Minimalistic multichip DIY LED build

[ReefUrchin]

First a replay to your post 2311. I'm sure Benanimal misunderstood my post which gave rise to his response. It is history now.

I do not often need a translator to read English (not even to read Americans ) but I need a bunch to write

First - You have probably already realized that not everyone here agrees with me. And that's lucky - for in that case, I probably wrong

What I'm saying: you need to evaluate by yourself and see if it makes sense.

1. I think that 20 watt with lenses will be enough for 22" and at least up to 25 - 26" It is also in the wavelengths where it is difficult to measure the effect of light (PAR and Lumen are not useful here). In the case of 420 nm, the question is easy because I have not seen any chip stronger than 20 watts of these wavelengths.

Personally, I think that 445 nm is most important - absorbance peak of chlorophyll C is around this wavelength. Maybe 2 pcs of 420/430 to 3 pcs 445 to 3 pcs 455 nm. I think also that you during "growth period" should run the because you get the blue wavelengths over 460 with the white chip’s..

2. I have tested 10 - 20 000 K from AC/RC - around 3 different batches. My impression was that 20 000 K gave a yellowier light than both the 16 000 and 14 000 K did. Then I looked at the light from the side and in the air. Not in the tank. My tank is based on biological principles, but I use ozone regularly so I have no yellow substances (humus) in the water.

I think that the best thing you can do is to bring home a 20 000K chip and compare for yourself if it looks good.

I have no experiences of any discoeffect in my set ups with one exeption. I tested a special rigg with a strong red “center” light over my aquarium. In this case – the small waves at my surface create a reddish skimmer in the water. With mixes of blue and white’s I never have seen this.

Sincerely Lasse

Thank you for the replay, but, now, hmmmmmmmm.........................

What also has me in a question is you say white light will give shades of blue. If we need such close but separate color chips, like 420, 430, 445 etc, how is white 20k light giving blue 460? This is one of the things bean was squawking about, not much spread. If it is there, that is great!

I think what needs to be done is to take a TOM, a really GOOD one, and put a spectrometer (or whatever it is called) in the tank and see what kind of spectrum the lights put out and, how much, to get the corals that are in that tank looking like they do. Then we take/make chips of various small divided colors and frequency's, divide that into separate tanks per color and cook (use word loosely) the corals. We will then see what grows what with no more guessing. Okay, well, at least a lot less guessing. We know somewhat the spectrum's that grow, but apparently not how much intensity of each we need, and, what others can be beneficial. If we use a good known example, I would think it will make it that much more easier to figure out?

Then again, maybe LEDs having such a narrow spectrum will be an Achilles heal and we will just have to wait for plasma to come to town? can we really fit 10nm times(x) 100 divisions of spectral color (just a guess) times(x) 100 LEDs? How big would a 10,000 LED chip be anyways?

 

[dartier]

Thank you for the replay, but, now, What also has me in a question is you say white light will give shades of blue. If we need such close but separate color chips, like 420, 430, 445 etc, how is white 20k light giving blue 460? This is one of the things bean was squawking about, not much spread. If it is there, that is great!

I am also wondering about the blue given off by the 20.5K light as used by the OP.

I have been out of the hobby for 20 years and am just returning. The move to LED lights is a bright spot for me compared to having to replace bulbs every six months, of which I have a dim view. :lolspin:

I am considering using the same 100W panels as the OP over a 125G (60x18x26), but I am also thinking about adding some 20W 420 nm as well. My plan is to use 2 x 100W and 4 x 20 420nm, with 3 Meanwell drivers (2 x HLG-120HB, 1 x LPF-90D-48). The 420 nm would be in series so they would all dim together. What I am wondering is would this be too blue visually. The OP's video looked pretty white to me (and that is with the hybrid Royal Blue LEDs) so I thought the 420 nm might be both good for the coral and help their colours.

Advice appreciated.

Dennis

 

[TropTrea]

Thank you for the replay, but, now, hmmmmmmmm.........................

What also has me in a question is you say white light will give shades of blue. If we need such close but separate color chips, like 420, 430, 445 etc, how is white 20k light giving blue 460? This is one of the things bean was squawking about, not much spread. If it is there, that is great!

The thing to keep in mind is that you cannot judge a light source strictly on the advertised Kelvin temperature. For one thing the K ratings especialy when you get that high are somewhat arbitrary. The general way of rating light sources are by determining the amount of red, blue, and green light they emit. The specs vary by manufacturer and can be as far as 10% off by some manufacturers on any of these color bands. Looking at the blue color band which we are mainly concerned about your realy looking at all the light emitted between 420 nm and 500 nm. One light source may be strong at 430 nm and produce noting at 490 nm. Another might be strong at 490 and produce nothing at 430 nm.

The only true way to judge the quality of the light is to either look at it through an accurate light spectrum anylizer system, or to get the manufacturers spectrum and hope it is accurate and consistant enough for our purposes. For most of us the later is the only feasable way since there are no non-bias testing labratory that has the money for equipment to give us accurate comparisons.

Looking at the plotting system mentioned earlier I strongly suspect that if three individuals evaluated the same light source only using different digital equipment they would get considerably different results.

 

[TropTrea]

I am also wondering about the blue given off by the 20.5K light as used by the OP.

I have been out of the hobby for 20 years and am just returning. The move to LED lights is a bright spot for me compared to having to replace bulbs every six months, of which I have a dim view. :lolspin:

I am considering using the same 100W panels as the OP over a 125G (60x18x26), but I am also thinking about adding some 20W 420 nm as well. My plan is to use 2 x 100W and 4 x 20 420nm, with 3 Meanwell drivers (2 x HLG-120HB, 1 x LPF-90D-48). The 420 nm would be in series so they would all dim together. What I am wondering is would this be too blue visually. The OP's video looked pretty white to me (and that is with the hybrid Royal Blue LEDs) so I thought the 420 nm might be both good for the coral and help their colours.

Advice appreciated.

Dennis

On a 125 gallon tank having two point white light sources to me is not the best idea. From any point light source you will get very hard defined shadowing. For a tank with this much surface area running more point sources will reduce the shadowing as well as assure a morte even distribution of light. your talking 200 watts of white light so rather than 2 100 sources I think 4 50 watt sources would be a big improvement, and 10 20 Watt sources would be even better.

On the Royal Blues and 420 nm chips this is a long debate that has gone on over and over. Idealy you want a strong blue light spectrum for your corals. However you do not want an excessive amount of light under 420 nm, and you want the spectrum to reach to at least 470nm or 490nm. No one LED ship covers this entire range. We each have our preferences on how to cover this range.

Royal Blues handle the spectrum nicely between 450 nm and 465 nm, True Blues handle the spectrum between 465 and 485 nm nicely. This leaves the spectrum between 420 nm and 450 nm that different individuals rate in various degrees of importance. But the royal blues to touch into this rang however do they give enough light in this range is very debatable.

 

[dartier]

Hmm, I could certainly do 20W units in place of the 100W. I expected the 100W would be too bright and would need to be dimmed. The same vendor also offers 430 nm 20W modules (which would help fill the gaps between the 420 and 455 RB).

He also has 405 and 395, but those seem too low to me and I don't want anything that *could* be dangerous.

I will think about doing it with all 20W. Thanks for the suggestion.

Dennis

 

[kuya_cesar]

If you start a new thread - please comeback with the link. If you not start a new thread - just take the discussions here but i could be good to have different spectra collected in its own thread. I did not mean that it was OT

Sincerely Lasse

Lasse
I did not start the thread, but since its on the same topic, I jumped in. He has a very very interesting link. A link with all the plots for LEDs with diff manufacturer, etc.

http://reefcentral.com/forums/showthread.php?t=2209884

Also because of this new tool, I ordered a bunch of AC-RC LEDs and LED-HK with different spectrums to test.

 

[jerpa]

I expected the 100W would be too bright and would need to be dimmed.

The brightness itself is not the problem. These multichip LED's are such a small point source that they create very distinct shadowing. The light will be very intense, wherever it can reach. No matter how bright they are the light won't go through rock.

 

[gtfcniles]

From ym experience the kelvin rating of LEDs is different to halide. 20k gives a very white light, for the blue look you need to go for 30k.

 

[tomservo]

It's true, some don't like the point source ("MH") look, but I prefer it, if you use tighter optics and hang the light higher (like a pendant) you get more natural sort of shadows. It depends on your aquarium's situation and what you want for a fixture/canopy/whathaveyou.

Keep in mind with 20w units you'll need a lot more heatsink/s or a big aluminum frame to put them on, and this can obstruct your access to the aquarium. It can be a real PITA.

 

[ReefUrchin]

The thing to keep in mind is that you cannot judge a light source strictly on the advertised Kelvin temperature. For one thing the K ratings especialy when you get that high are somewhat arbitrary. The general way of rating light sources are by determining the amount of red, blue, and green light they emit. The specs vary by manufacturer and can be as far as 10% off by some manufacturers on any of these color bands. Looking at the blue color band which we are mainly concerned about your realy looking at all the light emitted between 420 nm and 500 nm. One light source may be strong at 430 nm and produce noting at 490 nm. Another might be strong at 490 and produce nothing at 430 nm.

The only true way to judge the quality of the light is to either look at it through an accurate light spectrum anylizer system, or to get the manufacturers spectrum and hope it is accurate and consistant enough for our purposes. For most of us the later is the only feasable way since there are no non-bias testing labratory that has the money for equipment to give us accurate comparisons.

Looking at the plotting system mentioned earlier I strongly suspect that if three individuals evaluated the same light source only using different digital equipment they would get considerably different results.

I understand that and thats my point. Each chip is covering around a 10nm spectrum. How far away is 20k from 450mn? I am sure it is a couple of 10nm's spectrum width. If we are going by Beananimal, no way is 460 covered that far away (effectively) IMHO. But, to be honest, I am not a LED-ologist, soooo. If in fact a 20k LED CAN cover/go down to 460, then Beananimal is wrong and LEDs do cover the whole spectrum. We could also be talking about off wicked spectrum spikes too? Someone is wrong here, could even be me, lol.

Again, it should not be too hard to take a MH light and copy the spectrum using LEDs. We would just need a well proven volunteer (tank) and go from there. To get that dream chip, we may have to use green, yellow red and brown w purple polka dotted LEDs, but, once they are grouped together, we would know what % of what is needed. As for the visual test, i volunteer Bean for that. Coral growth I am handing off to Lasse. :smokin:

 

[Lassef]

Just wired it up real quick to see...... bright! But not so 20000k. More 11000k.

I have try to say that for at least the last 60 pages You can´t compare the same K from MH and LED

So what now? Order another not so blue and supplement? Let it burn in? I'm thinking 3 blue and 3 RB 3w Crees in a circle around it on a dimmable meanwell?

10/20 watts RB multichip

Brightness..... deff bright. Lit the bottom of my 25" tank like nothing. Without lens. Too spread without it though. The lens holder helped a lot. No really focusing it, more like containing it. Maybe instead of a 90* I will do a 120* if I can find it.

Comments on the blue, RB ring? I kinda need some direction.. 3 of each, or 4 rob and 2 blue?


Edit. Only 4 3w leds..... 2 and 2 blues or 3 and 1?

I prefer 3:1 but 12 watt will probably drown - test


Sincerely Lasse

 

[ReefUrchin]

One question, away from my multi color/chips rant...........

Is there a 90v-108v /2.7-3.3A single variable driver out there for 3 of these?
http://www.ebay.com/itm/SMD-100W-LE...416?pt=LH_DefaultDomain_0&hash=item3f17886d00

 

[Lassef]

We could also be talking about off wicked spectrum spikes too? Someone is wrong here, could even be me, lol.

Yes and yes.

All white (as I know) LED consist of a blue (often around 450 - 470 nm) LED covered by a phopsphorus layer. Basically - thicker layer, lower K. There are also various mixtures of phosphorus. This results in a sharp blue spike with the source LED's wavelength peak. Typically stronger the higher K.

See here for examples

Sincerely Lasse

Is there a 90v-108v /2.7-3.3A single variable driver out there for 3 of these?

Have not seen any

 

[Lassef]

I am also wondering about the blue given off by the 20.5K light as used by the OP.

I have been out of the hobby for 20 years and am just returning. The move to LED lights is a bright spot for me compared to having to replace bulbs every six months, of which I have a dim view. :lolspin:

I am considering using the same 100W panels as the OP over a 125G (60x18x26), but I am also thinking about adding some 20W 420 nm as well. My plan is to use 2 x 100W and 4 x 20 420nm, with 3 Meanwell drivers (2 x HLG-120HB, 1 x LPF-90D-48). The 420 nm would be in series so they would all dim together. What I am wondering is would this be too blue visually. The OP's video looked pretty white to me (and that is with the hybrid Royal Blue LEDs) so I thought the 420 nm might be both good for the coral and help their colours.

Advice appreciated.

Dennis

Send Maglofster a PM and ask him for his experiences since he set it up - he has been away for a while but he is back home now.

I have been using 10 and 20 watts because I want to fine tune my light. If you do not have that whish - it works good with the 100 watts - see them as halides in the perspective of point sources. But you need some blue, not only 420 because the 420 is great for many purposes but your eyes will not see it as a strong source,

Sincerely Lasse

 

[mr.wilson]

T
@ mr.Wilson: with full respect: Your argument mixes wildly between reflective colours and colours we see due to fluorescence (and why does Google all the time spell colours like colors). I think that we must separate them in the discussion. I do not know any fluorescence based on red wavelengths (i.e. "“ red is the ingoing photons wavelength). It might be "“ do you have any links?

Americans don't use the "u" in words that are spelled with an "ou", such as neighbour, harbour, and colour. It's some kind of budget cut

This chart and photo from the study on blue light show fluorescence from red light. It is less significant than that of blue light, but here nonetheless.



Other studies have shown that an excess of red light can bleach corals. MH lights have more red light than any LED fixture on the market so we are working within the safe zone.

I'm striving for fluorescence in corals. A little red light will help me see this fluorescence which is largely caused by blue light. If I had a coral farm, my protocol would be more geared toward light for photosynthesis.

 

[ReefUrchin]

Have not seen any

Grrrrrr, darn. Anything close, lol?

 

[jt41time]

I have try to say that for at least the last 60 pages You can´t compare the same K from MH and LED



That or 10/20 watts RB multichip



I prefer 3:1 but 12 watt will probably drown - test


Sincerely Lasse

How would u make it more blue using the 50w I have on a 4x4" heat sink? On a budget.

 

[BeanAnimal]

From ym experience the kelvin rating of LEDs is different to halide. 20k gives a very white light, for the blue look you need to go for 30k.

It is differnet because the MH lamp produces full spectrum light with distinct peaks, the LED emitter (or combination of emitters) produces ONLY distint peaks. "Kelvin rating" is a very complex subject and without knowing the specifics of the CRI and subsuquent CCT if the lamp, the number is as meaningless as "blue" "yellow" or "green" or "red"....

What was being illuminted to compare to a 20K black body color temperature? Was it a blue chip of paint, was it a fill CRI color sheet? Was it simply a judgement call based on the spectral plot? Who knows? All we do know is that there isn't a single kelvan rated lamp in this hobby that renders as expected

 

[Lassef]

How would u make it more blue using the 50w I have on a 4x4" heat sink? On a budget.

Now I´m not reading you. What type of 50 watts do you have?

Sincerely Lasse

 

[mr.wilson]

From ym experience the kelvin rating of LEDs is different to halide. 20k gives a very white light, for the blue look you need to go for 30k.

Absolutely. I would say...

LED 10k = MH 7k
LED 14k = MH 10k
LED 16k = MH 12k
LED 20k = MH 14k

Some higher kelvin LEDs are quite pink looking as well, especially when mixed with low frequency blues & violets.

If anyone wants a professional spectrometer, this one has good reviews and is relatively cheap.

http://www.oceanoptics.com/products/usb650.asp