Steve6443 wrote:
Strangely enough that is the same thesis about some other crashes, including 737s or similar – that the a/p hands you a plane which has been trimmed to such an extent, pilots have become spatially disoriented as the plane suddenly heads in a direction they weren’t expecting it, nor did they understand why……
There was one such accident in Russia where a crew totally lost the plot in a 737-500 during a go around. The 737 has quite a pitch up movement due to the position of the engines. They opened the throttles manually without pressing the TOGA button, leaving the Flight Director in APP mode but thought they were flying a coupled go around. Eventually they found that they were not and followed the flight director (which still tried chasing the GS) straight into the ground.
Under normal circumstances, most AP’s will involve autotrim and actually never should hand over the plane out of trim. This can of course change if the mode you think it is in is different from the one it is actually in and, particularly, if you try to overpower the AP.
One bit I’ve really learnt while dealing with autoflight is that if the question “what is it doing now” rises it’s ugly head, I’ve already lost the plot to an extent where a situation can escalate very fast indeed. The standard procedure for this (disconnect the AP and get the airplane levelled out with wings level) will not always lead to something nice either as the AP may very well be doing something quite reasonable while the pilot is way behind the airplane and has no idea what really should be done next.
Timothy wrote:
Similarly, all those stories of autopilots masking engine power loss, through being in VS rather than IAS mode, resulting in spins in IMC, typically in Bonanza.
Full agreement here. In fact, GA Autopilots are often more complex to handle and have much more inherent dangers to those who do not go through the trouble of understanding them fully, than current airliner AP’s. Which sometimes leads to the grotesque situation that an airline pilot used to his Airbus or Boeing DFGS will fail miserably running a simple S-TEC 55×.
Operations which are dead simple in an Airbus can involve up to 7 steps with a 55x, such as altitude change or getting it to follow a nav signal properly.
Most GA AP’s need quite a bit of training to use them properly. Most airliner AP’s are quite straightforward but of course also can catch you out, as the example of the A320 at Strassbourg showed, where instead of a flight path angle a VS was selected which flew the plane into the ground. But generally of course, airline pilots also fly much more and know their AP’s inside out, whereas most GA pilots need to rethink how to use it every time they fly.
Timothy wrote:
The other thing is that lots of power, whether in aircraft or cars, creates its own controllability problems (torque in the TBM and EFATO in the King Air.)
I used to fly twins and love them. But I would not in my current flying want to own one for the very fact that i would not have the necessary currency to deal with an EFATO or even one in cruise or worse descent or go around (which causes a lot of grief when people open up both throttles but only one responds). That requires constant training and vigilance.
And talking of torque, try a Lancair 4P… that particular airplane scared the living daylights out of me the only time I tried it. Take off involves very slow opening of the throttle as the airplane will not track straight on a runway with the engine at full power…. I shudder to think how that beast behaves in a go around from low speed.
Yes, complexity is a huge factor. If you consider that the original Malibu 310P proved too much for most pilots who flew it only when it came to engine management, it doesn’t bother thinking about pressurisation and other stuff which comes with it.
@Steve6443, opinions vary as to the exact sequence, but my point is that he was flying an aircraft with systems more sophisticated than he could cope with, and the result was tragic.
Timothy wrote:
The Dunkeswell tragedy (family of four killed) appears at first glance to be loss of control in IMC (stall, incipient spin, out of the clouds into the ground) but seems to have been caused by the pilot being unable to cope with the complexity of the PA46’s autopilot, which may have put him in IMC in the first place.
My take was that the accident was caused by the fact that he’d flown an approach with the autopilot still active, meaning he was pushing the nose down to descend whilst the a/p was still active and trying to hold altitude hence it was actively trimming the plane nose up. Once he did disconnect the autopilot properly, he was handed over an aircraft which was suddenly trimmed full nose up; once he released pressure from the yoke, the aircraft climbed sharply and that put him into IMC, leading to spatial disorientation……
Strangely enough that is the same thesis about some other crashes, including 737s or similar – that the a/p hands you a plane which has been trimmed to such an extent, pilots have become spatially disoriented as the plane suddenly heads in a direction they weren’t expecting it, nor did they understand why……
Timothy wrote:
I think that you are not sufficiently factoring complexity into the sequence that ends with loss of control.
You are probably right. One has to look into the details of the investigation reports to find these things, sometimes even read what is between the lines, not only look at the basic statistics.
Peter wrote:
In airlines it is very different – crashes like AF447 are directly due to not understanding what the boxes show and where it comes from.
Arguably that accident exposed a lack of the basics of stick and rudder flying, as to fail to recover from a stall even though there was plenty of height shows some serious lack of feeling for keeping the air flowing over the wings .
Mooney_Driver wrote:
Most people I deal with finding airplanes for them are looking in the <100k segment. These are people who have learnt to fly the hard way, saving for their next lesson and afterwards saving hard to get the best bang for the few bucks they have. These are not people who do their PPL out of the petty cash and order their first plane new before they even start.
Doing your PPL/IR e.t.c. on a tight budget and then finding a plane you can use in a long process is different than that scenario, a lot different. It allows more time to learn and to get things right, it also will not involve “new” models where few experience is available. in my experience this excludes several factors which have led to the Bonanza and Malibu disasters. With Cirrus, it is the shute which has saved several such people from becoming a statistic but did not prevent others.
Good explanation. I agree 100% on that.
LeSving,
Although I agree with the general thrust of your comments, I think that you are not sufficiently factoring complexity into the sequence that ends with loss of control.
The Dunkeswell tragedy (family of four killed) appears at first glance to be loss of control in IMC (stall, incipient spin, out of the clouds into the ground) but seems to have been caused by the pilot being unable to cope with the complexity of the PA46’s autopilot, which may have put him in IMC in the first place.
Similarly, all those stories of autopilots masking engine power loss, through being in VS rather than IAS mode, resulting in spins in IMC, typically in Bonanza.
Neither of these things is likely to have happened in a C172 with a wing leveller (not least because it would probably wobble its way out of a spin without pro-spin controls anyway.)
The other thing is that lots of power, whether in aircraft or cars, creates its own controllability problems (torque in the TBM and EFATO in the King Air.)
So I would not lay it all at the door of poor pilot basic skills. Complexity plays a big part too.
Flying in a too complex aircraft doesn’t seem to enter the statistics
From having flown with many people, and talked with many more, I think the reason that a lack of aircraft systems knowledge doesn’t cause many crashes is because the pilots avoid using the parts of the kit they don’t understand.
Many pilots avoid using their autopilot because it does funny things. Maybe it needs to be fixed (which is itself a major problem since the number of avionics shops in Europe that can really diagnose and fix autopilots can be counted on 2 fingers) or they don’t understand [all the] the modes.
I recall one syndicate SR22 which was flown mostly with an Ipad running SD.
Very few PPL-level instructors are of much help with this. I never found any who knew an HSI, let alone the KLN94. This was in 2002.
It’s called “risk compensation” 
People compensate in terms of both equipment usage and the mission profile. And it makes gathering and understanding data really hard.
In airlines it is very different – crashes like AF447 are directly due to not understanding what the boxes show and where it comes from.
