SparkFun Forums

[hpk]

I'm hoping to start a project that will consist of a 802.15.4 star network with the hub powered by USB and the end devices running off of LiPoly batteries. Rather than do the typical 802.15.4 thing where nodes sleep for a long time and only transmit bursts once and a while, my application will have the end nodes communicating fairly constantly (100Hz or so?) for at least an hour or two, hopefully longer.

I'd like to come up with an estimate for battery lifetime of the end nodes. Is this as simple as taking all my components, adding up their current draw, and dividing the battery's rating (eg, 1000 mAh) by this number to find the number of hours I can run on a full charge? Or is the process more complex than this?

I have a lot of experience working with micros, but have never done anything with batteries or wireless before, so any advice would be helpful.

Thanks!

[Philba]

Yeah, it's pretty much addition but you've got to find everything that draw power. Micros can be especially tricky because you have to know about every single option you enable or disable. For example, turning on weak pullups will increase consumption. You need to measure the real HW in a realistic scenario. There are a lot of things you can do to decrease consumption.

I'd put a spare DMM to measure the current of your dev target. You will learn a lot.

[bigglez]

I'd like to come up with an estimate for battery lifetime of the end nodes. Is this as simple as taking all my components, adding up their current draw, and dividing the battery's rating (eg, 1000 mAh) by this number to find the number of hours I can run on a full charge? Or is the process more complex than this?

A battery life estimate is just that - an estimate.
You can take the calculated route, but there will
always be some 'fudge factor' that defies calculation.

A pragmatic approach is to life test the prototype
and measure the useful battery run time. This
should be done several times with several batteries
and at both hot and cold temp limits to be accurate.

A good starting point is to examine the battery
data sheet. In particular, watch for discharge rates
(usually curves). A battery with a light load will run
longer than the same battery on a heavy load.

A very quick experiment is to load the battery with
a resistor (scaled to the intended load) and run a
timer until a predefined end point is reached (the
"under-cut" voltage). Further discharge will actually
ruin the battery.

Consumer goods with batteries (such as cell phones
and camcorders, etc.) embed a coulomb counter in
the pack along with a uC to report the charge remaining.

Those same goods typically use an optimistic estimate
based on minimal optional power draw (such as not
zooming the lens or setting the LCD backlight to
turn off if inactive).

YMMV.