Minimalistic multichip DIY LED build

[jerpa]

Buying 100 multichips without testing the spectral mix first is risky to say the least. My other suggestion would be to select a multichip and matching drivers that will give you room to dim individual colour channels 50% without compromising minimum PAR values. In other words, if your tank needs 100w multichipsfor intensity, you can't start dimming it down to 50% to get the spectrum you want.

This was kind of my point. I have no problem running twice as many LEDs at half power. I think it will also help with the shadowing you encounter with such a small point source. I just need to be confident I can get a color that is pleasing to my eye period.

My honest concern is the rendering of warmer colors under these emitters. This will greatly impact whether I see any need for lower Kelvin emitters. Obviously the Kelvin isn't everything as I've seen some incredibly yellow cool white LEDs that don't render pinks and reds for squat.

 

[Psionicdragon]

Sorry it got jumbled.

Which of these will work for a 50w ??

Model BG-50-12 BG-50-24DC output voltage 12V 24V
Output voltage error ±1% ±1%
Rated output current 4.2A 2.1A
Output current rage 0-4.2A 0-2.1A
Wave and noise 100mVp-p 100mVp-p
Inlet stability ±0.5% ±0.5%
Load stability ±0.5% ±0.5%
DC output power 50W 50W
Efficiency 80% 82%

Power Supply 1:

Model BG-50-12
DC output voltage 12V
Output voltage error* ±1%
Rated output current 4.2A
Output current rage 0-4.2A
Wave and noise 100mVp-p
Inlet stability ±0.5%
Load stability ±0.5%
DC output power 50W
Efficiency 80%

Power Supply 2:

Model BG-50-24
DC output voltage 24V
Output voltage error* ±1%
Rated output current 2.1A
Output current rage 0-2.1A
Wave and noise 100mVp-p
Inlet stability ±0.5%
Load stability ±0.5%
DC output power 50W
Efficiency 82%

 

[Lassef]

That does nothing in showing me how it would look in a multichip alongside 1:1 royal blue, on your wood floor, with a camera. The fact it isn't somewhat yellow means yes, there is plenty of blue in it. It also goes to show the 20000k label means nothing. If you want to use basic math skills, some blue + more blue added alongside it = lots of blue.

Your insults are starting to get really tiring. It does nothing to promote the discussion here. I'm not looking to start an argument, but I will unload on you if you continue with the insults.

Two more photos (published before)

1.5:1 ratio of royal blue 445 nm and 20 000 K epistar based chip from AC-RC

Here is both of them in the same spot - ratio around 1:2.1

Is not my wooden floor you should look at - its the color pattern of my carpet that is interesting. It shows how the different colors apear - and yes you can compare through this pictures because there is a reference picture - how it looks like in bright daylight. If your screen not is like mine, or your color perception is not like mine you got the same fault in both pictures and you can se differences between what colors the carpent have in daylight and in the light from the chips. It is not the actual colors that its important - it is the differences between the different pictures that is important. and you have to trust me when I say that the pictures is not manupulated.

Yes blue + more blue = much blue but in this case the other colors in the 16 000 K chip mask the blue wavelenghts so you do not se them with your eyes - if you are familary with the RGB system - you understand what I mean (Red + Green + Blue = White) The blue wavelenghts is there for the corals photosynthesis but you only se white if its ideal. Now its not ideal - there is a yellow tint and you need more blue to get a more white appearance.

What I have try to say all the time is in spite of that AC-RC chip of 10, 16 to 20 000 K content large blue peaks they are for the eye rather white because of the other colors in them. In fact they have all a yellow tint. Its seems like you do not trust what I´m saying and all the time refer to a 14 k chip based on Epistar of unknown orgin. I´m not talking about that chip and do not try to tell you that your are wrong on this. I relay on what you are saying but according to the chip´s I´m familary with the AC-RC chips- your statement is not true.

Sincerely Lasse

Not state before but the pictures are all taken with manual white balance and manual time and shutter so you can compare.

 

[Lassef]

If this thread is closed it will be due to the AC-RC commercial and "group buy", not over a white vs. blue argument.

I can't comment on other people's aesthetic preferences, but I will say that the three blue chips being considered here are all at the less visible end of the spectrum.

I like the 480nm "blue" chip because it is more white & bright. The best Windex bang for the buck is out of 465nm royal blues. You need fewer low nm (blue) chips when you use blues at the higher end of the spectrum (465-480). Of course a broad spectrum of blue is ideal, and I'm not saying your choices are not beneficial for coral growth and pigmentation, they just won't be as "blue" to the human eye as some are predicting.
The 15,000k white has a slight purple/pink tinge to it when run with blues, even when driven on separate channels. Your proposed mix may seem heavy in the blue spectrum on paper, but considering the low nm aspect of them, they will be overpowered by the whites.

Thanks - this I have try to say but obviously not succeded with.

For me - its first of all a chip to promote growth and because there is invidual legs I can mix as I wan't so it looks well also

Buying 100 multichips without testing the spectral mix first is risky to say the least. My other suggestion would be to select a multichip and matching drivers that will give you room to dim individual colour channels 50% without compromising minimum PAR values. In other words, if your tank needs 100w multichips for intensity, you can't start dimming it down to 50% to get the spectrum you want. .

Yes - thats one way - the other is to run with full power (to promote growth) when you are not at home and adjust for pleasant mix when you are at home

I would also like to offer a caveat with regard to fitting five individual colour channels on one multichip unless it is made on the much larger 300w frame. The blues you are considering are very close together so there is no need for individual colour control. If you aren't using reds, then you only need two channels. This will give you a dawn/dusk sequence of white & blue exclusivity.

Yes - but it will be possible to conect the different legs in that way but to have five separate legs is a choise of freedom

Sincerely Lasse

 

[sfsuphysics]

Most of us will only be using 80% or less of the chip once we get it dialed into where we want it. Some more blue, others more white, etc.

I was under the impression that we were ordering a chip that could be tuned to a WIDE range of color renderings so that it had a broad appeal, not a "pre tuned" chip designed to be run at 100% (or near) for each color

I'm with you on this 100% (no pun)

To compare to commercial products, this is basically what the Kessil A350 fixture knobs are for, they're not for dimming purposes but to dial in what you want as far as color output, and more often than not you will not use 100% of the total capacity. You want max power with a fixed color you go towards the Ecoxotic cannons or other similar products.

Why bother with 5 independently controllable channels of colors if you already have a chip that has the coloration that you want?

Seems the people are pushing for blue white blue white blue, which... they already make, in a non-custom priced color, now I will concede the shades are different, but it's not much of a color tunable chip if you changing shades of blue.

I'm not going to stomp my feet and declare I won't order unless I get my way, don't want to be "that guy" . However it is a question that others should ponder. What exactly are they wanting/expecting from a chip that NEEDS to be ordered in bulk numbers? Understanding the NEED for number means some level of compromise must exist... unless of course everyone just happens to agree on the same thing

 

[Lassef]

Having two seperate pictures does not realy give anyone a true confident comparison. The only true way would be to have these LED's in the same picture frame but seperated so the light from one would not effect the light from another. If you simply take seperate photos most digital cameras will automaticly correct for the color balance.

Forgott to say that I allways operate my camera in manual mode. IE I use a constant white balance in this pictures.

Now in reality you can take probably a dozen different brands of LED's that are rated at the the same color temperature and get extremly different results. The difference between a 14,000K and a 20,000K light source is so close that manufacturer specs will easily overlap these values. So it is very possibe to get a 14,000K LED from manufaturer A that is Bluer than a 20,000K LED from another manufacturer.

Claiming a 20,000K light source is not blue to me only means that the light source is not truely a 20,000K light source. The truest White light is in the 8,000K too 10,000K range. Anything above 10,000 K will move the spectrum more to the blue side as the rating increases. There is very fine line from the infinity point where all the light is blue and 20,000K. Simularly anything under 8,000K will move the spectrum to the red side. However there is much wider difference between a 1,000K light source and a 0 K light Source where there is no detectable Blue or Green Light.

True, my references has all the time been to the chips that I have experiences of.

Another thing to keep in mind is that your eyes will automaticly adjust to different color temperatures. If your in a room with all 3,200 K incadescent light bulbs it will appear to you that is all white light, Simularly if you go to room with all 12,000K floresent light it will pappear to you as all white light. But if you put the two light sources into the same room you will see the 3,200K light as Orange light and the 12,000K light as Blue light.

True

Sincerely Lasse

 

[TropTrea]

As I see this if the light is not prediminantly Blue then it is NOT a 20,000K LED. The manufacturer may claim it is bit in reality it is a much lower K chip. The fact that you say the tint is yeallow to me means that the light source is less than 8,000K.

If you go back to T-5 lights and have a fixture with many different K temp bulbs you will see that the bulbs themselves look like different colors. A 6,500K bulb will look very Yellow compared to a 10,000K bulb and a 14,000K bulb will look very blue compared to a 10,000K bulb. But if you put a 14,000K bulb next to a 20,000K bulb the 14,000K will look white and the 20,000K will look blue.

A lot of this personal eye preference. However a true 20,000K light source has very little Green or Red light in it. So balancing out Blue LED's with a 20,000K light source will require a lot of 20,000K light. For some people a 20,000K spectrum is too blue without the adittion of any blue light to start with. Idealy you want a light source roughly ezual in the Green and Red colors to balance against a Blue light source. The amount of the red and Blue is something that will varry between individual taste but you also need caution as excessive red light especialy 685 nm light will bleach many corals.

 

[sfsuphysics]

Oh, and I think Mr. Wilson had a fantastic point. This thread is way too good of an informational resource to risk having it closed because it gets turned into a group buy thread

 

[Lassef]

Sorry it got jumbled.



Power Supply 1:

Model BG-50-12
DC output voltage 12V
Output voltage error* ±1%
Rated output current 4.2A
Output current rage 0-4.2A
Wave and noise 100mVp-p
Inlet stability ±0.5%
Load stability ±0.5%
DC output power 50W
Efficiency 80%

Power Supply 2:

Model BG-50-24
DC output voltage 24V
Output voltage error* ±1%
Rated output current 2.1A
Output current rage 0-2.1A
Wave and noise 100mVp-p
Inlet stability ±0.5%
Load stability ±0.5%
DC output power 50W
Efficiency 82%

Need your chips specification also

Sincerely Lasse

 

[bhazard451]

Did you not read not even three posts above yours, where Troptrea basically (and correctly) pointed out that pics mean squat in the overall scheme of things?

Since you keep going on about this with pictures, I need to point something out to you... your own tank, shown with pictures.

I DO NOT WANT MY TANK TO LOOK LIKE THIS. It is so heavy on yellow, brown, blue, and pale green, that I don't think I can find a reddish hue anywhere if I tried. This is not what I want at all. This gives me further encouragement to include 6500k or even lower to achieve the look I want. You can fight me tooth and nail in your belief that the higher K ac-rc chips include enough red/green/rainbow unicorn in their spectrum for color, but if we are to go by pictures (which is almost pointless to begin with), your tank is proof to me this is not the case.

I started with a dull CW:RB fixture, and I will not revert in going back to one.

Here is some pictures from december last year (2011)

Sincerely Lasse

 

[sfsuphysics]

However a true 20,000K light source has very little Green or Red light in it.

Actually, they have a surprising amount of green and red. Perhaps as much as 50% of the blue energy in green (I'll admit I don't have a blackbody spectrum handy to look at exacts, but it is not insignificant), while red is less, there's still a bunch of red photons flying around from a 20kK object.

 

[mr.wilson]

Thanks - this I have try to say but obviously not succeded with.

My Mother is Swedish so I was able to understand half of what you were saying

Yes - thats one way - the other is to run with full power (to promote growth) when you are not at home and adjust for pleasant mix when you are at home

Yes, this a point that Bean brought up recently. LED gives you the control to offer both "show" and "grow" lighting at different times of the day. Many metal halide users run crisp white lighting (Iwaskai 6500k or Ushio 10,000k) during the day for growth and less realistic 14,000 & 20,000k metal halides at night when they get home from work. It balances form & function.

As much as I'm not crazy about red light during the day, I'll admit it looks super cool at night with only 10% power. Of course the fish look much nicer under a bit of red light. The LED phosphor coatings vary so much, I think a lot of people get hung up on what they know as 14,000k metal halide light. A 14,000k LED can look much different and I find they have a clear red or violet colour at times.

For most of us, it isn't a contest to grow coral the fastest, it's more about the general aesthetic, so we end up with some very unrealistic blue lighting.

 

[TropTrea]

Bhazard:

I will agree with you on pictures above being weak in the red part of the spectrum. However keep in mind everyone has there own color taste.

As far as the warmer whites go I use a Neutral White which is my preference. I like a ratio of 1 neutral white to 3 Blues. However I mix equaly royal blues with blues.

With some experimenting I did I found that a ratio of 3 to 1 with neutral whites (5,000K) gives the same effect as a ration of 1.5 to one with cool whites 10.000K, simularly if you use warm whites 3,000K a ratio opf 4.5 blues to one white gives a simular effect in the Blue Red balance. But this is my personal color taste. with 20,000K whites even all whites to me would lack the red end of the spectrum.

however keep in mind that if Laseff's LED's in these pictures are showing this much yellow I truely doubt that they are true 20,000K bulbs. Yellow is a combination of Green and Red as seen by the eye and a 20,000K should not have much red in it at all so the overall tint should be in the blue green wange rather the Yellow range.

 

[Lassef]

My Mother is Swedish so I was able to understand half of what you were saying

Sincerely Lasse

 

[Lassef]

Bhazard:

However keep in mind that if Laseff's LED's in these pictures are showing this much yellow I truely doubt that they are true 20,000K bulbs. Yellow is a combination of Green and Red as seen by the eye and a 20,000K should not have much red in it at all so the overall tint should be in the blue green wange rather the Yellow range.

There is no 20 000 K bulbs in these aquarium pictures only 14 000 to 16 000 K of the brand I use. It is also a ratio 2 white to 1 RB.

Sincerely Lasse

 

[gtfcniles]

That does nothing in showing me how it would look in a multichip alongside 1:1 royal blue, on your wood floor, with a camera. The fact it isn't somewhat yellow means yes, there is plenty of blue in it. It also goes to show the 20000k label means nothing. If you want to use basic math skills, some blue + more blue added alongside it = lots of blue.

Your insults are starting to get really tiring. It does nothing to promote the discussion here. I'm not looking to start an argument, but I will unload on you if you continue with the insults.

Hardly an insult.

 

[gtfcniles]

Let me clarify:

-There is potiential for a group buy. That is all. It's all potential. I am not asking for cash commitments or am keeping any list of potential buyers.

-For anything to remotely happen in terms of any buy we needed to choose a first color configuration. I believe we have reached that with the following color configuration. (This will not change):

1) 20 x 10000K
2) 20 x 455nm
3) 20 x 445nm
4) 10 x 420nm, 10 x 430nm
5) 20 x 15000K

-We must decide on second color configuration. I need solid input from all of you. Not pettiness. This is my choice for a far warmer white color configuration:

Copy and paste the format below into your response with your changes:
1) 20 x 3300K
2) 10 x 445nm, 10 x 455nm
3) 20 x 5500
4) 10 x 420nm, 10 x 430nm
5) 20 x 12000K

-The panel size will not change. We will be using the standard 100W multichip panel. Again, this will not change.



I log in a few times a day to gauge progress. I log in with my noise filter turned on very high. Unless you have a valid point it gets tossed into the noise pile. Outliers will be tossed as well. I have to drive this to an end simply because I want my custom multichip panel in my hands as much as you do.

I went into this knowing everyone will not be happy. Based on offline responses, the happy campers far and away outweigh the handful of unhappy campers. Keep it civil else I'll support the closure of the thread myself

So... about that second color configuration... let me see your input.

Ron

Put me down for two of option one please Ron (if you have a list).

Thanks,

Chris

 

[jerpa]

1) 20 10000K
2) 10 445nm+10 455nm
3) 20 6500K
4) 20 420nm
5) 20 10000K

That's my input at this time. Hopefully next week I will have more insight based on personal experience with these LEDs.:thumbsup:

 

[bhazard451]

Haha. There is no list. When the time comes for listing anything, you will know. I am only asking for input on a second color configuration using the format below. Before any list is even considered we must decide on a second color configuration:


Input goes here for the second warmer white color configuration.
Copy and past with your input:

1) 20 x ????
2) 20 x ????
3) 20 x ????
4) 20 x ????
5) 20 x ????

The first chip is pretty much in agreement with everyone. The second will be much harder to figure out, as you cant mix 10 + 10 whites per leg or have them close together. That adds a huge limitation in trying to reach a 2:1 or even a 3:1 ratio. The brightness of their lower k leds come into play as well, as the standard 2:1 ,3:1 most are used to may look totally off if these have whites have low output.

Some testing would need to be done before anyone commits to anything on it.

Traditional thinking trying to incorporate warm whites would need to be 60 to 20 to 20 b:w:v, but I don't think these warm whites will have enough output, and I don't think using only 3300 would look good. This below has the potential to look disastrous if the whites overpower the blue, or vice versa.

1) 20 x 3300k? 5500-6500k?
2) 10 x 445nm, 10 x 455nm
3) 10 x 420nm, 10 x 430nm
4) 10 x 445nm, 10 x 455nm
5) 20 x 5500-6500k? 10-15k?

The Maxspect R420r's 16000k is the type of setup I would like to mimic into a 14k multichip, excluding all the extra blue 465-485.. or not? The 10000k version uses double the cool whites and less blue, but I'd rather limit CW since we have a chip with it already.

6 382 lm @ 1300mA Cool White 8000K
6 1520mW @ 1300mA Royal Blue 450-465nm
6 294 lm @ 1300mA Warm White 3000K
9 63 lm @ 860mA Blue465-485nm
6 1520mW @ 1300mA Royal Blue 450-465nm
6 1100mW @ 1300mA Super Actinic 410-420nm

 

[BeanAnimal]

For the record, I am not trying to be petty, or take my toys and go home. I just don't think many folks understand the logic some of us are trying to put forth. Jerpa has articulated the point better than I have but we are pretty much on the same page.

Many of us are looking for something that can be versitile and replicate dawn, dusk, mid day, reddish skies, whatever. If we need more punch we will ad more chips. We are not looking for a pre-configured "dream ratio" that only needs to be ramped up and down for intensity.

So instead of fighting what is almost sure to be a losing battle... I will just try whatever the rest of you decide on. If it is too blue and can't replicate the colors that I want, then I will have to look elsewhere or take matters into my own hands with a group of like minded folks...

CHIP 1 is clearly BLUE on paper and can only go from bluish white to deep blue. I am not sure that is what many of us want, but I will certainly try one. I am not so closed minded that I wont give it a chance.

If CHIP one is already locked in, then why not throw the kitchen sink in CHIP 2 and make a truly broad spectrum chip?

That looks interesting: It also avoids that 6500K "dead flat" spectrum that many folks have complained about with the CW and favor NW and WW for instead.

1) 20 x 3300K
2) 10 x 445nm, 10 x 455nm
3) 20 x 5500
4) 10 x 420nm, 10 x 430nm
5) 20 x 12000K

I would even be happy with:
1) 20 x 10000K
2) 20 x 455nm
3) 20 x 445nm
4) 10 x 420nm, 10 x 430nm
5) 20 x 5500K

Or some similar variation. The above gives us both bluish white and the NW that people have had good luck with. I can add the "reds" and "turquoise" along side if really needed then.