-
ICMag with help from Landrace Warden and The Vault is running a NEW contest in November! You can check it here. Prizes are seeds & forum premium access. Come join in!
Is it possible for a 63W LED lamp to produce 58W of actual light?
- Thread starter pinstripe
- Start date
SupraSPL
Member
asde2, you can get 50+ lm/watt at 640nm from the osram GD+ reds. They are approx 30% efficient vs the 660nm ledengin mentioned above @ 17%. Of course running the LEDs at higher current levels will further reduce these numbers. Running them without the heroic cooling renz mentioned will reduce the numbers even further. Without proper cooling lumen maintenance will be penalized with time as well.
One problem with the osrams is that they are sold by groups of bins. Maybe if you call digi-key they can specify the bin if you are ordering an entire reel?
One problem with the osrams is that they are sold by groups of bins. Maybe if you call digi-key they can specify the bin if you are ordering an entire reel?
I'm guessing that's how it works. When I buy low qty power LED stuff from digikey/mouser/newark, the Vf of the LEDs is always towards the edge of the min spec, so lower power. They're prob a lower light output bins, people calling up for qty prob get closer to what they want.
Also when I was browsing low output (20mA) SMT LEDs on digikey last night I was pretty surprised by how many datasheet links were either 404, just one or two page sales brochures or links to a manufacturer web page with no link to an actual datasheet.
Like wtf are you not trying to sell your products or what...
Like wtf are you not trying to sell your products or what...
SupraSPL
Member
Originally Posted by LEDGirl: "For anyone wanting the "nitty gritty" as far as calculations go, here it is: The 1st generation 63W has an actual power draw of 60-63W from the wall, with 5-6W being consumed by the cooling fans. In all actuality, it creates 54-58W of actual light..."
To put some terms in to what Pinstripe was saying, the 60-63 watts is the input power, 5-6 watts consumed by the fan leaves about 57 watts input power for the LED system. If we assume the AC-DC driver is 90% efficient, that means the LEDs are dissipating 51 watts power.
As Pinstripe mentioned, affordable, high quality, high bin, high power, properly cooled LEDs have radiometric efficiencies of 15-40%. If we assume we are using LEDs with 25% radiometric efficiency and we have 51 watts of dissipation power, that means we have approx 12.75 PAR watts of actual light* and the rest becomes heat**.
When we measure grams per watt from an HID, we do not use the input power (ballast losses are not included), we use the dissipation power of the bulb. So I agree with LEDGirl's statement that gram per watt numbers we use to compare lamps should be based on the LED's dissipation power.
However, once that has been established to compare the effectiveness of the lamps, I believe the entire system efficiency should also be considered as well. The cooling fans used for cool tubes can draw hundreds of watts. Of course there is also air conditioning electrical consumption to consider.
*Infrared LEDs and a small portion of white LEDs output falls in the infrared (>700nm) and would not technically be included in the PAR watts figure.
**We could take this a step further by considering the color AKA wavelength AKA spectral power distribution (SPD), calculate the actual photon count per second (PPF). If we had a fancy light measuring device we could measure the actual amount of photons that are reaching the canopy (PPFD). We could then attempt to adjust for the photosynthetic effectiveness of the SPD.
To put some terms in to what Pinstripe was saying, the 60-63 watts is the input power, 5-6 watts consumed by the fan leaves about 57 watts input power for the LED system. If we assume the AC-DC driver is 90% efficient, that means the LEDs are dissipating 51 watts power.
As Pinstripe mentioned, affordable, high quality, high bin, high power, properly cooled LEDs have radiometric efficiencies of 15-40%. If we assume we are using LEDs with 25% radiometric efficiency and we have 51 watts of dissipation power, that means we have approx 12.75 PAR watts of actual light* and the rest becomes heat**.
When we measure grams per watt from an HID, we do not use the input power (ballast losses are not included), we use the dissipation power of the bulb. So I agree with LEDGirl's statement that gram per watt numbers we use to compare lamps should be based on the LED's dissipation power.
However, once that has been established to compare the effectiveness of the lamps, I believe the entire system efficiency should also be considered as well. The cooling fans used for cool tubes can draw hundreds of watts. Of course there is also air conditioning electrical consumption to consider.
*Infrared LEDs and a small portion of white LEDs output falls in the infrared (>700nm) and would not technically be included in the PAR watts figure.
**We could take this a step further by considering the color AKA wavelength AKA spectral power distribution (SPD), calculate the actual photon count per second (PPF). If we had a fancy light measuring device we could measure the actual amount of photons that are reaching the canopy (PPFD). We could then attempt to adjust for the photosynthetic effectiveness of the SPD.
Why are we assuming? No one knows? If no one knows or no one has said they're switch mode current regulated, I'm assuming resistors at around 75%.
BTW does anyone happen to have any teardown pics of any of the popular LED panels? I've wanted to see them inside out, get a peek on the regulators, the PCBs, and the thermal management on them since they came out.
Last edited:
yo supra,
i thought 50lm/w at 640nm are ~40% effciency, maybe im wrong. but anyways i dont need want more than 35 watts because i thought about supplementing my cmh just to test difference in yield, i can build 3 cmh setups with same specs but and wanna try different led additions to see end results using same rate of additional output watts making it possible to specify wavelenght efficiency more - i think that 660nm isnt as good as 640nm no matter the qe but since i seem to be alone with this theory i wanna try on my own - of course i would love to build an only led panel but most wavelengths i would need arent available or just suck in efficiency resulting in very very very high prices to get the same output
i thought 50lm/w at 640nm are ~40% effciency, maybe im wrong. but anyways i dont need want more than 35 watts because i thought about supplementing my cmh just to test difference in yield, i can build 3 cmh setups with same specs but and wanna try different led additions to see end results using same rate of additional output watts making it possible to specify wavelenght efficiency more - i think that 660nm isnt as good as 640nm no matter the qe but since i seem to be alone with this theory i wanna try on my own - of course i would love to build an only led panel but most wavelengths i would need arent available or just suck in efficiency resulting in very very very high prices to get the same output
SupraSPL
Member
You could be right asde, I'm not certain. I saw some bins that had 70lm/w+ at 625 nm on the OSRAM site. If 50lm/w are 40% efficient that would make the top bins 60%!
I expect KNNA knows and his spreadsheet could probably clear that one up for us as well. Does anyone know if the sheet will function correctly on Google Spreadsheets? I can't seem to figure it out.
I expect KNNA knows and his spreadsheet could probably clear that one up for us as well. Does anyone know if the sheet will function correctly on Google Spreadsheets? I can't seem to figure it out.
Important to remember that LEDs have a spectral width of anywhere between 20nm and 35nm. So a 627nm LED will produce light in the range of 617nm to 637nm but the optical power will be less at the edges. So when you think about multiple LEDs of different wavelengths also understand that the spectra will overlap given a broader spectral distribution.
I realize that everyone it is easier for most to discuss light output in lumens but that is only relative to the human visual system not for a plant. Much better to keep the terms in watts or energy over time.
I realize that everyone it is easier for most to discuss light output in lumens but that is only relative to the human visual system not for a plant. Much better to keep the terms in watts or energy over time.
i dont have any datasheet showing the different spectral distributions
means i cant use them to calculate but because of the small range you just mentioned fitting the 'usual' specs for red led i think the difference from calculating with the peak wavelength only isnt as biggish as 10%
means i cant use them to calculate but because of the small range you just mentioned fitting the 'usual' specs for red led i think the difference from calculating with the peak wavelength only isnt as biggish as 10%
@asde. Link on Luxeon Rebel LEDs below (see pg 6 and 11). To calculate the power distribution requires a little integral calculus but one can eyeball to get a feel. They are additive which is key. Furthermore, LEDs have a Gaussian distribution so the power (W) is the same from 617nm to 627nm as it is from 627nm to 637nm
http://www.philipslumileds.com/pdfs/DS65.pdf
http://www.philipslumileds.com/pdfs/DS65.pdf
someone should make or find a lumens@freq=watts calculator. It comes up a lot.
and maybe an easy way to turn 0 - 1.0 scaled spectrum profiles into areas (total power) to determine what power level the 1.0 is, so you can directly compare profiles for different LED and make accurate composite spectrum profiles for planning out LED array color recipes.
A CAD program can probably provide the spectrum profile to area conversion. Getting combined LED plots probably isn't so difficult; worst case it's an excel project.
and maybe an easy way to turn 0 - 1.0 scaled spectrum profiles into areas (total power) to determine what power level the 1.0 is, so you can directly compare profiles for different LED and make accurate composite spectrum profiles for planning out LED array color recipes.
A CAD program can probably provide the spectrum profile to area conversion. Getting combined LED plots probably isn't so difficult; worst case it's an excel project.
You have to be careful when normalizing data for comparisons. It's very important all of your units are the same and relative, otherwise you will get erroneous comparisons especially when it come to intensities, it's actually not that simple.
Go to this site for wavelength and lumens/watt under Luminous Efficiency. It's in Excel format. Also, whenever you see the power for a given LED it is the total radiometric flux, that is the area under the spectral distribution curve.
SupraSPL
Member
Very interesting point ASDE, I had no idea there was that much difference bewteen 625nm and 640nm. Thanks also to Dr. Light for the link to CVRL.
683 lumens per watt is possible at 555nm yes? The CVRL site says 625nm is 36% as efficient as 555nm so that is 245 lumens per watt max. For 640nm it says 19.4% so that would be 132 lumens per watt max.
So if we have a 625nm LED claiming 50lm/w, that is 20% radiometric efficiency. A 640nm claiming 50lm/w would be 37.8% radiometric efficiency. We need to be in the 40% range to outgun the big HID bulbs.
This raises a question for me because the I was told the LED I have peak at 640nm, but if I want to buy more from Digikey, they are listed as 625nm.
SupraSPL
Member
Also I see your point about the need for a bit of calculus to get more accurate percentages Dr Light. The manufacturer's sure could help us out there. It would be nice if they would list the output of LEDs in mw as well as lumens. They are doing that with the 660nm and some of the blues at least.
Z
Zombo
All I will add is that all this physics talk gave me a real Techno-Chub™. I used to study LASER/Electro-Optics, and did LED DIY stuff for reef aquaria. Good stuff.
Oh, and I love my HGL units.
Oh, and I love my HGL units.
