Lighting Bilbo's Brief Guide to Choosing Your LED.
- Thread starter Bilbo
- Start date
Can I just add 2 cents worth to this?
LED is split into 2 sections...AC your input side and DC your output side...
LED no matter what anyone says is typically inefficient in several categories, and the paramount driver is power in or AC, IF your manufacturers run their LEDs at less than 20v DC then you will almost certainly run into trouble within a year...your AC supply at 240v should be turning that into at least 1.5 times what it draws from the wall...high current, high DC voltage (but less than 50v) is good, but only a strong driver can handle this...gives you high output LED. High output, high efficiency LEDs are above 40% efficient, 80% of LED STRUGGLES TO GET THAT HIGH...
Next is wavelength...many wavelengths are horribly inefficient, 630nm for example is half as efficient as 660nm...so pay attention to what spectrum is in your light, your chosen manufacturer SHOULD be able to give you a spectral response diagram. Make sure you get a good response through your required curve, more red for flowering, more blue for growth, UV and far red into IR WILL help.
PAR is what drives your plants, if your chosen manufacturer cannot provide you with AT LEAST 200umols at 400mm (16") from your canopy, AVERAGED not spot, then reconsider! 215umols is considered optimal, more is good, but less is not so good...check out "inverse square law" for how light travels...an "at source" spot measurement is a con...
Mike
,
LED is split into 2 sections...AC your input side and DC your output side...
LED no matter what anyone says is typically inefficient in several categories, and the paramount driver is power in or AC, IF your manufacturers run their LEDs at less than 20v DC then you will almost certainly run into trouble within a year...your AC supply at 240v should be turning that into at least 1.5 times what it draws from the wall...high current, high DC voltage (but less than 50v) is good, but only a strong driver can handle this...gives you high output LED. High output, high efficiency LEDs are above 40% efficient, 80% of LED STRUGGLES TO GET THAT HIGH...
Next is wavelength...many wavelengths are horribly inefficient, 630nm for example is half as efficient as 660nm...so pay attention to what spectrum is in your light, your chosen manufacturer SHOULD be able to give you a spectral response diagram. Make sure you get a good response through your required curve, more red for flowering, more blue for growth, UV and far red into IR WILL help.
PAR is what drives your plants, if your chosen manufacturer cannot provide you with AT LEAST 200umols at 400mm (16") from your canopy, AVERAGED not spot, then reconsider! 215umols is considered optimal, more is good, but less is not so good...check out "inverse square law" for how light travels...an "at source" spot measurement is a con...
Mike
,
Could you explain that further, why is 630 only half as efficient spectrum wise?
Hi Xagor
the outputs on different frequencies are made up of different chemical dopes, no matter how it is made...most "modern" LED starts out blue for example...then is doped to make the colours...for some reason the compounds used to manufacture 630 and several other frequencies block so much light output that they put out very little light...or the chemical compound itself used degrades very quickly so a lower power or a higher grade doping is used...
I'm not going to state that I know exactly how this all works as I have a very clever man who does this for us...BUT I will illustrate...the following graphic is a 9 colour chip we had made, with the same power in 660nm and 630nm and this clearly shows the actual spectral response against the required response...the 550nm they couldn't "do" at the time...so it was actually 7 colours in the chip...shortly after this we changed to the phosphor spectrum...which is much smoother...
argh...trying to figure out how to get a picture up...give me a minute...again, trying to get a picture up!
the outputs on different frequencies are made up of different chemical dopes, no matter how it is made...most "modern" LED starts out blue for example...then is doped to make the colours...for some reason the compounds used to manufacture 630 and several other frequencies block so much light output that they put out very little light...or the chemical compound itself used degrades very quickly so a lower power or a higher grade doping is used...
I'm not going to state that I know exactly how this all works as I have a very clever man who does this for us...BUT I will illustrate...the following graphic is a 9 colour chip we had made, with the same power in 660nm and 630nm and this clearly shows the actual spectral response against the required response...the 550nm they couldn't "do" at the time...so it was actually 7 colours in the chip...shortly after this we changed to the phosphor spectrum...which is much smoother...
argh...trying to figure out how to get a picture up...give me a minute...again, trying to get a picture up!
Hi,
ah thx.
The led manufacturers are making it harder then it should be.
Looking at e.g. Osram Oslon SSL or Rebel series reds and deep reds:
630 they give its value in lumens and for 660 in milliwatts.
Both in mw and it would be easy to compare but no...
So its the same with them as with what I heard about green leds: electrically unefficient -while with 630 and 660 the difference isnt as drastic as with green.
But besides that, why not prefer white leds overall especially cobs?
With Cree CXB series e.g. 3070 or 3590 you can easily achieve e.g. 50% efficiency at power to light conversion.
Getting single diodes to run at this value you need a lot of them at low currents/wattage which makes it much more expensive compared to cobs.
ah thx.
The led manufacturers are making it harder then it should be.
Looking at e.g. Osram Oslon SSL or Rebel series reds and deep reds:
630 they give its value in lumens and for 660 in milliwatts.
Both in mw and it would be easy to compare but no...
So its the same with them as with what I heard about green leds: electrically unefficient -while with 630 and 660 the difference isnt as drastic as with green.
But besides that, why not prefer white leds overall especially cobs?
With Cree CXB series e.g. 3070 or 3590 you can easily achieve e.g. 50% efficiency at power to light conversion.
Getting single diodes to run at this value you need a lot of them at low currents/wattage which makes it much more expensive compared to cobs.
Hi Xagor
the manufacturers want to make it hard, and electrically the input is identical mostly, the lumen output doesn't matter to a plant. Electrical inefficiency is the curse of LED, output inefficiency is the curse of good marketing, and people who investigate these things
We have seen many times the high lumen output hides very poor PAR output or vice versa (look at far red, virtually zero lumen output). This is often the case with white LED as you mention. The original idea of using red/blue LED was that you are only putting power into useful wavelengths thus saving power and making the "energy in" efficient.
As you can see this simply is not the case when it comes to COB, indeed single diodes are often extremely inefficient which is why we don't use them. No matter what anyone says, in terms of pure efficiency, a COB design (or single chip multi diode) is always going to be more efficient. Especially over a spread of wavelengths.
HOWEVER, the trade off with using a white LED is that the dope used to create overall white lacks power required for effective PAR outputs (remember lumens for humans) so an overall white falls short of many very definite advantages of targeting wavelengths. We have seen and been involved in trials using white LED, indeed we have found a natural white (not a warm or cold white) to be quite effective, but in order for a white to be effective, it either needs to be used as supplemental lighting such as in a glass house OR has to have additional peaks in the output. Which will "taint" the outputs...making the white, no longer quite so white...you have to remember here, that 98% of white LED starts out as BLUE and is then doped...
Over the next couple of years we intend to put some research hours into this and see if we can't alter our bespoke phosphor dope that we have and make it more visibly white without affecting the PAR output, in the meantime, the best we can get to is a very light pink colour which gives the curve you see above.
Mike
the manufacturers want to make it hard, and electrically the input is identical mostly, the lumen output doesn't matter to a plant. Electrical inefficiency is the curse of LED, output inefficiency is the curse of good marketing, and people who investigate these things
We have seen many times the high lumen output hides very poor PAR output or vice versa (look at far red, virtually zero lumen output). This is often the case with white LED as you mention. The original idea of using red/blue LED was that you are only putting power into useful wavelengths thus saving power and making the "energy in" efficient.
As you can see this simply is not the case when it comes to COB, indeed single diodes are often extremely inefficient which is why we don't use them. No matter what anyone says, in terms of pure efficiency, a COB design (or single chip multi diode) is always going to be more efficient. Especially over a spread of wavelengths.
HOWEVER, the trade off with using a white LED is that the dope used to create overall white lacks power required for effective PAR outputs (remember lumens for humans) so an overall white falls short of many very definite advantages of targeting wavelengths. We have seen and been involved in trials using white LED, indeed we have found a natural white (not a warm or cold white) to be quite effective, but in order for a white to be effective, it either needs to be used as supplemental lighting such as in a glass house OR has to have additional peaks in the output. Which will "taint" the outputs...making the white, no longer quite so white...you have to remember here, that 98% of white LED starts out as BLUE and is then doped...
Over the next couple of years we intend to put some research hours into this and see if we can't alter our bespoke phosphor dope that we have and make it more visibly white without affecting the PAR output, in the meantime, the best we can get to is a very light pink colour which gives the curve you see above.
Mike
Hi Mike,
But when I said 48% efficiency of a white cob running at about 50W I meant PAR watt/ per watt for a warm-white suitable for flowering. Means at a consumption of 50W it puts out about 25W of PAR.
You can even get them up to crazy insane 59% for e.g. cxb3590 at 22W for 3000K warmwhite (even though its unimportant: 190lm/W).
Yes there are differences regarding quantum efficiency but not higher then 9% for blue vs green. At about +550nm green/ish becomes even better.
Looking at science you can even find an oxford study/paper stating that at high lightlevels green becomes even more efficient than red. -Due to the structure of leave tissue, locations of chloroplasts and the fast saturation with red (bc of the high absorbtion).
So with green beeing fairly important, crappy efficiencies of green leds: white cobs seem to be extrememly suitable -extremly efficient, good balanced spectrum, appealing to the eyes, cheaper -as white is more mass market.
Yeah I the highest value of lumes you get at 555 at the peak of human eye sensitivity.
But when I said 48% efficiency of a white cob running at about 50W I meant PAR watt/ per watt for a warm-white suitable for flowering. Means at a consumption of 50W it puts out about 25W of PAR.
You can even get them up to crazy insane 59% for e.g. cxb3590 at 22W for 3000K warmwhite (even though its unimportant: 190lm/W).
Isnt that more of a marketing myth caused by the response curve of algea/extracted chlorophyll, but not plants?
Yes there are differences regarding quantum efficiency but not higher then 9% for blue vs green. At about +550nm green/ish becomes even better.
Looking at science you can even find an oxford study/paper stating that at high lightlevels green becomes even more efficient than red. -Due to the structure of leave tissue, locations of chloroplasts and the fast saturation with red (bc of the high absorbtion).
So with green beeing fairly important, crappy efficiencies of green leds: white cobs seem to be extrememly suitable -extremly efficient, good balanced spectrum, appealing to the eyes, cheaper -as white is more mass market.
Last edited:
Hi Xagor
A white COB runs no more efficiently overall than any other colour COB, a white is just doped blue, some of the Cree chips are very efficient, some are not, but there are far more electrically efficient chips out there, Bridgelux for example make some massively efficient chips getting in excess of 70%. We know of one company getting nearer 75% efficiency...
The original work into plant based LED was done with regard to putting energy just into the wavelengths needed, there are quite literally hundreds of papers out there with conflicting information, many plant scientists say one thing, many others say things that are entirely different. This IS a huge problem. We know of one very eminent scientist that ran two identical trials, side by side, same room, same feeding regimes, same everything and got 2 different results..., and running the same experiments again got 2 more different results...purely in terms of food production (which is what we are primarily involved in) this causes huge issues...and if a scientist is getting differing results from 2 benches side by side, a man who is quite literally one of the forefathers of bringing LED into the plant biology world, what is the average man going to get?
So taking all this into account, it suggests that there is still a long way to go until LED and plant relationship is known, it's why we continually take our research and make small changes to our lights...tiny changes bringing reliability and repeat ability of crop (whatever crop that happens to be) production. We have a bespoke and unique technology, which is bringing some of these results home, but we can't see it in a public setting for at least 3 years.
But the fact is that the average LED that is available now must give 2 things, power and spectrum and those spectra must be agreed upon before LED technology becomes the norm rather than the exception. Then and only then will prices drop to make them attractive to every man and not just those that can justify current prices...
A white COB runs no more efficiently overall than any other colour COB, a white is just doped blue, some of the Cree chips are very efficient, some are not, but there are far more electrically efficient chips out there, Bridgelux for example make some massively efficient chips getting in excess of 70%. We know of one company getting nearer 75% efficiency...
The original work into plant based LED was done with regard to putting energy just into the wavelengths needed, there are quite literally hundreds of papers out there with conflicting information, many plant scientists say one thing, many others say things that are entirely different. This IS a huge problem. We know of one very eminent scientist that ran two identical trials, side by side, same room, same feeding regimes, same everything and got 2 different results..., and running the same experiments again got 2 more different results...purely in terms of food production (which is what we are primarily involved in) this causes huge issues...and if a scientist is getting differing results from 2 benches side by side, a man who is quite literally one of the forefathers of bringing LED into the plant biology world, what is the average man going to get?
So taking all this into account, it suggests that there is still a long way to go until LED and plant relationship is known, it's why we continually take our research and make small changes to our lights...tiny changes bringing reliability and repeat ability of crop (whatever crop that happens to be) production. We have a bespoke and unique technology, which is bringing some of these results home, but we can't see it in a public setting for at least 3 years.
But the fact is that the average LED that is available now must give 2 things, power and spectrum and those spectra must be agreed upon before LED technology becomes the norm rather than the exception. Then and only then will prices drop to make them attractive to every man and not just those that can justify current prices...
Similar threads
