I’m working on a plasma cutter to update my mpcnc capability. Shielded and isolated the control box with a filter and ferrite cores as well as grounded all the components now. Last week the cutter came, I tested everything and manually it works. The Plasma Cutter has a HF start, NOT a pilot arc.
That means, after the start and end cut settings, that open and closes the relais, I need to tell the machine to move to the cutting height. Now I can’t invert the toolpath depth in estlcam, it only accepts positive values.
How others solved this, that the head is touching off on AZ0 and then goes up, to let’s say Z-5?
I could just hack my way around this, maybe manually edit the gcode so it works, but if my part has inside futures, the probe would need to touch off again to start the arc. Same applies if I invert the Z axis, the first cutout would work, but then, the arc just drags on. Besides, I would need to edit every operation a start and end cut.
I seam not to find this answer in the how-to’s online so far. As most use the pilot arc or just not mention this point.
i found the laser/plasma setup tap in estlcam, so the start and end part is taken care of.
the other point is that i’m thinking of inverting Z axies, use the start level as 0 and cutting depth is then the optimal cutting distance. then setting in the tool libary a high cutting depth so it does not need to step up to it, and also, a fast z travel (maximum that is possible).
I wouldn’t invert Z direction. Estlcam is still very useful if you set it up correctly. Estlcam has a setting for how far it will lift the Z above the origin for travel moves. The idea is that you will set your origin to be the top of your work piece. If you invert your Z then it will lower your cutter for these fast moves and that isn’t what you want.
Instead here is what I do. I am using a laser that has the focus set to 65mm away from the lens. So when I start a job I need to be very close to that distance. What I do is square my machine by pushing/sliding the gantry into the front left corner and pressing home all. Then I use controls to move to where my origin X,Y will be on my work piece. Then I home Z. This could be done with a probe and an offset but I just use a custom plastic spacer I made I move the z until it is in the correct position. For me that is 65mm above the work piece. For you it may be 5mm I don’t know. Then I run my GCode.
My GCode starts with a G92 x0 y0 z0 so the machine thinks the current position is the homed origin.
In estlcam cam I have settings to lift the Z for fast travel moves set to 0.1mm. So it doesn’t take much time.
I have the fast travel settings that copy Ryan’s basic estlcam instructions. You should copy that as well.
I have the laser start & end settings setup to use the “S” value as a variable for laser power. So it uses the RPM setting from the tool list.
Then in my tool list I tell the laser that it cuts 0.1mm deep per pass. With an RPM of 255 usually. That way If I want the laser to make multiple passes I can set the engrave function to go .1 mm deeper per pass. Ie .5 mm deep means 5x passes. But my laser will only move down .5 mm so it isn’t likely to crash.
If you copy all of that you should get the best of what you are trying to do. Plasma is a lot stronger so you won’t need to use multiple passes.
I have not put much thought into this but would love to get a plasma soon.
Aaryn, you are basically saying use a touch plate the thickness of your arc offset (+0.1mm) and then use my pen trick to only move down 0.1mm for the cut? Easy!
I guess on top of that you need to start the arc, so your beginning gcode could have a little script to move down, start, wait, move up and go…I think.
Thats what i ended up doing, by editing for the start:
G1 Z0
M4
G1 Z5 (distance for cutting)
and for the stop:
M5
G1 Z10 (distance for Travel)
Furthermore set the cutting dept to minimum.
This seams to work, but now, i got the problem of RF interference to my controller, it goes blank the moment the plasma starts.
I will have to determine if it is cased over mains or if i need to create a Faraday cage around the controller.
even with the pilot arc you get a lot noise everywhere
i will try to cover the controller box with aluminium foil, lets see if that helps, if not, i’m going to mount a propper case. if that also fails, i have to look into shielding the steppers fully with enclosures.
but at the end of the story, i’ve seen cnc plasma cutter diy without all those shenanigans, and i wonder where exactly the problem originates
Yeah if you look at the links from the gallery. Some of them have had zero issues, others have fried boards. I intended on buying the exact cutter one of the successful guys had.
well, here is the solution to this problem. aint pretty, but it works;
You should ground your electronics over a longer distance
furthermore, a full shield as wide as possible around the controller. As i am not keen of taking the cabinet apart (yet) this is what i ended up doing; wrap the lunchbox
pilot arc, nothing else.
you have to be lucky to get it to work with HF, and the distance needs to be 100% spot on every time.
if not, you get EMI and that not little, the noise will reset the connection and maybe also kick your board around. if you got a rapsberry as controller or close by, scramble the SD card or kill you computer.
After this stories of nightmare, i’ve decided to invest (grrr) another 200 bucks into a pilot arc machine.
unfortuantly it won’t be for this christmas that the plasma is working, so you can see, my stars for the decoration will not be metal :((( but wood will do
the picture concludes also the way i’ve come, after a lot of trial and error, that’s how good i could make it work. it will work, maybe, but not reliable.
Hello everyone! A bit time has passed, i got a chinese arc pilot machine that was not working and had to source a new one after that, long story short:
i could cut out a pattern today successfully, but i had the arc started manually with a switch i wiped up.
after that, i hooked the relay up to the ignition and on the cnc shield to spindle output on the board, and interrupted the relays 5v power with a manual switch just as a safeguard. that side works now good.
now i adapted the estlcam config to start and stop the cutter, but here comes my problem:
before it begins with the cut, the controller has a delay before at the start position for a second or two while the torch is burning, before it begins the actuall cut. that cases the plasma to cut the metal, but then the arc does not have any more material to cut, increasing the ressistance/current. EM interference that crashed my board again is the result of this glitch. when i did it manually it worked, because the resistance was not so big, i could cut out the star (testpattern).
how to i get estlcam not to wait so long, but to start with the cutting out of the part right after it triggered M5. or can i include the start command right after the part is being cut out?
the delay just cases problems, beside that, it seams to work
There should not be a delay. The machine must be doing something. It should be executing one GCode command right after another. My guess is there is some movement command it is performing. Probably in the Z. That it is waiting to complete before it starts moving for the cuts.
You could test by unplugging the plasma torch and plugging in something like a lamp. Run the program and watch to see when the lamp comes on. What happens next. Then when it starts moving. Sharing a video of that might as well.
Have you opened the GCode? Does it show any other commands between the ON and the movement?
I suggest clicking through some of the plasma links in the gallery, there seems to be a lot of people doing it successfully.
It seems like you are too close to stop now? Get rid of the not needed Z move and then what issues do you have? I would have a hard look at your shielding and how you are doing it, or at what plasam machines they are actually using. Some are clearly better for this than others.
thanks, but i’ve had it. the arduino is not robust enough, and the boards/forums are filled with EM problems.
last step i was willing to take was an pilot arc start machine, i did that now, it kinda works, but as soon as there is no material to cut, and he needs to overcome a gap, the radio/em crashes the arduino, and that is the main issue.
thinking about it, why are commercial plasma machines based on LTP or older serial standards? why the electronics are not so fine? i think this is because it needs to withstand higher interference.
you see, i want to build steam boilers, and a cnc plasma would allow me to cut out complex pattern in one go. the material is copper, so it needs to be spotless, and if there is a slight gap or a controller crash that result of that gab a sheet of very expensive material is ruined.
i rather do it by hand
but thanks for the efford, and hey, i learned a lot. i did cut out with the plasma, but by hand just faster…
Not sure if this is a valid blanket statement, It didn’t work with the board and setup you had, but that is not the only way to do it. I am still pretty convinced it can be viable. You had the most basic 3 axis board that can be bought. I know other boards are far more robust but you are right, I have no idea if they can take it.
I am not trying to be argumentative. I just prefer not to make such a statement and have other think it is fact.
Wish I could be of more help but I just don’t have time to set up a plasma yet. It is on my wish list though!
I don’t think the above factors are actually significant to the EMI resistance of those commercial machines. The important difference is that those machines mount the electronics in a properly grounded metal enclosure, shield ALL of the wiring harnesses and ground the shields at the enclosure. The plasma cutter produces an extreme EMI environment. The wiring is actually a bigger problem than the control electronics itself, as the wires act like antennas, pick up the EMI and transmit it back to the electronics.
Yeah. Arduino is just a bootloader. The microcontroller in your desktop is way more susceptible to emi and noise than an atmega. The difference is in the case, and also in that the cables and connectors have proper shielding. If you put the arduino in a case and use rs232 cables, it would probably be fine. Meanwhile, the desktop wants to reboot to do updates or pauses the job to check your email…
Here is mine, uses a simple 5 axis bob for about 12$, tb6600 drivers 6$ each and a pilot arc machine. Just the wires were roughly 150$ as I didn’t want cheap stuff and EMI issues. Running on mach3. Everything shielded and grounding at box. Table has its own 8’ ground spike right beside it and all grounds shields go there. I will make a post when I finalize the rails and cable chain. That’s not to say I havnt had my fair share of problems to overcome.
just faster…