Opinions and advices for my new cheap idea. LUXEON COB GEN 4 (2.5 X 2.5)

[LivingLight]

Hello folks

I extrapoleted datasheet spectral distribution chart and elaborated to convert it in numerical data.
With numerical data, it is possible to integrate power over wavelenght in order to calculate PPF.

Below PPF for some COBs @ 1000mA, and assuming a LED efficiency of 30% (the capacity to transform electrical power in light):

L2C2_70_5700K
ppf: 48.03453837990068 umol/s

L2C2_80_2700K
ppf: 49.01280628712339 umol/s

CXB3070_80_3000K
ppf: 47.878880284205316 umol/s

CXB3070_90_3000K
ppf: 45.778220550503654 umol/s

Please consider that I've NOT measured the real ppf, and values reported are only results from theoretical calculations. Could anyone validate data for CXB3070? Is calculation close to the real output?
I'm not sure that 1208 COBs can provide enough power and the right light spectrum for growing plants.

Regards
Alex

Hey,

By integrating spectrums, you can calculate the real efficiency of your led depending on the luminous flux (lm/w)
Using a CIE luminosity curve. First you got to normalise your led spectrum's curve to 1w, so if u got 1 point every nanometer you just divide the whole data to get the total amount =1
Then you calculate : 683*CIE curve*Led spectrum and it gives you the amount of lumens the spectrum produce for 1 radiant watt.
For exemple the CXB 3070 3000k 80CRI produce 330lm for 1 radiant watt.
so if your led, according to the datasheets, produce 165lm/w that means for each electrical watt the led produce 0.5 (165/330) luminous radiant watt.
basically an efficiency of 50%.

Concerning the photon flux, a 3000k led usually produce approximately 4.8umol/j. So for each radiant watt you are supposed to get 4.8umol/sec.

[axef]

Hi LivingLight,

Lumens is a measure of the light emitted that best fit human eye response, so lm is useful for led comparison but is almost impossible to get umol from lumens.

Concerning the photon flux, a 3000k led usually produce approximately 4.8umol/j. So for each radiant watt you are supposed to get 4.8umol/sec.

That's the point, how could you know how much photons a multi wavelengths light source is emitting?
A 3000K light source means that the majority of the wavelegths emitted are shifted around the red, but it doesen't exclude the existance of others wavelengths light components at different colors.

I followed this idea to calculate umol/s:

1. the energy of a photon depends on wavelength:

1.png (474 Bytes) Viewed 785 times

where h is Plank constant, c is the speed of light, and lambda is the wavelength of the photon emitted.

2. to calculate the total number of emitted photones (u) by a light source at a given wavelength (lambda), you have to divide the total light power emitted (w*spd(lambda)) by the power of a single photon at that wavelength (E):

2.png (696 Bytes) Viewed 785 times

where spd is the normalized spectral power distribution, w is the light power and E is the power of 1 photon.

3. To calculate the total number of photones emitted in PAR spectral range, an integration of (2.) is needed over wavelegths between 380 to 800nm:

3.png (1.14 KiB) Viewed 785 times

4. to obtain umol/s, the total number of photones (u) have to be divided by the number of particles in 1 umol:

4.png (1.5 KiB) Viewed 785 times

Once got data from datasheet chart by an automated process, I wrote a program that perform above operations in order to calculate umol/s.

What do you guys think about? In your opinion, is this process sound?
Thnaks

Alex

[LivingLight]

Hi LivingLight,

Lumens is a measure of the light emitted that best fit human eye response, so lm is useful for led comparison but is almost impossible to get umol from lumens.

In fact it's really easy, eyes' response is well known and plotted on a curve . Maybe you didn't understand what i've said because i'm not explaining well. The CIE is giving curves that plot the lm vs the wavelenth for 1 radiant luminous watt. (CIE luminosity curve)
Actually, in order to measure the lm, u got to measure the photons or radiant watt and calculate the lm from this specific curve (that's the only way). To get the photons from the lumens you got to do the reverse process.

That's the point, how could you know how much photons a multi wavelengths light source is emitting?
A 3000K light source means that the majority of the wavelegths emitted are shifted around the red, but it doesen't exclude the existance of others wavelengths light components at different colors.

I got the data of the spectrums, actually i got around 100 spectrums' data (digital integrated data, not just a picture). 3000k 80CRI led spectrums are all almost the same and i've been calculating the photon flux for a lot of them.
Using basical equation E=nhc/lambda just like you


I've made a little program on a tab to calculate photon flux, efficiency, LER, phytochromes photostationary state. it's a kind of simulator with samsung strip leds, bridgelux vero, cree xp-e and cree cxb

[tazztone]

I've made a little program on a tab to calculate photon flux, efficiency, LER, phytochromes photostationary state. it's a kind of simulator with samsung strip leds, bridgelux vero, cree xp-e and cree cxb

care to share?

[LivingLight]

Sure but it's in french xD i could change for english but for now im a little busy
also it would need a tutorial but anyway
all light grey is what u gotta enter or choose (when you click on most of the cases it gives you several choices)
simulator is on page 5
i've made this on libreoffice i'm not sure it works on excell
https://www.catupload.com/download/015b ... 2cf70.html

[tazztone]

thanks so much.
i wonder how you got the data points for each nm wavelength of the spectrum

[LivingLight]

thanks so much.
i wonder how you got the data points for each nm wavelength of the spectrum

I'm using plot-digitizer (manual selection) that gives me a lot of data points (~10/nm) then i'm doing a little approximation on the tab using the average value of all the points selected in the range for each nm. It's more accurate than i expected at the begining.
Btw if you are interested in electrical efficiency calculation you got to change the selection of the C37 box in the simulator and choose "calcul à partir du flux"
efficiency values are given on M5 and N5