2DOF WiperSim+Pololu problem(solved)+traction loss

[yokoyoko]

Never ending modifications......

... Hi all,
I moved with my sim into another room. After some tips of sirnoname I changed several parts and the geometrie of the sim. I'm also trying to find the "perfect" PID-settings for my setup After several test and regarding the graphs in the Pololu configuration utility I now use P=6 I=0,25 D=8 - this works more like the SCN5 than my previous PID settings. A lot more micro shakes and curb movement BUT also alot more headache after some rounds on the Nordschleife

If you like have a look:


Best regards
yokoyoko

[motiondave]

Looks good there.
I cannot quite understand the PID settings
P = 6 and whats the other number
I = 0/25
D = 8 and whats the other number?
What max accel are you using?

Interesting to see the result of JRKS on wipers.

[yokoyoko]

Hi Motiondave,

Looks good there.

Thank you very much!

On PID:
Important is the result of P,I and D.

Unbenannt.PNG (8.49 KiB) Viewed 11023 times

Numerator and denominator are only for the possibility of getting decimal places.

On acc, max duty and co -> this depends on. Atm I use all @ 450 and current = 0.

Best regards
yoko

[motiondave]

that still didnt answer my question, i know its a math calculation, but what are the numbers please.
The picture shows nothing.
For example, mine is ,
P =4/4
I =0/3
D = 4/4
I have two numbers listed. I am curious as to what ALL your numbers are.

[yokoyoko]

Hi dave,

I have no access to my sim here. And if the Jrks are not pluged in - you see no values in configuration utility

Don't understand your point? My result for P is 6 ; for I is 0,25 and for D is 8 !!!

But if you fell better I will send you a screenshot with ALL numbers tomorrow

Please correct me but your I value is useless, or not? - if you use 0/3 or 0/12 or 0/100 all results are 0...

[motiondave]

Thanks, that will satisfy my OCD if you can post a pic of it when you have time.
For reference, I have worked on PID to get the motors to run in the utility slider with minimal errors, then I start changing things based on reactions from the motion of the game.
But thats how i do it.

[yokoyoko]

I watch the graph (plot) for one axis while playing the game. I let the graph be drawn (plot) for scaled feedback, target and error. You can see the over and undershooting while driving (in different situations of course) in realtime - cool feature of the Jrk's.

The adjustments I have done after "Ziegler and Nichols" method - easy and efficient!

Screenshot will come tomorrow.

[motiondave]

I watch the graph (plot) for one axis while playing the game. I let the graph be drawn (plot) for scaled feedback, target and error. You can see the over and undershooting while driving (in different situations of course) in realtime - cool feature of the Jrk's.

The adjustments I have done after "Ziegler and Nichols" method - easy and efficient!

Screenshot will come tomorrow.

mnmmmm...didnt think of that...I just drive, smash into some walls then adjust....

[yokoyoko]

As promised for Motiondave the screenshots of the again finetuned PID:

3.PNG (74.7 KiB) Viewed 11063 times

and the proof that just the results are important - they way you get them is your choice :

md1.PNG (2.35 KiB) Viewed 11063 times

---> 22/(2²)=22/4=5,5

md2.PNG (2.41 KiB) Viewed 11063 times

---> 88/(2^4)=88/16=5,5

Best regards
yoko

[motiondave]

Thanks, that makes more sens to me seeing the TWO numbers.
Interesting having P 11/1 , I 1/3 , D 17/1.