"Copper/nickel sandwich" buses for series connections

[spinningmagnets]

Yes, split copper with "infinite slot" has been done and it helps. More of the current of welding is forced down through the contact points, and less flows through the strip itself.

As a result, less current is needed to get a solid weld. That being said, many pack builders here are happy with 0.10mm thick copper, so the extra trouble of adding infinite slot was not necessary.

If going to thicker copper, it "might" be worth the effort. The gap can be tiny, and you could use a dot of super glue to connect the nickel-plated steel to the copper, to hold then in alignment for the final weld.

There are some "high amp" 21700 cells that could benefit from 0.15 copper on the series connections, but now we are getting into motorcycle territory.

 

[Voltspa]

Thanks spinningmagnets. Since this is a 20s12p battery made with cells that claim a 20A max continuous discharge and I'll be using a 5kw hub motor, AND I am a noob, I just want to feel comfortable with being able to take liberties with my speed and acceleration. If I understand correctly, this pack will be able to put out 240A if required, right? The intent is not to race, but to be assured that I wont ever overheat my battery. I don't mind overbuilding it a little if I can achieve that. Do you think 0.15mm copper would be fine?

Thanks again and I should start my own thread at this point I guess... here it is.

https://endless-sphere.com/forums/viewtopic.php?f=12&t=114924

 

[Amperry417]

Yes, split copper with "infinite slot" has been done and it helps. More of the current of welding is forced down through the contact points, and less flows through the strip itself.

As a result, less current is needed to get a solid weld. That being said, many pack builders here are happy with 0.10mm thick copper, so the extra trouble of adding infinite slot was not necessary.

If going to thicker copper, it "might" be worth the effort. The gap can be tiny, and you could use a dot of super glue to connect the nickel-plated steel to the copper, to hold then in alignment for the final weld.

There are some "high amp" 21700 cells that could benefit from 0.15 copper on the series connections, but now we are getting into motorcycle territory.

Thank you (and the previous user's post) for pointing out the infinite slot method again. I've seen pictures of one battery that used copper plates (for parallel and series), such as mine, but the plates were cut into multiple pieces and then put back together with the pieces being slightly separated... Creating an infinite slot at each cell tab while still keeping the overall "plate" design.
I agree that many, including myself, are having good results with .1 copper with no slot and nickel or plated steel as a welding tab aid. My welds are more than acceptable with this method however the thing that intrigues me about the slot in the copper itself, is the lower power needed to get the same result. Lower power means less heat to the cells, so even if the welds are great without a slot... We still may be able to achieve that same result with less power/heat being introduced to the cells.
Currently Malectrics V4 at 29ms weld time with .1 copper (no infinite slot) and .1 nickel plated steel tabs may be able to be reduced down to say 24-25ms if the copper was slotted without negatively affecting weld quality. Will definitely be doing some tests using this method. Unfortunately I have my plates cut already, so doing this may end up throwing off my plate shape if cut into pieces and then spaced, even slightly. Hmm. Excellent food for thought

 

[Pete's moto]

Hi all.
I've tried welding 0.2mm nickel and 0.2mm copper as a sandwich to the batteries using a kweld at different setting but doesn't do a strong enough weld.
Could this be a solution?
Welding the nickel strip to the copper first seems to make a strong weld and after the nickel to the batteries.
Any thoughts would be appreciated as I am building my very firs battery pack.

 

[Pete's moto]

Here is a picture of the sandwich welded to some batteries

 

[spinningmagnets]

Any length of the current pathway that you replace with copper is a benefit. It looks like you've replaced about 90% of the path, which doesn't hurt. I do have a question, though. What is the series count, the parallel count along with the continuous and peak amps you hope to draw from the pack.

 

[Pete's moto]

Batteries: 21700 p42a
20s14p peak 350a

 

[Frank]

Out of curiosity, how many joules were you trying to weld at in the original test?

 

[spinningmagnets]

If temporary peak amps are 25A per cell (*350A / 14P =), then...perhaps 0.15mm thick copper would be adequate?

If yes, you could weld directly to the cell and also reduce the number of welds.

Continuous amps might be 10A per cell during cruise? (*10A x 14P = 140A from pack)

 

[Frank]

If temporary peak amps are 25A per cell (*350A / 14P =), then...perhaps 0.15mm thick copper would be adequate?

If yes, you could weld directly to the cell and also reduce the number of welds.

Continuous amps might be 10A per cell during cruise? (*10A x 14P = 140A from pack)

I agree .15 mm would be adequate (assuming continuous sheet of copper for the series connections) at 350A peak.

 

[Pete's moto]

Out of curiosity, how many joules were you trying to weld at in the original test?

I tried it at 140J nickel copper to battery
Stuck but it popped off when prying it.

 

[eMark]

Soldering a Copper Sandwich seems to be frowned upon in comparison to spot-welding a Copper/Nickel Sandwich. Even if accomplished by a seasoned DIY builder that has the know-how, previous experience and soldering skill.

For someone that's budget conscious (salvaged cells), and an expert solderer they may not see any advantage to a spot-welded C/N Sandwich build. So am posting photos from a 2018 14S4P Copper Sandwich build using LG MF1 salvaged cells ... https://endless-sphere.com/forums/viewtopic.php?f=2&t=93576&start=75 ... Go to page 4 to see the following photos of what may be the very first DIY soldering Copper Sandwich build posted at ES ...

Perhaps the following was the very first Copper Sandwich DIY build thread at ES. Is this Copper Sandwich solder build deserving of meritorious approval as that of a Copper/Nickel Sandwich spot-welded build by that of an expert spot-welder building a C/N Sandwich battery pack using salvaged cells (pros/cons) ...

Got 3 LG-Chem hoverboard battery = 60 cells for 105 USD few months back, from alarmhookup.
Dissassembled everything. Had sand every terminal... and rewrap the LG MF1 cells (2150 mAh, 10A) because the old wraps were flaking off.... Then I assembled the new battery using some nice honeycomb type holder.
Got a 45 A BMS.

I dont have a spotwelder, so I plan to use my 140W soldering Iron....
Got some 0.005" (0.127 mm) 99.99% copper sheet. Will double thickness (0.25mm)... 10+ mm wide by 0.25 mm thick will be conductive enough... with the cut slits to melt solder through and aid rapid melted solder flow through holes/slits and thus good contact with cell terminals …

Finished DIY Soldered Copper Sandwich 14S4P Battery Build (439.5 Wh 52V 8.6Ah LG MF1 cells):

 

[spinningmagnets]

It appears that 0.15mm thick copper can be spot-welded using the resistance of a steel/nickel cap to generate enough heat to create a welded spot.

Many have found success using 0.10mm thick copper. The purpose of the nickel/steel cap is only to convert the welding current into heat at a concentrated spot.

When a cell is presented with a copper/nickel layer of ribbons for its series current, and both are at the same cross-section of 0.15mm thick...the fact that copper has four times the conductivity might suggest that 4/5ths of the current flows through the copper, and 1/5th of the current flows through the nickel.

In a 25A peak draw from a cell, this translates into 20A taking the copper path, and 5A flowing through the nickel.

However, in nature, energy appears to prefer to avoid resistors when an option is on the same path, and I suspect that almost no current flows through the nickel. Builders who use a tiny square cap over copper ribbon are not "missing out" on an extra current path.

Recent experiments show that steel ribbon has at least three-times a higher resistance than nickel, so using steel caps would reduce the amount of current needed through the welder to achieve satisfactory results.

Experiments have also shown that the thinner the steel, the less current is needed to achieve a good hot-spot. Therefore, I would suggest that 0.10mm steel ribbon as the current best welding cap option.

In recent history, some pack builders have attempted to buy 0.10 pure nickel ribbon, and they were sent nickel-plated steel.

On its face, this was a rip-off by a Chinese retailer who misrepresented their product. However, I now find myself in the odd position of wanting to buy 0.10mm nickel-plated steel ribbon, while fearing the retailer might ship me pure nickel. (*anybody have some to sell?)

The nickel plating is useful for resisting corrosion on the steel welding caps.

Nickel-plated copper ribbon is now becoming readily available for the series connections, and that appears to be the next needed experiment. I dont foresee any issues, but I want to be certain.

 

[john61ct]

round discs of 0.10 nickel plated steel

small enough to fit within the cell end positive center (negative on A123)

in conjunction with the protective insulating donuts to protect abrasion at the shoulders

with a 0.15 copper per-cell **strip**

someone with qood gear could do all the cells as a paid service


allowing for the copper strips to then be attached to their thicker copper buss material of choice

soldering for those who prefer, without risk of heating the cells

 

[Amperry417]

Battery is completed, BMS is live and all is well. Vsett 10+ scooter with dual Nucular 12F controllers. 50% charge with wheels off the ground, 59mph. Can't wait until the weather breaks so I can charge fully and ride. 30 minute test charge at 7a looked perfect.
Samsung 48X cells in 20s9p configuration for 43.2ah and 150a discharge. .1 copper sheets cut to custom shape and .1 nickel plated steel tabs. Main Bus Bars were L shaped pieces, soldered first, then spot welded to cell and finally folded on top of itself, secured the top fold with a few welds. Welder was Malectrics V4 on 12v marine battery at 30ms. Amazing performance from the Malectrics and welds are super solid. This was my first battery ever and couldn't have done it without the knowledge I gained from everyone on this forum.
Thank you for sharing the knowledge and hopefully the information I posted will help someone else.

 

[jonescg]

Not sure if this is of use to folks, but I have found this method to be really effective. Waterjet cut the copper busbars (0.9 mm thick) and solder nickel strips to them, such that the nickel strips can be pushed down onto the cell ends and spotwelded.


So far very reliable, especially when the whole surface is coated in filled epoxy and a layer of G10FR4.

 

[spinningmagnets]

G-10/FR-4 is a thermosetting industrial laminate consisting of a continuous filament glass cloth material with an epoxy resin binder. It has characteristics of high strength, excellent electrical properties and chemical resistance not only at room temperature but also under humid or moist conditions.

G-10/FR-4 glass epoxy laminate meets the specifications of Mil-I-24768/27
G-10/FR-4 also meet LP 509 & MIL P 18177 Type GEE
Maximum Continuous Operating Temperature: 285F / 140C
Commonly referred to as Micarta, Bakelite and more.

G-10/FR4 Sheets offer:
Dimensional Stability
Outstanding Insulating Properties
Moisture Resistant
Tremendous electrical properties
Meets Mil-I-24768/27 (GEE-F)

 

[john61ct]

So no copper touches the cells?

 

[Amperry417]

So no copper touches the cells?

Looks that way. Even though the distance from cell tip to copper, via nickel, is very short... I don't see how that method could be extremely beneficial. Also, solder from nickel to copper is an extremely poor conductor, so hopefully the nickel is making tons of contact with the copper before soldering. With the steel plated nickel tab and pure copper sheets method I think the benefit would be much greater and also much simpler. This method seems to be more work than the benefits involved. Good deisgn but amp carrying capacity may not be as much as one would think

 

[john61ct]

I think getting tinned copper tabs on the cells is the top priority.

Then the xPyS layout and buss power layout can be whatever the user needs for their use case.

Attaching the copper tabs to the power buss can be low tech

it's spotwelding the tabs to the cells that needs to be just right

And that way you can atomise the pack for cell level testing maintenance replacement does not need to damage the good cells.

Kind of half way to the solderless/no-weld solutions

 

[Frank]

@jonescg is pretty thorough, I'm wondering if the very small resistance in this scenario is "measurable" using hobby-grade VOM's?

 

[GalFisk]

I never could get the sandwich method working reliably with my capacitor-based spot welder and copper electrodes. I suspect that this welder is a bit anemic, because even at the highest pulse lengths and infinite slots in both the copper and the nickel-plated steel, the copper just didn't stick very well. Also if I held the probe wrong, I'd blow a hole in the nickel at higher energies. I'd used it earlier to weld 0.13mm slotless nickel-plated steel only, which worked very well.

However, I read somewhere else on the forum that copper could be welded directly with tungsten probes, so today I got a 1.6mm TIG welding stick of pure tungsten, and cut it into spot welding tips. I first tried cutting it with pliers, but it split - it's quite brittle. Working it slowly with a dremel cutting disc at full speed worked nicely. I made the tips flat, with just a small chamfer to get rid of the burrs from cutting.

This actually welded very nicely, especially with an infinite slot, using only 0.1mm copper. Getting rid of the sparks thrown by the steel was also nice.

 

[spinningmagnets]

sounds like you have used steel or copper, but not both at the same time.

If yes, perhaps try copper with infinite slot, and weld with the steel cap over the copper.

 

[GalFisk]

I used both at the same time; 0.13mm nickel-plated steel on top of 0.1mm copper to be precise. I tried infinite and non-infinite slots in only the copper, only the nickel, and both. I got good welds sometimes, but not consistently, on used cells with some nickel remnants. The best welds were often sticking to the top of the remnants. I have considered grinding the remnants off for the build itself, but if I can make the tungsten-copper method work reliably without doing so, it'll save a lot of work. 20s17p 18650 is the plan, and the cells will be loaded at <1C max, for a very long range (or long life, I'm not sure how my usage pattern will look yet) moped battery.

 

[GalFisk]

Here are a couple of rows of test welds on some junk cells using 1.6mm pure tungsten TIG stick probes. The offset welds on the lower row are just an artefact of sloppy cutting of the test ribbons. The upper row of welds is on the positive pole.

Edit: my inspiration for trying tungsten came from this post: https://endless-sphere.com/forums/viewtopic.php?f=14&t=84680&p=1584958&hilit=tungsten#p1584958

Edit 2: I tore off the strips, and made a few more passes. Most welds were ok, a few were great, and a few were duds. A few too many in fact. I'll experiment more with my technique, and maybe try some cell surface preparation as well. A bit of sanding or brushing. I don't need super strong welds, but I do need a bit better consistency.

Some observations:

• If the two copper strips happened to come in contact somewhere, they would spark at the contact point when welding, even if it was several centimeters away. A testament to the value of the infinite slot for copper.
• It's easier to identify the weldable area of both positive and negative poles when you can glean the position of the pole through the slot.
• The probe got a lot warmer than when using copper tips. Some pauses may be needed when welding lots.
• One of the tips eventually became copper coated and had to be sanded, because it started sticking to the ribbon. I don't know if this has to do with the tip, the polarity, or just random chance.