Servers by jDrones

Synthetic Airspeed malfunction (2nd time)

I experienced a malfunction of the airspeed reading today during a flight. It appears to be the same thing that happened last month which caused my Mini Talon to crash. This time it was my Nano Talon. I was lucky the behavior was just violent oscillation as if the plane was stalling and autopilot was trying to stabilize even though my speed was good. Now that it’s happened again, I’m wondering if this is a bug or if it something I am doing wrong in my configuration. 2 different planes, 2 different flight controllers.

Here is a video showing what happened https://youtu.be/VWV1hhhCCuU?t=90. The tlog is here https://www.dropbox.com/s/vx29mgabee3gubv/2019-10-06%2017-51-46.tlog?dl=0.

Also, previous thread with the same airspeed symptom on the Mini Talon here Crash on 3.9.9. No airspeed reading.

Any insight would be greatly appreciated. @tridge? A lot of people say no need for a digital airspeed sensor as the synthetic calculations are very good, but this has me a little concerned.

Do you have the onboard log from the microSD too?
It certainly looks like it did get the estimated airspeed way off, though I think what we really need is to detect the oscillations and back off the gain scaling.

Hi Tridge thanks for the response. Yes I do have the log from the SD. I will pull it later today and upload. The log from my mini talon crash last month is from the SD if you wanted to look at that to be ready for comparison. I’ll post the Dropbox link later today with the Nano Talon flight from yesterday. Thanks again!

1 Like

@tridge, here’s the bin file from the SD. This is the Nano Talon running Matek F405 Wing.
Arduplane 3.9.11

For quick reference, here is the other log from the Mini Talon that had the same airspeed issue and crashed:

Thanks

Thanks. I notice quite high EKF velocity innovations, I suspect the lack of compass data is impacting on the EKF yaw estimate and thus on the wind estimate.
Unfortunately the MatekF405-Wing is single IMU which makes some types of analysis (eg. looking for aliasing) not possible.
What I’m going to try to add is code to detect the oscillations and reduce the speed scaling of control surfaces. I’m not confident we can really improve the airspeed estimation any more, especially without a compass, so reducing the impact of the poor airspeed estimation seems like the best bet

@tridge thank you for looking into this. Was it the same issue with the mini talon that I had running a pixhawk 4 mini? What do you suggest as a more safe configuration? Would using a digital airspeed sensor alleviate this? My research indicates that enabling compass on a fixed wing is not necessary and just opens up the possibility of more issues. I just want to have a safe, stable system and the airspeed issue I’ve experienced twice in 1 month has me a little hesitant to fly. Thanks again for your help.

I swapped out the bn-220 for a bn-800 gps/compass. Enabled and calibrated compass and flew it for about an hour total in 10+ mph winds with 22-25mph gusts at the surface. Airspeed reading did not seem to malfunction. I did try to induce a stall which resulted in a brief wobble but nothing violent like before. It seems a compass really is necessary in fixed wing in windy conditions. Despite what most people claim.

Hi buddy
question why did you stay in fbwa do you have a manual mode setup as flight mode ?
i had an issue like that but i had input some bad pid settings when i went into fbwa i was lucky i just put in manual got home lol

I do have manual but prefer to launch in FBWA. My tune seems to be good. I have a few hours in now with the compass enabled and it’s been handling better. Going to try auto launch next.

Servers by jDrones