N00b question, so forgive me if this is something simple.

What I want to do is have two (or more) cells available, but have a circuit pull from only one cell at a time, switching to the next fresh cell when the currently used cells voltage drops to a certain level.

The purpose is to have the convenience of replacing the batteries less often.

In other words, three cells present, A, B, C. Circuit runs for x hours until A drains to certain level, circuit then automatically switches and starts pulling from cell B, and so on.

Is this possible? If so, what are the basic here?

Thanks!
Bob H

 

a voltage detection circuit to monitor the battery in use, as well as a switching circuit possibly a diac/scr combo circuit. the offline batteries could be trickle charging. by the way an auto battery has 6 cells.

 

You could use a microcontroller's A/D input (or use an a/d ic), couple that to a quad analog mux to give you four or more inputs. If each battery is less than 5V (for some a/d this is even less) you can use it directly. If not you'd need a step down for each cell. By scanning each voltage with the A/D you can see immediately which is high and operate a relay or transistor switch to change to the highest cell. This can be pretty fast and will evenly discharge the batteries. I think you could actually do this with opamps but am blurring on how.

You could also use something like this to do the sensing. It's I2C so you can daisy chain a bunch of them with the same digital i/o lines.
http://focus.ti.com/lit/ds/symlink/bq27500.pdf

 

Lets say you have 15 batteries, and you want the load to be
powered up about 15 times longer than it normally would on 1 battery.

How about adding 15 diodes (1 in series with each battery).

I think if these batteries you're wanting to use are high enough in
voltage, then the voltage drop across the diodes shouldn't bother you much.

This solution would differ slightly from what you asked for (because all
batteries would be depleted in unison), but it sure would be easy and cheap
to implement.

The main drawback I see about my idea is that extra wattage would be
wasted at the diodes - but again I offer how easy it would be to wire it
up for a test run in your system.

 

The purpose is to have the convenience of replacing the batteries less often.

Thanks!
Bob H

Any way to run them in parallel?

 

Yeah, what Mrwassman said.

Simply running them all connected in parallel will result in the least waste of power - that is, if all of the cells are in good condition with similar levels of charge to begin with.

What kind of cells are you talking about? Dry cells? Rechargeable NiMH? Lead/acid, as in automotive/marine batteries?

If you're considering using lead/acid, you should use deep cycle batteries, as standard automotive batteries don't stand up well to deep cycle usage; the plates are much thinner.

 

The question of system voltage is an important one (and I think we may be assuming too much about the abbreviation auto)

Did Bob mean automatic or automotive ?

If these "cells" are each of very low voltage then scrap my diode idea (it won't work).
But if they each are say over 10 volts, then I think the diodes might offer an advantage of allowing the "cells" to be at random states of charge before connecting a load.

That's one diode in series with each battery, and all battery/diode circuits in parallel.

If Bob is following along, I hope he'll let us know the battery voltages and currents being drawn by the circuit load(s).