Inverter Welder Strangeness

[Crystaleyes]

Re: Inverter Welder Strangeness

I'm thinking that the problem is with the PWM controller or circuitry, as having another look, but with a fresh head, reveals that the two outputs of the SG3525 are unequal when measured across to pcb main ground.

VCC to ground pins, reads only +0.02 volts, when according to the datasheet it should be between +8 and +40 vdc.

Granted the tester is not a Fluke but it generally gives a good idea

Anyway, taking an AC voltage measurement on the SG3525, it fluctuates around 2.5 v on both outputs, and the frequencies are around 40KHz on output A, and 35KHz on output B.

In 'Frequency/Duty Cycle' test mode however, the readings are fluctuating at just over 170KHz on 'A' and just over 160KHz on output B.
Both outputs float around 61% duty cycle in relation to ground and adjusting the control knob makes no difference at all.

When measuring between both pwm controller outputs, the duty cycle reads a stable 50% with the frequency also stable at 45KHz regardless of knob position.



My lack of experience aside, surely in order for this to work, the switching must be in unison?

So if anyone has any other ideas as to where and what to test, I am all ears.

[CapLeaker]

Re: Inverter Welder Strangeness

probably using wrong GND on your measurement.

[Crystaleyes]

Re: Inverter Welder Strangeness

probably using wrong GND on your measurement.

I had thought about that.

If I use the pwm controller ground then I get no readings at all. Simply zero on the voltage and frequency regardless of the position of the control knob


Unless anyone has any other suggestions I'm gonna order a new SG3525 and see what happens?

[CapLeaker]

Re: Inverter Welder Strangeness

I reread the the thread and it looks like your next logical step is to change the SG3525.

[Crystaleyes]

Re: Inverter Welder Strangeness

I reread the the thread and it looks like your next logical step is to change the SG3525.

It's in the post...


Thanks for the input. This is a steep learning curve in uncharted waters for me.

Nice one!

[Crystaleyes]

Re: Inverter Welder Strangeness

Ok, the parts arrived today. I didn't have the spare money to buy the main caps but have a new controller and IGBT's.

I'm planning on installing the controller first and connecting up through the dimbulb, but one question is,

"should I expact to see a variable output on the SG3525AP controller without the IGBT's connected?"

I can't afford to keep blowing them.

[CapLeaker]

Re: Inverter Welder Strangeness

Then leave the IGBTs out. The gate drive failed because the IGBTs went short on all 3 legs, dumping full power to the gate drive. Assuming you got all the parts and these are good parts, you should be at the same step when you fixed the welder first time go around.

[Crystaleyes]

Re: Inverter Welder Strangeness

Cool.

Wanna change the large caps too but first want to be sure that the chopper section is ok.

[Crystaleyes]

Re: Inverter Welder Strangeness

Ok, so installed the replacement controller chip and the welder seems to be doing pretty much the same as before.. perhaps the frequency drift range is wider (between 30 KHz and 145 KHz) but the control button on the front does fuck all. And yes, I have verified that it is working.

As I said, I have never looked into (nor even at) one of these inverter welders before, but from what I can see it appears to work using pulse width modulation.

Perhaps it only alters the pwm frequency under load?
Perhaps I am wrong about it using pwm and that is why I am stuck?

I spent some considerable time looking online for topology and functionality explanations, but couldn't find anything in clear English which had anything close to the layout of this circuitry.

Why are almost all the online electronic lessons in either unintelligible Hindu or Germanic versions of English?



Anyway, testing the A & B outputs of the controller, or testing the Gate signal of the IGBT's reveals nothing consistent nor affected by the amperage control knob on the front of the unit, which leaves me completely stumped as to what to test next.
One thing that seems strange is when measuring the A & B Outputs with reference to the controller Ground, output A displays a KHz signal whereas output B just goes to zero. If output B is tested with one lead only (no ground), then it too displays a fluctuating KHz output

It's getting to the point that even though the unit is brand new, unless there are any ideas as to what to look for, it might just end up being scrapped for parts and put down to experience.

One thing I HAVE learned, is that it is possible to use a bridge rectifier with only one AC input wire.
Still don't understand how that one works, but there you go...


So shall I just scrap it?

[CapLeaker]

Re: Inverter Welder Strangeness

Scrapping it or not, is totally up to you. Some repairs I made took me multiple attempts too. Sometimes it’s better to put it aside for 4-8 weeks and then get at it with a clear head.
So you got PWM back somewhat. Measure the resistance on output a and b against gnd. I wonder if there is something dragging output b down? There has to be an explanation on why output b isn’t doing anything.

[Crystaleyes]

Re: Inverter Welder Strangeness

Thanks for the suggestion.


So, had a few days break and just had a little look again just now.

Anyways, there is no short between output 'B' and Gnd. As with Channel 'A' it reads several MΩ.

Checking just the voltages in both AC and DC doesn't reveal a lot, other that output A is much less stable than B. Like completely wonky and all over the shop.

Really, I should read up as to how these controllers generate their frequencies. Up until now, I had just 'assumed' that it would be somewhat similar to a 555, in that it will use capacitors and resistors to set the output frequency.

It's late again here and I am only getting late nights free at the moment, but I'll get back on it after another think,..

[CapLeaker]

Re: Inverter Welder Strangeness

Seems like something might be wrong down the line on output A for some reason. I guess more tracing, drawing a schematic is in order.

[Crystaleyes]

Re: Inverter Welder Strangeness

Cheers for tthe suggestion.

The way the welder is designed is that 'after' the pwm controller, the two controller outputs go through what appears to be some kind of amplifying board (there is a mosfet on each channel) before the two yellow wires supply the IGBT section.
Disconnecting this board (and all the output circuitry) didn't resolve the pwm controller output issues, so I had a look at the IC next to the controller.
It is an LM358 twin Op-Amp and taking pin voltage readings appear to show that something is not quite right.

The pin layout is included below and the pin DC voltages are as follows:

1 - 14.7
2 - 0.17
3 - 10.54
4 - 0.041
5 - 0.04
6 - 0.017
7 - 0.083
8 - 15.8

I'm not quite sure where the Op Amp fits in with the PWM controller and it's too late here to start tracing it back but I'm not giving up just yet

Cheers

[CapLeaker]

Re: Inverter Welder Strangeness

See… that’s exactly how I did repair the plasma cutter. Couldn’t really understand what is happening either in the beginning, but after a few attempts you find something that is weird and maybe should not be and build on that.

[Crystaleyes]

Re: Inverter Welder Strangeness

Well, having a think on it after, I've decided that I'm simply gonna list and test every component.

There are only three IC's (the third one being an LNK626), maybe ten transistors, half a dozen diodes, then a total of 70 or 80 resistors or capacitors.
All the resistor values are clearly stamped, so it's only the cap values which I won't be able to verify although it will be possible to identify any short or open ones.

I've already drawn up most of the component list, and have tested (and noted) perhaps 30.

Gonna have a look at, and around, that dual op-amp to see why those readings were so wrong. Maybe remove it and check the inputs?

Thanks again for the encouragement

[CapLeaker]

Re: Inverter Welder Strangeness

I would see what that opamp does and how it fits into this puzzle. Then check it out of circuit with an ohm meter. Maybe your lucky and you measure something. Opamps are cheap enough and quickly replaced too.

[Crystaleyes]

Re: Inverter Welder Strangeness

Ok. My 8 weeks break is up and I'm ready to have another look at this machine.

I've just changed the LM358 and now we actually have even and equal output frequencies on the PWM chip, so woo-fuckin-hoo to that one! About bloody time...

So, now that the pwm circuit seems fine, I think that two new filter caps is in order, so I'll order them and get back to you here.



One other thing which I really want to examine is the input section, for as I said, the bridge rectifier only has one wire going to the AC pins.
How the bejesus that works is like all the covid jab deaths... Baffling!?!?

Anyways, feeling positive that progress has been made and I am definitely learning.. Which is why I bought this broken welder.

[CapLeaker]

Re: Inverter Welder Strangeness

Sweet, that the PWM works now as it should!
That AC rectifier mystery… can you take a picture? What is on the output of that rectifier? I think there is a broken off AC wire and that circuit doesn’t work. What is the part number on that rectifier?

[Crystaleyes]

Re: Inverter Welder Strangeness

Sweet, that the PWM works now as it should!

That AC rectifier mystery… can you take a picture? What is on the output of that rectifier? I think there is a broken off AC wire and that circuit doesn’t work. What is the part number on that rectifier?

There is no broken wire or trace going to the rectifier, as I said, the unit appeared brand new when I bought it, and it powers up and seemingly works with this peculiar configuration.


Below are some photos showing the space where the bridge rectifier goes.
It is a GBJ 5010

If you focus on the second hole from the board edge, it can be seen that there is no trace connecting to it on either side...

So how then, with or without LSD, can the current be rectified in this manner?

[Crystaleyes]

Re: Inverter Welder Strangeness

The welder has 127 volts AC on it's input. The output of the rectifier is a steady 305 VDC - which seems absurdly high!

Measuring across the two input AC pins on the rectifier, gives around 30v. Which I suppose is what happens when you only connect one wire of the AC mains. ????

Looking at these readings however, has revealed that with the main 2200uF caps being rated at 250v... (on a 305v supply).. That the original problem surely has something to do with the caps being underrated, (or perhaps more accurately, 'over-loaded').

Good thing I hadn't ordered the new filter caps yet...