Should I arrange the cells in series or parallel?

I’m planning on creating a small 3.3VDC PIC-based device that runs on three AA cells. Current draw will be very low and I’d like to maximize battery life. The way I see it, I have two options:

  • Attach cells in series and use a 3.3V voltage regulator.

  • Attach cells in parallel and use a DC-DC converter IC.

Using a converter seems like it would be more efficient. However, almost every AA powered device that I can think of has the cells arranged in series. So I’m wondering if I’m overlooking something?

The problem with connecting batteries in parallel is that each cell may have a slightly different voltage and internal resistance. That usually means that one cell will discharge before they other and the total capacity is not as great as you wish. But it may not be that much.

Then there is the danger of a consumer putting one battery in backwards which causes the two batteries to be a dead short which results in rapid overheating, rupture and maybe even a fire.

Are you using rechargeable batteries? If so which type?

The simplest is 3-AA batteries (3.6 - 4.5V nominal) and a linear LDO 3.3V regular. Remember that the voltage the regular drops times the current draw is dissipated as wasted heat. So 3-NiMH (3.6V) would be more efficient than 3-alkaline (4.5V) batteries.

For best efficiency use a good DC-DC converter. They come as either down-converter (high to lower voltage) or up-converters (lower to higher voltage) and some will do both (great when the new battery’s voltage is greater than the regulated voltage but drops below the regulated voltage when the battery is almost depleted).

Grab a few data sheets of devices and batteries you are thinking of using and do the calculations of each combination (an Excel spreadsheet is good for this).

Hope this helps.

waltr,

Thanks for the detailed reply! I’m intending for the batteries to last for at least a year (hopefully longer) in the device. NiCd and NiMH cells have too large of a self-discharge rate for that, so I’ll probably just use alkaline cells.

I’ll dig around for datasheets.

It’s tough. I’ve tried fiddling around with this as well. Unfortunately all the HiMH batteries I’ve tested are actually over 1.2v, usually around 1.35v, so three of those goes way over the max most 3.3v devices have, which is 3.6v.

I’ve been using a single CR123A battery, which is 3.2v. I don’t have a one year required lifetime, though.

If you can find a way to have a decent battery around 3.3v without regulation, or just really low quiescent current, I’d like to hear it too.

Yes alkaline cells are good for that. Also consider Li batteries.

To get AA sized cells to last a year you probably also need to put the PIC into sleep mode when it has nothing to do. MicroChip does have some app notes on doing this.

TheDirty:
It’s tough. I’ve tried fiddling around with this as well. Unfortunately all the HiMH batteries I’ve tested are actually over 1.2v, usually around 1.35v, so three of those goes way over the max most 3.3v devices have, which is 3.6v.

Would using a diode to lower the voltage a little bit be a good solution?

olivier_p:
Would using a diode to lower the voltage a little bit be a good solution?

I would need something to drop at least 0.4v. 0.5v to 0.6v would be preferable. If you have a diode that would do that, it would work. The general purpose diodes I have here only drop about 0.1v each.

I’m not sure what the objection is to using an LDO regulator. There are plenty available that have a very small quiescent current.

If you really don’t want to use a regulator, try using just 2 AA Alkaline cells in series. This won’t exceed the max voltage, and most of the components probably work down to 2V or so anyway.

The reason for an LDO regulator is that its output will remain regulated with a lower input voltage, ie battery voltage being near fully discharged.

Another is using a lower initial battery voltage so that there is less Vdrop times current power loss through the regulator.

Dropping the voltage with a diode is the same as using a regulator except for the small amount of quiescent current the regulator needs to operate. Any voltage drop times the current passing through it is power dissipated.

If the circuits will run properly over a range of voltages than direct battery power could be the must efficient. This is the biggest problem with batteries, their voltage varies with state of charge so the voltage regulation (or no regulation) design needs to be thought out carefully.

MagicSmoke didn’t say which PIC he plans on using or at what clock speed it’ll run at. Many of the PICs need a minimum voltage at higher clocks speeds which also draw more current. The minimum voltage and minimum current is at low clock speeds.

Walt, I should have been clearer - I like the idea of using an LDO, but I can’t figure out why you’d be talking about using a diode to drop some of the voltage. If you’re using 3 cells, put an LDO; if using 2 cells, no need to regulate or drop the voltage at all.

With the minimum voltage, yes, he will need to check the frequency - voltage curves from the datasheet. As for the reducing voltage as the batteries deplete, the Brownout detector should be programmed to a suitable level & he should be fine…

Someone else mentioned using a diode and I was just commenting on that.

The original objective of this thread is really ‘how to get the longest battery life’. This can be done in a number of ways but does take some analyzing and decisions about everything from the power source to the uC code.

It been a good thread so far. Hope more details are presented so the thread continues in a productive way.

cheers