Y Axis Creep

More
28 Oct 2020 18:23 - 28 Oct 2020 18:47 #187600 by squiresuzuki
Y Axis Creep was created by squiresuzuki
Setting up a Gerber Sabre 408 and it develops fairly significant creep in the positive Y direction after machining (as much as 1" after 10 minutes of machining). Seems to happen regardless of feed rate.

Using pncconf, Mesa 7i76, Clearpath servos, stepgen set up with step/dir, control type is '1' (velocity?). Removed stepgen maxerror as suggested by PCW on various forum posts. Standard setup with P=1000,I=0,D=0.

DRO does not show the creep.

Any help is appreciated, thank you.
Last edit: 28 Oct 2020 18:47 by squiresuzuki.

Please Log in or Create an account to join the conversation.

More
28 Oct 2020 18:33 - 28 Oct 2020 18:34 #187603 by PCW
Replied by PCW on topic Y Axis Creep
An accumulated error might result from incorrect step/dir timings
especially direction setup and hold or step pulse length.
Inverted step polarity can also cause issues.

As a quick check I would double the step length and set the
direction setup and hold timings to 20000 ns, and see if there is any
difference.

It could also be a mechanical problem.
Last edit: 28 Oct 2020 18:34 by PCW.

Please Log in or Create an account to join the conversation.

More
28 Oct 2020 18:59 #187604 by squiresuzuki
Replied by squiresuzuki on topic Y Axis Creep
That worked!

Original values:
step time 1500, step space 1500, direction hold 1500, direction setup 1500

Working values (your suggestion):
step time 3000, step space 3000, direction hold 10000, direction setup 10000
(pncconf didn't allow values larger than 10000 for direction values)

Why don't the original values work? They are above the minimum values from clearpath (step time 715, step space 715, direction hold 1000, direction setup 25).
And now how can I determine the optimal values? Just trial and error?

Please Log in or Create an account to join the conversation.

More
28 Oct 2020 19:04 - 28 Oct 2020 19:05 #187605 by PCW
Replied by PCW on topic Y Axis Creep
The optimum values are the largest that do not interfere with the maximum required step rate.
Last edit: 28 Oct 2020 19:05 by PCW.

Please Log in or Create an account to join the conversation.

More
28 Oct 2020 19:26 - 28 Oct 2020 19:27 #187608 by squiresuzuki
Replied by squiresuzuki on topic Y Axis Creep
Gotcha. So that would mostly be a function of step scale and max velocity?

step scale = 2540 steps/inch
max velocity = 400 inches/min

1/(2540*(400/60)) = 59055ns step period (time+space)
59055ns / 2 = 23528ns step time
Minus some margin, so perhaps 20000ns for both step values?
and keeping the 10000 direction values.

Just want to make sure that kind of value is in the right ballpark.

Thanks for all your help.
Last edit: 28 Oct 2020 19:27 by squiresuzuki.

Please Log in or Create an account to join the conversation.

More
28 Oct 2020 21:28 #187632 by PCW
Replied by PCW on topic Y Axis Creep
I would keep the step time shorter than the hold time just to guarantee
that the timings still work if the step polarity is not right

Since 1500 ns was marginal and 3000 worked, I would probably use
5000,5000,10000,10000

Please Log in or Create an account to join the conversation.

Time to create page: 0.075 seconds
Powered by Kunena Forum