putting three li-poly charger chips in a series config?

[frank26080115]

I have a circuit I am about to fab to recharge 2S or 3S li-po battery packs.

It uses three switching DC/DC step-down voltage regulators, and three MCP73831 charge management chips, each set to 500mA.

But the ground reference for each cell is dependent on the cell below it.

If a cell is missing, then the charger for that particular stage will see that there's no battery, but I am unsure what the stage above it will do, because the regulator's feedback will become referenced to the output of a charger that doesn't really have a cell.

At best, I don't think any damage will occur to that charger, and I don't think there's any reason for a charger to over-charge or over-current. So I think I am safe.

Is the kind of design I am doing safe and going to work? If not, please explain why not.

Thanks.

[KL27x]

But the ground reference for each cell is dependent on the cell below it.

Setting aside the "ground reference," where is the actual return path for the top two DC converters?

Are you connecting each stepdown converter to a separate, 24V isolated power supply? Cuz that's the only way I see this working. I don't see separate in/out grounds on your voltage converters, and the first word on the datasheet is "Non-Isolated."

I am curious to see what happens when you plug it in.

[richard.cs]

I've made a similar charger for four cells in series. The charging chip I used needed a 5V supply so the mains powered version uses two dual-output transformers to give 4 independant secondaries which are seperately rectified and feed four 7805s. Not the most elegant design perhaps but it does work.

There's a 12V input version I never finished which has a high frequency transformer with 4 output windings driven by a simple logic gate oscillator and a pair of mosfets in push-pull. The rest of the circuit is otherwise the same,  keeping the linear secondary regulators because the transformer has a fixed stepdown ratio.

[frank26080115]

The ground return path will be through a cell. Thus, that particular regulator will not see 24V at all, instead, 24 minus cell voltage, or minus 2x cell voltage for the top regulator.

Assuming the cell is present of course. If the cell is gone, it should be as if the ground wire got cut.

But I am simply worried about the constant current output mode of the MCP73831 chip screwing up the feedback signal on the next regulator, or the PROG pin of the next MCP73831.

[amyk]

You do realise that there are single-chip solutions to this, right? Laptops with 2S/3S/4S packs use a single charge controller with balancing.

[KL27x]

The ground return path will be through a cell. Thus, that particular regulator will not see 24V at all, instead, 24 minus cell voltage, or minus 2x cell voltage for the top regulator.

So there's no individual charging/balancing (which isn't necessarily a deal breaker, I know)? Unless the 78731 will also sink current, will the top charger charge all the cells below it with the same amount of current?

[frank26080115]

So there's no individual charging/balancing (which isn't necessarily a deal breaker, I know)? Unless the 78731 will also sink current, will the top charger charge all the cells below it with the same amount of current?

Hmmm... if all MCP73831 are in CC mode, would the bottom cell get 1500mA? err... there's a chance lol, hence why I wanted to review this stupid idea of mine

[KL27x]

It's hard for me to figure out. Maybe so. Or maybe the bottom cells would get 500mA as expected. But if the bottom cell were to enter CV mode before the top cell, it seems like it would still get 500mA until the top cell reaches 4.2V, too. The bottom two chargers might not be doing "anything" much. Damn, I can't follow the bouncing ball. This circuit is just too weird. I use that MCP device, too, but I don't know what it will do in this circuit.

In practice, I find cell balancing unnecessary even when the draw is pretty close to C. So if the latter, maybe this will work ok, in practice? I'm highly curious. If the former, maybe you could configure the bottom two 78731's to output <1mA max? Oh, that wouldn't allow charging of a 1S or 2S battery. Or would it? . You owe us some real world results. I'll be checking back

[mikerj]

You do realise that there are single-chip solutions to this, right? Laptops with 2S/3S/4S packs use a single charge controller with balancing.

No they don't!  Laptop batteries controllers provide under and over voltage protection, and gas gauge functionality, but they do not provide cell balancing.  This is one of the reasons that the battery capacity drops off as more charge/discharge cycles get put on the battery (in addition to normal cell ageing).  Very often you can crack open an old battery with very low capacity, balance all the cells and return most of the capacity back; I've done this to numerous laptop battery packs. It takes a few charge/discharge cycles for the gas gauge to recalibrate itself.

I can't see the OP's scheme working correctly since the lower cells will be seeing the charge current from all the cells above it, plus it's own charge current.  e.g. if  you charge at 1C, the top cell would be getting 1C, the next one down would be getting 2C, the next one 3C etc.  To make this work you need isolated chargers.

[Rerouter]

Use a dumb 3S charge controller, add 3 op amps measuring deferentially, and use them to turn on a balancing shunt resistance of 1.1C if the cells voltage goes slightly above what your intended charge voltage is, this way yes you would need to complete a full charge to balance the pack, but it will balance it, and you get the benefit of per cell OVP

[frank26080115]

well today was gerber out for a bunch of stuff and this gerber didn't make it as I am no longer confident in it. thanks guys

cheap 2S 3S balance chargers that are below 800mA are very rare but I did find one on AliExpress, I'll just grab like 6

[amyk]

You do realise that there are single-chip solutions to this, right? Laptops with 2S/3S/4S packs use a single charge controller with balancing.

No they don't!  Laptop batteries controllers provide under and over voltage protection, and gas gauge functionality, but they do not provide cell balancing.  This is one of the reasons that the battery capacity drops off as more charge/discharge cycles get put on the battery (in addition to normal cell ageing).  Very often you can crack open an old battery with very low capacity, balance all the cells and return most of the capacity back; I've done this to numerous laptop battery packs. It takes a few charge/discharge cycles for the gas gauge to recalibrate itself.

They do, but probably not provide enough balancing current if the cells are severely mismatched. Here's a common IC used in laptop batteries:

http://www.farnell.com/datasheets/1882624.pdf

It has integrated balance MOSFETs, but possible to increase balance current using external ones:

http://www.ti.com/slua420-aaj

[djQUAN]

Search for Turnigy-5011 lipo charger. It is very cheap and commercially available version of what the OP is planning to do but with LTC4054 clone chips.

Although I haven't checked if it is still currently available as I bought one a few years ago.

also see: http://belza.cz/charge/battr.htm

[frank26080115]

Thanks, that site had images of the PCB, I bet they use the transformer and diodes to isolate. Well I guess if the cheap ones all look like that on the inside, I'm not gaining any quality improvements over my DIY solution, and I should be able to just swap resistors to change the current.

[djQUAN]

The powersupply uses a flyback topology with three separate isolated windings with feedback taken from only one.

If I remember correctly the controller was a UC3842 or similar with external mosfet.