What are the tradeoffs when building a modular battery pack

Question

I want to build a 120 v / 128 Ah battery for an EV conversion project (small car with short range). I plan to base this on 18650 size lithium-ion batteries (recovered from old laptop battery packs).

To enable distributing the batteries around the car, I would like to do this with a number of modules wired together.

My first idea was to have 34 modules (34 x 3.6 = 122.4 v) of 64 cells (assuming ~2Ah capacity, so 128 / 2 = 64) wired in series.

I could also do this differently with fewer modules of higher voltage.

What are the considerations here and what are the benefits/drawbacks of choosing modules of different voltage/capacity? If the modules are spread around the vehicle, would my initial idea of 34 modules in 64p configuration lead to more resistive losses?

Answer 1

Your ageing of the weakest cell will be accelerated by the % of mismatched. They ought to be <<1% matched which is unlikely with used LiPo cells. I suggest you use active flying inductor full bridge switches to bypass cells that are over or under-charged, rather than the typical passive Zener loads that only protect overvoltage to some % of the max. cell power ( volt-amps ) during CV charge so it works during CC mode as well.

This is needed to prevent rapid cell death or low DoD of a low capacity cell.

The array S*P number depends on choosing the optimal power impedance curves to the load and discharge times. Matching cells in series strings is just as important as parallel cells if using the max C rate charge/discharge values. If relaxed, and heat sensors are used, then mismatch tolerance can rise.

The series ESR of battery cells rise and must be much lower than the motor ESR or DCR to reduce load regulation error and battery losses. So this is an impedance factor that demands you have a good battery tester that not only measures Vbat vs I but computes estimated capacity and Coulomb counting during charging and batter ESR, such as some Cadex models.