Minimalistic multichip DIY LED build

Wow, so your in agreeance with the white input from the other thread. I keep hearing about a new DC, maybe they need to cut back on white channel. Do you have your whites dialed up just at viewing times?
 
I do not know how much you know about LED and phosphorus technology but in general – if you want a lower Kelvin and higher CRI – you will lose more of the output from the original blue led in the transformation to a white look. If you look at Cree’s own table of the XM-L output you will see this. For Cool White (5000-8000K) they have 300 Lumen at 700 mA, NW (3000-5000K) – 260 Lumen and down to 185 Lumen for 85 and 90 CRI white (2700 K – 3200 K) My bottom line is that you lose a lot of energy when you transform the blue light to low Kelvin whites – therefore you can’t compare the way you do



I totally disagree with you - I have 4 years experiences with only high Kelvin LEDs over coral tanks. I have also experiences with Cree XLM on other tanks. You have to explain for me what makes a warm white LED better for growth compared with a High Kelvin LED. The last 5 months I have run 3 DC over the tank (420,430,445,455 nm blue, 10 000 and 16 000 K white). The growth has been very good.

It’s exactly what I have done for 4 years - and I have also seen what happens. And it is not what you indicate....
I'm asking about sticking only high-kelvin whites over a tank of coral, with no blues. It won't work. And when you use high-kelvin whites, you are required to use more to get a nice 14k or 20k color- 1:1 seems to work for many.
But, with lower kelvin whites, it doesn't matter if they are generally more inefficient. The fact that they are Cree, and the fact that you need to use less to get a nice color is a big plus.
So unless that 10k white is getting more than 2x the output of a single Cree neutral white(you can run NW in a 2:1 ratio of RB:NW), than it's not worth it.
Lumens are also not the compelte picture for whites. As I mentioned before, lower kelvin whites have a lot more "useful" spectrum in them than cool whites, especially 660nm red. It's not critical for coral growth, but it's there.

EDIT: Looking at this datasheet: http://www.wayjun.com/Datasheet/Led/10W Epistar 45mil Chip High Power LED.pdf
and comparing it to the Cree XM-L datasheet, we can see that the Cree cool white at 1000ma gets 416lm in the highest bin, minimum. It runs at about 3v at 1000ma, so we get around 138lm/w. By contrast, the Epistar at 1050ma and 9.6v gets a minimum of 810lm, for an overall efficiency of 75lm/w minimum. That's a TON less. And those are at about the same color temperature.

<WARNING> assumptions ahead. Most of the following is probably flawed.
Now, accounting for the changes in luminosity with color temperature, we see that the lumen output at test current is 300lm for cool white, 260 for neutral, and 220 for warm. That's from 2600k, to 3700k, to 5000k. About a 1200k difference per step. And you get 18% more output with warm to neutral, and 11% from neutral to cool. So every step, percentages are multiplies by 11/18 to get the percentage increase for the next step.
To get to 10000k, we can multiply that 1200k step by about 4 to get to 9800k, which is close. So we calculate.

1. 416lm * (0.11 * 11/18 + 1) = 443.
2. 443 * (0.06722 * 11/18 + 1) = 461.
3. 461 * (1.04107 * 11/18 + 1) = 473.
4. 473 * (1.02509 * 11/18 + 1) = 480.

So, a VERY rough estimate is 480lm for a Cree 9800k minimum white when run at 1000ma. That's about a 15.3% increase, although the numbers are off by a bit. Versus a neutral white with 360lm at 1000ma, that's a 33.3% increase. But because you only need half the number of neutrals to get the same effect on color as a cold white, it's a better bargain and you can have more blue, royal blue, and violet.
 
Last edited:
Wow, so your in agreeance with the white input from the other thread. I keep hearing about a new DC, maybe they need to cut back on white channel. Do you have your whites dialed up just at viewing times?

I think that the idea of using high Kelvin LEDs and Pacific Suns idea with not use whites at all (but get a "white" light through using RGB) adress the same problem - not to blow out the colours from the fluorescence with high intensity whites in the "Lumen" window. Our eyes is most sensitive for light with a wavelength of 555 nm. Se this page. To much of intensity in the aera between 500 - 600 nm will block out weaker sources of light (read fluorescence) They are there but you are blinded of the reflecting wavelenghts caused by the wavelenghts between 500 - 600 nm.

If you can create a white look without using the wavelengths in the optimum of your eyes sensitivity - its more likely you will see the colours from the fluorescence. High Kelvin LEDs has not so much of their wavelengths in this area and the Pacific Sun method use an other way to create a feeling of white light. Peak sensitivity for colours in our eye is 445, 535 and 575 nm. Your brain is doing a fast calculation of the intensity of those wavelengths and "guess" the original colour. Are they rather equal - you will see a white light (the RGB trick!) but you are not blinded out from other colours that will be sent out from the fluorescence.

I do not know if it is a good idea to change the configuration of the DC but I will try to do some experiment in the future and try to figure that out.

Sincerely Lasse
 
Last edited:
I'm asking about sticking only high-kelvin whites over a tank of coral, with no blues. It won't work. And when you use high-kelvin whites, you are required to use more to get a nice 14k or 20k color- 1:1 seems to work for many.
But, with lower kelvin whites, it doesn't matter if they are generally more inefficient. The fact that they are Cree, and the fact that you need to use less to get a nice color is a big plus.
So unless that 10k white is getting more than 2x the output of a single Cree neutral white(you can run NW in a 2:1 ratio of RB:NW), than it's not worth it.
Lumens are also not the compelte picture for whites. As I mentioned before, lower kelvin whites have a lot more "useful" spectrum in them than cool whites, especially 660nm red. It's not critical for coral growth, but it's there.

EDIT: Looking at this datasheet: http://www.wayjun.com/Datasheet/Led/10W%20Epistar%2045mil%20Chip%20High%20Power%20LED.pdf
and comparing it to the Cree XM-L datasheet, we can see that the Cree cool white at 1000ma gets 416lm in the highest bin, minimum. It runs at about 3v at 1000ma, so we get around 138lm/w. By contrast, the Epistar at 1050ma and 9.6v gets a minimum of 810lm, for an overall efficiency of 75lm/w minimum. That's a TON less. And those are at about the same color temperature.

<WARNING>assumptions ahead. Most of the following is probably flawed.
Now, accounting for the changes in luminosity with color temperature, we see that the lumen output at test current is 300lm for cool white, 260 for neutral, and 220 for warm. That's from 2600k, to 3700k, to 5000k. About a 1200k difference per step. And you get 18% more output with warm to neutral, and 11% from neutral to cool. So every step, percentages are multiplies by 11/18 to get the percentage increase for the next step.
To get to 10000k, we can multiply that 1200k step by about 4 to get to 9800k, which is close. So we calculate.

1. 416lm * (0.11 * 11/18 + 1) = 443.
2. 443 * (0.06722 * 11/18 + 1) = 461.
3. 461 * (1.04107 * 11/18 + 1) = 473.
4. 473 * (1.02509 * 11/18 + 1) = 480.

So, a VERY rough estimate is 480lm for a Cree 9800k minimum white when run at 1000ma. That's about a 15.3% increase, although the numbers are off by a bit. Versus a neutral white with 360lm at 1000ma, that's a 33.3% increase. But because you only need half the number of neutrals to get the same effect on color as a cold white, it's a better bargain and you can have more blue, royal blue, and violet.

You will have more blue wavelengths with using high Kelvin LEDs - you forget that a high Kelvin LED by definition has more left of the original blue wavelength. And we are not talking about only 10 000 K - we are talking about 16 000 K also. You also totally forget that colours of corals is not only depending of reflection - corals are unique because much of their colours come from fluorescence. When you use low Kelvin whites you directly hit our eyes more sensitive parts (that the reason why they are developed from the lighting industry) and you will blind out the weaker light sources (a coral that use fluorescence is indeed a light source). To get the beatuful colours of the fluorescence you need to tune down the low Kelvin whites a lot and you will get at very blue tint in your aquarium. With high Kelvin LEDs and especially with the Pacific Sun method you do not need to turn down the white light (at least not so much with high Kelvin and not at all with the RGB method) and you will also get the reflecting colours because you will still have these wavelenghts with.

lower kelvin whites have a lot more "useful" spectrum in them than cool whites, especially 660nm red

The 660 nm red is (IMO) of no value for corals - it has disappear already at a depth around 1.5 m in nature. If you are growing freshwater plants and macro algae - thats another question. Still I do not understand what you mean with more "useful" spectrum - please give examples



Sincerely Lasse
 
Last edited:
You will have more blue wavelengths with using high Kelvin LEDs - you forget that a high Kelvin LED by definition has more left of the original blue wavelength. And we are not talking about only 10 000 K - we are talking about 16 000 K also. You also totally forget that colours of corals is not only depending of reflection - corals are unique because much of their colours come from fluorescence. When you use low Kelvin whites you directly hit our eyes more sensitive parts (that the reason why they are developed from the lighting industry) and you will blind out the weaker light sources (a coral that use fluorescence is indeed a light source). To get the beatuful colours of the fluorescence you need to tune down the low Kelvin whites a lot and you will get at very blue tint in your aquarium. With high Kelvin LEDs and especially with the Pacific Sun method you do not need to turn down the white light (at least not so much with high Kelvin and not at all with the RGB method) and you will also get the reflecting colours because you will still have these wavelenghts with.



The 660 nm red is (IMO) of no value for corals - it has disappear already at a depth around 1.5 m in nature. If you are growing freshwater plants and macro algae - thats another question. Still I do not understand what you mean with more "useful" spectrum - please give examples



Sincerely Lasse

I'll address the issue of the red first.
images

Take a look at the above image. Chlorophyll A, which is prevalent in most corals, has a large peak in the 660nm. I don't know why, because as you said, red disappears fast in the sea, but it is there to take advantage of. That is the "useful" spectrum I speak, of - red. In graphs of 10k and higher leds, red is almost nonexistent. Not to mention there is a lack of green and yellow, but those are mainly for the human eye.

On the first point, with higher kelvin whites. While varieties in the 12,000k and 20,0000k are available, there is not much difference other than that there is more blue in the spectrum. I used 10k because is has marginally more red than the other two. Your point on the washing out of colors with low kelvin white is valid though- I have personally noticed it myself. However, in this case, using 10k and higher whites are no different; it's just that you get the "tuned down" color without tuning them down. That might actually be a problem, because you will be using more power to accomplish the same thing. Tuning down the low-kelvin whites still gives more red than cold whites, but uses less power. You can then add blue and royal blue to offset the lack of blue that comes with low-kelvin leds. Even though the high kelvin leds have a lot of blue in them, IMO a discrete blue is better for efficiency.

Of course, ideally, you would just use a combination of cyan, deep red, and cool blue to get white. Normal RGB leds that use green, red, and blue are not as good, because the 630nm and 430nm from the red and green are almost useless. But when speaking in terms of 10k+ white vs. 4k white, the topic is irrelevant.
This site sells custom multichips with 4 chips in each, you could use a combo of 660nm red, 400nm cyan, 470nm cool blue, and neutral white to get a nice white color, without making the corals looks bad.
On the minus side, $7.50 for 4w of led is not a good deal, particularly if they are of unknown brand. The same site sells 10w and 50w multichips, so those would be a better option if you would give up control (they 10w, 50w, adn 100w multichips are not individually controllalbe).
 
Summary: (from Photosynthesis, 6th Edition" by Hall & Rao (1999.
Cambridge University Press)
Chlorophylls:
Chlorophyll a .... 420nm and 660nm .... in all higher plants and algae
Chlorophyll b .... 435nm and 643nm .... in all higher plants and green algae
Chlorophyll c .... 445nm and 625nm .... in diatoms and brown algae
Chlorophyll d .... 450nm and 690nm .... in red algae

Carotenoids:
beta-carotene .... 425nm, 450nm, 480nm ... in higher plants and most algae
alpha-carotene ... 420nm, 440nm, 480nm ... in most plants and some algae
Luteol ........... 425nm, 445nm, 475nm ... in green and red algae and higher plants
Violaxanthol ..... 425nm, 450nm, 475nm ... in diatoms and brown algae

Phycobilins (water soluble):
Phycoerythrins ... 490nm, 546nm, 576nm ... in red algae and some cyanobacteria
Phycocyanins ..... 618nm ................. in some red algae and cyanobacteria
Allophycocyanins . 650nm ................. cyanobacteria and red algae

Chlorophyll B has not been detected in the zoox of corals - only chlorophyll A and C.

Of the other proteins - you should look for diatoms and most algae.

Most of our corals can be adapted (and live in nature) to depth greater than 3 m. The red wavelengths disappear fast - with the longest first. The intensity goes also down very fast. The blue wavelengths will on the other hand only lose around 2-5 % of its intensity the first 10 meters. In nature - therefore - the red wavelengths is not of any significance for corals to grow or not. At swallow areas - the 625 can be of interest, but sure - they can live and grow very good without them.

After I have used high Kelvin LEDs for more than 4 years I can say that if you use high quality high Kelvin LED developed for the aquarium market - you will get enough of red wavelengths to give a good colour reflection. I can take some pictures of my fishes to prove that. Further on - one of the companies directly addressing aquarium market has developed a phosphorus layer with a red tint and manage to get a CRI of 85 with a 16 000 K LED.

My experiences with different colour temperature is that the high Kelvin LEDs do not wash out most of the fluorescence colours the same way as WW, NW and CW. The photos in post 3843 is taken when my cannons was adjusted to give around 32 % for the white and 68 % for the different blues (of total 100 % energy output). As you can se - it is a good colouration both from reflecting colours and from fluorescence. Even the red discosoma show a good colouration.

Many times I have bought corals from shops with normal lighting (MH an fluorescents) - just to discover that they had total different colours when I put them in my aquarium. The fluorescence is there all the time - even when the whites are on. This is for most corals.


Sincerely Lasse

Edit. The photos are taken without flash and the only adjustment that sometimes has been done is auto correction of colours (Photo Shop) I have try to show exactly the colours I see with my own eyes.
 
Last edited:
Thank you Lassef. I do not rescind my comment on the usage of red by corals, but it is interesting to know that it does not affect your colors. I might give a couple of those higher-kelvin whites a spin if I had the money. Regrettably, I need some lower-kelvin whites for a seagrass tank, but after that a nice led project would be to use 10k whites, perhaps compare them with a cyan, deep red, and cool blue combo white.
Could you give a me a link to that 85 CRI white? Seems like a holy grail of leds!
 
Looking to try this again. It appears that a box or two was lost in my move that contained my leds and drivers. I cant even find the paperwork from the person I bought them from on eBay. Who is a good supplier for the chips and drivers? Looking for a good selection of colors and wattages up to 100 watts.
 
Lasse:

Why is it that you are unable to post without using subtle (weasely?) means to drum up business? Is that you simply have no respect for the rules or is it that you have no respect for the rest of us that follow the rules?

Instead of playing games and taking the time to CROP the screenshots AND ADD "CENSORSHIP" text to the tile, why not just crop the title bar out? One step instead of two? Showing the software you (or your friend, whatever) sells and pasting the word "CENSORED" over it is simply a means to generate interest said software/hardware... but you already know that.

While I enjoy your contributions to the discussion, your continued efforts (albeit subtle ones) to promote your products are getting rather tired.
 
Last edited:
Lasse:

Why is it that you are unable to post without using subtle (weasely?) means to drum up business? Is that you simply have no respect for the rules or is it that you have no respect for the rest of us that follow the rules?

Instead of playing games and taking the time to CROP the screenshots AND ADD "CENSORSHIP" text to the tile, why not just crop the title bar out? One step instead of two? Showing the software you (or your friend, whatever) sells and pasting the word "CENSORED" over it is simply a means to generate interest said software/hardware... but you already know that.

While I enjoy your contributions to the discussion, your continued efforts (albeit subtle ones) to promote your products are getting rather tired.
I don't think he's trying to promote his products. I personally see no real advertising there, unless it's subliminal. :P
And if the title isn't in there, I can't see anything about the product in question. Without the name, you would be hard-pressed to find what software that is. Lasse has done nothing wrong IMO.
 
While I enjoy your contributions to the discussion, your continued efforts (albeit subtle ones) to promote your products are getting rather tired.
Man, I've been away for 4 months and have now returned to RC and this discussion and you're still at it. If someone is getting tiring it's you, absolutely not bringing anything new to the topic, while seemingly the only reason you are here is to accuse Lasse of "subtle promotion of his products".

There is a "Report post" button, you know? Leave it up to the admins to decide whether Lasse has broken any rules, instead of pretending to be a Forum cop.
 
Guy´s

Do not be rude to BenAnimal
In spite of his nick - I think he will be covered by the declaration of human rights from 1948, especially the parts about freedom of speech and the right to protest

sincerely Lasse
 
I am very interested in this build....

My two 250watt Radiums have been online for about 3-months now and I am looking to go LED when the times comes to change the bulbs.

My question is, can I build a multi chip fixture that will match the Radium 20k color spectrum and have a little more PAR? And how much in terms of Watts could I expect to save in switching to this led setup.

If so, to save me from spending 2-months reading through close to 4000 post's, can some one Please post a parts list of the latest and greatest (led chip & power supply) for this type of build?

-Alex.
 
I am very interested in this build....

My two 250watt Radiums have been online for about 3-months now and I am looking to go LED when the times comes to change the bulbs.

My question is, can I build a multi chip fixture that will match the Radium 20k color spectrum and have a little more PAR? And how much in terms of Watts could I expect to save in switching to this led setup.

If so, to save me from spending 2-months reading through close to 4000 post's, can some one Please post a parts list of the latest and greatest (led chip & power supply) for this type of build?

-Alex.

MY new setup is 5 Meanwell LLD 700H Drivers (Powergate website)
Meanwell 48v powersupply
A 5-up O2Surplus designed board (Local reefer sold me a Coralux board)
Makers 12" Heatsink (There is a groupbuy site)
And one of the 5.1 chips (AC-RC on Ebay or the other version on the groupbuy site)


HTH,

Larry
 
Last edited:
Yep, I have mine put together for testing right now. It dose all kinds of neat things...have it rotating a couple feeders, have the DS18B20'S addressed, outlet switched for heaters and pumps, my Tunez pumps, ect,ect. I have a 4 unit relay board for the pumps along with some of these guys for different voltages.>http://www.ebay.com/itm/310569140784?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649 (I've always been a flashing light, buzzer, bell type guy)
I'm going to seperate my PS voltage with them. 24 to Tunze via a relay, 11-12V to Arduino, 3-4V to feeders via relay ect. FUN:dance:---Rick
 
Last edited:
Back
Top