I performed a test flight with INAV 8.0.0. The first flight went smoothly, and Loiter mode worked as expected. After landing, I swapped the propeller and took off again without disconnecting the battery between flights (I disarmed the system, changed the prop, and then armed it again). Telemetry data was recorded using INAV blackbox (with a 1/2 save rate).
As you can see in the video, I didn’t move the sticks when activating Loiter mode. I attempted to engage Loiter mode twice during the flight, but both attempts resulted in the same issue. As a result, I had to deactivate Loiter mode to avoid losing the plane.
The flight controller used was the Matek F405 VTOL, and the plane is a glider with a conventional tail. I don’t control the rudder directly, it is mixed into the roll.
GPS is M10 based and all 3 additional GPS constellations were selected. Update rate was by default set to 10 Hz. I’m going to only select Galileo constellation and increase GPS update rate up to 20 Hz. Additionally I reduced blackbox save rate to 1/8. Lets see if that will help
I reported this issue on the INAV GitHub, mentioning that I could share the flight log but only if there was a way to remove or edit the GPS coordinates. The response I received was, “then no one will take your report seriously.” It seems they didn’t understand that I was reporting an issue with their software, not with my own setup. Because of this, I closed the ticket and removed any information I had added.
UPDATE:
Today I performed another test flight. I started/landed 4 times in a row without disconnecting the battery. I performed many turns in different directions, was using different modes ACRO/MANUAL/ALTHOLD in addition to LOITER. Non of that resulted in a behavior like in the video and LOITER worked as it should every time I activated it!
What I changed:
– changed GPS UART speed from 57600 to 115200
– set blackbox save rate to 1/8 (by default it is 1/2 which is 2ms approximately and I thought that could spoil control loop timings)
– selected only one additional GPS constellation GALILEO to be able to increase GPS update rate up to 20 Hz (better get M9 based GPS to get 25Hz and all 4 constellations at the same time)
Conclusion:
I honestly don’t know what happened during that flight. Was it some kind of a software glitch or a GPS-related issue? I’m not sure. And as far as I do not know what is the source of the problem, I cannot be sure that the changes I made do really solve it. To properly diagnose it I’d need to revert the changes, make the problem reproducible and then apply the changes one by one until it is gone. But as far as INAV devs don’t wanna take it seriously anyway then I see no reason why should I bother.
However, there is one strange thing visible in the video: after landing, the gyro stops moving, which suggests that the telemetry is perfectly synchronized with the video. On the other hand, the distance to the home point keeps changing after landing, as if the GPS coordinates were somehow delayed or buffered and lagging significantly behind. Interestingly, this kind of GPS “delay” no longer occurs after the changes I mentioned earlier. Additionally, when LOITER is activated second time, the software completely looses YAW and also attitude isn’t always synchronized (or is somehow delayed) with the real horizon.
So, I believe that certain settings—or a combination of settings—could cause this strange behavior, which might even result in the loss of a plane in some flight modes like LOITER or GPS navigation.
[ad_2]
source
