attached the dimensions of the hangar as evidence that they could accommodate the aircraft for maintenance. On the 22nd of June, the maintenance company was granted one-time authorisation to carry out full maintenance and release of the Convair 340/440 to service, even though the maintenance company had still not addressed the issue with the training programme which had been highlighted in the original audit. This one-off authorisation did not include a validity period.
So the Aircraft Maintenance Organisation, having failed the audit, was subsequently authorised to do full maintenance on the Convair 340/440 as a one-time thing, which seems to have been swayed by the logic that there was no one else that could do it.
The thing is, the maintenance company still had only the one maintenance engineer for the Convair 340/440, the only person in the country who was licensed to do the work. This means that there was no one available to inspect the maintenance once it was done.
The engineer who performed the maintenance was also in charge of inspecting his own work.
The AMO had limited resources in respect of maintenance personnel to properly maintain the Convair 340/440 aircraft as it only had one licensed Aircraft Maintenance Engineer (LAME) who was responsible for the full maintenance including the last A, B and C maintenance checks carried out on the aircraft four days prior to the accident flight. Therefore, all duplicate inspection task(s) required in terms of Civil Aviation Regulation Part 43.04.8 were carried out by the same licensed AME according to the maintenance records provided to the investigation team. There were no records which suggested that the organisation had contracted an AME who was rated on a Convair 340/440 or any person meeting the requirements of CAR Part 43.04.8 for the purpose of duplicate inspections. This was in contravention of Civil Aviation Regulation Part 43.04.8.
This was, of course, that same engineer in the jump seat on the flight.
The damage to the cylinders showed that the left engine had not been operating efficiently for some time. How was it that he hadn’t noticed?
In February, five months before the accident, the left engine manifold pressure gauge had been reported as defective. At the time, the maintenance engineer removed the gauge, cleaned it up, signed it off and refitted it to the aircraft.
Then on the 5th of May, just over two months before the accident, the same defect was reported again. Again, the engineer removed the pressure gauge, cleaned it, signed it off and refitted it to the aircraft. Job done.
There was nothing wrong with the gauge. The left engine manifold pressure was low because of the damage to cylinders No 7 and No 13.
With no one else to question his decision or check his work, the engineer misdiagnosed the problem, not once but twice, both times removing the gauge and cleaning it before installing it again. It’s blatant that he never looked at the cylinders, where the carbon deposits would have made it clear. He also obviously never did the required compression tests, which would have shown the damage to the compression rings. A simple short cut that usually didn’t much matter, as long as everything was OK. A lot like not bothering with the checklist when you think you know what to do anyway.
The Convair 340/440 continued to fly with the damaged cylinders.
That day, as the Convair engines worked to provide the thrust to reach take-off speed, the faulty No 7 cylinder exhaust valve head fractured under the pressure.
Once that exhaust valve failed, the cooling fins overheated. The heat burned a hole through the housing, leading to the flames that the passengers saw licking the exhaust and the cowling.
With the engine at high RPM, the fire quickly began to weaken the front wing spar. The mounting rivets melted in the heat, releasing the aileron pulley attachments. The pulleys fell apart in the wing, slacking the aileron control cables, forcing the aircraft to roll to the left and then crash.
This is a clear illustraion of the “holes in the cheese lining up”, a quick way of saying that every opportunity to avoid this crash had failed.
If the carbon build up and the damage to the cylinders had been recognised, the cylinders could easily have been replaced. If the maintenance and inspections had been done according to regulation, then the misdiagnosis of the pressure gauge might have been caught. If the crew had aborted take-off at 50 knots, when the manifold pressure dropped, the fire would have been easily dealt with on the ground. If they had followed their checklist and secured the left engine immediately, the fire may have been arrested before the aileron control cables had failed.
Instead, the SACAA approved a maintenance organisation to offer maintenance on the Convair 340/440 even though there was no one available to inspect the work done. The Convair was repaired by the only licensed engineer in the country who twice misdiagnosed the fault. The aircraft was piloted by two pilots who hadn’t bother to apply for the validation to fly the Convair 340/440 in South Africa; however they were experienced and realistically, they were both perfectly qualified to fly the aircraft, the missing rating was simply a paperwork issue.
That’s not to dismiss their mistakes: they didn’t abort the take-off to find out what was wrong with the left engine manifold pressure and they didn’t prioritise extinguishing the fire once it was clear that the left engine was burning. But let’s also remember that they didn’t cause the engine failure and fire in the first place.
In my opinion, the report is a bit spotty and not very good at offering a linear version of events but it does try to break down these issues rather than focus slowly on the final seconds of the flight.
During take-off, the left engine caught fire and the crew continued with the flight without securing the left engine as prescribed in the aircraft flight manual (AFM). The crew declared an emergency and attempted to return to the aerodrome, however, they lost control of the aircraft and collided with power lines prior to crashing into a factory building.
This is an executive summary of the final sequence of events. I think what they are trying to say here is that the probable cause was a fire in the left engine.
3.4.1 Pre-existing damage to the cylinder No 13 piston and ring pack deformation and, most probably, the cylinder No 7’s fractured exhaust valve head that were not detected during maintenance of the aircraft.
3.4.2 Substandard maintenance for failing to conduct compression tests on all cylinders during the scheduled maintenance prior to the accident.
3.4.3 Misdiagnosis of the left engine manifold pressure defect as it was reported twice prior to the accident.
3.4.4 The crew not aborting take-off at 50kts prior to reaching V1; manifold pressure fluctuation was observed by the crew at 50kts and that should have resulted in an aborted take-off.
3.4.6 Lack of crew resource management; this was evident as the crew ignored using the emergency checklist to respond to the in-flight left engine fire.
3.4.7 Lack of recency training for both the PF and PM, as well as the LAME.
3.4.8 Non-compliance to Civil Aviation Regulations by both the crew and the maintenance organisation.
The report concludes with five recommendations, which pretty much come down to that the aviation authority needs to pay more attention. They recommend that the other aircraft maintained by the maintenance organisation should all be checked, which is sensible as their systems for checks and balances clearly is out of whack. But also, that the aviation authority needs to look at the effectiveness of its Part 91 operations, as it is the aviation authority, after all, who relented and agreed for the maintenance organisation to take responsibility for the Convair’s maintenance, although they clearly didn’t have the staff or the processes in place.
The engineer clearly made a mistake, one that he paid for with his life. When it comes to it, though, the lack of oversight absolutely was a contributing factor, allowing all of those involved to get away with a lax and unregulated approach to a passenger flight which should have been a wonderful day out in a historic aircraft.