Current Sharing in Pack Design

[Punx0r]

Can I have a sanity check please on a proposed 4P9S 18650 pack layout? Grey strips are parallel nickel strip, orange lines are interconnecting wire on the top of the pack, purple lines are the same thing but on the bottom of the pack. I'm looking for the simplest/neatest solution that ensures proper current share across the 4P cells.



No. 1 looks a bit complicated. No. 2 might be O.K. No.3 is the simplest, but I initially suspect it wouldn't share current as the connections are at the ends of the parallel cell group. However, thinking about it made me wonder if it isn't, because the connections are at opposite ends, so the pathway through each cell should be the same length, as per the different coloured lines in the sketch below?



Have I lost it from staring at the screen too long? Have I missed an obvious alternative solution?

Thanks

 

[Syonyk]

Three is how basically every pack that's parallel groups connected in series is wired.

Do that.

Any minor imbalance at high current draws will balance out between the cells in a parallel group when the current drops.

 

[Drunkskunk]

#3. All cells in a parallel group will share their current just fine. They have to. The instant one gets a fraction of a picovolt out of spec from the other three, the other 3 will feed current over to that cell.

Besides, #1 and #2 are nightmare fuel for an engineer. All those ground loops!
Yeah, it's doubtful those would ever cause a problem with such small loops, but it's cringe-worthy.

 

[Punx0r]

Thanks for the feedback. I had current sharing ringing in my ears, I think from that J.B. Strauss lecture Dnmun used to link to. I think if I ensure the parallel connections are beefy enough there won't be any cell getting hit particularly hard during discharge peaks.

I'm glad you guys said No. 3 is OK because the other two are a bit of a mess, even without balance taps

 

[markisses]

I posted something similar awhile ago. Maybe there's something there that would be helpful to your design.

https://endless-sphere.com/forums/viewtopic.php?f=14&t=75286

 

[Punx0r]

Thanks for that, the quote below was helpful:

Yea, if you have series conductor alternating on the end of Parallel string the conductor length/added resistance is equal for every cell. BTW overall conductor resistance, using same materials, will be lower in zigzag pattern.

 

[999zip999]

If using bms I also use a sense wire plug so to make an easy check of parallel strings with cellog.