Convair 340 (C-131D) ZS-BRV crash Pretoria, South Africa

[Old Fella]

Originally Posted by garpal gumnut

Without being too simplistic, is it possible that this aircraft was past it's use by date? No matter what inspection or maintenance showed or was performed some other mechanical gremlin was about to surface. As it did in this case. A fubar.

Yes gg, it is too simplistic. With proper maintenance there is absolutely no reason this aircraft should not have been flown. It was a relatively low hour airframe.

[atakacs]

Originally Posted by SLFinAZ

That had to be the PIC's responsibility and it needed to be done quickly with whatever input he required from the FO or LAME. The lack of a quick and concise summery and course of action from the PIC is perplexing to say the least....

They might or might not have briefed their contingency plans but from the "uber CVR" we had here it seems that execution was lacking, to say the least.
If what is described in the report is a fair representation I'd say we are faced with very poor airmanship. I don't know if that flight could have been saved (seems they had a lot going against them) but it would appear that amazing luck was their only redeeming quality.

[Dan_Brown]

Has carbinmonoxside been ruled out, as a contributory factor. Unlikely I know, insidious and no warning whatsoever of course.

[The Old Fat One]

I do hope that accident report (and the full one to come) gets the widest possible circulation wherever aviation takes place.

Once more, so many lessons...for those that want to learn.

[Chris Scott]

Thanks for the superb cockpit photos, megan, although an even closer shot of the engine instruments would be even better. I'm going to try and describe them in case it helps anyone and, perhaps, prompts a discussion on the problem of failure identification.

As expected, on the L/H side of the centre panel, the design and layout of the primary engine instruments - typical of the period - is an ergonomic nightmare.
There seems to be one BMEP gauge per engine (a parameter with which I',m not familiar), side by side.
Underneath the left BMEP gauge is a single manifold-pressure gauge, employing two concentric needles; one for each engine.
Underneath the right BMEP gauge is a single RPM gauge, employing two concentric needles; one for each engine.
All these gauges would be difficult to read from the co-pilot's seat, and subject to parallax error (particularly on the twin-needle gauges).

Further right on the centre panel, the gauges for the other engine parameters are of a similar size. Being for piston engines, they cover a far greater range of parameters than pilots have to deal with on modern turbine engines. Although a gauge - such as the one for cylinder-head temperature - covers both engines on one gauge, the twin needles are in most cases separate - typical of the period. This reduces the sweep of the needle, but reduces the chance of misidentifying which engine is which.

[jimjim1]

Originally Posted by A Squared

I noticed a a few discrepancies like that in the report. It stated the propellers were Composite, which they are not, they're aluminum. The position of hte accident was given as S26°67.031" E028°28.461" which is just gibberish, not a valid set of coordinates. The actual position of the crash was about S25 40' E28 17'

Re - Composite. As I recall the preliminary report goes on to detail that the hub was one particular alloy and the blades a different one. I think that they mean composite as more than one piece or material. Perhaps an older use of the word but perhaps one time common in aviation to differentiate from a one piece propeller.

Re - co-ordinates.
A typo here. Supposed to be decimal degrees I guess. The number are all correct, someone stuck in the degrees symbol and moved the decimal point. Degrees Minutes.DecimalMinutes is quite a common way to write coordinates. It was used in a sailing course that I was on recently.

S26.67031 E028.28461

Notice 67.031 is very close to 40/60 x 100 and
28.461 is very close to 17/60 x 100

All very confusing.

In the UK the best maps (not aviation or marine) use a local co-ordinate system that dates from 1936. https://en.wikipedia.org/wiki/Ordnan..._National_Grid. Must have seemed a good idea at the time. The maps were created for military purposes originally. Google Maps does not understand it at all:-) The maps though are very good.

[TWT]

jimjim1, those co-ordinates are also incorrect ( for what's it's worth).

The crash occurred at the Moorddrift Dairy Plant (long green roof building), see here :

https://goo.gl/maps/JWntJvbdipC2

[GordonR_Cape]

The coordinates given by avherald.com are S25.6711 E28.2838, which can easily be confirmed in Google Maps.

[aterpster]

Originally Posted by jimjim1

In the UK the best maps (not aviation or marine) use a local co-ordinate system that dates from 1936. https://en.wikipedia.org/wiki/Ordnan..._National_Grid. Must have seemed a good idea at the time. The maps were created for military purposes originally. Google Maps does not understand it at all:-) The maps though are very good.

And, not WGS84 compliant.

[A Squared]

Originally Posted by Chris Scott

There seems to be one BMEP gauge per engine (a parameter with which I',m not familiar), side by side.

Chris, the BMEP gauges are simply torquemeters. For whatever reason Pratt&Whitney decided to lable them as psi of Brake Mean Effective Pressure, an abstract theoretical quantity representing the average pressure in the cylinders during the power stroke. Ultimately I guess it doesn’t matter whether ithe units are psi bmep or inch-pounds, you learn the numbers and apply them in operation. Thats why i said earlier that it would be interesting to know what the bmep gauge for the left engine showed, bmep is the best indication of how much power is being developed, whether it was producing more thrust or more drag.

[A Squared]

Originally Posted by jimjim1

Re - co-ordinates.
A typo here. Supposed to be decimal degrees I guess. The number are all correct, someone stuck in the degrees symbol and moved the decimal point. Degrees Minutes.DecimalMinutes is quite a common way to write coordinates. It was used in a sailing course that I was on recently.

S26.67031 E028.28461

Notice 67.031 is very close to 40/60 x 100 and
28.461 is very close to 17/60 x 100

All very confusing.

Jimjim, good catch. I know about decimal degrees, but from the format of the coordinates it didn’t occur to me that’s what was intended. It’s worth noting though that S26.67 degrees is about 60 NM south of the crash site.

[Chris Scott]

Originally Posted by A Squared

Chris, the BMEP gauges are simply torquemeters. For whatever reason Pratt&Whitney decided to lable them as psi of Brake Mean Effective Pressure, an abstract theoretical quantity representing the average pressure in the cylinders during the power stroke. Ultimately I guess it doesn’t matter whether ithe units are psi bmep or inch-pounds, you learn the numbers and apply them in operation. Thats why i said earlier that it would be interesting to know what the bmep gauge for the left engine showed, bmep is the best indication of how much power is being developed, whether it was producing more thrust or more drag

Yes, as you say, the precise meaning of the numbers on a torquemeter are of little significance, provided the crew can confirm that they are within the normal range for the phase of flight.

I imagine the auto-feathering system would look at BMEP and throttle angle? On take-off, with the throttle at or somewhere near the gate, a critical loss of BMEP would presumably activate auto-feathering. But we know the captain commented that the auto-feathering light was not working [report, 1.1.12]. I'm unable to identify the light he was referring to on the photo posted by megan, but would there not be a separate one for each engine? Maybe the auto-feathering system was inoperative, at least on the No.1 (L/H) engine.

[atakacs]

A slight thread drift but in light of the potentially missing certification of one of the pilots: how would someone get qualified on a Convair 340/440 in 2018 ?

[Mach E Avelli]

Some people here are too quick to suggest poor airmanship. We do not know how thorough the pre take off briefing was, but given the background of the two pilots I expect it covered the usual engine failure actions. The stiffness of the rudder could have been due to some binding or friction in the aileron circuit (interconnect on this airframe) but given the age of this old beast perhaps not unusual in itself. Whether or not the autofeather light was working should not have interfered with the system. If the system tested OK they probably considered a faulty light acceptable.
But for autofeather to work the engine must lose power with the throttle near wide open. Autofeather does not detect fire, and fire does not necessarily cause a loss of power. Cross wiring of fire warning circuits, failure of circuits to detect a fire and false warnings are not unknown. I can attest to an engine fire in which the circuit completely failed within seconds of the fire starting, leading us initially to assume a false warning.
They appear to have had a combination of control degradation tending to turn the aircraft one way and fire on the opposite side. That would take some analysis which would not be easy while simultaneously trying to maintain controlled flight in an aircraft that neither had flown extensively. Most of us would struggle with that combination on an aircraft with which we were intimately familiar. I would like to spring this nightmare scenario as an unrehearsed exercise on keyboard experts who cry "poor airmanship" or "lack of command decision making" in the simulator to see what they would do.
While the Flight Manual says the first action of the engine failure/fire drill is to feather, given the control problem it is understandable (to me at least) that they would be reluctant to do this.
Most jet pilots would probably retard the throttle on the side that they assumed was on fire, as this is a fairly universal initial procedure on jets.. Do that in a large prop driven aircraft and not only will autofeather be inhibited, but it won't fly very well, if at all.
They did the best that they could to get it on the ground and were just unlucky because of the obstacles..

.

[roundsounds]

Originally Posted by atakacs

A slight thread drift but in light of the potentially missing certification of one of the pilots: how would someone get qualified on a Convair 340/440 in 2018 ?

The same way these guys did. They obtained a certificate of validation in SA, then did a type rating with an approved pilot / instructor.

[A Squared]

Originally Posted by Chris Scott

I imagine the auto-feathering system would look at BMEP and throttle angle?

Probably, no direct Convair knowledge but on the DC-6, which had the same engines and same props, the Autofeather actuated when BMEP dropped below a certain reading for more than one second with the throttles pushed up. It's pretty likely the Convair had the same system.

[MarkerInbound]

Originally Posted by Chris Scott

I imagine the auto-feathering system would look at BMEP and throttle angle? On take-off, with the throttle at or somewhere near the gate, a critical loss of BMEP would presumably activate auto-feathering. But we know the captain commented that the auto-feathering light was not working [report, 1.1.12]. I'm unable to identify the light he was referring to on the photo posted by megan, but would there not be a separate one for each engine? Maybe the auto-feathering system was inoperative, at least on the No.1 (L/H) engine.

The autofeather system is armed by raising the guard and then the switch under it found at the center of the pedestal aft of the prop controls. There is one green light to the left of the switch that indicates the system is armed. When the throttles are advanced to a point that equates to ~45 inches MAP and the BMEP drops below ~70 for more than one second a solenoid pulls the feather button in and the prop feathers as if the feather button was pushed. The system also then disarms to prevent autofeathering the other prop.

[A Squared]

Originally Posted by atakacs

A slight thread drift but in light of the potentially missing certification of one of the pilots: how would someone get qualified on a Convair 340/440 in 2018 ?

Same way get any type rating when there's no Simulator: Get some ground training. Get some flight training in in a CV 340 or 440. Take a check ride in a CV 340 or 440 with an examiner who has the authorization to issue type ratings.

Finding the airplane and the examiner might be problematic. I'm not sure if there are any recip convairs still operating commercially in the US, I believe the Commemorative Air Force operates a C-131. I would imagine that someone within that organization is a FAA designee for CV240/340/440 type ratings.

Originally Posted by roundsounds

The same way these guys did. They obtained a certificate of validation in SA, then did a type rating with an approved pilot / instructor.

Well, actually, one of them *didn't*

[Capt Fathom]

maybe tied to a lack of familiarity with flying a big piston, versus a modern jet.
Just because you're a superdooper jet jockey doesn't mean you can fly a vintage airliner

I don't think they were totally clueless!

From an earlier post:
Both have been heavily involved for a number of years with the Historical Aircraft Restoration Society (HARS) south of Sydney. Together, the aircraft they have flown are the DC3, Caribou, Convair 440, Lockheed Super Constellation, Lockheed Neptune, PBY Catalina plus others.

I would be surprised if the copilot had been flying a Convair for the last 2 years if not rated. Plus the upcoming ferry flight to Holland. Just doesn't make sense!

[A Squared]

Originally Posted by Capt Fathom

I would be surprised if the copilot had been flying a Convair for the last 2 years if not rated. Plus the upcoming ferry flight to Holland. Just doesn't make sense!

The FO had a valid Australian Air Transport, Commercial and Private Pilot Licence, however he was not type rated on the aircraft (Convair 340/440).

So, what's your working theory here? That the South African Civil Aviation Authority is clueless when it comes to determining which type ratings a pilot holds?