THCAD Environmental Fluctuations

rodw wrote:

Would you like me to lend you a THCAD so you can test by swapping out?
Send me an email and I can express it down tomorrow with TNT.

Thanks for the offer Rod but I have a spare THCAD10 on my desk that Ill put in to test the issue.

rodw wrote:

I'm not understanding your circuit. From memory your THCAD is in your control panel so you should be able to power the THCAD from one of the 5 volt signals on the 7i76e provided for that purpose.

For some reason when I first built my controller I mounted my THCAD in a small steel box outside of the control box, looking back it was a silly thing to do so that is why I have used a external 24VDC to 5VDC step down transformer.

thefabricator03 wrote:

rodw wrote:

I'm not understanding your circuit. From memory your THCAD is in your control panel so you should be able to power the THCAD from one of the 5 volt signals on the 7i76e provided for that purpose.

For some reason when I first built my controller I mounted my THCAD in a small steel box outside of the control box, looking back it was a silly thing to do so that is why I have used a external 24VDC to 5VDC step down transformer.

How far away? There should not be much voltage loss.

I think you should duck down to Jaycar and grab some shielded cable and power the THCAD as PCW said. If you use 4 core cable, use 2 wires for each side. That certainly eliminates the power supply as being the culprit.

If it happens again, measure the voltage on the THCAD input terminals coming from the plasma cutter. When I was chasing stuff down like this, once I grabbed some old multimeter leads, cut the ends off and permanently wired them in.

Actually, are the THCAD outputs shielded? If not, redo with shielded cable. In that case, just run the one shielded cable for both power and frequency signals

rodw wrote:

Actually, are the THCAD outputs shielded? If not, redo with shielded cable. In that case, just run the one shielded cable for both power and frequency signals

In fact, some CAT6 network cable would do. Use one twisted pair for +5v joined together, another pair for -5v but bring the frequency back on a single twisted pair for best noise immunity. See the mesa manual for the 7i76/7i73 connection if in doubt as Mesa powers the 7i73 the same way.

Well, I think I know whats going on.

I set up a 7i76e on my desk at home and decided to play with a THCAD-5 and a 24 Volt power supply I installed for Ohmic sensing.
I had not set up any scaling and was just observing encoder velocity
I was getting all sorts of crazy fluctuations. They went away if I disconnected both THCAD inputs.
As soon as I connected either +ve or -ve, the fluctuations returned.

So I plotted them in halscope

There is a perfect 50 Hz AC ripple!

So then I disconnected the THCAD from ground (being to a DIN rail via a clamp) and the signal was rock solid.
The mains was also grounded to the din rail

So I think that your problem is a ground loop emanating from somewhere in your building on hot days due to some random dodginess somewhere.

I think the solution might lay in what has been suggested earlier,
1. Remove the external power supply
2. Power from the 7i76e
3. Install a quality RFI filter on the power input (Jaycar may not qualify here!)

Eventually I'll try and add a mains RFI filter and observe the response.

Thanks for that Rod, I think your on the money. My first thought was the edges of the plates I was cutting then this first occurred is that they looked like a sine wave. They were perfectly cut in a wave pattern. I had a power cable near the THCAD box and I thought that it might be causing some interference somehow.

I will implement those suggestions and report back.Its been cooler here the last couple of days and I have not had the issue reoccur.

This actually takes me back to issues I had with noise here at home after I first started.
My test rig for reference.


I think the photo shows how little hardware that is required to set up a plasma machine. There is only 2 more resistors to fit!