kWeld - "Next level" DIY battery spot welder
- Thread starter tatus1969
- Start date
Infernal Bill
10 µW
- Joined
- Oct 17, 2019
- Messages
- 5
Doubling up the parallel portions of the .15mm tabs makes a lot of sense. Thanks!
I am not a fan of double layers of nickel strips (due to adding multiple bottlenecks), or the 0.30mm thick nickel strips (although the thicker single-layer nickel may be rated for higher amps, the total current will have to flow through the same area of weld-connections). Of course they both work, but the thicker nickel still remains a resistor in the higher amp user-profiles (converting high amps into heat).
If you have too much resistance in-line when attempting to draw high amps, you will experience some "voltage drop" under load. You mentioned 60A peak from a pack that I believe will be somewhat small, which suggests a high amp-draw from each cell.
I would recommend experimenting with the copper sandwich method using 0.15mm nickel and 0.15-ish copper.
0.15mm___6-mil__34 ga
0.20mm___8-mil__32 ga
0.25mm__10-mil__30 ga
If you have too much resistance in-line when attempting to draw high amps, you will experience some "voltage drop" under load. You mentioned 60A peak from a pack that I believe will be somewhat small, which suggests a high amp-draw from each cell.
I would recommend experimenting with the copper sandwich method using 0.15mm nickel and 0.15-ish copper.
0.15mm___6-mil__34 ga
0.20mm___8-mil__32 ga
0.25mm__10-mil__30 ga
ElectricGod
10 MW
spinningmagnets said:I am not a fan of double layers of nickel strips (due to adding multiple bottlenecks), or the 0.30mm thick nickel strips (although the thicker single-layer nickel may be rated for higher amps, the total current will have to flow through the same area of weld-connections). Of course they both work, but the thicker nickel still remains a resistor in the higher amp user-profiles (converting high amps into heat).
If you have too much resistance in-line when attempting to draw high amps, you will experience some "voltage drop" under load. You mentioned 60A peak from a pack that I believe will be somewhat small, which suggests a high amp-draw from each cell.
I would recommend experimenting with the copper sandwich method using 0.15mm nickel and 0.15-ish copper.
Spinningmagnets alluded to this, but didn't make it really obvious.
Your weld spots are the high resistance part of any welded pack. More welds is better than fewer to reduce weld resistance. Deep welds helps too, but be careful to not blow through the bottom of the cell. You'll notice in the image he posted that the top of the cell has lots of welds.
I agree with him about layered welds too. I'm not fond of this. This comes down to the welds resistance again. If I was building your pack (small with high amp draws per cell) I'd be using thicker nickel or trying to weld copper to the cells and using lots of deep welds. I can tell you for a fact that a single 7mm x .15mm nickel strip gets pretty warm with 50 amps on it.
You can avoid welds all together. I don't suggest soldering to 18650 cells as the positive cap has a seal under it. Melt the seal and the cell can leak. However Boston Power cells have an isolated cap that is somewhat distant from the seal. Unless you linger on the cell top a fairly long time while soldering, you won't melt the seal. Solder is a better conductor than welds. The bottom of the BP cell has a factory installed nickel strip. You can solder to these cells directly without issues. This might not be the best option if you need maximum charge density for the weight. They are not as good in this regard as the 30Q, but the cost is significantly less.
If you want to learn more about BP cells go here:
https://endless-sphere.com/forums/viewtopic.php?f=14&t=90938&hilit=boston+power
LOL, thanks for the Jackie Chan pic.
In the nickel-copper "layered" pic, the nickel is only there to focus the welding current, instead of dissipating it, like a "copper only" bus would do. In fact, the nickel only needs to be a small square over the cell-ends, because a full-length strip would have barely carried any current at all, since the current would take the path of lowest resistance, which is clearly the copper.
Copper is four times more conductive than nickel, so 0.15mm thick copper would have the same current-carrying ability as 0.60mm nickel, but with less resistance or waste-heat (22% vs 100%, IACS). The copper is also a much better thermal "sponge", to absorb temporary peak cell-heat, and then dissipate it over time. Copper is also cheaper than nickel, and readily available locally (and easily cut). https://www.bluesea.com/resources/108/Electrical_Conductivity_of_Materials
If copper is so great for using as a bus, why add any nickel at all? Pure copper alone is very difficult to weld onto an 18650 cell-tip by itself, unless you have a very expensive laser-welder like the kind some high-amp cordless tools use. In this application for nickel, a stainless steel strip would also work, along with common carbon steel sheet-metal (although carbon steel would be at high risk for rusting).
In the nickel-copper "layered" pic, the nickel is only there to focus the welding current, instead of dissipating it, like a "copper only" bus would do. In fact, the nickel only needs to be a small square over the cell-ends, because a full-length strip would have barely carried any current at all, since the current would take the path of lowest resistance, which is clearly the copper.
Copper is four times more conductive than nickel, so 0.15mm thick copper would have the same current-carrying ability as 0.60mm nickel, but with less resistance or waste-heat (22% vs 100%, IACS). The copper is also a much better thermal "sponge", to absorb temporary peak cell-heat, and then dissipate it over time. Copper is also cheaper than nickel, and readily available locally (and easily cut). https://www.bluesea.com/resources/108/Electrical_Conductivity_of_Materials
If copper is so great for using as a bus, why add any nickel at all? Pure copper alone is very difficult to weld onto an 18650 cell-tip by itself, unless you have a very expensive laser-welder like the kind some high-amp cordless tools use. In this application for nickel, a stainless steel strip would also work, along with common carbon steel sheet-metal (although carbon steel would be at high risk for rusting).
ElectricGod
10 MW
spinningmagnets said:
Having failed to weld copper directly, maybe you have just explained why I haven't succeeded. The addition of the nickel on top of the copper would do the trick.
Thanks for posting that detail. I'll try copper again.
Yeah...copper is lots better than nickel...just couldn't get it to weld so I suggested thick nickel.
I'm going to mess with .15mm nickel on top of copper over the weekend and see how that goes. I have copper sheet that's .5mm and some that's .2mm. I really want to see if I can weld it down!!!
Infernal Bill
10 µW
- Joined
- Oct 17, 2019
- Messages
- 5
Thanks so much for the replies and input on this. I don’t have my welder or materials, yet, so I won’t be able to experiment this weekend. Looking forward to hearing the results on technique and conductivity.
ElectricGod
10 MW
I have an idea...
I don't know if anybody has tried this or not.
A bit of solder paste like this product from Kester. It's ground up solder into a powder mixed with flux in a syringe. It's commonly used for re-flow soldering SMD components onto PC boards.
https://www.ebay.com/itm/Kester-EP256-Lead-Solder-Paste-63-37-Syringe-Dispenser-w-additional-tips/201842021162?epid=931646955&hash=item2efeb8d32a:g:5qcAAOSwBLdZq348
The KWeld makes lots of localized heat. Enough so that 2 fairly high temperature metals melt together at the weld tips.
The entire metal area of an 18650 is quite easily soldered to. The problem is the application of heat over time needed to get solder to flow and not damage the seals on the cell. AKA spot welding...which provides a LOT of heat for a very short duration so that the heat doesn't spread into the seal and cause the cell to leak.
My idea is to put down solder paste on the cells exposed metal. Then spot weld down copper or nickel. I'll get weld spots, but I'll also get solder flow around the weld spot. You should be welding down every connection to a cell at least 3 times. Surely 3 or 4 weld points will also make a really good solder connection along with making the spot welds.
The spot welds are the highest resistance point between the cell and the nickel or copper. The solder paste ought to mitigate this almost completely. I bet it also makes the weld points virtually impossible to break off too.
I don't know if anybody has tried this or not.
A bit of solder paste like this product from Kester. It's ground up solder into a powder mixed with flux in a syringe. It's commonly used for re-flow soldering SMD components onto PC boards.
https://www.ebay.com/itm/Kester-EP256-Lead-Solder-Paste-63-37-Syringe-Dispenser-w-additional-tips/201842021162?epid=931646955&hash=item2efeb8d32a:g:5qcAAOSwBLdZq348
The KWeld makes lots of localized heat. Enough so that 2 fairly high temperature metals melt together at the weld tips.
The entire metal area of an 18650 is quite easily soldered to. The problem is the application of heat over time needed to get solder to flow and not damage the seals on the cell. AKA spot welding...which provides a LOT of heat for a very short duration so that the heat doesn't spread into the seal and cause the cell to leak.
My idea is to put down solder paste on the cells exposed metal. Then spot weld down copper or nickel. I'll get weld spots, but I'll also get solder flow around the weld spot. You should be welding down every connection to a cell at least 3 times. Surely 3 or 4 weld points will also make a really good solder connection along with making the spot welds.
The spot welds are the highest resistance point between the cell and the nickel or copper. The solder paste ought to mitigate this almost completely. I bet it also makes the weld points virtually impossible to break off too.
I tried the same thing with just piece of hammered down 0.8mm solder and some flux paste mainly to keep it in place before welding. You can clearly feel it melting away with just one weld. Cant really recomended it since it was so much extra labor that next I rather try the nickel on top method. And I dont know about the resistance since solder should be rather poor conductor compared to copper..ElectricGod said:
I'm not convinced of that... Yes, they are tiny, but they are also very short. So I did some back-of-the-envelope calculations: spot height 0.5mm, spot diameter 2mm, shape of a cylinder, material composition pure steel, 4 welds per cell. That's maybe even too generous since the spots are probably much thinner and contain some nickel as well, but that way we get a worst case. The electrical resistance of this link is then 0.2 * 0.0005 * PI * 2^2 / 4 / 4 = 8 microOhms. That's pretty negligible, if you ask meElectricGod said:
EDIT: I made a typo, correct formula is 0.2 * 0.0005 / (PI * 2^2 / 4) / 4. Result is correct though.
ElectricGod
10 MW
tatus1969 said:I'm not convinced of that... Yes, they are tiny, but they are also very short. So I did some back-of-the-envelope calculations: spot height 0.5mm, spot diameter 2mm, shape of a cylinder, material composition pure steel, 4 welds per cell. That's maybe even too generous since the spots are probably much thinner and contain some nickel as well, but that way we get a worst case. The electrical resistance of this link is then 0.2 * 0.0005 * PI * 2^2 / 4 / 4 = 8 microOhms. That's pretty negligible, if you ask meElectricGod said:
That's a good point...
I just finished up building a 20S pack. That's a fiddling 160 micro ohms across the entire pack. All cells got 4 spot welds on them too.
ElectricGod
10 MW
ossivirt said:
Too many things going at once ATM...
I need to finish up this 60Ah pack first, but it's close. I have to finish up the BMS yet and then get it all in the battery box.
After that pack is finished, I can get back to insulating my welding probes and trying solder paste under the nickel or copper.
My next pack will be 400 cells or 20S20P. I'll have plenty of opportunity to do more spot welding soon!
Has anybody tried welding aluminum with the KWeld? I have 2 cases of LTO cells. I forget the exact number, but I think it's 240 of them. They are 11Ah pouch cells with aluminum tabs. Usually factory built packs are done with ultrasonic welding. This is quite expensive! The welding machines typically start at $4000 and go up from there. I was thinking aluminum is going to have a decent amount of resistance and probably enough that I can use the KWeld to spot weld the aluminum tabs on the LTO cells. I bought them to mess with, but also to use in a 45S arrangement in a scooter. LTO at full charge is 2.8v so it takes quite a lot of them in series compared to LION or LIPO.
That precision resistor I knocked off and lost from my first super cap board is on order, but Mouser says they won't be in stock until January. grrr! I'm half tempted to make up a resistor myself. It's not hard. I'm sure I can scrounge a resistor of a slightly lower value. Then all I need to do is file a V in the side of it and reduce the amount of resistive material a little until I get to exactly 6.75 Kohms. That will take maybe 5 minutes...instead of 3 months!
I did that for a customer to test kWeld suitability. Tested cells had alu cans - does *not* work.ElectricGod said:
https://www.digikey.com/products/en?keywords=RNCF0603BTE6K57CT-ND%20 and it's 6k57 :wink: Didn't scroll up, did I give the wrong info? If yes, then sorry!ElectricGod said:
Copper or aluminum tabs *might* be able to be spotwelded if using the sandwich method. The aluminum tabs should be pressed together between two outer squares of nickel.
The thickness of the nickel and the kWeld settings are unknown. This is only a suggestion for experimentation, and has not been verified on ES.
The thickness of the nickel and the kWeld settings are unknown. This is only a suggestion for experimentation, and has not been verified on ES.
ElectricGod
10 MW
spinningmagnets said:
It's the localized heat that does the trick. So the nickel square...probably .15mm will work...is the current and heat path into the aluminum.
I think the sandwich method ought to work. Aluminum liquefies at a much lower temperature than nickel or copper so bonding aluminum might be pretty easy.
There is a detail I learned from TIG welding aluminum that may be an issue here too. Aluminum oxidizes and makes a layer that interferes with TIG welding. You need to scrub off the oxide layer before TIG welding or else the weld is poor. The oxide layer is much more resistant to heat than is bare aluminum and it doesn't flow or bond into aluminum very well. This may be an issue with spot welding too. A small wire brush may be needed to scrub off the oxide from the battery tabs before spot welding.
parabellum
1 MW
Spiningmagnets is probably right. Watch the video below, nice trick.ElectricGod said:
[youtube]gcgC3V3mkcw[/youtube]
eee291
100 kW
why don't we just replace the copper tips with metal ones?
eee291
100 kW
Thanks for the correction but I think you know what I meant :wink:
Infernal Bill
10 µW
- Joined
- Oct 17, 2019
- Messages
- 5
Any thoughts on whether tinned copper will weld without the nickel sandwich? I have some on order and will experiment unless someone else has already done this.
