Acro mode! And more...

[Patscherpofel]

so I tried to fly it again with 1919
I noticed some very strong yaw corrections the quad would do by itself , after take off especially but also after the first yaw inputs
so I thought it was the deadband I put to zero and put it back to 40
I also changed the smoothing filter (low and high) to see what happens
thats when things got totally random again,...sometimes heavy yaw corrections in opposite direction of stick input(again all in between the deadband)
so it seams something is seriously wrong with those parameters
I really hope you can find the problem

[JussiH]

Can you try and reset all params and then uoload them again.

[Max]

[Patscherpofel]

shit...I missread your message, will try 1987 also

the mixing of deadband and smoothing filter could be a problem
so if I put both deadbands to 0 I should see what the smoothing filter really does and tune it to a way I like it
can I turn it completely off also?

yes I really think its asymmetric in some sort of way...super hard to see...should put a camera on it also
the right direction seams to start more directly ...then slows down and then goes faster again...I have to look at the logs again

[Max]

[Max]

Just for "fun" here is a test spreadsheet showing result of the triple filtering going on for angles and rates. Simply adjust the "tau" value and watch the resulting graph. Blue line is input signal (randomly generated), purple (z3) is final filter output (z1 and z2 are intermediate stages from 1st and 2nd level of the filter). For example you can see that when TAU matches DT (delta time, fixed @400Hz in AQ) then the output follows the signal.

-Max

[Patscherpofel]

so now I set to defaults and reloaded the params from 1952, (fw on the m4 is still 1919)
and to my surprise all problems were gone

so I set roll pitch and yaw deadbands to 4...and it reacted nicely to my sticks (with 50%expo)
also the flight performance is good...probably not any different to 1952

so I started to see what the tau filters do and changed the tau angle yaw to 0,3 (couldn't tell a difference)
and then I changed tau rate yaw to 0,05 ...and bamm same problems like before
random jumping yaw...and sometimes up to 45 degrees jaw correction on take off!
I rebooted several time...always the same
but when set back to 0,005 (same as roll pitch rate tau) it is all good again

...could it be that there is a problem if roll pitch and yaw rate tau are not the same ?

[Max]

Hi Pat,

I'm a bit confused because the default yaw rate filter TAU is set to 0.01. And it flew OK with defaults but not with 0.05 yaw rate TAU?

The yaw angle filter was removed in 1923, it was causing issues with nav functions (and not needed). Tilt/roll rate filter is only used in "acro" (rate) mode, otherwise the angle filter is being used.

-Max

[Patscherpofel]

I am confused too

I tired it again, 0,01 is also ok...0,05 is acting all weird...really hard to discribe what is going on...it seams like it wants to counterattack it in the wrong direction somehow
I can send you the logs if you want

[Patscherpofel]

another thing that is confusing to me how the smoothing filter for "tilt angle" really works
0.06 fels more reactive then the default , but it still is too slow to follow fast stick inputs...but lowering it even more and the tilting angle overshoots very strong..you can easily tilt the aircraft up to almost 90 degrees
I thought this should be limited by the pitch/roll scale , which I have set to 0.06
do I have to retune the pid controller to prevent the overshooting?