Li-ion cells mosfet switching circuit

[mrmp17]

I'm designing some kind of power bank with two 18650 cells, but the problem is I need 2s Li-ion voltage at output, but also charge it from USB. Booster circuit is too ineficient for my aplication and 2s charging too complicated, so why not make a circuit that switches cells from two in series for discharging to two in parallel for charging. I came up with a schematic using a few mosfets but I'm not sure if it'll work. So, what do you think about my circuit?

Mosfets used will AO3401 and AO3402. Output current is going to be sub 1A and charging current around 1A.

Schematic attached below.

[NiHaoMike]

A boost converter is more efficient than a one cell linear charger running from 5V. It would only be about 76% efficient over much of the charge cycle, while a boost converter can easily get 90% or more.

[mrmp17]

I meant boosting from 1s for discharging, charging efficiency is not a problem.

[NiHaoMike]

Still way overcomplicated for what you're trying to do. Just use a boost converter to boost to 8.2V (4.1V/cell) and that voltage is fine to leave connected to the cells.

[mrmp17]

Just a booster is not enough, Li-ion cells must be charged with CC/CV charging cycle and if there are two cells in series, you must balance their voltage to prevent one cell from over charging or discharging.

[onlooker]

The much easier solution is a DPDT center off switch. You get charge/discharge/off 3 positions and all may be needed.

[dmwahl]

I'd suggest a boost circuit to 8.2V for charging, as others have said, and to keep the cells balanced a TL431 configured as a "high power shunt regulator" across each cell (see datasheet example). You set the turn-on voltage of the TL431 with a simple voltage divider. Setting it to 4.15V (or just above your charge voltage / 2) should be about right to keep either individual cell from going over 4.2V. Mind the reference input current on the TL431, as I recall it's around 2uA which is enough to affect the set point. This is the same basic circuit that most inexpensive multi-cell BMS systems use.

[mrmp17]

That seems like a good option, but do you guys think that would be simpler circuit than a few mosfets to rearange cells for charging? I mean, you need boost regulator controller, a diode, an inductor etc. And on the other hand, mosfets are just some sot-23s that take up a very small amount of board space, and they are realy cheap, also 1s lipo charging is practicaly a one-IC solution.

[KL27x]

Neat idea. Worth doing it just for the novelty. Be sure to update the thread when you have results.

[dmwahl]

Depends on your goal for the project, if it's to be easy to use and simple, then I wouldn't use MOSFETs switching the orientation around. Since MOSFETs switch from full on to full off over a range of gate-source voltages (and even the same parts will have variance due to temperature, batch variation, load current, etc) you'll need to sequence the switching to avoid the situation where both are partially on. Search for "mosfet shoot through" if you're not familiar with the idea already.

The TL431 circuit is also simple, and you don't need to worry about shoot through. I'll even argue it will be more simple.

I do agree that the MOSFET version is a good learning exercise, but you'll find it will be more complicated in real life than on paper. Try building it in SPICE if you want an idea of how it will work. For the purposes of simulation, you can just use voltage sources as your batteries.

[mrmp17]

Ddmwahl, yes, I am aware of possible shoot-through, but maybe I could get away with it because of low voltages and very fast switching times, would deffinitely test it out before using tih circuit in my project. Do you know of any good circuit simulators? Anyway, I will probably try it out on a breadboard and see how it works, but it'd be nice to play with it a bit in a simulator.

[dmwahl]

I use LTSpice myself, reasonably easy to get started and has a decent library.

[mrmp17]

I tested the circuit on a breadboard today and it seems to be working  excellent! It looks like mosfet shoot-through isn't much of a problem, but I don't have a scope to measure it.

[Artlav]

simpler circuit than a few mosfets to rearange cells for charging?

How about a DPDT relay?
Solid state one if you are space constrained.
Wire it so that when it's off then it's in discharge mode, and when on it's in charge mode (to avoid wasting battery power on the coil).