cnc router is out of alignment after a few movements

[ShawnM]

That link worked just fine, thanks for the pics.

It doesn’t matter much which drive type selection you make in the wizard, it’s just a jumping off point to get you started on some basic settings. You can then tweak it from there to suit your machine and drive configuration. When you get everything dialed in simply create a custom drive profile for your drives.

With your report we can see how you have the limit switches wired to what inputs and how they are labeled.

But again, after you home the machine to your home switches and have your soft limits setup as you do you shouldn’t be able to jog any axis and trip a home switch again.

Right after you home the machine is it square? How do you know?

[vincent-visser]

Right after you home the machine is it square? How do you know?

Well,after auto-squaring it is as square is at can get right now, untill the problem is solved and the machine doesn't loose it's place. We've cut a square en both diagonals are within half a millimetre over a distance of 150cm of each other. Still not very confedient with that because after that we drive back to 0,0,0 and it hits the poximity sensors again.

Can you understand what our problem is after all those massages?

[cnckeith]

when you say "drive back to 0,0,0" what did you actually do?

[cnckeith]

thanks for the photos. please add some close ups of the prox switches. the prox are hard to see.. but,
typically you mount prox switches so that if there is any over travel the machine will not crush/hit the switch.
so the two i can see on the gantry axis should be mounted 90 degrees in orienatation from thier currect configuration

[ShawnM]

As Keith stated "best practice" when mounting limit/home switches and or prox sensors is to mount them at 90° from the direction of travel. Then mount a steel flag on the moving part to pass in front of the sensor. It's a bad idea to mount any switch inline with the travel motion so it can get damaged.

Try slowing things down, lower your fast jog and max rate to half of your 15000mm/min that you have now. What feed rate are you cutting your test square at?

Increase your Accel/Decel rate in the wizard from .3 to .6 and try again.

And be sure to tell us how you "drive back to 0,0,0"

[vincent-visser]

Allright, got some progress on what and when the problem is happening.

I changed the mounting positions of the proximity sensor by rotating them 90degrees en that they never can get damaged anymore. Adjusted the sensor positions so that after squaring y-master and y-slave are exactly the not distance from the beginning of the linear-rail. ( I know this doesn’t mean the machine is square but that’s for later on).
After that we started the machine and auto-squared all axes, finishing up with machine coordinated 0,0,0.
Then we zeroed also the wcs.
Next up we put an MDI G1Y314,3007F300. Our Overall Turn Ratio is 31,43007mm so means that the stepper motor does exactly 10full circles.
Then I measured the actual traveled distances of both y-master at 318,… mm and y-slave 314,…. mm. (the average of both is close to what it should be).
Good to know: Y-master is in normal direction and y-slave is in reverse direction.

Then I power cycled everything, homes and zeroed all axis and did the same MDI code but with 7500mm/min feedrate instead of 300mm/min and now both axis y-slave and y-master are exactly spot-on at 314,300mm.

This means that at low feedrate something happens what causes an extremely bad alignment.

Next I changed out motors from both axis to see if problem moves but everyting stays the same.
After research we decided to rebuild the power supply of stepper drives from 5V to 24v and retried everything but no changes.

See photo for new mounting positions of prox. Sensors

[tblough]

My guess is that the side with the shorter travel is binding and then loosing steps. Make sure your guide rails are parallel and that you can turn the screw by hand without binding.

[cncsnw]

If I read that right, the slave axis moved the correct distance, as near as could be measured; and the master axis moved too far.

If an axis is moving too far in an open-loop step-and-direction control system, I would speculate that noise on the step command wire is introducing extra steps.

Since, in your test move, each motor is supposed to turn exactly 10.0 revolutions, it should be simple enough to mark the motor shafts so you can see whether a motor moved too far, or not far enough.

You could also do the same test moves with the motors uncoupled from the rack, to see whether load or binding is causing the issue. If it is an EMI interference issue, you will still get the wrong amount of movement on one or both motors, even with no load.

[ShawnM]

There are certainly several ways to trouble shoot this as already posted. There could be mechanical issues and/or electrical issues. I also noticed from your Google album that none of your control wires are shielded. This is not "best practice" when wiring a control panel in such a noisy environment.

Uncouple the motors from the rack and check for smooth, non binding movement over the full length of travel on all axis.

You also never stated how you "drive back to 0,0,0". Please explain. The fact that you are tripping the home switches when doing this is an issue considering your soft limits are setup. We need to know how you are doing this exactly.

Please post another current report.

Those are 5v stepper drives, wiring them to 24 volts and then adding the resistor is the same as leaving them at 5V without a resistor. That wont change anything.

[vincent-visser]

Hello All,
We were 2 weeks off with Christmans and New Year so it's been a while.

We disconected the pinion from the rack , marked the position, zeroed XY and did the same MDI G1Y314,3007F300. (exactly 10.0 revolutions of the steppermotor). In normal direction the motor did exactly 10.0 revs but in reverse direction, if we do MDI G1Y0F300 the motor did approx 9,75 revs. So this is the problem.
This also explains why Y slave and Ymaster have a different deviation , 1 in positive and 1 in negative direction, because Slave is mirrored.
We also noticed that the steppers make more sound when driving in reverse direction.
We changed steps per revolutions to 200(no microsteps) and now the problem is also visible, every few secondes the steppers locks for a short period (like 0,25sec or something) and the turns further, it sounds like electrical lock and not that the power is completely gone cause is still makes noisy when locked.
This problem is the same for X Y-master and Y-slave. Z axis has no problems.

Only difference between Z and the other motors is the OverallTurnRatio. For Z is 5mm/rev and others 31.43007mm/rev.

Then we measured the voltage on the Dir cables between Acorn6 and stepper drivers. With an feedrate of 380mm/min it's between 23.5 and 24.0Volts. At this speed we don't haven problems with loosing steps and the motors make same sound in both directions.
When feedrate drops beneath 380mm/min the problems comes up and we measure lower volts al the way to something like 5.0volts with 30mm/min.

So a lot of things are clear right now, but not the cause of the problem haha.
no mechanical issue, no problems in normal direction, no problems at higher feedrates. Only problems at low feedrate in rev. direction.

Is it correct that Voltage drops at lower speeds? What can cause the stepper to lock if the voltage on Dir. is lower then 23,0V?
I'll try to add a movie for the sounds in both directions.

Thanks,
Vincent