How well is your machine grounded?
I remember someone posting that their wall outlet ground was bad causing an issue.
1 Like
Having eliminated most noise sources, it has to be the stepper, wiring, or controller. If its changing which stepper then that kind of rules out stepper and wiring unless its intermittent and all of the ones are bad which I doubt. That would leave only one item left, the controller. I would double check that your outlet is actually grounded etc otherwise thats what direction I would head but maybe Redline will reach out on here they seem pretty active.
Edit: I dont see how it can be the PP/Programming but that said still happy to run the file myself and you can try it out.
1 Like
Its all brand new wiring where the panel goes to a standard ground rod. Solid copper wire from the machine points to the ground.
1 Like
Sometimes solving a problem means ruling out what it isn’t.
Make a checklist of things you have done/not done and or eliminated.
So ground from panel to outlet is good.
Now check for continuity of grounded elements to outlet using a multimeter. Then you have ruled out all grounding of equipment.
1 Like
I completely agree with that. I tested the continuity all over the machine back to where it grounds in the outlet. Theres solid ground all over. Mind you the black powder coat doesnt work. However everything is connected through a bolt which works. From the back of the pc to the y rail from the y rail to the outlet, From the dust extraction hose to the spindle mount and again where the y rail goes to the outlet ground.
2 Likes
The new smaller test file again loosing steps on the Z axis. Also you cant turn the spindle control box completely off as it throws an alarm. You can turn the spindle itself off though which Ive tried and didnt help.
1 Like
I assume this kills the power along the spindles power cord which is all I was really going for incase there was EMI from it.
I dont have the Redline so I dont know how detailed it is but when you load the file does the little 3d preview look correct? Once again I dont have the controller so I dont know how much you can zoom in or tell on there if the preview would show a shift in the axis
1 Like
The instances where its being ran and tested for lost steps, there is no alarms that happen. It simply just progressively gets deeper the longer it runs. On the full 3d relief file thats about 8in across it dropped 3/4in into what would be my spoil board. Completely loosing steps and position with no alarms from the closed loop system.
Previously I was asked to try it with the spindle off. I can turn the spindle off but not the spindle control box because that throws an alarm, spindle error. I mentioned this because its still a source of power nearby even though the spindle was off. The controller is still on.
The 3d preview looks correct every time in any of the files. Even in 3rd party software.
The next test I was asked to try is to swap motors. So I did that and its running now. In the dark and the spindle turned off. No need for extraction because its not cutting anything really. Now we’ll see if the problem follows the motor or stays with the Z axis.
1 Like
So after switching the two motors and running the smaller test file which is only 3.5in square while cutting nothing. Switched the braking z with a y motor. The x axis seems to have lost 1 step. The y lost 5 steps. And the z 2 steps. Its better but still showing some loss. Each step being .005in. Next I’ll try the longer job again to compare that with previous attempts.
1 Like
If all axis are loosing steps I still feel like the common here is the controller. If it followed a single motor or a single wire then so be it but it seems like its all over.
3 Likes
I did see one or two off previously as well. Fairly negotiable. Anything more than that is definitely worth noting. Especially when it totals 3/4in. Either way i have to wait on support. I can run tests faster than I get replies. Also if nobody else is having issues then its not the PP.
1 Like
How are you determining the recorded step loss when not machining?
I work in metric, and if my X axis is set at 2000steps/revolution, which is 2000 steps/16 mm (ball screw pitch)…
then that is 125 steps/mm which means 1 step is equal to (1/125) or 0.008mm.
Above you mention each step loss equivalent to 0.005”, or 0.127mm.
If using the Redline controller, what are the motor resolutions set to?
4 Likes
You are correct in that the motors should be 2000 steps per revolution. There really isnt a good way to measure that. What i did was note where the ball screws set screw was and put a line there. Zeroed the machine. Ran the file. Sent it back to zero. Then used the incremental jog to get back where it should be. Each jog increment is .005in. Theres probably a different number of actual steps per jog increment but thats what I did anyways. It definitely does not take pressing the button 2000 times to make one revolution of the ball screw.
1 Like
And your motor setting is 125steps/mm (X and Y) and 200steps/mm (Z)?
Correct, step signals and jog presses will not have a 1:1 ratio.
If money and interest permit, I would invest in a dial indicator of reasonable quality, as it would help you be more accurate in determining movements on the scale we are discussing, but it can be useful for determining what degree of backlash you have on your axes. it can also be used to measure homing repeatability and much more again if your interests take you in those directions.
Do you own calipers? They can be useful for checking the accuracy of your machined parts, and in your current situation might give clues as to what is happening if you ran various test files designed specifically for checking accuracy and tolerances of the final machined part.
1 Like
Im not even trying to get that accurate. I just wanted to make some parts. The x and y are set to defaults 3175 steps per in. Z is set to default 5080 steps per in. The problem is it cant keep track and is drifting for some unknown reason at this point. Yes i own calipers and micrometers. If the square its told to make comes out a diagonal theyre not going to help though. The code says make a square.
1 Like
Here’s a checklist for out possible causes
- Mechanical impediments: Physical obstructions, dirt, grime, or a lack of lubrication on the machine’s rails or leadscrews can prevent an axis from moving freely.
- Overloading: The motor may be overloaded by too aggressive cuts or a toolpath that is too demanding for the setup.
- Incorrect settings:
- Acceleration and velocity: Setting the acceleration or velocity too high for the motor can cause it to lose steps as it tries to ramp up or reach maximum speed too quickly.
- Motor current: If the motor current is set too low, the motor may not have enough torque to overcome inertia and friction.*
- Electrical issues:
- Loose connections: Check all connections to the motor, drivers, and limit switches.
- Poor wiring: Unshielded wires or faulty wiring can cause electrical interference that disrupts the signal to the motor.*
Software/Controller problems: Issues with the CNC software, controller, or driver board can cause unpredictable behavior.
Troubleshooting and solutions
- Clean and lubricate the machine: Ensure all mechanical components are clean and properly lubricated.
- Slow down your cuts: Reduce the feed rate and/or take lighter cuts to reduce the load on the motor.
- Adjust acceleration settings: Increase the acceleration setting gradually, if possible, or decrease it if you are already at the maximum. A smoother ramp-up is more likely to prevent step loss.
- Check and secure all connections: Inspect all cables and connectors for the motors, drivers, and limit switches. Ensure they are secure and properly seated.
- Verify motor and controller settings:
- Steps per unit: Calibrate your machine to ensure the steps per unit in the software are accurate for each axis.
- Motor current: Ensure the motor current is set to the correct value for your motors. It should be high enough to provide adequate torque but not so high as to cause overheating.
- Inspect for physical obstructions: Check that nothing is physically interfering with the movement of the machine, such as a hose, cable, or a piece of debris caught in a belt or pulley.
- Consider using shielded wire: If you are experiencing electrical interference, switch to using shielded cables for your motor and limit switch wiring.
- Check the Z-axis: If only the Z-axis is losing steps, it may be because it is undersized or has a different configuration compared to the X and Y axes.
Software & Configuration Issues
Incorrect software settings can command the machine to move in ways its hardware cannot physically handle.
- Incorrect acceleration and speed settings: Trying to accelerate or move too fast can cause the motor to lose synchronization.
- Solution: Lower the maximum velocity and acceleration settings in your CNC control software (e.g., GRBL, Mach3).
- Mismatched steps per unit: The software’s calculation of how many steps equal a unit of distance (mm or inch) may be incorrect.
- Solution: Recalibrate the steps per unit for each axis to ensure accuracy.
- Driver/controller signal mismatch: The pulse width timing from the controller might be too short for the specific motor drivers being used.
- Solution: Consult the driver datasheet and adjust the pulse width settings in your control software
1 Like
Sorry for speaking up but, isnt the purpose of closed loop steppers to prevent mis steps? and if it does misstep isnt it supposed to stop or error out? or am i confused?
5 Likes
My guess is the unshielded wiring. Support says well it must be emi and no hobby machine has shielded cables. Which is false…even chinas genmitsu cnc now has them. Its a standard and has been for over a decade. Ive built machines out of wood, black pipe and skate bearings and still used sheilded cables almost 15yrs ago… Im not a fan of throwing away money at software that doesn’t fix the problem or tearing the machine apart numerous times just to try something else. Itll never go back 100% and opens the door for more errors.
1 Like