I am (attempting) to create a device the size of a credit card, which will need 5v and 12v from battery power. The power requirements for the 5v will need to run an avr + a couple led's (20-30ma max I would say), and for 12v 1-2ma should be plenty. The device will normally only run for 5-10 seconds at a time, and probably just be turned on with a momentary switch (powered as long as switch is on). Usage would not be very frequent.
Cheap as possible is good. Low parts count is good. Through-hole parts is good.
After much messing around with coin cell lithium batteries, button cells and small alkalines like A23's (~12v), switched capacitor doublers, voltage triplers, reading battery datasheets, etc., I have finally decided (I think) to just use a single AAA alkaline and a dc switcher. The AAA has enough power to run a switcher, is common and inexpensive, the battery holders are cheap enough, and it will nicely fit on the credit card sized pcb.
Since I want through-hole parts, that eliminates lots of switchers. Since I want cheap, that eliminates lots more (I don't want to pay for a switcher that costs more than the avr, and sometimes its double the price of an avr). Since I want to run from 1.5v and below, that eliminates about the remaining ones left.
Somehow I ended up finding the TL499A from TI (cheap, and available in pdip, down to 1.1v input). It seems to be working as planned so far (have not powered an avr with it yet). This is what I have wired up so far (this was from memory, so it may not be exactly correct)-
http://www.mtcnet.net/~henryvm/T...
I can power led's from both outputs (15ma/2ma) for at least several hours from unkown AAA's just laying around, so at least I'm heading in the right direction.
Its all standard datasheet values/connections, except for my added D1/D2/C3/C4 doubler and C2 being 100uf instead of 470uf. I have R3 adjusted (a 100k pot) so I can have ~12 at V2 (V1 is then 6.? volts), so R3 value in schematic is not actual value. The zener business at V2 output is just something I tacked on the schematic but have not tried (limit to 12v/2ma with 14v in if I am thinking correctly).
Anyway, the plan for now is to have the TL499 ref from the V1 output so as to make V2 ~12 volts. Then drop the ~6.? volts down to 4.5-5.5 via a diode or two. I will also need to control V2 on/off, so that may change what I need at V2 (and change what I need to do to drop V1).
If anyone has suggestions, comments, improvements, hints, tips, etc., fire away. I could be doing something obviously dumb here, and would like to be kept on the straight and narrow path.
(I'm not thrilled at what happens to the voltage output for 100-200ns when the switching takes place, but maybe that's normal, or maybe I just put my scope away and just use a multimeter)