Question: More Amperage = Higher K rating?

[Daimyo68]

So I'm looking at the Cree XP-G2 Cool whites. Spec sheet shows:
Cool White, 5000K - 8300K

I'm assuming running them at the minimum amperage (350mA) would produce the 5000k, and running them at max amperage (1.5A) would produce the 8300K.

Am I thinking correctly, or should I just go to bed?...

[jerpa]

They are referring to the range of the various bins available. If you order a generic XP-G2 it could be anywhere within that range. If you order a specific bin it will have a much tighter tolerance for the Kelvin rating. While different currents will have some effect on the spectrum it shouldn't be noticeable.

[Daimyo68]

OK, so according to the Cree Spec sheet (linked below) I am assuming the Cool White, XPGBWT-L1-0000-00H51, R5 is 8300k ?

http://www.cree.com/led-components-and-modules/products/xlamp/discrete-directional/~/media/Files/Cree/LED%20Components%20and%20Modules/XLamp/Data%20and%20Binning/XLampXPG2.pdf

And also with the XTE Cool white, R5, XTEAWT-00-0000-000000H51 is 8300k ?

http://www.cree.com/~/media/Files/Cree/LED%20Components%20and%20Modules/XLamp/Data%20and%20Binning/XLampXTE.pdf

Appreciate the clarification :)

[uncleof6]

Mentioning Kelvins (apparent color temperature—or how the eye preceives the light) in the same sentence as "spectrum" or the wavelength distribution, is about the same as speaking about the Mariana Trench, and where it is on Mount Everest...Mt. Everest might fit inside the Mariana Trench with room to spare, but there is no relationship between the two...

Varying the current will simply alter the wattage, if the voltage stays the same. Ohms Law. The light source will either get brighter, or dimmer, but it will have no affect on the "K" of the light source or the spectral distribution of the light source. The spectral distribution determines the "K" of the light source; the spectral distribution is determined by the mixture of junk (phosphors etc.) inside the light source, that the current excites so it radiates light—the same junk that is inside fluorescent lamps. Physics.

The problem here is folks buy light sources according to the "K" rating, rather than the spectral distribution. Kelvins refer to how the eye sees the light. The spectral distribution is how the photosynthetic critters see the light. The eye sees the light exactly opposite to the photosynthetic curve: peaks in the green and yellow, and lows in the blue and red.

The bottom line is a light source needs to have strong peaks in the "blue" wavelengths (for marine critters; red for terrestrial), and what the Kelvin rating of the light source is, does not matter a toot. Simply because it may look "blue," but not have any "blue" radiation output at all...Physics.

[Daimyo68]

I'm attempting to get a "crisp" white look instead of a "yellowish" look from my whites. The whites I run now, make the tank look very "drab" to my eyes if I run them more than 45% power, even with a 2:1 ratio of Blue & Royal Blue to white.

My thoughts are I can accomplish gaining the crisp look by either replacing the current whites with the Cree's, or add them to my current whites, and run the bridgelux at a lower %.

My current Bridgelux produce a 6500k
The Cree produces an 8300k (if the bin numbers I posted earlier are correct).

Their spectral output (curves) appear to be the same.

http://i1244.photobucket.com/albums/gg578/68Daimyo/LEDS_zps6cbfe6bb.jpg

[zachts]

this is the biggest issue with leds for reef lighting. they lack the red and green wavelelnths in the right combination to reproduce the crisp white light we might be used to from metal halide or flourescent. they instead focus on the comercial lighting spectrum used for area and room lighting not reef tanks as that is the main market using them presently.

you'll have to do some playing around but adding to your current setup I would suggest adding not the cree chips you mention, they will be still too drab. either replace with neutral white Crees in place of you whites or cheaper option supplement with a few luxeons and use a couple of the 2700k 90cri warm whites to help fill in some of the missing red spectrum. maybe one per 3 other whites as a start and see if your eyes need more.....

you could also try supplementing with red and green leds on dimmers to blend a "whiter" light. rgb strip lights might work for that purpose otherwise you usually need some sort of difuser or the red and greens will produce nasty dico-like shadows if you only use a few of the high power type.

much of it is as already mentioned up to personal perception and you just have to figure out what you like.

[Daimyo68]

I guess I'm trying (as I'm sure lots of people are) to replicate a halide. 14k pheonix.

I look at my friends Razor R420R 16k, and I like the way the color looks, it's the closest I have seen so far from a manufactured led fixture, imo. I looked into what they are using and it has the XTE Cool whites @ 8300k, Warm whites @ 3000k. RB, Blue and "actinic"... But then I look at my buddy's DIY using Cree Royal Blue, Blue and Cool whites, and it seems almost identical to the Razor.

So with that in mind, I'm assuming I can get that same color "perception" in my DIY, and still have most of the full spectrum covered (less the reds@660nm and greens@520-530nm, which I could add at the same time also).

The only other "outside of the box" led I have on my fixture are 4 medical grade UV's@395nm

It will be easy to expand on the fixture, as I built the "expand-ability" into it thinking I would probably need to play around anyways.

[asid61]

2 royal blue : 1 neutral white nets around 14k.
Best spectrum is 4:2:2:1 royal:neutral:violet:cool blue (470nm).
You do not want UV leds. The peak for violets is 428nm in corals, so 395nm is far less effective, regardless of grade. 420nm violets are common enough.
You don't need green, or red. The neutral white gives enough green, and you can add a warm white to get red spectrum (the neutral white already gives enough, but adding a discrete red would be too much).

[zachts]

2 royal blue : 1 neutral white nets around 14k.
Best spectrum is 4:2:2:1 royal:neutral:violet:cool blue (470nm).
You do not want UV leds. The peak for violets is 428nm in corals, so 395nm is far less effective, regardless of grade. 420nm violets are common enough.
You don't need green, or red. The neutral white gives enough green, and you can add a warm white to get red spectrum (the neutral white already gives enough, but adding a discrete red would be too much).

Best spectrum is of course a matter of personal opinion and aesthetics at this point in time. the only sure things that most all the discussions have agreed is that the base of a build should start with royal blue and neutral white and that some amount of violet and true blue are beneficial. beyond that it's all up to personal preference.

I spent last winter playing with hundreds of LED combinations on one of my fixtures and to me the "best spectrum" at presents starts with a 3:1:1:1:1 ratio of Violet:Royal Blue:Blue:5000k white: 2700k 90CRI white. that of course is with all leds runing the same current on one channel.

I am however never satisfied and am still fiddling around with different shades of violet and toying with how to implement more precise red and green wavelenghts, however further playing requires dimmability to fine tune things...... :D

[Secondsbest]

I'm attempting to get a "crisp" white look instead of a "yellowish" look from my whites. The whites I run now, make the tank look very "drab" to my eyes if I run them more than 45% power, even with a 2:1 ratio of Blue & Royal Blue to white.

My thoughts are I can accomplish gaining the crisp look by either replacing the current whites with the Cree's, or add them to my current whites, and run the bridgelux at a lower %.

My current Bridgelux produce a 6500k
The Cree produces an 8300k (if the bin numbers I posted earlier are correct).

Their spectral output (curves) appear to be the same.



Someone with more of a physics background can feel free to correct me, but I believe LEDs are designed to only emit certain wave lengths (color). Current adjustments will change the intensity of LED light (lumens), but wavelength will only shift a few nm. Using your posted graphs, and particularly for white LEDs, pay close attention to the emitted color line in the short wavelength/ high temperature area around 400-500nm. There's where the largest difference of color output will effect how your eye will perceive those LEDs. Higher blue curves will have less yellowing effects.
Keep in mind that color temp in kelvins, and wavelength are tied but inversely proportional. Blue flames are hotter than yellow ones.
Any light source's spectral curve can tell you how accurately it can reproduce natural light. Search the solar spectral curve and you will see the ideal.
Edit: Lighting stated as warm is "comfortable", not higher in color temp btw. Always confusing that one...

[asid61]

Best spectrum is of course a matter of personal opinion and aesthetics at this point in time. the only sure things that most all the discussions have agreed is that the base of a build should start with royal blue and neutral white and that some amount of violet and true blue are beneficial. beyond that it's all up to personal preference.

I spent last winter playing with hundreds of LED combinations on one of my fixtures and to me the "best spectrum" at presents starts with a 3:1:1:1:1 ratio of Violet:Royal Blue:Blue:5000k white: 2700k 90CRI white. that of course is with all leds runing the same current on one channel.

I am however never satisfied and am still fiddling around with different shades of violet and toying with how to implement more precise red and green wavelenghts, however further playing requires dimmability to fine tune things...... :D
Those violets send tingles up my spine! I'll make sure to try that for my next build.
Actually, I just found that using 660nm, 400nm, and warm white to use as a white replacement works really well to boost white without lessening the actinic effect, but has a slight green tinge.

[Epicreefer]

The led builds I've seen, my own included, that use only crisp white tend to look a little off, while those with warm white tend to look a little more natural. I wouldn't use all warm white but 2 cool to 1 warm to 4 royal blue seems to help even out the color. One key is adjustable ballasts, if they are adjustable it's much easier than just trying to swap out LEDs and what not to get the color you want. LEDs generally seem to be lacking red so adding red LEDs instead of warm whites could be another idea.

[oreo57]

Someone with more of a physics background can feel free to correct me, but I believe LEDs are designed to only emit certain wave lengths (color). Current adjustments will change the intensity of LED light (lumens), but wavelength will only shift a few nm.
Possibly yes and no.. some LED's use basically some phosphors to get the light "temp" they want.. In this case you could shift the color a bit more than you think if there is a mix of phosphors w/ varying efficiencies of emission based on power input. Of course this also means they also change w/ time, as do all phosphor based lights.

Apparently even in monochromatic LED's this is an issue (albeit it seems minor to me)
Full "on" to almost off has an approx 14nm shift
:

While this dimming scheme is simple and works well for lower-end displays, the drawback with analog dimming is that an LED’s color shifts with changes in forward current. Figure 1 shows a true green LED’s color variation with changes in forward current. This LED’s full brightness is specified at 20 mA. Analog dimming to 25% brightness shifts the color spectrum from 525 nm to 531 nm. This color shift becomes unacceptable in displays requiring a true color representation.

http://www.ti.com/lit/an/slyt238/slyt238.pdf

Includes a "case" for PWM vs analog dimming....

[Daimyo68]

OK, so after some reading here, and further looking into the Spectrum instead of just the K rating, it seems that my current whites have the biggest output in the Blue region (image above)

If I was to add a warm white that peaks in the red region, would/could I produce more of crisp white by filling in more of the color spectrum? It seems that I have plenty of light in the Blue spectrum, but not enough in the red/green spectrum.

Looking at the Cree XTE Warm Whites, they have the peaks in the Red Spectrum while not adding much to the Blue, and a little to the green spectrum.

I do have individual control of each color by way of Dimmable Meanwell Ballast's and the Apex.

I appreciate the responses here, and hope I'm understanding them correctly.