nfhiggs
Well-Known Member
You seem to be a broken record, stuck on Mono/RGB. Cobs do not use RGB emitters, and that's the fly in your ointment.But you did not do the math. You talk about wall watts in, yet nothing about radiant watts out, thermal watts out, or quantum efficiency.
So you think 200 lm/W is high efficiency for green? It's twice as good as what is currently being sold. What is done in the lab is one thing. What is economically feasible to mass produce is another. Currently the green LEDs being sold by Cree, Lumileds, and OSRAM are native green, not phosphor. While some green phosphors may be more efficient at producing photons, they are not yet economically viable for mono LEDs. White LEDs have to use green phosphors.
The only common mono LED color being sold is PC Amber. I have never seen a PC green LED.
Can we agree that a direct narrow band deep red and deep blue are more efficient than a phosphor pushed red and blue? Deep blue is the no brainier because they are used in white LEDs except covered with phosphor. The white paper I posted from Lumileds says direct red is more efficient than red phosphor. With red it's a matter of economics whether to use mix technology where a direct red produces superior CRI with phosphor blue and green vs. RGB phosphor.
So if we were to make a grow light using mono RGB that produce 5 µMoles each of red, green and blue. How efficient would that be? It will be more efficient than a phosphor pushed RGB, correct?
The efficiency 36.6% using the most efficient Red Green and Blue LEDs on the market calculated at their Maximum rated output run at their most efficient current at an unrealistic 25° C.
And that is why I do not believe a CoB is 60% efficient.
Let's do the "simple maths"
Currently the most efficient on the market
Cree XP-G3 Royal Blue .730 radiant Watts @ 350mA 2.82V,
Cree XPE green 113 lm/W @ 350mA 3.2V,
OSRAM Olsen SSL Hyper Red .530 radiant Watts @ 350mA 2.15V
To produce 5 µmol each of RGB we will need 2 blue 1.8W, 5 green 5.7W, and 2 red 1.7W.
Datasheets attached so you can check the "simple maths".
The column lm/W is the number of lumens required for 1 Watt radiant flux.
The column µmol/W is the number of µMoles in one watt radiant flux at that wavelength.
flux is in watts
View attachment 3966939
There is a reason manufacturers do not use RGB emitters to make white light - as you just demonstrated, its just not efficient. Instead, they use highly efficient Royal Blues and phosphor coatings to convert a portion of those efficient blue photons to red, orange, yellow and green photons. So in fact, two thirds of your table simply does not apply when talking about COB efficiency, because there are no green or red emitters in a cob. You're simply trying to compare apples to oranges, and declaring 60% COB efficiency to be impossible based on the efficiency of RGB monos.
If you want to determine the potential efficiency of a white light COB, you need to look at the efficiency of its emitters, and subtract the Stokes shift losses for the converted photons. OK, now lets look at your table again - by your numbers in the last two columns, the Stokes loss can be deduced from the Radiant flux difference between the Blue/Green and the Blue/Red - And since ALL photons created by the royal blue emitters cost the same amount of energy to produce, the wall watts column for blue, green and red photons should be the same - 1.82W. The total wall watts for a white light COB with equal amounts of red, green, and blue photons (5 uMoles of each) is only 5.42W, not 9.24W. Now divide the total radiant watts by the total wall watts (3.38/5.42) and we get - 62.3%. All you have proved is that COBs are inherently more efficient than R+G+B white light.
Suddenly that 60% efficient COB looks quite reasonable, using your own numbers applied correctly and logically. It is simply not logical to try to make deductions about COB efficiency, by looking at RBG mono efficiencies.