Welder rectifier

First off - Hello All.
I'm a relativley new member that's been reading the forum but never posted.

I'm not an electronic engineer, I just hobby dabble a bit as I need to.

So - I've been trying to find info on using FETs instead of diodes in a full-wave bridge for a welder. I want to turn my old AC welder into a capable DC unit and a power source for a TIG add-on.

Why FETs? Because big (>100A) diodes get very expensive and I suspected that I could get plenty of MOSFETs to parallel-up for current at a better price (surplus) than suitable diodes.

Be gentle.

Theta Engineering: Design Ideas: FET Bridge Has some neat ideas. Not sure if thats exactly what youre looking for, I just googled FET Bridge Rectifier

Many thanks ParkingLotLust - try as I might, I didn't find that link myself.
It's amazing how you change the Google results with a small change in words.

It's pretty much what I was looking for. The use of P and N type FETs to use instead of diodes was what I was looking for.
I suppose (because I don't know) that I could parallel the FETs as well to increase the current capacity?

From the data sheet at Theta Engineering: Design Ideas: FET Bridge, it looks like the limiting applied voltage is 20 Volts...peak...This would be about 14 Volts RMS. You might want to measure the output of your AC welder and see how this compares .... If your AC welder has an adjustable output, maybe it would be just a matter of keeping the output below a certain level.

Hi there,

The general idea is called "Synchronous Rectification" and doing a search
on that will turn up quite a few results.

MrAl said:

Hi there,

The general idea is called "Synchronous Rectification" and doing a search
on that will turn up quite a few results.

Click to expand...

It was used in old valve car radios - they used a vibrator to feed a transformer to generate the HT voltages. The vibrators came in two types, non-syncronous (which required a rectifier - usually a metal type), and syncronous - which had extra contacts on the vibrator that mechanically rectified the output.

Funnily enough a friend of mine has one for repair at the moment, out of an old Aston Martin, and the owner want's it repairing, cost no consideration.

Hi bill-gj,

I don't recommend a full wave rectifier for a welder. A full wave rectifier requires both N-channel and P-channel MosFet transistors. P-channel transistors generally have a much higher RDSON than N-channel types which restricts current flow. Also the correct timing is crititcal with a full bridge rectifier. Resistors slowly age and out of tolerance an N-channel and a P-channel transistor might conduct simultaneously which means a dead short.

I made a two-way rectifier circuit using two N-channel transistors and an operatiol amplifier. It works pretty well.

You can have a schematic and a PCB layout via email, if you want.

Boncuk

Hello Boncuk, yes, please, I'd like to see the circuit if that's OK.

The prospect of just using N-types is interesting as, unless I am mistaken, they're more commonly available and cheaper than P-types (?)

Hi bill_gj,

here is the schematic of the circuit. This example can deliver 5A at 15V (transformer output voltage 12V) without heat sinks on the transistors. The transformer used must have a center tap (AC2).

Using a solenoid transformer and two IRFZ48N you can go as high as 27V output voltage (with load connected) at 10A. In that case the transistors require a small heat sink. (The max voltage is the supply voltage restriction for safe operation of the OpAmp, (32V))

Using a TLC272 instead of the AD822 OpAmp use the original TI-manufactured IC. It is the only one which switches to zero, which is a must for the circuit function.

Measuring points 1, 2 and 3 are given for best adjustment. Adjust 50 - 100mV (no DMM! Oscilloscope only!) between 1 and 2, respecetively 1 and 3 with pots P1 and P2.

The circuit uses the parasitic diodes in the transistors to gain the supply voltage for the OpAmp, which is already present after the first AC cycle.

And yes, N-channel MosFets are more widespread than P-channel types, hence they are to be cheaper.

If you PM me your email address I'll send the Eagle files, including schematic, PCB-layout (single sided) and BOM. I'll translate the circuit description from German to English for you, too.

Nigel Goodwin said:

It was used in old valve car radios - they used a vibrator to feed a transformer to generate the HT voltages. The vibrators came in two types, non-syncronous (which required a rectifier - usually a metal type), and syncronous - which had extra contacts on the vibrator that mechanically rectified the output.

Click to expand...

I'm no expert on the matter but wasn't the transformer and vibrator one module?

I would've thought it'd be easier to wind an HT secondary round the coil of the vibrator rather than having a separate transformer.

I'd imagine that they would be very noisy, all the EM emissions would've made the rubbish AM radio totally unintelligible.

Hero999 said:

I'm no expert on the matter but wasn't the transformer and vibrator one module?

I would've thought it'd be easier to wind an HT secondary round the coil of the vibrator rather than having a separate transformer.

I'd imagine that they would be very noisy, all the EM emissions would've made the rubbish AM radio totally unintelligible.

Click to expand...

Hi Hero999,

in Germany the vibrators were separate units and plugged on an octal socket, easy to replace when the contactors were worn out. The vibrators themselves were pretty noisy (audible), but there was no interference with radio reception since the filter network was quite impressive.

Hans

Hero999 said:

I'm no expert on the matter but wasn't the transformer and vibrator one module?

Click to expand...

No, they were separate - vibrators been mechanical were replaceable items, and like Boncuk said, were often plug-in items. We used to stock a range of them at work, before my time really, but we had various ones left in stock when I started (I probably threw them away years ago).

I'd imagine that they would be very noisy, all the EM emissions would've made the rubbish AM radio totally unintelligible.

Click to expand...

Well filtered, no problems - but (again like Boncuk said) they were mechanically noisey.

Nigel Goodwin said:

We used to stock a range of them at work, before my time really, but we had various ones left in stock when I started (I probably threw them away years ago).

Click to expand...

What a waste of good material!

You could have converted them to factory break bells and sold them!

Hans