Just carried out my first autotune on my Extended F550 (680mm motor-motor) hexacopter. Completed in just over 10mins and I must say the copter feels a lot more dialiled in now. Just got home to hook it upto MP, and to my surprise, the Stab P values are both higher than 12 (I had them at 4.2. before). Just wondering if that seems normal for my build (I realise every build is different, but a general ball-park would be good if possible).
This is my build:
F550 frame with extended arms to 680mm.
Carbon fibre battery tray and landing gear slung under the copter holding 2 x 5800mah 3s Zippy Compact 25c LiPos wired in parallel (960g)
Turnigy Multistar 2213-935kv motors, spinning 11" carbon props (well balanced and very stiff)
Hobbywing X-Rotor 20A ESCs
and these are my autotune results:
It certainly feels pretty stable and responsive . Much more decisive quick stops too - used to wobble quite a bit when stopping forward motion quickly. Do these look acceptable?
Thanks, Paul
Replies
OK,I just did an autotune on my brand new build Tarot 680Pro, again with very strange results.
Stab Roll/Pitch P values are over 14 (despite 12 being the max value for these). I know the image shows the Stab Pitch P as 7 but that appears to be a bug - on the field, after the autotune, I checked these values and both Roll and Pitch were both showing 14.625. I get home connect up the copter to Mission planner, and it shows as below - very weird!
For the rate PID values these differ vastly between Roll and Pitch, as you can see below:
My copter does have 2 batteries slung under it on the camera rails, these run sideways, one forward and one rear, so there is more weight distributed front to back than there is side to side (hopefully that makes sense). So wondering if this is why there's a difference in Pitch and Roll gains. Here's a pic:
Me, after autotune, changed stab roll and pitch around 7 and works well, you have to test yours until you feel well. (I have less Kv motors)
I think what's weird about the results of the AT (and I have tried this on another hexa with similar results) is that it sets the P and I values of roll axis the same as one another (and the same for the P and I of pitch). Its highly unlikely that a proper tune would end up with these values being the same, as far as I know.
I just did an autotune on my new hexa which is a 4s Tarot 680Pro running Sunnysky 690Kv motors (just over 3Kg), and the values it chose for Pitch Rate P and I were above the maximum values for these fields in Mission Planner - it set them to 0.3499999 (max being 0.30), whilst the values set to the roll axis were only 0.03 - very weird! Although I'm not completely certain that the AT finished. It did the roll axis first then moved onto the pitch, but I had to reposition it quite a few times due to a slight breeze, and the last time I did this, on releasing the stick, it didn't seem to continue with the AT - as if it had finished, but I didn't hear the spoken message on my radio this time (which usually happens upon completion) - as I have Taranis with Teensy 3.1 (running Wolkstein code from here https://github.com/wolkstein/MavLink_FrSkySPort/blob/s-c-l-v-rc/REA...) converting to FrSky telemetry, so get these voice prompts. So it possibly chose to abort the AT before completion. Need to try on a calmer day.
For what it's worth : I get similar results - often STB_RLL_P of 16 , same for STB_PIT_P - reasonable setting is about 4.
rate tuning results seems to work well.
To be honest, I know what PID loops are theoretically, but wondered why 2 sets of P values in APM - one for Stabilize and one for Rates. Any ideas how these work together? Is it that in all modes except rates mode these 2 sets of P values are added together, and in rates mode only the Rates values are used (so Stabilise P values ignored)?
Andre, Once you get these high values, do you tend to manually back them off down to around 4?