Welcome to LinuxCNC.org

Hi,

I have a 24x24 router machine with G540 controller running LinuxEMC.
Machine has made a lot of good projects but just recently has started having a problem.

I am cutting a shape that has arcs and I am doing multiple passes to make this happen. I am noticing
significant drift in the negative y direction when the additional passes are done.

Forinstance, I cut "air" for a test and when the program was done, had the tool return to the home position, it is significantly off in the negative Y-direction.
The X-diretion is dead on.

I slowed down acceleration, feed rates, and velocity, but still have the same problem. Each time I do a depth pass, it drifts in the negative Y position about .05" ish.

What is going on???? I am lost. Especially when it was working and I didnt make changes. Am I losing steps? Even when cutting air?

I did some basic tests where I home it, send it in positive Y direction and bring it back and that test looks ok. It is something to do with making a compound arc where X and Y must move together.

I am off to clean the machine now, maybe that will help...can't hurt.

Any thoughts would be appreciated!

Justin

difficult without knowing the build of the machine , but if the machine was ok , and is now drifting , i'd first go looking for any backlash in nuts or loose belts
and any unwanted movement . if nothing has changed electricaly or in linuxcnc files etc , the likely hood is that it's mechanical wear , or loose pullys etc

Also check for loose electrical connections, broken wires etc. It will be mechanical or electrical.

John

The machine is a Zenbot 24x24 that uses pulleys and belts. I checked the Y-axis and there is zero slop, nice and tight.

Belt and pully systems are notorious for belts stretching , and could account for the problem , even though the belt is tight , check distances between teeth with caliper .

Have you done a stepper test on the Y axis?

linuxcnc.org/docs/html/common/Stepper_Diagnostics.html

Could be binding up somewhere...

Got a photo of the machine?

John

cpy911 wrote:
Mark every pulley, coupler and shaft interface with a marker pen. The most likely thing is a slipping coupling.

If it isn't that then it could be a phantom step problem from noise. Maybe a shield connection has come adrift, or an earth tag has come loose.

Did anything change since the problem appeared. Anything at all. It might be that you think it is irrelevant but we can say, "ah, well, you see, X always does that because Y" (Though admittedly it is more likely we will look puzzled)

Ok. Still frustrated here.
Here are some tests I have run:

1. Set home position, run a program to machine a circle of about 8" diameter. Raise Z-height to not actually cut anything, tool takes multiple passes at .0625" depth pass.
Result: tool returns home

2. Set home position, move y axis +10" and return home, move tool to random y and x locations and return to home.
Result: tool returns home

3. Set home position, run a program to machine the letter "D" of about 12"x12" diameter. Raise Z-height to not actually cut anything, tool takes multiple passes at .0625" depth pass.
Result: tool returns home to X position, Y position has drifted by .250 or up to .50" in the negative Y direction. It is drifting more and more with every pass. (If I took fewer passes, I would see less drift, if I did just one pass, it goes back to home...the more time it runs the more drift I see).

I am baffled!!!! I can machine a perfect circle or move the tool to locations and I get a good return to home position, but if I try to machine something with arcs, it drifts or the Y-position like it is not seeing the Y-direction commands!!!

Of course, this all started when I was machining a sign with various letters, I noticed each pass the y axis was drifting so it took more and more of the material I did not wanted to machine.

Some things I tried....checked wiring and fitting for the Y-axis. Made sure they were secure.
Slowed the machine down both acceleration and velocity.
Rebooted the machine.
Checked belts for tension and y-axis was secure and not loose.

Thanks!

have you changed anything .. stepper motor .. stepper driver ... power supply ,
acceleration or velocity or even stepper current settings

could although i doubt it be a bad motor , so try swapping motors and or stepper driver and see if the fault then moves to the other axis

Hello,

You might try the following ngcgui sub at different speeds to check if the problem only occurs when more than one axis is moving.
The idea is to move the axis back and forth at ever shorter moves and then return to the start location. At high speeds this should put some stress on the axis.

www.linuxcnc.org/index.php/english/forum...929-check-axis#27516

What are you using for G64 settings? A circle and a sharp corner may put different stress on the involved axis.

Is it possible that your power supply is marginal and cannot not provide enough amps for combined moves?

Are you using any backlash compensation?

I am surprised that different speeds and acceleration in your test did not produce different results, did you make these changes dramatic?

Having said all that following cncbasher's advice should help pinpoint the problem.
Rick G