Runaway torch THCAD-10 + PlasmaC + Hypertherm 45XP

Hi,
I have an issue with recently installed THCAD-10.

During the cut, the torch is moving higher and higher and eventually stops cutting and just blows a bit of material from the top surface.
The result of the cut from the video is attached below.
I'm totally new to this and not even know where to start the investigation. I suspect that the system thinks that the voltage is incorrect an tries to compensate by moving the torch.

I've used 3mm steel, according to Hypertherm manual I should have 117V on the arc, I'm using 1:20 divider and THCAD with 1:64 divider, config attached ( ).

Some cuts from the same settings are fine, so I do not know why it does work on some cuts and does not on other.

Looks to me like one possibility is that you're cutting too fast, and then the voltage goes up once it stops blowing the material through and the torch raises to compensate.

Another thing I (and probably everyone here) would tell you to do is to use "auto volts" instead of the specified voltage in the Hypertherm manual. You have no easy way of verifying that you are cutting the exact same type of material, under the exact same conditions, and that your THCAD is calibrated the same way to match the way Hypertherm took the voltage readings. Fortunately, Phill made it easy, and PlasmaC has the option to sample the voltage after the THC becomes active, and then shoot to maintain that voltage throughout the cut. If you're cutting height is 1.5mm, and the voltage is say 125v for your system instead of the Hypertherm value of 117v, in the end you don't really care, so long as you're cutting at 1.5mm. Make sense? Auto volts lets you achieve that without all the fancy testing.

That said, you should also verify that your THCAD reads voltage correctly by supplying it a known voltage and making sure it displays correctly in PlasmaC.

Actually, now that I see your home screen on the other post (sitting at -30 volts), it suggests to me that something isn't set up right. If you are reading a higher voltage or a lower voltage than you should be, and you are trying to hit 117 volts, it would cause the torch to raise or lower to compensate (raise in this case). You need to make sure you properly calibrated your THCAD. Why did you use 1/64 vs the normally recommended 1/32? Also, use this calculator to verify your calculations: calc

Snowgoer540 - thank you very much for your suggestions.

I will take multimeter and bench power supply to verify voltage reading. Without knowing that the other spreadseeht exists - I've created my own:
docs.google.com/spreadsheets/d/1dGBwp7ia...h2Wdn9UeY/edit#gid=0

I check voltage reading and see if it does read correct values.

The photo on which you can see -30 is from my Dell laptop. This laptop is not connected to the plasma. I've booted the same copy of LinuxCNC on it to diagnose the issue with "Preview" screen, so the photo is completely misleading.

radek_marko wrote: Snowgoer540 - thank you very much for your suggestions.

I will take multimeter and bench power supply to verify voltage reading. Without knowing that the other spreadseeht exists - I've created my own:
docs.google.com/spreadsheets/d/1dGBwp7ia...h2Wdn9UeY/edit#gid=0

I check voltage reading and see if it does read correct values.

The photo on which you can see -30 is from my Dell laptop. This laptop is not connected to the plasma. I've booted the same copy of LinuxCNC on it to diagnose the issue with "Preview" screen, so the photo is completely misleading.

Copy that, I see you are also using a parallel port, which is likely why you chose 1/64.

To echo what Tommy said in the pther post, you should do a ton of reading. Start here, http://linuxcnc.org/docs/devel/html/plasma/plasmac-user-guide.html

Read it all, a TON of time went into putting this together. It also covers features that most people don't know, or forget exist.

All that said, we are happy to help. Sounds like you have some checking to do though. The joys of "easy" plasma tables

I did a quick test using a bench PSU and the voltage is stable at the lower range, but totally unstable at the higher range.

I didn't have a chance to do more diagnostics, but most likely that's the source of the problem of run-away torch.

My guess is that you are exceeding the count rate capabilities
of your encoder input

I think I will reduce settings on THCAD-10 from F/64 to F/128 and see if it will help.

Might I recommend a nice Mesa 7i96 or similar

Snowgoer - are you delicately suggesting that LPT might not be capable of performing well on counting impulses from THCAD-10?