Yeah.I thought the vibes were ok-ish.Just that Z axis stuff.I’m all set for another autotune but unsure of myself doing a manual.It’s one of those subjects that just makes me glaze over whenever I read a how to.It just doesn’t stick.Maybe I need to get my hands dirty with it to get the feel.
I’ll see how the AT goes and take it from there.
Totally understand it seeming like alot at first with the PID’s. Once you get into it a bit though theres nothing to it. I think you are spot on with getting your hands dirty, I know personally when I first got into it PID made no sense at all, but once digging into it and seeing exactly what your parameter changes do things became much clearer. If you start with a low set of parameters for P, I & D and pick P term and assign it to CH6 tuning then bring up until ossicilations start then back off and continue with other parameters it goes pretty smoothly. There are warning in flight behavior that cue you to possible ossicilations and if you take it slow there is little risk in something going catastrophically wrong. D term can get bad quick, butbonce again, taking it slow in small steps negates that as well. In the end you can end up with a great flying frame that flies like “you” want it to while still providing the auto modes with what they need to function properly.
I follow this fairly straight forward procedure posted some years ago by Leonard Hall. It will take some time but it has worked for me on many multirotors. As Tim says the Live Tuning Pot is your best friend!
Take off in stabilize.
Increase Rate D until you get oscillation,
Back off Rate D until you don’t get oscillation (plus a little buffer if you like)
Increase Rate P until you get oscillation,
Back off Rate P until you don’t get oscillation (plus a little buffer if you like)
Set Rate I to be equal to Rate P
Increase Stab P until you get oscillation
Back off Stab P until you don’t get oscillation (plus a little buffer if you like)
Then try an Auto Tune if you like.
Cheers Dave.I’ll give it a go.
Just so I’m clear on this - do you land each time to change parameters or do it literally on the fly ?
Also so I don’t pancake it through ignorance.
I think he meant to assign a live control to a pot on the radio.
Yep I got that bit.But changing from rate D to rate P to rate I is what I meant.I can only use one channel for that (ch 6) so need to swap them over during the tune.I’m not clear if I can hot swap them whilst it’s in the air or if I have to land to reassign the channel each time.I’m guessing land but I’ve been wrong before.
You will have to change the tuning function each time. I asked once here about adding the ability to assign 2 channels for Live Tuning but it got no traction. Plenty of other priorities no doubt which I get!
I setup a ground station where I can easily see the screen (in the house usually), make a flight adjusting one parameter, land, disarm and then change the Tune function via telemetry, then arm, take-off and adjust that one. Repeat as long as the battery lasts!
Thanks Dave.Just what I needed to know.I did suspect it was like that but suspecting and knowing are a way apart in this game.
It’s nice and still and raining this morning.
Great to see the conversation and feedback here. We’ve reverted those filter settings in -rc9 so hopefully this doesn’t bite anyone else.
Thanks Randy.
I’ve sucessfully done an autotune after resetting that filter to 20Hz.It seemed to fly better from the off and is certainly much improved now,to the extent that I can trust myself to control it in the air.RTL worked ok too but seemed a bit sluggish on the way down.
The only continuing worry is the Z axis vibrations which seem to be a bit high and give me some clipping in free flight.It didn’t really notice in the YAW autotune (although it was there) but did when I tried the new settings out just afterwards.Not sure what I can do differently to calm those numbers down so am open to suggestions.
I may get to do another autotune today with a bit of luck and should have plenty of battery for all three axes together.I may drop down in battery capacity and weight as well to see if that makes any difference to things.I also have to get to grips with the current measurement stuff.
A work in progress but that’s better than it was before (a scary monster).Yaw AT and post AT free flight logs.
Managed another AT yesterday after dropping the set values by around 10%.It’s still got a slight wobble in POS_HOLD and vibrations seem to be quite bad,or the clipping does,but there seems to be a dearth of info about those numbers so I haven’t actually learned what these numbers actually mean.
I’m also less than impressed with barometer #1 which has some serious deviation in it,and I’m talking metres and not centimetres.Again I’m not sure what this means or if I can switch barometer preference to #2 which is a lot smoother.Maybe there’s a filter for it but that I doubt.
So still coming to terms with 3.5.8 terminology and values and with the Pixhawk 2.1 and HERE hardware but at least it’s flying ok.LAst flight was 25 minutes or there abouts with plenty of battery left.Power consumption figures still need a lot of tweaking as thery’re nowhere near right.
Jagger and others. It looks like you went through a lot to get this 2.1 with HERE to work. Does anyone know if they have possibly tweaked and/or fixed the firmware to eliminate going through what you went through. We have EXACTLY the same problems? I would like to go back to the older pixhawk but I think the HERE+ RTK doesnt’ work on that model. I guess i’m asking if there is any “new” on this subject
I heard that they’d put some new parameters in for the HERE but I ended up junking the Pixhawk 2.1 and HERE in favour of some stuff that works.So can’t help a lot.Sorry.
Hello my friends in this thread.
@LeonardTHall
@proficnc
I flew my S1000 w/Pixhawk 2.1 again today. We got it to behave "decent’ in alt hold mode and flying gentle. have it tuned decent in pitch, but ROLL is NOT good. still very sluggish and wobbles. The pitch and roll PIDs are just about the same and yet the roll behaves much different.
Pitch PIDs
ATC_ANG_PIT_P,12.34839
ATC_RAT_PIT_P,0.28
ATC_RAT_PIT_I,0.28
ATC_RAT_PIT_D,0.0057361
ATC_RAT_PIT_IMAX,0.5
Roll PIDs
ATC_ANG_RLL_P,12
ATC_RAT_RLL_P,0.28
ATC_RAT_RLL_I,0.28
ATC_RAT_RLL_D,0.005
ATC_RAT_RLL_IMAX,0.5
How can those setting above be so close and yet behave SOOO different? Pitch = snappy and nice. Roll = jello
ALSO, we changed just about every value in the Roll settings and saw ABSOLUTELY NO CHANGES in response. So, either i have somethign else to turn on/off to make a difference or the Roll PIDs are being IGNORED?
I am posting my logs today where you can see i changed the Roll PIDS and no real changes. Flew inside in EXCELLENT conditions in Alt hold because stabilize was a bit tough. The throttle in stabilize is VERY VERY sensitive and I could not get it to back off. Any advice there would be great also.
then we flew outside to see how “loiter” was going to work. When we pushed it hard in roll it went into the wobble of death. Didn’t crash but had to hit the throttle and flip back to alt hold to save it.
I would REALLY love to figure this out. 
Help would be MUCH appreciated.
Then we put it in loiter and pushed it to roll, and it began to wobble and overcompensate and almost crashed.
here are my log files today.
also, the throttle in stabilize mode is VERY VERY sensitive. not even one mark on the throttle stick and it is ready to take off.
https://drive.google.com/open?id=1cGp74zZA6piQAuZYChUeST-hvaaByEHg
Follow up question… when tuning, I see references to tuning for all flight modes? If that is the case, then why would my stabilize throttle be so much different than any other mode? I realize the stablize mode has a more aggressive stick, but where is the setting to cool that off a little? Not even one notch of throttle on the TX and it is flying. Same for coming down.
Thought I would give a positive update. I was given the suggestion to look at ATC_ACCEL_R_MAX setting. and compare that setting to the some one for the Pitch. For some reason it was much different. I’m guessing it got changed during an autotune. Even so, i am wondering if this setting is also one of the settings you can change in the tuning window? And why one would NOT be there if it is not? I thought it was one of the settings at the top that gave the range that you could tune withing. But this seems to be a CONTROL over the whole thing. Anyway, it is nice to have that fixed, but I am nervous about trying any further auto tunes. It would be a very RARE case to have most of the pitch and roll tune settings different. if ANY settings related to pitch and roll, it would be nice to "limit’ the difference.
