Setting up a plasma cutter and looking for some help.

Birdman3131 wrote: As for paying attention to that part it has been one of the bigger headaches of this ordeal.

I am glad you did, no one (including me) has noticed it before.

Ended up using a not instead cause I was not understanding your code at the time although I am fairly sure I know where I messed up and could make it work now

Rather than using a not component you could have used: But, in the end they achieve the same result.

EDIT: Its not very clear that it is units per second not minute. Seems to be working now. Now to figure out how to cut something.

So one rewire later I have the ohmic probe working mostly. (Pull low for the relays it seems.)

However now I have an issue with Z. I had set its max accel and velocity when setting it up. Its a stepper so needs more care. However it seems to ignore the values. the jog value in plasmac is at 117 in/min and it will gladly overdrive it. (60 is closer to right.) The ini file has it was lower even but it does not seem to be effecting it.

would need to see the full ini, but check your overall max velocity. It may be capped below your max for the individual axis’

Well I finally cut a thing. Got 2 issues.

1. Firing up the plasma trips the limit switches. I ran it with them unplugged and it worked. I can likely fix this but its still odd as they are normally open and should not be that susceptible to noise. But I will try a couple things and see what I can do.

2. The bigger issue was some jerking while cutting. This occurs in dry run as well so i likely have to tweak my motor settings. You can see it on the wrench here. Not sure yet what it is. Im gonna look at backlash and mess with the motor accel settings.

the edge of the wrench

Tripping the switches is a sure sine you have problems with grounding/shielding.
On the phone so i can not do much, but do a search for proper grounding, there is some very useful info somewhere on linuxcnc pages.

I am rewiring them to NC rather than NO to see if it helps. (Assuming noise on a nc won't trigger it whereas I can see noise on a NO triggering the input if it induces a current along the wire.)

It does not help that they are unshielded wire. (Everything else is shielded. The wires to the limits are the only ones not and I reused the old wires.) If that does not work ill go through the grounds a bit more. If nothing else I have the shielded wire to re run them. Just would rather not.

I have seen the grounding doc you mention. I should be able to find it again if needed.

Just be aware that you are risking burning something to a crisp, the amount of induced field when firing the torch is staggering.

If you feel you've done all that you can to mitigate the noise problem on the limit switches. You could add a debounce to the limit inputs in your hal file. It won't get rid of the noise or induced voltages on the input circuits (or the potential damage they could cause) but it could successfully filter the signal within Linuxcnc.

What voltage are you using through your limit switches? There is a big difference with noise immunity with 24 volt v's 5 volt.

I would go back to Clearpath about the drive tuning. Clearly something is not right if you are using shielded cable

Also make sure the shield on any cables is only connected to GND at one end.

So been a fun couple days. I keep breaking things. Have not fired the plasma side of things up again.

Zhome/breakaway switch died. Not sure if it got zapped from the plasma or when I knocked a pair of 24v wires loose. it has been replaced by a normal mechanical switch.

I believe the limits were on the 12v rail at the time. I could switch them to 24 but have not atm. I have switched them to NC though. (Also I forgot to mention but i got rid of the paraport board a while back as I ended up having sufficient inputs on the mesa board. Think I have 1 input free now.)


I don't believe it is an issue with the clearpath motors. Fairly sure it is a mechanical issue involving the rack and pinion gear system. I had noticed mild quirks moving it by hand but had ignored them as it is easier to be willing fix those if we know the machine works. The actual mechanicals are from the early 80's I believe. If I get a chance monday ill get a better video of this to show.

I will probably have the oscilloscope recording next time I fire it so as to verify any spikes or lack thereof. Its rated for 400V P-P on the high voltage inputs so should be safe for it unless the voltages go way above what the max torch voltage is.

Most of my time was either on other projects or swapping the axises on it. I had set it up the way it was labeled but according to my coworkers it was a stupid setup and required flipping the sheet in autocad. And if that is gonna be done now is the time to do it. Had it mostly figured out and then managed to find a specific route that broke one of the jack chains that was supporting the plasma lines from the ceiling. Everything is ok but I have to re hang the lines to they don't drag.

So on the grounding side of things I do need to go back over all that but have not as too much other stuff has sidetracked me.

We are off friday so ill get back to it monday