My CNC is now losing steps, and think its due to EMF

I built a new cabinet and installed a new computer system for the CNC.

Now i'm losing steps. I'm thinking of pulling everything and lining the back with tinfoil
and installing some sort of a divider plate between the CPU and CNC electronics.

I had no issues with the old setup.

Thoughts?

EMF can be tricky to resolve "if that is truly the culprit", but in general, the trick is to find the cause, reduce / contain / eliminate it.
I would suggest simply  boxing the pc & the complete  controller which can be simply a cardboard box lined with Al foil. No need to take things all
apart and then try.

One can "sniff" to find the culprit using a small battery operated radio and just scanning around using the antenna as the noise will create
interference on the radio when near the field. DO NOT actually touch  anything with the antenna ( place some electrical tape over the end of the antenna).

Above is  about as simple as I could explain things and better understanding of EMF  would be a book.

If you truly had no noise  problem before then effectively duplicate what you had before on the new cabinet!

RICH

Most typical noise issues are on the E-stop or limit switches lines. Based on the photo I suspect the E-stop wire that is looped around the back of the monitor, (inherently emits high emf & generated voltage of over 300 volts to drive the screen). Or If you have a separate monitor try to unplug the one in the cabinet and use one off to the side to see if that monitor is an issue. But id suspect it is something simple. Not like you changed anything in the computer connections. Just rearranged things.

Thanks for the ideas.
I'll test out a couple of things (I'll have to find an old radio)
Being as xmas just past, I have a few spare boxes now

Oh, I already have small caps on the stop and limit switches now, can't remember the size though.

Oh, I already have small caps on the stop and limit switches now, can't remember the size though.

They need to be 0.1uF

Tweakie.

Mogal,
I hate to say it but you have pretty much broken all the rules regarding electrical noise. First is non-metallic backplane, that doesn't provide a good ground plane. Second the signal cables should all be shielded cables. Third all components should be oriented so that signal and power wires can be routed away from each other. Mounting all 3 drives vertical would have helped routing, and then the heat sinks would work properly. Vertical heatsinks generate a flow of cooling air as warmed air rises. Experience will let you recognize a mistake, when you make it again.

Sigh, newbie errors I know 

The CNC electronics are on an aluminium backplate, and the computer MoBo and components are still mounted on the 'case'
If I can't soon find and fix the issue, I will completely redo the system (after learning more on EMF)

But! I think I found my issue.
I went to load in a test file, so moused up to 'File' and noticed the very top corner of the screen was 'vibrating'
I didn't notice that before. I uncoiled the power cable and wrapped it around the monitor instead.

The screen isn't vibrating any more, and my test file was within 0.0005 on all axis after a verifying.
I'm currently running a new program and will report back on what happens.

Thanks,
Chris

Okay, delay last, I'm still losing steps.
After a 30 minute run, the X and Y axis are within 0.001" and Z was 0.020" out from a verify.

K, back to the drawing board.

Well, here's a weird one.

I had my pulse freq set at 35,000. I'm loosing steps
I switch to 25,000 and I'm not losing steps (Verify shows approx 0.0005" - acceptable margin of error?)

BUT!

I have my steppers set at a speed of 140ipm.
The DRO reads out 152ipm when I rapid move.

Turns out the frequency is actually 27,300 pulses (Not 25,000)

So my question is, is there any way to fix this?
I'f I'm not losing steps, and my accuracy remains, can I assume all is well?

I think I've also narrowed down my interference to the relay controlling the router.
I also sucked it up and bought some higher end shielded wires and power cable for the router.