Fitting a 14 pin remote control to an engine drive welder

That seems reasonable. I wondered the same thing but couldn't see through most of the details. A DC motor speed control module could probably be adapted to do the same job. Possibly it could be made to do both the regulator's job and the welder control's job by selecting either current set pot or output voltage error amplifier as signal source. That's for another day...

I took a look at S1B today and it makes even less sense 8-) The panel labels agree with the schematic markings, which make no sense. Even better, the switch is open! The contact slider for S1B appears to be missing. Everything else is there, S1A is fine.

That's my impression also. The welder is open loop, mostly just a fixed voltage on the current setting choke. Any feedback would be too slow.

That has me stumped as well. I have been told it's possible to build a PWM circuit using SCRs, either by gate-turn-off or current diversion, but either approach seems to require more parts than are shown.

For the moment I'm tempted to leave well enough alone. As Iggy pointed out, the last think I want to do is mess up a working welder. In the meantime you've given me some good things to think about. I'm probably better off practicing than tinkering. Google found a few examples of folks hooking up remote rheostats, so when the carbon pile arrives it makes a low risk experiment. At this stage just keeping the tungsten out of the puddle takes most of my attention; another control to think about could easily make matters worse.

Thanks for all your help!

bob prohaska

I'm rather surprised: Seems like there's be less inductance to fight on the primary side. The secondary side is _designed_ to maintain an arc. Leakage between primary and secondary would give the primary a little isolation.

bob p

You're welcome. Welding machines interest me a lot. A slightly different tack on this might be to try building your own current control/HF system to run off the weld outputs of your Legend, something like:

That way, your machine would stay unmodified and retain what value it has now.

I have wanted to build my own design inverter TIG machine for a long time, but I wanted to have one that would always work and not be a perpetual project in the way of just going out and welding. As a result, I bought a Miller machine. Maybe after I retire...

Good Luck, Bob

That's an astoundingly ambitious project! Given what could be bought in 2011, it might have made economic sense. Given what can be bought today it's not so obviously worthwhile. The challenge might be satisfying but I'm not sure there's any functional gain to be had.

The machine I have is no collector's item. There are a few photos here:

Mostly I like to play with old machinery. If it can be made easier to use on the cheap I'm not afraid to make changes. What the AEAD-200LE's taught me so far is that the faults are in the bum on the seat....

There are times when it's possible to build better than one can buy. It was true of hi-fi speakers in the '70s for a short time, maybe ten years. Probably true of welders when IGBTs first appeared, but now?

Thank you!

bob p

snip

I don't have any illusions about being able to build a better machine than I could buy. Mostly, it would be a very cool technical challenge. Building ambitious projects has been one of my favorite ways to learn new stuff and increase or keep current with stuff that I can use in my day job or maybe my next day job.

BobH

Ok, understood. Might it make sense to try much higher than normal working frequency, tens of kHz or even higher? I wonder what effect it would have...

bob prohaska

The Dynasty box that I got will run 250Hz on AC and you get significantly better penetration on aluminum at 250Hz. It also sounds like a hive of really angry bees. I don't know how far upward that extends in frequency though.

BobH

There's a remark in one of the welding forums that Lincoln built a

I was thinking in terms of considerably higher frequency, high enough so that electrode capacitance, rather than electron emission, contributes to the discharge. I'd think that would be in the hundreds of kHz if not higher. Admittedly, I do not know if it's worth the trouble. But, inverters weren't worth the trouble twenty years ago, now the trouble is much less and the virtues seem considerable.

Once the frequency goes over 20kHz at least you won't hear the bees....

bob p

Sounds like a dandy radio frequency jamming device...

My Hitachi GP2 TIG goes to 500Hz IIRC it was made around 1998.

The FCC regulates emissions down to 10 KHz.

What would be the advantage of capacitive coupling? I suspect that the plasma in a normal TIG arc is so conductive, that the difference would be lost. But then I don't understand why the higher frequency AC penetrates better. I'm not criticizing here, I am just trying to piece it together.

It just occurred to me that the voltage drop between the electrode and the work times the current flow is dissipating a LOT of power, that may be the primary source of the welding heat. This is pure speculation on my part though, I am just an engineer, not a plasma physicist.

Curious, BobH

Do you notice much difference between 250 and 500Hz on aluminum welding?

BobH

Shielding would likely be a significant problem. However, if the advantages justify it the work could be done in a Faraday cage. They can be made of screen or metal covered wood and aren't really very exotic, apart from being low production items. Think of a carefully costructed screened porch. Or, a sandblasting cabinet.

There are some FCC allocations for industrial heating applications. One at 13.56 MHz is used quite widely, though that frequency may be impractically high. Induction heating is lower frequency, but I'm not sure how or if it's regulated. Microwave ovens have a slot at 2.4 GHz, but I'm pretty sure that's too high for convenience, if not for effectiveness.

bob prohaska

The power dissipated is distributed between the electrode, the workpiece and the arc itself. I think the major effect of raising the AC frequency would be to change the distribution of power dissipation among those components. If you could build a 13.6 MHz welder I'm pretty sure it would behave very differently than a conventional one. Better or worse, and for what purpose, I'll admit to being unsure.

In a DC arc electrons have to be emitted by the cathode and collected at the anode. Most of the current is carried by the electrons and most of the power is delivered to the surface they land on. To make the cathode emit electrons either ions must strike it with enough energy to dislodge sufficient electrons or the cathode must be hot enough to emit thermionically.

This remains true in an AC arc up to the frequency where electrons don't have time to transit the arc before the voltage reverses. Then, they're trapped, sloshing back and forth in the arc, collisionally ionizing the gas and ensuring an adequate supply of charge carriers to keep the discharge going. The result is that the electrodes don't _have_ to supply the electrons. This allows the cathode to run colder than required for thermionic emission and avoids the erosion caused by ions hitting the cathode. More power goes into the arc, less into the electrode and workpiece. That opens an opportunity to manipulate where the heat goes. At the same time, ion cleaning of the workpiece would decrease, perhaps to the detriment of the welding process.

The frequency at which electrons don't have time to transit the arc isn't known to me and is probably rather hard to estimate accurately. Ordinary fluorescent lamps have an efficiency peak in the tens of kHz, electrodeless lamps have been commercially available in that frequeny range. Microwave frequencies have been tried but far as I know they were a technical success and a commercial failure.

I'm not sure plasma physics is much help in a puzzle like this; it's a very messy problem. Direct experimentation is apt to be easier and more persuasive. Quite possibly experiments have already done, probably for somebody's Defense Department during the 1950's. They'd have been stuck using vacuum tubes, which would have made the project much harder than it would be today.

Apologies for the length, thanks if you read this far!

bob prohaska

The speed of electrons in a solid wire was one of my college physics homework problems.

You'd have to know the free electron density in the plasma channel to figure it for an arc.

--jsw

This gives around 200 m/s as the plasma velocity in a 10mm long arc, which corresponds to one wavelength at 20 KHz.

Yes, and the various electron collision frequencies (or equivalently, the mean free path between those collisions). I think one of the assumptions of the copper wire case is that there is one conduction electron per atom and negligible scattering. In an arc there are many more neutral atoms than electrons and electron-neutral collisions dominate transport.

An estimate could be made, but an accurate estimate would likely cost more than a simple experiment.

It occurs to me that selecting a driving frequency which maximized the electron-neutral collision frequency would effectively raise the resistance of the arc, allowing more power to be delivered with less current and a lighter, smaller torch and electrode. That might be of some value.

thanks for reading,

bob prohaska

That's an interesting paper, I wish I could understand the math better. It seems to treat only the DC case, however. I wonder if there's an AC treatment around somewhere. It'll surely be much messier! One of the guiding assumptions is that the arc is a resistive fluid obeying Ohm's law. Arcs generally don't obey Ohm's law.

As the authors note, the paper is a "useful intermediate step".

Thanks for writing!

bob prohaska