jonescg's cylindrical cell build thread

[jonescg]

I've got a LiPo build thread in the for sale section which folks are welcome to check out - It's a long running study in the best (or rather, least worst) way of assembling a LiPo pack with multiple cells in parallel.

I have recently started building battery packs using cylindrical cells because sometimes, they are the more appropriate option. Of course for eBikes you can just spotweld them together using hexagonal capture kits that click together and nickel strips, and the ones offered by KarT_PS are prety good: https://endless-sphere.com/forums/viewtopic.php?f=31&t=92223

But I am building big batteries fairly often, so I've started using capture plates and bus plates, and this will allow me to make 50-100 Ah packs with lots in series. So this is my latest effort:







Unfortunately my old work who bought a wirebonder closed up shop, so I can't get them wirebonded by robot. So I will resort to spot-welding. Nickel plated aluminium will take a spot weld, but in all fairness, soldered connections to the busplate are probably a bit better. The nickel strip would then be spot-welded to the cells below.



The goal is to build four of these for the Prelude conversion. They won't be liquid cooled - after the pouch cell pack was totalled by a coolant leak, I have decided this will have to deal with the heat. But it will be fully potted inside a polyurethane encapsulant which will make it waterproof and rugged.

I'll add more photos as I go along.

 

[DogDipstick]

yeah that coolant will rot em.

I take it your customer does not care about the cost being an issue?

I can never imagine a scenario where.. the.... "battery packs using cylindrical cells because sometimes, they are the more appropriate option."


Looks expensive, and with negligible gains, in my opinions.. but hey, yall know my take.

Why the small cell, when there are so many larger, LTO, Lifepo, and Lion cells in larger formats? Fr moving cars, the application, a few thousand cells vs a typical pouch cell EV. If I was tooling up, I guess 18650 would be an option.. but, they are overpriced, underpowered, and typically more dangerous when installed by someone who is not professional or explicitly trained.

I would love to see the whole IR of your (Cyl. cell) pack vs your ( the pouches and all that.. ) packs. I have friends building with 21700 that love them, good 4-6Kw power, traction packs, and do custom collector plates that are copper and wire bonded.

Boy oh boy, thats gonna be half a grand in epoxy. Keep up the good werk.

 

[jonescg]

The customer needs 24 kWh of battery in the smallest volume and lowest weight, at the best price.
In this instance, cylindrical cells wins. The cost of cells in this instance is AUD$8200, and the extra stuff to make it all fit will probably bring the final price up to about AUD$13,500. That's $540 per kWh, excluding BMS and other stuff, but that's markedly cheaper than LiPo pouch cells. The first lot of LiPo cells cost closer to $12,000 in cells, and another $5,000 in extra parts like copper, PCBs and enclosures. But critically, I can fit these four modules where the fuel tank was and it frees up all the space in the boot (trunk) and it weighs less.

The small cell is very configurable for the volume available. LFP wouldn't fit, and an LFP battery that would fit didn't have the capacity we needed.

Total DCIR of this battery will be 0.25 ohms. Not great, but it's not a particularly powerful motor either.

So yeah, your opinions aren't correct. Batteries are expensive, some are more expensive than others.

 

[Frank]

I'm interested in this switch Chris and have been moving in the direction of cylindrical cells myself lately. There seems to be many advantages.

 

[fechter]

That looks great!
What did you use to cut out the bus plates? Soldering might be a challenge. It will act like a giant heat sink.
The potting stuff makes it really resistant to all kinds of failures, but nearly impossible to repair. It does help with heat dissipation too, but if it's a low powered application that won't be an issue.

 

[spinningmagnets]

Jonescg, have you seen the threads using the copper/nickel sandwich method for series?

I dont know if it would be appropriate for this build, but I thought it was worth mentioning.

 

[st35326]

13s modules with various size P groups?

 

[jonescg]

24s23p modules. I can make anything of any size with enough planning.
Ron - Yeah been following that one. I think soldering nickel strips to the nickel plated aluminium will work OK, but for the next one copper might be the better option in terms of current capacity, corrosion resistance and the like. The issue is that the laser cutters don't have a fiber laser which would cut copper. Currently get them water jet cut and they don't like ding such thin material.
Frank - It's out of necessity really. I've been struggling to get reliable supplies of any cells out of China lately, and the cylindrical cells seem the least-hardest to access.

And yep, the potting will help absorb some sensible heat. Mainly doing it for waterproofing...

 

[neptronix]

This is looking good already :thumb:

 

[DogDipstick]

Jonescg, have you seen the threads using the copper/nickel sandwich method for series?

I dont know if it would be appropriate for this build, but I thought it was worth mentioning.

I have always wondered about the long term integrity of such, bare copper used anywhere in an assembly. I feel like the necessity to plate every bit of bare copper is imprtant. I have begun think about custom nickel plating even. How tough can it be? A few washes, a few vats, people have been doing little garage plating for years.

I can see how the aluminum can be cheap, and readily machined. Aluminum is forgiving, and, is very low on the scale. Like Zinc.
What metal is that in the pictures? Did you plate it yourself?

Why al the complicated zig-zags? ( .. and that one blank in the middle? whats that serve? Lol. )

 

[jonescg]

These are laser cut aluminium, but I will be getting them nickel plated. The process is electroless, and uses a zinc stage in between. I'll be paying someone else to do it...

The zig zag shapes are so that I could fit the maximum number of cells in the space available.

 

[DogDipstick]

These are laser cut aluminium, but I will be getting them nickel plated. The process is electroless, and uses a zinc stage in between. I'll be paying someone else to do it...

Yes. That was mentioned. Yes many ways to plate. I was wondering the grade, honestly.

ok. Still wondering about that blank hole in the middle.

 

[DogDipstick]

Lemme guess.. Fill hole?

Yah packing efficiency is what it is.

 

[spinningmagnets]

Beautiful work, thanks for posting.

 

[jonescg]

These are laser cut aluminium, but I will be getting them nickel plated. The process is electroless, and uses a zinc stage in between. I'll be paying someone else to do it...

Yes. That was mentioned. Yes many ways to plate. I was wondering the grade, honestly.

ok. Still wondering about that blank hole in the middle.

Sorry, I was replying on my phone and I couldn't see what had already been written. It's 5005 series aluminium which is apparently a decent conductor. In hindsight copper would have worked, and possibly easier to solder but the quality of the cut was the issue. Aluminium and it's propensity to corrode when near other metals always bugs me, but hopefully if the resin does it's job all will be well.

And yep, the blank hole will be used as a filling point, but the funny thing is I never planned for it to be there, it just happened to wind up right in th emiddle. It felt like playing checkers and colouring in dots until the board was full

 

[jonescg]

Dropped the busbars off at the nickel plater - I reckon there's about $1000 worth of work there. Things we do eh?

 

[jonescg]

Minor update - busplates are back from the nickel plater. These will take a spot weld quite nicely, but in the worst case, they can be soldered.

 

[jonescg]

Well my little Sunko spot welder is not up to the task of sticking the nickel to the aluminium busplates. I can zap it at full power 3 or 4 times and it still peels off with relative ease. Really don't want to have to solder them

 

[Hillhater]

Is it obviously the weld failing on the Nickel, or is the Nickel/zinc layer separating from the Al ?
I guess some late night reading of the spot welding threads may be needed ?

 

[jonescg]

It looks like there's no fusion between the aluminium and the nickel tab - and peeling it lifts the microns thick layer of nickel off. I do know it can be welded with enough energy, but the little jig isn't up to the task.

Here's an example of a Sunstone welding nickel tab to nickel plated aluminium:
[youtube]rI27bDJGexU[/youtube]

 

[drdrs]

Looks very nice.

Are the holder plates polycarbonate? How are they machined?

 

[jonescg]

Looks very nice.

Are the holder plates polycarbonate? How are they machined?

Yes, polycarbonate, and machined by normal CNC. The 4 modules (8 capture plates) involved about 20 hours of machine time.

 

[Frank]

Chris, have you tried "split" nickel, or nickel-plated steel? Or is this not possible with your configuration?

 

[jonescg]

Not as convenient in this case, but what might help is clamping the busplate and nickel strip in a pincer type welder and zapping them that way. Would result in a larger weld spot rather than 4 or 6 tiny spots.
Something like this:


At least just to attach the tabs, then the plates can be fixed to the module and the cells welded with my Sunkko.

 

[jonescg]

Well I tried one of these and it did absolutely nothing. Even put stainless steel on the other side and it did nothing. So my Junkko welder will have to do, and solder wherever it can't.