High Frequency TIG welder causing UNO issues

Hi there
To learn this wonderful world of Arduino i tasked myself with throwing together a “cold wire feeder” for my HF AC/DC TIG welder.
A cold wire feeder puts welding filler wire right into the weld puddle as a pulse sequence through a UNO board.
The code is really simple and I am using Pots to vary wire advance and pause delay triggered by a switch.
For DC welding the result is first class welds every time with steel and stainless steel.
In AC for aluminium the TIG runs HF constantly and this is triggering a HIGH input without button input which makes things interesting (difficult).
I thought it might be shielding so I put it all in a alloy box and have noise suppression on all the wires. I then thought it may be the MOSFET but the same result with simple Relays.
Isolated separate power supplies for the UNO and the wire feed motor.
This issue does happen for one pulse when I start on DC also so I know it is the HF that is doing it.
It runs well enough to use but I have to start the welder hotter than needed as the wire starts immediately with no time to develop a nice weld puddle before triggering the wire feed sequence. I am also concerned that it may not be great for the UNO either.
Has anyone had this problem and know of a solution?
Any suggestions much appreciated.

Unsurprizingly you have vast amounts of interference to deal with. How are you protecting your inputs at the Arduino end?

I’m surprised you say that on AC welding the HF runs continuously. My AC/DC welder certainly doesn’t so maybe you have a fault in your machine.

On my Miller Syncowave 250 you always turn the HF switch to continuous when you are welding aluminum with AC. What kind of machine do you have. Maybe something is wrong with yours. I had the same problem when I was trying to build a wireless control pedal using Uno's and RFM69 transceivers. Worked great until the HF activated and then they would just go crazy. Got kind of disappointed and other projects got in the way so never pursued it any further.

One single wire that picks up HF energy and enters the Arduino box can ruin the functionality.

EMI testing and resolving is a very tricky business. Often one needs spectrum analyser and some specialized probes and a very rigorous setup.

All your wires should be shielded to a single solid ground point that itself does not introduce a common mode interference signal.

Preferably each wire that enters your mumetal (magnetic shielding alloy) box should then have an appropriately sized ferrite bead or toroidal core to block HF energy from entering sensitive inputs or even being picked up by PCB traces. Beware, even your power supply should be thorougly isolated from the possibility of picking up or transmitting HF energy to Arduino. Your mains will be infected by all kinds of noisy signals due to the industrial grade TIG emissions.

Not an easy task that needs a very thorough design strategy.

Look for EMI, EMC and ESD topics related to high powered devices to learn more.

Hi all.
So I am not going crazy==?
For the welders: no she is on the low end (JASIC) and certainly no Miller. Had it 10 years reliable , stable and i love her.
For others: When I am looking closely at the business end of a TIG on AC with aluminium i am thinking “World Engine” from the Superman movie! It is most certainly a seriously weird source of noise. However a nice lapping shiny alloy weld bead is worth all the weirdness!
Maybe i am doing it wrong?
I have ferrites on every wire and have paid attention to grounding. I was thinking that the HF noise was traveling down the wire to the large spool and through the wire feeder chassis but simple continuity checks show that the drive chassis is isolated from ground.
I am not sure if this should be or not. I guess there is one way to find out!?
My research is heading down the CAP path as i remember 20 years ago when I was messing around with guitars and big amplifiers we used caps across switches for noise. This was a long time ago and I am going to have to do some reading.
I guess a simple solution is to accept the problem and for AC welding use a simple dedicated secondary trigger switch to the power of the Arduino. At least this way I can get the weld puddle started then boot the arduino and let it start freaking out then kill it at the end of the weld as it seems I still have my pot pulse and delay working its is just the HF triggering the switch HIGH.
The trouble with this is that I will know I have not resolved the problem and have only created a not-elegant “backyard” workaround and will have to live with the shame. I will then move on to my next Arduino project and wait for the solution to hit me.
Thank you so much for your comments as this was my first post on any forum ever so consider that this issue is seriously bugging me. If there was a virtual Bar nearby I would buy you all a virtual beer!
Attached macro photo of a weld bead run on plain old mild steel which I see as a partial win for this project.
Makers unite!
int_hayes

There's an excellent article here from Miller which explains why "modern" inverter welders no longer require continuous HF when doing TIG AC welding.

https://www.millerwelds.com/resources/article-library/tig-frequency

Thanks jakrae Miller know what they are doing. Here in Oz I think they are re-badged as CIG WELD or WIA not sure. For me tig is the one I can do in my slippers. Clean, precise. We need to find a way to shield arduino from our industrial scale noise. I am a perfectionist and this one will not defeat me although it may take a while. I did run the switch as a power switch today with no problem other than the 2 second boot time before the arduino powers up. Works really well but as a said before it is not the solution I am happy with.

Out of interest - why do you actually need an Arduino. What is it measuring and what computations is it required to perform during the operation. Could it be that you are "over-engineering" the project.

When you think about it, spark radio transmission was all that existed when Mr Marconi first started developing long range radio communications, so trying to build a detection and control system that won't react to the radio noise being produced by the unscreened welding arc will be quite a challenge.

jackrae: There's an excellent article here from Miller which explains why "modern" inverter welders no longer require continuous HF when doing TIG AC welding.

https://www.millerwelds.com/resources/article-library/tig-frequency A Syncrowave 250 is a transformer welder, not an inverter. The reason the inverter ones can get away with this is because they run the arc frequency much higher than a transformer ones. On a transformer one you are pretty much stuck at 60 HZ arc frequency.

Is this like .... a false trigger issue?

See what happens if you put a big electrolytic capacitor eg...few hundred microfarad (or more) between the supply pin of the Arduino and the ground (0V) pin of the Arduino.

detown:

Indeed, but is the OP’s machine a transformer or an invertor

Hi all Jackrae Today i was standing at my desk at work which happens to be at the end of of my big CNC HD plasma fiddling with a model on my 3D CAD platform. Occasionally I would check the NC bandsaw and CNC router and then do a loop around and tell a few jokes to the welders words of encouragement and direction. The whole time I am considering your question. People probably said to mr Marconi "what the hell is wrong with writing letters under candlelight?" (Actually sounds like a good idea right now) I desperately want to learn micro's and the way I learn fast is to throw myself into a challenge and cram what I need to get it done. Not really into making LED's flash or laser limbo with my children although that is lots of fun and we all laugh a lot. I have a notebook of ideas for micro's but they are all for the industrial side where there will be plenty of noise. I don't know where you are but manufacturing in Australia is tough and I have this romantic vision that the way of the future is ultra smart fabrication through innovation. To me this means embedded systems where there has been none. So my cold wire feeder is nothing new until I put the heads up display in my welding helmet!!!?? Sorry for the essay but you asked. My little tig is a square wave inverter. Southpark yes absolutely it is a false trigger. Thank you I will try that cap. I have an impossible deadline to meet at work but I can't wait to try it. Thank you all int_hayes

I know little about welding but, there is a fab shop close to me and arduino projects that work at the weekend suddenly stop working at 9 AM thereabouts on monday morning.

The mains earth becomes amazingly dirty at this time.

Boardburner2

Welding is great fun just don't do it in shorts and a t-shirt. 20 years and I never learn. Ok right. So maybe battery to power the board?

On the grounding issue I mentioned previously that I was thinking of grounding the wire feeder. I am thinking this may be BAD as the HF is seeking a continuity to start the ionised arc and so it may decide that the wire is a nice easy route rather than the target metal. Perfect for that other welding process called MIG but I think awesomely bad for tig. So a cap on the switch and battery to the board? All righty then. Thank you

int_hayes: Boardburner2

So maybe battery to power the board?

This is my first and often successful method.

However in practical use there is often a connection to earth somewhere, that's when things get difficult.

I prefer to use a locally derived earth apart from the mains earth, a separate earth rod can help although for mains stuff this requires an isolating transformer.

int_hayes: Hi all.

I have ferrites on every wire and have paid attention to grounding. I was thinking that the HF noise was traveling down the wire to the large spool and through the wire feeder chassis but simple continuity checks show that the drive chassis is isolated from ground.

A picture/diagram of the setup would be helpful.

My problems are ground borne but yours may well be pickup from a strong rf source. Shielding and filtering is probably the answer. More detail about the welding process and circuit would help for those of us not familiar with the process.

Boardburner2: A picture/diagram of the setup would be helpful.

My problems are ground borne but yours may well be pickup from a strong rf source. Shielding and filtering is probably the answer. More detail about the welding process and circuit would help for those of us not familiar with the process.

Sure Give me a few days. Thanks

I am guessing from the pic that you are using a wall wart . the leads from those make an effective aerial. Using twisted pair or shielded cable may help.

Using batteries inside the box may tell you if this is where the problem is occuring. Given the setup you may have magnetic field pickup as well, a steel rather than alloy box may be better.

The default configuration of most Arduino pins is INPUT which is high impedance, so any unused pins would act as antennas bringing the interference right in to the chip's die. I would suggest configuring all unused pins with INPUT_PULLUP.