Today I wanted to escape the Fog and take some pictures from a nearby Mountain. Copter is a Tarot 680 Heavy (means it has enough Engine Reserve 6x TMotor 4010/475 with 13x4.4 Props but weights about 4.7Kg).
Takeoff surface was agriculture land which was not complete level. I tried to arm the copter and just saw a short flash of a Preflight Check fail Mavlink Message on the OSD immediately followed by Black Box recording started. Tried various times but no help. I fetched my Laptop from the Car and figured that the Gyro Inconsistency Check was this time preventing Arming - interestingly BlackBox Recording started anyhow but Motors did not spin.
So I did a Gyro Tuning on this not sooo flat pice but did not touch the Vehicle. This was successful and I could arm and start to fly. During the first attempt the Vehicle almost tipped over as it was a rolled to the left on the ground. I stopped i an repositioned the Vehicle and was able to start. The Vehicle felt a bit sketchy in Stabilized so I switched to PosCtrl. Despite having about 17 Sats it was completely unstable and was not able to maintain position nor altitude (Variance by 2-3m - my estimation).
So I tried to land in Stabilized which turned out to be rather tricky. Finally I could land. I Power cycled the Copter.
I tried to start again. This time there were no Alarms and I could take off. The copter felt much better and PosCtrl Mode worked as I expected. I was flying around for about 6-7 minutes and landed in PosCtrl Mode. As I was unsure how it would behave I flipped the kill switch as soon as it was more or less stable on the ground. Landing was detected by the FC but it did not disarm itself nor was I allowed to disarm in PosCtrl Mode.
So here are my current findings/problems with 1.5.2:
- Preflight Fail Messages are not shown on MavLink Msg for OSD because immediately followed by BlackBox Message
- Still Gryo Issues. Every time I startup the Flightcontroller I have warnings about this. When I analyze I can see some minor deviations of the 3 Gyro Axes. This time it was even preventing arming. See: https://github.com/PX4/Firmware/issues/6033
- No stable flight after Gyro calibration / ok after Power Cycling
- Landing detected but no disarming in PosCtrl Mode
Log & Msg Files are available.
Thanks for the very detailed flight report. That’s always very helpful!
Could you try current master and see if that improves things? We have improved many, many things in the last two weeks including the ICM driver on Pixracer and I’m wondering if that could improve your experience.
And it would be great if you could upload the logs to log muncher (http://logs.uaventure.com) and share the link in a PM or alternatively send them to mail email address. I’d like to make sure I understand the root cause of the issues you had.
-Lorenz
@LorenzMeier here are the logs. I appreciate looking at them. For sure the first flight felt strange and I lost a bit confidence for the second, which turned our rather ok, IMHO. Would have loaded Logs immediately but there was a congestion with LogMuncher 
1st Flight: http://logs.uaventure.com/view/xNDz6n6RLf4oRMdXaA5vHZ
2nd Flight: http://logs.uaventure.com/view/7sfYD7mKesWL7LFcCHJGjN
I’m a bit amazed by the amount of Y-Vibration… probably due to reaching the Frames Limits.
Regarding current Master I was a bit skeptical using it on a live system since the Issue I reported that where the FC would not boot up completely. But I can try it again. The FC used was an AUAV-X2.
Regarding the GYRO Problem, could this be a HW related issue?
Ok, thanks! Did you fly the ground trajectory in the 2nd log? There are two distinct straight lines from your start position. Is this what you flew?
Other than that you do have some pretty heavy vibration on the frame which can induce non-zero mean offsets of the accelerometer, which in turns means the attitude won’t be level. I think you might need to consider dampening it.
Not sure what you mean… in the 1st log the two peak to the east must be wrong… I was for sure never that far east and very sure not in the later stage of flight.
Not sure which lines you mean in the 2nd Log.
About dampening I’m using 4x 1cm^1 gel Pads about 5mm think. I’m already considering adding lead to the FC Case in order to dampen it. The AUAV-X2 is just to light and there are to many cables. I think DJI made a smart damping system in the NAZA.
Its definitely too light, yes. Ok, so those two excursions in position then explain everything already. Most likely the heavy vibration meant the filter did trust the GPS more and the GPS spikes set off your attitude. @Paul_Riseborough any insights? Related: https://github.com/PX4/ecl/issues/218
Some more feedback from todays Flight with a lighter Gimbal (roughly 4Kg TOW).
I did loosen some cables from the FC to check weather this would reduce vibrations (I think there is not much improvement. I was thinking about replacing all Cables from conventional Type to silicon type and use i.e. only one Ground for all 6 ESC Connections and combine the Grounds external from the FC).
I tried to take off in PosCtrl with the not so new Start Feature. I stopped the attempt as I was afraid, that the Copter might Tip over. So I started a new flight in Stabilized mode. Flight was smooth so far with actual Master. Still some minor observations:
- Once I decided to stop the Start in PosCtrl Mode, I landed by using the Kill-Switch but was unable to Disarm. I had to switch back to Stabilized Mode to disarm. From a User Experience Point of view the Handling of Start & Landing is still far from DJI. At least this is my impression.
- Flight was basically smooth, but I noticed a new feature in my OSD (MicroMinimOSD Running the following SW: GitHub - night-ghost/minimosd-extra: Full rework of minimosd-extra with lots of news) that suddendly in the Coordinates Field I got “Stale Date” or “Stale Connection”. I do not remember having seen this before. 57600 Baud Connected to the Telemetry as well.
- I also noticed a few drops (gaps?) in the Logging of Flight Params.
Appart from this I think it was ok.
@LorenzMeier Link to Logs (PosCtrl Start & Flight) follows via PN.
