"Copper/nickel sandwich" buses for series connections

spinningmagnets

Moderator
Staff member
Joined
Dec 21, 2007
Messages
12,952
Location
Ft Riley, NE Kansas
There is a popular thread called:

"Spot Welding Copper Strips to 18650 Battery Cells"
Spot Welding Copper Strips to 18650 Battery Cells

It is the place where the copper/nickel sandwich method was first reported, and enough interest has been shown that some members have asked that this method be given it's own thread. In short...copper foil is inexpensive and readily available as a bus material, but it is notoriously difficult to weld directly to the ends of 18650 cells (currently requires an expensive laser-welder).

If you overlay nickel (*or steel) ribbon on top of the copper, any one of the common "pocket welders" (kWeld/Malectrics...kWeld is more expensive, but is capable of higher currents) will be able to weld this combination onto the cells. The nickel only facilitates the welding, so it is not a conductor in this configuration, since current will follow the path of least resistance, which is obviously the copper (copper is four times more conductive than nickel).

The common and affordable 0.15mm thick nickel has proven to be adequate, and the nickel-plated steel ribbon will also work quite well.

The parallel connections can be just about any material that is conductive, since the current is quite small for all cells (less than 1A under all conditions). Nickel, and nickel-plated steel ribbon will work quite well for paralleling. They act as a resistor to slow cell equalization.

However, the series connections handle the current of the pack, so using copper there is the best material.

COPPER FOIL

one-mil is 0.001-inch

0.10mm___4-mil__38 ga
0.15mm___6-mil__34 ga
0.20mm___8-mil__32 ga

Pic courtesy of ES member "kdog"

file.php


Resistance of materials in Ohms

0.7 x10^-6 Ω.cm COPPER
2.6 x10^-6 Ω.cm ALUMINUM
6.0 x10^-6 Ω.cm ZINC
6.9 x10^-6 Ω.cm NICKEL
9.7 x10^-6 Ω.cm IRON (STEEL)

Conductivity in IACS

100 Copper
61 Aluminum
27 Zinc
22 Nickel
5 Iron-Steel

xxxxxxxxxxxxxxxxxxxxxxxxx

There is enough evidence in now that the optimum method has been identified.

[Darren2018] the combination that yielded the highest [tearout] strength with the least amount of energy was 0.10 nickel plated steel with 0.10mm copper...I have managed to get extremely strong welds on 30J with 0.10 copper and fully split 0.10 nickel plated steel

[Rio Cole] I am using 0.10mm copper sheet and the nickel plated steel is 0.15 thickness. I am using 60J with this material. I think a wide range of Joule would likely work fine. I strongly believe from my short bit of experience doing this that the sharp pencil-like shape of the electrodes is the key to being able to have full confidence in each weld

[ian M Molina] Welder: kWeld @ 110J
Battery: generic car battery 59533 (740A 95Ah), 0.15mm Nickel + 0.15mm Copper

[vanturion] The combination of 0.20mm copper + 0.15mm nickel-plated steel went much better with consistently strong welds at only 70J

file.php
"...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. Thats awesome! .005 copper with .010 nickel..." Shea Nyquist, Facebook e-motorcycle
 
Last edited:
Molicel 21700 P42A cells, 0.10mm copper sheet, plus 0.15 nickel squares only to facilitate welding. 20S / 6P

kWeld used, setting is 50-Joule

file.php


From serious_sam

"...I didn't have any issues with the cell/copper/steel sandwich method, but I only used 0.1mm copper and 0.2mm nickel plated steel. I regularly pull 150A through this pack with no problems, so I don't think I'd change my method at this stage..."

file.php


By "Tommy L", 0.10mm copper, 0.20mm nickel on top

90 joules @1150-1200 amps.

BatteryCopper1.jpg

copper for the high series current, nickel for facilitating the welding, plus the parallel current (which is very small)

BatteryCopperSeries1.jpg


From "devmonkey" arduino welder , 800cca and 25ms, Battery is a Numax XV31MF Heavy Duty Ultra Deep Cycle Leisure Marine Battery - 12V / 105Ah / 740EN / 925MCA
https://endless-sphere.com/forums/viewtopic.php?f=14&t=84680&p=1607061#p1607095

40T cells, 16s3p

file.php


https://endless-sphere.com/forums/viewtopic.php?f=12&t=110873#p1627835

Molicel 21700 P42A in a 20S / 8P, 72V. It has cell level fusing via 28 awg copper wire on positive end of cell

BatteryWeldCopper3.jpg

From Robbie Andrew Wright on facebook.

BatteryCopperNickel1.jpg

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

https://endless-sphere.com/forums/viewtopic.php?f=14&t=84680&start=500#p1654650

Member "pbert",

copper = 0.15mm
nickel = 0.2mm
spot welder = malectrics
spot welder setting = 26 ms pulse

The welds are solid, and I feel better having the 0.15mm copper than the 0.10mm even if the 0.10mm may have done the job

file.php
 
spinningmagnets said:
Molicel 21700 P42A cells, 0.10mm copper sheet, plus 0.15 nickel squares only to facilitate welding. 20S / 6P

kWeld used, setting is 50-Joule

file.php


From serious_sam

"...I didn't have any issues with the cell/copper/steel sandwich method, but I only used 0.1mm copper and 0.2mm nickel plated steel. I regularly pull 150A through this pack with no problems, so I don't think I'd change my method at this stage..."

file.php

I’m assuming the nickel didn’t stay or was pulled off?
Thank you so much for this!!!
I have 1,000 2170 cells I need to do right!!!
I got the Kwelder and pure nickel. I need the copper. 😊🙏🏻
 
I am also curious about the effect of using the thinner 0.10mm thick nickel ribbon for the welding cap, cheaper than 0.15mm, and easily cut with scissors. I would also have wanted to use 0.15mm copper because that could handle the max currents of the highest-amp cells, so I am looking forward to how that would weld up.

ES member ridethelightning has built a pack using the "infinite slot" method, using high energy. No nickel, just copper, but I am curious about if this might help the weld to be solid at a lower power level. (copper / nickel sandwich plus infinite slot)

https://endless-sphere.com/forums/viewtopic.php?f=14&t=68005&start=100#p1134835
file.php


Another variation of the infinite slot shown below, The copper strips have a dab of glue to hold them onto the nickel bus. These would be flipped over so only the copper touches the cell-ends

BatteryCopperSeries2.jpg


Cordless tool packs have a slot between the two welding probes to try and force all the welding current to take the pathway through the cell-end. They also use very high energy. If some of the welding energy passes sideways through the bus ribbon, then you have to use a higher welding current to get "enough" current to take the desired path.

Cells21700_11.png




SpotWeldCurrent1.png
 
Informative thread.

Where do you source your copper and nickel strips? Once I get a kweld, I plan to make a 24S4P pack of A123 26650 LiFePO4s and draw 96A peak from the pack with a Phaserunner controller. If the copper can handle 150A, then 96A will certainly not fail if it is built right.
 
Last edited:
0.1mm = 0.004” = 38 gauge.
And 110 copper is 99.6% pure

Do I have this right????

And we know copper is 4x more conducive than
Pure nickel.....

I hope I’ve got this right. Lol.

Where are some great copper suppliers?
Edit: just saw your above post Ron with ref to eBay copper. :oops:

And what’s the best thickness for the Kwelder?


🙏🏻


I found this handy chart. Maybe useful. 😊

https://metalwerx.com/gauge-inch-mm-conversion-chart
 
First-off how do you decide whether to have the cells as close/tight together as possible as in this photo or instead decide to have them arranged in rank and file for a specific use ...
spinningmagnets said:
Is that a 10S5P you can split into two 5S5P packs for paralleling together for balance charging with a 2S-6S balance charger :thumb: ... what size cells did you decide to use ?

Good to see you've still got your thinking cap and posting another informative thread :thumb:
 
eMark said:
First-off how do you decide whether to have the cells as close/tight together as possible as in this photo or instead decide to have them arranged in rank and file for a specific use ...
They were tight to allow unfolding.
DSC_3766_1.JPG
 
It's encouraging !
I'm wondering if it's possible to weld 0.2mm copper (or 2*0.1mm) with this method and the kWeld
If someone is willing to do the test, I'm interested
 
Copper to copper? Probably not, too conductive.

8mm x 0.2mm nickel strip (four wide) to 0.8128mm and 2mm plate copper is what I am doing. Then nickel to nickel (cell).
 
hallkbrdz said:
Copper to copper? Probably not, too conductive.

8mm x 0.2mm nickel strip (four wide) to 0.8128mm and 2mm plate copper is what I am doing. Then nickel to nickel (cell).

0.2mm copper strip with a 0.1mm or 0.15mm nickel strip
 
serious_sam said:
They were tight to allow unfolding.
https://endless-sphere.com/forums/download/file.php?id=278999"
About as tight together as you'd want them to be ... very professional :thumb:
Having the cells touching each other should be avoided ... when possible
 
Do we know how this works? Thinking out loud:
Do we think that the resistance through the layer of nickel heats up the Nickel and thermal transfer melts the copper touching it to the underlying nickel covered steel of the cell? If so it may take more time to weld than other methods (risk).
The melting temperature of copper is 1085 C and the melting point of Nickel is 1455 C so this seems possible but I believe it's a different mechanism from that of typical spot welders. More similar to resistance brazing (I've seen it used on motor windings)
In the copper nickel sandwich system, the hottest point might as easily be between the point of the copper probe and the nickel as the nickel to the copper sheet.
Do we know how much of the heat comes from the steel? if the cell case is 0.3mm thick and the combined nickel and copper are as thick but lower resistance then far less current will travel through the cell which may mean less heat to the cell (good) or poorer bonding to the copper (bad) and may be highly dependent on the relative thicknesses of the nickel and copper. split stripps become more important possibly. Interesting stuff.
 
I’m in the process of building a 20s7p pack using Sony VTC5D 18650 cells. They’re rated at 25A, so 25x7 = 175A for the pack.
I chose to use .2mm copper sheet to minimize losses. Using kweld, I experimented with various welding options. I ended up punching slots in the copper sheet and using two strips of nickel on top, straddling the copper slots. This gave me the best welds. I settled on 95 joules. Some of the welds look a little burnt on top, but 85-90 joules wasn’t enough for consistent robust welds to the battery cap.
Since I needed the slots to get the best welds, I’m guessing it’s not just a thermal contact weld from the copper heating up. The weld current going through the battery cap must be doing the bulk of the cap weld.
Once I finish the build, I’ll post pics. The application is a 2nd pack for my dirtEbike. Primary pack is a purchased 20s10p that fits in the standard slot in the mid section. The pack I’m building goes in a case I 3D printed (PETG) to fit under the seat behind the mid section.
 
You could well be right or it could be the battery cap heating the copper from the other side if I understood you right.
I'd imagine the battery is 0.3mm steel with maybe 10 micron nickel plate. The 0.2mm copper would conduct the heat accross it's cross section without the slot punched and even with perfect electrical connection, 70% of the power would go through the copper instead of the steel as it has less resistance. If the steel doesn't melt the copper, the copper doesn't weld to it.
With slots, most of the current may go through a point location in the cell, so the copper may be heated from the steel of the battery and melt to it. It's being heated by a spot in the nickel and a small ish area of steel - may make enough of a difference.
I don't know, just thinking out loud - waiting for copper samples to arrive
 
ggHawk said:
Forgot to mention - the nickel strips I'm using are .15mm.
If you use nickel plated steel (instead of pure nickel), you get less heat conductivity, and more weld per joule.
 
Did some experiments tonight with this technique using a KWeld...

Welding 0.20mm nickel (8mm wide with a slit) to some scrap 0.52mm copper plate took 80J consistently. No issues whatsoever. Great bond.

Then I tried the 0.20mm nickel with 0.81mm copper plate and I couldn't quite get a strong weld. Last successful was at 150J, 158ms and 1649A. Then I tried 175J and got a timeout. That copper thickness will be for most of my bus bars on the pack (nickel copper sandwich).

After that I will also need to spot weld 0.2mm nickel strips to 2mm copper plate as well.

Need More Power! :D
 
Was this a nickel-copper-cell sandwich or nickel welded only to copper? I'm guessing that slitted strips will help a lot with welding to the cell.
 
From Rio Cole on Facebook. 0.10mm copper, 0.15 nickel-plated steel. kWeld setting 60 joules. Rio states that the steel ribbon welds easier with a lower setting compared to pure nickel ribbon. This makes sense because the higher resistance of steel would get hotter per the same amount of amps at the point of contact.

I am using 0.10mm copper sheet and the nickel plated steel is 0.15 thickness. I am using 60J with this material. I think a wide range of J would likely work fine. I strongly believe from my short bit of experience doing this that the sharp pencil like shape of the electrodes is the key to being able to have full confidence in each weld; every 20 - maybe a person could get away with 30 - or so welds, the tip-shape should be addressed. Glad to share. Best, Rio

BatteryWeldCopper1.jpg
 
spinningmagnets said:
From Rio Cole on Facebook. 0.10mm copper, 0.15 nickel-plated steel. kWeld setting 60 joules. Rio states that the steel ribbon welds easier with a lower setting compared to pure nickel ribbon. This makes sense because the higher resistance of steel would get hotter per the same amount of amps at the point of contact.

I am using 0.10mm copper sheet and the nickel plated steel is 0.15 thickness. I am using 60J with this material. I think a wide range of J would likely work fine. I strongly believe from my short bit of experience doing this that the sharp pencil like shape of the electrodes is the key to being able to have full confidence in each weld; every 20 - maybe a person could get away with 30 - or so welds, the tip-shape should be addressed. Glad to share. Best, Rio

BatteryWeldCopper1.jpg

I have been using this technique myself since I read your post on it. I have managed to get extremely strong welds on 30J with 0.10 copper and fully split 0.10 nickel plated steel. The strength of the weld is not only incredibly strong but the structural integrity of the strip itself is also massively increased.

I have tested nickel, split nickel and thicker nickel-plated steel but the combination that yielded the highest strength with the least amount of energy was 0.10 nickel plated steel with 0.10mm copper.
IMG_4392.JPG
 
Back
Top