Repair an inverter welder

[edugimeno]

Hi! I got this (see pics) welder which wouldn't work at all at the welding leads. I tested output voltage and got 0v
Opened it and found a diode that was busted (missing resin...). Ref RHRP1560.
I believe this diode protects one of the IGBT's that drive the transformer
His twin diode wasn't physically damaged but measured OL on both polarities so both will be replaced
Now, after watching some videos on welder repairs, it seems that the IGBT's usually fail together with the diodes, is this true? I tested IGBT with multimeter-diode mode, still on board, and got OL on all pin combination, both polarities.
Should I go ahead and swap IGBT+Diodes?

Thanks!

PS: Rest of electronics "could" be ok, at least I checked +12v at the largest electrolytic in the control electronics area...

[edugimeno]

Re: Repair an inverter welder

Also, there are 2 large diodes after the IGBT's output, driving each side of the wave to the transformer. I checked them in a DMM and they look good.
Anyone?
Thanks

[edugimeno]

Re: Repair an inverter welder

Anyone? Thanks

[Crystaleyes]

Re: Repair an inverter welder

Hola Edugimeno.

Some more photos will help, but without more details, I can tell you for nothing that 12v on 400v main filter caps is way too low (at least if the caps are in a standard configuration).

You first need to confirm that your 220v AC mains is getting rectified to DC, then you need to confirm that there is also your low DC voltage to power the chips, this is likely via an LM7805 (although I can't see one in the photos).

Check the IGBT's for shorts and if shorted then replace them. I wouldn't change things just because you can, but you do need to verify that second RHRP1560 diode.
Your vertical doughnut inductor is also burnt and has a loose wire.


The way these inverter welders often work is not too dissimilar to a Switch Mode Power Supply, only the output handles a much higher current than your average smps.
There are various ways in which these machines are configured so I won't try to guess what is happening in yours without more details.


It will make it easier to receive help if you can give the any make and model details.

[edugimeno]

Re: Repair an inverter welder

Thanks a lot for your reply. Before I go ahead and do some of the tests you mentioned, just some points:
1.-The 12V I was refering was not the one after the rectified AC bridge, that one has the right voltgage (>300v); i meant the smaller capacitor near the electronics which is probably after a volt regulator (7812 expected but I didn't check)
One of the recitifiers was physically blown as in pictures and the second onewas shorted (DMM tested)
I realized the output inductor has one burnt wire but it still has 100% continuity (with reduced max current theorically) as this is a single coil with 10-15 wires going all together in parallel. If I get all this back to work I will try to solder that loose wire
I will check all the rest when Im back from vacation, thanks a lot!

[edugimeno]

Re: Repair an inverter welder

Ok I was able to take a quick check and pictures from the board before I leave for vacation
I tested both IGBT (still on board), nothing in continuity (not shorted) and got nothing either in diode mode (combining all leads and all polarities). Are they all open? I can't see the model until I desolder them
The relay coil is 12V so I guess this is a good voltage for the electronics (I measured 12v on the control side capacitor)
I took some close pictures from all sides. Note that the protecting diodes for the IGBT's are removed from the board. The other 2 output diodes are testing right with the DMM

Thanks!

[CapLeaker]

Re: Repair an inverter welder

I can see on your very fist picture first post there is an open inductor on the very bottom left!

[Crystaleyes]

Re: Repair an inverter welder

That SG3525 is your PWM controller and if you have a frequency option on your DMM you can easily verify if the chip is working properly, by measuring outputs A & B to Gnd.
The LM339 is the comparator which monitors the PWM chip.

There is some serious dirt on those heatsinks so wash all that off too.

[edugimeno]

Re: Repair an inverter welder

I can see on your very fist picture first post there is an open inductor on the very bottom left!

Yes, thanks for checking, read comment #5 (it still has continuity for tests)

[Crystaleyes]

Re: Repair an inverter welder

Do you know the make and model?

[CapLeaker]

Re: Repair an inverter welder

Did you make sure the wires wound on the inductor aren’t shorted? Kind of hard to do using only a DMM, since you really can’t put power to it to see if the insulation broke down.

[edugimeno]

Re: Repair an inverter welder

Do you know the make and model?

Yes, I just checked, its a Deca MOS 170

Thanks

[edugimeno]

Re: Repair an inverter welder

Did you make sure the wires wound on the inductor aren't shorted? Kind of hard to do using only a DMM, since you really can't put power to it to see if the insulation broke down.

It doesn't look too bad, the cut wire is not stuck to the rest of the wires. The rest of wires don't look dark at all and moving them with my fingernail they all seem loose

Thanks

[edugimeno]

Re: Repair an inverter welder

Ok today I returned and had time to check all this again
I removed the IGBT's. Both are OPEN. One of them has a missing chunk of resin. The other one looks fine outside but still measures OL on all pins all polarities (OHM and DIODE mode), so I guess they both are blown
IGBT's are G30N60A4
Diodes are RHRP1560
Should I go ahead and orden these 4 parts?

Thanks!

[CapLeaker]

Re: Repair an inverter welder

Before you go and order stuff, check that gate drive circuit for the IGBT’s. Because if it’s blown, your new IGBT’s gonna blow again!

[Crystaleyes]

Re: Repair an inverter welder

Does your multimeter have a Hz/% option?
If it does, with the machine switched on, put one DMM lead onto pin 11 and one onto pin 14 of the SG3525.

You should get a reading of 10's of KHz and around 50% (depending on the quality of your meter).




And how did those diodes test?

[eccerr0r]

Re: Repair an inverter welder

Be careful about duty cycle, it shouldn't be producing pulses or be near 0% duty cycle if there's no load.

Speaking of igbt's I recently replaced two in an inverter, one was shorted all ways, other was shorted C-E. I'm surprised of these transistors failing open all ways, then again if it blew a hole, okay I can take that going all open.

[edugimeno]

Re: Repair an inverter welder

Crystaleyes Yes I do have a multimeter with Hz and I also have a digital oscilloscope so I will try that tomorrow.

I guess I can go straight and read the IGBT's gate pin right?
What diodes do you mean? The protecting diodes at the IGBT were all blown. The large rectifier diodes at the output of the inverter tested good with DMM in diode mode (0.5 or 0,6 in in direction, OL in opposite)

Eccerr0r I can't use any load for now as this board doesn't have any working IGBT's. I want to see what I get in my oscilloscope screen... I will probably send a picture here with different AMP settings in the front panel

[eccerr0r]

Re: Repair an inverter welder

Ah okay maybe it should produce something as it tries in vain to bring up voltage, not sure if it has a shutdown if it fails to do so or not. Just a warning that you may not see what you expect.

[Crystaleyes]

Re: Repair an inverter welder

Crystaleyes Yes I do have a multimeter with Hz and I also have a digital oscilloscope so I will try that tomorrow.

I guess I can go straight and read the IGBT's gate pin right?
What diodes do you mean? The protecting diodes at the IGBT were all blown. The large rectifier diodes at the output of the inverter tested good with DMM in diode mode (0.5 or 0,6 in in direction, OL in opposite)

Hi Edugimeno

Forget the IGBT's for now.

Before you replace any parts, you need to identify at which point in the circuit you have a problem.

'In my opinion' a good place to check is the output of the PWM controller, which is before the IGBT's and output diodes.
You know which PWM controller pins to test, and you know what a good controller output will be.
'If' the controller is bad, then replace that before you start dumping money into new IGBT's and switching diodes.

I'm not sure what else to suggest until we know what the pwm controller is doing?

Please do the tests and let us know.

If you need a clearer explanation then ask...