"Copper/nickel sandwich" buses for series connections

[ThomDE]

I bougth a Kweld last week and assembled it last evening. Did some test on very old dirty cells to try the sandwich method, I tried 0.1mm copper together with 0.15 nickel with great success on 70 jules with a 750ah car battery pulling 1100 amps. Impossible to pull off by hand.

Now off to building 2x 6s2p pack for rc use with the sandwich method.

it's nog a great picture but just to prove the point.

 

[Cowardlyduck]

I am finally giving this a crack! Using 0.15 copper + 0.15 nickel. I am using a JP spot welder with 2 x 850CCA old car batteries and 2S2P (7.2V x 2) NiMh cells from a Prius. Even with that, I have to crank the JP all the way up to the max to get decent welds out of it.


So far so good, nearly finished the pack above now. I learned the hard way not to just spot weld the nickel onto the negative cell end at the same power level...had electrolyte leak out on that one!

Cheers

 

[spinningmagnets]

Try the "infinite slot" method on some old flashlight batteries. It should help.

Pic from ES member Willow

Spot Welding Copper Strips to 18650 Battery Cells

 

[999zip999]

I need to bump up this thread to see if anyone has anything else to add and I'm still reading it many pages.

 

[Cowardlyduck]

Well, I'm about to do my second build using this method. Lessons learned are that it's important to clean with isopropyl alcohol and that while it's hard to line up the positive side 'blind' it is possible. What's hard is paying attention for that many welds. It only takes one mistake to ruin your day!
I also found it works well to use copper at the terminal ends, where I still put nickle over top and soldered the main discharge to that.
I like not having to do cut so many nickel strips!

Cheers

 

[spinningmagnets]

From the e-motorcycle facebook, Shea Nyquist

"...This whole copper under nickel looks pretty cool! I took an ohm reading and it went from 10-mOhms to 1-mOhm across the busbar. That's awesome! .005 copper with .010 nickel..."

 

[Cowardlyduck]

Here's a few shots of my latest sandwich including thermal testing under extended moderate load which showed no issues








Cheers

 

[Frank]

Does anyone know if the tape inhibits thermal transfer? (ex. to airflow over the surface)

 

[Cowardlyduck]

Does anyone know if the tape inhibits thermal transfer? (ex. to airflow over the surface)

It probably does, and that would be a concern if it weren't for the many other layers also on top of this....padding, heat-shrink, more padding, more heat-shrink, and then finally this battery is going in a back pack. So ultimately, a bit of tape on the cells isn't going to make much difference.

The best way to prevent heat build up is by oversizing the pack. The build above is 15S9P with Samsung 50S 25A cells. Capable of 225A, but it will only ever see 100A max with less than 50A continuous.
If oversizing to that extent isn't possible (because it often isn't) just keep an eye on temps. A properly build pack with quality cells, ample bus bars and wiring will generally not heat up that fast unless run at or over it's max rating.

Cheers

 

[cutlaszc]

Is it normal for the tips to wear out? I am welding 0.15mm copper with 0.1mm pure steel sandwich method at 50J. My source is a 4s8p 26650 lifepo4 Lithiumwerks battery. I'm new to this, it's my first time working with the kweld

 

[amberwolf]

If there's enough resistance between the tip and the welded surface to allow arcing between them, or enough heating to weld the tip a little to the surface, then they'll be damaged by it.

If the resistance and heating is only in the welded surface itself, the tips shouldn't be damaged by it.

I don't know if heating in the tip/surface contact point is necessary to create the welds in your setup, though.

 

[ta0s]

Great thread.

I'm using Molicel P42A 21700 cells with max continuous discharge of 45 amps. I'm doing an unusual battery configuration, similar to a "bosch powertube", with 10 21700s lengthways in an offset fashion, resulting in a cylindrical shape (3, 4, 3). The resulting battery config will be 20s8p, with 2 "10 packs" in series x 8 of these in parallel. Reason why I'm building such an unusual battery is to conform to frame tubes in my scratch-built BMX inspired emoto.

The challenge is I want to be able to pull ~320amps for short bursts from this battery @ 40amps per series connection.

I've been studying Matador's Ampacity of battery metal strips (thanks!)

I'm looking at 0.2 copper + 0.1 nickel plated steel, but this still only gives a theoretical continual of ~30 amps. Anything beyond looks difficult to hone in with the kweld.. I'm not keen to invest in getting 1800amps+ to do thicker copper (which is my understanding at this point)

Other options I'm considering is:

- soldering 3.5mm copper desoldering braid to nickel strip, and spotwelding the nickel on.
- https://www.aliexpress.com/item/1005005651952418.html?pdp_npi=3@dis!USD!US $1.27!US $1.27!!!!!@2103011217010806245588491e2dd2!12000033892821771!im!! (very expensive!!)

Anything I'm missing from anyone running 40amps through their copper nickel sandwich? Am I over thinking?

 

[ta0s]

I've opted to try the factory made copper nickel link off AliExpress. No factory data available on it's rated amps, but uses a 0.15mm nickel plated copper tab for welding, and 0.5mm copper for the actual link plate. Seller is claiming 200a, which seems very high. I only need to achieve 40a, so will test and report back.

 

[jonescg]

For reference I used 0.15 mm nickel and 0.07 mm copper foil (available from here in Oz: 5mm x 15m Copper Tape - Altronics)

I used a transformer-based spot welder (Sunkko 709) and it worked a treat. I can pull 100 amps from a single cell and the link barely gets warm. The cell is suffering in agony and getting pretty hot, but at least the connection is sound.

Biggest difference in my case is how I connect the nickel-copper tab to the busplate, which is by soldering. I suspect the copper is taking the bulk of the current and the nickel is largely there for structural support. Although, it surely takes about 1/5 of the current.

 

[cutlaszc]

For sandwich technique, Ni(Fe) + Cu + Cell cap, you'd use only Ni(Fe) "Nickel-plated steel.

After I had problem with my kweld probe shorting on one pad of the during-assembly-pack and blew a TVS diode. Replaced it, but still not welding from 50J upto 180J. Asked Frank, he suspected some problems with internal resistance of the switching MosFet ? So he was kind enought to send me a new pcb. Replaced it and did some mods on my kWeld - it works again. Now I can achieve 1700-1800A (140J) on actual 0.15mm Ni(Fe) + 0.2mm Cu for the remaining (about 80%) of the pack of 20S3P. The pack is finished with welding just today. I use 2 parallel 12V30A cheap PSUs with 2 CC-CV programmable regulators that I supply 14A each simultaneously (through power diodes). It works fine except that one regulator is crappy with oftenly flickering display. I had to remove the box and clean internal connectors with contact cleaner spray several times already.

Spot welding pure copper requires at least 2,000A to start melting, better 3,000A and higher. Current kWeld cannot do that directly - without any mods or technique like sandwiching. Frank is woking on a higher version of kWeld, up to 4,500A. Should be dued next year.

Problems with spot-welding Cu is that it has very low R (loss as heat, hi R= hi heat) and conducts current too well that it is not heated up enough to directly weld it. According to Frank, we will need to dump very high current in the shortest time possible (Q). While Ni has about 4X higher R and requires less amp to create adequate heat and weld successfully - but still high. Luckily for sandwiching method, there is Ni(Fe), which is steel (5-6X higher resistance than copper) with microns of Ni plated only on its surface. When spot-welded this Ni(Fe) become very hot (much higer R) - enough to also melt copper next it. Also with the help of slot (or slit) on Cu and/or Ni(Fe) separating both probes from each other that will force majority of current to go even deeper through cell cap instead of just bypassing through Cu (low R) or Ni.

There are details of other mods / tricks I made with my kWeld, i.e : big cables, probes, bus bars, V set, fuse and etc. Each contributes to the success of this welding Ni(Fe) + Cu on its own. I will share the details I mod sometimes later. It will be lengthy though. Busy with so many things going on now.

Cautions : we are dealing with very higher current here with kWeld. Accident can happen. Be very careful. I now cover the pack with insulation and open only the pads that I am welding to prevent short that can damage kWeld permanently. That one single short cuase me a lot of headache and 6 weeks to get it back on track again.

I am grateful to Frank (Tatus1969) for his help and time and truely believe that he really stands by his product and reputation. Thank you, Frank. Actually, this post is to contribute to what I have learnt a lot form ES and all the generous guys here.

Also with special thanks to the people who discover sandwich-welding method and share it here. This technique enables us to use considerable lower current (with much less expensive welding machines) to successfully weld Cu.



View attachment 325541



View attachment 325542



View attachment 325543




** for some reasons, I cannot attach MP4 VDO clip for now. Otherwise, I can show my display and some of you guys can see what's with my system. Would anyone tell me how to do it, pls.

Hello, you can upload the video to YouTube and share the link. I would really like to know what modifications you made to your Kweld and what you use as a power source?

 

[cutlaszc]

The one on the left is 0.1mm pure steel and the one on the right is 0.15mm nickel-plated steel, the copper is 0.15mm, the kweld is at 80J and the power source it uses is a Lithiumwerks 26650 lifepo4 battery.

 

[Cowardlyduck]

Anything I'm missing from anyone running 40amps through their copper nickel sandwich? Am I over thinking?

I would say you are overthinking it. I would just add an extra layer or two of nickle on top and it should be fine.

In the past I've build large packs with only nickle and many (6+) layers to increase current capabilities. Noting with each additional layer it is not a linear increase it still helps enough to allow huge (120A+) draw without issue. I see no reason the same could not apply to the copper/nickle sandwich.

Cheers

 

[chuyskywalker]

Is it normal for the tips to wear out?

Yes; they are wear items. You can grind them back to a point and round it out again, though, to get more use out of that probe.

 

[ta0s]

I would say you are overthinking it. I would just add an extra layer or two of nickle on top and it should be fine.

In the past I've build large packs with only nickle and many (6+) layers to increase current capabilities. Noting with each additional layer it is not a linear increase it still helps enough to allow huge (120A+) draw without issue. I see no reason the same could not apply to the copper/nickle sandwich.

Cheers

Thanks. I've opted with the factory made link with 0.5mm copper bus and 0.15 copper/nickel plate connector as shown in the photo attached to my last post.

 

[riba2233]

Molicel P42A 21700 cells with max continuous discharge of 45 amps

just fyi, they are closer to 30A continuous (and that is in open air, less in the pack), 45 is peak and they will heat too much at that current. Be careful

 

[TheRealJohnnyB]

Thanks. I've opted with the factory made link with 0.5mm copper bus and 0.15 copper/nickel plate connector as shown in the photo attached to my last post.

What are you using to weld .5mm copper?

 

[floydr]

You weld the .015 mm nickel tabs that are inside the .05 mm copper bus
later floyd

 

[pbowes02]

Hi, I have found this forum incredibly useful for choice of busbar material and wanted to press a query about an EV battery I am currently working on which has been passed onto myself. On the welds between busbars and 18650 cells the distance between the spots on the positive terminals of the cells are different than the distance between the spots on the negative terminals of the cells. I am not sure of the reasoning for this if someone could provide an explanation?

 

[argo]

Thanks. I've opted with the factory made link with 0.5mm copper bus and 0.15 copper/nickel plate connector as shown in the photo attached to my last post.

As said multiple times - great thread!

Hey ta0s - did you end up using those Aliexpress sourced nickel/copper links?
I'm only needing to pull a max of 25Amps per cell, but those links do look like an easy to weld solution.

 

[No_Shorty]

You weld the .015 mm nickel tabs that are inside the .05 mm copper bus
later floyd
View attachment 345053

Thanks for posting this... just found them on AliE from WELLGO (AliTools plugin is great for reverse image lookup). Is there any disadvantage to these over the sandwich technique? These seem to be perfect.