Interference Problems

[gaton12]

Hello!
New here, just created my account so excuse me if I am in the wrong section.

I am currently building a CNC plasma cutter using a Neje laser master 2s max for the frame. Now it is a pretty weak frame but seems to work alright for its price.
I have bought a cheap (I guess this was my wrongdoing) handheld plasma cutter and have attached it to the CNC table and hooked up the signal wire to a relay controlling the plasma cutter.
When I was using the board that was included with the Neje CNC I could successfully cut things with a maximum of 6 corners, more than 6 and the CNC would just stop without any errors in the console, I assume this has something to do with a buffer. I tried grounding things like madman with no success. I could even place the CNC outside at a 10 meter distance from the cutter and the same problem still occurred.
So after this I thought maybe the board is very susceptible to interference (it has a bluetooth antenna on it which I thought maybe were the culprit) so I have bought an Arduino with a CNC shield and I am running GRBL. The steppers move fine and I can now control the CNC/arduino with ease, however, It is still the same problem that as soon as I run the plasma cutter, the CNC just stops. My latest change was changing all the cables to shielded cables but still with no success.

Im 2/3 done with my education as an electrical engineer so this is incredible frustrating, I should be good at this stuff dammit!
Any help is appreciated, I am lost.

[WyoGreen]

What program are you using to generate your cut files?

[acourtjester]

In the past you could use Demo software but it was limited to about 250 lines of code. Does your system have a display showing the G-code as it is running. Or you could count the lines to see when it stops. Are you only using one cut operation or different ones. Does you system use an Arc Ok signal, you plasma cuter may be running out of air and dropping the Plasma, which stops the table. I does sound odd to stop after 6 corners, have you tried cutting circles.

[adbuch]

--- so I have bought an Arduino with a CNC shield and I am running GRBL. The steppers move fine and I can now control the CNC/arduino with ease, however, It is still the same problem that as soon as I run the plasma cutter, the CNC just stops. My latest change was changing all the cables to shielded cables but still with no success.

Welcome fellow EE. What software are you using to feed your G-code to the Grbl/Arduino? And what software did you use to create your drawing and export the G-code from? I would try doing a "dry run" with the torch off and see if it still stops after 6 corners. If it does, then this could possibly be due to some limitation with the trial version of your software. I know that the Mach3 trial version has a limitation, as do probably some of the others.

I have a small DIY CNC router/plasma cutting table that uses NEMA 23 stepper motors on all axis. It is controlled with the Grbl/Arduino setup, and I use Candle software to send the G-code file to the Arduino. I don't believe there is any limitation with Candle as to how many lines of G-code it can handle. I have generally exported my G-code from Fusion 360 and have never encountered any limitation as to the file size.

I will look forward to knowing exactly what software you are using.

David

[gaton12]

I will look forward to knowing exactly what software you are using.

David

First off, thank you all for replies. At the moment Im running the paid version of Lightburn (which I think is designed for laser-cutters but hey, they are pretty close right).
If I run the GCODE without the plasma cutter connected or just turn it off then it runs just fine and finishes the job, it is only when then cutter is connected that it stops.

Does your system have a display showing the G-code as it is running. Or you could count the lines to see when it stops. Are you only using one cut operation or different ones. Does you system use an Arc Ok signal ... have you tried cutting circles.

I dont have display other than a console in the software (Lightburn) where I can see the GCODE being sent. Only one cut operation at the moment. No Arc Ok signal. And when cutting circles it finishes only about 30% of the circle then stops.
However!, the being able to only do six corners problem I had only occurred when using the board that came with the laser cutter. After changing to the Arduino board it instantly stops the second the plasma cutter turns on instead, so no six corners.

In my latest test I changed my USB-cable to one that is more beefy (stole it from my microphone) and that seemed to have improved the distance I could bring the table and cutter to each other. Perhaps that is where the interference is coming from? Im thinking about buying the best shielded USB I can find or is there something else I could do to shield it?

Again thank you guys for the replies and Im sorry if I have butchered the quotes function.

[acourtjester]

You may be fighting a few problems, (1)your plasma unit may be sending out noise. (2) I'm not sure is Arduino has software that will run plasma, but the main differences is the Torch fire code. Don't really think that will a problem, the plasma noise may be doing that. I know there are member here that use Lightburn for projects but again no sure if they are plasma type. If your Plasma unit is causing the problems you will need to change to a BlowBack type ignition unit.
With plasma the Arc OK is used to tell the software the plasma is operational, if you lose the arc it stops the software. Normally this would happen if you move off the metal you are cutting.

[weldguy]

Definitely sounds like high frequency interference from the plasma cutter and you could fight that until the end of time with your setup. If you are happy with your motion controls and want to get cutting things I would get a new plasma cutter with a blowback type start system to eliminate the hf interference.

[tcaudle]

You are fighting multiple problems :
Cheap import plasm acutters are NOISY . There are two types of noise . Radiated (though the air ) and conducted (though common connections.
The workclamp on the plasma is NOT ground, its positive (noisy ) votage0 If the table is not properly tied to some form os noise "drain" it makes the whole thing a noise antenna. Note I did not say a "ground rod" Its not to establish an AC electrical ground its to shunt noise to a low impedance media (the earth). So you deal witht he transmitter first (the plasma) and try to get its commissions lowered. Now, look at your setup and see if any wires connected to the table also connect to to your electronics ....like circuit grounds. Bad things Man! You need isolation on inputs and outputs . I don't know how you are reading (if you are) the voltage from the torch but if you are using a divider it needs to be AT the plasma and filtered and not pump high voltage noise around the shop. Any switches or inputs need to be via opto isolators with noise filtering.
Radiated noise is pretty easy to fix. You move things appart because air is a poor transmitter of HF noise . YOu use ferrite cable chokes on signal wires . You make sure the table common is not connected to the circuit common except though high impedance . Never connect circuit aground to the table structure . You poor little Aurdrino does not take much noise to just overwhelm it and lock the program

Conducted noise is harder and takes isolation and careful wiring .

I have seen where just disconnecting the table from a building ground (same \ground shared by the PC ) and giving it its own "Noise Anode" locally stopped noise issues. I have seen where users though it a good idea to mount the controller and pC in a metal cabinet tied tot he table and wonder why it did strange things . True STAR grounding takes everything int he systems out to one central ground and with more or less equal ground lengths. Thats usually not practical so the isolation is the next best thing

Dont forget that two grounds 6 ft apart can be at different potentials If your electrical safety ground is 20ft from your local grounds there can be volts of difference.


Strange things happen when you pumps high current at chopped frequencies through wires close to each other! Good Luck