pc rebooting when plasma is on

[freesteel]

Hi everybody, i have a problem with my plasma router, i'm debugging it and this is my first cuts, but when the plasma is on, my pc goes away like rebooting or so. i use a cnc3ax control board and a 35A plasma cut machine, i thought that it would be probably some isolate pb from high electricity in the routerr made by the plasma ground??
you can see my router in my pics, it's all made in stainless steel and the plate to cut is laying on the metal frame.
thanks for your help!

[Al_The_Man]

Your Plasma table will be at ground potential as well as the PC power supply.
It is important to make sure these ground points are bonded together at the system ground input for both PC and plasma.
Al.

[freesteel]

You're probably right Al, the router is on 4 rubers weels that isolate the router from ground and there's no ground wire on the frame, i'll add one and try the same work i did.
But another thing i'm thinking is the place where the plasma machine is sitting, it's just under my pc on a shelf of a furniture, does the plasma machine spread any high frequency or so making my pc above it crashing?
Thanks for the help.

[jgramlich]

It really shouldn't...but it could depend on your pc case as well. Is the case a metal box or is it plastic? I know my unit is fully enclosed in sheet metal, acting like a faraday cage. (http://en.wikipedia.org/wiki/Faraday_cage).

[Al_The_Man]

Most desk top PC's have the incoming ground conductor connected to case, or any metal portion of the PC, the MB generally has one or more screws that carry the 5v ground plane through to the case, hence in these PC's, the P.S. will be at ground potential.
If yours is like this and there happens to be a difference in potential between the plasma ground and PC ground due to poor bonding then this can be the cause.
In any event, it is wise to make sure of the ground bond.
Al.

[jgramlich]

All the chassis parts should be grounded together, yes, but if the sides of the case are plastic, noise coming off the plasma could be inducted into circuits directly on the board, it won't all get trapped by ground if the unit is not completely surrounded by metal. This *can* happen, doesn't mean it is. :-) I'm not terribly familiar with the type of RF noise generated by a firing plasma torch, but I'd guess it's a lot....not to mention the motors and other electronics inside the plasma power supply itself.

[jimcolt]

Some plasma systems....such as Hypertherm't Powermax plasma systems....do not use a high frequency discharge to ionize the gas for arc starting.....these units, known as "blowback start" usually do not cause electrical noise interference with sensitive electronics.

Many older design plasmas and recent imported units as well as high end industrial plasmas...use a very powerful high frequency discharge to fire the arc...for a few seconds at the beginning of each cut cycle. These systems work well with cnc controls, drives and electronics that are hardened to work under these conditions....but often have problems associated with noise on machines that are controlled by a PC.....and that have wiring to encoders, drives and motors that is not properly shielded and grounded, as well as having input/output filtration and isolation.

The start of the best solution is to drive a ground rod as near to the cutting table as possible.....and ground the chassis of every machine component (individually, not "daisey chained" to the star ground source.....also the work connection from the plasma should directly connect to the point. Another heavy cable should connect the cutting bed slats to the ground rod. Torch leads should be as short as possible, and separated as far as possible from any other data or power cables on the machine. Data cables should be shielded, and the shield drain connected to the ground rod at one end only. This may help.....if the problem persists...then opical isolation of PC and Drive I/O may be necessary.....or switch to a Hypertherm Powermax brand plasma!

The high frequency circuit generally operates in the 10 to 15 kvolt rang at a frequency of approx 10 megaherts.

Jim

[freesteel]

Thanks guys for your help, i'll tell you if it's better with all frames connected to the ground. If it doesn't work, i could try to connect my pc on a UPS too. It's obvious that my plasma is not as efficient as an Hyperterm, the cost is not the same! mine is a regular imported 35A plasma!!
See you!

[jimcolt]

Ground rod is extremely important......I would suggest this as a first step...then attach the chassis grounds to it. Connecting chassis grounds to an electrical panel ground can/may make the problem worse.

And.....if the plasma does not work...it probably is not really less expensive!

Jim

[WSS]

It really shouldn't...but it could depend on your pc case as well. Is the case a metal box or is it plastic? I know my unit is fully enclosed in sheet metal, acting like a faraday cage. (http://en.wikipedia.org/wiki/Faraday_cage).

One of my favorite detective shows (Midsomer Murders) featured a Faraday cage in a episopde. They were using it to isolate stray RF to see if certain people really had second sight. Well someone with malice hooked up the cage to the mains and the subjects were charged with something like 200amps of 220V, making a big mess.

So be carful how you hook up your ground!!!

WSS

[Alan Bradford]

The first thing to do is see if your Plasma is a High Frequency Starting type.
As Jim said most of the cheaper units are.
One of the biggest problems I have seen with the PC based hobby class of plasma cutting tables is they were never designed for HF type plasma units.

Before you all jump me.....Hobby grade is the machines that cost less then $30,000 and the Industrial units usually usually run around $50,000 to $500,000 The plama systems usually add another $20,000 to $80,000 per torch.

The RF can get into the plastic cases, but I suspect that the cables that run to the motors, limit switches, keyboards, mice and whatever you have on the machine are not shielded. These work as an antenna for the HF and NO amount of grounding will fix that.
You can look at ALL the cables and try to see if the conductors are twisted in pairs and/or if there is a foil shield between the wires and outer insulation.
Sometimes you can remove a shell on the ends of the cables to look how they are made. Most, (but not all) cables with molded plug ends are not shielded.
Try to keep the cables away from the torch leads, and if they have to cross do it at 90 degree angles to reduce capacitive coupling.

You may need to replace the cables with shielded ones. If you make up the cables then ground the shields at one end only. This helps to reduce ground loops and lowers the niose level in the cables.

I typically find the problems are with the Cross and/or Z axis drive motors. These are usually located closer to the torch and run along with the torch leads. If so try to isolate them as far apart as you can get them.

Most of the modern industrial grade cutting machines are filtered,shielded, and isolated to run with the big High Frequency Monsters (Like Hypertherm Max 200/HT-2000, and Thermal Dynamics Merlin series come to mind as the noiseist).
You will find all of the cables are twisted pair, and shielded. This offers the best protection from becoming antennas. The CNC's are highly filtered and optically isolated on all the I/O's. Almost half of the cost of the Industrial Grade CNC ($10,000-$20,000) alone is due to filtering.

So it may be cheaper to go with a non HF plasma unit. Cheapest isnt always cheapest.
Ask my brother in law who insisted on buying recap tires. He couldnt afford new ones, because he was wearing them out twice a year!!!

Hope this helps.

[HuFlungDung]

If it will make you feel any better, I've got the same problem with my PC retrofitted lathe. It is using an old Athlon 550 CPU. It reboots every time the plasma is fired Kind of annoying but I think it is just a weakness of the motherboard in that PC. The distance is 50 feet away. The PC is in its own typical case, and is inside a metal cabinet with some openings in it.

My office PC is unaffected by the plasma.