Building cylindrical battery pack for Dummy!

[GGoodrum]

I get a lot of a123 pack-building questions, and lately I've been referring people to this thread, which describes the easiest way I know to build a123 packs quickly. It always takes me awhile to find it, which is annoying. Mitch suggested I make this thread a "sticky", which makes wonder why I didn't do it earlier.

[BMI]

Wow, this is great. I have been playing around with trying to find simple "bolt together " solutions for building a123-based packs for, well, for years. This is about as simple as it comes, great job.

You can't get a much simpler end plate solution than machined nylon end plates held together with threaded rods.

DSC00017.jpg (38.87 KiB) Viewed 993 times

DSC00016-1.jpg (40.14 KiB) Viewed 993 times

[GGoodrum]

But how are the connections handled? Are these assuming spot-welded tabs are used? If so, these nice-looking end plates won't work for the cases where loose cells, without tabs are used. The technique described in this thread will work with loose cells just fine.

-- Gary

[MitchJi]

Hi,

Wow, this is great. I have been playing around with trying to find simple "bolt together " solutions for building a123-based packs for, well, for years. This is about as simple as it comes, great job.

You can't get a much simpler end plate solution than machined nylon end plates held together with threaded rods.

Your machined nylon end plates look very good! I'd like to see some pictures of the inner side of the plates. How are the cells held in place? Are there recesses for the ends of the cells?

But how are the connections handled?

I'd also like to see some pictures that show how the cells are connected.

Thanks!

Mitch

[katou]

Leamcorp's idea is valuable because of it's simplicity, ease of construction, and low cost.

If your idea with nylon plates works as well, I'm all ears. How did you make them? Pix pix!

One thing I'm not sure about with Leamcorp's construction method, if the clips are not soldered on, can high amperage (say 100a) be passed by the small contact spring bumps?

I'll find out the hard way I'm sure, but surely someone has a sensitive enough ohmmeter to test the resistance between the copper bar and the clips?

Katou

[BMI]

But how are the connections handled? Are these assuming spot-welded tabs are used? If so, these nice-looking end plates won't work for the cases where loose cells, without tabs are used. The technique described in this thread will work with loose cells just fine.
-- Gary

No, there is no welding involved. Link bars join the cells and they are held in place by pressure alone when the threaded rods are tightened.

A123 Frame Kit.jpg (41.18 KiB) Viewed 2290 times

[deVries]

No, there is no welding involved. Link bars join the cells and they are held in place by pressure alone when the threaded rods are tightened.

Looks great.

Now, we just need an ES member discount...

ES discount pricing?

Why can't the other "white" 26650 cells fit and work too?

Thanks!

[MitchJi]

Hi,

One thing I'm not sure about with Leamcorp's construction method, if the clips are not soldered on, can high amperage (say 100a) be passed by the small contact spring bumps?

100a per cell is over 40c so that would drain the cell in about 90 seconds. Unless you are planning a dragster its not an issue.

No, there is no welding involved. Link bars join the cells and they are held in place by pressure alone when the threaded rods are tightened.

Thanks for the pictures and information. Nice work! At first I wondered why you posted this on a DIY thread but then I realized it would be pretty easy to do a DIY version. Off the top of my head I came up with three methods:

Drill holes for the cells part way through the nylon (wood would work also), then drill out the portion between the cells using a smaller bit for the metal strips.

Use two pieces (an inner piece and an outer piece) for the top and bottom. Drill the holes for the cells and cut the slots for the metal strips in the inner piece.

Use two pieces (an inner piece and an outer piece) for the top and bottom and drill the holes for the cells but don't bother cutting slots for the metal strips. Just sandwich the strips between the inner piece and the outer piece of Nylon or Wood.

Leamcorp's method is probably a little easier but preparing Nylon or Wood in a similar way to the BMI method would not be that hard. Which method is best might depend on the desired size and shape of the final pack.

I had eliminated 18650 cells from consideration due to the extra work involved in assembling the larger number of required cells but both of these methods are easy enough that I don't think thats any longer an issue.

Two simple and elegant DIY solutions.

Thank You Leamcorp and BMI!

Mitch

[999zip999]

Can you use a 48v ping charger on the A123 pack ?

[katou]

This method (BMI nylon plates) is already DIY. If you notice the large # of holes in the plate?

Those holes are from the pilot point of a forstner or similar type bit.

The holes might be the other way too. If you drill all of those little holes accurately spaced (template or something) then you can just put the pilot point of the forstner in the little hole and know that the big hole will come out the right spot. Set the depth gauge and off you go.

Still have to saw out the nylon plates and then finish the edges I guess. Probably be about 10 min each in batches of 20 or so.

Did I get it right BMI?

Katou

[cell_man]

Those plates are not made or supplied by BMI, they are something that an Australian fella makes and supplies AFAIK.

Looking at the plates, I'd say they couldn't be made so easily without at least a milling machine. Now I'm no whiz on a milling machine, but I'd say there is a fair amount of work involved in machining those plates and terminals to the required tolerances. IME soldering A123 26650s isn't so difficult with the right tools, a bit of practice, the right technique and a bit of soldering skills. Now I reckon a thin sheet of copper (I used 0.2mm, same as used in the Killacycle spot welded pack AFAIK) is better than a big thick lump of copper. With the latter you need to put alot of heat energy into the copper to get a good join and that then bleeds back into the cells over a prolonged period. With the thin sheet, much less energy is put into the joint and as soon as the iron is removed, the thin copper sheet very quickly dissipates the heat and pulls heat away from the cell. I also use a pair of pliers pushed against the joint to ensure there is minimal gap between copper and cell, and this further helps to draw heat away from the cell once the iron is removed.

[katou]

I'm almost certain that it could be done without a milling machine. I say "almost" because I haven't tried to make one. I don't want to be all absolute unless I can actually prove it.

If anyone has the desire to make a set of these, and wants help though, I'm as sure I could help them be successful - PM me anytime.

I'm actually working on a battery pack using Leamcorp's ideas right now. I'll post pictures when it's nearer completion.

Katou

[veloman]

I found some really cheap brass strips on amazon. $7 shipped for 12 feet of 1/2"x .016" thick.

http://www.amazon.com/Brass-Strip-ASTM- ... B00137WIJU

My plan is to just stamp out tabs in the strip - should act like a spring to keep tension on the cell for proper connection.

This should be a LOT cheaper than buying those strips and separate tabs listed by the OP.

I have some small cardboard boxes from shipments, I might try fitting my cells in, then wrapping it all real tight with duct tape.

[MitchJi]

Hi,

Looking at the plates, I'd say they couldn't be made so easily without at least a milling machine.

I'm almost certain that it could be done without a milling machine. I say "almost" because I haven't tried to make one. I don't want to be all absolute unless I can actually prove it.

You are correct. Using a drill press with a depth gauge would easily do it. But using two layers for each plate would be even easier (a solid end piece and a piece with holes or slots that gets fastened to [glued or bolted] to the piece with holes).

Also with two pieces, rather than a series of round holes, it would be much easier to drill a hole at each end and connect the holes (slots with round ends) or even rectangular slots that are the correct length. Its easier to work with cells that are hot glued together into strips anyway so putting a strip of cells into a slot would be easier than putting a strip of cells into a series of holes or putting individual cells into a series of holes.

My plan is to just stamp out tabs in the strip - should act like a spring to keep tension on the cell for proper connection. This should be a LOT cheaper than buying those strips and separate tabs listed by the OP.

I have some small cardboard boxes from shipments, I might try fitting my cells in, then wrapping it all real tight with duct tape.

The cost of the springs is trivial and springs will retain their spring much better than brass strips.

If you are planning to use duct tape on the brass strips I think it's conductive. You might want to use gaffers tape.

[The Mighty Volt]

Hi there............does anyone know where I can get the boards to make a BMS to feed and charge my cylindrical pack?

Thanks.

[PedalingBiped]

I'm coming out before I'd plan to: I going to make a new thread in a couple of weeks when I'm done.
I am building a pack this way but I have some issues I had to overcome.
The clips are $25/hundred. So 15p12s = 180 cells # 2 = 360 clips or 400. $100
The span for 15p was too much so I added rubber bands to keep the straps together. Didn't work. Some cells didn't drain on my first ride. Then I found cable ties at "big lots". 100pcs 11" for $4. Priced at big box hardware $8 - $12.

Here's the latest plan. Made a 'U' shape jig. The uprights are only 5". Lay cells down, screw the straps with clips into the risers. Feed cable straps every 3 cells are so and pull. Straps don't twist (like they do when "free handed" and you get a solid "flat"
Your idea with "dimple" is perfect!!!!! Some of my clips are flattened from too much pressure. "dimples' would be the best solution. Lets KISS.

I started with screwing the straps into a wood box. Each one individually fitted. Then I hooked short pieces of wire to the screws to make the pack in series.

No more. I got a new interconnection way that I'll post when I'm done.

The beauty of this system is that once all the hard work is over, it is done! New technology comes out with new cells, just swap out.

With Konions, they self balance so no BMS.

edit: just bought brass strips for $1.60 from the Amazon link above. My copper staps were about $6 and came in 6 ft lengths that I had to cut to 12 inches. The brass are already 12inch length.

[katou]

I'm totally interested to see your new system. I am working on the original idea so far.

My price for the clips from mouser was not 25/100, it was far less than that. I think my cost was more like 10$/100 shipped.

I wonder about the thickness of the metal you are using. If you are using the Amazon stuff from the previous posting, it was 0.016" thick right?

I would not use less than 1/8" or 0.125" thick. With such thin metal, I figured it would not stay straight, the spring pressure would push it into a curve.

Did you find this to be a problem?

Katou

[MitchJi]

Hi,

I wonder about the thickness of the metal you are using. If you are using the Amazon stuff from the previous posting, it was 0.016" thick right?

I would not use less than 1/8" or 0.125" thick. With such thin metal, I figured it would not stay straight, the spring pressure would push it into a curve.

Katou

Just run a bolt or two through the strips (nylon bolts if that helps).

[PedalingBiped]

I bought the clips and copper from the links in the original thread.
You would have to have very thick straps to hold the clips in place with out help. I thought of plastic bolts to hold the straps together but couldn't find any locally (didn't search that hard)
Plus the plastic bolts would compete with the cells for space. Thus your "flat" would be longer.
I only have enough copper for 12s so I bought bronze from Amazon to see what its like and maybe go a few for more s!

I am not sorting the cells by capacity. Don't have time or resources. I am making up "flats" and riding around 5-7 miles then I break it all apart and sort cells by voltage. Then I make up new "flats" and do it again. Did this 3 times and it looks like the "flats" are staying balanced now. After I put this together more permanently, I am thinking about breaking the pack down in week then a month to test the cells and balancing.

I've found connectors in the Electrician section at the hardware store. Uses a set screw to hold the wire. (or grub screw)
I can't find one that's small enough so my final connection solution won't be as elegant I fear.

[katou]

Well, it sounds like you know what you are doing. Your sorting method is interesting because it uses the easy remove/replace feature of this construction as a way to homogenize the quality/capacity of the cells. Nifty! I look forward to seeing the pictures, should be a good read.

Katou

[Doctorbass]

I found a great illustration of the basic arrangement of battery in serie parallel pack

The Tesla Roadster Battery pack

Doc

[PedalingBiped]

I used 12 inch copper straps for every p. With 16 cells the straps are a little over.
So, 12 inches hold about 20 cells.
69 cells = 1 meter, at 1.5ah 18650
Since the cells are bigger they would be even longer.

[veloman]

The cost of the springs is trivial and springs will retain their spring much better than brass strips.

If you are planning to use duct tape on the brass strips I think it's conductive. You might want to use gaffers tape.

You're right, the brass has no 'spring' to it, won't work. I'm trying to find a cheaper source for these clips but I may just order them from mouser.

I'll look into that tape issue.

[PedalingBiped]

My next build, after I finish this one, I was think of using a chisel to cut an X in the brass and then punch or peen a dimple that would connect, at least with the negative side. Might have to use a clip on positive unless the dimple was good enough.
I use cable ties to pull everything together. $4 /100 at big lots. Usally $8 - $12 at hardware stores.

[methods]

I am seriously considering dumping my 2500+ emoli cells just for the fact that they are such a bugger to deal with. At least with prismatic cells I can lay out a PCB to hook them together. With these 26700's I will probably end up spending more time and money on building the pack than I will on the cells themselves.

I have 950 of them in a "temporary" configuration in my car but now that I am looking at doing the full pack of ~2500 the prospects are looking dim.

5000+ spot welds ? no thanks

Big racks?
for 34S 70P I would have to build 34 individual 70P racks - or 17 2S 70P etc
All of that sounds big and heavy and lame
Tons of interconnects, lots of expensive materials, a real production.

I guess I understand why people choose the "easy route" of 3C Thundermug cells for large scale EV's

I wonder how much I could get for my emoli cells if I sold them in "small bunches"
Like hell am I going to characterize them though... Some are new and some are old - but my labor is way to expensive to test them all.

You guys should buy them all from me

-methods