[ozzysmarie]

I build a 4x8 ft cnc router using the cnc router parts system. The control system is KL-4030 set to 1/64 microstep, the power supply is 36V 9.7Amps, with a c10 breakout board. The problem i am having is that my Y axis is bouncing all over the place. The cuts are varying between 0.030 over and 0.25 under . I dont believe the problem is mechanical because if i cut over the same space the accuracy still changes. some cuts are with in +/-0.005 and then the next cut will be way off again. Im using shielded cable grounded at one end. all of controllers are grounded as well. The machine is grounded to a steel I beam in the building. The steppers are KL 23H286-20-8B with 425 oz-in torque. ive switched controllers between my z and y thinking it was a bad controller but it still happens.

Im currently at a loss as to what is causing this problem.

[ger21]

What are your velocity and acceleration settings?
Cut them in half and see if it helps.
In most cases, lost steps are simple due to trying to go faster than your motors are capable of. The motors you have would run about twice as fast on a 72V power supply.
I'd also recommend changing the microstepping to 1/8.

[ozzysmarie]

My velocity is 100 ipm with an acceleration of 3. i can understand how a low voltage could account for lost steps, but how would it cause a cut to be longer?

[CarveOne]

I'm with Gerry. I think most of us use around 8 to 10 microsteps depending on motors and drivers, or in some cases, less than that. If your driver can handle 50vdc, use a 48vdc supply. If it can handle more voltage, use the next lower voltage standard power supply.

Running on 36vdc your motors won't develop full torque and they probably aren't even able to respond to those very short pulse widths per microstep reliably.

[ger21]

With an accel of 3, you shouldn't be seeing any issues due to low power.
Change your microstepping to 1/8 and see if it helps.
Also, what do you have the step and dir pulse widths set to on the motor tuning page? Try setting them to 10 or 15.

[ozzysmarie]

This afternoon i was able to play around with the machine some more. I reduced the velocity to 10 and the accel to 1, i ran 6 tests just moving the y axis a known distance and measure with calipers, rinse and repeat. Every test was within 0.005 or less. The motors i am using are the KL23H286-20-8B. I currently have them wired in bipolar serial mode this is listed at 27.2 mH. if i rewire them to be in bipolar parallel mode at 6.8 mH would i see a benefit? (my controllers can handle the increase in amperage) My understanding is that my required maximum voltage would decrease. I think my pulse widths are set to 9 currently. But i will change that tomorrow. If the general opinion is that i need to increase my voltage, What PS/driver setup would you recommend?

[ger21]

Definitely run them bipolar parallel, and you'll see increased performance.
And again, change the microstepping to 1/8.

You might be having resonance issues.

You should be able to run them at several hundred ipm with an accel of at least 10-20 I would think.

[ozzysmarie]

I was able to rewire the stepper into bipolar parallel, changed my controller to 1/8, changed the pulse width to 10. After some tests im still seeing a positional error of +/-0.015 more often im gaining distance than being under. Does this still appear to be a voltage issue? Is it possible Im having signal issues from the computer itself? Are there any possible measurements that i could take to better narrow the possible cause. If it is a low voltage issue what power supply/controller setup would you recommend.

[ger21]

I would say that no, it's not a low voltage issue. That would only show up at higher speeds.
What velocity and accel are you testing with?
What kernel speed do you have Mach3 set at?
What are the specs of the PC you're using?

[ozzysmarie]

i was running it at 100 IPM with a accel of 10. The computer is a 2.6 Ghz single core with 4 gigs of ram, running windows xp. I havent changed the kernel speed of mach3. I think the default is 25K.

[ger21]

If you run Drivertest.exe in the Mach3 folder, do you get a very smooth flat line, with no spikes?

[ger21]

Can you put a mark on the stepper shaft when it's at 0, and then move around, and go back to 0 and see if the mark is off?
I'm wondering if the error you're seeing is mechanical? Have you checked the setscrews on the pinion? Loose setscrews have caused similar issues for a LOT of people.

[ozzysmarie]

I ran the driver test. I'm not sure what qualifies as smooth, but to me this seems rough. What would the effect of this be on my machine? And if this could be causing the problem, what is the fix.

[ger21]

The line does seem a bit rough, and could possibly be an issue.
I've attached an XP optimization guide for Mach3. If you have a clean install of XP, then using Standard PC instead of ACPI will usually make a large improvement, but results can vary greatly between PC's.
Note that this may not make any difference at all.
Tracking down issues like these can be very difficult, as there are many things that can cause the issues you're seeing.