After a lot of trial and error with my own version using over 370mF of capacitance (similar setup to Lynneandpeter4947) I can't recommend building this setup. The SCR module simply doesn't provide enough control to the pulse duration and strength, compared to FET control systems. Maybe if you happen to have a large capacitor bank on hand, otherwise I've spent ~$70 on this project and for a bit more I could have gotten a pre built more reliable super capacitor welder. Thanks "Wojciech "adalbert" J" for the guide however, it's well made!

What are you using for wire to / from the capacitor + the probe tips? The probes + wire can have a HUGE effect on the whole device basically making or breaking it.. Plus all the interconnects between your capacitors also need to be low inductance and resistance. Look at this video for example ;) https://www.youtube.com/watch?v=-79YL4pStlU&t=366s

Also you can easily use a MOSFET with your capacitor bank if you think that will improve it - or a much larger SCR or other switching device like the simple relay types. Infact this combination of normal capacitors + large 200A relay (from car starter) is probably the best simplest solution of all.. 

Using a 370000uF capacitor bank (37 cheap, unbranded Chinese 35V 10000uF aluminium electrolytic capacitors 18*35mm, low ESR) and the same 12v dual relay, DC to DC converter, timing capacitor and RC pulse control described in Adalbert's original project I am successfully spot welding .3mm pure nickel strips (with the capacitor bank charged to 30.7v), .2mm pure nickel strips (with the capacitor bank charged to 21v), and welding .15mm pure nickel strips with the capacitor bank charged to 15.8v; but the up-to 7kA pulse produced by the larger capacitor bank did cause the T90RIA120 thyristor to fail (probably because the max ITSM of the 90A thyristor was exceeded by ~ ×2, even though the pulse would have been ~1ms compared to the 10ms limit in the 90A thyristor specifications). The damage to the 90A thyristor caused it to be permanently switched on.

I replaced the damaged T90RIA120 thyristor with a stud thyristor (Semikron SKT300 series with an ITSM of ~11kA). It is bulky and therefore external to the 160 x 110 x 90mm plastic case into which I had squeezed the 37 x 10000uF capacitors, the original T90RIA120 thyristor and the other components, but it is effective. The stud thyristor is inserted inline with the negative probe cable, and the cathode cable is bolted to the same connector inside the case as the original T90RIA120 thyristor. The thyristor gate is triggered by the same 12v RC circuit described in the original project.  The stud thyristor has further lowered the voltage required to spot weld (see my stats for welding .15mm nickel strips using the original 90A thyristor and 370000uF capacitance in my earlier comment). I suspect the discharge time using this larger thyristor has been reduced to <1ms (probably .00055ms), assuming the amperage needed to spot weld specific nickel strip thicknesses using Adalbert's design is much the same regardless of the size of the capacitor bank and capacitor bank voltage. I don't have an oscilloscope to confirm my theory.

To protect the DC to DC step up converter from unintended discharge from the capacitor bank I inserted a 40v 3A diode between the capacitor bank and the converter, notwithstanding the converter  specifications include output reverse voltage protection, and the diode causes minor forward-direction voltage drop.

In conclusion, as Adalbert contemplated, I confirm it is feasible to use cheaper capacitors; and increasing capacitance by 250% enables spot welding of .3mm pure nickel strips at 30.7 volts, and thinner strips with significantly reduced voltage; however, increased capacitance has led to deviation from Adalbert's original imperatives of compactness and low cost; the heavy duty stud thyristor cost twice as much as the T90RIA120.

I'd love to know more about your setup. I'm also using a ~370mF capacitor bank and otherwise the same setup as the original guide. However I'm having problems with tip fusion when welding pure 0.2mm nickel.

Using 8AWG weld probes sanded down to 1mm conical tips, I am successfully welding 0.15mm pure nickel strips at 30V. I'm really interested how you are welding 0.2 at the same voltage! Are you sure your nickel strips are pure and not plated steel?

Final result using a 370000uF capacitor bank (still with max 31v input):

.1mm nickel strips 21v 5.6kA

.15 nickel strips 24.5v 6.3kA.

Excellent welds; no sparks or tip fusion etc. (3mm diameter copper probes have 1mm conical points and I apply moderate pressure).

There is no indication the 90A thyristor can't handle the high pulse current, although its use has been limited so far. I will post a final comment once I have tried welding .2mm nickel strips. (Extrapolation suggests .2mm will require 29v with a pulse of 7.4kA.) 

I'm stll waiting for another 20 x 10000uF capacitors (and have 6 redundant 10000uF capacitors soldered in place awaiting to be paired when the final order arrives) but have connected the circuit in the interim to test components. With only 14 x 10000uF capacitors in the circuit (close to the capacity of the 142500uF capacitor bank in the original project) I am able to weld strong joints with .1mm pure nickel strips at 26V, and .15mm pure nickel strips at 28.5V (3kA). I have 14 square mm of cable to each probe ((doubled up AWG10 (3mm dia.) silicon copper cable)), so resistance is negligible. Probes are 3mm dia. copper wire with 1mm dia. tips.

I'm still concerned the 90A thyristor will not cope with >4.7kA pulses (even for 1mS) with sustained use. If 200000uF will weld .2mm strips using <31V I'll probably just use a 200000uF bank, which (only!) generates 4.1kA.

My 12v power supply is an 18v Makita battery pack with a buck converter I had. I tried an L7912 to regulate power but didn't have the correct capacitors to control the input and output, but will try again later.

I had assumed the key criteria for welding increased thickness of nickel strips was current. The original project estimated ~3kA using 14.2mF, and he said he could easily weld .1mm nickel strips (I presumed he meant pure nickel). 20mF should produce ~4.1kA. I was concerned about whether the 90A SRC thyristor could handle repeated surges of 4kA, even for 1ms with a couple of seconds delay between each weld.

I take your point about trying the penetration capability using any 12v power supply before introducing the 18v and L7912. I'll let you know how it goes in a few weeks (I'm on vacation).

Thyristor is 50RIA120 bought here https://www.amazon.nl/-/en/Thyristor-50RIA120-Rectifier-Convectors-Controlled/dp/B08CDF573F. So far so good ;).

And the cap bank is 100mF, using this https://www.amazon.de/-/en/dp/B09JB94CDX?psc=1&ref=ppx_yo2ov_dt_b_product_details to be clear.

Good luck with your build and please share here your outcome! 

I am using an 18v power source (Makita battery) and therefore putting an L7912 regulator on the negative line to drop voltage to a continuous 12v. The 10 uF cap is inserted after the voltage regulator to filter out any noise or high-frequency (ac) signals that may be on the DC line. This may be unnecessary because the circuit probably isn't that sensitive. I also think that having an L7912 should do away with the need for a fuse for the 12v supply, but I'm not sure about that.

Hi jandy123,

Has your 50mF set up performed as expected? Any problems with the amps (50A) of the thyristor, and has it held up to sustained use? 

I intend using a bank of 20mF/35v (20 x 1mF) without making any other changes to the posted project; I want to achieve at least .15mm welds on pure nickel strips.

In the end I used a 100mF cap. bank and followed the images for setting  the relay delays.  See my comments below. So far so good, I can solder 0.1mm nickel strips just fine, between 28 and 30V. Lower than 28V, I do not achieve proper bound. I would strongly recommend, allowing the possibility of larger than 30V (this imposes caps requirements and boost converter).  Unfortunately I cannot try >30V, and I also do not have 0.15mm strips.

20mF will not be enough  to weld 0.15mm strips. You will probably need at least 150mF @35V.

I am in the process of replicating your project.

The image of the bottom of the HW-279 relay board shows a 330uF/25v between Gnd and TRIG; but the circuit diagram shows a 100uF/35v. Did you make this imcrease to a 330uF/25v to tweak the timing delay?

I am adding a single LM7912 to regulate voltage from 18v to 12v, provided the 12v circuit is <1.5 amp. I plan to include a 10uF output capacitor. Are you able to confirm a single LM7912 will suffice?

Yeah, I faced this too. In the end I used two 330uF caps on both triggers.I Btw.,  I "limit" current with two series 12V halogen bulbs allowing for 2amps in series with a cheap non-isolated, non-CC boost converter. I power everything with a beefy xbox360 (at 12V capable of 10amps).  

What do you mean with 10uF output cap???

For the boost convertor (12V to about 30V) the 1.5amp may not be nearly enough, unless you use a current limiting boost converter and a very small cap. charge current. That will result I think in large delays between successive welds.

I would first get the circuit running from 12V (without the LM7912), make sure the cap. bank is large enough for 0.15mm strips and then worry about powering the circuit from 18V.

Very nice project, quite smart actually, compared to the many fakes around "capacitor spot welder". 

I'm trying to build my own, using for now a 50mF capacitor bank. (No clue if this is large enough, but time will tell.) For now, I have the cap bank powered by a 18V/1A trafo, with which I can blow out strips, scissors, etc., but no proper binding/welding on batteries; using even 0.1mm strips.  However, as far as I understand, either i) energy is not enough (easy to solve -- add more caps???), or time of contact/discharge/melting is not appropriate. So, I thought to address  the latter,  I should give this  nice project a shot. After looking at the info, I see that the schematics changed compared to the detailed info regarding the relay mod.  Could you please let me know which I should follow? Some more/up-to-date info from your own findings, about settings, energy needs, etc. would be greatly appreciated! 

Anyhow, great, simple and smart solution!

Thanks in advance!  

Answering to my self (in case it helps anyone). I did it using the relay mod schematics, for the rest following the "main" schematics.  I used a cap bank of 100mF made of 10x10mF 35V cheap low-esr caps, a cheap non-CC boost module (used two halogen light bulbs in series to limit  the current somewhat) and a cheap 50RIA120 thyristor clone. It works quite well, in the sense that I'm able to spot weld 0.2mm strips at about 28V.

*** Correction 0.1mm strips ***

Definitely better than my 4V battery powered device (which run out of battery) and than most youtube "capacitor spot welder" fakes.

Your schematic says "Low ESR Capacitor bank (~1mΩ total)" but the Nichicon UBY 7500uF/35V caps are like 20mΩ  How does that work?

Capacitors are 18mΩ each and there are 19 of them, and because they are connected in parallel, I just assumed that total ESR equals 18mΩ divided by 19 = 0,947mΩ

Ah. So if I had 3x 47000uF caps with 18mΩ handy and used  them instead of the 19x 7500uF, I'd have about the same capacitance but 6x the ESR.  With the Thyristor's 2.5mΩ and 2 foot of 10 gauge being about 2mΩ, it would be about 10.5mΩ vs 5.5mΩ.  If the weld material isn't a significant resistance, then maybe the higher ESR capacitor bank might have half the peak current for the same charge.

I got a lipo-powered spot welder from ali last year. It was reasonably nice, produced not-so-bad welds, but I was a bit worried about having a *big* lipo in there. However, it turned out the control circuitry was badly designed, so the lipo was completely flat after it was on the shelf for a couple of months...

However, your design sparks an idea: Use your capacitor bank as a power source for the welder I have (so I can still use its pulse timing, connectors, etc.). Will have to properly check allowable voltages and currents, though...

Thanks for sharing this!

I'm in the same boat! Have you ever managed to recycle the thing, apart from replacing battery?

Great idea. 

I was considering doing this with a bank of supercaps from a discarded project each 3Farad but at 3V - just not sure if that would work though.

Thanks! Not sure if 3V wouldn't be too low for a thyristor, as it has some noticeable voltage drop. And charging low voltage supercaps would be slower.

It should also be kept in mind that in my design all of the energy from capacitors is dumped into the nickel strip instantly, at once. Oversized bank of supercapacitors could store so much energy, that it might melt the nickel strip completely. Some kind of pulse time control would be needed, which is not possible with a thyristor (a thyristor will stop conducting only after the capacitors are discharged completely, or the circuit breaks in another way). Maybe it would work if you used only 4 or 5 supercaps in parallel, so they won't be too powerful, but now the question is would they be able to put enough amps into the welding pulse.

Of course It certainly is possible to use supercapacitors in a spot welder, as I saw some existing designs, but they usually have precise pulse control with MOSFETs and microcontrollers, and these supercaps are treated in similar way as batteries. Regular electrolytic capacitors are easier to use without sophisticated pulse control system.

Yes all noted. I recently worked on reparing a large plasma torch and replaced some very high current MOSFETs which I now have some spares. Also regularly use a two transistor 'spike' generator which could work. Always fiddling around witn stuff here but have access to a huge bag of these supercaps so worth a try.

Thanks for the feedback.

I recently put together a k-weld. There wasn’t much else out there, other than some sketchy Amazon products. Your design is a better starting point, I think, for someone who needs a few spot welds now and then.

Thanks for the comment! This design doesn't have any batteries, so even if you use it once in a year, there shouldn't be anything that will degrade over time (i hope the caps will last for years and hopefully decades), and there are no self-discharge issues to worry about.