Single lipo cell low voltage load disconnect circuit

Has anyone made a circuit that disconnects a single cell lipo when it's voltage gets down to ~3.0V to prevent over-discharge? I'm trying to figure out the best way to do it for the following system (it's only a simple block diagram):



The system can be powered either from the USB port (also used for charging up the lipo) or from the single lipo cell.

The device could be sat unused for a long length of time, and it's possible that the micro in sleep mode eventually discharges the lipo down until it's internal low voltage protection trips. Now that's fine from a safety point of view, and I've tested that connecting the charger will re-connect the battery and charge it back up, but the data sheet says that they recommend that the battery is not discharged down below 2.75V and the built in protection kicks in ~2.5V.

To try and prevent that I'd like to have an extra layer of protection and disconnect the battery from the circuit at 3.0/3.1V ish. Then there is just the voltage monitor circuit discharging the battery, which should mean the battery doesn't get discharged down to the actually batteries low voltage cut off protection too quickly.

Once in this state, and the battery is disconnected, I'd like it to latch in that state so if the voltage creeps back up with no load on it it doesn't re-connect.

Once the device is plugged into USB again, the circuit will power from USB and the uP will know the battery has been disconnected (battery voltage is monitored by an A2D at the input to the board), so it will reset the latch in the monitor circuit and re-connect the battery so the charger IC can re-charge the battery.

All this is to hopefully stop batteries being damaged when units are not used for a long time.

Has anyone done anything like this?

Thanks,
G

Just use battery pack with protection circuit, it will be protected from over current, over/under voltage etc.

Or if you want to build the protection circuit yourself then use special ic for that, something like **broken link removed** from seiko instruments - or there are some others BQ series by TI.

I've looked extensively at the dedicated chips available but I haven't found one that latches, and I'd really like this circuit to latch off and disconnect the battery. I don't really want the no load voltage creeping back up and reconnecting the battery.

I know there is protection built into the battery pack, but I'd like to use that as a last resort. I'd really like to disconnect the battery at a higher voltage to give the device more shelf time before the battery gets deeply discharged and potentially damaged.

There is a slight complication as to where I connect the charger to the battery. I can see how it would be possible to connect as shown, but the charger I am using is the bq24073 that you connect the battery and USB power to and it sorts out the power supplied to the rest of the power supply.

I've been prototyping a little circuit (I've attached a picture of it and added system blocks around it to explain it better) and it seems to work, but it now raises some other questions.



The MAX835 monitors the voltage, disconnects the battery and latches off. I've now connected the clear line to the +5V from the USB, because it won't be possible to charge the battery again without connecting the device to a USB 5V source, so that might as well clear the latch and re-connect the battery.

However I've realised that because of the charger used the load switch needs to sit between the charger and the battery. So..... to charge the battery current needs to flow backwards through the PFET. I've tried it and it seems to work, but I'm not sure it's the best solution.

So is reverse current through the PFET a bad idea? Will it damage it?
Would a dedicated load switch be better?

I'd also like to minimise the quiescent current that the monitor circuit uses as ultimately once the battery is disconnected that is what affects the 'life' of the battery.

I'm really enjoying working on this but I find myself just spending endless time going round and round in circles looking at data sheets!

Thanks,

G