In theory, "one motor for 2500/400 = ~6 hours" would be correct. In practice, it's not quite that easy due to factors such as the battery's discharge curve, rate of discharge and such. But's close enough and for the most part gives a relative picture.
I just rule of thumb it and shoot for about 50% head room. I'm going to see if I can design a multi-charger like I described in my prior comment. It could have nine battery connections to charge all nine individually, but discharge them in series/parallel for 12V out.
Max - My biggest fear with putting LiPoly batteries in series is the risk of fire if they aren't charged or discharged evenly. It may not be a rational fear, but none the less, it makes me nervous.
The charger chip I've been sing is a 3mm X 3mm part with almost no external components needed. I should look at building a triple charger to mitigate the uneven charging risk.
I'm not exactly sure how to put it all together though. The power in to the charger chips would have to be parallel, but the Vout from the battery would need to be in series. Has anyone reading here done such a thing?
@Duane: How about stacking a set of three in series to get 12V, then connecting three of these stacks in parallel to give me some "oomph" -- all of these could be located in the "engine room" -- then use this "super pack" to power everything?
@Duane: For my smaller bots, I tend to use LiPoly, like this one from Adafruit (http://www.adafruit.com/products/328). I've desined a charger chip into some of my robot boards so it's pretty convenient. But they're 3ish Volts, so won't do for 12 Volt motors.
These look very tasty -- certainly small enough to fit in my "engine room". The Adafruit page says "The output ranges from 4.2V when completely charged to 3.7V" -- so I could strap three in series to give me the 12V required. So I could have theee sets of three -- one for each motor.
The thing that scares me is how much I've forgotten over the years. For example, the Adafruit pages says the battery has a capacity of 2500mAh. My motors draw ~400mA. Does this mean that three batteries (a set of three stacked in series) can power one motor for 2500/400 = ~6 hours? That seems to be pretty impressive.
Actually, the Adafruit page says "This battery has a capacity of 2500mAh for a total of about 10 Wh." Now Power (in Watts) = Current x Voltage, so what does "2500mAh for a total of about 10 Wh" actually mean?
Max - I'd try to avoid separate batteries for the different systems. Pololu, whom you mentioned, has some innenxpensive DC/DC converters. This one here (http://www.pololu.com/product/2110) will give you 3.3V or 5V at 3.5 Amps.
It's bigger brother (http://www.pololu.com/product/2111) does the same at 7 Amps. The also sell a number of smaller DC/DC converters that will fit into a standard three terminal "7805" regulator foot print. Use one of those with a 12 V main battery and you'll be all set.
I use 12V SLA (sealed lead acid) for any of my larger robots - anything that might be drawing an amp or more - because they're pretty safe and simple. They are heavy and bulky though.
For my smaller bots, I tend to use LiPoly, like this one from Adafruit (http://www.adafruit.com/products/328). I've desined a charger chip into some of my robot boards so it's pretty convenient. But they're 3ish Volts, so won't do for 12 Volt motors.
I would look to the world of radio controlled vehicles for battery packs. My first inclination would be to power the electronics from small li-po cells for weight and space, but I'm not sure how well mixing batteries does.