Thanks Lano for that explanation about the Yaw data. I’ve used MagFit on my helis and had intended to use it on this model too.
I’m now making the changes to the parameters file that have been suggested in this thread, and will now be printing a replacement model to test it on.
Once again, thanks for all the inputs.
It’s happened again!
Yesterday after taking off in manual mode and flying several anti-clockwise circuits in autotune mode, without any issues, the plane suddenly banked and headed vertically into the ground from about 60ft. This time I think the log is intact, for it shows massive variance between desired and actual roll and pitch at the end of the flight. But what should I be looking for to tell me what has caused the failure?
The log and parameter files are below: Logging frequency has been reduced, as suggested in earlier posts, so the file easily fits in memory. The model is as before (with previous damage reprinted) and the electronics are as before – FrSky RX4R receiver and Matek 405WMN FC.
Model B after fence type changed to altitude 120m.param (18.9 KB)
It looks very likely to be a tip stall. You have applied less and less throttle (at the end 17%) and at the same time pulled when initiating the turn.
Rolf
I agree the failure mode looked like a very abrupt tip stall, but I find that difficult to believe because (a) the model is a ‘trainer’ with straight Cub-like wings, (b) I was flying sedate circuits at a relatively constant throttle and (c) I’ve encountered tip stall in other models with tapered wings but it’s never been so abrupt as this was.
In the graph I can see that throttle input is decreasing, though I wasn’t conscious of the model slowing down (and it was close enough to me at that point to see if it was slower than normal); but DesRoll diverges suddenly from Roll and after that my reaction was then to shut the throttle.
So, as far as you can see from the log there was no indication of a failure of receiver, flight controller or GPS/compass? If that’s the case I’ll have to suspect pilot or aerodynamics failure 
You weren’t aware of this because the tailwind at that side of the “circles” made the aeroplane fly faster for an observer on the ground.
Rolf
Possible, though the wind was near-zero and I thought I was too experienced to be fooled by a tail wind 
So, given the evidence of the log, would you be happy to re-use the electronics in another model?
Sorry, the groundspeed change (approximately the same altitude in the last ‘lap’ )
doesn’t look like almost no wind, certainly not in relation to the slow-flying aircraft:
Rolf
Thanks for taking time to look into it Rolf. I see what you mean from the graph, but it certainly didn’t feel like that on the ground. I know the trees in the bottom right of the picture can create different air patterns (I’ve been flying there 30+ years) so maybe that’s a factor.
