18650 DIY powerwall 48v

[Korishan]

At 60% SOH, or so recorded as such, I'd go ahead and do a capacity test on each module. You don't want to inadvertently create an imbalanced battery when you could have made sure beforehand.
Granted, it's possible it could actually have more life left in it than is recorded. You'll definitely not be abusing it no where near as hard as a vehicle does.

 

[OffGridInTheCity]

Totally jealous!

 

[Fukuokian]

I've received my 2nd nissan leaf battery pack (60% SOH) for free!! I'm currently working with opening the harness and the busbars.

I'll use this in the same pack as the 1st: 7 stacks connected together in series. I also got some more individual modules I'll add.
View attachment 34217

I'm considering capacity test every module or just charge them to the same voltage before making the 48v pack... What do you think?

Free is great! These are gen3 modules, so have you given thought to how you're going to wire them... They come in pairs, but the neg & pos alternate... like this:

The modules are glued together with epoxy and riveted/welded case. I definitely wouldn't recommend trying to separate them.

You can still connect them together in parallel stacks, but you'll need to customize the busbars.

Alternately, you could series connect the cells in stacks, and then put the stacks in parallel. (...I'm not sure how you'd use a BMS in this case)

 

[kje]

Free is great! These are gen3 modules, so have you given thought to how you're going to wire them... They come in pairs, but the neg & pos alternate... like this:

View attachment 34218

The modules are glued together with epoxy and riveted/welded case. I definitely wouldn't recommend trying to separate them.

You can still connect them together in parallel stacks, but you'll need to customize the busbars.

Alternately, you could series connect the cells in stacks, and then put the stacks in parallel. (...I'm not sure how you'd use a BMS in this case)

Good point. I haven't thought about that. I have to figure out how to customize the bus bars.

I could make a lot of wires with lugs like this:

Or I could try to separate them like this:

View: https://youtu.be/eTZkEXdGOtM?is=Lofd0Fr89W5UqHPF

 

[Korishan]

I would go the route of the guy in the video. It's not much work and heck of a lot cheaper. Lugs/wires are more expensive than bus bars in both time and money.
That would also allow you to parallel all the cells together, bring them up to full charge and then you can capacity discharge them 1 at a time. This would also save you time. The other thing about paralleling them makes them become balanced with each other and you can find a bad/weak cell a lot faster. After you charge them up, let them sit for a day or 2 to full absorb, then disconnect the busbar and let them sit for a week. Then do voltage/IR checks (do IR checks before connecting as well as IR can change based on SoC; this gives you a baseline number). If you have a regenerative charger you can just discharge into your other battery pack or after you've discharged 1 cell you can then recharge it from discharging and just work your way down the line that way.

 

[kje]

I would go the route of the guy in the video. It's not much work and heck of a lot cheaper. Lugs/wires are more expensive than bus bars in both time and money.

Yes I was most tempted to this method

That would also allow you to parallel all the cells together, bring them up to full charge and then you can capacity discharge them 1 at a time. This would also save you time. The other thing about paralleling them makes them become balanced with each other and you can find a bad/weak cell a lot faster.

How can I find a bad week cell faster?

After you charge them up, let them sit for a day or 2 to full absorb, then disconnect the busbar and let them sit for a week. Then do voltage/IR checks (do IR checks before connecting as well as IR can change based on SoC; this gives you a baseline number).

Yes this is a great way to do it

If you have a regenerative charger you can just discharge into your other battery pack or after you've discharged 1 cell you can then recharge it from discharging and just work your way down the line that way.

I'm not sure I wanna do all the work with discharge and cycle all the cells. Maybe I end up with just charge them, Let Them rest, checking Voltage and IR.

All the cells has 3.55 Voltage except to modules that seems different. You can see they have slightly different color. maybe they where changed at some point? All the modules has numbers except These to modules:

 

[Fukuokian]

I would go the route of the guy in the video

After watching the video, I agree that this is the way to go! So much easier when all the pos. terminals and all the neg. terminals are aligned.

A little armchair critique...

Be really careful separating the modules. Around the 2:50 mark of the video, he's already run his plastic chisel down the front part, then he switches to the back and hits downward (ok), but then switches and hits into the middle of the pouches. (that makes me nervous)

 

[Korishan]

I'm not sure I wanna do all the work with discharge and cycle all the cells. Maybe I end up with just charge them, Let Them rest, checking Voltage and IR.

With cells that are that old, they really should be capacity tested to verify their integrity. By doing a capacity test you can make sure that you can match cells so that each group has the same capacity. Which means you may end up moving a pack from the current group 3 down to group 12, for example. Capacity of each group needs to be absolutely as close as possible in all groups. Otherwise you make the BMS work for no reason and it would cover up a potential degrading cell (s)
Now if these were relatively new cells then I would agree, just slap them together and put the BMS on it and then find out overtime that way. But with them being with the age and the SOH that low I wouldn't go off of blind expectations that they are all equal.

Another option is to leave the cells so that you have one cell in each group. So basically a 14s1p configuration. And you would use 2 BMS because one battery pack would become two battery banks and you need a different BMS for each bank. Unless you use something like Batrium, then you just need enough -mons for all the cells

But I personally would go with capacity testing and go with 14s4p, or 8p. The more cells that are in parallel final battery the more stable voltage curve will be, especially under sudden heavy loads

You have to remember that you are not dealing with LFP you are dealing with NMC. They are nowhere near as forgiving and have a greater chance of becoming a fire hazard if misused. Just puncturing the cell and a possible internal short could lead to rapid expulsion of the cell out the window situation. So by doing the capacity test helps to reduce the risk of over or under charging the cell.