[murilosouza.ms]

Hello guys!

I´ve made a routing machine wich get a PMDX-125 breakout board with 3 CNC AC Servo Drive ASDA-B on it and I´m using the Mach3 software.

Well, everything is working fine, but when I´m doing some work with the machine, I get some mistakes on measures.

For exemple, I do a mirrored part on the same drawing and the final product gets almost 1mm of differecence. I can see on the CNC code that is everything ok, but on the machined part I get this difference.

What can be the problem? I guess that my xml for Mach3 is ok, but I think that I got some noise for the board and drives and I don´t know how to eliminate that problem.

I hope you guys help me!

[frog]

Check your Mach3 configuration, your encoders may need fine tunned.

[murilosouza.ms]

What you mean with this? Wich configurations on my encoders you think could make this problems to ocurr?

[H.O]

It can be a lot of things. I'd start by trying to narrow it down a bit. Finding it while running complicated production parts can be very hard.

Setup a test where you run one axis back and forth a certain distance (100mm?) a certain amount of times (100 times?). Set up a dial indicator and whatch if it's loosing a little bit each time it "comes back" to the indicator.

Here's a few things that could cause it:

• You have the step-polarity setup wrong, (Mach3 set to active low while the drive steps on the rising edge)
• Too low value for the direction prechange (check drive docs for timing requirements and check BOB docs for any delays it might introduce to the signals)
• Too low value for the step-pulse width, (again check drive docs etc)
• Noise on the encoder lines, (properly terminated, shielded, twisted pair cables etc?)
• Noise on the step-lines. (Are the inputs on the drive opto-isolated?)
• Encoder not fixed to motorshaft properly.
• Faulty/flaky encoder (like the USDigital E7P...aaargh....)

It can be a real PITA to find so you need to work logically and try to narrow it down as much as possible.

Good luck!
/Henrik.

[mvd535]

I just finish my new home made CNC routing & milling machine. Travel capacity; X=500, Y=500 and Z=200. My hardware configuration was got it from Keling Inc. Package is:
Stepper motors: KL 23H276-28-4B.
Power supply: KL-350-48 48V/7.3A
Drivers: KL6050
Question: Even everything is working fine in terms that the three axis are moving OK, I have a very weird situation. Two of the three stepper motors, one more than the other, are unexpectly turning some steps (about 40 to 60) every 10 or 15 seconds. This is ocurring even when I do not have the Mach3 software sending any pulses. Worst, its happen when the Mach3 is not running at all. Why these motors are attempting to move without any instruction. Anyone knows why??. Thanks.

[H.O]

Does it do it even if you disconnect the cable from the PC? If not you you have some software on the PC accessing the LPT-port - a hardware dongle driver or a flakey printer or scanner driver etc.

Uninstall everything you don't need and turn off everything in the startup tab of MSCONFIG. If it doesn't help re-installing Windows is your best bet but I know it's a PITA.

[mvd535]

Does it do it even if you disconnect the cable from the PC? If not you you have some software on the PC accessing the LPT-port - a hardware dongle driver or a flakey printer or scanner driver etc.

Uninstall everything you don't need and turn off everything in the startup tab of MSCONFIG. If it doesn't help re-installing Windows is your best bet but I know it's a PITA.

Henrik,
Many thanks for your quick reply.
Actually you are totally right, its must be something with my PC. I found this out when, last night, by accident, I change the turning off sequence of my system, and I turned my PC off before my drivers and breakout board system, and nothing happen with the stepper motors !!!. So, It is very clear that some software in my PC is sending out unexpected information through my parallel port all the time. I will fix that in my PC. Many thanks because, this morning, I can confirm that with your appreciated answer.
Mario