1S4P battery split into remote halves

I’m designing an evolution of a product and the ME’s have proposed an enclosure (for the sake of size reduction in a particular axis) that would split the current 1S4P 18650 (3.7V) based lithium battery pack into two halves, with about 8 inches between them. I know this is rather unconventional…

Current design specs/needs related to power:

  • AC input voltage 12Vdc+
  • Ability to slow charge from an external USB 2.0 host (5V)
  • Main system power rail is 3V3 and below so just using a buck converter
  • Currently using a TI BQ24192 single cell charger and BQ27621 gas gauge

My concerns off the bat:

  • Both 1S2P packs would now include their own battery protection circuit
  • Extra impedance of the wires and voltage drops (use bigger gauge wire to mitigate?)
  • Cell balancing (right off the bat after assembly, and long term)
  • Will the fuel gauge be anywhere accurate?
  • Total available energy
  • Other safety concerns

My gut tells me that joining the packs in SERIES would alleviate the balancing issues, but he only reason I would want to keep the packs in PARALLEL is to avoid implementing a boost converter when charging from the external USB 5V. I rather not use some unicorn IC, especially giving the supply constraints we face right now.

Another option would be to treat the two packs as separate power sources and switch between them with an ideal diode setup of some sort, but I rather avoid that complexity.

Also, this isn’t very cost sensitive so I prefer simple & reliable versus elegant & cheap. :slight_smile:

My availability to test equipment is limited right now so I can’t do any testing, and I’m not too well versed on simulations.

Any thoughts or suggestions? Thanks!!

I’m actually doing a similar design right now.

Why is this a concern? Besides a mild cost adder, they should work in parallel relatively well.

I’m assuming it’s the same wires for the 1S4P as the 1S2P, and if so, the voltage drop should be half.

So, Cell balancing for packs in parallel isn’t really much of a concern. Instead, if you go in series, it will mandate cell balancing.

If it isn’t cost sensitive, then the diode OR might be the better solution. However, that means having fully parallel batteries and charger ICs.

Hope that helps,

To add to @Slawek 's answer, parallel cell balancing really only matters when you add or remove cells from the pack. During normal operation it takes care of itself.

The scariest case is when a user decides to replace one of your two packs, leaving you with a combination of charged and discharged cells. If the total impedance in the path is high enough they’ll balance without issue… if not, you’ll need some additional limiting. Maybe a pair of e-fuses with an intentional leak path?

The gas gauge should be fine, just treat it like a normal 1s4p.

Total available energy should be similarly fine.

1 Like

As mentioned, biggest concern will be the attachment of two unmatched packs. You will equalize packs as fast as physically possible, which can be damaging or even catastrophic (hopefully with built in protection its just damaging). If the internal impedance varies too much between the packs you will get uneven charge and draw. However, it may not be completely noticeable or even an issue. Just don’t mix brands or cell series. Have you tried reaching out to your battery pack vendor?

Thanks for the responses and reassurance.

I guess my concern was that the voltage drop would be different because in my application, one set of wires may be 3" and the other 7" but I’m thinking if I should just standardize the length to 7" for the sake of limiting to one custom SKU and just tuck the extra wiring somewhere.

Thankfully, these won’t be user serviceable in my application, so I’ll take care to incorporate a “battery pack introduction/conditioning” step prior to having the separate packs joined on the PCB.

That’s a clever idea. :+1: