Since I've gotten a few requests to add this information to this thread here goes.
I've been playing around with the DIY CREE solderless LED kits for a few months now and have been really pleased with the results, so much so that I'm starting to replace T5 fixtures with LEDs when the bulbs burn out instead of just replacing the bulbs.
A few things to keep in mind
LEDs CAN PRODUCE A HUGE AMOUNT OF HEAT
Not all LEDs are created equal, the ones you see at the chain fish stores/hardware stores are not the same as the ones I'm talking about here
LED heat is directed into the heatsink and not so much into the plant area
LEDs are spotlights
LEDs are expensive short term, cheap long term
I recently setup a 40 breeder nursery tank
http://www.terraforums.com/forums/showthread.php/135880-Terrarium-tank-stand-build-log and when it came time to figure out what type of light I would use here is how I ended up with my answer.
From past experiences
http://www.terraforums.com/forums/s...le-Short-Term-LED-Experiment-with-a-few-Cephs and
http://www.terraforums.com/forums/s...uild-a-Cephalotus-grow-space-would-love-input! I "feel" that 8 CREE or similar LEDs powered at 350mA provides slightly more light than a single T5HO bulb. I came to this assumption watching Cephs/Helis coloration changes when moved from a T5 setup to a LED setup. Without expensive test equipment and more knowledge than I have I'll just have to be happy with my assumption.
Hobby light meters are basically useless for measuring plant light since they are only measuring the spectrums they are calibrated to measure. Basically if you take a light meter and measure a cool white LED then measure a blue/red combo LED the reading would say the white LED was something like 50X brighter than the blue/red combo. But to the plants they are much closer.
I'm using a 48" T5 bulb as the basis for all other lighting as I feel it is the most cost effective method out there besides LEDs today.
Running with the assumption that 8 CREE LEDs = 1 48" T5 I can start to do some basic math.
48" / 8 LEDs = 6" per LED - Now figuring that LEDs are spotlights I "feel" that is too large a space between LEDs to provide even coverage. 4-5" would be be better in my option, even down to 3" if you really wanted a lot of light.
Add some readings from a power usage meter.
Moonlight Kit 4 LEDs running for 12 hours read .06 KWhrs = 5-5.83W (1.25-1.46W per LED)
28 LEDs dimmer set at about 50% running for 5 hours read .19 KWhrs = 38W-40W (1.36-1.43W per LED)
2 single bulb T5 fixtures running for 2 hours read .19 KWhrs. = 47.5-50W
Here are the electrical costs for me @ $.114 per KWH
Running 2 48" T5 12 hours a day (48.75W*2)*12 = 1.17KWH = $.133 per day running costs or $48.68 per year.
Running 16 LEDs @ 350mA 12 hours a day (1.355W*16)*12 = .26KWH = $.03 per day running costs or $10.83 per year
So running T5s would cost me an extra $37.85 per year in electrical costs. This is a very important number as we can use it to figure out the "break even" point for the LEDs.
For a while you could get a 4 bulb 48" T5 fixture for $60, I will use that example even thou I was not able to locate that fixture for sale a few weeks ago.
Setup costs for the T5 system would be $60 (bulbs included/shipping included), it was/is a great deal.
Setup costs for the LED system would be $153
Using the break even number we can figure out that the LED system would be $93 more right out the gate. So the extra setup costs of the LEDs divided by the extra electrical costs of the T5 system gives us the # of years it would take to break even. $93/37.85=2.46 years
At 3.14 years running the T5 system you have paid enough in electricity to buy the LED system.
At 4.04 years running the T5 system you have paid enough in electricity to buy the LED system and run it for those 4 years. This bit of info is useful if you already have a T5 system running.
Since I'm in this for the long run, T5s just don't make sense anymore for me. The main exception I can think of is if you wanted the heat from the T5 for the plants/animals in the terrarium.
This is a very simplistic example.
I'm not counting the extra learning/setup time the LEDs require
I'm not counting any mounting required by either system
I'm only using a single LED (XP-G) solderless DIY kit system for an example, as I have no other LED knowledge
I would use 20 LEDs in that example instead of 16 due to LEDs being a spotlight and my desire for more even coverage
Electrical cost vary from state to state, higher electrical costs would reduce the break even point, lower costs would raise it.
I never figured in replacement of any T5 bulbs
Hope this is helpful to someone. YMMV