Detecting Remaining Battery Life

[ElijahRockers]

Sorry for not using the question template, but I don't have any relevant information, like equations or solution attempts at this, since I literally have no idea where to start. Google has failed me.

This isn't necessarily a homework problem, but it is for part of a project I am working on. Long story short, I will be very happy if I can find some way to monitor the remaining battery power on a 15000mA battery, two 5V outputs (one 2A and one 1A) and one 12V output. If I can build some relatively basic circuit to detect this, then hopefully I can use a microcontroller to transform an analog signal into digital data, so that eventually I can have a digital readout of the remaining battery life.

I have been told by a professor that this is extremely difficult, to the point that we should give up on it... but I am not quite ready to do that, so I have come here to seek a second opinion.

Thanks!

 

[rcgldr]

Depends on how accurate the remaining lifetime needs to be. The voltage versus remaining energy curve is initially somewhat steep, then becomes nearly linear until the battery is nearly depleted. The approximate remaining liftetime could be based on this curve, even though there may be variances between batteries and/or dicharge / recharge cycles. The curve is also affected by the load (the amount of current), so assuming you take that into account, it's possible.

I have an electric shaver that displays the number of minutes remaining when not under a load (when it's off), and while in use, it just displays a timer for time used. It's old enough that it's overestimating the number of minutes remaining, and occasionally goes into cycles where the number of minutes remaining cycles downwards even though there's no load, when the battery is getting low on charge.

 

[berkeman]

Sorry for not using the question template, but I don't have any relevant information, like equations or solution attempts at this, since I literally have no idea where to start. Google has failed me.

This isn't necessarily a homework problem, but it is for part of a project I am working on. Long story short, I will be very happy if I can find some way to monitor the remaining battery power on a 15000mA battery, two 5V outputs (one 2A and one 1A) and one 12V output. If I can build some relatively basic circuit to detect this, then hopefully I can use a microcontroller to transform an analog signal into digital data, so that eventually I can have a digital readout of the remaining battery life.

I have been told by a professor that this is extremely difficult, to the point that we should give up on it... but I am not quite ready to do that, so I have come here to seek a second opinion.

Thanks!

It also depends on what kind of battery you are using. Is it lead acid (it probably is, based on the size you quoted), or some other battery type?

There are "battery gas gauge" circuits and applications that can help you accomplish this function:

https://www.google.com/search?q=bat...CPOHFigKO3YGwBg&ved=0CGgQsAQ&biw=1082&bih=789

You can also use your uC to monitor energy useage (and recharge cycles), and estimate the remaining bettery energy programatically.

Have a look at the google images links, to see if any of those gas gauge circuits and applications look like they might work for you.

 

[ElijahRockers]

It does not need to be terribly accurate, but it does need to be able to to detect when it's getting low, say 10 or 20%. The battery is Li-ion rechargeable. I'll take a closer look at some of those circuits soon. Unfortunately I'm more on the digital design and programming side of things so it might be a bit. Thanks.

 

[CWatters]

If this is for a one off project perhaps try calibrating the battery. eg measure the discharge curve and set threshold voltages accordingly.

In the real world/mass produced products there are several problem to overcome. For example battery voltage may depend on the temperature, or age of the battery. If the discharge curve is very flat a tiny change in voltage due to temperature change may imply a large change in charge state. Manufacturing tolerance may also be an issue - I was once told that for NiMh cells capacity can vary by 20% batch to batch.

I might be wrong but I believe modern battery monitors in laptops etc try to measure the capacity during charge and discharge cycles to calibrate the meter. Then they measure the flow of current into and out of the battery to try and estimate charge state. Not easy as people don't always fully charge/discharge a battery.

 

[ElijahRockers]

It is indeed for a one-off project. The battery has three (or four, can't remember now) small blue LEDs on it that act kinda like a battery life gauge. As far as measuring, what exactly am I measuring? I know I can measure amperage, but shouldn't the output voltage stay relatively the same during battery discharge? The power guy in our group said he'd have to rip apart the battery and build a circuit around each cell in the battery, but that sounds a little excessive.

 

[skeptic2]

There are several approaches you can take. I think taking a voltage measurement of a lithium battery is the least practical. If I remember correctly the voltage of lithium batteries is constant until the very end of its life.

A better approach might be to monitor the current drawn from the battery and trigger your lights based on the amp-hrs used as they approach the amount that you've determined to be the life of the battery.

 

[CWatters]

This diagram illustrates one possible problem with just using the battery voltage...As the battery warms up the voltage varies quite a bit, possibly fooling meter into thinking the battery is empty or at least messing with the accuracy quite a bit...

http://blog2.optibike.com/wp-content/uploads/2009/12/lithium-battery-temperature-vs-capacity.jpg