DIY LED lighting brighter than Solaris? Prove me wrong!!

jma1978

Premium Member
Ok, so I've been thinking about this for a while and have read everything I can find about this. I don't see why it won't work; but I don't know why nobody has done it yet, either.
Here's the plan:

An array of 10mm LEDs over my 70 gallon bow front. I've seen people try the 5mm LEDs in the past, but give up because it's too expensive. Well, I see the prices have come down a lot and the brightness of the LEDs has gone up! BestHongKong seems to have the best prices. I've heard the brightness fades after about 1000 hours, but that was with overdriven LEDs.

For example, I can get 130,000 mcd white (7000K to 10000K range) White LEDs and 40,000 mcd blue Blue LEDs.

1000 mcd equals 1 candela equals 1 lumen. So.... each white LED equals 130 lumens and each blue is 40 lumens. With LEDs this bright, you wouldn't have to overdrive them, and they should last the rated life.

They have a 12 degree viewing angle, so at 24", the spot would be 5.1 inches radius.

I want to use this board for mounting them:
dark-4_b.jpg

Available for $5.59 .

I'll mount them 0.8 inches apart. This should give me great overlap, and the beam (at 12 degrees) will touch the one next to it at a water depth of 3.75".

I'll use these drivers:
DXD60U.jpg

They cost $19.99 at BallastWise

So my 36" X 12" array will have 1734 LEDs and will draw 141 Watts, which means I need three of the drivers. This is with 100 ohm resistors in series with 3 LEDs. You can see the schematic here: LED array wizard (source voltage is 12V, diode forward volatge is 3.4V, diode forward current is 20mA, number of LEDs is 1734).

So if I use a ratio of 3 blue to 1 white (just to even out the intensity difference, and because Solaris seems to favor the blue spectrum), then I'll have 1301 blue and 433 white. This will be a total of 52040 lumens from blue and 56290 lumens from white. Since my total area is 3 square feet (.28 square meters, round up to 0.3), I can now calculate lux. 108,330 total lumens divided by .3 meters equals 361,100 lux. Seems like quite a lot!!

Yes, it's a lot of soldering, but I theoretically have a huge amount of light for only 141 watts, and very easy to cool (just blow a fan over the top of the boards.

So here's the cost breakdown:
LEDs: $153 shipped for White. $271 shipped for Blue.
Driver: $100 for three shipped.
Project boards: $44.72 for eight shipped.
Solder: cheap
Solder gun: already got
Wire: cheap, because I'll use the LED leads for most of it.

Total: $568.72, plus labor and solder. Labor is priceless to me, so I won't even bother factoring it in. But just for fun, figure 3 LEDs per minute, equals 9.6 hours. Whew!!

So for $568, I have a 36" X 12" led light that is way brighter than Solaris. I can dim the arrays using my X10 controller and program it for sunrise, sunset and moonlight.

Ok, now everybody, tell me why it can't be done!!
 
you will want to incorporate some kind of reflector to get the most out of your investment as well add some more cash for the cooling fans and power supply
 
And if you want to make one for me maybe we can work something out. Let me know. Thanks.
 
Just make sure before you tell everyone how great it turned out, you have PAR ratings to back up your statements :)
 
Thanks for crushing his dream Rich! lol, anyhow, i was wondering the same thing about teh LED's. They have come down alot in price and there doesn't seem to be much to it. Get alot of them in the right color and have a good reflector, fans, mount them and fire em up. Much like a retro for T5's or Mh's. Am i missing something as well?
 
Ok, thanks for the help. The error in my math does kinda rain on my parade (I blame google calculator for it!). It seems the 1:1 conversion is for a sphere, so you need to figure the 12 degree beam to get the true number. So the new calculation gives me 12,426 lux, which is much less impressive. Although if I change my ratio of 3 blue : 1 white and instead use 1 blue : 3 white, then I end up with 21,393 lux. Still lower, but much more believable and realistic now.

As far as PAR and PUR, I've seen people arguing that on the Solaris thread, and nobody seems to know for sure how they factor into LEDs.

The fans will be 12V computer fans, so they'll be powered by the LED driver and will blow over the top. Since most of the heat travels to the leads, I think having the fan blow over the leads should cool it down...

A reflector is not necessary, because the all the light is already pointing down. The high power LEDs do need a reflector and also produce much more heat. That's why I was thinking these low power LEDs might be better.

Another approach might be to calculate my lumens per watt. I come up with 6418 lumens divided by 141 watts equals 46 lumens per watt.

I think a MH bulb puts out about 80 lumens per watt, so now it's not looking so good.

But what about that beam angle thing? Since a metal halide is radiating light around 360 degrees, you need to figure how much of it actually is radiated down into the tank. I hear that the good reflectors have 90% efficiency. What does that mean?
 
Rich,
That's why I was planning to use the 10mm LEDs. They only draw 20 mA, which is considerably less than the 700-1000 mA of the high power LEDs. I'm just not sure of the intensity of the 10mm LEDs.

All the threads I can find where people have tried it, they used 10,000 mcd LEDs, then gave up and went to high power LEDs (and high price, high cooling, etc). But now the low power LEDs are rated at 130,000 mcd. So are the 130,000 mcd LEDs really 13 times brighter, or is the manufacturer stretching the truth?
 
But that's 250 Watts and the coverage area is probably not nearly as large. Plus now I'm using .0288 as many LEDs (50 instead of 1734), so each LED needs to be 34 times brighter. Is a 5W LED 34 times brighter than these 10mm LEDs?
 
Back
Top