CREE CXA 3000°K/80CRI spectrum analysis.

[stardustsailor]

-RELATIVE POWER SPECTRAL GRAPH ,supplied from Cree datasheet ,was used to analyse further the
3000°K 80 min CRI CXA series led arrays.

-Graph was cleaned up ,enhanced and processed with Adobe PS CS4.
-Then digitized with Engauge (ver.4.1 ,for Win ).





Y values obtained per X nm ,were copied & then pasted to a spreadsheet ...





Relative spectral power means that for every nm ,there is a 'factor' with which
a " x " power value is multiplied ,to express Radiometric Power values per Nanometer .
(Absolute Power Spectral Graph . )

For example if a rel .factor of 1 is given to ~603 nm ,then it mean that this wl is the most powerful wl .
all the rest are expressed as % of this wl 's power ....

So if 1 * x = the absolute power of the 603 nm quanta ..
Then the sum (total ) of " RP factor * x / nm " will be the total radiometric output power of the array .

While 'rel power factor' is given by the rel .power graph ...
(and the digitized values obtained ..)
The " x " value depends from actual total output power . (not a constant value ) ...

( But it can be expressed as a 'constant ' in the form of " x W/nm per Radiant Watt " )

Complicated ?

Not so ...

For example ...

4x Cree CXA3070 @ 2000mA & Tc= 55°C ,will radiate 4x 25Watts =100 Watts of light Power .

Radiant Power value placed on the spreadsheet ...





Some info there for 100 Watts of light ,the 4x CXA3070s are emitting ....

Constants (for 3000K 80CRI CXAs ):
(for the 'average' 30F / 30 H 3000K bins )


-For every radiant Watt , the quanta flux is ~ 4.963 umols/sec
(multiply x Watts,with 4.963 of radiant power to convert into umols/sec )

.....


More info here ....









-The "tri -band " of radiant power distribution is as follows :
Blue 400-499 nm : 13,44%
Green 500-599 nm : 27,31 %
Red 600-700 nm : 59,25%
(More analytical in a "7 band" distribution ,at the pic )


-X RP factor ,per radiant Watt is ~ 0.0024794571 W/nm


-Peak excitation wl is 454 nm ( ~ 48.5 % )
-Peak emission wl is 601-604 nm . (100% )

ChB 642 factor : 0,706 ( ~ 70.6% )
( *0.0024794571 W/nm= 0.0017 Watts - 1.7 mW - of 642nm per CXA's radiant watt )

ChA 662 factor : 0,471 ( 47, 1 % )
( *0.0024794571 W/nm= 0.0011 Watts - 1.1 mW - of 662nm per CXA's radiant watt )

 

[SupraSPL]

Wow awesome data SDS thanks! Looks like the blue % is higher than we thought (13.44% vs 10%) and that is actually very good news!

 

[CaliWorthington]

Simply outstanding!

 

[stardustsailor]

Well ..
Some more careful -manual - digitizing was done ..
(over 3300 manual points placed ..LOL ..)


Now,the rel spectral graph ,in the spreadsheet ,looks far more precisely and accurately digitized .




So ,the calculations are far more precise (more accurate ..)
than in the previous example ....


So a single CXA3070 bin AB 30F /30H ,driven at 1400mA and with aTc = 45°C ......

-Rad .Efficiency = 40,11%
-El.Pwr =52.052 W
-Heat Pwr Q = 31,172 W
-Light Pwr Φ = 20,88
-Tj = 69,94°C
-Total PAR Flux =100,663 umol/sec
-Radiometric to PAR flux Conversion multiplier = 4,82 umols/sec / Radiant Watt



-Blue % = 14,05
-Green%= 44,57
-Red% =41,38

Distributed in wl ' bands ' as shown :




-Peak excitation wl : 454 nm ( 0,486 )
-Peak emission wl : 603 nm (1 )
-wl range over 50% rel.Power( 0,5 <=> 0,5 ) : from 524 nm to 659 nm

Cheers .

 

[epicfail]

3300 manual points!! your awesome man.

Today Salmonetin posted a video about some software that could do that for you possibly.

 

[Bueno Time]

Very good SDS!

 

[Positivity]

Spectrum looks nice..really broke the colors down on that.

Now where to get those ab..

Broke one of my cxas..lifting it from sink I think. Can see a crack in the base..still fires. Tough little buggers

 

[Rooster99]

Simply awesome!! Well done SDS.
Where abouts on the spectrum is that little bump in the red area? Looks to be around the 660nm but I cant see the numbers unfortunatly.

 

[rufus01]

Now we have just one small problem, find this shit in Europe

 

[stardustsailor]

Simply awesome!! Well done SDS.
Where abouts on the spectrum is that little bump in the red area? Looks to be around the 660nm but I cant see the numbers unfortunatly.

695 nm ...

Cheers !

 

[stardustsailor]

Actually ,once digitizing is done slowly and manually ,a very good copy and analysis can be done
from a Relative Power spectral graph ...

I'm still learning how to use that digitizing software (freeware ) ,which actually is pretty simple and easy ...

This is a new rel .Power spectreal graph ,digitized from a "traced " line onto the Cree's graph ,
and the y values obtained ( RP factors per wl ) ,they were plotted in a spreadsheet ....(1 nm distance at x axis ) .



It has some "noise " ,but the basic y values are more or less ,right on spot ....
Compared with the original Cree Rel.Power Graph ...(Layererd over )



Some 'maths' and useful equations ,for those who would like to make their own spreadsheet ...


The "area " under the Curve ,is the total Output light power ( in W ) Φ .
Say it is a given value of i.e. 10 Watts ....

The area is equal to the sum of every y value ,for all the range of x values ,we're interested in ...

Cree's graph is from 380 nmm to 780- nm ....
Output light power (radiometric ) obtained by the luminus efficiency / LER ,is for PAR range of course ..
(400-699 nm,no lumens below 400 nm and / or beyond 699 nm,as there's no visible light ,over those 'limits' ....

So 1) Φ= ( lm/ W / LER ) * Vf * If , is the Φ (radiometric light power )

From graph :
2) Sum y 400=>699 = ( y400 + y401+ y402+....+y699 )
and ....
3 ) Φ= Sum y 400=>699 * i

Where 4) i = Φ / Sum y 400=>699
..
To calculate umols/ sec ,per wl ....

PPFλ = λ (nm ) / 119,708 * ( i* yλ )

=>

PPFλ = { λ * yλ (<= the y value- factor obtained by digitizing the rel.pow. graph , for every given wl λ ..) / 119,708 } * i


PPFλ = { λ * yλ / 119,708 } * { Φ / Sum400=>699}

Total PPF = Sum PPF λ 400=>699
( ppf400 + ppf 401+ ppf402 + ...+ ppf 699 )

Maybe it looks kinda complicated ,but ain't really ,
once you get to understand what actually is a Rel. Power graph ...

 

[stardustsailor]

Wow awesome data SDS thanks! Looks like the blue % is higher than we thought (13.44% vs 10%) and that is actually very good news!

After three-four sessions of digitizing and recalculating ,yes...
The CXA 3000K ,80 CRI has 14% of it's output power (W ) at the 400-499 nm range ,
44,7 % at the 500-599 nm and the remaining ~ 41,3 % at 600-699 nm range ...

Analytical in 12 "bands" ,four for each wl range :



Blue Range 400-499 nm
-Violet 400-424 nm : 0.37% Φο (W ) / 0,27% PPF (umol/sec )
-Blue 425-449 nm : 3.36% Φο (W ) / 2.57% PPF (umol/sec )
-Light Blue 450-474 nm : 6.25 % Φο (W ) / 4.99% PPF (umol/sec )<= Dom.
-Cyan 475-499 nm : 4.07 % Φο (W ) / 3.44% PPF (umol/sec )

Green Range 500-599 nm
-Turquise 500-524 nm : 7.05 % Φο (W ) / 6.27% PPF (umol/sec )
-Green 525-549 nm : 9.58% Φο (W ) / 8.92% PPF (umol/sec )
-Lime 550-574 nm : 12.39% Φο (W ) / 12.08% PPF (umol/sec )
-Yellow 575-599 nm : 15.64% Φο (W ) / 15.92% PPF (umol/sec )<=Dom.

Red Range 600-699 nm
-Amber 600-624 nm : 16.24% Φο (W ) / 17.21% PPF (umol/sec )<=Dom.
-Red 625-649 nm : 12.75% Φο (W ) / 14.05% PPF (umol/sec )
-Deep Red 650-674 nm : 7.95% Φο (W ) / 9.1% PPF (umol/sec )
-Hyper Red 675-699 nm : 4.36% Φο (W ) / 5.18% PPF (umol/sec )

 

[MarCus M.D]

Awesome. Where you can buy these CREE, at the best price in USA?

 

[medicinehuman]

Most of it's over my head but what I understood was excellent.

 

[stardustsailor]

Most of it's over my head but what I understood was excellent.

Wait until you hear about the 'experimental ' unit ,(still under development ) .
A unit aimed to quantify a light's " Photosynthetic Potential " ...

(A unit that involves Light Spectrum,Power and Duration (time ) ,along with Chlorophyl
absorptance and action Spectrum (rel.Quantum efficiency ..)

You want to hear the units ?

Liters or grams (mass or volume ) of CO2 ,processed by plant(s) ,per electrical Watt !!!

( Actually is umolsCO2 / Watt ,but since 1 mol of gas ,under Standard conditions for temperature and pressure,
is 22.4 lt of CO2 (~44 grms of CO2 ) ,it can be expressed as mass or Volume of CO2 absorbed and utilised ,under
a given electrical power of a light .... )

{ PFF /nm * RQE/nm * Absorptance/nm (and not absorbance ) } / 10 = mols of CO2 used ( or mols of O2 evolved )

{ PFF /nm * RQE/nm * Absorptance/nm (and not absorbance ) } * Vf/ ( 10*If ) =
Photosynthetic Potential per nm .

*PPF/nm = { Radiometric efficiency * Vf*If / Sum y400-699 } * y/nm* λ / 119,708.

.....
Try to imagine the rest ...

http://photobiology.info/Gorton.html





Simply ,it's a unit , that shows how much of the electrical energy used ,from a light fixture ,
can be (potentionaly ) turned into plant matter ...

Rather 'cool' unit,I should say ...

 

[CaliWorthington]

I noticed in one of the papers SDS posted, it says "Taken from KJ. McCree", and you guys refer to them as "fried eggs". Now I wanna change my username to "Cree McMuffin".

 

[bicit]

Well any hope of getting anything productive done today is shot. This is the thread that I've been needing for a while now.

I'm off to see if I can replicate this. I'd like to see what the vero series is like since I'm partial to them over CREE.

ETA: Question, while this is effective for prototyping and 'bench testing' so to speak. Is it possible to take similar measurements with data from a par meter like Apogee sells? Or would that require a full fledged spectroradiometer?

Any hope you could do one for a vero 29 4000k 80cri at 1400ma while I try and fit all of this data into my brain?(Also so I could check my work)

 

[stardustsailor]

Well any hope of getting anything productive done today is shot. This is the thread that I've been needing for a while now.

I'm off to see if I can replicate this. I'd like to see what the vero series is like since I'm partial to them over CREE.

ETA: Question, while this is effective for prototyping and 'bench testing' so to speak. Is it possible to take similar measurements with data from a par meter like Apogee sells? Or would that require a full fledged spectroradiometer?

Any hope you could do one for a vero 29 4000k 80cri at 1400ma while I try and fit all of this data into my brain?(Also so I could check my work)

If I will manage to find some free time ,I will gladly do some calcs for the Vero you've asked .
..

(Do you want the spreadsheet I use ?
I can email to you .- It runs on OpenOffice Suit.-free to download )

Cheers.

 

[stardustsailor]

Now ....
Prepare yourselves ,for some ...
FGS ,pay some attention and I'm sure you'll understand everything ...
...



Ok ...

1 ) Light can be assumed that is actually a Flux of "quanta" ( "packets of energy" aka photons for light )
Thus quanta=photons from now on ...

2) One Watt of Radiant or incident light power ,is actually a flux of energy ...
Thus energy units per time units ..Thus Power / Quanta(Photon) Flux = Joule /sec

3) Since Photons/quanta are "packets of energy " ,then Power / Quanta(Photon) Flux = quanta/sec

4) Because they are many of them ,around ...The Avogadro's Number is used as a " Divider " / numbering unit ...

So N quanta = 1 mol = 6,023 * 10^23 (yes ..23 zeros ...Super huge number ... )
1 umole = 1mol/ 1 000 000 = 6,023 * 10 ^17 mols ( 17 zeros..Still a huge number ... )

So when we say a led radiates a Photon Flux of 1umol per sec ,we actually mean that
602.300.000.000.000.000 quantas are radiated per second ...

Is 1 umol/sec = 1 joule/sec ?
No ,of course not ...
(It takes some calculations to convert ... )

But ..Not all wavelenghts (colors of light ) carry the same amount / number of quanta per radiated
Joule/sec or aka Watt ...( 1 Watt = 1 Joule / sec ) ...
The equation is pretty simple ...

λ / 119.708 = umol/sec /Watt ...

Where λ =is the wavelength in nm units

As we can understand ,the number of quanta ,per given Radiated Watt ,is wavelength dependant ..

For example a blue ideal monochromatic led of 480nm ,for every 1000mW ( 1 W ) of radiated light ,
would output a quanta flux of 480 / ~120 * 1= 4 umols/sec
( 4 * 6,023 *10^17 !!!! ... )

A FR led of 960 nm ,for every radiated Watt ,will output a photon ( ) flux of 960 / ~120 *1 = 8 umols / sec ....

The equation of course gives a " linear " graph ....

So ,at PAR range ( 400-700 nm ) umoles per radiated watt ,have a lowere value of 3,34 umol/sec
at 400 nm violet and a highest value of 5,84 umols / secat deep red of 700 nm .....


But power is 1 watt in case of both blue and FR led ...
How come the photons are less in numbers in blue and more in red /FR ?
What is different ?

Difference is tha blue photons carry more energy ,than the red / FR ones ...
Fewer in numbers but loaded with energy ...
(Thus UV is so dangerous ...And the lower wavelength of E/M radiation ,the worse for living organisms .... )

Does photosynthesis use tha quanta energy ?

No ,of course not !
Photosynthesis works with "jumping electrons " of Magnesium as it
gets "excited" by quanta of light ...

By numbers ,and not by the energy the quanta carry ...(this becomes heat ,ultimately ... )
Thus the red light of many photons ,drive better the photosynthesis than the blue light of few but
"enegy loaded " photons .....

By numbers .
Keep that in mind .
It is important .
By numbers of quanta .
That's how photosynthesis is driven .
From Quantum/ Photon flux of great numbers ...

Is that enough ?

Well no ....
For many reasons ...
But let's think of the most basic one ...

Plants do not absorb all wavelengths with the same and the they do not
utilise all photons of various wavelengths as efficiently ,to drive photosynthesis ...


Absorptance is equal to 1 - ( Reflectance+transmission )

It is a value between 0 and 1 ( 0-100% ) ,that expresses the actual percentage of Quanta absorbed to be used further ..

So .You have the blue led of the previous example ......
480 nm blue .
Say it radiates 1000mW (1 Watt ) of light .(not it's electrical power ... )
About 4 umoles /sec of quanta .....

Digitising this :

The "leaf " marked curve ......(The "averaged" absorptance of an alive intact"averaged" leaf of "averaged" different species of higher plants )

Digitising will reveal an absorptance y value of 0,92 or 92% at the 480 nm ....

That means that from the 4umols/sec from the led ,the 92% is potentialy absorbable from the plant ...
It means that if the plant will had a canopy all over,and that all of the 4 umols of Quanta ,were incident to
leaf canopy ,still the plant ,for that given wl ,can only absorb 92% of the photons ...The rest are either reflected
(from pigments,waxes,etc ) or transmitted ,throught out the leaf to next lower ones ...

So ...For our led ...0.92 * 4 umols/sec = 3,68 umols /sec ...
The maximum potential absorptance-by plant(s) of our led's light ...
( in that particular wl of 480 nm and for 4 umol/sec ...)

From the 3,68 umols/sec the plant absorbs ,not all of them drive photosynthesis ...
Each wavelength's quanta have a different Relative Quantum Yield factor ..
Also known as "Action Spectra " .....

Digitising the curve for the chamber grown plants ...
(I 've started doing it ,and then stopped ,as I was watching the numbers of each manual digitising point ..)
it will be exactly the same, as doing the opposite action of 'graphing' those values here :





They are the same thing ....


In the spreadsheet RQE curve is made using the RQE table's values ,
while absorptance values where obtained through digitizing the intact's leaf absorptance graph .

Notice at ~485 or something where absorptance starts to decrease ,
the RQE of the same,of decreasing
absorptance ,wavelengths/ photons ,begins to increase ! ...
Life and evolution ! Amazing mechanisms !

So ...
Where were we ?
A h yes ...
The 3,68 umols/sec that the plant absorbs ,but not all of them drive photosynthesis .
RQE of the 480 nm quanta is 0.69 ...

So 69% of them will be used for driving photosynthesis ....

0.69 * 3.68 = 2.53 umols ,maximum potential number of quanta to be absorbed and used for photosynthesis
From the total of 4 umols /sec ....


Wait ...
I've not even started to go 'deeper' yet ...
It gets better ...

Cheers

 

[SupraSPL]

According to Mr Flux method, the Vero 4000K curve is 325 LER with 13% blue. At 1.4A (Tj 50C) it would be 37.1% efficient. Of course there a few problems with that measure. It is using the minimum numbers, which Mr Flux pointed out is more beneficial to the CXA than the Vero line, and Tj 50C might not be realistic in some applications at 1.4A.

It would be interesting to see how the SDS method compares. If a bunch of us can learn how to digitize these curves, we can share the burden