Struggling with LiPo load sharing circuit

[ebastler]

I might have a mental block here... Trying to get my head around this load-sharing circuit from https://blog.zakkemble.net/a-lithium-battery-charger-with-load-sharing/



The idea is that the battery charger IC U1 (MCP73831) should not see the additional current draw from the load -- to limit the current draw on U1, and to allow the charger to correctly monitor the battery charging curve. So when external USB power is supplied, the load is powered directly from that via diode D1. So far, so good.

But isn't p-channel MOSFET Q1 connected the wrong way round? Unless I am missing something, when there is no external USB power, Q1 is turned on, but will still only conduct via its source-drain diode? (With a less-than-ideal voltage drop.) Might as well just use another Schottky diode instead of Q1, right?

Flipping the source & drain connections for Q1 should get a better result for the no-external-power situation, with Q1 conducting the battery voltage to the load with a lower voltage drop.  But with USB power applied, Q1's source-drain diode would then charge the battery directly and in an uncontrolled way -- not good.

So what is the right way of doing this? Is there an accepted standard solution using MOSFETs and/or diodes, which minimizes the voltage drop when supplying the load from the battery? Or am I just misunderstanding the circuit? Thanks for your help!

[2N3055]

For current not to go from battery, voltage drop on mosfet should be bigger than on the Schottky diode.
Which is true for intrinsic diode in Mosfet.

And when there is no outside power, Mosfet gets to conduct, decreasing voltage drop to almost zero, so you have full battery voltage.
Make note Mosfet is variable resistor, it will conduct in both directions when conducting.

Similar mosfet circuit is used for polarity reversal protection...


https://www.monolithicpower.com/learning/resources/designing-a-reverse-polarity-protection-circuit-part-i

[ebastler]

And when there is no outside power, Mosfet gets to conduct, decreasing voltage drop to almost zero, so you have full battery voltage.
Make note Mosfet is variable resistor, it will conduct in both directions when conducting.

Duh, I did not know that.   I always assumed that the source/drain junction can conduct in one direction only. So, as long as the body diode does not get in the way (which it doesn't in this application), one can use MOSFETs "backwards"?! Great, then the load sharing circuit makes sense.

Well, at least I know what I don't know -- asked this in the Beginners section for a reason. I can find my way around passive networks (which were tought at uni), and have a decent handle on digital and op-amp circuits (mostly self-taught, starting top-down with computers). But inbetween there is this mysterious, uncharted valley of single transistors...

I should probably make another attempt at filling that gap systematically, rather than just cobbling stuff together ad hoc. In the meantime, many thanks for educating me!

[2N3055]

And when there is no outside power, Mosfet gets to conduct, decreasing voltage drop to almost zero, so you have full battery voltage.
Make note Mosfet is variable resistor, it will conduct in both directions when conducting.

Duh, I did not know that.   I always assumed that the source/drain junction can conduct in one direction only. So, as long as the body diode does not get in the way (which it doesn't in this application), one can use MOSFETs "backwards"?! Great, then the load sharing circuit makes sense.

Well, at least I know what I don't know -- asked this in the Beginners section for a reason. I can find my way around passive networks (which were tought at uni), and have a decent handle on digital and op-amp circuits (mostly self-taught, starting top-down with computers). But inbetween there is this mysterious, uncharted valley of single transistors...

I should probably make another attempt at filling that gap systematically, rather than just cobbling stuff together ad hoc. In the meantime, many thanks for educating me!

There is no shame in asking questions..

First time I saw this reverse protection circuit i was also WTF.... But after a minute I was "DUH but of course!"..

I actually very much enjoy discrete electronics, sooo many clever things are possible by going out of the box.

Also some old ICs are also fun. For instance, once I made a panic light controller charger that had complete (simple though) state machine by use of 555 and TL431, using 555 internal bistable multi and dividers and TL431 as comparators and references.
Today that is job for a 1€ 8 pin µC.
But not as much fun!!

[Peabody]

The direction of the body diode is key here.  With the mosfet turned off, the body diode and the schottky form a simple diode-OR circuit, where the supply with the highest voltage will supply all of the load current (because the other diode will be reverse-biased).  And since battery voltage is always lower than 4.7V, the external 5V supply always wins.

But in this orientation, the body diode also prevents unlimited current from flowing through the schottky diode, then through the mosfet, back into the battery, which you don't want.  Only the charger should be charging the battery.  If the mosfet were oriented "normally", that flow would become possible as the battery discharges.

Here's a more official source for this design:

http://ww1.microchip.com/downloads/en/AppNotes/01149c.pdf