Missing Steps in circular interpolation

[ShawnM]

What kind of drives and motors? Are they identical?

He mentions the Gecko G210X drive in his first post but nothing about the motors.

[kman1967]

Response to 'ShawnM'
They are wired to the headers. Linear moves are good I actually checked that out with a dial indicator

[kman1967]

'ShawnM'
Wired to the headers. I checked the linear moves with a dial indicator and they were good
'Centroidsupport'
Specs for motors attached. They come with the Avid Machine. drives and motors are identical.

[kman1967]

'ShawnM'
Wired to the headers. I checked the linear moves with a dial indicator and they were good
'Centroidsupport'
Specs for motors attached. They come with the Avid Machine. drives and motors are identical.

[Gary Campbell]

That chart appears to have added "2000 PULSE" to the full step (max) torque for the motor. It is my guess that when 2000 p/rev is selected (IMO the worst choice) torque is reduced by ~70%..or more! Add that to the fact that mechanical reduction is lower than desirable, you will have to be prudent in feed/speed selections.

Were the actual cutting feeds/speeds ever mentioned? Does the issue persist if feed is cut in half?

[kman1967]

'Gary Cambell'
I have tried tests with different feeds and speeds and the problem persists. I mentioned in my first post that I ran tests with no material. Basically I started from a 0 point, made a linear move, made a circle then back to zero with another linear move. The spindle was out from the original zero by about 0.010". If I repeat the sequence the error accumulates. If I do the same test but reverse the direction of the circle the error starts to subtract.

If I do only linear moves all is fine.

[CNCMaryland]

Backlash or slop in the axis.

[Gary Campbell]

'Gary Cambell'
I have tried tests with different feeds and speeds and the problem persists. I mentioned in my first post that I ran tests with no material. Basically I started from a 0 point, made a linear move, made a circle then back to zero with another linear move. The spindle was out from the original zero by about 0.010". If I repeat the sequence the error accumulates. If I do the same test but reverse the direction of the circle the error starts to subtract.

If I do only linear moves all is fine.

I've seen this before. I suggest connecting to the DB25 and setting the step pulse to 100Khtz or less.

[kman1967]

'Gary Campbell'
- Is there any specific hardware you are recommending to use on the DB25 as a breakout?
- Are the headers not the same as the DB25? I am curious why they would introduce this problem.

[cnckeith]

in general, screw terminals are for AC brushless drives
DB25 connector is for stepper drives.
search "gecko" here. https://www.centroidcnc.com/centroid_di ... browse.php
to see examples of other gecko drives connected to the DB25

DB25 breakout board we use is here.
https://www.centroidcnc.com/centroid_di ... _gear.html