G540 + EMC2 = missing steps

Hi,

Still same problem after changing the PSU. See attached photos. I reduced the accel from 15 to 10 and test run.
Thanks

I am not so sure this is step loss. Happens at the same spot every time.
Did you check for backlash or other mechanical problems?

Hi,

How do you check the backlash?
Thanks

Normally you use a dial indicator. Put it on the spindle carriage with the business end on a fixed object on the table. Zero the indicator. Move a distance away from the fixed object (one inch, for example), and then the same distance back again. the dial indicator should return to zero.

If you do not have a dial indicator checking for a gap will work too. Put a round piece in the spindle, move a fixed object to it so it touches and there is no gap, and do the away/return thing. Check gap with a flashlight.

Also check the ballscrew/acme spindle bearings and it's nuts for play.

Hi,


This is what I got:

An increment movement of .01 on y axis : from 0 to .50 = getting .498 the emc2 screen says .500

An increment movement of .01 on x axis : from 0 to .50 = getting .498.5 the emc2 screen says .500

difference of .002 would that be enough to do it ?
Thanks

No, it won't. That gives you 0.002 deviation in the distance, and I see more in your photos.

If you do a long spiral, what do you get?
If you use a pencil instead of an endmill, what do you get?

Hi

ArcEye wrote:

Give it a try and see if it cuts properly.
If so you can add things back until it errors and you will get a good idea what is causing problems.

Did you ever do this?
Your PSU was under powered, but if that was not the root cause, then something else affected the problem

The thing against this being simple binding, is that the modified code worked properly.

Was it setting G61?
If so that tends to go against binding too, exact path mode cannot affect that.

Andy Pugh wrote:

How does the hole size compare to what is being requested? I could imagine what you see being due to machine flex. (especially if the Z rocks during a plunge)
Tangential arc lead-in moves would be nicer.

Did you try this?
Change the code to begin at a point outside of the shape to be cut and plunge at a tangent down to just above the workpiece.

What your original code was doing was the opposite, it started at X0 Y0 and then plunged towards the outside extent of the shape, meaning any overrun through flex would be outside that extent not within it.

My changes in breaking up the moves, brought the cutter above the plunge point and then lowered the Z axis, removing this outward tangent.

Might also be time to share exactly what machine this is, we can guess it is pretty small and probably router-ish, but not much beyond that

regards

Hi,

This is my machine:
K2cnc capable of cutting 13" x 25" with 296 oz of stepper motors and a porter cable router .See attached photos.
software by Cut2d.

Thanks

Hi,

What do you mean by spiral ?

Thanks

Fredasiong wrote:

Hi,
This is my machine:
K2cnc capable of cutting 13" x 25" with 296 oz of stepper motors and a porter cable router .See attached photos.
software by Cut2d.

OK so it is a small router type machine.

How about the other questions re the code, what did you try and what were the results?

regards