kWeld - "Next level" DIY battery spot welder

[flywire]

Right click to open in new tab

I'd have expected the screws holding the electrode to be towards the bottom of the electrode holder not the top so they can still hold worn electrodes.

 

[flippy]

Thanks for the correction but I think you know what I meant :wink:

yes, but it would not be as funny.

on a more serious note: higher resistance metals that can take higher temps is a cirular fail as you need more power as the electrodes take up more power then the weld itself so you get even more heat and everything else needs to be bigger and better as well and the only thing you are doing in the meantime is wasting power. lower resistance is arguably the most important metric for electrode choice.

 

[ElectricGod]

Thanks for the correction but I think you know what I meant :wink:

yes, but it would not be as funny.

on a more serious note: higher resistance metals that can take higher temps is a cirular fail as you need more power as the electrodes take up more power then the weld itself so you get even more heat and everything else needs to be bigger and better as well and the only thing you are doing in the meantime is wasting power. lower resistance is arguably the most important metric for electrode choice.

Steel rod is readily available. Make new tips out of steel rod. I'm thinking you might need to clean off the tips a bit more often, but it ought to work OK so you don't need to lay down nickel on top of the copper or aluminum.

I know this will raise the resistance, but it's not like you are going from 1 mOhm to 10 ohms. I'm just guessing since I haven't checked, but the steel probes will add a couple mOhms. It probably won't be enough to impact welding any more than adding a thin piece of nickel on top of the copper or aluminum would. The important detail is getting heat focusing at the weld spots.

Brass rod may be even better. It's going to be lower resistance than steel, but maybe enough at the tips to focus heat?

 

[eee291]

Thanks for the correction but I think you know what I meant :wink:

yes, but it would not be as funny.

on a more serious note: higher resistance metals that can take higher temps is a cirular fail as you need more power as the electrodes take up more power then the weld itself so you get even more heat and everything else needs to be bigger and better as well and the only thing you are doing in the meantime is wasting power. lower resistance is arguably the most important metric for electrode choice.

They don't need to be large, you can get 3.2mm wide tungsten rods that only need to be maybe 15mm long with just 5mm protruding from the copper handle.
Anyways I'm just throwing out ideas it's not like I want to weld aluminum... maybe copper someday in the future.

 

[tomjasz]

why not use electrodes made for commercial welders?

https://sunstonewelders.com/electrodes/

 

[ElectricGod]

The idea is to be able to weld copper or aluminum with the KWeld and not need an intermediary metal...like a square of nickle.

Maybe those copper and tungsten electrodes might work.

Need to get the heat focused at the probe tips so the metals melt together.

 

[flippy]

why not use electrodes made for commercial welders?
https://sunstonewelders.com/electrodes/

first reason would be that they have "commercial" prices.

second would be that they need serious grinders to do the finishing the electrodes to the right shape.

 

[tomjasz]

first reason would be that they have "commercial" prices.

$20-$30? My POS dead Sunnko was $250. My new KWeld will be $234.

The real question is will they perform better?

 

[flippy]

The real question is will they perform better?

considering the kweld electrodes are also made from "the right stuff" i doubt you will see any difference with the copper based electrodes. if you want to go hardcore you can get the moly of tungsten electrodes but using them for nickel is not recommended.

 

[ElectricGod]

The real question is will they perform better?

considering the kweld electrodes are also made from "the right stuff" i doubt you will see any difference with the copper based electrodes. if you want to go hardcore you can get the moly of tungsten electrodes but using them for nickel is not recommended.

The point is to focus the heat like a little section of nickel does on top of copper foil.

Obviously the probe tips provided with the welder are pretty darn good for welding nickel, but welding copper, you need to do like SpinningMagnets said...add a section of nickel strip on top of the copper to focus the heat.

I've tried welding copper directly and completely failed to get anything more than very weak welds even at very high current and joule settings. What SM said was the key. You need a bit of higher resistance (like nickel strip) at the point of current path to create a hot spot so the copper welds down.

The idea is can you use some other form of probe tips so that you don't need the added piece of nickel on top of the copper sheet? It might completely defeat the purpose by adding too much resistance in the probe tips...hence discussing various metals like steel or brass for the welding tips.

I have another idea:
Why bother welding at all? I already mentioned using some reflow solder paste between the nickel or copper and the cell. As the weld is made, it creates a lot of localized heat. Enough heat to melt solder and short enough duration to NOT heat up the cell. This may be a great way to solder the nickel or copper to the cell without worrying about welds at all.

I'm still busy with the closing details of that battery pack build and a controller mod, but I'll soon be reporting what I find out.

 

[ElectricGod]

first reason would be that they have "commercial" prices.

$20-$30? My POS dead Sunnko was $250. My new KWeld will be $234.

The real question is will they perform better?

Make sure you have a good dump current source:
High C rate LIPO packs
Super cap module
Lots of parallel battery capacity in lower C rate LIPO

If you look back to page 33 in this thread, I tried a variety of lower C-rate LIPO options in parallel and I could not get good welds.

 

[PaulD]

I have another idea:
Why bother welding at all? I already mentioned using some reflow solder paste between the nickel or copper and the cell. As the weld is made, it creates a lot of localized heat. Enough heat to melt solder and short enough duration to NOT heat up the cell. This may be a great way to solder the nickel or copper to the cell without worrying about welds at all.

I've thought about this a bit, but never tried it. I thought maybe the solder wouldn't have enough time to flow and get a good bond, but I guess it doesn't need to flow far. Googled resistive soldering for some interesting applications.

 

[ElectricGod]

I was super curious to see how well solder paste would augment a spot weld so I pulled out the KWeld and a few dead cells to find out. I didn't clean up anything. The cells was pulled at random out of a bag of dead cells and the piece of .15mm nickel has been banging around my bench for several months for experimenting with various KWeld related things.

This may look like a lot of solder paste, but it's actually enough to make a solder bead maybe 2mm across. THe actual solder is powder in a flux paste form. I tried to stick with a similar amount of solder paste for every test. Here's the stuff I used. Everything below is 2 spot welds.

I didn't bother taking a picture of the cell at all. This was 2 sets of spots at 10 joules. The solder melted to the nickel, but not to the cell at all. There was almost zero adhesion between the nickel and the cell.

30 joules. I'm definitely getting an actual weld as expected since I already know that 30 joules all by itself will make a weakish weld with .15mm. Pulling this apart was not too easy. The solder was doing 80% of the work and doing it quite well. Those weld spots are a lot bigger thanks to the added solder paste than you'd ever get with just a weld.

This is 40 joules and I have to say the result is VERY GOOD! Not only do I get a spot weld, but the solder weld is very strong and spreads out a fair bit. To get nickel to tear a hole through, I need at least 50 joules. I'm getting that weld strength plus a bigger weld spot with 40 joules and a little solder paste.

In this image you can see the 30 joule weld to the right. It almost tears through the nickel. The welds seem to really dent in the nickel thanks to the added paste under the nickel. The actual weld on the cell top is at 40 joules.

Tearing off the 40 joule weld was doable, but it was a really GOOD weld! All the crusty stuff is the left over paste that didn't melt.

Again 40 joules and .15mm nickel. No tears in the nickel. Just a nice big round circle of nickel pulled out and remained on the cell top. I've never pulled off welds and gotten a weld spot this large before without applying 50+ joules to the weld.

Now I'm switching to 8 spot welds. The idea is to use up more of the solder paste and see what happens at 40 joules. I'm not sure what is happening, but I'm getting solder welds now and no spot weld. These joints are strong and I'd say quite reliable, but not like the above 40 joule weld which was both a spot and solder. Still...I'd trust this long term. Here's several consecutive tries and no actual spot weld.

Reference paste blob and nickel.

Random cells were used so I had equally dirty and flat surfaces for this test. This is just solder welds. They are pretty darn good, but I want a spot weld too!

After cleaning off a solder joint.

This is a single 50 joule spot weld on clean nickel and a clean battery top sanded with 220 grit. It's definitely stronger than the above solder only connection. It did weld strong enough to tear holes through the nickel when pulling it loose.

I took another random cell and cleaned off the bottom with 220 grit sand paper like I did for the 50 joule weld above. That fixed the solder only and no spot weld issue. We all know clean surfaces aids welding. This weld will pick up my 3 pound panavise and not tear at all.

After cleaning up the cell bottom, the soldering is great and the spot weld is like I had for that original cell at 40 joules. When I broke the weld, I didn't tear through the nickel strip this time, but the joint felt stronger than the 50 joule weld on clean metal. The flux residue can stay there for years and won't corrode anything. I'm not afraid to leave this connection with some amount of unused paste under the nickel. It will dry out in the air within a few hours and turn to inert white chalk.

Both surfaces cleaned up. I see no evidence of holes through the nickel, but I can tell you this joint is really strong.

Weld resistance: I'm sticking to 40 joules and .15mm nickel for the solder/weld test. Spot welding at 40 joules is too weak so I'll compare to a 50 joule spot weld like I have found works really well for .15mm nickel. I'll use 2 cells with the bottoms sanded clean and a spot halfway up the side sanded clean. The only difference will be use of paste and the weld power. All nickel strips are 1" long. I'm going to keep my spots the same distances apart for all welds...about 1mm gap between the weld tips.

2 cells sanded clean bottom and side.

I'm starting with the solder paste weld at 40 joules.

50 joule spot weld only.

Spot and solder weld at 40 joules. Notice the positions of the alligator clamps out on the ends of the 1" sections of .15mm nickel. I want all possible variables to be minimized. That's .00875 ohms. This is less weld energy, stronger joint and less resistance than a 50 joule spot weld. That's compelling!

Spot weld at 50 joules. That's .00919 ohms. This may not seem like a lot...just about 1 mOhm of difference, but it all adds up.

I tried to twist the 1" tabs sideways. Usually I pull straight up from the point of the weld to test strength. Spot welds on .3mm nickel at 100 joules will shear off with twisting sideways on the weld. 40 joules with .15mm nickel will shear off. I could not get any of these welds to shear before the nickel crumpled. I tried several times to get a shear failure and both weld types held up nicely.

Right now, I'm saying that solder/spot welding is better than just spot welding in every way.

Stuff yet to be tested:
1. Copper solder welded vs copper with a nickel square on top.
2. .3mm nickel solder/spot welded vs spot welded.

 

[tomjasz]

[



I have another idea:
Why bother welding at all? I already mentioned using some reflow solder paste between the nickel or copper and the cell. As the weld is made, it creates a lot of localized heat. Enough heat to melt solder and short enough duration to NOT heat up the cell. This may be a great way to solder the nickel or copper to the cell without worrying about welds at all.

I'm still busy with the closing details of that battery pack build and a controller mod, but I'll soon be reporting what I find out.

Thanks, I'm anxious to read the results!

 

[ElectricGod]

I mentioned in a previous post that I had accidentally dropped a screwdriver point first onto one of my two super cap boards. I knocked off a micro small precision resistor which went flying who knows where! I bought 20 6.75k resistors from Digikey among a few other parts I use fairly often. The parts arrived Saturday and tonight was the first chance to get my second super cap board running again. It's already modded with an LED volt meter and an external balance cable.

This is the 20 resistors. Yeah those tiny black specks in the white paper...that's them.

I'll let you do the conversion to mm, but that's a part 1/16" long.

NOTE: ALL work seen below was done with my Hakko 936 solder station and a small soldering iron tip.

The screw driver tip tore up a tiny solder trace that went to the IC and 2 resistors.

First step is to make a tiny L out of a single strand from a 28 awg wire to replace the missing trace. It's maybe 1/8" long. It has a bit of a solder blob on the resistor end, but that will smooth out in a minute.

The new resistor soldered in place. Now I need to clean up a bit of flux, add conformal to this spot and it's ready to go again.

 

[ElectricGod]

Now that I have that resistor replaced in the original super cap module. You know...the one I deliberately blew up just to see what the super caps could deal with? Yup...that one. Well it's back in service now so it was time to parallel the 2 modules. The old plastic shell is pretty thoroughly buggered up with the chemicals from the super cap explosion, but it still works. I've scrubbed the plastic parts with soap and water, but the plastic is permanently discolored.

The first thing to do was to take apart the new module and remove the input wires so they could be mounted to the original super cap module. The new module will always be the "front" module going into the KWeld.

The new module reassembled with shorter input wires and 8mm long bullets going into it. The 1.5" of cable I cut off got used for the output wires from the second module and they too get 8mm long bullets.

Nearly all assembled. Just doing a test fit before closing it all up.

Both modules closed up and about to charge.

Not bad...both are at the same voltage. I've checked all the super cap voltages and they are close enough to the same to not matter. I wonder if I can do 2000 amps now for more than a millisecond?

I did notice one detail about charging up. I can't bypass my precharge resistor bank at 8.7 volts anymore. I have to wait til I get to 9.2v now. I guess I don't really care since the resistor bank is designed around passing 50 amps at 9.44v. I might as well leave the bypass switch open all the time.

 

[tatus1969]

I'd have expected the screws holding the electrode to be towards the bottom of the electrode holder not the top so they can still hold worn electrodes.

That was done deliberately, to create a longer path that the picked up heat can travel before reaching the holder. I have quite a few of these in stock at the moment, therefore not easy to change that

 

[tatus1969]

My POS dead Sunnko was $250. My new KWeld will be $234.

The real question is will they perform better?

Never owned a Sunkko, but I can tell that I have quite a few customers who told me that they have dumped their Sunkko after having used kWeld 8)

 

[tatus1969]

I was super curious to see how well solder paste would augment a spot weld

Your results look very promising! One remark that matches your observations: solder paste contains not only flux but also chemicals that remove corrosion, which is why the recommended reflow profile usually starts with a preheating period at a lower temp like 150degC. This allows these chemicals to remove corrosion from the SMD components and the PCB pads, before raising the temp to the reflow level. As this isn't possible in the short welding time here, it explains to me why you found sanding necessary.

Notice the positions of the alligator clamps out on the ends of the 1" sections of .15mm nickel. I want all possible variables to be minimized. That's .00875 ohms.

That 8.75 milliOhm reading seems to be way too high to my 'electrical intuition'. I assume that your meter uses 4-terminal measurement, right? If that is the case, then I would use the two clips for the current terminals, but solder the voltage sensing leads directly to the nickel strips.

 

[tomjasz]

My POS dead Sunnko was $250. My new KWeld will be $234.

The real question is will they perform better?

Never owned a Sunkko, but I can tell that I quite a few customers who told me that they have dumped the Sunkko after having used kWeld 8)

A miserable experience.

Would you mind commenting on probes. I imagine somewhere in the thread you have, but I''m at a loss finding that. Are any of these materials a potential improvement? https://sunstonewelders.com/products/welding-electrodes/

Thank you!

 

[ElectricGod]

My POS dead Sunnko was $250. My new KWeld will be $234.

The real question is will they perform better?

Never owned a Sunkko, but I can tell that I quite a few customers who told me that they have dumped the Sunkko after having used kWeld 8)

A miserable experience.

Would you mind commenting on probes. I imagine somewhere in the thread you have, but I''m at a loss finding that. Are any of these materials a potential improvement? https://sunstonewelders.com/products/welding-electrodes/

Thank you!

I still have my Sunnko. It never did good welds...not once. I had it set to the maximum settings to get it to weld .15 nickel and even then it was just barely making welds.

The KWeld is doing the right thing...putting down controllable current like you need it where you need it.

One of these days, I'm going to do a review of the Sunnko and the KWeld. I won't have anything good to say about the Sunnko and lots of good things to say about the KWeld.

 

[ElectricGod]

Your results look very promising! One remark that matches your observations: solder paste contains not only flux but also chemicals that remove corrosion, which is why the recommended reflow profile usually starts with a preheating period at a lower temp like 150degC. This allows these chemicals to remove corrosion from the SMD components and the PCB pads, before raising the temp to the reflow level. As this isn't possible in the short welding time here, it explains to me why you found sanding necessary.

I can do pure spot welds without sanding...that's generally not much of a problem. I will sand an old cell from time to time just to see if it helps or if the cell is really dirty, but usually when I'm messing around with an old cell and trying different spot welds, I just use the cell as is. I'm not sure why it made a difference for the actual spot weld in conjunction with solder paste.

It seems likely that the solder paste just squishes out of the way when I press down on the welding probes. If it didn't the thin nickel just sparks like when you barely touch the probes to the nickel. I think I'm making sufficient contact with the cell through the paste.

You are right about the paste...it never gets a chance to burn off much of anything before the joint is cool again. What do you call this? Maybe... "High speed reflow soldering"

That 8.75 milliOhm reading seems to be way too high to my 'electrical intuition'. I assume that your meter uses 4-terminal measurement, right? If that is the case, then I would use the two clips for the current terminals, but solder the voltage sensing leads directly to the nickel strips.

In my pictures of the LCD you can see the model number for my mOhm meter. It's a real 4 wire meter and I zeroed the test probes before testing anything. I even clean off the inner surfaces of the alligator clips to make sure they don't add spurious resistance before zeroing.

My measurements are across a 1" long .15mm x 7mm pure nickel strip, across half a cells outer casing and then to another 1" long strip.

I think the measurements are valid. I think adding the cell wall, the welds and the specific nickel strip used accounts for the resistance I measured. The only thing I'd maybe do different is try another set of 2 wire test leads...which I don't have.

These were my resistance measurements for 12" long sections of 3 types of strips. Notice the resistance for the .15x7mm nickel...same thing used for these tests. This stuff doesn't exactly have a low resistance.

SS/Ni/Cu .1mm x 9mm 11.1 mOhm
Ni .15mm x 7mm 20.9 mOhm
Ni .3mm x 8mm 9.85 mOhm

I don't know when I'll get to it, but I want to try pure copper and .3mm nickel. They will both show significantly less resistance than the .15x7mm nickel does.

I guess a better test would be the side of a cell. Weld on either end of the side. This keeps the distances exactly the same. If there are thin spots in the corner area of the bottom of the battery casing that's eliminated too. I want to eliminate as many possibilities as can be accounted for. Maybe use a 2" length of .3mm nickel and weld to the ends of that to eliminate any variables is cell walls. The idea was not to just show resistance, but also weld strength on an actual battery.

 

[spinningmagnets]

When using copper strips with nickel over it, I am certain the nickel does not need to be as thick as 0.30mm, but of course it certainly would not hurt.

"...would you mind commenting on probes. I imagine somewhere in the thread you have, but I''m at a loss finding that. Are any of these materials a potential improvement?..."

Sunstone makes industrial welding products for factories that will run continuously. Of course their products for the electrode tips will work, but many have a proprietary interface, and even their copper tips are more expensive than the kWelds.

The kWeld electrode holders are solid 14500 copper, and rather than the replaceable tips being made so that they can only be purchased from Keenlab, they are made from pure soft copper, easily sourced locally at any major hardware store as grounding wire, and cheaply sold by the foot.

Pure soft copper has very low resistance, and it will produce the lowest possible waste-heat from the welding current. That being said, a frequent welding pulse can make them hot. Possibly so hot that the very sharp tip might soften enough to stick to the workpiece.

In that case, a tungsten tip would be more expensive, but it has a very high melting temperature, eliminating to issue with "sticking". However, tungsten has more resistance than copper, so the same amount of welding would cause much hotter tips. It would also require recalibration, and a higher joule setting.

Welding tip rods can be found that are half tungsten and half copper as a compromise. However, the higher cost of replacing worn tungsten tips may suggest that having a second set of kWeld probes would be the cheaper option, over the long term. One set working, and the other set cooling off.

 

[garolittle]

My POS dead Sunnko was $250. My new KWeld will be $234.

The real question is will they perform better?

Never owned a Sunkko, but I can tell that I have quite a few customers who told me that they have dumped their Sunkko after having used kWeld 8)

I am one of those who dumped the Sunkko after using the kWeld. 8)

 

[flippy]

I am one of those who dumped the Sunkko after using the kWeld. 8)

i stopped using after the transformer blew up all over the workbench.