Telwin digital welding machine

[hansdampf]

Hello EEV forum!

I recently got my hands on a digital welding machine but unfortunately it appears to be dead. I would appreciate any help to figure out what's the cause so it can be saved from the junk yard. First of all, I'm an electronics beginner with some sort of idea how stuff works but I'm far from claiming I know what I'm doing here. Please bear with me if I get things wrong.

The machine is a Telwin Digital MIG 180. It's a MIG/MAG welding machine with digital control panel that automatically adjusts certain parameters such as gas flow, welding wire speed and welding current. Unfortunately the control panel doesn't do anything. From a Youtube video I found out then when powered on the three 7-cell digits should light up and do some sort of self test and the LED's should light up too. Also the buzzer should make some noise. None of that happens.

The board is powered by 15V AC and 24V AC. They are converted on the board to 5V DC and 12V DC. I measured the output of the voltage regulators and they are working fine.

The main controller is a Motorola HC908GP32CB. I checked the power supply pins and they receive 4.95V, which is what the chip expects.

There is a white component glued on top of the main controller and it's connected to power supply + and RST pin (thermal fuse of sorts? It has printed on: 47nK100). It's not shorted so I don't think the chip is resetting all the time. I measured between the legs 4.95V so RST is pulled down.

I measured all capacitors on the board and none of them are shorted.

And from here on I am out of ideas what else I can check. I'll attach pictures of the board and I have the manual with a block diagram of the whole machine. Unfortunately there is no schematic of the board itself and the information online about this machine is very limited. If there is anything that might help please let me know.

Thanks for all ideas and suggestions in advance!

[hansdampf]

I found an updated version of the block diagram with more information on it. Attached below.

[PaulAm]

The fact that the readout is dead is not a good sign.  Can you source a replacement board from the manufacturer?  If you can find a part number, you might find something on one of the Chinese sites.

Worst case, these types of welders are not rocket science.  The voltage is set by the combination of L1/L2 and the control board probably sends a DC PWM voltage to drive the motor, using the current shunt as feedback.  There is nothing inherently complicated about that.  For one cheapo welder I built a replacement wire feed unit including a motor driver with opto feedback on the motor speed.  Ended up working really well and I got as good or better welds compared to a Miller 250 I later picked up.  Still works too well to get rid of.  Because I could, I added a spot timer, gas purge switch, etc. etc.

Varying the motor speed based on welding current likely helps improve the weld, but you'll still have a good welder if you can just get a speed control with some feedback to keep the speed constant.  Some controllers used back EMF for that, so there are a number of options available.

[Gyro]

There is a white component glued on top of the main controller and it's connected to power supply + and RST pin (thermal fuse of sorts? It has printed on: 47nK100). It's not shorted so I don't think the chip is resetting all the time. I measured between the legs 4.95V so RST is pulled down.

That's a 47nF capacitor, probably tacked onto the reset line as a factory mod to cure a noise pickup problem (or to extend the power on reset time).


P.S. Welcome to the forum.

[mikec]

Had a search for some schematics, no luck though.

This company appears to sell control boards for that range of welders

https://www.bricoutensili.com/en/telwin-spare-parts/8213-front-panel-module-kit-for-welding-machine-telwin-digital-mig-220-222.html

This is not the exact version you have, though the LED layout for the front panel is the same and afaik the connector layout is the same as well. Its 264euro however so not cheap.

It might be a contact to get schematics or you could contact Telwin I have seen other schematics and repair documentation (scope plots etc) from Telwin so they may be open to providing a schematic
or maybe a cheaper refurb for your model of welder

https://www.telwin.com/en/assistenza/richiedi-informazioni/


The fact that nearly 20 leds, the display and buzzer are all dead on power up and you have verified at the least the power rails are correct points at the micro-controller. From what I can tell all of these
are likely driven by the micro-controller (they provide information about the settings of the welder on the front-panel) and unless they are tied to a common supply which is not working then the issue seems to
be around the micro-controller.

Basically at this point I'd like to see some life from the micro-controller, for example that large relay on the
back of the PCB used to turn on the welding circuits does that respond when you start welding (presumably pressing a button on the remote control)?
I notice the TX/RX pins of the micro-controller appear to run over to the black connector beside the potentiometer, are those pins available to connect a +5V TTL Serial-to-Usb adapter, any coherent data been output here by the micro-controller (i've done this before, when all else failed at least i could be confident the micro was alive  ), they could just be configured and used as I/O off course.

Is that a ROM chip to the top-left of the micro-controller (can't read the part number) if so how is it connected to the micro-controller?

[hansdampf]

Thanks for all the replies and welcoming me to the forum! Appreciated!

Even though I was hoping there would be a simple solution, it narrows down that the problem is the micro-controller. That's where this projects ends I'm afraid due to a particular lack of skills on my end I'm trying to get in touch with the manufacturer and the supplier mikec mentioned. Maybe I get get some insight from there or a replacement for a good price. In case they would provide schematics I'll let you know.

Thanks again!

[hansdampf]

@mikec above the MC next to the displays are two ULN2803A (transistor arrays). Below them are two HC373 (OCTAL TRANSPARENT D-TYPE LATCHES) and left of the MC is one HC541 (Octal Buffers and Line Drivers). There is an eeprom underneath the displays: https://pdf1.alldatasheet.com/datasheet-pdf/view/23723/STMICROELECTRONICS/24C04.html But I can't tell how it connects to the Mc. There are two traces going that direction but the go under the MC and I can't see where they end up.

If that would be broken I assume just replacing it wouldn't help as it needs to be programmed, wouldn't it?

[mikec]

So that ROM is an I2C ROM, the MC has some I2C physical compatibility on the SPI port so those 2 wires could be routing over to pins 17 and 18. That ROM is likely storing the settings of the welder, how the MC/Welder would respond to a new blanked ROM is anyone's guess tbh. In fact how the MC/welder would respond to a faulty ROM would be guess work as well, one would like to think the firmware on the MC handles it by at least putting out a fault code on the display but ....

That said one of those lines is a clock and could be used as a diagnostic to see if the MC has some sort of alive pulse.. so if you got a clock signal on Scl (of the I2C) that would suggest MC is powering up and initializing.

Do you have an Oscilloscope/logic analyser btw? Also importantly some of the safety training/self thought/experience required when working with mains equipment?

Just to add to a previous post regarding the LEDs/7-segment displays, i made the point that the LEDs and 7-segment displays are driven from the MC and therefore the issue points strongly to the MC UNLESS there was a common-supply. Looking a bit further and trying to work out how the LEDs and 7-Segment displays are driven that point about a COMMON-SUPPLY is something that I'd be inclined to check.

1. As far as i can tell the LEDs/7-segment display are actually part of the same drive circuit, on the 7-Segment displays you you have a total of 3 x 8 = 24 leds effectively. If you look around the board you can count 20 status/configuration Leds populated and 4 un-populated = 24 (or visa-versa 3 more 7-segment displays with period)

2. There is no way the MC is driving all of these separately, there wouldn't be enough pins on the MC to do this. The MC is probably using 8 pins to drive all LEDs and the 3 7-segment displays. (you can look up circuits on how this is done if your not familiar). This brings back the point about the common-supply to these LEDs/7-segment displays, if that supply is dead than you get nothing on the welder front panel.

3. One notable place where this common-supply could be is on the ULN2803A transistor arrays (Pin 10 on the datasheet). I'd certainly probe pin 10 on these ICs to be sure. There does appear to be zener-diodes located on the PCB very close to these transistor arrays, these could be providing the common-supply voltage, its not impossible that one of these is dead.

Without schematic and the board at hand I am assuming the transistor arrays have anything to do with the LEDs/7-segment display, i am really just basing it on their geographical location on the board. Its tricky to trace PCB tracks with pictures only 


Regardless I think you have gone the right route for now in contacting a supplier and hopefully getting the support needed to get the welder working.

I would be interested if you got any response from that supplier, were you able to get a decent price for a replacement or even a schematic?

[fzabkar]

I measured between the legs 4.95V so RST is pulled down.

According to the datasheet, /RST (pin 6) is active low (see page 39):

https://datasheetspdf.com/pdf-file/578958/Motorola/HC908GP32CB/1

[hansdampf]

Hey guys,

Unfortunately I wasn't able to solve the problem with my board. However, I ordered a replacement from Italy through the website mikec found and I'm happy to say that the machine is working again. I wish I would have been able to trace the problem down to a faulty component and fix it just for the satisfaction of it but maybe next time. Thanks for all the help! Take care!