Design mistake

[nerginer]

Hi,
I build up my new cnc plasma cutting machine. Every thing is good but I have a problem. I put the pc and the drivers everything behind the gantry. When the plasma came near the pc it restarts the system when M3 command fires the plasma. I can not change the design. Is there a way to prevent the system from interferance

regards,
Nuri Erginer

[totally_screwed]

Nerginer
I know nothing about plasma cutting.
From first principles
First of all you need to establish how the interference is getting into your electronics. Interference can be radiated [Radio Frequrency] or conducted ,[via wires]. I assume your PC and the Plasma drive circuitry are both separately and fully enclosed in conductive metal boxes that are separately earthed. Are you running the PC off a separate supply, [separate from the drive to the Plasma cutter]? If not, it could be that the electrical load imposed by the plasma source or interference is either dropping the CPU supply voltage and causing a reset, alternatively interference could be generating reset signals.

An initial first stab would be to fit ferrite toroids [beads] to every wire [including mains supply] that enters and leaves your PC, except those that have moulded plastic cylinders, [these already have ferrite toroids fitted], e.g. monitor, some usb cables & etc.
It would also be helpful to do the same for the plasma drive PSU and all the signal interface wires.
It also might be an earthing problem. However these can be awkward to resolve.
If that doesn't work, it's going to need a field engineer, unless someone else has go any bright ideas.

Hope this helps!

[totally_screwed]

Nerginer,
I 'googled' for plasma cutting and it appears that it's highly likely that it's the high frequency interference emanating from the pilot arc. These are apparently well known for causing this kind of problem. This is potentially good news because it suggests that it isn't an earthing problem. If I were you, I would concentrate on screening your PC and fitting those ferrite toroids on every wire that enters and leaves your PC's cabinet. It would be worth considering converting all signal-ground wire pairs to 'twisted-pairs'. It is worth fitting earthed metal mesh screens [Faraday screen] over any holes in the case such as the fan vents, open expansion card slots and any cut-outs for DVD, CD, and floppy drive & etc. You will also need to ensure that access covers are properly grounded. Star-earthing is probably essential.
If the problem persists, the cables could be replaced with double foil screened cables.

Have you got access to an oscilloscope or a spectrum analyser? These should give you the possibility to do before and after tests. Look at the signals with plasma arc off, the switch it on and see what happens!. Make sure you don't blow up the SA.
Hint: with spectrum analyser [SA], connect it via a capacitior to DC or signal rail. Some SAs can be blown-up [no shrapnel, just dead SA] by feeding DC into the input! [expensive]

An alternative would be to try and have a chat with someone who has real experience of this kind of problem. Their advice would probably much more specific to your situation.

Once again good luck!

totally_screwed

[apcnc]

What plasma system do you have? Some like the Hypertherm Powermax series use an inverter instead of high freq. If you are using a system with HF pilot arc, do you have shielded leads? The plasma system, lead shield & table should be tied to grnd rod.

Micah

[linatrol god]

This is typical plasma noise. Isolate all your electronics from the machine. Run a thick braided wire from the gantry to the ground rod, hopefully you have a ground rod. If not put one in and make sure the ground rod itself makes ground contact to earth. Shield any and all cables that you can, separate the torch lead from all other cables and shield it, run this shield to the leg of the table closest to the earth ground. Connect a thick wire from this same leg to the ground wire.
Plasma noise, the radiated noise emitted from the high voltage, high frequency firing of the torch can cause all sorts of problems.

[Torchhead]

All new plasma systems are "inverter" systems. That has nothing to do with the way the arc is started. The Hypertherm 1000 series and smaller TD machines use a "air-contact" start system. It fires the torch and uses the burst of air to drive the electrode back from the tip causing an arc. Older systems uses an high frequency start proscess that is like a spark generator (ignition coil) and it radiates huge amounts of RFI and EMI. The contact method contains the arc noise at the tip. The HF puts it everywhere.

Evidently the air-contact start system only works up to a certain size and after that the bigger machines are back to HF start.

The worst are older plasma systems like the Stac-Pak and old Hypertherm MAX series. They will jump an arc over 1/2" and send out a noise pulse that would probably bounce off the moon!

The important thing is to have the plasma and table both grounded to a local ground rod and NOT have any other ground between the PC side and the table side but the one back at the main breaker box. Having two ground paths between two systems is fine for 60HZ and safety grounding. Two paths for higher frequency causes a ground loop and can send sensitive electronics into reset or worse.

The problem with the older HF start units is that the start frequency is relatively low so that it's harder to filter (ferrite clamps alone have no effect) and of greater amplitude.

For noise grounding the SURFACE area of the gounding wire is more important than the diameter. Stranded wire is better than solid. Braided wire is better than stranded. Noise travels on the surface of the wire (Skin effect). Establish a single point close to the table as the central ground (close to the ground rod) and star (run separate runs from each section to be grounded) back to the point. Don't daisy chain grounds. Bearings may not be good ground HF conductors so use a separate strap on the gantry.

All plasma cutting (and arc welding) generates a level of noise but with proper grounding it can be suppressed. Older HF start systems present a larger challenge and everything has to be perfect for them to work with modern digital controls.

If your motors have a case ground connect that out at the table. DON't run it back into the drive cabinet or ground it to the AC ground on the PC side.

With using shielded cables only ground them at one end (ususally the "driven" end)


Tom Caudle
www.CandCNC.com