CNC Conversion Woes!

[LacL]

I have always used the same speed and accel values for all 3 motors. I was not interested in pushing each drive to max so I just left the values the same.

Its clearly loosing steps during dir/accel changes but still unclear why. Some things I'm investigating are:

1) win 7 lpt port issues. I'm going to shut down as many processes as possible and see if that makes a difference.
2) grounding issues and/or noise on the lines. Both the PC and CNC are connected through a power surge protector.

Is anyone else using G203V drivers? I never adjusted the trimpot on these since the motors are quiet when idle and never get very hot even after full test runs. Is it necessary to adjust them for proper operation?

[arizonavideo]

I have the 203 and you must adjust the trim pots to counteract the ring that all steppers have.

Follow the instructions or just changes speeds until the motors are loudest and adjust the pot until they quiet down. You can hear the ring fairly easy.

[LongRat]

In my experience it will be pretty obvious when the machine is running if any steps are being lost. The sound of the motors will change and it will be pretty unpleasant.

[LacL]

Thanks for the input arizonavideo. Right now I'm rewiring the grounds so they all meet at one point. I had the 203s gnd daisy chained. I will tune the 203s once I get that done.

Someone in the MACH forum suggested connecting the grounds to earth ground essentially bypassing my surge protector. Is that a good idea?

@ LongRat, I have been able make the motors loose steps when I turn my MPG very fast so I am familiar with the sound, but never heard the same sound during the cutting. Then again I'm still getting familiar with all things CNC so what may sound normal to me may in fact not be.

[Fastest1]

Thanks for the input arizonavideo. Right now I'm rewiring the grounds so they all meet at one point. I had the 203s gnd daisy chained. I will tune the 203s once I get that done.

Someone in the MACH forum suggested connecting the grounds to earth ground essentially bypassing my surge protector. Is that a good idea?

@ LongRat, I have been able make the motors loose steps when I turn my MPG very fast so I am familiar with the sound, but never heard the same sound during the cutting. Then again I'm still getting familiar with all things CNC so what may sound normal to me may in fact not be.

This is a common and accepted practice, known as "a star ground".

[photomankc]

What are you using to make the GCode? I would doubt there is an electrical noise issue that is causing you to cut the exact same undersize each time, nor would I see it being lost steps. Seems like something is off in the production of the code to get a repeatable error like that. What software did you use to go from drawing to CAM to GCode.

The min/max X values I see are:
-0.3968 / -3.214

That's a difference of 2.817" between the tool paths then subtract the end mill that's 2.6922" of material between the cuts. Is that the X extents dimension of the part? If not then your tool path isn't right. Lets start there. What is the part supposed to measure in the X direction and what does it actually measure?

[Mad Welder]

The part which is a square is 45 mils too small on both x and y once cut.....every time..

Hi there Lacl, I normally run for help to the guy's here too as I’m not at this long either but guy's is there any chance that the Inches/MM settings are getting mixed up here as all calibration measurements are referred to in US imperial (inches) but the repeatable error is suggested as 45 mils (mm's) or is mils another aphorism for a commonly used imperial measurement term?

[RTTIM]

Hi there Lacl, I normally run for help to the guy's here too as I’m not at this long either but guy's is there any chance that the Inches/MM settings are getting mixed up here as all calibration measurements are referred to in US imperial (inches) but the repeatable error is suggested as 45 mils (mm's) or is mils another aphorism for a commonly used imperial measurement term?

Eoin,

The "mil" is a milli-inch = 0.001". It's a commonly use term in the US....so he's saying the part is 0.045" too small.

[LacL]

Additional things I tried, but no change in results.

1) I ran MACH3 in XP SP3 mode.
2) I turned off everything I could think of in windows including firewall and virus checker (after disconnecting from the network of course).
3) Wired GNDs to one junction (star ground).
4) Tuned the 203 drivers as per manual.
5) Made a small offering of blood when I accidentally scratched my hand on the end mill.

The min/max X values I see are:
-0.3968 / -3.214

That's a difference of 2.817" between the tool paths then subtract the end mill that's 2.6922" of material between the cuts. Is that the X extents dimension of the part? If not then your tool path isn't right. Lets start there. What is the part supposed to measure in the X direction and what does it actually measure?

This was the first thing I verified when I first noticed the problem. The values you posted are not the extents. If you scan through the file, you'll notice the min/max values are -0.3968/-3.6032. That's a difference of 3.2064" minus tool dia of 0.125" and that should leave a part that is 3.0814". Actual part should be 3.081" and I consistently get 3.036" +/-0.002

The values on the MACH3 DROs match those in the file during the test runs at the extents.

[hoss2006]

What version of Mach 3 are you running? The fact that you said the axis calibration wasn't working points to the problem.
The dros could match the gcode as it's running but the machine isn't physically moving the same as the dros.
Try the newest lockdown version and do the again.
Hoss

[Fixittt]

here is a suggestion, in ports and pins, enable sherline mode. Then try again.

[Fastest1]

Have you gone thru all of the Mach optimization steps?