[kerrymann]

I have documenting my miata build in the "All Conversions and Builds" section and on my blog at:

http://electricmiata.blogspot.com/

I realized that my pack build up might be worth it's own post as a summary for others to build.



I went through at least a dozen design iterations and each time tried to simplify it and make it as DIY friendly as possible. I think the result is a pack that is about as light as you could possibly make it and you can still run max power without any additional life or performance hit. For my car I will only be driving in nice weather so cell cooling was my only concern. But it you wanted to put it in a insulated enclosure you could also heat the cells by circulating warm air.

The basic design end up being really just a DIY version of the A123 modules.

The cells are stacked together with cooling fin sandwiched between every other cells. They are restrained by end places made of .120" aluminum sheet and 1" aluminum angle. The assembly is clamped together by 1/4" steel all thread.

For the electrical connections the tabs were bent and sandwiched together with 3 cells connected in parallel.

I punch holes in the cells using a simple block I made that fit into a 3 hole punch.

The holes went very quickly but unfortunately removing the glue from the barcode stickers did not. You might have been able to leave it but adhesive is typically very non-conductive. To reduce the current density at the connections I interleaved the tabs.

The pack was a little harder to assemble but the result was three times as much surface area for the current to flow through. The hardware was just zinc plated 1/4 bolts, washers and locking nuts.

The BMS will connect to the extra threads on top of the nuts.

[Jordysport]

Interesting design, But how are you allowing for the swelling of the cells? and what BMS system are you using? this appears to be what rwaudio wanted to do with his pack.

Edit, just noticed your bolted just the tabs together are you not worried about the Tabs failing when you've tightened them up (small surface area the bolt is acting on)? Maybe have a sandwich block on either side to clamp it up, maybe 2/3mm thick copper on each side with the same two bolts.

What are the dimensions of the bolting face? as i am looking at making a 3P pack.

[rwaudio]

Originally Posted by Jordysport

Interesting design, But how are you allowing for the swelling of the cells? and what BMS system are you using? this appears to be what rwaudio wanted to do with his pack.

Edit, just noticed your bolted just the tabs together are you not worried about the Tabs failing when you've tightened them up (small surface area the bolt is acting on)? Maybe have a sandwich block on either side to clamp it up, maybe 2/3mm thick copper on each side with the same two bolts.

What are the dimensions of the bolting face? as i am looking at making a 3P pack.

That is very similar to what I proposed but with the addition of clamping blocks like you mentioned. This should work well and the trade off for not using the clamping blocks is probably small with regard to voltage drop and does simplify the required hardware. I would speculate at the performance of the pack as slightly better than any pack using a separator block. (IE my pack or Jack R's design)

Well done Kerrymann!

[Jordysport]

Originally Posted by rwaudio

That is very similar to what I proposed but with the addition of clamping blocks like you mentioned. This should work well and the trade off for not using the clamping blocks is probably small with regard to voltage drop and does simplify the required hardware. I would speculate at the performance of the pack as slightly better than any pack using a separator block. (IE my pack or Jack R's design)

Well done Kerrymann!

Less required hardware yes but the question of longevity on the tabs work hardening under vibrations etc combined with a small contact surface area from the clamping bolts.

you would only need 2-4mm of copper/Al sandwich plates just to make the N/mm^2 less and to achieve a higher clamping force with less risk of failure.

But those are minor details, Looks a great pack nice clean and tidy ready to drop in! will for sure be keeping an eye on your thread

[NikiL]

Great job. Nice and smart designed A123 20Ah battery pack. In that way or similar I would build it also.

[kerrymann]

Originally Posted by rwaudio

That is very similar to what I proposed but with the addition of clamping blocks like you mentioned. This should work well and the trade off for not using the clamping blocks is probably small with regard to voltage drop and does simplify the required hardware. I would speculate at the performance of the pack as slightly better than any pack using a separator block. (IE my pack or Jack R's design)

Well done Kerrymann!

Originally Posted by Jordysport

Less required hardware yes but the question of longevity on the tabs work hardening under vibrations etc combined with a small contact surface area from the clamping bolts.

you would only need 2-4mm of copper/Al sandwich plates just to make the N/mm^2 less and to achieve a higher clamping force with less risk of failure.

But those are minor details, Looks a great pack nice clean and tidy ready to drop in! will for sure be keeping an eye on your thread

Thanks to you both. I went back and forth on using plates and seperators even building a few cells up. I originally made Al plates to increase the surface area but I found washers gave almost the same area and when by interleaving the tabs the area was much greater:

Plates

My plates were 6mm thick and 45mm x 20mm giving 900mm^2. Neglecting current flow through the bolt area (12.5mm^2 *2 bolts) that gives you 875mm^2. Max current you can expect from these cells is 1200amps giving 1.37A/mm^2 which is an acceptable current density

Washers

The washers are 20mm OD and ~6mm ID giving each a area of 314mm^2-12.5mm^2. So each washer has ~300mm^2 of area so for two washers it is 600mm^2 total. This gives a current density of ~2A/mm^2 which is still acceptable but getting marginal. Now all this is assuming the current is just flowing through one face. So I interleaved the cell tabs so the current could flow through 5 faces. This gives me a surface area of 1500mm^2 and dropping the current density to a low 0.8A/mm^2

The clamping force is going to be constant based on the diameter and grade of the bolts. A higher clamping pressure is actually a good thing because it makes a better electrical connection. I am really not worried about fatigue in these tabs. Copper and aluminum are the most susceptible materials for fatigue but there is no really load on these tabs, road vibrations The shaker table we have the lab at my work is the size of a semi so it will be to sneak in some vibration testing. It's not shown but there is a lexan cover that protects the terminals and you could put a bead of silicone down between the cover and the top of the tabs. It would restrain it but still peel off when you wanted to service the cells. In my opinion it isn't necessary but only time will tell.

I am actually more worries about any potential internal damage to the cells from making the packs in the first place as some of these cells were removed and installed 3-4 times...

I have shot video step by step for building the pack. If there are people that have cells and are really interested I can try to make up a tutorial and throw it on YouTube.

[kerrymann]

I missed your question on the BMS. I am using a Lithiumate Lite system.

[danh818]

Great job Kerry! I think if you posted the video we would all appreciate it.

Any worries about all that conductive metal surrounding the cells?

[danh818]

Also, what is supporting the bottom of your module? I have seen the picture on you build thread of the module in the stock fuel tank position; it seems to be hanging. What is keeping the cells from slipping out the bottom other than the tension of the cables?

[Jordysport]

Originally Posted by kerrymann

I missed your question on the BMS. I am using a Lithiumate Lite system.

Thanks, looks a good setup, might consider it myself. just a few questions on it sorry to be a pain. I assume you have bottom balanced the 3 cells in parallel before assembling, then you can use the BMS to balance them in series? Do you have any more pics of the BMS system as this is one area i have no clue about atm (sorted controller, cells, motors, etc etc)

Thanks Again
Jordy