Is this really backlash?

[igull]

Hi
I recently installed the new 2012 controller, and am having great success with it.
However, I noticed something that on the face of it looks to me like classic Y backlash, but I'm fairly certain isn't.
The issue initially showed up when cutting some concentric circles (the inner was held with 'bridges' to the outer) - the inner was offset in Y. I tried various mechanical cures with absolutely no effect, so attempted something more scientific.

I made up a rectangular 60mm vertical 'snake' pattern which has has six 10mm segments. The left pattern starts at the top and cuts down. The right hand pattern starts at the bottom and cuts up. When cut, the starts and ends of both patterns are aligned - the segments in between are not !
The first section (bottom) of the right hand pattern is short by about 0.5mm, but both start and finish lines are aligned.
There's also something else that's strange - both axes use exactly the same mechanics (different steppers, but both 200 step) - over 60mm, the X is dead nuts on, the Y is 0.5mm short. Both axes have the same setup in software.
I know the mechanical setup is ok because I have been using it with my lasersaur controller and I have done many tests (I can't change the controllers right now for other reasons unfortunately). It's the same over the whole of the cutting area (1200*600mm).

Any thoughts or magic spelles ? - it smells like software to me

Main board 7.12.04.01
LCD 5.12.02.29
LaserCAD 5.93

I've attached the snake pattern that you can try. Make sure that you DON'T have 'auto order' set, otherwise it won't work - it should start top left, work down to the bottom, hop across to the right snake then work it's way to the top.

I'll post this across on buildlog.net in case anyone else has seen it.

TIA

Neil

[Tech_Marco]

Can you post a picture showing what the problem is?

I'll test it after finishing my workshop cleanup


Marco

[igull]

Hi Marco

Thanks for that, I've posted a whole write-up across on buildlog.net. Rather than repeat it here, here is the link ...

http://www.buildlog.net/forum/viewtopic ... 581#p10581

I can post it here for posterity if you like 'though.

Cheers

Neil

[Tech_Marco]

Per programmer, trying changing the "Falling edge" or "Rising edge" setting on System setup

If you have "Falling" selected, then change it to "Rising". Or, if you have "Rising", then change it to Falling.
That should take care that issue


Marco

[igull]

Per programmer, trying changing the "Falling edge" or "Rising edge" setting on System setup
If you have "Falling" selected, then change it to "Rising". Or, if you have "Rising", then change it to Falling.
That should take care that issue
Marco

Hi Marco
Sorry, no change at all - still does exactly the same. I have been doing a lot more tests today with the 'backlash snake at 45deg - results are perfect, zero deg perfect - only at 90 deg does there seem to be an issue.
The stage always returns within 10 microns or so to the same spot.
There's more write-up across on buildlog.net
I'm pretty certain that it is software/firmware, but I can't get anyone else to repeat my results, so I'm at a loss for an answer.
I wondered if it was tied into the previous scanning issue.

Cheers

Neil

[Tech_Marco]

I haven't had a chance to test on my little K40 as I'm busying doing inventory sorting. But I doubt that it's software or hardware issue. I'm thinking that it could be the motor issue instead. Did you try it in slow speed say 10mm/s for the same file? Are you running K40 or other machine? Did you apply enough current to the motor? A folk in buildlog said that he had to set the belt as tight as a guitar string which I don't think so. If it worked out that way, that means something wrong with the machine itself, but not the DSP card (software and hardware). When the machine is making a 90 degree turn, the momentum from the motor is the greatest as it need to stop and lock itself quick then turn. The motor seems losing step and the only thing that it come to my mind is either the motor is bad or the current is low. Just my though

Marco

[igull]

If it worked out that way, that means something wrong with the machine itself, but not the DSP card (software and hardware). When the machine is making a 90 degree turn, the momentum from the motor is the greatest as it need to stop and lock itself quick then turn. The motor seems losing step and the only thing that it come to my mind is either the motor is bad or the current is low. Just my though
Marco

Hi Marco

Yes, I agree completely with you - the belts just need to be taut, not tight!
I did some more testing late last night and changed the rectangular 'snake' for a snake made from arcs - problem showed up right away - at the start of a run - any direction, the arc was flat at the start - a sure sign of backlash. I'm certain it's not mechanical backlash 'though, as my setup is extremely smooth and tight - there is no problem in X - it uses a nema17 size motor for weight. I have a hi-torque nema23 in Y - however, the stepper drivers are keling 4020's - max 2A - I have the Y cranked up to max current, but perhaps I am asking too much of the 4020 and need something with a touch more 'beef'. The steppers are both brand new and seem to operate correctly, hold position ok etc They both run fairly warm as expected.
It has no problem moving the stage around. I'm going to order a new stepper driver with double the current and see where that leads.

Thanks for the useful comments Marco. I'll post this across on buildlog too.

Cheers

Neil

[whmeade10]

If it worked out that way, that means something wrong with the machine itself, but not the DSP card (software and hardware). When the machine is making a 90 degree turn, the momentum from the motor is the greatest as it need to stop and lock itself quick then turn. The motor seems losing step and the only thing that it come to my mind is either the motor is bad or the current is low. Just my though
Marco

Hi Marco

Yes, I agree completely with you - the belts just need to be taut, not tight!
I did some more testing late last night and changed the rectangular 'snake' for a snake made from arcs - problem showed up right away - at the start of a run - any direction, the arc was flat at the start - a sure sign of backlash. I'm certain it's not mechanical backlash 'though, as my setup is extremely smooth and tight - there is no problem in X - it uses a nema17 size motor for weight. I have a hi-torque nema23 in Y - however, the stepper drivers are keling 4020's - max 2A - I have the Y cranked up to max current, but perhaps I am asking too much of the 4020 and need something with a touch more 'beef'. The steppers are both brand new and seem to operate correctly, hold position ok etc They both run fairly warm as expected.
It has no problem moving the stage around. I'm going to order a new stepper driver with double the current and see where that leads.

Thanks for the useful comments Marco. I'll post this across on buildlog too.

Cheers

Neil

I am using the 4030s and they work great, but I am only using NEMA 17 steppers. I think you are on the right track getting a driver that will provide more current to the NEMA 23 stepper.

- Bill