Electronic welding machine in short circuit

[coppercone2]

Ok then you certainly found the problem.

It sounds like something is telling the chip to turn off. A fault signal perhaps, or something that got busted.

When you are sure about what logic inputs the chip should have to function, you can determine which one is different, and what might be the cause of that. I think this might mean your power section is fine, because its not even being turned on.

[coppercone2]

TTL is 5V. If you have 9V there is a problem, I think. A ttl signal is 0V low 5V high (ideally). 9.7 is very strange.



It's not connected to Vz on your schematic. So that means the TTL input is getting double the usual.

Usually the absolute maximum voltage input you see on a TTL input is like 7V. 

[coppercone2]

But the chip has a absolute maximum of Vz for all inputs, even if they call it TTL. Its enhanced TTL I would say.



Can you measure the Vz ?

I don't like it anyway, because the data sheet basically says, your not allowed more then Vz on any input, meaning... that 9.7V is 0.7V above the worst case 9.

I would replace that chip for sure. It stands out. Its mad drifted


This is like Treez specifications. How much can you go above absolute maximum lol?!

[DisasterNow]

I also think that the chip inhibits the output from QE pin 14.
The datasheet for EN pin 10 actually says that it accepts the TTL signal, so it is assumed that it should enable the output with 5v and disable it with the low signal (this evening the yellow LED did not light up, it should lower the signal on this pin to zero I imagine. But as mentioned I have 9.7dc (seen the signal on the oscilloscope).
Looking at the track of pin 10 (it is the signal that then passes over C4), it arrives at pin 5 of the LM324 (which is part of the operational on pins 5, 5 and 7).
Vz is as per data, perfect, stable 8.4dc.
I'll check further tomorrow. Thanks for everything.
(I attach the TDA1060 datasheet)

[coppercone2]

Ok, you might wanna check with a meter (scope is not great for DCV).

But the fact is, the datasheet says absolute maximum on a input pin is Vz. But your getting 9.7 which is way above 8.4V, which you measured. 

and it should be 5V to be in TTL spec!!!!!!

[coppercone2]

something is not right there unless the designers dont care at all

absolute maximum is absolute maximum, unless your boss is going to fire you before you get a respectable job

[DisasterNow]

Yes, the voltages are all checked with the multimeter) with the oscilloscope I just wanted to see the type of signal carefully).
Thanks, I'll do more checks tomorrow. I would have liked to try to figure out how to temporarily enable the output (with some tricks), just to understand if it's not an internal chip problem).

[coppercone2]

ehh check all the circuits leading up to that 9.7V however you can before you do that.


Theoretically if you get 9.7V on a absolute maximum 8.4 V pin, its toast! I would consider that chip damaged at this point.

[DisasterNow]

Yes, I had seen that the inputs should not go beyond Vz (therefore beyond 8.4dc). At this point, but I'll check better tomorrow), the EN input is over 1.3v, but also at pin 5 I measured 12.6 dc! The others are inferior.

[DisasterNow]

I'll check everything tomorrow, thank you very much Coppercone. Good night

[DisasterNow]

I checked the voltages on the TDA1060 pins again, I confirm those I reported yesterday.
For example, at pin 10 (EN) there are 9.7v (which does not comply with the datasheet). Below in the sketch I drew the part of the circuit that affects that pin, and what could influence it.
Unfortunately I do not understand what could inhibit the output of the signal on pin 14 QE of the TDA1060 (it should be a rectangular wave).
I think I will have to abandon this repair, my skills are finished 

[coppercone2]

Well you found the problem, it looks like its a damaged chip. For all you know the chip broke.

If its going into a INPUT of something else, chances are its busted!

The non inverting input of op amp does not consume current in that configuration, its just measuring!


replace all the silicon stuff on the control board. It should only cost a few dollars. those are cheap parts. That op amp is the most cheap op amp in the world, and I doubt that controller is too expensive.

All that shit should run you less then 15 euro.

It could be as simple as some kinda problem with the power supply causing the chips to break.


Anyway, its alot less of a problem then trying to replace all those power components, and cheaper.



I would


1) remove all the silicon cheap parts from the PCB (including the transistor and diodes, if you can figure out which diodes it is, or at least measure them)
2) test all the passives with silicon parts removed, including the capacitors
3) install new parts
4) hope



See I suspect, the op amp is getting something from a 310V rail. It probobly mal functioned. maybe something arced over like dust or whatever.

310V near a LM324 is like if a house near a firing range got hit with artillery, its not very safe. or like a car ending up in the front of a store... (blues brothers)

and check that temperature sensor.

I would not put anything past a high voltage circuit getting into control silicon



If high voltage gets into a low voltage circuit, what it does is basically totally fucking random. it can break the whole board, it can  find a weak part, it can jump around and degrade a bunch of things. it can do anything from explosions to almost undetectable damage.Its common, look up requirements for protecting keypads and touchscreens and such on industrial equipment, they want like double or more isolation! The keypad needs to have like a coating resistant to HV (1 isolation), and then have it electronics connected through another isolator (i.e. opto, transformer, etc). I think I Have seen one where they wanted double isolation and a protective film (triple isolation) for some medical weirdo thing



What happen to that 310V rail if the power grid spike to 500V while your welding with a fast pulse or something? power section might take it but it can fry the digital side, especially if its cheap.

[DisasterNow]

As you said, the 310 volts pass through two resistors (they are soldered vertically), R16 and R17, which lower the voltage to about 5.6 volts, which is certainly used to communicate to the control board the presence of an adequate voltage on the power part.
I checked the thermal sensor (with hot air), when the temperature goes above 80°, it opens the circuit and the yellow LED lights up, so it works perfectly.

The problem is that I do not have a reliable schematic of the small hybrid board (there are some sketches found online, which I had attached previously, but they do not match much ...), so apart from the two chips and the mosfet, the rest would not be easy to identify. It must also be said that for example the TDA1060 and I believe also the LT324 are obsolete and perhaps can not even be found anymore. This is a big problem.
Would it be possible to unsolder at least the 3 largest chips, and check their efficiency on the bench in some way? This would make sense, for example if I could verify that the TDA1060 comes out with some signal from pin 14, then it could be efficient.
Thanks

[DisasterNow]

I attach the only sketches I found online of this small control board (but as you can see, on mine many components are not mounted, all the resistors for example etc.)

[coppercone2]

 

LM324 is standard part, still available. The other part I see you can buy from website for 5 euro.


If you are hell bent on testing a 2$ lm324 chip, you can look at the data sheet for performance tests. You would need to desolder it, put it on a adapter to soic, and do all the performance circuits on a breadboard

the other chip IDK, its complicated.


now for the ceramic card, I never worked on one, I won't give you advise. I just know if its silver on ceramic, you want silver bearing solder so you don't dissolve the traces. this card is a module purchased from someone to do a complicated function drop in. I have seen some people like 'mikeselectricstuff' repair these kind of cards. Theoretically if something is soldered to a ceramic, its the same thing as a PCB, other then that silver related nuance.

in practice its brittle. If something did fry on there or it over heated before, it could be damaged in such a way that makes the ceramic extra fragile. Also ceramic is a good heat conductor to PCB, so it might need alot more wattage to remove parts. You would need to do research on this, I never tried.


I have made my own ceramic silver pCB traces, the only thing I noticed is that while somewhat strong, its not that strong, never ever yank at anything on there.


I think you need a second opinion on that chip. I wonder if there is a special extended voltage range series or something,  but I think its broken. those cards are not invulnerable and a welder is exactly the type of equipment I would expect it to break in lol because their evil


If I just decided to randomly try it, I would get very thin solder braid and see if I can suck any solder of those chips. you need to be precise with the iron not to scratch or knock something off

[coppercone2]

if you want to persue the theory about the ceramic card, the hard way might be to remove the whole card carefully from the PCB, wire it into a bread board, and then make a mock up circuit that tests all its functions. bit of engineering effort . you just need to map out all the signals that it expects and test it for realistic performance on dummy loads. Not sure how easy they are to remove though.


I am out of ideas on this one for now


Oddly enough, the IRFD110 was one of the part that broke on my miller. They had like 10 of them in a bunch, and the PCB literarly made a hole from sparks near those chips. I had to replace that card because it was totally messed up with way too much arc damage. it looked like it got shot by a ray gun

[DisasterNow]

Of the 15 pins on the logic board, those that reach at least one of the two main chips, and which could therefore inhibit the signal, have voltages in the rule. This means that the output signal from QE does not come out due to a fault somewhere on this hybrid board, and not on the main board. As you also said, desoldering and soldering on this ceramic board is not easy, I would also add that there is a layer of protective lacquer that would increase the problems. I could also do a desoldering test (I would use the hot air desoldering station); but at this point I can say that I have done everything possible, the fault has been (more or less) identified, but the repair is too complicated.
It was still a nice 'journey', I studied a circuit that I had never had the opportunity to repair. A heartfelt thanks to Coppercone for all your advice received, you were very kind. Thanks 

[coppercone2]

It's not that bad to repair it. You need to suck out the solder to remove the chip and hot air should get rid of the op amps. You can pack the board with putty or use tape to protect the other parts.

The conformal coating is light weight, it should come off no problem, especially after all this time.


Its easy to try, I would not recommend giving up at this point. You might be surprised how well it goes.

I would only give up if you plan on giving it to someone else to repair

my welder just got destroyed by a beetle and I need to figure that mess out



The military has giant ceramic boards (2x2 feet) that they used to repair for jet airplanes maintence systems and stuff

[coppercone2]

for removing parts SMT there is also the "pakistan style" trick where you bend copper wire around the chip so it can screw into the barrel of a soldering iron. After enough time your thick copper wire gets hot enough to heat up all the contacts at once. It might actually be useful if your scared that hot air is going to be too hard to control (which it often is not, usually the problem people write about hot air comes when there is tiny parts around, i.e. 0402 or 0201 (that you need a magnifier to see). All your parts are big



Practice desolderin a chip on a normal PCB, then you know you have the right settings, then do the same thing for ceramic pcb, possibly you might need to increase the heat a little bit or do some very gentle pulling (no yanking) when its hot if there is some glue

[coppercone2]

Oh yeah, but there is another method you can use to increase your reliability to 100% for the repair


you cut all the parts off at the base of the part with flush cutters (fine tip), remove the part, then heat each lead individually and pull it with tweezers. it wrecks the old parts but its what I basically always do because its just better

I would only bother desoldering a part properly if its worth at least $10 and I have a feeling that its ok (i.e. to isolate part of troublsome circuit)

this is more then good enough
https://www.amazon.com/Xcelite-170M-General-Shearcutter-Diagonal/dp/B0002BBZIS


this is total overkill
https://tronex.descoindustries.com/Tronex-Catalog/Cutters/Tip-Cutters/5049/



then you solder on the new parts with a soldering iron, you can even glue it down with a tiny dot of hot glue to make it easy. Wick out the holes with fine solder braid, or a solder extractor, or if those two fail, you can use a air blower to blow out the solder to the other side (messy and requires very detailed inspection to make sure you did not short something with a errant solder blob)



You only have a SOIC part. The tear drops might start when you have a heat sinked no-leads part, then you can't cut them off. A SOIC is just a squashed dip, you can cut them off! it has legs




And Ceramic is not that weak, i mean its weak compared to fiberglass but I have been using this ceramic substrate (ceramic utility knife blade) for my experiments

it survived multiple drops on the concrete floor and around 5 heat cycles, one to 1300C and 4 to 1000C, before it broke when I wanted to see if it still cuts into a bottle cap after all that  (I forgot that I had initially sanded the edge off with diamonds to make it safe, otherwise it would probobly be fine). I even dropped it hot when it was like 600C the first time I took it out of the oven, because it slid off my oven plate when I put it into my cooling insulated brick

https://www.sliceproducts.com/products/ceramic-utility-blades

they work for knives, I use them. I am telling you, your repair is easier then you think . Those ceramic substrates are not shitty clay pottery, its a hard core engineering material, they do all sorts of things a typical ceramic person does not do, like x-ray? porosity tests, density tests, cracking inspection, etc before it gets approved for circuit card use.

[DisasterNow]

I admit I thought about giving up on the repair, I spent 1 month on it. We understood together where the problem lies. However, your further encouragement convinced me to continue. I have enough manual skills on how to unsolder or solder these components, perhaps I could have some problems in soldering the LM324, the TDA1060 will not be a problem.
I will proceed like this: first I have to unsolder the entire small hybrid board from the main board, there are 12 solder pins (3 are not used); so I can then have more manual skills, and I will start by unsoldering the TDA1060 and the LM324 calmly and patiently... If the work is done well, with all the pads still usable, then I will try to order these 2 chips from the Chinese.
That seems sensible to me, right?

Be careful though Coppercone: let's not forget that in any case there remains that occasional short circuit that I have not yet unfortunately identified (unless it is a consequence of these 2 chips, but I do not know). In the meantime, thanks again, I will proceed calmly, but I will update the situation here 

[coppercone2]

yeah it should work so long you don't drop it or get forceful with the heater

[coppercone2]

just don't forget you might want to use the silver solder. I don't know exactly what that board is, but it could be silver glass, meaning it will dissolve in to normal solder, and it can leech

https://www.ebay.com/itm/116004675563

a few inches worth should do, if you can find it in smaller quantity.

they are very cheap with silver, so you don't have much to work with. I never noticed anything on my DIY ceramic silver PCB with the leaching, but I trowel it on, I know someone would get shot if they tried to put that thick a layer in a normal factory lol



I do have some piece of shit russian inductor (well actually its kind of nice) with metalized silver ceramic pipes to feed wires into a chassis. I noticed when I start feeding regular solder into that, the "silver" on the pipe walls started to disapear. But I assume a soviet worker ripped them off and carried out most of the silver paste on the bottom of his hat, leaving the absolute bare minimum on the walls of the ceramic feedthrough
. To get wires into holes, they would use a ceramic pipe, metalize the interior and exterior, solder it into a hole in the chassis, fill it with silver bearing solder, then use it like a interconnect. This was for a hermetic sealed part with a big gasket on the bottom.  Then it could be filled with dielectric gas, or be water proof, or whatever they needed it for

[DisasterNow]

I preferred not to unsolder the hybrid board, in addition to the normal soldering on the back side, all the pins had the soldering cone also on the board side, and being in the part that faces the voltage regulator heatsink, it was impossible to work. I then unsoldered both the TDA1060 (which came off immediately), and also the LM324 (more difficult, it was completely drowned in the protective lacquer that prevented the heat). No damage to the pads, let's say that I then cleaned lightly with the braid but not much, and scraped the pads from the lacquer. I protected the other areas, but I see that in the lower part of the small chip there was a lot of lacquer that has taken on a brown color, but I don't think it's a problem.
I only have lead solder wire at home, I would try with that maybe.

Looking at the photo I have a doubt (the board is in the lab and I'll see it tomorrow): has the base of the small electrolytic SMD capacitor melted?
But I don't think so, I was at 370°, but it was all covered by a protective aluminum layer.

[coppercone2]

those bases suck their  easy to melt i doubt it damaged the pad you can probe it with a LCR meter near it if you wanna see if its still connected with low ESR but I think its OK