So my friend has a broken TIG welder, an Lorch IT 250 GW to be specific. It's quite old, water cooled, has 250A continuous welding current and is powered by 3x230V mains. It was sitting for about a year in his garage and when he tried to connect it, something blew up inside. After inspection it turned out that two out of four power modules were shorted. This is what it says on the sticker: INV 30/1000-1.4 P. Capacitors in their drivers were also shorted and bulged. After much of searching we've found nothing about them. I've even contacted the company if they have them but this welder being this old is no longer supported. After some reverse engineering it turns out that there are 2 main transformers, and 2 modules are for each one. They are connected in two switch forward topology. After fixing 2 drivers and connecting two working modules I was able to get voltage on the output. Of course it won't weld on just 1 transformer (I didn't tried but I'm fairly sure). We've decided that we can make this switching modules and their drivers from scratch. I've disassembled one of them and they consist of BJT transistors as main switches. It is all encapsulated in resin so I can't really fix them. I've decided that we're gonna use IGBTs as switches. Basing on the label I assume that they are 30A 1000V so I've chosen FGL40N120AND as it is for 1200V and 40A @100*C. As for the recovery diode I went with RHRG30120. These are the schematics that I came with for high side and low side transistors:



The +15V are NOT connected together. There are separate power supplies for them and they were originally in the welder. HVCC and HVGND are all connected together and it is about 600VDC from rectified 3 phase mains input. Inputs are taken from driving transformer, which has 2 secondary windings - one for each driver. It outputs about 20V 20kHz square wave. I stepped it down to about 15V. D6 and D9 are just in case. C7, C8, R13, R14 were originally connected like that. Is my schematic correct? I've tested it on the breadboard with 15V supply, 4ohm resistor as load, 20kHz ~50% square wave and connected in common emitter. I didn't connected D1 and D2 for this test. I didn't saw any ringing on the gate. The rising and falling edge were about 150ns. There was quite large overshoot of 120V peak at turn off. But I think when it's connected in two switch forward it should be biased by D1 and D2. Am I right? Also when I was calculating power dissipation it shows switching losses of about 31.5W and conducting losses of around 100W. I assumed 35A of current, 150ns switching times, 20kHz, 600VDC and saturation voltage of 2,75V. With ambient temperature of 40C and RθJA of 0,5C/W the max temperature is 115*C. Is all of that correct? Or am I just dumb and it's all going to blow up? Thanks in advance!

  • \$\begingroup\$ How are going to sustain 250A? With a 40A rating IGBT? \$\endgroup\$ Jan 5, 2022 at 18:47
  • \$\begingroup\$ 250A is on the output of the welder. This is on the primary side. Stock wires for the transformer primary winding are nowhere near thick enough to carry 250A. \$\endgroup\$ Jan 5, 2022 at 18:51
  • \$\begingroup\$ Current and voltage ratings ought to be derated 50% of expected load \$\endgroup\$ Jan 5, 2022 at 18:56
  • \$\begingroup\$ So if I expect 30A on the primary should I choose IGBTs that can handle 60A? \$\endgroup\$ Jan 5, 2022 at 18:58
  • \$\begingroup\$ Right on...... same for voltage due to inductive transients \$\endgroup\$ Jan 5, 2022 at 19:00


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.