LED for main lighting?

[utahsaltreefer]

In researching this here's what I found. Please correct me where I've made mistakes.

1 250 watt MH 6500K bulb is approx 14000 lumens
1 bright white LED is between 5000-10000 mcd

1 mcd = (lumens*1000)/12.57 so
140 8000mcd white LCDs = 1 250 250 watt MH 6500K bulb

The cost of the LEDS is about $1.49 each so it would cost $210 for the bulbs which is way more than a MH bulb. However the power consumption may mediate this. Ignoring loss due to the transformer:

1 LED uses 20mA and 3.3V
Since a watts are volts times amps I figure that
140bulbs * .02A * 3.3V = 9.24 watts

According to the manufacturer the LEDs keep their potency for about half their MAX life which is 30000 hours. So 15000 hours / 12 hours per day is about 3.5 years of life. That would use about 3600 less kWh in its lifespan. Even at $.05 a kWh that would be $185 less energy. At higher energy costs it would be even more beneficial.

The image I'm attaching from Super Bright LEDs shows the wavelength distribution for the bright white LEDS.

 

[taikonaut]

I have seem brighter LED with the same current and voltage.

At any rate, one important factor that missed by LED people is that the beam of a LED is focused (it got build-in reflector), while the MH is almost point source. This will give you big issue when you try to convert from lumen(lm) to candela (cd), since cd is directional, but lm is not. You see engineers convert one from another, but intrinsically, you can't since they are different unit.

It is the right idea to use lm/W for measuring efficiency, and the number is around the following:
incand. bulb: 15 lm/W
MH bulb: 50 lm/W

LED and PC is between the two, largely depends on color. Red LED probably has around 40lm/W, and blue/white is less efficient.

With that say, you will realize that you can't save much power when replacing MH with LED, since MH is still more efficient than the semiconductor in generating light from electricity. The real neat thing about LED is the smaller package and lower heat, which can be very useful in small (nano) aquariums.

 

[utahsaltreefer]

Thought I'd post the specturm analysis for the blue bulb from Super Bright LEDS also.

 

[sl33stak]

ok, so ya have MY ATTENTION NOW!! I like the idea of that! I have read on other forums people who use LED's for moonlight, and they say if you take a peice of sandpaper and just "rough up" the tip of the LED a bit it will help to diffuse the light so as not to create such a "pinpoint". If you go ahead with this project PLEASE keep us updated!

 

[utahsaltreefer]

I've ordered a handful of 10000mcd LEDs to do some tests with. I'll do some tests on a small 20 gallon tank first to see what kind of results I get. If that goes well then I'll order the large number of LEDs that would be required for a larger tank. My initial calculations say I'll need about 250, but we'll see what really happens It'll probably be a while before I have anything to report, but I will do so with either good or bad news. Thanks for the interest.

 

[taikonaut]

>... if you take a peice of sandpaper and just "rough up" the tip of the LED a bit it will help to diffuse the light ...

Yeah, using sand paper not only make the light more diffuse, it also will cut down the amount of light pass thru the lense, making it less efficient.

>... 1 mcd = (lumens*1000)/12.57

utahsaltreefer, I don't know if I had make my point explicit enough.... you can't convert the mcd and lumen with the above "conversion" since the unit is not compatible between LED and MH.

Depends on the type of reflector and lense, you will need about 1000-1500 LEDs to get as much lum as one 250W MH. That will suck up about 300W of electrity, not 9W in your calculation. I don't want you to disappointed by your project after you spend all the money on it.

 

[Mihai]

OK, then I guess that we need to know how many mcd are radiated from a typically MH lamp + reflector. Taikonaut, do you have any idea? Numbers? Besides less heat and smaller form factor, they also don't need a balast: a simple DC converter would do. There is one more thing taiko: the LEDs are focused, but this is good in our case, as we do want all the light in the tank and not all around the hood like it happens with the MH.

All the best,
Mihai

 

[utahsaltreefer]

taikonaut I understand your agrument and am not discounting it. I ordered about $20 worth of LEDs to experiment with so its not a major investment at this point. I realize that 10 leds are not going to be a large amount of light overall, but I want to test how much they light up a 6 inch by 6 inch by 16 inch deep area. If the tests don't pan out, not much is lost (at least not much for a hobby like this :lol: ) If they don't turn out to be useful as a main light, I think at least they may have some value as a cluster for a spotlight on corals or clams. It may be nice to be able to hit a small area with intense actinic light.

 

[Bobzarry]

10 high output leds will give you about as much light as new hologen-zenon automotive headlight. This is just from observation of the led replacement bulbs for headlights.


Bob

 

[taikonaut]

Mihai, good question.

1 lumens = 1 cd*4_pi. pi = 3.14159, so the ratio is ~12.57 (see utahsaltreefer's formula).

The important thing is that ratio. The difference between cd and lm is that lm is over the whole unit sphere (surface area of a unit sphere is 4*pi*r, where r=1, so the ratio is 4*pi.) While cd is over a square patch on the unit sphere.

For LED people, they purposely pick a specify patch on the sphere that give them maxmium reading, especifically, right in front of the LED. There is nothing wrong with this approach, since who care about how bright is an LED if you look at it from the back. But the ramification is that if you use the LED specification to compare with MH, you will be mislead to believe that LED is better than MH *all* the time (LED, mind you, is great for certain application, such as nano tank). So in order to compare apple to apple, you need to find out the delta angle (depends on the lense and reflector design fo the LED) that they use when they give you that mcd reading. With narrow angle that typical LED measurement is used, you need to put a factor of anywhere between 20-50 on the LED reading when compare to MH.

Wiht reflector for MH, you can increase the brightness of MH by at least a factor of 3 or 4. Take a look at a common MH reflector such as Spider light. With the bulb off, and you look at it at most common angle, you will see at least 3 bulbs (one real, the rest is the reflection). So with a reflector, you increase the brightness by a factor of at least 3 when compare to a MH in a totally black hood.

 

[Mihai]

Taikonaut, that explains a lot. Thanks for enlightening me. This makes sense. Except for one part:

Wiht reflector for MH, you can increase the brightness of MH by at least a factor of 3 or 4.

I disagree that the multiplication factor is equal to the number of bulbs you see: if you would put 1000 little mirrors you will not have 1000 times the intensity. Rather, I'd say that the multiplication from a reflector cannot be more than double the one from the black reflector:
With a black reflector almost half of the entire energy goes to the tank (the energy radiated in the lower half), with a perfect reflector 100% goes to the tank.

OK, maybe slightly more than twice because you can have less than half hitting the tank (maybe 30-40%).

Anyway, it becomes apparent that the main missing ingredient in the equation is the lumen quantity for the LEDs, which would be equal to the integral of the (variable) mcd over the entire sphere around the LED. This is something that cannot be figured out without the help of the guys that manufactured the LED (or without expensive light measuring equipment, by measuring all around the LED).

Now that we decided that utah's formula may be off by a lot, the next question becomes: does anybody know how many lumens can a LED
put out?

I think that a reef lighted by LEDs would be extremly cool (also literarly ).

Mihai

 

[taikonaut]

>... if you would put 1000 little mirrors you will not have 1000 times the intensity.

That's right, Mihai. This explains why those LEDs have such a high value for brightness when in mcd unit. See, the spot where all the mirror directed the MH light to is 1000X brighter, but that is only *one* spot, while the rest of the tank is pretty dark. If you can't picture this, consider holding a magnifing glass under the sun, or more appropriately for our discussion, the solar heater out in the desert that boils water for generating electricity. You got a very, very bright spot, at the expense of dark area.

If the specs use lm instead of mcd, the LED will have a very disappointing, albeit unfair, value. They are not trying to mislead you, since it is the right type of unit for the application, but when you use it to compare LED with MH, you mislead yourself due to ignorance.

>... I disagree ...maybe 30-40%...
He he... so 1/30% is around a factor between 3 and 4, isn't it?

Think of a unit ballon.... now cut the bottom out hortizontally with a knife. That is the amount of light going to the tank. See the remaining part of the ballon? That is at least 3 time larger in surface area compare to the bottom part that you cut out. This is why most reflectors has the bulb within it, to capture the light that goes sideway.

 

[Mihai]

Agreed completly (except for the ballon - that would be the 1/2 mentioned above, but I still think that less than half gets into the tank directly).

Anyway, without some measurements we can not proceed with the comparison (either cdm for MH or lumens for the LEDs).

Awaiting with interest utah's measurements.
Mihai

 

[taikonaut]

>... except for the ballon - that would be the 1/2 mentioned above...

Well, what can I say!

Actually, if the LED specs states the light angle, you can calcuate the factor you need to convert from cd to lm.

 

[Mihai]

Yeah, that would be neat. Utah, do your specs specify anything like a solid angle (even non solid would work) for the LEDs? Something like 90% of all light is concentrated in x degrees/radians/steradians?

Thanks,
Mihai

 

[utahsaltreefer]

Sorry I stopped reading my on post for a while. The angle on the LCD is 15 degrees. I don't know the distribution over the 15 degree angle. The LEDs will arrive sometime this week.

P.S. taikonaut people like you are why many people don't post questions. Being called ignorant after asking someone to check my formula is uncalled for. If I missed something fine. That's why I posted it, to find out.

 

[taikonaut]

Everybody is ignorant of one thing or an other, unless you are omniscient. I am sorry if you take my statement to be too negative, and please accept my apology if you feel that way.

 

[taikonaut]

Alright, back to discussion....

A cone from center of sphere with angle of 15 degree is a bit difficult to visualize for me, so I am going to use a pyramidal cone instead. Let me know if the following math is wrong due to my ignorance.

2 * tan(7.5 deg) = 0.3633 = side of py.
Area of base = side * side = 0.06933 ster. radian.
Full sphere = 4*pi ster. radian, so the factor is:
4*pi/area_of_base = 12.57 / 0.06933 = 181.

 

[utahsaltreefer]

Its been quite a while since college physics, but I wonder what effect the angle of entry into the tank will have with respect to efficiency of the lighting? The air water interface has quite a severe difraction ratio if memory serves. That is exacerbated by the motion and ripples at the surface which diffuse the light. I don't have a physics book handy, but the entry angle of a focused light source should negate some of the difraction shouldn't it? I realize that is the purpose of reflectors, but can't recall an analysis. Many reflectors claim to give you 100% plus light directed down at the tank, but each time the light source is redirected it would lose some energy. Maybe this is minimal, I dunno. Just more ignorant speculation probably :?

 

[utahsaltreefer]

After doing much searching of manufacturers websites It seems that LEDs vary considerably in luminous efficacy. MH is anywhere from 60 to 125 lumens per watt and LEDs are anywhere from 20-55 lumens per watt. In the next year or so they are expected to pass 100 lumens per watt though. The LEDs I ordered according to the manufacturer generate about 1 watt of luminous radiance and efficacy of 35. This is the number that you plug into the lumens per watt rating. So to equal the total output of a 14000 luminous radiance MH you would need about 400 LEDs The manufacturer did point out that the small angle should be to my benefit though. The 14000 lumens of the MH would have a much lower percentage of luminous intensity that actually made it to the tank. Of course they didn't have any numbers on this.