THC lifting off sheet?

After kicking around the idea of building a CNC plasma, I eventualy bought a secondhand (new) EZrouter machine via this site, last year - it was only 15 miles away and I was cutting within a day, even though my bank balance was somewhat worse for wear!
In general this has been a good machine, the odd teething problem and a couple of build mods have been required, but generally good.
The problem I have is that occasionally (1 in 50 or so small parts) the THC will lift the torch clear of the work such that the arc fails (approx 40mm) at this point the x and y motion continues, but this part is obviously ruined. The next part to cut is usually ok again. I am only cutting small pieces at present, but it worries me that if I were trying to cut large metal art I could end up wasting a lot of time and material. This problem has been present from the start, but I havn`t done that much with the machine and have not worried about it until now.
The THC is a THC300, and I am running Mach3 after processing with sheetcam. Any suggestions of where to look for issues would be greatly appreciated.

Thanks

Geoff

Well if nobody else is willing to give it a go...

Up means Volts too low. The cutter tries to maintain a constant Ampage between hafnium bead and workpiece. If the resistance drops so do the Volts, the THC assumes it is too close and moves up.

As far as I can make out, most problems seem to result in increased resistance and the torch plummeting to it's death. So how to get a decrease?

If not a loose connection, my only suggestion is return current via the bed and you suddenly improve the connection by "welding" the workpiece to a slat.

It's a puzzle.

Thanks Robin,

I found the tip was somewhat worn, so replaced and had no problems for a while. (cut about 3 or so sheets since, changing tips regularly). I have a feeling theres a wiring fault somewhere as it was having problems starting the arc more recently.

I intend to go right through the machine in Jan and check out exactly what components have been used (drivers etc, check opto isolation exists etc) and perhaps rewire some of it - for example I had a problem with the X axis stalling and found it to be a connector going OC in the cable tracking.

As stated before, the mechanics of the machine seem good, but the wiring and attention to detail is a bit lacking. Example: had to cut a flat on the motor mount for the leadscrew of the Z axis as the grub screw was not secure enough and hence the motor was stepping, but not turning the screw!

Would be interested in hearing anyone else`s experiences of these machines.

Geoff