Feasible

What a good idea and a refreshing new approach. Best are LED's in the wavelength range of say 400-500 nm (nano-meter or 10^-9 meter), which is ultraviolet and blue, and 650-690 nm, which is red. This would give near 80% or more conversion of light energy into plant growth. In case the process needs some high and low frequency, we could mix say 60% diodes at the shorter wave length and say 40% diodes at the longer wave length. This is assuming cost is not a factor.

If it takes only one sec/diode to solder them together this then is 39 days of 8 hr/day labor for each person. A population of 300 would take 31 years for one person to solder this up. If it takes 2 sec/diode then it will take about 2.5 months of labor for each person and 62 years for 300 people. Finding LED arrays would defiantly save time. Just something to consider if it works.

LED's work in a limited current range and in most applications need a currant limiting resistor. I figure voltage fluctuations could be up to .3 of applied voltage. For example: voltage regulators for 12 volt car batteries can allow charging voltage to get up to 15.5 v during full charging. During a discharge cycle one might let it go as a worst case to say 10.5 volts. Thus one would need to limit current in these LED's for a voltage charge of approximately ..3 of maximum applied voltage. The internal resistance of the batteries or filament type light bulbs could be used to limit the current.

Offered by Mike.