Switching inverter welders

[MorRobotz]

You ever look at the specs on these things?

I found this build log http://danyk.cz/svar_en.html
Crazy stuff.

I know you can buy Chinese ones for less then 300 bucks that do 90+ amps DC.

Anyone ever think about building their own? It's been haunting me for a while now. 

No need to lecture me about HC and HV safety. half of my dreams involve isolation slots and thermal cut offs. Other half of the dreams involve larger capacitor banks. Soo... thoughts?

[MagicSmoker]

The site you linked to shows a two-switch forward converter - an excellent choice of topology overall due to the low stress on the primary switches and extreme ruggedness. There's a few things in the schematic that are a bit questionable/wonky, but nothing that is a show-stopper. Go ahead and build it.

[NiHaoMike]

It would be very difficult, to say the least, to build one for less than the cost of buying one, especially after accounting for development time and blown components. Now if you're after something exotic that either doesn't exist on the market or is considerably more expensive, go for it. Even then, it might make more sense to modify an off the shelf device.

[Miyuki]

Im now building one but more complex with pfc, full bridge topology and mcu controlled to generate "constant voltage" for GMAW
I have now done PFC module controlled by IR1155, but must be checked at load, AUX supply, its simple flyback to generate 2x12V and output rectifier which is bridge to simpler design, both sides of transformer is then clamped

to do:
Must write control software and put transistors to power stage and wound transformer and output inductor

[najrao]

You ever look at the specs on these things?

I found this build log http://danyk.cz/svar_en.html
Crazy stuff.

I know you can buy Chinese ones for less then 300 bucks that do 90+ amps DC.

Anyone ever think about building their own? It's been haunting me for a while now. 

No need to lecture me about HC and HV safety. half of my dreams involve isolation slots and thermal cut offs. Other half of the dreams involve larger capacitor banks. Soo... thoughts?

Old thread, but new to me!
I am considering repurposing telecom rectifiers as inverter welder. These are extremely well made, and very low cost in the local market. I have may be more than a dozen, of different ratings; the ones in contention are 54V 83A continuous. Also have 27V offerings if more suitable. Using three to be fed by 3-phase 400V, one in each phase, would make for an output of 250A!

Intended as lead acid battery charger, the rectifier has settable inverted L current limit, and a voltage adjust range of 42-59V: still needs to be studied for how to do these via the logic pins at the connector. On an older rectifier, which had somewhat more open layout, I have indeed modded to adjust CV and CC, zero up each. Pretty difficult because pcb's are multilayer, and with very small smd's, all in extremely compact packaging. The wide input range PFC would reduce current demand from the mains, as would 3-phase input (only for high power). Efficiencies are out of this world, 94% on the specification sheet. Resonant mode ZVS.

I should think I will still need an additional external inductor and freewheel, and do not know how to size these.

Just concept yet, so open to all suggestions. Will work on this if I have what it takes.

[soldar]

At the prices these things can be found I would just buy one unless you want to spend a lot of time learning and cussing and want to spend a lot more money in parts.

[Zero999]

Be careful with the layout, paying attention to creepage and clearances.

The main thing which concerns me is there's no Y capacitor, so this thing could be an RF bomb.

[NiHaoMike]

The main thing which concerns me is there's no Y capacitor, so this thing could be an RF bomb.

A welding arc is a huge spark gap transmitter, EMI from the inverter would be the least of your worries.

[najrao]

Thanks for the responses received.
What I proposed was not to be a commercial product. Telecom rectifiers are cheap only because they are released by the first users without even checking for faults if any. They use them in thousands I should think. They are certainly not cheap to manufacture.
That said, repurposing easily accessed surplus electronic modules can only be -- shall I say, invigorating?
The 'cheap' welders available are not built to such standards. I doubt if they even employ PFC, essential to minimize current draw. The telecom rectifiers have excellent input protection and EMI suppression. And you get up to 4kW output from a package not bigger than a quarter of a cinder block. The suggested three phase input should make for a very powerful welder source, 250A continuous. And don't forget the 94% efficiency, and correspondingly very low heat rise.
Issues I sought answers for are: do I need an additional inductor in series, and how to design this and any inevitable freewheel diode. There is also the major issue of paralleling the outputs of three rectifiers, and sharing the current. (They use up to 6 or more in parallel, but I have no access to their backplane, nor control unit which forces equal current sharing). Data for modding is also impossible to gather.
I appreciate any help, thanks friends.