Minimalistic multichip DIY LED build

What do you mean string of 405, 420 and 430. So these are multichips or linear strips like the ones on ebay. And where did you bought them from?

By string I mean group of multichips on one driver. It has 2x 405 2x 420 and 1x 430 20w LEDs on it I am only running them with 10w each. Its the same as my 445 and 453 strings just with 3 different LEDs on it not just the one type.
 
This maybye can help somebody

View attachment 202933

Sincerely Lasse

I love it! This should be a sticky.

Except for the fact that some of the light in the cone below the surface gets reflected back into our box of rain by the glass. How do we account for that? My 8th grade trig can't and my college physics is too rusty to remember.

But now I'm talking about intensity not footprint.
 
Last edited:
Btw, this will be on a 4'x2'dx28"h

That seems like a bonkers light setup to me, although I understand you already own some of the 3-up emitters.. You'd do a lot better to get a pair of something like a 50-60w hybrid 45mil multichips and drive them at 100-130w each. If you add blues, reds and greens like in your diagram what you will get is loads of crazy disco lights. If you have a lid the condensation will act as lenses and make for strange colored spots and dots too.
 
That seems like a bonkers light setup to me, although I understand you already own some of the 3-up emitters.. You'd do a lot better to get a pair of something like a 50-60w hybrid 45mil multichips and drive them at 100-130w each. If you add blues, reds and greens like in your diagram what you will get is loads of crazy disco lights. If you have a lid the condensation will act as lenses and make for strange colored spots and dots too.

First thing, what hybrid ones are you talking about(link to them and the drivers please;) )
Second, I have all of the stuff on hand for what I would do(except the 50w). I went out and did an inventory of what i have right now. Except the triple pucks, I may run the rest of the leds without lens, which would help with spotlighting) Also, I have enough room in the canopy to raise it up fairly high if needed. I don't want to not use them, as I already spent good money on the lights, but I want to use them so I can customize the color and hit all the spectrums. Even if the led shows up as 20K and it looks close, I want that fine adjustment in color that only separate channels will give me.
LEDMockup-1.png



Thank you for the advice though! I will do my best to minimize the disco...unless I am at a party:bounce1:
 
First thing, what hybrid ones are you talking about(link to them and the drivers please;) )

Well, it's not considered appropriate to link to sellers - you could check popular online auction house for hybrid 45mil and probably find them. I have one here that is a 60w - what that really means is 60 45mil(3w size) chips. In this case it's 30 x 10000k white and 30 x 453nm blue. The end result is something like 12000k. It's substantially less blue than say a radium 14000k. The kelvin rating of white LEDs, especially in the higher color temps, is not directly comparable to MH or T5. Partly it's because the kelvin is a computed number, and there is some skew because the royal blue color is less visible to our eyes. There's also a 50w version that is I think 30 x 20000k white and 20x 453nm blue. The drawback is that there is less tune-ability but the color is already quite good - I have compared this one I have directly to my own 4000k white + 445nm + 420nm fixture and I actually prefer the hybrid, although I do feel it could use some 420nm. Versions with more than one controllable channel have been proposed and will eventually be widely available. I think Mr Wilson indicated that Orphek had made him some custom 100w chips but I don't know anything about that, pricing or minimum qty or even if they'd ever do it again.


Edit: I thought it would be helpful for demonstration purposes to include this link, there are side by sides 20000k multichip with phoenix 14000k http://www.reefcentral.com/forums/showthread.php?t=2199428&highlight=multichip
 
Last edited:
I too am building a multi-chip solution for my 180 reef tank.

Currently, I have three separate arrays on heat sinks with 2@50w 10,000K and 6@20w 450(ish)nm. The heat sinks are from a company that only makes heat sinks in the USA, and they are relatively local to me. They are 10" wide, and 18" long. I am putting one of these arrays per 2 cubic feet of water column.

I am using PWM drivers as I have an embedded solution (.NET on an ARM processor) that is driving the daylight and moonlight cycles. Using a one-wire thermal probe and a relay, I am intending on running my own reef temperature monitor/controller. I am still considering what ATO i want to run. Might just get a DC motor that runs at 24V and use some of the spare amperage I have from one of the converters to run the pump off of the TLC5940NT PWM Multiplexer.

Why the MUX? Because, I want to be able to simulate actual sun rise on a linear plane in the tank. The sun rises in one cardinal direction and sets in the other. I also want some "cloud" cover as well. The moonlight cycle is tied to the actual moon cycle. New moon is 0% duty on the moonlight LEDs, full moon is 100% duty cycle. All of this is being controlled by either an Android app, or by a touch-screen on the aquarium.

I have two 24V 14.7a AC-to-DC converters that will run them. Positionally, they will be about 11" off the surface. I am still deciding on optics. It's a 6'X2'X2' standard 180.

I purchased a single 50w with driver and installed it on an old AMD CPU heatsink I had lying around and man it gets hot. I decided to order proper heat sinks.

I have the layout in sketchup for now. The heat sinks are about 9lbs each. I wanted to cut weight on the fixture, but it seems like it's going to be 40lbs or so.
 
Last edited:
I love it! This should be a sticky.

Except for the fact that some of the light in the cone below the surface gets reflected back into our box of rain by the glass. How do we account for that? My 8th grade trig can't and my college physics is too rusty to remember.

But now I'm talking about intensity not footprint.

You can use this tool for the reflection. I have use it to get the refracted light angel in my calculations

If you look at the pull down meny just after reflection s and reflection p you can set % there. Now you have the reflection in % at each polarizations The reflection of non -polarizations light is around the average of this two.

But remember if you have the light source in a 90 degree angel to the surface - is only the outermost portion of the light beam reflected as much as this figure. The internal components ranges from 0 (center) to this value (outer part)

This is theoretical values ​​but gives an indication how much goes away. Same with my excell sheet - it gives you an idea

Sincerely Lasse
 
Last edited:
Ok camera has been charged and I took a few more photos. I also gave the tank a quick clean... You can actually see in the tank now. This is only a 4x1x2 that I have some live rock in.

50w 10000K running at full power.
9KXvvl.jpg


5x 20w 453nm Blue running at 10w each.
9XVZhl.jpg


50w 10000k dimmed down and Blues at full power
TeLX4l.jpg


50w 10000k full Blues dimmed
21NTLl.jpg


Both Dimmed or both full power sorry i cant remember but both look very similar.
gm77hl.jpg


I am currently running the 50w dimmed all the way down and the blues at a fraction above all the way dimmed down.

Anyway let me know what you think.
 
Folks:

Maybe it bears reminding that photographs are 100% useless for discerning the color rendering of light fixtures. There are simply far too many variables that come into play. Ignoring all of the source variables. each of our monitors is vastly different and none of us are seeing the same thing. Reach up and press the color temperature menu button on your monitor. Which setting is correct, they are vastly different?
 
I minimized the source variables by taking the photos in manual mode, I did not change any setting between shots or even move the camera.

Monitor variation is a well known problem and is true for every picture and video uploaded to the internet but that does not stop people from post hundreds of pictures and videos every minute of the day.

What other option do we have? Until something better comes along this is the only option, people are already well aware of the problems you mentioned but it doesn't stop people spending thousands of dollars on that "rare" zoanthid polyp...

Here are some links to the full size pictures for those that are interested.
http://i.imgur.com/zoBia.jpg
http://i.imgur.com/21NTL.jpg
http://i.imgur.com/gm77h.jpg
http://i.imgur.com/TeLX4.jpg
http://i.imgur.com/JKb5y.jpg
http://i.imgur.com/9XVZh.jpg
 
I minimized the source variables by taking the photos in manual mode, I did not change any setting between shots or even move the camera.
Even in a perfect world, what the eye sees and what a CCD see are very different. To drive the point home, we are talking about LEDs and as such our eyes (the cones in our eyes) average (for lack of better easy terms) those peaks to derive a "color". The CCD works very differently and actually sees the peaks and does not average them. That is why folks have so much trouble taking photos with scenes illuminated by LED.

Photos are 100% useless for the use of discerning the color rendering of a light fixture or scene, even if the camera, software and viewing device are calibrated.


Monitor variation is a well known problem and is true for every picture and video uploaded to the internet but that does not stop people from post hundreds of pictures and videos every minute of the day.
25 billion photographs and videos don't change the physics. Nobody said photos and videos were not pretty to look at or useless for all purposes. They are useless for the purpose of conveying the real world color rendering of a light fixture.

What other option do we have? Until something better comes along this is the only option, people are already well aware of the problems you mentioned but it doesn't stop people spending thousands of dollars on that "rare" zoanthid polyp...
Maybe if they were better informed and understood the problem (the reason I posted), they would not rely on photographs for color rendering data.
 
Coulor rendering data? I just asked what people think. I am not the first to post pictures in this thread (the first post has 3) and I wont be the last, this is a thread about multichip LED builds, I am currently completing a build and wanted to share my results. The method I used to take the pictures it the best I could use to minimizes the variation the equipment I have would impart on the pictures I was taking. Until there is a better method for providing coulor rendering data or just plain old pictures I don't understand what you want me and other people to do. Should every picture on the internet have a disclosure statement underneath or even embedded within it stating the facts about there monitor colour rendering problems and the conditions the photos were taken in?

I was well aware of the problems you have mentioned well before you raised them as many people are I just don't understand why you feel you have to tell everyone about it here and not on other threads, possibly you have and I have just missed it? You could also just be a troll and enjoy this type of "debate" again I may have just missed it.

I would like to keep the thread on track and go back to talking about Minimalistic multichip DIY LED builds myself, maybe you could start up a new thread and talk about the problems with every photo and video on the internet not looking identical to the moment in time they were taken. You can even use the photos I posted if you like I really don't mind.
 
Coulor rendering data? I just asked what people think. I am not the first to post pictures in this thread (the first post has 3) and I wont be the last, this is a thread about multichip LED builds, I am currently completing a build and wanted to share my results. The method I used to take the pictures it the best I could use to minimizes the variation the equipment I have would impart on the pictures I was taking. Until there is a better method for providing coulor rendering data or just plain old pictures I don't understand what you want me and other people to do. Should every picture on the internet have a disclosure statement underneath or even embedded within it stating the facts about there monitor colour rendering problems and the conditions the photos were taken in?
There is no need to get defensive, as you are not being attacked. The idea here is to help inform people and help them to make decisions based on good information. Helping those following along to understand the limitations of photographs (esp those posted to the internet) is part of providing good information.

A Disclaimer? Sure, in situations like this it is important. Most folks are blissfully ignorant of the limitations and complexities of photography and color rendering (in an analog or digital world) and simply trust what they are seeing as "fact" or "reality". Why? Because nobody has taken the time to explained to them why a photograph can't be trusted :)

You mention this thread as context, and it is in fact a perfect example. Much of this thread has been heavy on color rendering discussion and how different combinations of LEDs will actually end up color rendering a fish tank. The ONLY accurate (or even reasonable) way to convey this is in person. The photos are pretty, but 100% useless for anything more than "that is bluer than the other one" and even that can (and is) misleading.

I was well aware of the problems you have mentioned well before you raised them as many people are I just don't understand why you feel you have to tell everyone about it here and not on other threads, possibly you have and I have just missed it? You could also just be a troll and enjoy this type of "debate" again I may have just missed it.
You may be well aware, but the fact is that most folks are not AT ALL aware of the "problem" and simply trust what they see. To that end, most folks posting photos to depecit color end up saying something to the effect "It looks just like this in real life"... Or (as you pointed out) end up buying a coral based on a photograph, only to be upset when they see it under actual real world conditions.

With regard to "other threads". Yes myself (and others) tend to point out the limitations of photographs when we see them posted in a situation where they can (or are) being used to convery specific color rendering of lamp combinations. Again, there is no personal component to this and the information is simply posted to help others following along.

I would like to keep the thread on track and go back to talking about Minimalistic multichip DIY LED builds myself, maybe you could start up a new thread and talk about the problems with every photo and video on the internet not looking identical to the moment in time they were taken. You can even use the photos I posted if you like I really don't mind.
The thread is perfectly on track and the conversation is perfectly in context to the bulk of the conversation (color rendering of combinations of LEDs).
 
Figured I would share in this thread. Here are some pictures of my NEW light pendents. What do you think. Thanks for everyone who has posted info in this thread!
396569_4604393077957_900702509_n.jpg

418935_4604392397940_655290630_n.jpg

418905_4633405523250_1887272970_n.jpg

296556_4604391477917_520632799_n.jpg
 
Last edited:
mr. wilson, I have a 72 in (length) x 36 in (width) x 24 in (height) with SPS, LPS, clams, etc. What multichip would you recommend, optics, and how many?

1) 100-250w chips
2) 120˚ lenses
3) 3 fixtures ,mounted toward the front of the tank pointing slightly back, unless it's viewed from both sides.

The chip brand depends on your personal aesthetic preference for the tank. I like to use a higher wattage chip so I can dim it down to where I need it. Some lights meet minimum intensity expectations, but fail to do so once you start dimming down to your colour mix.

Red is very dominant, so that channel should be low intensity (<5% of the total chip). You could omit red entirely if you have enough red in your white chip.

Blue is easily drowned out by white, so go with more blue than you think you need. Having said that, I personally like a crisp white look.
 
Does anyone know what kind of spread a 100w LED multi-chip has 24" from the surface with no optics? Is it narrow enough to even measure, lol?

The spread is good, it will light up your walls & ceiling, but intensity isn't suitable for even 10" of water. I will take readings again, but if I remember correctly, I was getting less than 100PAR in 10" of water with the 100w chip 24" above the surface. Let me check again in case I'm remembering wrong.
 
Back
Top