SparkFun Forums

[alexc]

I am designing the power system for a board that will run on a 3.7V lipo battery, all of the IC's are 3.3v; naturally I want to squeeze as much battery life as possible. The power system includes a charger for the lipo battery and which also allows the board to be used while plugged to a usb port (for debugging or using while charging).

I expect most of the IC's will be on power down mode for 70% of the time, and active for 30% of the time. When the system is active/inactive is controlled by external interrupts and not by a user controllable switch. The maximum current draw when active is 250mA, but for the most part while active it should be around 120mA while active; when in power down mode the current draw is of 30uA (last sentence edited for clarity).

I've been debating with a few choices.

1. Should I use an LDO regulator or a buck-boost regulator?
2. Should I disconnect the battery from the load when operating with USB power?

The MPC1700 is an LDO regulator with a very low quiescent current of 1.6uA and a voltage dropout of ~200mV. Problem is, when battery drops below a certain point the whole system will shut off even if the lipo still had some life left in it. An alternative would be the TPS61200 which is a buck-boost type regulator, it has a higher quiescent current of 55uA but will let me use the whole battery range. Unfortunately, in that case I would have to worry about under-voltage protection and the added components will also consume power.

Here is the circuit I came up with (open in new tab since it gets chopped by the forum)

Thoughts are welcome.

[uChip]

Your circuit looks fine, but with a power down current of 120mA the 55uA of the regulator is not going to make a significant difference. On the other hand the buck boost regulator is overkill for this application since the regulation voltage of 3.3v plus drop out 0f.2V is pretty close to the bottom of the battery voltage. Instead of squeezing the last 100mV from the battery, you'd get better battery life by finding a way to reduce your power down current.

Good luck with your project!
- Chip

[alexc]

Thanks chip.

I realized my post wasn't clear, when active the current will fluctuate between 120mA and 250mA (mostly 120mA). But when in power down, the current consumption is of just 30uA.

As far as I understand, the recommended limit for most 3.7v lipo batteries is 2.8-2.9V. Won't the LDO stop working @3.5V? (which means I only get to use the battery while in >= 3.5V, instead of >= 2.9V). Do you feel this is negligible? I am asking because I have no experience with lipo batteries.

[jremington]

Most LDOs don't stop working when the input voltage drops below the minimum required, they just track, and I could not find anything in the MCP1700 data sheet that indicates differently. So you still need to worry about battery protection.

This step up/step down regulator seems like a reasonable choice: http://www.pololu.com/product/2122
You can buy battery protection PCBs quite cheaply from all-battery.com http://www.all-battery.com/batterypcbpcmbms.aspx

[uChip]

My experience with the batteries available from SparkFun (e.g. https://www.sparkfun.com/products/341) is that they have an internal cutoff around 3.3 to 3.4V and they roll off very fast after about 3.7V.

Here is a graph of a lipo discharge I measured.

BatteryGraph.PNG

[alexc]

Thank you chip!

I think that plot is the most useful piece of information I have come across! I now understand your point about LDO vs buck-boost, the buck-boost will only operate ever as a step-down regulator anyway, so its not worth it.

If it is not too much to ask, could you also comment on the necessity of a mosfet to disconnect the battery from the load while charging? I've seen most designs don't include this, so I am not sure if it is overkill or if it is worth the effort.

[alexc]

Most LDOs don't stop working when the input voltage drops below the minimum required, they just track, and I could not find anything in the MCP1700 data sheet that indicates differently. So you still need to worry about battery protection.

This step up/step down regulator seems like a reasonable choice: http://www.pololu.com/product/2122
You can buy battery protection PCBs quite cheaply from all-battery.com http://www.all-battery.com/batterypcbpcmbms.aspx

Thanks @jremington, I had seen that pololu regulator before, unfortunately due to size constraints I have to roll my own. Aside from the ready-made convenience, could you comment why you would prefer the pololu step up/down regulator from a custom built around the TPS61200?

I appreciate the warning regarding low voltage with LDO's, I was just assuming they shut-off on their own. I based this assumption based on the observation that all products I've seen (including in sparkfun) that use lipo batteries and an LDO regulator have no undervoltage lockout circuitry. Whenever I find a schematic that uses a lipo and has an undervoltage lockout circuitry, it is using a buck-boost regulator. For example look at the arduino fio board schematic http://dlnmh9ip6v2uc.cloudfront.net/dat ... io-v30.pdf

PS: However, I have to say that Chip's plot has almost convinced me that fiddling with a buck-boost isn't necessary, given that I will be using sparkfun lipo batteries. It also explains why most circuits don't have undervoltage protection, this is included in at least some lipo batteries, perhaps most of them (I have no idea).

[uChip]

On the subject of the MOSFET, I would say that yes, it is needed. The 73831 has no load current management so trying to charge the battery and drive the load at the same time can confuse the charger. Alternative to the MOSFET you should look at the 73871 which has load current management built in.

[jremington]

Many LiPos do NOT include cutoff circuits (I personally have several such) and in that case, the battery will be destroyed if the voltage is allowed to drop below about 2.5V.

I prefer Pololu products for three reasons: they seem to be among the best engineers around and they extensively document their work. They also offer an extremely responsive customer support service -- probably the best in the entire hobby robotics business.

[alexc]

@uChip, thanks for your input on the mosfet, I'll take a look at the 73871 as an alternative.

@jremington my experience with pololu has also been very positive, I love their orangutan series, and I am a big user of their ISP programmers (cheap, small, and customizable). That being said, in this case I can't go with the ready-made solution, and one thing I don't like about pololu is that they not always provide schematics for their products (but I understand that may be motivated by economical reasons to prevent cheap knock-offs on ebay).