How reliable are "glass" avionics

Nowhere in any statistics do software failures feature prominently.

The various GA populations and their intersections make research hard.

Taking US + worldwide GA, only a few % have an IR and actually use it. Those who don’t fly IFR are not likely to even notice a subtle avionics issue, let alone bother to report it.

The % of G1000 (etc) planes which actually fly places (i.e. non FTO stuff) is small.

The % of pilots who have detailed avionics systems knowledge is very small.

The % of pilots who understand engine management (and know the meaning of EGTs CHTs etc) is still today only a few % of pilots.

Then you have good olde Europe where most pilots are not comfortable speaking English and are hardly going to report an issue to the manufacturer (same goes for flight planning software – the principal players don’t get tested and bug-reported from beyond the UK and a few other countries, and you won’t get many reports from Spain France or Germany).

I have correspondence from an aircraft dealer in the USA who some years ago correlated a number of US CFITs with defective terrain depiction in the KMD550/850 and with a similarly defective depiction in Jepp Flitestar/Jeppview. He had no proof, not least because everybody involved was dead.

It’s true that the bulk of PPL-level fatals are simple stuff like stalls and spins on base to final turn, or CFITs in circumstances where the “V” in “VFR” was purely wishful thinking, but that doesn’t mean there isn’t stuff going on which kills people and we don’t find out – because there are no survivors. Also CFITs are “so obviously” pilot error that nobody investigates the “how”.

Also, autopilot failures can, at the right moment, jack up the workload massively, and a lot of people do rely on the AP a lot. I know pilots of certain types who tell me they fly on the AP practically the whole time. I use the AP all the time too but hopefully get enough hand flying practice on my weekly local Class G messing about.

Even the two above were clear pilot error even though they may have been preceded by some mechanical failure.

That’s true for almost everything. A perfect pilot could save almost anything.

I should have added one more factor: where one flies to. For example I had several autopilot failures over a point in France, over a number of years. Documented in my linked writeup above. So you could easily find another owner of the same type whose AP never failed… this is susceptibility to EMI.

You must be able to fly by compass, dead reckoning, and by backup instruments

IFR?

Of course the G1000 is reliable. That and the G500 have topped the informal US surveys by a factor of 2x or 3x

But that’s a different point to be making from saying that it never fails, or that failures have never been heard of. I gave reasons for why a lot of failures can go unreported. And let me give one more really BIG one while I am at it: the vast majority of aircraft owners do not talk about problems, especially on the internet – because they want to sell their plane at some stage, and they don’t want the prospective buyer to find the posting. Only the other day I read of such an event…

While not everybody has detailled knowledge down to soldering level like you, Peter, that doesn’t mean none of them know their avionics and engine management.

“Soldering level” is a completely daft analogy. I don’t know any of this stuff to “soldering level”. But if I owned say a G1000 plane I would want to learn all the features and which features are dependent on what – see this thread which is more appropriate to this debate anyway. And I know for a FACT that many pilots are not that familiar with the kit.

And the “fly places” part is completely wrong. Most of these planes ONLY “fly places”.

The TPs and jets, yes, obviously. At €500-1000/hr DOC you aren’t going to be going around the Isle of Wight. One almost cannot get decent photos through the windows. But why is it that when I fly somewhere off the beaten track, even just to Mali Losinj (which for Germans is what Le Touquet is to Brits) I see mostly “old stuff”? Example 1 2 3 4. (I could have spent hours doing those screenshots, from further out airports – like this or this. Once you get away from the burger runs, these scenes really do represent European GA.) In the US, the new stuff does fly distances but for some reason the distances flown here in Europe seem to be far less, on average. It’s clear what while some mfgs have sold a fair number of new stuff into Europe, the bulk of the fleet is still made up of the more traditional types. Not surprising – they can be bought for 1/10 to 1/5 of the money!