New DIY LED light questions

[dacianb]

Hello everybody,
more than a year ago I switched my FW Juwel tank to SW as a “let’s try sw” but slowly I became addicted to this hobby.In the meantime I am very unhappy with the current light system on Juwel tank, I even bought a LED fixture (according to leaflet I should be able to keep SPS with it, but fishes don’t even wake up when the light is on). I looked on all possible led systems on the market, but most are overpriced, some are really bad designed and yes, the acceptable ones are waaaay too expensive.

So I decided to start designing my own LED fixture.

No, I am not crazy – I am a professional led lighting designer since more than 15 years and I designed lights for most advanced industries from advanced microscopy to defense and aerospace… everything with LEDs, of course.
I will try to keep posting the steps I am doing on this DYI project, but will not be fast as I try to do things as should.

I know lots of things about LEDs and light, but still have plenty of questions for amazing people on this forum, as I am very new to sw and even newer to corals.

1. What is the best spectrum to cover requirements of all marine life forms from a tank? – is very confusing looking at what is on market. Here, my friends, I am totally lost as my sw knowledge is extremely limited.
2. If you would have possibility, what feature would you add to a led light on your tank? Honestly, I really don’t want to control the spectrum from my phone and I will not pay hundreds of $ for such gadgets. I want to make a light using all my knowledge, but have to be simple, efficient and as cheap possible. Let’s say, I will be able to offer myself a 5 years guarantee on this light
3. Did anyone saw on an aquarium LED fixture something related to IEC 62471-1 Photobiological safety of lamps and lamp systems?? Warning symbols?? I can explain further if someone is interested by, but I can say that blue/white leds are more dangerous than you think…

I would like to thank you all for the support already

Excuse my English, but is not my mother tongue

[farkwar]

Ya, try not to look into these LEDs when they are on.

 

If thats what IEC 62471-1 is about.

 

I believe they are classed or are equivalent to a Class II laser.

 

Also, Its doubtful that you can save money building DIY LED fixtures, over commercial units.And their resale value is horrible.

[jedimasterben]

You haven't even mentioned what size tank you have

 

As for your first question, this is the spectrum of the sun:

 

 

That's what everything exists under in the wild, so that is technically the 'correct' growth spectrum. That being said, most corals look like shit in the wild because they're desperately trying to avoid overillumination and death, having to use different pigments and fluoresce the light away.

 

This is the spectrum of the Iwasaki 6500K metal halide, regarded by many (and many coral farms) to be the single best light for coral growth:

 

 

And this is the spectrum of the Radium 20K metal halide, regarded by many (and many coral farms) to be the single best light for coral color (with stellar growth, as well):

 

 

That should give you your base of what you should want to mimic for better growth or for better color.

[farkwar]

Excuse my English, but is not my mother tongue

Thats ok, but didn't you know that we are all living in Amerika?

 

 

 

Rammstein - Amerika:

 

http://youtu.be/4NAM3rIBG5k

[fredfish01]

Output of a the ATI coral plus bulb:

 

Everyone has a slightly different idea of what the 'ideal' spectrum for a reef tank is except that it is heavier on the blue spectrum. A peak of some sort in the 550nm range appears to be needed to give the tank proper brightness and a more balanced look. A peak in the red (usually around 625nm) appears to be desirable to bring out the reds.

 

The big issue I see with all current lights is that they either show colour separation (disco effect) because light from separate LEDs does not blend properly or not enough light spread from 'LED cannon' products like Kessil. Nobody has yet found the right ballance between clustered LEDs, to avoid the disco effect, and tightly clustered LED arrays that do not properly spread light in the tank, but eliminate disco effects.

 

My sense it that you need a larger number of smaller clusters (7 to 9 LEDs) of various colours to give you a spectral output similar to the popular Metal Halide and T5 aquarium bulbs meant for reef aquariums.

 

Jedimasterben appears to have found a good combination, but his current setup is too much like a light cannon (not enough spread). It relies on T5 suplementation to avoid too much shadowing. Look at his very long thread in this forum for ideas.

[jedimasterben]

Jedimasterben appears to have found a good combination, but his current setup is too much like a light cannon (not enough spread). It relies on T5 suplementation to avoid too much shadowing. Look at his very long thread in this forum for ideas.

If I did it again, there would be two identical strips of LEDs with more clusters, running at a lower overall current. That being said, the LEDs do cover the tank pretty much end to end (but there would eventually be issues with stony corals and coloration on the sides).

[fredfish01]

If I did it again, there would be two identical strips of LEDs with more clusters, running at a lower overall current. That being said, the LEDs do cover the tank pretty much end to end (but there would eventually be issues with stony corals and coloration on the sides).

That's what I was trying to say.

 

I want to get to a similar combination to your evil cluster, but using 7 to 9 LEDs. I have a spreadsheet with about a bajillion different combinations I could try (paralysis by analysis ). I'd love to try them all out, but have not yet won a lottery.

[farkwar]

Check this out..

 

http://www.nano-reef.com/topic/346217-my-take-on-the-cluster/?p=4818759

 

Maybe he can sell you some.

 

Looks like an awesome cluster array.

[jedimasterben]

That's what I was trying to say.

 

I want to get to a similar combination to your evil cluster, but using 7 to 9 LEDs. I have a spreadsheet with about a bajillion different combinations I could try (paralysis by analysis ). I'd love to try them all out, but have not yet won a lottery.

Luckily for testing it doesn't take many, just order 2-3x of each LED that you plan on using in different combos and see how they all do, then buy the rest when you decide. That's what I do

[dacianb]

Thanks to all of you for so many great answers. And to try to answer a bit to some questions here...

- It is impossible to have with LEDs identical spectral output as with MH - just because those extremely high narrow peeks MH have. Even single color LEDs have wider bandwidth than those from MH

- I have some nice tools so I can start combining LEDs in a pure theoretical way until I get the right output - will keep you posted about results

- Also, jedimasterben - the spectrum of the sun you showed there is in midday - I am agree with you. But during the day, sun spectrum change from 2000K at sunset and sunrise to 10 000 K at noon

- and I have a Juwel Rio180 tank (approx 50G), but hopefully to upgrade it soon

[dacianb]

Looking around on different articles and other online sources I found some graphs regarding Zooxanthellae Light Absorption (on image bellow on yellow dotted line). In the same graph I added the VLambda ( human eye sensitivity graph) - red dotted line; and also typical spectra of a Neutral White LED - green line. All curves normalized to 1.

 

1. As I know PAR meters uses the same VLambda curve for measurements, so are those really necessary, as seems that most efficient wavelengths for corals are not even measured

2. Looking at the Zooxanthellae Light Absorption, Deep red range is 60% of the deep blue/violet range - but I never saw such a light on market. Is there any reason?

 

 

[jedimasterben]

- Also, jedimasterben - the spectrum of the sun you showed there is in midday - I am agree with you. But during the day, sun spectrum change from 2000K at sunset and sunrise to 10 000 K at noon

The sun always emits the same spectrum, and spectrum doesn't correlate to kelvin color temperature. Depending on the angle of incidence, though, our atmosphere can change the spectral quality. At any rate, the one I posted is the 'standard' for it

 

1. As I know PAR meters uses the same VLambda curve for measurements, so are those really necessary, as seems that most efficient wavelengths for corals are not even measured

PAR meters do not care about sensitivity to the human eye. They measure the total amount of photons falling on the area measured to their spectral response curves.

 

 

Picture from Advanced Aquarist: http://www.advancedaquarist.com/2013/2/equipment

 

2. Looking at the Zooxanthellae Light Absorption, Deep red range is 60% of the deep blue/violet range - but I never saw such a light on market. Is there any reason?

Because people don't understand red light, put simply. Seawater filters out red light fairly quickly, so most do not think it is necessary.

[dacianb]

Thanks a LOT jedimasterben. Things start to light up in my head

[dacianb]

And here is the first concept on my light....

 

300x190 mm size. 72 LEDs on 6 channels. Cooling will be capable of handling 200-250 W (if necesary, but I dont think I will go in high range). And some other features I am working on.

 

 

[Horerczy]

Neat concept design.

[dacianb]

I am fully agree that the solar spectrum dont change, but light spectrum reaching the sea level change at different hours.

 

I dont want to debate it, but look at following image - such descriptions you can find in different photography books

 

 

 

The sun always emits the same spectrum, and spectrum doesn't correlate to kelvin color temperature. Depending on the angle of incidence, though, our atmosphere can change the spectral quality. At any rate, the one I posted is the 'standard' for it

[jedimasterben]

I am fully agree that the solar spectrum dont change, but light spectrum reaching the sea level change at different hours.

Very little light enters the ocean when the sun is low in the sky.

 

 

See Advanced Aquarist: http://www.advancedaquarist.com/2005/8/aafeature

 

I dont want to debate it, but look at following image - such descriptions you can find in different photography books

White balance settings on cameras and the spectral output of the sun are different subjects that are loosely attached.

[farkwar]

Unless that fixture has only one or two different color LEDs, youre going to have horrible color separation, strip clubbing light.

 

That style layout works fine for a white/royal blue light. As soon as you add other colors they will cast different colors and shadows, noticably.

[dacianb]

Unless that fixture has only one or two different color LEDs, youre going to have horrible color separation, strip clubbing light.

 

That style layout works fine for a white/royal blue light. As soon as you add other colors they will cast different colors and shadows, noticably.

Thank you Farkwar for the tip. Strips are quite close one to each other and the LEDs will not use extra collimators - so color blending should be quite effective.

 

Anyway, I will test this possibility too and will see what happens.

[dacianb]

Unless that fixture has only one or two different color LEDs, youre going to have horrible color separation, strip clubbing light.

 

That style layout works fine for a white/royal blue light. As soon as you add other colors they will cast different colors and shadows, noticably.

Farkwar,

 

your post remembered me about a technology I used in the past on different application to blend the different colors LEDs very smooth - it is called LSD or Light Shaping Diffuser.

With such the hotspots of the LEDs are completely eliminated and the colors will blend very uniform. With small losses in transmission a LED fixture can have a similar light output with a T5 tube.

 

http://www.luminitco.com/sites/default/files/LEDLightingApps_5_21_12_lo.pdf

[Halo_003]

Why not order a couple of Nanobox clusters from Dave Fason?

[jedimasterben]

Why not order a couple of Nanobox clusters from Dave Fason?

It would take more than a couple

 

And some want more control. I would have no problem using Nanobox pucks if I had better control.

[Halo_003]

It would take more than a couple

 

And some want more control. I would have no problem using Nanobox pucks if I had better control.

 

 

Ah I missed that he had a 50G tank.

 

That makes sense of course.

[dacianb]

I am still far to finish the lights, but I believe that will be a good light and cheaper than commercial versions. starting this threat Ilearned a lot about what have to be done.

[jedimasterben]

39"x 16" x 20" is the size of the tank. I would not run a single fixture like you have planned above. I would use two separate fixtures with several clusters of LEDs on each instead of keeping them in strips.