Minimalistic multichip DIY LED build

[tomservo]

Whoops, I was looking at the 5 channel one, noticed i had 5 colors so... well ok I screwed up


1) 25 - whites ("10000k"), 5 - red
2) 20 - 420nm
3) 25 "royal blue" 455nm
4) 25 "blue" 445nm

There!

Groups of 10 are mandatory unless you want to mix colors onto different channels... It's a 10 rows x 10 chips setup. So you have total 10 possible different strings. I seriously doubt they would even consider mixing more than one color on a row, even if it wouldn't screw up the voltage, especially with red chips, which run at ~2.4v.

I've seen colors wash out firsthand with the use of only 10k and higher epistars. If I didn't, I would otherwise be in agreement with you. Unless a multichip somehow makes the same led perform differently, I'd have to assume the same would apply.

Actually, epistar only makes single color chips; it's up to the (assembler?) to apply a phosphor coating to make a blue chip into a white LED. So yes, one company's epistar is not necessarily the same as another.

 

[bhazard451]

Groups of 10 are mandatory unless you want to mix colors onto different channels... It's a 10 rows x 10 chips setup. So you have total 10 possible different strings. I seriously doubt they would even consider mixing more than one color on a row, even if it wouldn't screw up the voltage, especially with red chips, which run at ~2.4v.

Actually, epistar only makes single color chips; it's up to the (assembler?) to apply a phosphor coating to make a blue chip into a white LED. So yes, one company's epistar is not necessarily the same as another.

This is true.

Why though, is the consensus here that a low CRI 10k 15k led, would be better off than including some lower kelvin, higher CRI leds? The "biological and aesthetic" reasoning, being that the lower K led is closer to natural sunlight. The counter argument is that everything is mostly blue under 30ft anyway, but that doesn't work for an aesthetically pleasing tank. Its a personal preference thing, as I also believe it will boost red and green output and visuals without requiring extra red leds on a chip.

There IS a noticeable effect when red leds are introduced into cw:rb fixtures that don't have them, but if you can include more of the red spectrum within the whites being used, you can achieve the same effect without spotting or overpowering other leds. You cut down the chances of growing out algae quicker this way as well.

 

[tomservo]

Well, high CCT kelvin doesn't mean there is too little red off hand; You can have a high kelvin temperature and have nothing but blue and red spectra. With LEDs and fluoros (to some extent) you really need to see kelvin and CRI both to get a grasp on the spectra. For example, a 70 CRI 5000k vs 85 CRI 5000k, the 85 CRI is the same color temp but has a lot more emphasis on red/orange while the 70 CRI will be more green/yellow. Both likely have a similar amount of blue but it wouldn't be a guarantee.

You can't assume that a high kelvin light source is insufficiently supplied with red (or whatever) just based on that one calculated value.

 

[Lassef]

I understand you reasoning but then why do we have an emitter dominated by blue? This ratio would be just as washed out no?

In a way your might be right. The ratio in the suggested chips is around 1.5 blue to 1 white and with my experiences I prefer at least 1:1 when using high Kelvin whites. But my experiences with the 420 cip´s is that you do not see so much of this blue wavelenght - for your eyes it does not dominate - for the eyes this ratio with 1 pcs 420, 1 pcs 445 and 1 pcs 455 to 1 pcs 10 000 K and 1 pcs 16 000 K will be around 1:1 - if you understand what I´m trying to say. In my fixture I use mostly 16 000 K and in this chip it is more of the 10 000 K. Because the different colors also will be invidual managed - you are able to take down the blues a bit if the light not please your eyes. I´m convinst of that this will be a good start.

@ bhazard451. You do not understand what I´m saying. No one take a 6 500 K MH bulb and compesate with Actinic just to get the right light for his or her taste. Most of the reefers start with a 10 000 K or higher MH bulb. If we have the possibilties I belive that we should do the same with LED´s.

I have tested with 6 500 K (and 3 500 K) also and I´m not pleased with the light either in a biological or aesthetic way.

Sincerely Lasse

 

[gtfcniles]

Any news ron?

 

[Sweeper65]

Wow. First want to say thanks for this tread. Took a long time to read but extremily informative.
I am trying to get back into this hobby after a long time out and this is very helpful to me as I plan on going LED for my lighting and mutli chip seems like the way to go.
With that said you guys are lossing my on the spectrum.
As I read this it seems like 10k falls in the middle and meets the bio needs, so is all 10k not appealing to the eye? do you needs the 6500k for a bio reason?
Any insite you can provide to help bring along those of us that are not as versed here is apprichaited.
I know there are tons of threads on spectrom on the forum but it seems like choices are different for LED and the other rules may not apply. From what i'm reading even single bulb LED and multi chip are different.

 

[gtfcniles]

I've had confirmation from AC-RC that they can provide a 5 leg 100w chip...!

 

[tronyony]

Lasse,

The color I'm trying to achive in my tank is what maglofster's Elos 120 looks like in his video. Do you think with the LED's I purchased I can achieve this or do I need to purchase additional LED's?

Thanks,
Tony

 

[bhazard451]

Ron,

I'll commit to one or two of the chips you create with ac-rc. The ability to use 1 chip over my 34 gallon tank with that mixture is very intriguing. It'll be a good test alongside my 180w 3w Cree unit. If I like it better, I'll commit to another 4 for my 120 gallon.

 

[gtfcniles]

A 5 separate channel multichip panel has been confirmed by ac-rc. Good news!!!

Ron

Yeh he emailed me too (see above), so what colour combination are we going for? I really do think that this would give us the best results but i'm open to peoples views:

1) 20 x 445nm
2) 20 x 10000k
3) 20 x 420nm
4) 20 x 15000k
5) 20 x 455nm

God I want one to test now! :spin1:

 

[BeanAnimal]

I think the 5 leg multichip is much closer to something I would be ready to order, and use 4 of over my 120 gallon tank. It is still lacking in whites though. I will never run Epistars above 10k again, as they simply do not look as good as using warm white and neutral variants mixed with Royal Blue and Violet.

If Ac-rc would make the chip for me with 6500k and 3500k instead, 2:1 blue to white and some violets I think I'd try two of them on one side of my tank. Actually, I'd order all 4 to help the group buy.

I think my point may be being missed.

• The problem with "linear" or "clustered" discrete builds appears (no pun) to be "disco" and/or uneven color rendering.
• To me, the ONLY fixtures that cast pleasing light are those that have 5, 6 or more colors.
• If you want mnore WHITE than run the white legs at higher current than the blue or visa-versa.
• You can always add more CHIPS, in doing so you can get the color balance and intensity you need AND better coverage edge-to-edge

What I see (again not picking on anybody) is a group of folks going through a lot of trouble to replicate (in multi-chip) what most of us already know is not pleasing in a discrete emitter design.

So I ask again, why design this thing only to have to add other colors around it in the form of single eemitters? I would rather buy and run (4) chips at 50% and have the ability to dial up ANY color and INTENSITY (due to the extra headroom of a 50% design) than run (2) chips at 90% with limited spectral adjustment available.

To those who don't want "red", then don't drive that channel...

The point being that we are creating broad spectrum point sources and can use as many as needed to dial up the color and intensity we need, with no single emitters and no disco.

It follows, that if I were going

 

[Sonny n Colleen]

Ronreef,

I sent you a PM.

 

[ronreef]

Let's narrow it down to the following:

Option 1:
1) 20 x 445nm
2) 20 x 455nm
3) 20 x 420nm
4) 20 x 10000K
5) 20 x 15000K

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

Option 3: Many arguments for and against 6500K. So, let's compromise and add 6500K. It adds a little red to the spectrum and overall color depth to the spectral output.
1) 20 x 445nm
2) 20 x 455nm
3) 20 x 420nm
4) 10 x 10000K, 10 x 6500K
5) 20 x 15000K

Option 4:
1) 20 x 445nm
2) 20 x 455nm
3) 10 x 420nm, 10 x 430nm
4) 10 x 10000K, 10 x 6500K
5) 20 x 15000K

Please provide your vote.

Option 1, 2, 3 or 4?

I will wait to here all input. One thing we have is time. Time to make the better decision.

 

[gtfcniles]

Let's narrow it down to the following:

Option 1:
1) 20 x 445nm
2) 20 x 455nm
3) 20 x 420nm
4) 20 x 10000K
5) 20 x 15000K

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

Option 3: Many arguments for and against 6500K. So, let's compromise and add 6500K. It adds a little red to the spectrum and overall color depth to the spectral output.
1) 20 x 445nm
2) 20 x 455nm
3) 20 x 420nm
4) 10 x 10000K, 10 x 6500K
5) 20 x 15000K

Option 4:
1) 20 x 445nm
2) 20 x 455nm
3) 10 x 420nm, 10 x 430nm
4) 10 x 10000K, 10 x 6500K
5) 20 x 15000K

Please provide your vote.

Option 1, 2, 3 or 4?

I will wait to here all input. One thing we have is time. Time to make the better decision.

Why in that order ron for option 1? (see my earlier post).

 

[BeanAnimal]

Why are the only options to vote on all blue?

 

[bhazard451]

I think my point may be being missed.

• The problem with "linear" or "clustered" discrete builds appears (no pun) to be "disco" and/or uneven color rendering.
• To me, the ONLY fixtures that cast pleasing light are those that have 5, 6 or more colors.
• If you want mnore WHITE than run the white legs at higher current than the blue or visa-versa.
• You can always add more CHIPS, in doing so you can get the color balance and intensity you need AND better coverage edge-to-edge

What I see (again not picking on anybody) is a group of folks going through a lot of trouble to replicate (in multi-chip) what most of us already know is not pleasing in a discrete emitter design.

So I ask again, why design this thing only to have to add other colors around it in the form of single eemitters? I would rather buy and run (4) chips at 50% and have the ability to dial up ANY color and INTENSITY (due to the extra headroom of a 50% design) than run (2) chips at 90% with limited spectral adjustment available.

To those who don't want "red", then don't drive that channel...

The point being that we are creating broad spectrum point sources and can use as many as needed to dial up the color and intensity we need, with no single emitters and no disco.

It follows, that if I were going

What other colors were you thinking of? Everything useful I can think of is being covered aside from red. I requested a 6500k option not for more white, but to get the added red spectrum without needing separate emitters. 10 reds could be very overpowering even if turned down, and I'm not sure 5 red/5 something else could be done on the same leg.

 

[bhazard451]

Let's narrow it down to the following:

Option 1:
1) 20 x 445nm
2) 20 x 455nm
3) 20 x 420nm
4) 20 x 10000K
5) 20 x 15000K

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

Option 3: Many arguments for and against 6500K. So, let's compromise and add 6500K. It adds a little red to the spectrum and overall color depth to the spectral output.
1) 20 x 445nm
2) 20 x 455nm
3) 20 x 420nm
4) 10 x 10000K, 10 x 6500K
5) 20 x 15000K

Option 4:
1) 20 x 445nm
2) 20 x 455nm
3) 10 x 420nm, 10 x 430nm
4) 10 x 10000K, 10 x 6500K
5) 20 x 15000K

Please provide your vote.

Option 1, 2, 3 or 4?

I will wait to here all input. One thing we have is time. Time to make the better decision.

My first choice is Option 3. I would also go with Option 1 if its decided on.

 

[gtfcniles]

What other colors were you thinking of? Everything useful I can think of is being covered aside from red. I requested a 6500k option not for more white, but to get the added red spectrum without needing separate emitters. 10 reds could be very overpowering even if turned down, and I'm not sure 5 red/5 something else could be done on the same leg.

Here's the reason why you can't have red (a email from a couple of weeks ago from AC-RC)

Chris, I don't think so that is a good color mixture. Let me expalin to you....the red chip which input voltage is 2.8V max but any other color are 3.6-3.8V max. If mix with 4% red so that all of other chips should decrease 26% voltage to accommodate the red in order to avoid the happen of burn it out. For example of 100W panel, it just can runs on 74W max under the color mixture you propose.

I don't know why people are keen on the 6500k, this range would promote algae growth, there isn't a single tank in the world that runs with a 6500k MH and even in a T5 fixture the lowest i've ran is a 10k.

 

[BeanAnimal]

What other colors were you thinking of? Everything useful I can think of is being covered aside from red. I requested a 6500k option not for more white, but to get the added red spectrum without needing separate emitters. 10 reds could be very overpowering even if turned down, and I'm not sure 5 red/5 something else could be done on the same leg.

fair enough...

I was under the impression that the turquoise and red were somewhat important for color appeal and balance to the eye and rendering of the corals.

Not complaining, just trying to advance the conversation

 

[BeanAnimal]

Here's the reason why you can't have red (a email from a couple of weeks ago from AC-RC)

Chris, I don't think so that is a good color mixture. Let me expalin to you....the red chip which input voltage is 2.8V max but any other color are 3.6-3.8V max. If mix with 4% red so that all of other chips should decrease 26% voltage to accommodate the red in order to avoid the happen of burn it out. For example of 100W panel, it just can runs on 74W max under the color mixture you propose.

Not if they were on their own leg... they would have nothing to do with the other legs

That said, the point about not being able to dim 10 of them enough is valid I suppose and maybe not worth the gamble.