Powering a project off LiPo and charging it

[Entropia]

Hey guys. I'm developing a small GPS tracker project that would work off a single cell LiPo-battery. During normal operation the microcontroller, GPS receiver, LCD and SPI Flash memory would be powered from the LiPo using a buck converter. A small buck that outputs 3.3V @ 200mA should do just fine, and since the LiPo cells don't really have much juice after going below 3.3V cutting off there is just fine by me. I however do need some advice on how to cut off the battery at, say, 3.2V. Some buck converters have a shutdown/enable pin that I can use in conjunction with a voltage supervisor but I'm worried this might cause oscillation. When the cell voltage drops below 3.2V, the supervisor will shut down the buck and cause an increase in the cell voltage which would enable it again and so on... So, any thought how I might realize the UVLO (under-voltage lock-out)?

The other thing I need help with is the charging. I have settled on using a MAX1811 for charging from USB port. The MCU will have an integrated USB peripheral, so that I can ask for permission to draw 500mA from USB bus. Conveniently the MAX1811 has a pin for toggling between 100mA and 500mA so this is not a problem. But what if the UVLO has kicked in and MCU is off? I figured I'd pull charging current to 100mA with a 10K pull-down and pull the enable pin to high with a 10K pull-up. This way it would start charging from USB even if the MCU was not energized. When the system powers up, the MCU can take over the charge monitoring, request more power and so on. Also I think it's wise to minimize all current consumption if the charging current is 100mA (switch off LCD backlight and GPS which draw around 150mA together)- I've drawn a simple diagram (link below) to illustrate the system.

So far I only have this in my head, so no schematics have been drawn yet. I'm open for suggestions, especially regarding the 3.3V regulation in the system. Also, any other ideas/thoughts are highly appreciated.
 
Simplified diagram of the system is here: http://users.jyu.fi/~mitarist/dropspot/gpsblockdiagram.png

[8086]

I am working on a similar project. I'm using this to get 3.3v from a single lipo cell: http://www.ti.com/product/tps61200

And it has programmable UVLO.

[amyk]

How much capacity does your battery have?

[Entropia]

Hey guys.

8086: TPS61200 looks nice, but there's one problem. I can't solder it.

amyk: Not decided yet. Probably 800 to 1000 mAh.

Actually some time after I posted the thread I decided on using MAX710 for the regulator. It has a built in UVLO but I'm a bit confused about how it works. The device has an active-low SHDN-pin that can be used to disconnect the regulator and any load from the battery. It also has the aforementioned Low Battery-comparator and an internal reference. Ideally I would set the threshold voltage with a voltage divider and run the active-low comparator output straight into the SHDN pin. The part I'm confused about is that the SHDN pin says when it's active the "entire circuit is off" but then comparator is drawn "separately" in the functional diagram. It might still get its power from the regulator output, though. So, does this active SHDN include the LB-comparator and reference voltage as well? If that's the case, then what's the point of the whole comparator?

I was also thinking about using a discrete comparator and a voltage reference of 3.1 to 3.2 volts but that would bring the static current consumption up by hundreds of microamps, if not milliamps. I would have to leave a safe enough margin in the battery for the comparator circuit to draw after the actual cut-off voltage. That could be something like 10-50 milliamp-hours so the system could last 10-50 hours without charging. Ofcourse the battery will have its own under-/overvoltage protection circuits but knowing the volatility of the battery type, I'd rather not leave it up to the cheap Chinese battery manufacturer's chance.

[amyk]

Not decided yet. Probably 800 to 1000 mAh

That's the sort of capacity you'll starting needing dedicated charging/discharging circuitry for.

The part I'm confused about is that the SHDN pin says when it's active the "entire circuit is off" but then comparator is drawn "separately" in the functional diagram. It might still get its power from the regulator output, though. So, does this active SHDN include the LB-comparator and reference voltage as well? If that's the case, then what's the point of the whole comparator?

The entire IC is powered from the PS pin (see page 6). I believe this includes the comparator too. On page 2 you can see the "shutdown quiescent current" is at most 5 uA.

[baljemmett]

(Gah, I replied there based on one of the later messages; have just re-read the original post and you've already considered the voltage supervisor-based design I'm using.  Whoops, sorry for the noise.)

[Entropia]

amyk: I have the MAX1811 for charging, do I have to somehow control the discharging as well?

baljemmett: If you have succesfully used a voltage supervisor as UVLO with a buck/boost-regulator and a LiPo battery, please share your experiences. :-) Any thoughts on the oscillation thing I was wondering about in the OP?

[baljemmett]

baljemmett: If you have succesfully used a voltage supervisor as UVLO with a buck/boost-regulator and a LiPo battery, please share your experiences. :-) Any thoughts on the oscillation thing I was wondering about in the OP?

Ah - I deleted my earlier post because it seemed pointless after re-reading the thread!  Basically I have exactly that breadboarded up at the moment for experimentation; I'm using a REG710 buck/boost to get 3.3V from a LiPo cell.  Since it'll happily work well below the cell's minimum voltage and I didn't want to rely on the cell protection circuit, I'm driving the regulator's enable pin with an MCP111-300 so it cuts out at 2.93V; as an added bonus this means I can hang an LED between Vin and the enable line to act as a 'terminally low battery' indicator.

It seems to work very nicely; quiescent current draw is about 65uA when the regulator's up and running, or about 1uA when it's in low-battery shutdown (plus the current through the warning LED if fitted, of course).  Of course you'd need something a bit beefier than the REG710 (limited to 30mA), so that quiescent current figure is probably useless, but I'd guess it's a reasonable ballpark figure.

Regarding oscillation, unfortunately I haven't managed to characterise the behaviour in that much detail yet.  I've been doing my tests from a bench PSU which doesn't offer massively precise setting of the output voltage (just a single-turn pot), but I've verified it cuts off at the expected voltage on the way down and seems to have a small amount of hysteresis (about 0.1V) which should prevent any problems.  At the weekend I'll probably try it out on a real battery, but since the one I intend to use has a 2000mAh capacity and the entire system is drawing less than 1mA on average I shall have to come up with a safe way to run it down before I can get to the interesting section of the discharge curve!

The other part of the battery circuitry I've got in mind is a MAX1555 for charging; I'm not sure if it's quite what I need (to be honest, I'd intended to use the MCP73832 in the project I originally wanted the battery for!), but I needed something that explicitly supported leaving the system load connected whilst charging the battery.  Hopefully I'm not going to blow myself to smithereens with it...

[Entropia]

Thanks for the input. I think the only way to really find out how it goes is to try it out on a breadboard. :-) Regarding the LiPo charging, you might also want to take a look at the MAX1811 which I'm planning to use. It's a bit bigger (package-wise) and you can program the charging current with a logic input. It will let you power the system whilst charging as well, though you might want to keep tabs on the current consumption so that the charger doesn't end up supplying your complete circuit while no current flows into the battery.

[baljemmett]

Thanks for the input. I think the only way to really find out how it goes is to try it out on a breadboard. :-)

There's a certain amount of satisfaction involved in wiring it up for a test and seeing it does exactly what you wanted it to, as well! 

Regarding the LiPo charging, you might also want to take a look at the MAX1811 which I'm planning to use. It's a bit bigger (package-wise) and you can program the charging current with a logic input. It will let you power the system whilst charging as well, though you might want to keep tabs on the current consumption so that the charger doesn't end up supplying your complete circuit while no current flows into the battery.

I'll check it out, thanks -- I see it's designed to operate from USB, but I don't see any reason I couldn't hook its input up to a DC jack instead (or maybe it's time I bought a modern phone with a USB charger, hah).  I think my system load is negligible for all but one minute in 24 hours, so hopefully that won't pose an issue...

[Entropia]

I see it's designed to operate from USB, but I don't see any reason I couldn't hook its input up to a DC jack instead

No problem using a DC jack, so long as you keep inside the allowable input voltage range. See MAX1811 datasheet page 7.

[amyk]

amyk: I have the MAX1811 for charging, do I have to somehow control the discharging as well?

UVLO is good, and probably something to limit current too - a 800-1000mAh LiPo can deliver massive currents if shorted. A simple fuse would be good protection against that.

[annejan]

Would there be any problem using a XC6206P332MR 662K LDO 3.3 V regulator in a similar application?
They come in 250 mA and 500 mA flavour for less than 10 cents..
That combined with the max1811 would probably be cheapest way to do a 3.3V single LiPo sollution.
Does anyone see any pitfalls in that plan?

[poorchava]

MCP73832 is my part of choice for that kind of stuff. Cheap, available, does the job.