Acro mode! And more...

[Patscherpofel]

and one thing I immediately noticed is that the different firmwares have very different performance on the quad
I recently changed to esc32v3 and tried them first on the new fw...it was shking badly on take off so I thought it was the faster esc and retuned it...tried a lot but only thing that helped was reducing both f values to around 0.18 ,but still it never flew super stable...(this is pid mode)
now back on the old fw I tried for fun to retune it again and pretty much ended up on my old values wich fly much better again

hmm so I have a bit of a dilemma going on

old 52 fw
+flies fantastic
+yaw contoles according to my stick input and not so erratic and robotic like on the new one
-it has this horrible stick delay on roll and pitch...which is an absolute nogo for me
this delay make it very mushy to control ...for example if you want to land it spot on, or also hand catch it
the new smoothing filter in the new fw was helping a lot (giving a sharper control when reduced) so I really thought I would be happy from now on

new fw
+has more direct stick input if you lower the tau filter
-never found a decent pid value for good flight performance
-has the described yaw problem in between the adjustable deadband (very noticable if you reduce it to almost 0)

[Max]

Hi Pat,

I'd like to look into these problems more. Just to be clear... which firmware are you calling the "old 52?" 7.1.1852? And also which is "new?" This "acro mode" one (7.1.1923)?

What about 7.1.1897 (the last "next" branch build)?

Thanks,
-Max

[Patscherpofel]

which firmware are you calling the "old 52?" 7.1.1852?

yes correct


And also which is "new?" This "acro mode" one (7.1.1923)?

yes correct

What about 7.1.1897 (the last "next" branch build)?

never tried it so far, should I?

thanks for the help!

[Max]

[Patscherpofel]

I'm not sure about the jump thing. Does it jump right or left?

It felt like as soon as I exit the deadband it was doing an erractic correction on something
this was very noticable on right stick inputs ...but I didnt notice it on left
like the deadband wasnt centred to my stick somehow

I general I dont like that deadband
if I want smooth controls I use expo on my tx but not a high deadband...of course I dont want a jittery remote signal entering the controller...but the deadband used on AQ is much higher then it could be right?

[Patscherpofel]

I guess the best would be to use this well working m4 with 1852 and update the fw
just to be sure...can I simply update and keep the parameters and mag/tare calibrations , or do I have to put in everything from scratch to make sure everything is correct?

[Max]

[Patscherpofel]

thanks Max, great help!

I never tried the 1923 on a bigger quad so far, had this m4 with that fw for a while on small race quad
so I dont really know...but I will update the one thats in there right now and compare it...especially the different pids I got
frome the change list I would guess that they should pretty much have the smme pids right?

[Max]

None of the versions should have any difference as far as PID (or Quatos) settings. That would be a "highlighted" change for sure. That's why I'm concerned/surprised that you're seeing any difference with 1923. The only (intentional) change in regards to PIDs in 1923 was the addition of the separate rate-only mode settings.

-Max

[Patscherpofel]

so now I installed 1919 and it seams to fly fine...would need to test more to find out why 1952 was shaking so badly
could be the other m4 or some mistakes I did with the pid also

the problem with yaw is still in there, but I couldnt reproduce the yaw into the wrong direction
one thing is very clear that going in and out of the deadband is not smooth...and it somehow feels that it reacts different to left and right...maybe that is the problem with the code you found , I dont know

one question...did the1952 had deadband on yaw at all?
even a deadband of only 2 is very noticable...maybe there needs to be some sort of speed ramp between inside and outside of deadband?

did it have a default smoothing filter on 1952 (roll/pitch/ yaw angle and rate tau)? or is this something implemented recently to bypass the pid controller?