Minimalistic multichip DIY LED build

[MrShroomy]

but back to my original question would that work and has anyone done it?

 

[daks]

Thanks daks! My intention is to have 4 drivers on completely separate plugs into a reefkeeper. Each individual driver would power five 20w LED modules. I would be able to dim and time all 4 drivers with different color LEDs independently allowing for excellent dawn/noon/dusk color control. I have purchased all LEDs from AC-RC on ebay. evidently they are banned links as my previous post was deleted. They are all 45 mil Epistar 20w colors are 445nm, 453nm, actinic hybrid, and 10k (10000k).

ballast 1 powers five 445nm
ballast 2 powers five 453nm
ballast 3 powers five hybrid
ballast 4 powers five 10k

It sounds like I may have been misguided on the ballast selection. I hope this is not the case. :debi:

K I think I found the chips you are talking about on ebay, they had the same link to the individual led pdf you posted earlier and they had this info in the Ebay add for the 445nm version (seems to be the same info for the other versions I looked at;

DC Forward Voltage (VF): 10-11V DC
DC Forward Current (IF): 2000mA Max
Rated Power: 22W max
Viewing Angle: 120 Degree
Intensity Luminous (Iv): 500LM
Life span 50000 hours guarantee

The answer is YES that driver WILL work if you wire 5 of those chips in SERIES with each driver. i.e. - of the first chip connects to the V- of the driver (black wire),
then the + of the first chip connects to the - of the second chip,
+ of the second chip connects to the - of the third
and so on...
then the + of the 5th chip will connect to the V+ of the driver (red wire).

 

[mr.wilson]

I like the car analogy, thanks Daks.

 

[mr.wilson]

Just out of curiosity has anyone used these multichips with t5 lighting? I know some people have trouble coloring corals with leds but if you had t5 lighting doing the coloring that seems like it would work. Maybe have leds deliver the white spectrum and t5 deliver the "blue". That way you get the best of both worlds, any thoughts on this?

Some people are very happy with T5 lighting and T5/LED combos. ATI makes a hybrid fixture with both. You could have the best of both worlds, but if you aren't careful, you may end up with the worst of both

Personally, I've never liked the quality or quantity of light from T5 lighting, so I don't see what it would bring to the table. Fluorescent light lacks shimmer, needs to be close to the water surface and is subsequently difficult to keep clean, and has a cold clinical/fake aesthetic. These are characteristics that some LED fixtures share.

If you want an LED fixture that doesn't look like LED, then a MH/LED hybrid might be for you. There are some great 70w HQI MH lamps out there (Aqualine, Ushio).

 

[mr.wilson]

who has reported trouble getting coral to color?

Achieving optimum colour from coral is a challenge that we will always face. Whether it is water quality, water chemistry, or lighting, corals wax and wane in colour intensity.

LED fixtures that use only two or three colours of chips often cause poor coral pigmentation. Some parts of the spectrum are beneficial for reflecting colours while others provide the right spectrum for fluorescence (colour conversion). You simply cannot achieve this to its full capacity with a daylight white and royal blue chips.

For example, 611nm light causes red fluorescent proteins to fluoresce. This wavelength is not present in most LED fixtures. GFPs (green fluorescent proteins) are excited by blue light which should include 420 violet, 440 deep blue, 460 royal blue, and 480nm true blue. Standard LED fixtures include only one or two of these blue colours. High kelvin white chips include a broad spectrum of blues, greens and reds that you will not see in task 6,500-10,000k task lighting.

 

[mr.wilson]

I've only seen it reported that colors are different, same exact thing happens when you change halide bulb types, ie 10k to 20k or a change in t5 spectrum

it's not a led thing at all

One of the benefits of LED lighting is spectrum. Unfortunately, it is also one of its shortcomings. While a full spectrum of colour is available, many DIY and mass produced LED fixtures only use a narrow band of the full spectrum.

One of the limitations of MH is the inability to limit spectrum. MH is only inefficient because of the unnecessary spectrum provided. You get IR (infrared), UV (true UV) and a mix between that is limited by phosphor application and interaction.

LED is a superior technology, but only if implemented properly. There is no definitive optimum spectrum for reef aquariums, but we can make a pretty good guess and estimate the right proportions of each colour of light. You certainly can't see quick results by dimming different colour channels, but at least we can satisfy our own personal aesthetic. Dimming has it's threshold and step limitations, and to make matters worse, LEDs can be difficult to colour mix due to placement and lens technology.

The narrow beam and hard shadows of LED, in comparison to the flood of 360˚ light of MH & T5 is a limitation we must respect. Multichips offer better colour mixing than individual emitters, but even with large dome lenses there is a colour separation if you are using uniform rows of one colour (e.g.. white on the right and blue on the left). While LED may not cast light everywhere in the tank like MH, it is nice to channel that light efficiently and keep it off of the glass and out of your room and eyes.

I am using four colour multichips in my shop (violet, royal blue, blue, 16k white). The light is predominantly blue so yellow or red corals do not show well. This is a reflective issue, not necessarily a pigmentation issue. When I use an LED with more white and red chips, the red, pink and yellow corals look awesome. There are two issues at hand "show" and "grow". You need to simultaneously balance growth, pigmentation, and presentation. I agree that LED is the way to accomplish this, and multichips are the best delivery system.

 

[megadeth72]

This is a 26 $ hybrid 50w multichip, a little twoo white for my tsate, it's on my sons tank

http://spectralworkbench.org/analyze/spectrum/2468

and here it is comapred to my dream chip how I have the color setup for me, I dont like too much white, the dreamchip has too much white for my taste if all are ran full blast, I could swap a whole white channel for more blue and be happy

I have 1 white at 20 % 1 white at 50 % 445 at 100 % 455 at 100 % and 420 at 100 % all channels using 1000ma

http://spectralworkbench.org/sets/show?id=70

 

[TropTrea]

For example, 611nm light causes red fluorescent proteins to fluoresce. This wavelength is not present in most LED fixtures. GFPs (green fluorescent proteins) are excited by blue light which should include 420 violet, 440 deep blue, 460 royal blue, and 480nm true blue. Standard LED fixtures include only one or two of these blue colours. High kelvin white chips include a broad spectrum of blues, greens and reds that you will not see in task 6,500-10,000k task lighting.

I will disagree with the comment that High Kelivin White LED's cover all the blues 420,440,460.and 480. In many cases they are nothng more than 6,500K chip with the longer wave lenghts filtered out. Remember when the rate a chip they are rating it on the ratio of red, green, blue. It is much eaier to get a higer amount of 460nm Blue than it to get an apreciable amount of 429 nm "blue"

This is why I cannot understand why somone will use a 20,000K chip then suplement it with Red chips and even in some cases amber and yellow chips.

Yes there is an isue with LED lighting with wave lenghts of 440 nm or less. There is also an issue with the wave lenghts in the 490 to 500 range as no chips put out a strong spectrum in that range. But with Royal Blue 454 (444 to 464 and Blue 460 (450 to 470) we are covering a majority of the blue range and the addaptibility of coral can easily compensate for anything missing near that range.

If you truely want to cover that range better the thought of combining HO T-5 bulbs like ATI Blue Plus, or URL Atinic plus White will do a better job than any individual LED.

 

[mr.wilson]

I will disagree with the comment that High Kelivin White LED's cover all the blues 420,440,460.and 480.

I agree. What I said was "High kelvin white chips include a broad spectrum of blues, greens and reds that you will not see in task 6,500-10,000k task lighting.". Also 420 is violet, not blue.

In many cases they are nothng more than 6,500K chip with the longer wave lenghts filtered out. Remember when the rate a chip they are rating it on the ratio of red, green, blue.

They are rated by how "blue" they are. Phosphor coating does not filter out light. It converts it to other spectra. Phosphor coated high kelvin chips are native blue chips, not 6,500k chips. They simply lack the phosphors that are cherished for their bright green colour. Blue light is converted to green and red to create a white chip.

It is much eaier to get a higer amount of 460nm Blue than it to get an apreciable amount of 429 nm "blue"

Of course, violet chips (429nm) are not blue chips with phosphor coating, they are a dedicated chip.

This is why I cannot understand why somone will use a 20,000K chip then suplement it with Red chips and even in some cases amber and yellow chips.

I have only seen yellow chips used with 6,500k fixtures which are already too yellow. I'm not sure what the logic is behind amber or yellow chips. The people who use them haven't given any reasoning that I have come across. 20,000k white chips have adequate yellow light.

Yes there is an isue with LED lighting with wave lenghts of 440 nm or less.

I don't know what issue you are referring to. Cost and availability is the main reason why these (violet) chips are absent in many mass produced fixtures.

There is also an issue with the wave lenghts in the 490 to 500 range as no chips put out a strong spectrum in that range.

490 - 500nm chips are readily available.

But with Royal Blue 454 (444 to 464 and Blue 460 (450 to 470) we are covering a majority of the blue range and the addaptibility of coral can easily compensate for anything missing near that range.

I don't want to single out any brands, but there are widely reported deficits with blue, royal blue and 6,500k chip mixes. You can keep a reef tank with that mix, but it will not equal the spectrum of metal halide, fluorescent or natural sunlight. There is a reason why 6,500k MH and fluorescent lights haven't been popular for over 20 years. They are popular in the reef LED industry due to heavy marketing for task lighting. High lumen/watt ratings sells lights. Unfortunately, this metric is based on a scale of human eye sight and not coral requirements.

If you truely want to cover that range better the thought of combining HO T-5 bulbs like ATI Blue Plus, or URL Atinic plus White will do a better job than any individual LED.

Fluorescent lighting has a broad spectrum, but there are many faults such as colour banding, no shimmer, artificial/clinical appearance, short lifespan, electrical hazard and maintenance issues (too close to the water surface), surface glare, and excess light cast upon the viewing panels (algae growth).

LED is clearly the easiest way to provide the desired broad spectrum and colour mixing.

 

[miatawnt2b]

Thanks for the plots megadeath72! You don't happen to have a 14000, 20000 and/or 25000 you could plot do you?

I put a 50w 25000k chip on my 18" cube and I really like the color... not too blue IMO, though I am thinking about getting a 14000k to try just for PAR sake. I am not running any optics, chip is about 7" from water surface. Corals bleached at first, but are starting to color now. (been running about 2 weeks) I switched to LED from a 175W hamilton bell pendant with a pretty old 14k bulb.

-J

 

[miatawnt2b]

Thanks for the plots megadeath72! You don't happen to have a 14000, 20000 and/or 25000 you could plot do you?

I put a 50w 25000k chip on my 18" cube and I really like the color... not too blue IMO, though I am thinking about getting a 14000k to try just for PAR sake. I am not running any optics, chip is about 7" from water surface. Corals bleached at first, but are starting to color now. (been running about 2 weeks) I switched to LED from a 175W hamilton bell pendant with a pretty old 14k bulb.

-J

 

[Psionicdragon]

so are there really difference in the chips that I should be looking for what I purchase them on ebay? like some RB 10w are $5 while others are $15...

 

[Whitebeam]

megadeth72: I'd love to know how you produced those plots please?

Peter

 

[Psionicdragon]

also what spacing do i need if i use 20 10w led to avoid the spot light effect?

 

[mr.wilson]

Here are some spectral plots of some commercially available LED fixtures. These may help as a comparison. Thanks to Sanjay Joshi for providing them...

 

[megadeth72]

Thanks for the plots megadeath72! You don't happen to have a 14000, 20000 and/or 25000 you could plot do you?

I put a 50w 25000k chip on my 18" cube and I really like the color... not too blue IMO, though I am thinking about getting a 14000k to try just for PAR sake. I am not running any optics, chip is about 7" from water surface. Corals bleached at first, but are starting to color now. (been running about 2 weeks) I switched to LED from a 175W hamilton bell pendant with a pretty old 14k bulb.

-J

I dont have any that high, I'm not a fan of those chips, I woudl rather use a blue and 6500k'ish white

I used a DIY spectrometer, they are hobbiest grade, not commercial but do a great job at comparing lights

I built my own using this guide and added a cosine diffuser

http://publiclaboratory.org/tool/spectrometer

 

[Whitebeam]

I used a DIY spectrometer, they are hobbiest grade, not commercial but do a great job at comparing lights

I built my own using this guide and added a cosine diffuser

http://publiclaboratory.org/tool/spectrometer

Yes, I did the obvious thing after posting the question and looked at the root of the links you provided. I have already built my own out of card and an old webcam. It needs some more work to make it consistent, but this is what I've been looking for all through my time reading about these LEDS, something that will actually allow me to scientifically check out what the (let's face it, cheap and perhaps not so precise?) Chinese suppliers claim for their chips.

At what point did you insert the diffuser and how much difference does it make BTW?

I wonder where (in the UK) I can get a scrap of coside diffuser?

Peter

 

[megadeth72]

I got some from us plastics, same type used for my DIY par meter, it's cheap

I cut it and stuck it right on the outside of the conduit box, it makes the reading much more stable as angle changes dont drastically change the reading intensity, the actual graph does not change at all, just makes it easier to take readings

I still have some focus issues to work out with mine, some of the peaks are are probably a little broader than they are in reality

 

[Whitebeam]

I still have some focus issues to work out with mine, some of the peaks are are probably a little broader than they are in reality

That may be a function of the width or the fuzziness of the slit that lets the light in. An old trick is to make the slit in the cardboard (in this case) very wide and then tape or otherwise fix two razorblades over it, sharp edges nearly meeting, to define a very precise, very narrow slit.

I may see if I can make a 'better' version with a 'proper' diffraction grating, a razor slit and a fixed camera - mind you, I have about 6 other projects on the go at once, so may never get around to this one ;-)

Peter

 

[Whitebeam]

BTW - Does anyone know how to take a spectrum like the ones Megadeth72 posted and convert this into the equivalent Kelvin rating? I suspect that may be the wrong sort of question - but I'll ask it anyway ;-)

Peter