Hi,
I need your help to solve an issue I got twice with 4.6b4
It’s all about LAND after battery failsafe.
My copter flown away just after FS kicked in.
First time I managed switched it back to Stabilize in RTL phase. So quit fast but it was at about 10m alt. Had some time.
Second time I had not change to act. It was LANDing from just 2m maybe then got power and skyrocket to the nearby tree. Good it was there. In that case RTL was disabled so only LAND could kick-in on critical batt failure.
What may be or not important is, I used to power drone on workbench, then put it on the ground. So my arming ALT is almost always lower then power-on ALT. Or even worst I power it on table on the balcony then go to the garden 3m below.
Below is a ALTs scenerio as was removed by anonymization.
It looks like the LOG_BITMASK has been changed so that the CTUN message isn’t displayed which means we can’t easily compare vs the throttle level. Still, it seems that when the RCOU (motor level throttle levels) goes above 60% the vibrations increase dramatically.
We have some advice on the wiki about vibration isolation so hopefully that will help. It’s a frame issue though.
Thank you, will look closer to vibration but I understood values at 20-30 are good.
Right side are a tree hit vibration. Same day I made few landing tests but without batt failsafe, it went good no problems.
Anyway, will look into vibrations.
30m/s/s ~ 60m/s/s is the grey zone where there may or may not be problems
below 30m/s/s normally works
I think the average I see on vehicles is about 15m/s/s.
I definitely would not say that 20 ~ 30m/s/s of vibration is “good”
I think vibrations in the X and Y axis actually need to be much lower than in the Z-axis. So 30m/s/s in the X/Y axis is definitely not acceptable. The limit is more like 15m/s/s max in X, Y I think.
We could update the recommendations on the wiki. We have four pages on this subject
Hi Randy - that sounds fair.
I originally quoted fairly conservative values since a lot of the older accelerometers were easily affected by vibrations. New flight controllers mostly have better accelerometers now with higher limits.
Usually when checking logs I look at the flight controller type to see if the vibrations levels are appropriate, just approximately.
Guys,
Here I posted the failed Autotune graphs.
With those vibs copter do fly quite well but can’t pass autotune.
Therefore, at least for my very old build, it seems to be on the edge, as I understand the situation.
Or here are last flight vibs. Again, very old build using SimonK loaded ESCs and crash tested propellers. Once it will be safe to install better props vibs will go down for sure.
So the limits you mentions are not to low even for poor builds.
With @randyV advice I replaced just one foam. That’s it. It’s not so complicated to go down by 10-15 points.
It’s not just high vibration levels that will prevent auto tune from running. A poor base level of tune will also. Are you getting “failure to level”messages?
So an ~11" on 3S, that’s where the default PID’s originated from. Noise in the Rate controller and the tune isn’t great but to be expected on defaults even with the “nominal” craft. Y-vibes are a problem as has been noted.
Why don’t you disable GPS logging so you can post a .bin that can be reviewed for the notch filter? I would probably disable the Notches configured and go from there. Perhaps abandon the FFT notch and use a simple throttle based one (1). But that remains to be seen.
