5000k or 3000k lights?

kdt15

Active Member
https://en.wikipedia.org/wiki/Color_temperature

To the extent that a hot surface emits thermal radiation but is not an ideal black-body radiator, the color temperature of the light is not the actual temperature of the surface. An incandescent lamp's light is thermal radiation, and the bulb approximates an ideal black-body radiator, so its color temperature is essentially the temperature of the filament. Thus a relatively low temperature emits a dull red and a high temperature emits the almost white of the traditional incandescent light bulb. Metal workers are able to judge the temperature of hot metals by their color, from dark red to orange-white and then white (see red heat).
Many other light sources, such as fluorescent lamps, or LEDs (light emitting diodes) emit light primarily by processes other than thermal radiation. This means that the emitted radiation does not follow the form of a black-body spectrum. These sources are assigned what is known as a correlated color temperature (CCT). CCT is the color temperature of a black-body radiator which to human color perception most closely matches the light from the lamp. Because such an approximation is not required for incandescent light, the CCT for an incandescent light is simply its unadjusted temperature, derived from comparison to a black-body radiator.
ok so if i understood this right, the color temp usually corresponds to light emitted by thermal radiation of something very hot. LEDs dont emit light this way, so a "cct" was made for LEDs just to match the 'normal' light sources from heating?

also, does this mean that 5000k cct for LED is exactly hte same as something burning at 5000k emitting light? does this imply that LED's are made just to match it? and theoretically i could have a completely random spetctrum made from LEDs that dont follow any color temp? (ex: extremely uniform across all wavelengths, something i presume is unnatural)
 

kdt15

Active Member
https://en.wikipedia.org/wiki/Color_temperature

To the extent that a hot surface emits thermal radiation but is not an ideal black-body radiator, the color temperature of the light is not the actual temperature of the surface. An incandescent lamp's light is thermal radiation, and the bulb approximates an ideal black-body radiator, so its color temperature is essentially the temperature of the filament. Thus a relatively low temperature emits a dull red and a high temperature emits the almost white of the traditional incandescent light bulb. Metal workers are able to judge the temperature of hot metals by their color, from dark red to orange-white and then white (see red heat).
Many other light sources, such as fluorescent lamps, or LEDs (light emitting diodes) emit light primarily by processes other than thermal radiation. This means that the emitted radiation does not follow the form of a black-body spectrum. These sources are assigned what is known as a correlated color temperature (CCT). CCT is the color temperature of a black-body radiator which to human color perception most closely matches the light from the lamp. Because such an approximation is not required for incandescent light, the CCT for an incandescent light is simply its unadjusted temperature, derived from comparison to a black-body radiator.
furthermore, can we assume that 3000k and 5000k mixed together evenly (say 1 of each strip) basically averages out to 4000k? or does it not work that way? (definitely curious since 3000k and 5000k seem to be opposites in peaks, yet 3500k has 2 peaks at 450 and 600nm, like 4000k

was looking at CCT's and cant figure out how the distribution works - 3500k seems to have 2 near maximal peaks at 450 and 600nm, but does that mean its equally as strong/intense at those wavelengthd as other cct's that have same y-axis values? or is it something regarding proportions and area under the curve?
 

mahiluana

Well-Known Member
area under the curve
if you imagine the area under a Spectral graph of Power Distribution of 2 white led chip (5000 +3000K) with the same radiation output

the simple addition of each wavelenghts` intensity is your new graph and many different curves
can interprete the same cri of 4000K
 

kdt15

Active Member
if you imagine the area under a Spectral graph of Power Distribution of 2 white led chip (5000 +3000K) with the same radiation output

the simple addition of each wavelenghts` intensity is your new graph and many different curves
can interprete the same cri of 4000K
that kind of went over my head. since the y axis is maxed at unity, i presume we cannot add each graph and superimpose it, since itll reach a potential max of 2 instead of 1

as for 3000k and 5000k mixed together, does that indeed create a 4000k effectively, or does it not work that way?
 

nfhiggs

Well-Known Member
something regarding proportions and area under the curve?
You are VERY close.

Have you ever seen trend charts for stocks? The price graph will go up and down erratically, but there is only one single straight line that divides the graph line "equally" so to speak, so that if you colored in the areas between the straight line and the graph, the colored areas above and below the straight line will be equal, like this:
trend.jpg

It works sort of like that - the slope of the line determines the CCT number. Higher kelvin numbers (like 5000K) will have the left end of the line higher - meaning the spectrum is overall bluer, while a lower kelvin number (3000K for example) will have the right side higher, meaning a redder spectrum. 4000K is considered a balanced or neutral spectrum.

As Mahiluana points out - various curves can produce the same temp - but as far as white LEDs go, they are all pretty much made with the same or very similar processes and materials, and so no matter who makes it one 4000K spectrum is going to be very similar to another if they are also the same CRI.
 

mahiluana

Well-Known Member
or does it not work that way?
More generally, the term spectral power distribution can refer to the concentration, as a function of wavelength, of any radiometric or photometric quantity (e.g. radiant energy, radiant flux, radiant intensity, radiance, irradiance, radiant exitance, radiosity, luminance, luminous flux, luminous intensity, illuminance, luminous emittance)

https://en.wikipedia.org/wiki/Spectral_power_distribution

- the whole area under your graph can correspond to a value in watt, lumens, lux.....or just relative intensity,
where 100% is given by the peak wavelenght

adding 2 identical monochrome (450nm) royal blue chips...your resulting graph will have
the same relative intensity, but double in w, lm,....

below you see the resulting graph of 7 different chips (1x10w-UV400nm/ 2x3w-425nm/ 2x3w-470nm/ 4x3w-490nm/ 1x10w-2200K/ 2x30w-3500K/ 3x50w-1800K)

i only measured Vf + If of the chips and calculated the total lumens with the datasheets of the diiferent leds.

my intention was to close the gap (~500nm) of white phosphorous led light and mimic a more even daylight of the sun between 400-700nm.
as i added a good portion of blue(400-500nm) without adding too much green -
i hope to see good results for veg + bloom without switch - my new peak is 450/643nm

but as it is imposible for me to concentrate the 7 chips in one LES -
i will receive a slightly "moving" SPD in different areas under this lamp.

if you mix 5000K + 3000K cobs - i recommend to use 2 dimmable drivers and bring both cobs
as close together as possible to have a better uniformity of the mix.

S6002129.JPG
S6002090.JPG
 

kdt15

Active Member
You are VERY close.

Have you ever seen trend charts for stocks? The price graph will go up and down erratically, but there is only one single straight line that divides the graph line "equally" so to speak, so that if you colored in the areas between the straight line and the graph, the colored areas above and below the straight line will be equal, like this:
View attachment 4087176

It works sort of like that - the slope of the line determines the CCT number. Higher kelvin numbers (like 5000K) will have the left end of the line higher - meaning the spectrum is overall bluer, while a lower kelvin number (3000K for example) will have the right side higher, meaning a redder spectrum. 4000K is considered a balanced or neutral spectrum.

As Mahiluana points out - various curves can produce the same temp - but as far as white LEDs go, they are all pretty much made with the same or very similar processes and materials, and so no matter who makes it one 4000K spectrum is going to be very similar to another if they are also the same CRI.
thanks for the science lesson! now can you explain how CRI factors into this and discuss what is more important, and if its signficantly noticeable? i hear cri90 all the time but couldnt find light strips of cri90 for my first grow so yeah, went with the 80s samsung f-series strips :\
 

kdt15

Active Member
More generally, the term spectral power distribution can refer to the concentration, as a function of wavelength, of any radiometric or photometric quantity (e.g. radiant energy, radiant flux, radiant intensity, radiance, irradiance, radiant exitance, radiosity, luminance, luminous flux, luminous intensity, illuminance, luminous emittance)

https://en.wikipedia.org/wiki/Spectral_power_distribution

- the whole area under your graph can correspond to a value in watt, lumens, lux.....or just relative intensity,
where 100% is given by the peak wavelenght

adding 2 identical monochrome (450nm) royal blue chips...your resulting graph will have
the same relative intensity, but double in w, lm,....

below you see the resulting graph of 7 different chips (1x10w-UV400nm/ 2x3w-425nm/ 2x3w-470nm/ 4x3w-490nm/ 1x10w-2200K/ 2x30w-3500K/ 3x50w-1800K)

i only measured Vf + If of the chips and calculated the total lumens with the datasheets of the diiferent leds.

my intention was to close the gap (~500nm) of white phosphorous led light and mimic a more even daylight of the sun between 400-700nm.
as i added a good portion of blue(400-500nm) without adding too much green -
i hope to see good results for veg + bloom without switch - my new peak is 450/643nm

but as it is imposible for me to concentrate the 7 chips in one LES -
i will receive a slightly "moving" SPD in different areas under this lamp.

if you mix 5000K + 3000K cobs - i recommend to use 2 dimmable drivers and bring both cobs
as close together as possible to have a better uniformity of the mix.

View attachment 4087383
View attachment 4087384
W, lumen, lux, relative intensity, these are all somewhat similar right, so all of these y-axes would be fine as plotted and not misleading?

"
adding 2 identical monochrome (450nm) royal blue chips...your resulting graph will have
the same relative intensity, but double in w, lm,...."
so the graph doesnt change at all, but i would just get more blue light in general for the plants, even though the graph doesnt change?

very cool experiement you did there, did you just add up and superimpose the values on the graph to form your new one?
 

nfhiggs

Well-Known Member
thanks for the science lesson! now can you explain how CRI factors into this and discuss what is more important, and if its signficantly noticeable? i hear cri90 all the time but couldnt find light strips of cri90 for my first grow so yeah, went with the 80s samsung f-series strips :\
CRI is how accurate the light will render colors as compared to natural light. The higher the CRI, the more natural colors will look under it. The sun has a CRI of 100.
 

mahiluana

Well-Known Member
The higher the CRI, the more natural colors will look under it.
most of white low(70) CRI leds (~...6500K - 3500K ) lack red wavelenghts that plants need for a perfect blooming.

the higher the CRI of those chips - the higher is the relative proportion of these reds inside the SPD of 3500-6500K leds range.
A 1800K pink chinise chip has a ratio of ~ 1blue : 3,5red and peak @ 655nm you can use it for veg + bloom.

so look to your SPD graph and subdivide the wavelenghts into 3 sections:

// blue400-500nm,// green500-600nm// red600-700nm //

The size of the area under your graph can give you an idea of the red-blue ratio , and if your spectrum is perfect enough to grow ~ healthy plants.

In the Mc Cree par-curve you notice,
- that red light is a bit more efficient than blue light.
- that there is a peak @ ~ 630nm ( ...and may indicate, that a white phosphor led with a peak in this region is a good option to use it as a basic spectrum for veg & bloom.)

We know that blue/red ratio is a tool to influence the photomorphogenetic of a plant.
Cryptochrome, Phototropine -, Phytochrome --- are keywords to find the different functions of blue & red light.

1 : 2,5 blue/red is my personal, average ratio to favorize - but all inside 1:1 <---> 1:10 can grow plants as we know. Some people here on riu grow weed with only red or only uva wavelenghts -
and it works somehow.


CRI 100 for natural sunlight is also a bit relative - and photographs know, that there is a wide range of light temps (30000K ---> 3000K) depending the hour and place.
Can`t imagine, that in the golden light of a sunset you still will see the lush green of a plant
and measure CRI 100.

Mccree-par-blue-red-ratio-1-2,5.jpg
 
Last edited:
Top