kWeld - "Next level" DIY battery spot welder

[tatus1969]

That looks really good, almost resembles a bullet  Is that brass with a copper electrode ?

Yes, these are 5mm copper rods, crimped to XT150 male "bullets". The electrode holder is CNC machined to fit AWG8 on the other end. Just two crimps to finish everything, one minute of work :-)

In the mean time my project has come to a standstill, I am awaiting the final results of which supercaps I need to succesfully weld the tabs, is there already a final outcome?
And my 45V/70A PSU has a major issues, which I am still trying to solve, it is still open for repair.

After having finished the second kWeld batch, I will go forward with a 2S2P 3S2P BCAP0310-P270-T10 configuration. That has already shown to deliver >1000A in my lab. Your 70A PSU will be perfect for them. The kWeld rev.3 will have an input voltage range from 4V to 30V (SEPIC converter) and will be able to run directly from these capacitors. I have also substantially improved the power switch, I have just made an explicit overload test of the power switch. It is rated at 2000A, and I tested it with 1000 successful pulses at 2200A, followed with 800 pulses at an extreme overload of 2800A (!) before it broke down. Also the inductive kickback handling should be better now, I will test that today.

But I did do some new stuff, I build a jig that will help me do some other welding operation: weld two tabs together, for which I need a vertical welding jig.
So this is how that looks, I found most parts of that jig in the dumpster a few years ago and now have a good use for it.
The electrodes I got where 7mm copper bars for the staggering price of €1.- a piece at my local ironscrap yard 
On top there is a 6mm thread to mount the weldingwire and the bottom is a 5mm tapped hole for your 5mm electrodes I bought from you with the kit.

[beduino]

I have also substantially improved the power switch, I have just tested it with 1000 pulses at 2200A, followed with 800 pulses at 2800A (!) before it broke down. Also the inductive kickback handling should be better now, I will test that today.

Old school transformer 230VAC -> 2.5V which means at 10A at primary pulse 1000A on secondary few turns of copper pipe with simple SCR based switch on transformer primary to turnoff at zero crossing much more reliable, however copper costs and transformer weightw, anyway some people buy gold for savings, the more copper we have the richer we are and forget about stuff like this-this is good for selling - lower weight etc, but less copper means lower weight means less reliable lower quality.

High quality products might weight while made of pure metals not plastic crap

[tatus1969]

Old school transformer 230VAC -> 2.5V which means at 10A at primary pulse 1000A on secondary few turns of copper pipe with simple SCR based switch on transformer primary to turnoff at zero crossing much more reliable, however copper costs and transformer weightw, anyway some people buy gold for savings, the more copper we have the richer we are and forget about stuff like this-this is good for selling - lower weight etc, but less copper means lower weight means less reliable lower quality.

That will only allow to chop the pulse duration into multiples of 10(8.3)ms, which was too much of a drawback for me when starting this. The kWeld has a resolution of 10us and a minimum pulse width of 40us, which gives it the bandwidth to weld 0.3mm nickel strips equally well as thin copper strands.

[Kjelt]

After having finished the second kWeld batch, I will go forward with a 2S2P 3S2P BCAP0310-P270-T10 configuration.

Thanks. Where do you get the supercap protection pcb's ? Or did you design them your self ?

I have also substantially improved the power switch, I have just tested it with 1000 pulses at 2200A, followed with 800 pulses at 2800A (!) before it broke down.

Please remind me, are you talking about the powermosfets on the pcb or something else? What made it break down, overheating? If I only use it once in awhile for ten welds this is not an issue is it?

[tatus1969]

Thanks. Where do you get the supercap protection pcb's ? Or did you design them your self ?

I am using them "free air" for now, manually balancing the voltages. I will design a protection PCB soon.

Please remind me, are you talking about the powermosfets on the pcb or something else? What made it break down, overheating? If I only use it once in awhile for ten welds this is not an issue is it?

Yes, it is the MOSFETs blowing up in an extreme overload stress test. Keep in mind that 2800A is 800A above the overcurrent protection (I have disabled it for the test), and power dissipation wise it is twice the maximum (resistive losses are current squared). For the test, I was running the unit directly shorted and had two batteries connected in parallel. This is far from reality, but I wanted to find the limits. Nothing to worry :-)

[beduino]

For the test, I was running the unit directly shorted and had two batteries connected in parallel. This is far from reality, but I wanted to find the limits.

I can't see any supercaps there - anyway I do not think they are rated for 1000A or so to use them for discharge at such extreme conditions, but while I haven't seen their specs or maybe plenty of them they are used in parallel . lets say 10 each rated at 100A discharge ?

I've already seen another project some time ago where someone used bulky IGBT to switch currents at range of 1000A at spot welder at he was not able to manage induction spikes which kiled many not cheap IGBTs, so what can I say, good luck 

[tautech]

I've already seen another project some time ago where someone used bulky IGBT to switch currents at range of 1000A at spot welder at he was not able to manage induction spikes which kiled many not cheap IGBTs, so what can I say, good luck 

You need read this thread from the start, all this has been taken care of to reach a commercial product.

It's not at all unreasonable that tatus1969 would want to do destructive tests to push his product to a new level or just make it more robust.

[Kjelt]

I will design a protection PCB soon.

Great!  Also hopefully sell them ?

Yes, it is the MOSFETs blowing up in an extreme overload stress test.
For the test, I was running the unit directly shorted and had two batteries connected in parallel. This is far from reality, but I wanted to find the limits. Nothing to worry :-)

I think it would be good to explicitly mention this in the start of such posts before potential customers might think there is a liftetime limit or other limitation.
Great to hear that the design is so robust (made in Germany  )

[tatus1969]

I can't see any supercaps there - anyway I do not think they are rated for 1000A or so to use them for discharge at such extreme conditions, but while I haven't seen their specs or maybe plenty of them they are used in parallel . lets say 10 each rated at 100A discharge ?

Maxwell specifies a continuous discharge current of 250A and a short circuit current of 1200A, with a side note "Current possible with short circuit from rated voltage. Do not use as an operating current." They do not specify a maximum pulse current, which leads me to the assumption that these parts are basically thermally limited. I will do a long term test to validate this, but I am very optimistic. They do not show any measurable effect in ESR or capacitance after a few hundred 900A pulses.

Great!  Also hopefully sell them ?

That's the plan!

I think it would be good to explicitly mention this in the start of such posts before potential customers might think there is a liftetime limit or other limitation.
Great to hear that the design is so robust (made in Germany  )

Thanks :-) I noticed this after your post, and have just edited my post above. I am doing destrutive tests because I am preparing for a batch of 100 and do not want to sit here all day repairing units. I didn't mention that I did this destructive test at 4 pulses per second (!)

The last test yesterday was high inductive load at 2200A and it also survived this 1000 times without damage, although I had to reduce the firing rate to once every two seconds to limit heating of the TVS diode. Here is a picture... (The small ugly board left to the kWeld is the new internal 3-40V power supply that was receiving all that inductive spiking, and it keeps all its voltages well within their limits - should be rock stable now.)

[beduino]

I can't see any supercaps there - anyway I do not think they are rated for 1000A or so to use them for discharge at such extreme conditions, but while I haven't seen their specs or maybe plenty of them they are used in parallel . lets say 10 each rated at 100A discharge ?

Maxwell specifies a continuous discharge current of 250A and a short circuit current of 1200A, with a side note "Current possible with short circuit from rated voltage. Do not use as an operating current." They do not specify a maximum pulse current, which leads me to the assumption that these parts are basically thermally limited. I will do a long term test to validate this, but I am very optimistic. They do not show any measurable effect in ESR or capacitance after a few hundred 900A pulses.

We are talking about this Maxwell HC series ultracapacitors? http://www.mouser.com/ds/2/257/Maxwell_HCSeries_DS_1013793-9-1179788.pdfhttp://www.mouser.com/ds/2/257/Maxwell_HCSeries_DS_1013793-9-1179788.pdf

On page 4 in  PDF linked above we have BCAP0150 with short circuit 190A BUT: this is in Safety section, and probably you are reading its datasheet in wrong way, since this rating in this section as short circuit only means that in the case of some hazards one can expect such huge discharge current and for example part of PCB can disapear, BUT tread again side note second part: "Do not use as an operating current." which means... those caps are not designed to make short circuits at regular basis at such huge currents, while probably overheats them and when you look carefully at thermal cycles at higher temperatures you have 1000 pulses or so in comparision to comfort temperature usage at 25*C with 10 years life span.

A few hundred 1000A pules is nothing for spot welder ... at 50Hz AC with transformer based spot welder is is less than ... 20 s (seconds) of welding with adjustable pulse energy based on time offset between zero crossing and choosen delay 



Could you give us link to datasheet of these upltracaps you are testing? I can't find any Maxwell rated at continous 250A discharge rate....

Bedu.

[Southerner]

What about using some boostcaps?  Something like these:
http://www.goldmine-elec-products.com/prodinfo.asp?number=G17930
http://www.goldmine-elec-products.com/prodinfo.asp?number=G22499
They look expensive but might work if enough seriesed or one with voltage rating for the lithium battery used.  Another possibility would be the multi farad automotive audio capacitors that used to be so popular for teeth jarring bass automotive audio systems.

[ovnr]

A few hundred 1000A pules is nothing for spot welder ... at 50Hz AC with transformer based spot welder is is less than ... 20 s (seconds) of welding with adjustable pulse energy based on time offset between zero crossing and choosen delay 

To be honest: This is a hobbyist-level product. I don't think tatus1969 plans to sell these to battery pack manufacturers. And if you like transformer-based spot welders so much, go buy or build one?

I want something small, reasonably cheap and controllable. Transformer-based options only tick one of those boxes.

[ovnr]

What about using some boostcaps?  Something like these:
http://www.goldmine-elec-products.com/prodinfo.asp?number=G17930
http://www.goldmine-elec-products.com/prodinfo.asp?number=G22499
They look expensive but might work if enough seriesed or one with voltage rating for the lithium battery used.  Another possibility would be the multi farad automotive audio capacitors that used to be so popular for teeth jarring bass automotive audio systems.

Well, the original design calls for 6x BCAP0310-P270-T10 at $12.17/pop @ Digikey, or $73 total. Since the design is for a 3S2P pack, you'd need at least 3 of the (used!) $45 caps, for a total of $135. Almost twice as much.

[Kjelt]

Could you give us link to datasheet of these upltracaps you are testing? I can't find any Maxwell rated at continous 250A discharge rate....

He is talking about the bcap 0310 which is cheap and does not have to survive years
http://www.mouser.com/ds/2/257/Maxwell_BCSeries_DS_1017105-4-1179684.pdf

If you want more like 2500A buy the 3400:
http://www.mouser.com/ds/2/257/Maxwell_Technologies_k2_2_85v_ds_3000619en_1-1179621.pdf

Maybe, you do not remember what happened about 75 years ago - hopefully, not everything made by Germany was good:

i remember well since my family had some real tragedies in that period but as you said that was 75 years ago, over and forgotten and that has nothing to do with the life we are living today and the generations that are living today. I find it rvery offensive that you bring something like this up on an electronics forum.

[Simon]

Great to hear that the design is so robust (made in Germany  )

Maybe, you do not remember what happened about 75 years ago - hopefully, not everything made by Germany was good:

Polish 303 Squadron - Battle of Britain


"Squadron 303 (Dywizjon 303) RAF No. 303 "Ko?ciuszko" Polish Fighter Squadron (Polish: Warszawski Dywizjon im. Tadeusza Ko?ciuszki) was the highest-scoring allied unit in the Battle of Britain.[1] It was formed in Britain as part of an agreement between the Polish Government in Exile and the United Kingdom. ..."

There is no relevance in that to this discussion, please stay on topic.

[anishkgt]

A few hundred 1000A pules is nothing for spot welder ... at 50Hz AC with transformer based spot welder is is less than ... 20 s (seconds) of welding with adjustable pulse energy based on time offset between zero crossing and choosen delay 

To be honest: This is a hobbyist-level product. I don't think tatus1969 plans to sell these to battery pack manufacturers. And if you like transformer-based spot welders so much, go buy or build one?

I want something small, reasonably cheap and controllable. Transformer-based options only tick one of those boxes.

Or better buy this pcb only https://rover.ebay.com/rover/0/0/0?mpre=https%3A%2F%2Fwww.ebay.com%2Fulk%2Fitm%2F282466833886


Www.Georgehobby.wordpress.com

Equipments: DSO104z, Hakko FX888D, Brymen BM869s

[tatus1969]

I am happy to announce that the next batch of kWeld is in production. I accept pre-orders through my shop: https://www.kicksurfer.de/index.php/product-category/kspot-welder-kit/  All pre-order products are on sale with a 10% discount until Nov 30!

New major features in revision 3:
• replaceable electrode system with minimum assembly effort
• extended input voltage range from 4V to 30V allowing the use of ultracapacitors without an auxiliary power supply
• improved protection against inductive kickback of both power switch and voltage regulator, tested to survive currents in excess of 2500A
• enhanced user interface featuring a rotary encoder with push button

Cheers
Frank

[Leo Bodnar]

This is one funny looking j-link.  Did Segger change the design recently?
Leo

[Fraser]

Hi Frank,

Is there any chance of an upgrade kit to add the push button rotary encoder and additional kickback protection to the MK1 unit please. I do not need the broader range regulator though. It would, of course require some reworking of the PCB by me, but that is no problem. New firmware would also be needed but I have your programming cable already.

It would be great to be able to update the Mk1 to at least the essential parts of the Mk2 design.

Fraser

[tatus1969]

Is there any chance of an upgrade kit to add the push button rotary encoder and additional kickback protection to the MK1 unit please. I do not need the broader range regulator though. It would, of course require some reworking of the PCB by me, but that is no problem. New firmware would also be needed but I have your programming cable already.

That is possible. Here is a picture of the additional protection elements. All parts are available at Digikey:
SLD8S28A (large 28V TVS, Littelfuse)
SMDJ28A (28V TVS in SMC package, Littelfuse)
2920L030 (resettable PTC fuse, 300mA, Littelfuse)

The first one is new, and it goes in parallel to the MOSFETs. The other two replace the "quick fix" that I recommended before. I didn't like the resistor in the current path, that fuse also adds resistance.

The rotary encoder is not that simple, because it uses new microcontroller pins and has a different footprint. If you still want to do this, the part no. is PEC12R-4220F-S0024 (also Digikey). I have attached the relevant schematic diagram page here as well.

[Fraser]

Thanks Frank,

That does not look too hard to achieve I will order the parts.
Thank you for the schematic. Will you make the new firmware available as well at some point in time please ?

Best Wishes

Fraser

[tatus1969]

Will you make the new firmware available as well at some point in time please ?

Yes of course :-) Otherwise the firmware update dongle would not make much sense ;-) I plan a kWeld product page that will contain links to available firmware versions together with the revision history.

[tatus1969]

This is one funny looking j-link.  Did Segger change the design recently?
Leo

That is an older version I got long ago from ebay. Hopefully not a clone?

[Fraser]

Frank,

I have received the V2 to V3 upgrade information pack.

Thank you so much for going to the trouble of producing such a detailed document with such excellent pictures. I had not realised that the MOSFET drive had also increased.

Best Wishes

Fraser

[tatus1969]

I have received the V2 to V3 upgrade information pack.

Thank you so much for going to the trouble of producing such a detailed document with such excellent pictures. I had not realised that the MOSFET drive had also increased.

No problem, you're welcome! Less than 2hrs of work.

Yes, I forgot to mention the MOSFET drive improvement. I realized that this was necessary when going the route towards 2800A. With the standard configuration and the recommended Turnigy it is not necessary because the current is "only" 1400A, but it doesn't even increase the kit price. I was afraid of excessive ringing from overly fast switching at the very beginning, but that has proven not to be problematic at all.

The inductive spiking countermeasures have the same rationale. The rev2 has no problem without it at 1400A, but I expect that many people will have their own and more powerful source. I had one customer who wasn't able to calibrate the unit. It turned out that this was due to the overcurrent protection constantly tripping.

I have attached the manual here for everybody. I would be willing to make a group purchase of the necessary components, just send me a PM. I will also make the entire upgrade free of charge, I would just want to have the shipping paid.

UPDATE: the parts are 7.78€ per set, so I would need at least 7 requests to get above Digikey's free shipping threshold. I will ship in simple envelopes for 0.90€.