I'm trying to teach myself electronics hence my beginner skills in design. I want to design a circuit to power another with two 18650 li-ion batteries in series. Charging them in series seemed like a lot of work with load balancing - and li-ion charging ICs are cheap and abundant, so I reasoned that I could charge each battery individually with a separate IC for each li-ion battery, from USB power.

The idea is when USB is connected, the batteries need to charge. Each battery is disconnected from the other and charged via the charging IC. When USB is disconnected, the batteries are connected back in series and power the circuit. Below is a basic circuit that seems to work, but I'm wondering if there's some kind of thought process or analysis I'm missing. In the circuit I changed USB from 0V to 5V and observed the output changing from 7.2V to close to 0V. Any help is appreciated. :)

LTSpice circuit of possible power circuit

  • \$\begingroup\$ What does M4 do? \$\endgroup\$
    – Andy aka
    Dec 11, 2021 at 15:05
  • 1
    \$\begingroup\$ The trouble with MOSFETs is that they also have a build-in diode between Drain and Source. That means that the MOSFET doesn't behave like a proper switch. Also, the Gate-Source voltage needs to be high enough for the MOSFET to fully conduct. For a beginner, these are traps you will fall into. It is much easier to study similar circuits designed by others that have proven to work and then use those. That will result in much less headaches for you, I guarantee it. \$\endgroup\$ Dec 11, 2021 at 15:39
  • 3
    \$\begingroup\$ Also: I have over 30 years of circuit design experience and I would absolutely not consider switching cells in series like you propose. I would 1) permanently connect the cells in parallel (balance them first), if I need more than 3.7 V, I would use a boost converter. Or 2) connect the cells in series but then treat that as a single 7.4 V cell. So charging that will require 9 V (optionally from a boost converter). Why? Because there are simply too many disadvantages to do it like you propose. And as a beginner, things will go wrong/blow up in your face, I guarantee it. \$\endgroup\$ Dec 11, 2021 at 15:46

1 Answer 1


The idea is when USB is connected, the batteries need to charge. Each battery is disconnected from the other and charged via the charging IC.

OK, I have a little observation but maybe later down...

When USB is disconnected, the batteries are connected back in series and power the circuit.

OK, here's the problem as I see it - when USB disconnects, the gate-source capacitances of the MOSFETs will hold that state of charge and nothing really happens - the MOSFETs remain in the state they were previously and er.. everything stays the same except...

As the gate capacitances naturally discharge, the MOSFETs might (and probably will) reach a point where the series elements (M1 for instance) are/is still pretty much "ON" and, M2 (the grounding MOSFET) also starts to activate. At this point there will be smoke and fire so beware.

To solve this I'd want to sequence the MOSFETs under control of comparators to ensure that the MOSFETs switch on or off very quickly and, with some dead-band to prevent the much less likely opportunity (but still finite possibility) of fire and smoke.

This addition will also be useful when USB is reconnected. I don't believe a simple pull-down resistor is enough without dead-band circuits being implemented.

I have no idea what M4 does: -

enter image description here

You also need to make sure that the P channel MOSFETs are sufficiently turned on when the USB voltage is low. Your source voltage may be as low as (say) 2.7 volts so, choose a MOSFET that has a low \$V_{GS(THRESHOLD)}\$.

  • \$\begingroup\$ Hi, thanks for the fast response. When you say use comparators for sequencing the MOSFETs, I assume I would use USB voltage on the input and use battery voltage and ground on the power rails? I included the last MOSFET, M4, because when simulating if I didn't include it, when USB was connected, the output didn't drop to close to 0V, it dropped to 3.6V. Perhaps I should consider Bimpelrekkie's words and scrap the idea and rather opt for 2 batteries in parallel and just scrap this idea for now - seems its not the best approach. \$\endgroup\$ Dec 11, 2021 at 20:07
  • \$\begingroup\$ That's up to you. I've tried to give you an answer to your question without second guessing your actual skill levels. You got the MOSFETs right as far as I can tell. Just the control and M4 need thinking about. Parallel charging followed by a booster is a safer move so, maybe, with a few schematic nuances you can ensure your pcb can service both options. You only learn by trying but, have that duality option up your sleeve. \$\endgroup\$
    – Andy aka
    Dec 11, 2021 at 20:15
  • \$\begingroup\$ @IanEdwards are we done now buddy? Do you still have any questions or concerns? You would need to use USB voltage to power the comparators that feed the P channel MOSFETs in order to ensure they turn-off correctly. \$\endgroup\$
    – Andy aka
    Dec 12, 2021 at 9:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.