Next level DIY battery spot welding 

thanks :-) There's a new cherry: the latest model also supports an anti arcing feature that protects you from burning a hole in the battery when accidentally triggering a plasma arc due to dirt, corrosion or too little contact pressure.

Just check out the video description, there's a link to the shop where I am selling these.

Thanks for the Kudos!

Where can we find PCB schematic, firmware etc ? :)

I am afraid but this one is not open source.

Thanks, and yes a good product should be continuous improvement. Do you have an idea of what could be added?

Thanks a lot! I am working on sourcing the next batch of 100 kits, I will open for pre-orders soon. I will inform about this through the (spam free) newsletter at kicksurfer.de, I suggest that you subscribe to it.

@@wearethewatt2950 how could I get one?

ship to USA?

How much it cost to USA???

How can I order this or the parts to build it to Chicago, USA?

great! Let me know if you have any issues.

I put it together, but won't be able to use it until next month when I move into my home. If I have any issues then; I'll let you know, thanks.

It works well with lead acid batteries, some users are using them. I am collecting info about suitable supplies here: www.kicksurfer.de/index.php/2017/08/29/kweld-power-supply-recommendations/

thanks :-))

no, can't do copper. Thanks for the Kudos :-)

Almost! It is doing this continuously while welding, at a very high update rate thanks to the powerful microprocessor.

Impressive! I might order one when they're back in stock ;)

Sorry, I confused the battery that is welded on, with the Lipo power source. I thought, why should I apply pressure to the latter? Makes more sense now :-) My system does not require consistent weld spot force, and it does neither need nor use cleaning pulses. That is because it works differently from all other spot welders that I have seen before. I'll try to explain: Commercial spot welders fire pulses with a given duration. If there is not enough pressure, or if there is a bad contact from corrosion etc, then the spot's ohmic resistance is different from what it should be - the weld fails. My design measures the instantaneous power that goes into the spot, integrates this to obtain the so far deposited amount of energy, and stops the pulse when the set value is reached. Al this is done every 10 microseconds. This ensures that the amount of energy, or heat, that is dumped into the weld spot, is *always constant*, no matter how much initial pressure you apply, how much corrosion there is, how well the surfaces tough, and so on.

Have you noticed that I have a US reseller? Infos right at the beginning of the product page.

I just ordered mine from the USA.

Where did you see high voltage in a resistance spot welder? For this kind of welding, you just need insanely high currents. The resistance of a weld spot is low, maybe one milliOhm. And you need to heat it up such that it melts.

many people use a microwave transformer- but actually I guess in adding a large secondary they are actually using a low voltage at the weld. So the big difference then is just using wall current versus battery current, and to produce the hundreds of amps needed a relatively high primary to secondary ratio is needed. But I always thought that arc welders use high voltage- but maybe no.

Yes, the microwave designs also produce low voltage, around 1-3V on the secondary. But hundreds of amps is not enough yet, also these are able to do >1000A when made well. Their advantage is that it is much easier to switch the current, but their disadvantage is that they can only chop the welding time into 10 millisecond intervals (one half AC sine wave). I need much better resolution to implement my energy based pulse control and therefore chose DC supply and MOSFET switching. I think that arc welders are using higher ignition voltages, but this one is not an arc welder. This is called resistance spot welding.

oh ok thanks for explaining- interesting that yes- it must be that 60hz is only 16ms I see your point about a higher switching frequency.

I'm based in Germany and therefore sell in EUR, so that depends on your exchange rate. The shop is at www.keenlab.de/index.php/product-category/kspot-welder-kit/

It gets hot after some time, because it receives heat from the weld spot naturally. But you don't burn your fingers :-)

Thanks for your suggestion. Sadly that electrode holder is already obsolete, I didn't like the machining effort that had to go into it and found a simpler solution. I hope to upload the next video by this weekend showing this and other improvements.

p.s. yes I am in mass production, if a first batch of 25 can be called that :-)

what sheet metal thickness do you aim at? 2.5kA is not a lot for this. Professional sheet metal welders can easily go up to 15-20kA. BTW, my design is calculated for 2kA max for 200ms.

Hi there, i doubt that a LiPo-battery can give you 2 kA - did you check the amps? I know that industrial spotwelders can deliver up to 30 and 40 kA, but i used an old welding transformer and replaced the secondary by 4 windings of 120 mm2 copper cable giving me up to 4 V at around 2,5 kA (for a few seconds). That should be enough for 1 - 1.5 mm stainless stell i guess. More would be better tough...

I measured the amps through a defined section of (brand quality pure copper) AWG8 wire with known resistance. Modern high amperage race Lipos are well capable of delivering these currents. I just received a fresh Turnigy nanotech 3S / 5Ah pack. That is rated for 130C discharge current, which is 650A. "Continuous"!

It's the current handling capacity (2000A) that makes the components expensive, and the reliability that my design stands for. Battery spot welding needs at least 1000A to work properly. The amount of energy is varied by adjusting the pulse duration. However, the numbers in this first videos are wrong (current measurement was inaccurate at the time of making it). The series product can do up to 200J, and needs 50-100J for 0.3mm nickel strips.

in kWeld, this is handled by an array of large TVS diodes. As you have experienced, the MOSFETs alone cannot handle the amount of inductive kickback at these current levels.

kWeld has actually been featured in his one of his most recent videos: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-b2sBhDxmPmA.html as I had send him one for his ebike work.

yah i know that was great to see.05:50 ;-) i also so him trying to create his own hahaha..was cool to watch.. There are lot's of cheap Chinese bulky and not so trust worthy solutions out there. and since i live in a apartment your compact solution has a high wife approval factor ..also designed tested and build in Germany finishes the stamp of approval .. for once i'm happy to live in The Netherlands and not having to pay customs crap..But the kit will have to wait for a while until i finish my current projects studies and lot's more...my compliments btw..

Thanks! When I saw him trying to build a spot welder that was actually when I decided that I had to donate one to him ;-)

Thanks a lot :-)

Yes it is :-)

Could have also used a 12v starter relais. These things are for high start currents at 12V of around 250A and are really cheap like 2 dollars per piece.

can you please supply a link? i looked for it and have not found it. thanks

Couldn't you tell by the accent? LOL.

this project reminds me of another one my some years ago, ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-dd-Vb1LqkLA.html

The cells that I use in this first video were not appropriate, and I had made quite a bit of research to come up with proven alternatives. The kWeld product page lists these in the text. And I am working on a capacitor solution that turns kWeld into a CD welder.

WeAreTheWatt uhm using supercapacitor?

That's already done and for sale in my shop, just didn't have time to cut the video footage that I have made weeks ago :-/

As this is welding, all corresponding precautions apply here. Protect your eyes, make sure that your hands don't get burnt, and work in a fire safe environment. This system runs from a 11.1V battery, which means that there are no dangerous voltages involved, hence no risk of electric shock.

Thanks for your reply. I had assumed that such high amperage would be dangerous, but I guess not at these voltages. For eye protection would you suggest welding goggles (darkened glass) to be necessary or will regular clear plastic eye protection glasses be sufficient?

Only amperage that goes through your body is dangerous, and at 12V this just doesn't happen. The only danger here is getting burnt. For eye protection, use clear glasses as there is no intense light involved.

@@wearethewatt2950 you are ignorant if you think 12V can't hurt you badly. 12V at the correct amp can kill you if it goes across your heart.

I'm not. Please inform yourself before making false claims. For example, en.wikipedia.org/wiki/Electrical_injury#Body_resistance

AFAIK those Lipo jump starter packs are basically just the battery and a schottky diode for reverse polarity protection. There is no power switch, you just connect them to your battery. The current is typically also lower, a few 100A. This unit is much more complex, as it has to reliably switch up to 2000 A.

if i was the iventor , i just throw the clones my self on ebay first ;D. the price is to high for a piece of electronics. Sure its very useful and engeneered. But just my opinion a very experienced buyer, the value must be higher than the price you must pay. And this is not the case here. (Microcontroller,Mosfets, a PCB cant justify 150/213€) i would say at max around 66€ and a on "sale" price of 50€ would get him way more sales.

This is not true, as a service engineer. you would need something portable. If we would use this more frequent than this is perfect. just if would have a nice case and isolate probes. This is really a interesting device.

Its pretty interesting and compact. I could see it's use, but I think my diy mot will work great for what I want. I might consider something similar in the future, but at his price I would probably just make my own. I wouldn't say that it is highly overpriced, but I enjoy doing it myself and currently I still have the time to waste.

The assembled PCB is 110€, the prices that you state are for the full kit I presume. That includes all mechanical parts, some of them custom CNC machined, and all cable assembly plus work plus packaging. If you would want to sell that assembled PCB for 50€, then you would certainly have done your calculations wrong. Just the 6 MOSFETs, the microcontroller, and the LCD account for a bit less than 40€ at Digikey.

the fets alone will run 25 bucks at the moment. i really wouldnt hold my breath on that one. if you have a lithography set up and etch tank at home you might make it for close to that price...assuming you also have a battery or cap bank already

I'm working on sourcing the next production batch, I'll inform through the kicksurfer.de website.

yes but new material is ordered. I expect new stock by end May 19.

zu dem Zeitpunkt hatte die Firmware noch falsch gerechnet, das ist längst behoben :-) Die richtigen Werte habe ich dann im Manual aufgelistet.

you can't get shocked from something that only uses 12V, and your assumption is absolutely right. It's Ohms Law that applies here of course.

WeAreTheWatt thank you fo the help I understand now

thanks :-)) Yes the firmware is monitoring the battery voltage and will display a warning when it goes low. The voltage at which is does that is programmble, to allow for different chemistries and cell counts.

That would be a good idea, but it would need to be one with an extremely low internal resistance. The battery that I am using currently has only a few milliOhms, and the capacitor would need to be well below that in order to make a contribution. Anyway, I am working on a stationery supply for the welder, which is built around such a capacitor.

Oh, you are right, I forgot about the importance of the internal resistance the power sources have in relation to each other. Perhaps an array of parallel connected capacitors would help a bit... I am looking forward to your power supply!

Fast. Das Material an den Elektroden hat einen Widerstand von 2 Milliohm. Bei 1000A ergibt das einen Schweißvorgang. Ich würde allerdings mittlerweile davon abraten, das mit einer 5Ah / 50C Lipo-Batterie zu machen. Für eine annehmbare Lebensdauer sind mindestens 500A Nennstrom notwendig, also zwei Packs davon parallel oder einen entsprechend dickeren. Hobbyking verlangt z.b. für einen 5AH / 130C (peak) Pack etwa 50€ (hobbyking.com/de_de/turnigy-battery-nano-tech-5000mah-3s-65-130c-lipo-pack-xt-90.html), selbst wenn man zwei davon parallel nimmt, ist das immer noch weitaus billiger als eine "klassische" Lösung mit Ultracaps. Für den notwendigen Kondensator wecher lange genug 1000A liefert zahlt man gerne einmal mehr als 200€.

you're welcome!

Danke :-) Der Preis wird noch deutlich runtergehen, wenn es klappt und größere Stückzahlen möglich sind. Im Moment mache ich alle mechanischen Arbeiten selbst und muss den Preis dafür relativ hoch ansetzen, um nicht überschwemmt zu werden. Es macht jetzt noch keinen Sinn, irgendetwas rauszugeben. Es lohnt sich hier selbstbauen, die Rohteile insgesamt dafür kosten dafür etwa 149€. Viele haben aber bereits einiges davon da, z.B. Fußtaster, Kabel oder Schrauben, dann lässt sich der Preis auf 99€ reduzieren. Vorsicht bei dem einen oder anderen Bausatz, vieles ist sehr knapp kalkuliert und "stirbt" schnell. Dieses Gerät ist deshalb teurer, weil es mit Reserve dimensioniert ist und auch noch 2000A klaglos schalten kann. Und die Energiemengensteuerung ist auch einzigartig im DIY-Bereich.

That'll liklely put the battery on fire :-)

Das nicht, aber es ist ein neues Produkt und ich habe zunächst einen ersten Satz von 25 Stück herstellen lassen, und diese per Pre-Order an bereits wartende Kunden verkauft. Es wird eine Testphase geben, während der ich mögliches Feedback/Verbesserungsvorschläge einarbeiten möchte, wonach der reguläre Verkauf im größeren Rahmen stattfinden soll, dann auch mit sofort lieferbarer Lagerware. Ich denke ich werde etwa Mitte/Ende September neu Lager aufbauen und werde über den Newsletter informieren.

I'm sorry but kWeld is not open source :-)

it was down because I maintained the products and started pre-orders! :-) You can check the shipping rates online, the shop has a calculator.

As I explain in the video, it works differently. It automatically determines the required pulse time based on the given conditions at the electrode tips, you just need to set the desired amount of energy. This ensures very consistent welds even with varying levels of corrosion or other dirt.

www.keenlab.de

This is a custom design that I have made, and I have them manufactured at reputable electronics assembly specialists. My current manufacturer is: www.sk-tronic.de/

The gate signal trace is on the top side, the PCB is double sided 1oz. That is enough because the power connection traces span the entire MOSFET group width, and are less than 1mm long. The bus bars quickly dissipate the heat after a pulse, I can barely measure a temperature rise. I calculated it for fun, the cross section is larger than that of the 300A fuse. I've added vias, and the same amount of copper on the bottom side, but that may not even be necessary. Don't route the current through vias anyway. A 0.3mm via can carry 1A, so you'd need >300 vias in total. Also keep in mind that minimum track widths and spacing is larger for 2oz copper, I would not go below 0.4mm for each. I have stopped making/drilling my own boards a long time ago, professional services are just too cheap nowadays. I would be able to sell the bare board of my welder for 10€, and sill make some profit.

It was the first board i designed and only 2 months ago, but I wouldnt have done it if not for realizeing I had a huge pile of plexi in my garage and a friend I could consult with chemistry on. through her advice on how not to die, I managed to produce efficent and refreshable cupric chloride from chemicals easily available. cost was about $20 and if i wanted to i could make 1.5 gallons but cupric grows as you use it so i started with 600ml for my thin tank. Acetone welds plexi if the edges are smooth. I already had a miter saw. Picked up a 80tooth 10in blade for 30usd and was able to construct a vertical bubble tank for the cost of the saw blade. I had been doing phone repair so I have a vacuum hot plate where heat and vac are switched independently. using more plexi and some disassembled light fixtures I was able to suspend a $35usd high efficency 365nm led at a variable z and y axis over the "floor" of the exposure box. drilling a few holes and using weather stripping allowed me to pull my glass plate down with vacuum so the transparency is compressed tight on to the uv resist film on top of the board. I have low iron glass coming this tuesday for even quicker exposure as normal glass reflects a lot of uv When i looked at proto services in the US the cost was about 111usd and a 1 week wait for the size of board i made. you get 3 for that price but my worry was how quickly cost would bankrupt me if I got it wrong with the first design. I would never do a production batch at home but the ability to prototype at home has opened the world of making products instead of repair them. I didnt end up spending much more than the 1 week service would have cost. After drill press is complete it was probably 250-300 spent. had i spent more time in research i could have done it for 200 but much of it was trial and error so I sometimes bought things that needed to be changed. now its together though i can get double sided 4x7in cc for a dollar a board. Anything i make in the future is now very cheap to proto. Later today im going to pick up some chemicals to try a via plating bath.

mine wasnt just a via, it was a copper rivit with the source lead stuck through and soldered onto the wide pad x6 n-fets ...irl83034

That's a quite impressive tool range :-) It deserves to be used. For PCB proto, you need to trade off cost and time. Companies like Elecrow or Seeedstudio will make them at low cost, but they don't give you order today, receive tomorrow. Only when I have to go fast, I order from local suppliers.My preferred German supplier PCB-Pool has a budget option with 8 working days and 43€ for 100x100mm 2layer 1oz.

thats not a bad price. i tend to be a rather impatient person lol. When I dont have to rivit through holes and can just plate them im thinking I will be able to do a 2 layer design in about an hr with 30min of that being hands on time. Less if I get help at some point and can teach someone to clean/prep and laminate a bunch of the CCs then insert them into a light proof bag. At night its no issue as my place is lit entirely by LED, but windows become an issue in the day. Right now im too swamped to prep a lot myself so I just do it as needed which adds to time/probably isnt as efficent

Yes it does, which is why I switched to a different model that is much more capable. I am discussing them in this video: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-rQnODV4VQjU.html

I don't understand why that would be meaningful? That would contradict the design concept of controlling weld spot energy.

Sorry, I was not clear.... The time I would want to control is how long one gets to position the probes before welding is triggered, when in automatic mode.,,, In the video the automatic mode appears very useful but the weld happens very soon after the probes make contact and I'm not sure everyone could position them accurately and press down ready for the weld to happen in time. Hope that makes it clearer.... nothing do do with the welding process itself.

Thanks for clarifying. Yes, this appears to be a very useful feature! I've added this to my todo list. But that also implies the decision to move from a potentiometer to a rotary encoder with pushbutton switch. One of the rationales behind this system is to make it as simple as possible. Not decided yet... But have you also watched the more recent videos of my channel? I have improved the auto mode to make it more usable.

Daves Tech no, the display reads 54.4ms.

Both aluminium and copper would require tin or silver plating to prevent oxidation, and brass is also easier to machine.

@@wearethewatt2950 thanks!

www.keenlab.de

;-) But your guess is a bit off, there is one zero missing - the current is 1400A.

Absolutely. These forums have lots of users that have kWeld and can give you feedback: www.eevblog.com/forum/projects/guesses-on-what-i-am-attempting-here endless-sphere.com/forums/viewtopic.php?f=14&t=89039

I have no references for this from other customers, everybody uses 3S Lipos or 12V starter batteries. Depending on the accumulated ohmic resistances in the system, your battery may deliver too much current causing the overcurrent protection trip at 2000A. A higher voltage also increases the likelyness of arcing, which in turn causes excessive electrode wear.

WeAreTheWatt OK. I'll keep this design in mind once i have enough spare cells to make dedicated battery, or have a reliable way to recharge my truck batteries paralleled.

www.kicksurfer.de

The first production batch had a 10% discount, which this one doesn't. But $260 seems to be a too high number. The kit with cables for self-assembly is €156 plus shipping. You can also order just the electronics module for €110 and provide or make all other parts. This kit is a bit more expensive than other similar kits, but I have also invested into parts to make it more robust. Just the six MOSFETs cost roughly €30 at Digikey.

$165 for the unit $51.48 for DHL shipping a little pricey for such a small item. $34.56 Mwst.? Why? $216.48 total which thanslate to over $260 US

The €34.56 is the VAT (German: Mehrwertsteuer, Mwst) that I have to deduct, and €43.26 is the business rates that I get from DHL. I assume that you are located in the USA, that's the DHL price for an overseas parcel. None of these go into my pocket. I have a US reseller though, please contact me through the website contact form and I'll give you his contact. Part of your high price is the recent isolation policits in your country, which is weakening the US dollar. Only a year ago the US$ / € conversion rate was much better for you. But that's an entire differen topic...

just check the video description, it has all the links. The main shop page is www.keenlab.de/index.php/product-category/kspot-welder-kit/

Please explain what you mean exactly.

He means the crimp on screw terminals I think.

check out my newer videos on this channel, I have already made one :-)

Hi Ray, I have been answering all your emails but all my emails get rejected by your hoster (IP blocked by EarthLink). I'm glad that you tried it this way now, as I'm sure that you can read my reply here. I am busy building up new stock, and expect to receive new units from my electronics manufacturer in 6-7 weeks from now. And sorry for not answering the phone, but my mobile has a cheap contract and it would be very expensive for me to make international phone calls (if that is the number that you dialed). Please use the phone# in the impressum, and leave a message if I'm couldn't answer.

ok, thank you... i am anxious to get a kit...

Hallo Grustig ! Hey, Herr Frank, i am supposed to be coming to Germany / Austria soon. by any chance you have kits again yet or not ? i was hoping if i was there i could buy a kit. please let me know, Danke.

Hi Ray :-) I expect new kits by the end of this June. By the way, I received another email from you a while ago, and my response to you got rejected by your email hoster *again*.

i do not know why it is rejecting you, very strange ? which email are you using from me ? .net or .com .net is the only email i had gotten from you before..

I have a prototype of the capacitor charger running at 60A continuous (80A short), heating up a lot of 50W halogen lamps. The capacitors need a base PCB with a balancer, I plan to continue this project in January.

ok that sounds great, and when i already have the spot welder i will be able to just hook it up to the new power supply as is?

yes, it will be an "official" accessory.

ok sounds excellent, ...on your waitlist now :-)

For kWeld, that would not be as easy, because the firmware of that is closed source and I have no ZIF adapter for the microcontroller that I could use to preprogram one for you.

WeAreTheWatt I see I thought this is just an stm32 in bootloader mode and I could program it with the firmware update package from you

That's true, however I am not using the built-in bootloader but a special one that I developed myself and that uses encryption.

that's possible, but it must be a very big one with an extremely low internal resistance (4..5 milliOhms max), in order to deliver the required current of at least 1000A (better 1500A).

all thanks, thought of 36v and 11ah. That would have to go?

The voltage is too high, kWeld is rated for 30V. But a lower voltage like 12V is advisable, otherwise you get too much arcing and electrode wear. And as I said, the internal resistance is important - you haven't given that number.

Alright thanks. Have looked a cell has already more milliOhm as it is forced. Then I will test it with lipo.

I am waiting for a definite date from my electronics assembly firm, they are purchasing components right now and it depends on their availability. If there are no stock shortages, I expect to receive material in 4-5 weeks from now. The kWeld shop page allows you to join a waitlist, and you get notified as soon as I have news. (You need to select a completion grade first, in order to see that button.)

I'm selling these :-)

Just do the math and calculate their electrical resistance (they are 12x4x75mm).

Energy is voltage x ampere x time, and this is what the unit measures.

when the fet is on? how fast do you need to sample to get energy in that short time

Typical pulse times are in the 1-10 millisecond range, that is no problem for a microcontroller ;-)

have you seen the link in the video description? It's all there www.keenlab.de/index.php/product-category/kspot-welder-kit/

we ship internationally! :-)

What do you mean by that?

What do you mean with that, do you mean physical pressure? I am not touching the battery.

YES THE PHYSICAL PRESSURE. EVAN THE ONE FROM CHINA HAS THIS. THANKS

Ah, now I get it. I confused the power source battery with the battery to weld on. Makes much more sense now. My welder does not need a constant pressure, and it does not need cleaning pulses. It measures the energy/heat that it creates, and keeps that constant, no matter how good or bad the contact is. That's the big novelty that AFAIK no other welder has, even the professional ones for 10000€.

It's not that easy, aluminium is highly corrosive and the "rust" is non conductive. At these cross sections (currently 4x12mm = 48mm^2), conductivity is not an issue, I concentrate more on cost and usability. Brass was most easily available for me for testing, but meanwhile I tried laser cut aluminium as well. I was not happy with the result, and will change to brass again. Check out my other videos, this one shows only the first evolution step :-)