Prototrak Plus Retrofit with Mesa 7i77 & 5i25 cards

Many thanks for all the help!!

Todd I will try this.
""Next time your working on this, open a HalMeter window, and If you are working on the Y axis check the value of the pin "ini.1.max_acceleration" (or for X use ini.0.max_acceleration).
Then using the "Test HAL Command" box in ShowHalConfiguration you can use the command "setp ini.1.max_acceleration 15" (or what ever number you choose) to test a new acceleration value for that axis.""


This is vel at .8 and acceleration at 18, closer.

When using "hal scope" there are times when the "trigger" will not work in "normal" need to use "auto" and jog when "trigger bar" is moving. I can tweak .ini file and come back and everything works, what am I missing?
Where the teal (light blue) overshoots the blue, what value am I looking for, or just try to keep the "pid.n.error" to the min amount as possible. I was thinking that lower velocity to .75 from .8.

Todd, I have reread your advise and can see how that saves time.

Many thanks!!

Set the trigger on hal scope to Source Chan 4 (axis.1.joint-vel-cmd) and you should get more consistent triggering.

I think increasing your FF1 should bring the following error (pid.y.error) while moving more in line (I think FF0 should probably remain 0)
Maybe a little more D and/or a little less P to try to tame the over shoot when reaching cruising speed.

Many thanks for all the great help!!!

This screenshot D = 24 P = 3300, too much D and pid.y.error goes way up.



this is D = 27 P = 3300 (this was .72 vel & 17 acceleration), there are several below, read the (72v17a is .72 velocity and 17 acceleration in .ini file) name of file for description of change.



The pid error is still high correct? How close am I?

Many, many Thanks!!!!

Interesting. Now try less FF1. Maybe 0.01

Many thanks for all the help!!

The screen shots have description in the file name. I believe this is the best, (I will need to look at these side by side myself) compared to others. The pid.n.error gets lower as the pics are arranged in order.




Many, many thanks!

Edit: The scale for pid.n.error was change between the pic, form 1m/div to 2m/div. The best screenshot I think is Sshotyf1.01f23.8Db.00003.png just did not edit well.

Just for shits and giggles try FF1=-0.2

Todd thank you!


this is FF1= .2 and other is .0009



the last 3 look the same to me, when I look at these traces, I am trying to keep pid.n.error low, and trying to keep the light blue (teal) trace (.encoder) from over shooting the dark blue. I can revisit the .ini file, I don't have notes in front of me to include that data now.

many thanks!

Now try a negative ,2 in FF1 (the font used by the forum made the "-" sign in my last post very small so maybe you didn't see it.

Hold on, Also try much larger FF1 valuses, like 1, 2 ...
To see the trend here. Don't change any onther things. Only make one change at a time to get a feel for what is actually doing what.