Lifepo4 cylinder 32650 cells simplified stationary design idea?

[orangezero]

I just registered here and read through 12 pages of people discussing copper/nickel spot weld sandwiches.

I am in the process of upgrading a 8s40p 32650 lifepo4 24v lifepo4 battery pack test run into something much larger. I have 1200-1400 new 32640 lifepo4 cells, some 0.15mm x 10mm nickel strips and a malectrics spot welder. I was considering 0.10mm or 0.15mm copper to connect the cells in series. This build is not for a vehicle but will be for a stationary off-grid solar system, currently using a 4000w inverter and lots of smaller DC loads. The draw on these cells, per cell, should end up being quite low. The max my solar could theoretically charge is 120a at 30v, but I only have 5.5kw (3kw in use) of solar right now. I would see using maybe 200-300amps at some point in the future if I upgraded my inverter.

My super simple plan was to have a rectangle/square of cells, negative on bottom, stacked in series above. Eight layers of cells, 8s120p. It would basically be 120 cells parallel, with another 120 cells stacked on top, with another 120 cells stacked on top in a very boring fashion. I am not sure if 0.15mm x 10mm nickel strip spot welded to each cell is enough for parallel connections, which is why I was reading about copper strips. I've read over and over that this parallel connection is basically to keep each cell connected in parallel at the same voltage, and this wouldn't need much even as it is drawing 100a from the battery pack.

I found a similar setup with battery born 12v batteries (and a few on here today), and it looks as if the full nickel sheet of one positive end is literally just laying against the nickel sheet from another group of cells, which is then compressed. So, in my case, would gravity almost be enough to provide a connection, similar to putting multiple AA or D batteries in a remote? Oddly enough, I went with cylindrical cells so I would avoid having to compress them, but I could also see how running some insulated threaded rods top to bottom wouldn't be difficult and keep things more stable.

I did some mass charging of cells, 8 in series and didn't bother connecting them in parallel at all, and was amazed how well they all fit together. No heat increase at the interconnections of the cells based on a thermal camera. Long-term I'd definitely want parallel connections... just not sure on a proper/efficient way to do it. In some weird thought experiment spending a lot more on 1/4in copper plates to go between each group in series doesn't seem like it would have the benefits I was originally envisioning.

I could see the bottom negative and the top positive would likely extend an inch or two into a busbar of sorts, allowing other connections to my inverter/fuses/solar.

Also, with my suspected use, it still seems I should be putting the cells pretty close together rather than leaving more space between them for airflow/temperature control.

And if anyone read this far down, feel free to throw out comments about arranging my cells 10x12 per layer vs something more rectangular like 6x20 or 3x40. Lots of powerwall diy I see they do 4x40 and hang them on a wall. I realize this will be 250-300lbs of just cells, and my concrete floor seems the best place to take that weight.

Final thought is if I need to purchase the specific positive end paper covers, or if I could laser cut some kaptom or other material for a similar result.

Appreciate any thoughts.

 

[orangezero]

Further reading and thoughts led me to this...

Copper between layers of battery would work, but is expensive and would need nickel plating. I don't think thin 0.15mm copper mesh would work well for the two end series connections, and I ultimately wonder about the cell to cell connections with mesh. Expanded copper would be nice, but is super expensive. Nickel strips of 0.15mm nickel would be a lot of work and need to be ultimately soldered together (solder also adding some slight resistance). And any experimentation down that line ruins the flat surface of my new cells.

Aluminum is a lot more electrically conductive than nickel, and is a LOT more affordable in the sizes I would need for each layer. It would need to be around 2x as thick to equal copper's conductivity, but also weighs a bit less than copper for equal ampacity.

So, I'm thinking of getting some 19 gauge, 0.912mm thick perforated aluminum baking sheets. They are way less expensive than anything else I've looked at. I can get a full baking sheet (18in x 26in) for around $8. Half sheets are too small and I can't find 2/3rd sheets that are perforated. I believe the parallel connections would be more than sufficient. It looks like I'll be able to get a 10x16 rectangle of 32650 lifepo4 cells (160) on each layer. They would be around 13in x 20.5in.

From all my reading, most recommend laying the cells flat for a bike or ev. I feel in a stationary setup this maybe is less of an issue, but never read the exact reason for wanting them flat. They fit together quite nicely in rows and allow a bit of an air gap between the cells. Butting them up against four side walls should allow them to stay in alignment. Any heat produced would be able to rise in those air gaps and up through the perforated holes of each aluminum sheet. I could have fans blow air up from the bottom if needed.

My current setup is only 40 of these cells in parallel (premade, nickel strips connecting them with soldered wires and ring terminals connecting each pack in series). They can draw 100amps without much in the way of temperature rise (measured through the shrink wrap via a thermal camera). I'm not sure if it is a waste of time to even consider the perforations, but it only adds around $15 and I figure someday I may want to pull a lot more amps at once.

The full size baking sheet would be big enough to allow some variation in connecting the - and + at the two ends. It appears the best solution would be to have the negative bottom end on one side of the sheet, and then the positive end on the other side so that all the cells have approximately equal path lengths as they pass through the cells.

I am not sure about these two end connectors, but I assume I'd want to have a thicker piece of aluminum sandwiched to the baking sheet on the top and bottom. It would be somewhere around 0.912mm thick x 330mm wide so the cross sectional area would be around 300mm2, also reduced by the perforations. I thought I read aluminum would be around 80% as a rough guide, so all things considered it may work at low loads without any extra sandwiched aluminum.

Lastly, I believe each positive cell should have some dielectric grease applied to the positive end to reduce corrosion/oxidation over time. If I sandwiched a thicker aluminum piece on the ends of the pack I'd also apply a thin layer of dielectric between the baking sheet and the aluminum. I haven't yet read of conclusive reasons to add anything to the grease to increase conductivity.

In case anyone is reading this, I have acquired some 80mm x 80mm thick walled aluminum extrusion to keep it all from warping. I like the idea of poron or silicone to help act as compressive force to keep all the cells connected.

Maybe I should add, don't live near an ocean and this will be in an area that has steady dehumidification (need to use excess energy on something...)

 

[jonescg]

Aluminium is not a good metal to spot weld to steel. Like copper, it would need a less conductive metal such as nickel on top before hitting it with a pulse, but it's got the added problem if higher reactivity, and thus corrosion potential. The weight penalty of copper is really quite minor, especially considering the corrosion and poor contact risk.

 

[Bbbbrass]

First, welcome! So much great info here.

Question, are these the Headway cells with the threaded ends? Spot welding isn't nearly as good as the mechanical connection if that is an option.

I am fine using kapton tape. Fish paper is another option, or (for between cell walls) cheap dollar store flexi cutting mats. 3D printed holders are awesome, and depending on your location I would be happy to ship you some. Also, I agree the floor is best for the weight you are talking about but make sure to elevate it a bit both in case of any flooding (pipes do leak/break!), to make it easier to maintain, and to minimize setting items on the top.

 

[orangezero]

Thanks for the responses.

My goal with aluminum and vertical mounting/clamping was to avoid spot welding and soldering completely. I have the malectrics spot welder, just not convinced I want to use it. I was planning to have ring terminals on some wires, bolted onto holes in the aluminum plate. I've taken a liking to ending my big wires with ferrules, so I may try to fashion some type of ferrule holders with thicker aluminum and a set screw.

These are a bit smaller than headway. They are 32mm x 65mm. I purchased from batteryhookup. They are covered in black shrink plastic with a small chunk of it hanging over the edge of the flat negative end. I'll likely need to remove a bit of that plastic to be sure the negative is as flat as possible on each sheet. The positive end has a raised area similar to a 18650 but it is a 11mm diameter central flat part. There isn't much difference between the central raised area and the plastic around the outer edge, which really makes me lean towards using the thinner Kapton material to ensure a good connection from flat part to the next sheet.

I was reading about nickel plating aluminum last night, but it seems you have to zinc it first and clean with some acids and bases and that isn't the easiest or cheapest. I certainly have enough nickel strips to use as donor material. I am not sure if nickel plating would end up being superior to using just the aluminum and coating the contact areas with ox gard or similar. Or if it is still recommended to ox gard over the nickel plating? I already have some super fine graphite powder but haven't discovered any results of people improving connections with a mixture of dielectric and graphite, other than perhaps the concern to not get it everywhere and cause a short.

I actually found a semi-local person selling 12 of the aluminum perforated sheets I've been eyeing for a really decent price. That would leave me a few to bake some cookies or mess up a few times. One reason I like this idea is I'm not the one drilling all these holes like if I purchased solid baking sheets and needed to cut holes for compressing or air. As I'm typing this I'm guessing it would increase the rigidity and ampacity of the top and bottom sheet if I put two sheets back to back on the bottom, almost making a ring of I beams around edge. Not sufficient, but better. Not cutting up the sheets at all would be the easiest way to do it, but I may end up wanting it to take up less space.

I'm not really sure if there would be a downside to having holes in the sheet. It would lower the ampacity somewhat, but the cells should easily all stay the same voltage based on my use of them so far. It seems a bit silly to stress over all these perforations and then end up covering it with kapton tape most places, eh?

I'm sorry, this has to be pretty boring. I need to get some pictures of what my silly words are trying to describe. I was going to find some aluminum I already own and line up an 8s string of these tomeasure the resistance with and without aluminum in between.

 

[jonescg]

I was reading about nickel plating aluminum last night, but it seems you have to zinc it first and clean with some acids and bases and that isn't the easiest or cheapest. I certainly have enough nickel strips to use as donor material. I am not sure if nickel plating would end up being superior to using just the aluminum and coating the contact areas with ox gard or similar.

I did this with the busbars of the Prelude battery pack. Do not recommend. The nickel layer is very thin, and any welded tabs pop off with ease.

 

[orangezero]

I did this with the busbars of the Prelude battery pack. Do not recommend. The nickel layer is very thin, and any welded tabs pop off with ease.

Thanks for the info.

In a press-fit or compression mount, (still can't remember the correct wording for what I'm attempting) I feel the ultimate test is setting something up and watching a thermal camera for heat at higher loads, correct? Some setups could make that difficult to do if a lot of connections are hidden by other cells.

 

[jonescg]

Screwing or bolting the two materials together is perfectly fine, provided you have something to screw or bolt to. At some point you need to go from cell to busbar. Steel to copper (or ally). The change in material properties from one end to the other is stark.