The disadvantage to mixing warm white with cool white

[churchhaze]

Let's assume you are starting fresh.

You decide on half 2700k 80cri leds and half 4000k 80cri leds. Assume you picked a vero series cob. Both the 2700k and the 4000k use very similar phosphors. The main difference is one has more 450nm while the other has more "wide-band" output.

Here's is a breakdown of this 50-50 strategy:

1. Put the 2700k cri phosphors on half the blue dies to convert most of the blue to the wide band.

2. After converting half of your blue photons to a wide band, you decide there isn't enough blue so for the other half of your blue photons, use 4000k 80cri phosphors so less blue is converted.

Here is the problem with this strategy. There is a loss when converting higher energy photons (blue) to lower energy photons (wide band centered around yellow). If you know in advance you will need more blue photons, why would you convert the first 50% of your blue photons to wide-band?

It would make sense in this case to get all 3500k up front, assuming that's the amount of blue you want. Getting a cool white later makes sense if you decide in the future you need to supplement blue, but it doesn't make sense in up front planning.

 

[PetFlora]

I consider 4000k as NW

>= 5000k as CW

 

[churchhaze]

I'm really questioning my original reasoning on this. I can't tell if my argument actually makes sense.

This may be another case of stoner logic, but something about getting 2700k leds just to mix with blues doesn't settle with me.

Sorry for this pointless thread!! Nothing to see here!

 

[getsoutalive]

No, I think you are on point. Better to choose the proper color temp for all cobs than trying to use multiple colors to mix to your desired result. Mixing from different chips will by definition result in color temp "hot spots."

I doubt there is anyone here who can be very sure that, just as an example, 3500K is not going to be functionally equivalent to mixing 3000K with 4000K. That is assuming that they could be blended evenly over the canopy, which would require them to be on top of each other.

Don't think it is necessary to think about blue conversion efficiency to come to this conclusion.

 

[ballist]

If you wanted to increase the blue in the spectrum wouldn't the most efficient way to just add discreet blues? It should be more efficient than getting it from phosphor conversion.

 

[tenthirty]

The uniformity is just as important as spectrum. Adding discrete frequency leds cannot be good for the uniformity of light in both intensity and color variance.

 

[PetFlora]

If you wanted to increase the blue in the spectrum wouldn't the most efficient way to just add discreet blues? It should be more efficient than getting it from phosphor conversion.

I did just that, erring on the side of caution, but the primary diodes in the S600 I am using are 5000-6500 & 3000

I added 4 ~ @ 450/bar + 3 @ 625/bar

Was this necessary? Dunno, but it was safe

 

[MrFlux]

If you wanted to increase the blue in the spectrum wouldn't the most efficient way to just add discreet blues? It should be more efficient than getting it from phosphor conversion.

The blue light from white LEDs is the portion of light that didn't get absorbed by the phosphor. There is no efficiency loss really, it's just as good as getting the blue from monochromes.

 

[PetFlora]

The blue light from white LEDs is the portion of light that didn't get absorbed by the phosphor. There is no efficiency loss really, it's just as good as getting the blue from monochromes.

Not exactly

There is a big difference in the blue spectrum peaks in a CW- NW, or WW, as well as each manufacturer's white diodes bin can be higher or lower

 

[MrFlux]

About the mixing of whites, I can think of three ways to look at it.

1. When mixing whites you put a little bit more phosphor on some LEDs and a little bit less on the others. Like CH said you could have put some in-between amount of phosphor on all the LEDs to get the same effect.

2. Using the Planckian black body locus: When mixing for example 2700K and 5000K the result will be somewhere on the blue line:
The blue line hardly deviates from the Planckian locus so there is very little difference with mixing whites of picking some in-between color temperature.

3. Doing the math on the various spectra. Here are some examples

Here you can see that it can make a difference whether to mix or not. It's not that one option is better than the other just different.

 

[MrFlux]

Not exactly

There is a big difference in the blue spectrum peaks in a CW- NW, or WW, as well as each manufacturer's white diodes bin can be higher or lower

That goes for monochromes too, not sure what your point is. My point was that it's just as efficient to get your blue from whites as from monochromes.

 

[PetFlora]

It depends on what the white's blue range is

 

[Fiveleafsleft]

After looking at spectras of several white cobs and also combination of 5000+2700 etc I found that CXA 4000 as blue base gave me much bang for buck in terms of lumens, par-values and spectral distribution. What I didn't quite notice, was that the blue band is a kind of narrow. So how bad is this? Will now pay my price for cheating with mixing! time will tell I guess but theorizing is a part of the fun!

 

[getsoutalive]

Make no mistake, your light will perform very well. You will not likely notice much if any difference.

My point was more that since we still do not have a real answer as to what the perfect spectrum is, there is little to be gained from mixing and with the various color temps available, it is simple enough to just pick the one color that most closely matches what you hope to achieve and not have to worry about blending.

Until we get more info on that perfect spectrum, if it even really exists given the many variables in everyone's gardens and strains, it is all just educated guessing. Sure, more blue will likely keep your plants shorter and more red is wanted after the stretch, but beyond that, I believe that intensity is more important than the slight color variations between COBs as long as most wavelengths are being covered.

This is why the early red/blue lights weren't worth a damn. Small LEDs even in great number really cannot give the intensity required, and buying even 7 different spectrums of low powered chips spinkled among hundreds of the red/blues is just not effective in flower phase.

 

[MrFlux]

After looking at spectras of several white cobs and also combination of 5000+2700 etc I found that CXA 4000 as blue base gave me much bang for buck in terms of lumens, par-values and spectral distribution. What I didn't quite notice, was that the blue band is a kind of narrow. So how bad is this? Will now pay my price for cheating with mixing! time will tell I guess but theorizing is a part of the fun!

The width of the blue peak is kind of typical compared to other white (and blue) LEDs so that's not something to worry about. It has proven to work very well.

Some slight hairsplitting is the location of the blue peak, at 455nm. The blue action spectrum has a local dip there which means that the blue light will be somewhat less effective with blue-light mediated responses like phototropism and stomata opening (see graph below). Since the CXA 4000K has plenty of blue that is again not something to worry about.

I hope somebody will try the CXA 4000K (with no additions) for flowering sometime. It should work very well, especially for higher irradiation. Ought to be a nice change from all the 3000K grows here and the 4000K can be gotten two bins higher.

Source: REQUIREMENTS OF BLUE, UV-A, AND UV-B LIGHT FOR NORMAL GROWTH OF HIGHER PLANTS,
AS ASSESSED BY ACTION SPECTRA FOR GROWTH AND RELATED PHENOMENA

 

[salmonetin]

i missing an graphic for tomatoes or cannabis, and an comparative graphics with CXA 4000 for each diferents bins ...¿posible?

for better visual understanding

sorry my bad english

saludos

 

[Fiveleafsleft]

The width of the blue peak is kind of typical compared to other white (and blue) LEDs so that's not something to worry about. It has proven to work very well.

Some slight hairsplitting is the location of the blue peak, at 455nm. The blue action spectrum has a local dip there which means that the blue light will be somewhat less effective with blue-light mediated responses like phototropism and stomata opening (see graph below). Since the CXA 4000K has plenty of blue that is again not something to worry about.

I hope somebody will try the CXA 4000K (with no additions) for flowering sometime. It should work very well, especially for higher irradiation. Ought to be a nice change from all the 3000K grows here and the 4000K can be gotten two bins higher.
View attachment 3253088
Source: REQUIREMENTS OF BLUE, UV-A, AND UV-B LIGHT FOR NORMAL GROWTH OF HIGHER PLANTS,
AS ASSESSED BY ACTION SPECTRA FOR GROWTH AND RELATED PHENOMENA

Thank you for that very clarifying answer MrFlux! I was trying to match the data from your thread Cree CXA analysis with prices of available cobs, and find the best deal. To me it looks like 5000K has a better spectrum of deeper blue but on the other hand i would need a ridiculous amount of red's to make that to a good flowering lamp.

I've built 8 cobs with 4000K as a base. They will be used from seedling to harvest in a perpetual 12/12 from seed grow. Four of them, the more flower orientated, are mixed with 2 660's and 2 630's on each panel. The other ones are built with 2 630's, one 660, and one blue.

The reason for the blue diodes is that i hope they should help against stretch in early flower. Height is a concern for me since I'm limited to having about 65 cm-flowers maximum, height wise.

But if it would contribute to the knowledge on this forum, i would gladly strip the more veg orientated panels of 3 watters and leave some plants to finnish there in order to see how they perform by them self..

 

[Fiveleafsleft]

Make no mistake, your light will perform very well. You will not likely notice much if any difference.

My point was more that since we still do not have a real answer as to what the perfect spectrum is, there is little to be gained from mixing and with the various color temps available, it is simple enough to just pick the one color that most closely matches what you hope to achieve and not have to worry about blending.

Until we get more info on that perfect spectrum, if it even really exists given the many variables in everyone's gardens and strains, it is all just educated guessing. Sure, more blue will likely keep your plants shorter and more red is wanted after the stretch, but beyond that, I believe that intensity is more important than the slight color variations between COBs as long as most wavelengths are being covered.

This is why the early red/blue lights weren't worth a damn. Small LEDs even in great number really cannot give the intensity required, and buying even 7 different spectrums of low powered chips spinkled among hundreds of the red/blues is just not effective in flower phase.

Ok. Thanks for those words to! Sometimes when you read all these threads, with really clever theories, you get the impression that everything hangs on getting "that special, magic, scientifically proven, yet kind of esoteric spectrum". Don't get me wrong, I'm a kind of drawn to this esotericism, but sometimes it's also very nice and kind of refreshing with these more down to earth statements!

 

[MrFlux]

i missing an graphic for tomatoes or cannabis,

This was the closest I could find. The blue action spectrum seems to be very similar for most plants though, just like the McCree curve.

and an comparative graphics with CXA 4000 for each diferents bins ...¿posible?

for better visual understanding

Okay. Let's choose from the CXA3070 group, lowest CRI, highest bin. So it will be 3000K AB bin vs 4000K BB bin. They cost about the same on Digikey and seem to be equally hard to get. The conditions are the minimum flux levels at 25C and 1.925A drive current.

Result:
3000K has a radiometric efficiency of 37.1% vs 42.1% for the 4000K,
PPF is 1.73 µmol/J vs 1.91 µmol/J,
7.9% blue photons between 420-480 nm vs 15.8%

 

[MrFlux]

But if it would contribute to the knowledge on this forum, i would gladly strip the more veg orientated panels of 3 watters and leave some plants to finnish there in order to see how they perform by them self..

IMO just through every light you have at your plants...