5S LiFePO4, 12v, no BMS

[Schludermann]

The conventional wisdom for managing a 4s 12v LiFePO4 battery using BMS for longevity and safety. Now that costs and availability have significantly improved, there may be an alternative method. What if an addition cell is added to create a 5s, but still is only charged to 4s voltage, 14.6v. With 5 cells in series, this would result in an 80% SoC. Since the Cells are not fully charged, only to 80%, would operating without BMS be feasible? What other current limiting devices should be considered in the 5s, no BMS configuration? I've tried searching for discussions of this configuration and haven't located any. Perhaps, I'm not using a suitable vernacular. My application is to have point of use generation and storage in distributed locations around the property using a common panel and battery modules. perhaps even a mobile unit on rolling wire shelf. Once the batteries are charged there would be ancillary applications for the surplus power. Such as Compressed air storage, atmospheric water harvesting and others.

 

[amberwolf]

The problem a BMS is there to manage is that cells are typically quite varied in their internal properties, so they have different capacities, resistances, etc., and so charge and discharge at different rates.

The more varied they are, the more likely a condition is to happen over time where the worst cell empties before the best or even average is halfway down, and that can lead to overdischarging and damaging that cell.

The same happens where that worst cell is charging, and fills before the others and then overcharges, damaging the cell.

This damage could lead to a fire.

The differences in properties between cells typically grow worse with time.

A BMS will monitor each cell or group's votlage and disallow further charge or discharge in the event of excessive differences between cells, or if one exceeds a limit.


Also, a BMS is not a current limiter, in the general sense. If designed to do so, it can monitor current and turn off hte ihnput or output completely, but it doesn't do what a charger, controller or inverter would do to limit current but let it still flow by lowering the voltage.



You can run without a BMS as long as you periodically manually check for these differences, and manually correct them when they happen, or replace the cells that are different from the others to eliminate the rpoblem.

Using a higher s-count and charging to lower SoC stresses the cells less, but doesn't prevent the issues above.

I don't know anyone that's done this in the specific situation, but I use a BMSless 4s NMC pack for lighting on my SB Cruiser, and only charge it to 16v (4v/cell) instead of the 4.15v/cell it's meant to. I do this for pack longevity and becasue I don't need 40Ah of lighting capacity. (I also do this with the 14s traction pack). But the cells I use are well-matched in properties and have stayed so as they aged, so I don't get the variation that causes problems BMSes are meant to prevent (but I still periodically check things under load, charge, and statically).



While here, I'll also mention that if you are using prismatic or pouch cells, make sure to use a compression structure on it's large flat faces; see the various posts about cell compression for examples. Some of the used-ev-module sellers have 4s / etc packs built with this structure already, that you can disassemble and reuse the cells and structure.

 

[bananu7]

addition cell is added to create a 5s, but still is only charged to 4s voltage, 14.6v. With 5 cells in series, this would result in an 80% SoC

This might be a problematic assumption. I think it's going to be very hard to accurately manage SoC with the lower voltage limit, and what might end up happening is that you'd stop the charging way earlier if the battery voltage rises too quickly, thus ending up with a lower capacity with 5 cells than you'd otherwise get with fully utilized 4.

The cell operational limits aren't really random, and while staying slightly below max voltage can be beneficial, I would recommend switching the BMS to a 5S capable model instead, or simply not bothering with such modifications.