Can we agree that there are two groups, one that believe green is useful and necessary and one that believes that it is not?
How about we agree there are more colors between 500 and 640nm than green. I like red, white, and blue. My thinking on the white, it is a catch all. If there is a magic potion in the 500-640nm then white has it covered. I also look at the economics, the number of photons for the buck.
I do not think green is the magic potion. Yellow is a contender.
This thread was on the right track. This is a significant topic.
I measured Quantum PAR for two CoBs. CLU028-1204C4-303M2K1, 3000K 80 CRI and CLU028-1204C4-273H7K4, 2700K 97CRI
The measurements were made with a StellarNet BlueWave Spectometer with the NIST LED profile calibration traceable to NIST standards.
The measurements were taken on May 20th 2017, and I do not remember the distance between the detector and the CoB.
3000K 80 CRI 484/477 µmol/m²/s
2700K 97CRI 430/414 µmol/m²/s
Where the slash is total PAR / PAR between 400 and 700. The far red is removed in the smaller number.
Plants do not care about the wavelength of the photon. Let's agree to disagree there. That is what my textbook says. There is a thing about blue having more energy per photon than green and red, but the plant blows if off as heat and the resulting bio-chemical action is the same for protons of any wavelength.
But no need to get stuck arguing over it. Let's just move on and you can make your choice of 70CRI vs 97 CRI in the privacy of your own home.
Each CoB was measured at the same distance from the detector and driven with the exact same Mean Well HDD-H driver with a fix current.
The CoBs were mounted on a 1.5" wide, 0.125" thick, 24" long bar of copper suspended on a Gorilla Grow Tent.
The graphs were created with a custom PHP app that reads the values from a data file created by StellarNet's SpectraWiz Spectroscopy Software and converted to an SVG image.
These images use quantum µMole measurement, not radiometric watts as in the datasheet's SPD.
3000K 80 CRI
________________________________________________________________________________
2700K 97CRI
________________________________________________________________________________
Here is the difference and at the heart of this topic. The photon difference between the 3000K and 2700K came from an increase in the blue green yellow, and decrease in red.
If you want to weigh that with McCree, that is simple math, so I will multiply the value in the top two images by the mean values from pages 208 and 210
and redo this same set of images and PAR values.
3000K 80 CRI minus 2700K 97CRI
I hope this at least puts an end to the CRI red debate. Red is the dominate color in CRI. Texas Instruments has an LED driver especially for CRI adjustments. It has an heavy duty efficient switching buck step down LED driver for white LEDs and a linear current regulator for a string of red.
________________________________________________________________________________
Superimposed.
________________________________________________________________________________
These are the values from pages 208 and 210 highlighted in yellow.