The question really is: Are the light source(s) actually efficient enough to overcome any other inefficiencies in the rest of the circuit? And do so inexpensively? And reliably?
To begin answering the questions above, I referred to the Radio Shack and Wikipedia sites for LED diodes and AA batteries.
A) I'll assume the Radio Shack red LED has a nominal forward voltage drop of 1.3 VDC and draws (worse case scenario) 20mA of current.
B) The small, high efficiency solar cell gives out .3 A and .55 VDC in the noonday sun.
C) So, 6ea. cells (Radio Shack #276-124, ~=$30.00) in series give 3.9 VDC @ .3A
and 3 LED's in series should require 3.9 VDC @ ~20 mA. 5 strings of 3ea. (15 total ~=$14.00 + tax) in parallel should require .3A @ 3.9 VDC.
The solar cells should be driven to their power limits by the LED's.
And, the LED's, placed close to the cells, should irradiate them pretty brightly, assuming the cells can accept red light wavelengths efficiently.
D) Another way to do this is: Take one of the encapsulated solar cells (Radio Shack #277-1205, ~=$15.00) and power two red LED's (=$1.70) with it.
This is less money, but makes for a lower-powered circuit as well.
The power requirements of the batter(ies) and its dedicated cell(s) will be addressed in due course.
Okay, here we go:
Industry standard NiCd batteries will be considered because they're designed to be recharged; it's just that without a trickle charge to maintain them, they'll lose about 1% voltage/day.
The usual power rating for 'AA's is either 600 mAh or 700 mAh per battery. Assume, by independent reports from users, tat the Radio Shack 6 VDC encapsulated solar cell give about 20 mA of current under a light bulb, then _ think one solar panel will charge 5 'aa's in series but take at least 6 days to charge 3Ah batteries at .02 mA of direct current. By its Ah rating (50mA), it would take a little over 2 days.
That's assuming the batteries were downright dead flat to begin with.
To prove at least a COP=1.0, the batteries might be fully charged and then be maintained at a float, trickle rate by the solar panel. One could always add more LED's to that part of the circuit in order to increase power level by panel luminous irradiation, thus allowing an extra load to be applied to that part of the circuit.
Okay, the descriptions should be done with. I have a very good place in my apartment to experiment in, since I don't use my stove oven for anything to cook in. It's also light sealed from the door's inside surface and it's grounded, in case I ever need that.
Parts shouldn't be terribly expensive and are widely available. I get paid at the first of next month, but a medical bill is looming on the calendar horizon. The circuit I'd use would be substantially like the one in the YouTube video at the top of this page.
Questions? Comments? Append them to the end of this thread by posting them, as usual.
--Lee
Reedit, 26 Sept '11: Spell check.