Thread: any up to date elctronics guys here/ broken welder

RTV? Is this stuff solid then?

I'm used to HTC (heat transfer compound) which looks and feels like silicone grease but it doesn't solidify. It's used for mounting transistors onto heat sinks etc.

The white goop looks like it's being used for strain relief, as they're running the wires straight into/onto the board with the risk that the wire will rip away the pad from the board. Better quality designs would use headers and plugs, but that would add expense in terms of components and terminations, and effort to route the relevant connections out to a pin header.

There's a guy on the OCAU electronics forum who services welders - dunno if he does board-level work tho, as there's probably no money in it.

The white goop looks like it's being used for strain relief

I'm with that.

Especially if the assembler/wave solder inadvertedly caused some wicking in the wires.


An old school mentor from a previous life reconed he had come across some strain relief stuff in consumer type electronics (VCRs) [mainly used on large caps] that was going low resistance over time. This guy knew his stuff; wish I was still in contact with him and has access to his shop.

Good luck with it.

Thanks for the replies. Now you've pointed it out, it should have been obvious because its mostly on cables and plugs right enough. What threw me was it was also on the transistors I wanted to remove for testing, and they didn't have wires. However, its the same stuff as on the wires, so I'm not going to worry about it.

This thing is denitely not made to be serviced. There are lots of wires soldered directly between boards. It may not be the best plan, but I ended up labelling and cutting the wires rather than unsoldering one end and trying to mark where it should go back later. They used a lot of connectors as well, it doesn't make sense to me, surely either use connectors or don't.
They still use that non-setting thermal heat paste between the transistors and the heatsinks. I seem to remember from the old days that stuff was quite toxic. Being chinese, it probably still is.

The repair shop said it just wasn't economic in his time to try to trace circuits, so hence the reason they only swap boards. From the little I've found that people have written on what it cost to repair their welders, one guy said his miller welder cost $1500 for a board replacement (the same as my welder) another guy said a shop repaired his chinese one for $600 and another said he was quoted more for the repair than a new one. Considering they will charge you just to look at it, I think its either me that fixes it, or I think about getting a welder which I have found diagrams for.

"getting a welder which I have found diagrams for"

Thats a good idea, can you post the links

Ken

Originally Posted by Gringo

"getting a welder which I have found diagrams for"Thats a good idea, can you post the linksKen

Sorry, I have been looking for days and haven't kept any links. I suspect that the service manuals I did find should not be free anyway, so maybe shouldnt be on the forum.For you info, the files names of the most useful wereMasterTig ACDC Service Manual.pdf - Kemppi welder. Costs about $6k. This had the closest digrams to my circuits and was the most useful.tig250gas_service_manual.pdf - Snap on welder. Good manual for generic overview of what the boards do, but no detailed circuit diagramsDocLib_2821_0-5001 Transtig 200 ACDC Service Manual_AA.pdf - CIG welder. Really only useful if you've got this welder. it shows you how to get at each board, but not so much cetail on the boards themselves.I found a miller one and some other things but can't find them now but they weren't so detailed, or too detailed on that specific model with fauilt codes etc. There is some russian genius who designed his own welder and there are diagrams for that, but I found them of lmited use. I think any of these would be all pretty easy to find if you search. There is a yahoo forum on chinese welders and if you have a harbor frieght welder, I think there were some diagrams in there that might help. Its still hard to know what to buy if need a new one. I could buy four of my welders for one Kemppi. If I was going to buy a Kemppi I would probably buy a miller instead, which is even more expensive. Mine stopped dead when I was using the plasma cutter. I "stuttered" my finger on the trigger and did three on-offs in short succesion by mistake. I still hope I can fix it but I don't think I'll use the plasma section again. Cheap plasma cutters seem very prone to dying, from what i have now read. Maybe there's something to be said for everlast and their 5 year warranty. Their one, equivalent to mine, is about $700 more than mine was off ebay. So you pay $700 for insurance for 5 years.

I'm assuming this is an inverter type welder, which has a switch mode main power supply

Yes, it had AC up to 250Hz ac output, pulse etc. It worked great, when it worked. I'm still hopeful I can fix it, just waiting for parts.

Well, not an "up-to-date" electronics guy, but a DIYer who has fixed a few electrical/electronic things, often both with and without the aid of circuit diagrams. Seems usually the culprits are: electrolytic caps, diodes, transistors. But the network switch that died after an electrical storm needed a surge protection device.

I have found more equipment can be helpful. I read on an electronics forum about ESR capacitor testers, apparently invented by Bob Parker a Telstra guy. So I found Peak electronic UK, who make a load of nifty test gadgets that will do what a multimeter can't. So I have four of the Peak testers for various things - Atlas ESR, LCR, SCR, DCA, . I figure if the gadget can be bought for not a lot more than a repair quote fee, I'm happy to keep fixing my own stuff. Recently got an oscilloscope to play with (still learning). The dearer ones have more of the white stuff holding the components - presumably to control movement and inductance or capacitance between components. I assume it is white neutral cure silicone.

see What kind of glue should I use for PCB-mounted components to avoid vibrations? - Electrical Engineering Stack Exchange

EG This one will test IGBTs. (I have the lower model DCA that won't, so might have to upgrade when the welder decides to die.) Peak Electronic Design Limited - Atlas DCA Pro - Advanced Semiconductor Analyser - Model DCA75

Peak will ship direct for not a lot extra, plus you get to miss paying VAT. Prices for their testers if findable locally are a whole lot more - check out RS Australia for instance. RS can be good for hard to find parts, ship for free - I think they want an ABN to join up.

Absolutely not worth paying for board replacement when it's just a cheap component. Mostly takes a bit of time and a few dollars worth of parts.

Seems a bit of a cheap/nasty board with those long solder tracks.....

You've more faith than me for finding a fault without a diagram then. It's not easy at all to trace out circuits when the thing is not designed to be taken apart easily and there are so many wires dissapearing into the chassis. To get to the board with the faulty diode, I had to cut about 20 wires, dismantle the chassis so it was all loose and try not to let the weight damage other boards/wires while pulling the board I wanted out. I have replaced the diode and wirewound resistor but its still not starting up. They may have been a red herring because they haven't blown again. I've let it run now, and done some tests and its the control board that is shutting it down, so I will start with the osciloscope this week. Time I've spent on it already... and I hate electronics and wouldn't do it for a hobby.

Radio spares is great. They were always really expensive, but you could always find what you wanted. Nowadays, their components are not too bad, but don't even think of things like tool steel or whatever. To be honest, if you just type the part number into ebay, you can find most things direct from china or hong kong for single components even. I've been taking a risk on items building up in price for ages now, as long as the feedback is 99.5+ and everything has arrived so far. A bit slow sometimes, so I would check RS first and do a price comparison anyway. Best way to treat ebay is consider what you've saved already and decide if you can afford to lose what you are about to pay while looking at the big picture. By now I'm $1000's ahead.

Another thing to watch with RS, I remember one of the test room guys ordered 100 x M5 nuts. Only they were sold in bags of 100 so we ended up with 10,000. When you ordered things with paper, it wasn't so obvious because they list the unit price and mention they are only sold in units of 100 elsewhere. I've still got a couple of bags. There's only so many M5 nuts you need in one life.

Originally Posted by sossity

You've more faith than me for finding a fault without a diagram then. It's not easy at all to trace out circuits when the thing is not designed to be taken apart easily and there are so many wires dissapearing into the chassis.

As an electronics engineer, you have a much better chance than me.

My mug's method is to assume a single or small number of component failures. Assume the mostly likely suspects for failure are the higher power ones. Test the components. Replace what doesn't seem to be testing to specification. Being able to test in circuit helps - I love my ESR tester. Maybe a faster desolder gun would make life easier....

Historically I've found RS expensive but comprehensive. I do love that I can buy a couple of dollars worth of parts and they will ship for free in 1-2 days - easier than physically going to the local electronics parts place to find they are out of stock.

Looks like the Peak meter prices are a bit friendlier now at RS. IIRC they were asking $480 when I got mine direct. Now $142.
Same day despatch on LCR Meters parts | RS Components

I don't suppose anyone cares quite as much as I do, but I managed to fix the fault. I did a lot of welding yesterday at 230 amps to give it a good run. It's running as well as ever. I couldn't source the corrrect parts and when I entered the part number in Google I didn't get even one hit. Bing however gave me several pages including a chinese one with an english datasheet, so I was able to work out an equivalent part from RS. I sometimes wonder if google isn't getting so far up its own bottom with all its "value add" that its losing some of its core ability, because there's been a few times I have come up with nothing on google now and i've found stuff with Bing.

If you have an electronic welder, I would recommend opening it up and blowing out all the steel dust. Mine was quite bad after two years and I don't do any grinding closer than 10m and never when the welder is on, although I usually weld just after that. I see some of the manufacturers will consider this neglect and that if a fault is caused by this, then the warranty is null. It would make sense to fit filter material over the fan in that case, if they know its a problem. If you want to take the cover off your welder, best to wait for 5 mins after switching off because the first thing you will find is 315V dc. Although there are bleed resistors to drain this voltage, its not unknown for them to go open circuit, so don't assume its safe.

Also, if you have HF tig or plasma cutter, I would recommend checking the spark gap. I saw Everlast who sell a similar welder to mine, say they use 0.75mm gap. Mine was twice that, which puts a lot of unnecessary stress onto the electronics. I could pull a spark a couple of inches before. After adjusting, it is much tamer but has an advantage for me too. Previously, if trying to weld at 5~10 amps, I noticed if everything wasn't dead clean, it was easy to get what looked like a plasma cloud (like in those plasma ball things) which would wander all over the place ruining any polished surface, until you increase the amps and it snaps into a small arc. That effect is much reduced with the smaller spark gap.

Anyway, i'm a happy welder (hack) again

Originally Posted by sossity

I don't suppose anyone cares quite as much as I do, but I managed to fix the fault.....

Good job! I think it's worth posting the method you used to debug and the actual component you needed to replace. One day, someone, somewhere will benefit from your experience, and at least one less repairable welder will end up in the tip.

PS Google hasn't failed me yet. But next time I will give Bing a go.

[QUOTE=blouis79;1718462]Good job! I think it's worth posting the method you used to debug and the actual component you needed to replace. One day, someone, somewhere will benefit from your experience, and at least one less repairable welder will end up in the tip.

I'll try to put a guide to one of these welders together because I wish I had found something to just show me what all the compnents were and how one of these things works. Took me ages and I made a few mistakes dismantling the thing. Might take me a while though.

I was a bit suprised at the cost of components now, it mounts up quickly. I checked out your links to those testers you provided, but in the end I decided its still best to change any components which I was suspicious of because of their relation to a known faulty component. In the old days, if I took out a standard component to test it, I would always replace it anyway. The problem is something can appear OK when you test it on the bench, but it can break down under load in the machine. That's the problem with those testers, they don't put much current through the device. I never checked to see if they can test at high voltage, which would make them more useful. They would save a lot of money if you were confident they would give you 100% results, but I can't decide how much I would rely on them.