Dividing lifepo4 cells

[Solr]

Hi,

I have ordered 4 3.2v 310ah lifepo4 cells for use as a camper van leisure battery.

Because of space constraints, ideally I would like to split them in to two groups of two; so join two cells together with bus bars then join them to a second group with cables about 50cm in length.

I would then use a single bms to manage all the cells as a single battery.

I realise the cable could cause a volt drop so hoped by oversizing it significantly could eliminate any problems, the volt drop could be less than 0.5%

I realise that this wouldn’t be an ideal situation, but is it doable?

I also thought about a separate 6v bms
For each pair of cells, but couldn’t find anything suitable.

Your advice and thoughts would be much appreciated.

Thanks in advance

 

[snoobler]

Welcome to the forum.

Not uncommon to split a battery, but 2X 6V BMS bad idea. Managing all 4 cells with one BMS is way better than 2 BMS managing 2 cells.

Just treat is as a super long bus bar. Pick a wire gauge suitable for your current.

Example: 2 gauge @ 50cm, 12V and 200A has a .1V drop:

Voltage Drop Calculator

This free voltage drop calculator estimates the voltage drop of an electrical circuit based on the wire size, distance, and anticipated load current.

www.calculator.net

 

[Solr]

Thanks for the quick reply, that’s good to hear. It’s amazing to be able to fit such a high capacity battery in a small space

 

[DLTooley]

I’d presume you’ll need to extend the wires on the bms. No reason to think that’s a problem, but that’s the thing I’d keep thinking about, and watching.

 

[curiouscarbon]

i’ve seen some people increase balance lead thickness and fuse them. might be worth considering if the distance is increased relative to a single pack. awesome idea btw

 

[Solr]

Good suggestions, thanks. I haven’t started looking for a bms yet, probably my next question

 

[houseofancients]

andy has a pretty good explainion in rather simple terms why there are proS and cons to each approach.

i personally would never parallel within the same pack unless they are fully mached cells, too much chance on drift of 1 cell taking out a complete pack, rather spend a little more on an extra bms

 

[DJSmiley]

How many amps are you're planning to draw? The wire gauge has to be fine for that, if you're only using 10A or so, basicly anthing will do.

For the balance wires: the balance current is 50mA or so for the JDB/Overkill BMS, so even 20AWG would be way overkill (But i'd go for that)

As said, make sure the balance wires are secured properly to prevent a short. I don't prefer fuses for the balance wires since they will add resistance (more than the actual cable). If you're using not too thick wires to extend the balance leads. If, somehow, it still goes wrong, even tho you have secured the wires, the balance cable will just burn (basicly act as a fuse-with-smoke). Too thick wire and that won't pop, so I won't recommend 10AWG as balance wires

You might replace all busbars with similar length wires, even the ones which could be busbars... that makes the voltage drop between all cells equal.

 

[ereams65]

I've always heard the parallel wires need to be the same length for resistance sake. Something to keep in mind.

 

[snoobler]

I've always heard the parallel wires need to be the same length for resistance sake. Something to keep in mind.

I've heard the same, but I disagree. These are sensing wires. They pass very very low levels of current. Assuming 32awg wire, 5mA current (likely 1-2 orders of magnitude higher than actual), lead length of 1m and the outlier lead at 2m, they will read 0.005V difference.

That's noise.

The answer is different if the sense leads are used for active balancing (rare).

 

[ereams65]

I've heard the same, but I disagree. These are sensing wires. They pass very very low levels of current. Assuming 32awg wire, 5mA current (likely 1-2 orders of magnitude higher than actual), lead length of 1m and the outlier lead at 2m, they will read 0.005V difference.

That's noise.

The answer is different if the sense leads are used for active balancing (rare).

I was talking about the parallel connections by moving half the cells further away. If that info is accurate you would need the same length wire even on the cells that are next to each other, correct?

 

[snoobler]

I was talking about the parallel connections by moving half the cells further away. If that info is accurate you would need the same length wire even on the cells that are next to each other, correct?

There are no parallel wires in the main connections. The only place "parallel" can be applied is in the BMS wires, so that's what I thought you meant.

These are all series connections. I provided an example of a 50cm 2awg "bus bar" having only a 0.1V drop at 200A (2400W).

 

[ereams65]

There are no parallel wires in the main connections. The only place "parallel" can be applied is in the BMS wires, so that's what I thought you meant.

These are all series connections. I provided an example of a 50cm 2awg "bus bar" having only a 0.1V drop at 200A (2400W).

Yeah I meant Series. I think he needs 2 of the 4 cells in a different location

 

[Just John]

i’ve seen some people increase balance lead thickness and fuse them. might be worth considering if the distance is increased relative to a single pack. awesome idea btw

You need to be careful. A fuse adds resistance. Ideally keep all balance/sense wires the same length and thin enough not to need a fuse, the wire is the fuse.

Edit since Snoobler pointed out, keeping them close in length works. Just don't make one 20 meters and another 1 meter.

 

[snoobler]

Thin wires are their own fuses.

 

[DJSmiley]

The added resistance of a fuse doesn't really affect the reading of the BMS.

Adding resistance doesn't cause any issues, it is the corresponding voltage drop that will cause misreadings. But this voltage drop is determined by the resistance AND the current.

Does it matter? That depends on the BMS. The current used for the sensing of the voltages is almost none, so for that purposes it's unnoticable.
But the balance leads are also used for balancing. If the BMS has only a low balance current, the voltage drop over a fuse is still barely noticable. However, if you have a BMS which can do for example 5A of balancing, it might indeed affect the readings (during balancing)

If a BMS doesn't have separate voltage sense and balance leads, balancing will affect the readings. Noticable? Depends on the current and total resistance. (I haven't seen any).
For high-amp power supplies this is more common practice: 2 wires carrying the high current for the loads, and 2 wires just for voltage sensing (So the powersupply can compensate for the voltage drop in the wires)

As said, using a thin wire (also when extending) is the easiest and still safe way, the wire will just melt.

Just make sure you're mounting the balance wires properly, so they can't rub against busbars or sharp corners or so which could damage the insulation.