Spot Welding Copper Strips to 18650 Battery Cells

hallkbrdz said:
Or... this approach looks very robust and easily made to carry the 350A surge current I'm designing for (borrowed from spiningmagnet's ebike site :D )

Spot weld the strips to the buss bar, then attach the strips to the cells.

BatteryBusStyle1.png

Nice ! The shear mass of copper here looks like an absolute overkill though. The amount of copper itself looks like it could carry a 1000 amps!
 
Love the copper. This may be a dumb question but would not the current be limited by the resistance of the nickel which connects to the copper? It is true that the copper can carry tremendous amounts of current but it has to get to the Copper first. That’s where the resistance of the nickel would limit the current. Correct?
 
It will, but the current can flow in both directions from the Nickel to the copper, so the distance it travells through Nickel is much lower and the cross section of Nickel effectively doubled. Most of the distance is in Copper. The answer partly depends on what you mean by limited, efficiency will suffer before you hit maximum safe power. The power from the battery at the far end would have to go all the way from one end to the other in Nickel otherwise, heating the whole bus bar. That's some chunky copper so there will be minimal losses once the current is safely in the bus bar.
 
BobBob said:
It will, but the current can flow in both directions from the Nickel to the copper, so the distance it travells through Nickel is much lower and the cross section of Nickel effectively doubled.

Did not think of that. Interesting information. Really appreciate the response.
 
The current flow is not great through the per-cell connections, nor the parallel ones.

Only the series connections need to be very low resistance that's where the amps get high.
 
Generally yes, but we were talking about the 24P1S packs above with great chunks of copper going to tabs at the corner.
In this example, the only series connection is the parallel one :)
 
The point about the relative current between single cells and the rest remains.

That **here** is the only connection where the nickel layer is relevant.

But a more general statement is still useful.
 
Yeah, I agree, it's a neat solution to the copper / Nickel problem.
With such low parallel resistance I imagine cells would need to be new and from the same batch?
 
So I was googling the topic of welding copper battery tabs and this thread came up. Absolutely fascinating for me, as I am contemplating my 3rd EV build, which will have a Tesla Drive unit in a Factory Five 818 Kit car. The car has been done before, I'm looking to do a custom battery. Present thinking is 112S38P using low impedance 26650 LiFePO4 cells. The plan is to build single cell "blocks" of 38 cells each. These blocks will look similar to the pics posted above by Matador. The block will be a sandwich with 2 copper plates, and the 38 cells need to be welded to the copper plates. The plates will be roughly 4x11 inches with tabs on top for interconnections. I had found the Sunstone cell welder several years ago, but I considered it way too pricey, especially for a "manual" machine. After a lot of pondering, I came up with the idea of the plates with holes over the battery terminals, and then very short "jumpers" of either thinner copper or nickel, as pictured above. The issue is the Tesla drive will draw around 1100 amps peak (for only the few seconds it will take to hit any legal speed). If my math is right, to keep the overall losses of the 244 series-connected bus plates down to 6 volts or less at 1000 amps, I'll need to use at least .060 thick copper plate, I might go a bit thicker. Instead of round holes, I was thinking of squares, to accomodate the essentially square pattern of the weld points. Based on what I've seen discussed here, I should be able to weld .15mm thick pure nickel to the bottom of the copper plates with either a MOT welder of one of the 4.3kW Sunkko units like a 737G+ or a 737DH. With an aperture of around .4" x .4", I will have current flow in all 4 directions, keeping the losses in the nickel "jumper squares" very low. I guess the good news is that the tough weld will likely be welding the nickel to the copper plates. But I won't have to worry about damaging any cells. Then lay the plate over the cells and weld through the .4" x .4" hole in the copper. My question is this: What equipment would anyone recommend for the welding? The Sunkko 1200 amp units look like they should handle the nickel to cell welding, but I'm seeing references that seem to indicate more than 1200 amps may be needed to weld the nickel to the copper plate. I was thinking of using parallel nickel strips to form the "infinite slot" to reduce the current needed to weld. The kWeld unit seems like a great design with closed loop control of energy, but to do the large number of welds in this project I'd probably need a full size power source, not a battery. I'd also want a stable and adjustable voltage for weld consistency. I'm also concerned that the current limit on it is 2000 amps, I'm not sure how much over 1200 amps I could go without risking hitting the 2000 amp limit. Using batteries as a power source makes the peak current limited only by the series resistances in the weld path.
Any thought or suggestions welcome!
 
TL:DR

Learn to use hard return for white space.

And write more concisely, ideal one comment/question per post.
 
67carguy said:
The kWeld unit seems like a great design with closed loop control of energy, but to do the large number of welds in this project I'd probably need a full size power source, not a battery. I'd also want a stable and adjustable voltage for weld consistency.
I believe that there is an ultra-capacitor array available for the kweld, which provides the instantaneous power required for welding, and is continually charged by a mains powered charger.
 
i can recommend this
https://www.yousun.org/product/spot-welding-72xmosfet-switch
here's mine-
https://endless-sphere.com/forums/viewtopic.php?f=14&t=100435
its not expensive and will do the job you want to do id say. i power mine with 3s 80p liion mongrel battery. it can weld copper 0.15mm with nickel 0.15mm to cells without breaking a sweat.


i think it will take way more power than kweld
 
Relevant background on where I'm coming from: I have only used large cells with post or bolt terminals, connected by wires with crimped ring terminals, each battery terminal only connected to one other terminal, same wire gauge for serial vs parallel.

So I really have a very hard time visualising the current flow through these connection "grids and plates" layouts, where the same terminals are "communicating" with multiple other terminals, both serial and parallel at the same time.

What about this layout / design for example 10P14S:

14 separate sub-pack batteries

each a separate "box" using this welded grid/plate technology internally only one big parallel group of 10 cells each to get to the 10P

each terminated from the grid/plate internally, to a [-/+] pair of threaded posts to the outside.

These "group boxes" then wired in series to 14S

using insulated stranded heavy copper wire, say #2 gauge, to carry 100A bursts, 50A continuous.

Now, my puzzled question

are the "intra parallel" connections **within** each of those 14 separate sub-pack batteries

truly low current flow?

Take that 50A continuous "through" the 15 serial wires. Given the 10P group of say 2000mAh cells, each cell-terminal-to-strip junction is getting an evenly divided 2.5C rate (5 amps) current flow?

So that is what makes using only 0.15 nickel strips OK? really OK?
 
Learn to use hard return for white space.

And write more concisely, ideal one comment/question per post.

Sorry about that, I didn't realize how big my first post became! Just wanted to put my question in perspective...
 
are the "intra parallel" connections **within** each of those 14 separate sub-pack batteries

truly low current flow?

I've been confused by other discussions elsewhere on this topic about series vs parallel currents. I think it's all relative.

Looking at the photo of the plate, you could assume that the plate distributes the current evenly among the cells, or the cells all feed the plate in parallel. However, I'm looking at the total voltage "sag" of the pack. I'm looking at cells that have a DC resistance of only around 7-8 milliohms, and I'm putting 38 of them in parallel to make that number as low as possible. The drive will draw 1200 amps, so at that current, there will be a voltage drop from the cells at the bottom of the copper plate compared to the cells at the top of the plate.

I approximated the number by assuming all the cells were connected on a line across the middle of the plate. So the 1200 amps gives me a 27mV drop, which doesn't sound bad. But I'll have 112 of these blocks in series, and each block has 2 plates, one on each end. So my .060 copper plates are giving me 2 x 112 x 27 = 6V! That's a good chunk of my voltage sag budget, without figuring the cell to plate resistances.
 
by ridethelightning » Sep 14 2020 6:56am

i can recommend this
https://www.yousun.org/product/spot-wel ... fet-switch
here's mine-
viewtopic.php?f=14&t=100435
its not expensive and will do the job you want to do id say. i power mine with 3s 80p liion mongrel battery. it can weld copper 0.15mm with nickel 0.15mm to cells without breaking a sweat.


Yup, that's a LOT of FETs and nice heavy busbars. They should carry the current AND dissipate whatever heat is generated. My issue is that I really wanted to find a reasonably priced welding solution already done. I was very impressed with the kWeld design. When you close the switch it starts calculating the energy delivered. I imagine the weld consistency is excellent. I see you used the control section from another welder: good idea, I think. I should contact kWeld. I wonder if I could use a remote FET bank with it. Or maybe I could parallel another bank of FET's right on the kWeld board. If it can drive them. It doesn't sound like I'd need more than the 2000 amps the board can do, It would be mostly for margin if I came close.

And the big cap bank with smaller adjustable power supply would get me an adjustable Voltage source for the kWeld.
 
Interesting, I didn’t realize they made a “manual” square punch. My assumption is that since I need 112x2=224 plates it will make sense to go out to a local shop and have them made to order on a punch press or a laser cutter. Of course I may go into shock when I see the price....
 
I used the emachineshop.com CAD program and setup my main bus bars with square cutouts. For a quantity of 38 it would be $7.43 each with a $141.21 setup fee. The 24x24 copper sheet I bought was about $95 with shipping, which was enough for those plus 2 each of two other similar parts with some left over. For just the main 38 parts (1 setup fee) that makes it about 5x as expensive to have them laser cut. This seems about typical, but possibly someone local is a little less expensive.

SendCutSend was $8.28 ea with no setup fee for the main 38 parts using 1.8mm copper (not sure what type and thicker than I need).

I did find a less expensive punch and die, I may try that out to see how it works at only $32 plus some jig to hold it for either the hand or hydraulic press.

https://www.amazon.com/dp/B088FNPQW...olid=2JYQRE30POK8Q&psc=1&ref_=lv_ov_lig_dp_it
 
I'm on Long Island (NY) and we have a lot of local shops that would do this sort of work. With 224 plates I'm hoping for a quantity discount! It'll be a while before I have a final design that I can have quoted but I'll post them when I do. I ordered my kWelder tonight, and some nickel sheets. I'm going to try to get away with 0.1mm nickel spot welded in a sandwich to the 10mm x 10mm hole perforated copper sheet. I may go with maybe 2mm thick copper. I plan to order some sample sheets and try welding the nickel to copper plate.

Anyone have any recommendation on electrodes? Are the standard copper electrodes going to be a practical choice for doing 68,096 welds?
(112S cells x 38P cells x 2 ends x 2 cell welds ea x 4 nickel to copper welds ea)

I'm thinking tungsten may be harder and wear less, but may get too hot? Does the heat help me weld the Nickel to the heavy copper plate? Would they work well for the nickel-to-cell weld?
 
Back
Top