5 axis CNC Router/3d printer

[George Butty]

Hi Tom, moving the bar by hand will deflect position but it bounces back to the position it was at before deflecting it+-. If at zero bounces back to
zero. If at error position bounces back to error. The only value changing on DOR is the C angle. C rotation is very smooth and quiet
and not aggressive enough to move main frame. Machine anchored to concrete floor. The only thing moving is C Axis.
I rotated small angles. Started with +- 1 deg. and gives approximately .375" movement from ball +-. Comes back to zero both ways.
Tried +- 5 deg. noticed a small shift from center on the return of error side. Tried +- 90deg. error shift approximately half of what it
would be running C180+-. So as the angle increases so does the return error.
Yes I can see this being a bitch!

George

[TomKerekes]

Very strange. But you need to check the KMotion Axis Screen not the DROs. Those are the raw motor positions.

[George Butty]

OK and also I can check the Teknic motor positions as well by connecting to it.
Will do this tomorrow.

Thanks
George

[George Butty]

HI Tom, found problem. It is backlash but not obvious. Kmotion is going back to zero and so is the motor.
I had to push lot harder than I thought to get the backlash to show. How does backlash comp work. Would this
help improve problem?

Thanks
George

[George Butty]

What does backlash rate mean?
George

[George Butty]

hi Tom , found how to setup backlash. Will give it a try.

George

[TomKerekes]

Hi George,

Software backlash correction rarely works well. It assumes the backlash is constant but this is more dependent on distance moved.

The Backlash Rate is the rate the Backlash correction is applied. For example if the Backlash Amount is 100 steps and the rate is 5000 steps/sec then the Backlash will be applied linearly over 100/5000 = 0.02 seconds

You might check in the Teknic Drives to see if it is an issue with servo tuning or it is purely mechanical. Do the drives show 0 position error? Hopefully its an issue with tuning and can be corrected by the drives.

[George Butty]

Hi Tom, I agree backlash comp sucks. At this point it does look mechanical because both Kmotion and Tecnik zero
all the time. I have been playing with the backlash comp and it seems to be working. Does the MoveAtVel account for
the Backlash value? It seems like it does not. Is there away to have it accounted for when when I zero axis. Right
now I zero to the backlash position then in MDI I run C-90 C0, and then it goes to the right position. Still haven't run a program
so not sure how this will work.
Thanks
George

[TomKerekes]

Hi George,

At this point it does look mechanical because both Kmotion and Tecnik zero all the time.

So the Teknik internal encoder position matches the commanded position? I wouldn't expect there to be zero error. How much error is there?

Does the MoveAtVel account for the Backlash value?

Yes it should

Is there away to have it accounted for when when I zero axis. Right now I zero to the backlash position then in MDI I run C-90 C0, and then it goes to the right position.

Not sure I understand this. But a call to Zero() disables the axis and re-enables the Axis. Enabling the axis clears the current backlash correction.

Moving in the positive direction the BacklashAmount is added.
Moving in the negative direction nothing is added.

So I think if you move in the negative direction before calling Zero() the backlash should remain properly applied.

HTH

[George Butty]

Tom , if you are talking about encoder error in the Tecnik, then the errors run average 0 to +- 3 counts on
a 8000 cnts per rev. encoder when based on Kmotion sending a 0 position output. I will double check this number again
because it does seem crazy low. One thing about those motors is they are amazing once settup.
I have had a crazy time Zeroing up the B and C axis. The backlash is 285 cnts and it had zero error in one direction and the full error in the other
when I would Zero the Axis.This was driving me crazy since I could not figure out why things where good one why and crazy out the other. So now
I managed to split the error + and - to give me an even error both side. Haven't finished settup yet. Hoping tomorrow.
All I can say Tom is I have built hundreds of custom machines in the last 40 years and this one was by far the most challanging
to settup! Can't overlook anything and order of opperation is very critical.

Thanks
George