Inverter Welder Strangeness

[Crystaleyes]

Re: Inverter Welder Strangeness

Nope. But you can remove the IGBT's and check the gate drive.

So that is what I did...

Without the IGBT's connected the fan and display once again return to life..

None of the resistors, nor diodes on that part of the Gate drive show any obvious deviations.


Up until the two yellow wires leave the small horizontal board, the frequency and voltage measurements are very similar.

Once the two yellow wires connect with main pcb however, the two channels' readings deviate away from each other.

As it is Silly O'clock here now, I'll post some actual readings and photos of the components in the morning...

Definitely something strange going on here

[CapLeaker]

Re: Inverter Welder Strangeness

Nope. But you can remove the IGBT’s and check the gate drive.

[Crystaleyes]

Re: Inverter Welder Strangeness

Cool. I'll check all that later.

As for the pwm controller, should it have a detectable output, with the shorted IGBT's still in circuit?
I'm thinking of whether to just replace it before confirming it is fecked?

[CapLeaker]

Re: Inverter Welder Strangeness

#991 is what I am looking for. So on the gate drive are these resistors and a diode. Check these very carefully and make sure they are within their tolerance. If there is anything not right in that stage, the IGBT will blow.

[Crystaleyes]

Re: Inverter Welder Strangeness

Yes. PWM, all component's between it and the IGBT.

Roger that!

Let's give it a go... (again... as the pics refuse to upload - grrrrr)

The first pic #981 shows the view from the SG3525AP pwm controller across to the board which sends the signals to the IGBT's gates. IC2 is an LM358 (I believe)

#998 is the view from directly above

#985 shows that the two pwm outputs A & B (pins 11 and 14) cross through the board to the other side.

#994 shows the hole on the reverse side and that it connects to pin 1 of the perpendicular 'comms' board

#995 shows the PWM output B connecting to pin 5 of the same board.

#996 shows the board in question. Pin 1 is at the bottom. VT1 and VT2 are 4688 MOSFETs (datasheet below). The two yellow wires are the outputs to the inductor before the IGBT's.

#997 is just another angle

#989 shows where the yellow wires go to, and that one leg passes through capacitor C26

#877 shows the 6 pinned transformer/inductor. The two yellow wires connect to the right hand side.

#991 Shows the underside of the 6 pins. The left hand pair connect to the yellow wires, Without confirming, I believe that the centre pair connect to the IGBT Emitters, and the right hand pair connect to the IGBT Gates via two SS14 diodes and some resistors

#803 shows the board with the IGBT's and heatsinks removed. The two diodes are the SS14's



More to come but it is late here

zzzz

[CapLeaker]

Re: Inverter Welder Strangeness

Yes. PWM, all component’s between it and the IGBT.

[Crystaleyes]

Re: Inverter Welder Strangeness

Is there a way to see a pic of the IGBT with the gate drive?

I would imagine so.


Not quite sure what you mean by 'Gate Drive'?

Are you referring to the circuitry between the pwm controller and the IGBT gate?

[CapLeaker]

Re: Inverter Welder Strangeness

If the IGBTs shorted to the gate, then it dumped full load to the gate drive circuit. So you have to inspect everything in that gate drive circuit, including the PWM. Is there a way to see a pic of the IGBT with the gate drive?

[Crystaleyes]

Re: Inverter Welder Strangeness

I personally would replace PWM if I were trying to repair this device unless you for sure that it is functioning correctly

It possible that the IGBT module might have taken a shit because of the PWM controller chip malfunction in some way or some other device malfunction

Agreed.

Earlier today I'd decided to replace the pwm controller along with the IGBT's and both of the fat caps, so shall get on with that. One of the recommendations in the manual, for low power output, is to change the large capacitors,


Correct me if I am wrong, however it seems to me that this welder uses pwm switching in much the same way as an SMPS would, and therefore, if there is no (or limited) 'pulsing' then the full circuit current will go through the IGBT's and fry them.

One thing I have to admit that I don't understand is the AC mains input not taking a direct path to the bridge rectifier, nor even connecting to one of the AC input pins on that rectifier.
That is a new one on me, and I'm struggling to see how the DC supply can be acheived in this way? As in, creating a two wire +ve and -ve DC supply from one AC wire?

As far as the PWM controller goes, would I be correct in assuming that controller would have a measurable signal on Outputs A and B, without the IGBT's being in circuit?

[sam_sam_sam]

Re: Inverter Welder Strangeness

I personally would replace PWM if I were trying to repair this device unless you for sure that it is functioning correctly

It possible that the IGBT module might have taken a shit because of the PWM controller chip malfunction in some way or some other device malfunction

[Crystaleyes]

Re: Inverter Welder Strangeness

What’s the make and model of this darn thing?

It's called a Pro Euro GP-155A - 127 volts

Sold in Brazil - Made in China (manual included below)

Seems to me to be a basic generic design. It certainly isn't the most robust piece of kit, but I bought this to learn how to repair it.


The situation update is that as mentioned, the two IGBT's blew for no apparent reason, whilst the unit was running. Last night I installed two new ones and ran the unit on the bench for a good hour without issue, at first through the dimbulb and then through the normal socket. The display was fine and the fan was blowing , whilst there was 30 VDC across the output terminals. When I first got the unit which wouldn't arc, there had been around 70 VDC.
This morning I took it to the garage to see if it would now weld and as soon as I switched it on, there was a flash the wall socket and the unit went dead. A quick test shows that the IGBT's are once again shorted, and checking the plug, it had been burnt in the socket.

The only difference between running the welder last night on the bench, and this morning in the garage, is that the garage socket is a bit oxidised. Whether the unit is so fragile that a poor plug contact would blow the transistors is surely doubtful, so I need to start thinking about what other fault would cause the IGBT's to blow when not under load? And seemingly an intermittent fault...
Do PWM controllers fail intermittently? As mentioned yesterday, taking readings last night showed no variations as the current controller was adjusted and after the incident this morning, there is no longer any output from the PWM controller.

Anyway, I now have to reorder the transistors which is a fucking pain.


Onwards and upwards

[CapLeaker]

Re: Inverter Welder Strangeness

What’s the make and model of this darn thing?

[Crystaleyes]

Re: Inverter Welder Strangeness

As long as she works, but having a weak output, I would suggest the big caps are leaky. If there would be a short on any of these fets, the unit wouldn’t be working like that. Problem with properly testing these kind of caps, you’d need to do a leakage test. ESR wouldn’t do you any good here.

Leaky big caps could well be the case. One of the 2200uF caps measures only 2000uF..
.. In fact looking online, they, along with the IGBT's are commonly sold as the repair solutions for these welders.


One other thing which I was wondering is whether the SG3525A PWM controller might be bad?
I've never actually tested this part of a circuit in a working SMPS however this controller is reading around 160KHz, with around 30% Duty Cycle, on both of the outputs, and doesn't change as the current control knob is adjusted.

Surely that should change?

I would have expected to see some change in the output of that chip, but I'm here to learn.

If you want you can read up on a plasma cutter that I repaired. Maybe that gives you some hints.

Had a good read through that, thanks, and unless someone declares that the PWM is working correctly, then I'll have a probe around on the board for bad components.

[CapLeaker]

Re: Inverter Welder Strangeness

As long as she works, but having a weak output, I would suggest the big caps are leaky. If there would be a short on any of these fets, the unit wouldn’t be working like that. Problem with properly testing these kind of caps, you’d need to do a leakage test. ESR wouldn’t do you any good here.
If you want you can read up on a plasma cutter that I repaired. Maybe that gives you some hints.

[Crystaleyes]

Re: Inverter Welder Strangeness

Did any caps split open and "VEHT" ?

Split and vent.. no

That is not to say that one or more of the caps isn't bad on the ESR side of things. I'm kinda limited with capacitance testing right now.

check continuity from the bridge rectifier to the mains plug.

There is no continuity.
The mains input either goes to one side of the relay contact (and then on to one pin of the bridge AC), or to one side of both caps.

caps looked connected as normal to me .

Filter caps are normally found 'after' rectification, no?



And speaking of that... How can the AC supply be 'rectified' if it is not actually connected to the rectifier?

In the following photos, it can clearly be seen that the AC terminal next to the -ve pin, goes precisely nowhere, nor has any input.


How on earth does that work?

[eccerr0r]

Re: Inverter Welder Strangeness

Did any caps split open and "VEHT" ?

[petehall347]

Re: Inverter Welder Strangeness

check continuity from the bridge rectifier to the mains plug .
caps looked connected as normal to me .

[Crystaleyes]

Hola peeps.

I'm placing it in this topic, because from what I can see, it is essentially a modified SMPS, but please move if there is a 'correct' forum section.


Recently, I bought an essentially brand new, but defective, 155 amp inverter welder.

The unit would turn on fine, with the digital display and fan doing their thing, however when attempting to strike an arc, the rod would merely spark or just get stuck onto the steel.

Now I know that many (if not most) in here would simply not have bought it, but I am still learning this trade and had never even looked inside one of these before, so for the equivalent of 25€, it is a good opportunity to learn.

Other than not striking arcs, the only other noticeable oddity was that after switching the neon illuminated mains switch off, the neon would remain on until the large capacitor(s) had discharged. And the plug pins would also give a sharp shock

Normally, the lack of arc symptoms could be simply a dirty or poor connection on the main leads, and there was actually some play in the two large bars which connect the pcb to the hefty 16mm cable terinals on the front panel, although tightening made no difference.

So without removing any parts, I looked for any obvious damge and basic tested the obvious for shorts... The large caps, the two IGBT's, the PCT, main relay, the two output diodes, etc. There is no fuse protection. One resistor next the the 7805 has almost no markings left on it.

After several attempts at welding down in the garage, I put the open unit on the bench to have a probe around and take some voltage and frequency readings.
When the welder had been turned on a few minutes, but before I had done any readings, there was a puff of smoke from the area of the relay and the unit switched off. Including the previously ever-glowing neon on/off switch.

Now, I haven't yet understood how this welder should be working, but with my somewhat limited understanding of SMPS design... IF it is based on the switching supply system, then the PWM controller not functioning correctly would limit the main current output, and any arcing capability.
Anyway, this puff of smoke event, took out the two 45 amp IGBT's and their two SS14 smd diodes.
The only chip in the region of the smoke is an LNK626 'ffline switcher' which I don't really understand what it is doing, but it isn't the PWM controller as that is an SG3525 which is located towards the end of the board,

So today, I put the unit back on the bench with the intention of drawing out some kind of circuit diagram in order to see if a better understanding could be gained.


What stood out immediately is that the two large electrolytic filter caps are connected before the bridge rectifier. This is a new one on me...
Even more unusual in my world of limited electronics understanding, is that one of the bridge rectifier AC terminals has no AC going to it, nor is the terminal connected to any other part of the PCB



I have drawn a very basic sketch of the first part for people to have a look at.

What is going on with the bridge rectifier, why are the caps before the bridge, and does anyone have any suggestions as to what else to check?

Thanks in advance