billyjojimbob11
Member
Knna, what whites are you speaking of? Cree XR-E top bin? What color?
knna
Member
I wasnt refering to any specific model, but in general top bin whites of good manufacturers. Cree, of course, as the efficiency leader, but Nichia, Osram, Lumileds, Seoul SC...
All of them have currently blues surpassing 35% of energy efficiency. Cree is currently over 40% on the top bin.
Although that efficiencies drops with conversion to white (or any other range, as currently there are phosphors for any visible waveband), due to the Stokes effect (loss of energy converting an high energy short wave photon (UV, blue) to a lower energy longer wavelength photon) and because conversion is not perfect. But for some ranges where LEDs are still inefficients, as green and yellow (at lower degree), using a blue chip with a phosphor is more efficient than using directly a LED of such inefficient color.
Its clear for green LEDs, profitable with yellow ones, and doable with amber (with the added benefit of more stable spectrum emission with temperature changes). With reds, currently depends of the wavelenghts. As longer, more profitable to keep using reds, but as blue chips are expected to reach efficiencies about 60% on just some years, while red's improvement is sttoped, surely soon its going to be profitable to use color converted blue chips for all.
Quantum dots (QD) are very promising, as they have quantum efficiencies (QE) of 100%, no losses on conversion. While current phosphors used are trying to reach 80% of QE. But QD have some problems to be used on LED packages, specially the inability of using them on densities high enough.
But ive seen reports on nanophosphors with 98-99% QE. And this type is useable almost inmediately. Using a converter with QE of 1 (100%) or very close to it, no matter if its a QD or a nanophosphor, is going to allow to get as much photons as the blue chip is achieving. Im almost sure this target is going to be reached soon, 1-2 years, so we are going to get a minimun of 1.5 uE/W (at 350mA) of any color we choose to use. Thats aprox the efficiency of 600w HPSs, so using 1w LEDs we are going to have at least same quantum efficiency than them, but without having aditional losses by reflectors, which averages for 25% of the light output.
ZnO still have a long road to be able to reach those efficiencies, AFAIK. Maybe ive missed some breakthrough of this tech.
Of course our first goal must be to know what colors we need and on what ratios: determining the best SPD for growing MJ. Then we will look for the most efficient way of giving it.
But at the same time, we must look for what is technically doable. Finding an ideal SPD that is not possible to achieve using current avalaible LEDs, or that is doable but at the cost of too much reduced efficiency is only useful for the case that in the future it could be done. But unuseful today.
But i think you are missing what i wanted to mean when I insist on the importance of knowing LED's efficiency. Its not about of saving more or less watts. Although its esential for a comercial LED lamp, i dont mind at all whats the power of an experimental setup.
No, if i insist on the need of knowing LEDs efficiency is because its the only way to determine whats the SPD we are using.
For example, say you are using 3 types of LEDs: blue green and red. You use 5 blues, 5 green and 10 reds, all at the same power (1watt)
Case 1. Efficiency of them: B=40%, G=5%, R=20%.
Case 2. Efficiency: B=20%, G=10%, R=30%.
In case 1, the SPD used is 47% blue, 6% green and 47% red.
In case 2, the SPD used is 22% blue, 11% green and 67% red.
As you can see, same LEDs configuration gives two very different light spectrums depending of the efficiency of each LED.
The question is: how do you know what color distribution are you using if you dont know the efficiency of the LEDs used?
It would depend of how mch of each range we can/should use. Blues are the most efficient LED today, so if we can use them, good. But as whites are already giving blue, using blues together with whites will depend if the best mix have more blue respect the other wavelenghts emited by the white than its actually emitting.
Most people at the testing group is going to use a little blues together with whites, apart of the reds. Personally, i think its good, but results will say if im right or wrong on it.
But what i know is that excess blue harms more than helps, aswell as any other bandwith. So we must find the point up each color becomes damaging (or not helping), and use the most efficients ones just up that point.
All of them have currently blues surpassing 35% of energy efficiency. Cree is currently over 40% on the top bin.
Although that efficiencies drops with conversion to white (or any other range, as currently there are phosphors for any visible waveband), due to the Stokes effect (loss of energy converting an high energy short wave photon (UV, blue) to a lower energy longer wavelength photon) and because conversion is not perfect. But for some ranges where LEDs are still inefficients, as green and yellow (at lower degree), using a blue chip with a phosphor is more efficient than using directly a LED of such inefficient color.
Its clear for green LEDs, profitable with yellow ones, and doable with amber (with the added benefit of more stable spectrum emission with temperature changes). With reds, currently depends of the wavelenghts. As longer, more profitable to keep using reds, but as blue chips are expected to reach efficiencies about 60% on just some years, while red's improvement is sttoped, surely soon its going to be profitable to use color converted blue chips for all.
Quantum dots (QD) are very promising, as they have quantum efficiencies (QE) of 100%, no losses on conversion. While current phosphors used are trying to reach 80% of QE. But QD have some problems to be used on LED packages, specially the inability of using them on densities high enough.
But ive seen reports on nanophosphors with 98-99% QE. And this type is useable almost inmediately. Using a converter with QE of 1 (100%) or very close to it, no matter if its a QD or a nanophosphor, is going to allow to get as much photons as the blue chip is achieving. Im almost sure this target is going to be reached soon, 1-2 years, so we are going to get a minimun of 1.5 uE/W (at 350mA) of any color we choose to use. Thats aprox the efficiency of 600w HPSs, so using 1w LEDs we are going to have at least same quantum efficiency than them, but without having aditional losses by reflectors, which averages for 25% of the light output.
ZnO still have a long road to be able to reach those efficiencies, AFAIK. Maybe ive missed some breakthrough of this tech.
Of course our first goal must be to know what colors we need and on what ratios: determining the best SPD for growing MJ. Then we will look for the most efficient way of giving it.
But at the same time, we must look for what is technically doable. Finding an ideal SPD that is not possible to achieve using current avalaible LEDs, or that is doable but at the cost of too much reduced efficiency is only useful for the case that in the future it could be done. But unuseful today.
But i think you are missing what i wanted to mean when I insist on the importance of knowing LED's efficiency. Its not about of saving more or less watts. Although its esential for a comercial LED lamp, i dont mind at all whats the power of an experimental setup.
No, if i insist on the need of knowing LEDs efficiency is because its the only way to determine whats the SPD we are using.
For example, say you are using 3 types of LEDs: blue green and red. You use 5 blues, 5 green and 10 reds, all at the same power (1watt)
Case 1. Efficiency of them: B=40%, G=5%, R=20%.
Case 2. Efficiency: B=20%, G=10%, R=30%.
In case 1, the SPD used is 47% blue, 6% green and 47% red.
In case 2, the SPD used is 22% blue, 11% green and 67% red.
As you can see, same LEDs configuration gives two very different light spectrums depending of the efficiency of each LED.
The question is: how do you know what color distribution are you using if you dont know the efficiency of the LEDs used?
It would depend of how mch of each range we can/should use. Blues are the most efficient LED today, so if we can use them, good. But as whites are already giving blue, using blues together with whites will depend if the best mix have more blue respect the other wavelenghts emited by the white than its actually emitting.
Most people at the testing group is going to use a little blues together with whites, apart of the reds. Personally, i think its good, but results will say if im right or wrong on it.
But what i know is that excess blue harms more than helps, aswell as any other bandwith. So we must find the point up each color becomes damaging (or not helping), and use the most efficients ones just up that point.
G
Gr8fulDen
This forum has become a den of mad scientists! Botanists, agronomists, futurists...the "ists" go on and on! If college was as relevant as some of these posts I would have paid more attention. Thanks so much for the edge u cayshun!
Jay_Druthers
Member
Deleted
knna
Member
Yeah, color adjustable arrays is the best way to go on experimental setups. Modules we are building are current adjustable with a potenciometer. Some are going to wire all colors in series (driver is current controlled), but most are going to wire each color with its own driver, so its going to be very easy to change the light spectrum, using the same array.
We are working too on a cheap DIY quantum meter able to show the uE/m2 independent of the spectrum used. There is already a prototipe that will measure light density, total and on the 3 mains PAR ranges separatelly (RGB), temperature, relative humidity and CO2 level. Components cost about 30€, but we are seeking for the right photodiodes, as we need ones that sense the light density almost linear with changing wavelenght, and its being difficult to find one for the blue range.
The measuring module is intended to be used controlled by a computer, so we can get a log of the varying levels along the day. This module would allow to know light density and roughly, the spectrum distribution, of any lighting, so people using unkown efficiency LEDs may know what spectrum they are using, without the need of having an expensive spectroradiometer.
We are working too on a cheap DIY quantum meter able to show the uE/m2 independent of the spectrum used. There is already a prototipe that will measure light density, total and on the 3 mains PAR ranges separatelly (RGB), temperature, relative humidity and CO2 level. Components cost about 30€, but we are seeking for the right photodiodes, as we need ones that sense the light density almost linear with changing wavelenght, and its being difficult to find one for the blue range.
The measuring module is intended to be used controlled by a computer, so we can get a log of the varying levels along the day. This module would allow to know light density and roughly, the spectrum distribution, of any lighting, so people using unkown efficiency LEDs may know what spectrum they are using, without the need of having an expensive spectroradiometer.
Jay_Druthers
Member
I don't know if this helps but the tcs230 8 pin ic is a interesting color sensor.
billyjojimbob11
Member
jtk707, thats great! I think blue leds have already defeated the purpose of using Metal Halide.
Knna, what 640 reds are you using?
Knna, what 640 reds are you using?
knna
Member
Finally, im using the Osram Golden Dragon Plus, bin JX (45-52lm@400mA). Actually, they arnt 640nm, but 635nm on operating conditions. Aprox the same perfomance than Crees, SSC, Lumileds...but way better price.
Im using too the Osram Power TopLED, that are mid power SMD leds. They are 645nm, and still a bit longer. Im using the bin AB (4.8lm@50mA). These devices emits on a excelent wl and are pretty efficients (over 31% at 50mA, 1.62 uE/W), but i still dont know if their intensity is high enough. Experimental setups will tell. But they have the best perfomance/cost ive been able to find. They seem the best choice for bottom areas suplemntary lighting, as their cooling needs are almost nill, so they can run directly on a PCB hanging between plants, requiring no space (3mm thick the full module). Im experimenting too with a reflective coating (98% average on PAR) on them, to avoid they absorb light.
OperationIvy
New member
knna
knna
I've been following your posts and threads for a few weeks now. Is this forum and thread the one I want to follow for the latest and greatest? I noticed that you've been on a couple different sites and have info spread out over several threads. The internet is full of nonsense regarding artificial lighting for cannabis but I've found your threads and posts to be insightful and as well researched as can be. I fully realize the potential of LEDs in an LST or SCROG setup and would like to participate in the research and development of an array. I would like to invest some money into building an array towards the end of the summer and would love you input along the way. I've done some research myself and am currently reaching out to a couple of friends in the universities for some additional help. One who I haven't met is a DR. out in California and has plenty of experience working with LEDs in his research. I don't yet know if any of his work has been horticulture related but perhaps he could become an asset for aquiring data on some newer LEDs. I am starting a new job here next week and am in transition but will be back up towards the end of the summer and looking to experiment again. I look forward to your future posts.
knna
I've been following your posts and threads for a few weeks now. Is this forum and thread the one I want to follow for the latest and greatest? I noticed that you've been on a couple different sites and have info spread out over several threads. The internet is full of nonsense regarding artificial lighting for cannabis but I've found your threads and posts to be insightful and as well researched as can be. I fully realize the potential of LEDs in an LST or SCROG setup and would like to participate in the research and development of an array. I would like to invest some money into building an array towards the end of the summer and would love you input along the way. I've done some research myself and am currently reaching out to a couple of friends in the universities for some additional help. One who I haven't met is a DR. out in California and has plenty of experience working with LEDs in his research. I don't yet know if any of his work has been horticulture related but perhaps he could become an asset for aquiring data on some newer LEDs. I am starting a new job here next week and am in transition but will be back up towards the end of the summer and looking to experiment again. I look forward to your future posts.
OperationIvy
New member
knna,
have you made use of the Cree Product Characterization Tool provided on their website? I was wondering if you found it helpful at all. I'll be playing around with it tonight. I really wish that I could get my hands on a copy of Radiant Imaging ProSource or LightTools. I have some good theoretical schematics for a high density array but with some computer aid I think that it could be put to a simulated test and help optimize the array. It may be worth plopping down the 1K for the software if others could put it to good use.
have you made use of the Cree Product Characterization Tool provided on their website? I was wondering if you found it helpful at all. I'll be playing around with it tonight. I really wish that I could get my hands on a copy of Radiant Imaging ProSource or LightTools. I have some good theoretical schematics for a high density array but with some computer aid I think that it could be put to a simulated test and help optimize the array. It may be worth plopping down the 1K for the software if others could put it to good use.
MrNebulized
Member
Awesome post! I didn't think that there would be a huge difference, but it is interesting that some of the strongest bud comes from areas that have heavy UV light, Afghan is a phenomenal example.
I am about done perfecting my nutrient up take(true aeroponics grow), LED lighting was initially on my list first. Forced to build a bigger system due to inheriting some new plants. Didn't want to do it twice, figured I should just do it right this time around.
I want to build 2 100 watt LED lights.
The numbers I seen you post elsewhere are:
70% Red(625 NM- 685 NM)
16-18% Blue (440 NM - 450 NM)
12-14% White LEDs(green-yellow)
Ok now KNNA is there a website/book you could refer me to to learn how to build a light? Give me some general idea of how the actual LED light works electrically and how to wire one up.
I have plenty of info on what products I should choose from reading the post, but it would be nice to understand the actual putting together lol.
Thanks in advance.
knna
Member
I didnt see it until know. Thanks for pointing me out of it. It seems very useful. Ive done spreadsheets to calculate that same parameters, and its a very good way to know what LED model and flux bin to use better, aswell as determining the max current to use which each model.
Me too
But Osram has put on the web photometric models of their LEDs for the free Dialux platform. It allows you to do simulations of light spreading for any setup. So you can find the best configurations for your arrays, and where to place them in the grow chamber/room. You may get a simulation that gives you the light density at any point, in 3D.Simulations are very good, as they save lots of work. But notice they are only that, simulations. You can avoid doing large errors by design and simulating arrays before building them. But at the end, the only way to check how it works is building it and taking measurements.
Specially about thermal management, its very dangerous to put your money on a simulation. There are so many materials, ways of building the arrays,etc, that its almost imposible to do a realistic simulation of it. So try to leave always some margin for temps and take measurements of a prototype, in order to check real performing it isnt going too far from the simulation.
If you dont read on spanish, yes. I wrote very deep post about LEDs and plant lighting at Overgrow and Cannabisworld that have been lost. I deleted all indeep info i posted at this forum some ago, and im reluctant to write again about it. I dont like moderation of this forum at all
But sometimes is hard for me to not answer to people i see with good intentions.Anyway, i feel more confortable at the Garden's Cure, that is my home english forum. I would prefer to talk about advanced questions there. Ive posted there advanced info about plant's lighting, and open some threads to discuss ways of doing DIY LED's arrays, but there was little interest on them.
Anyway, currently im very busy and im having little time to post. Tomorrow i plan to build a LED array and doing and tutorial with pics of how to assemble it. Its intended for my spanish forum, in order all the people wanting to do their own LEDs arrays knows how to do it, with the materials we are going to use. Surely when i have some time ill translate to my poor english and post it at the Garden's Cure.
The most difficult part when doing a LED array is designing and building the driver. There are a lots of different options of doing it. If you know electronics, there are several Integrated Circuits (IC) designed for doing LED drivers, that allows you to make the array exactly as you want using just some components: some resistors and capacitors, an inductance (for some topologies), a bridge if you are going to run them from the mains. Its the cheaper and most efficient way, but not all the people have the electronic knowledge to do it. For them, the best way is to use a comercial driver and design the array adapted to it.
The other esential point is designing the thermal management. But its different depending of the LEDs used, materials and way of assembling and the current used.
After that, it only consist on attaching LEDs to heatsink and soldering wires.
I really hope to offer the materials needed to do your own LED arrays before the summer. Im testing heatsinks, drivers, LEDs, etc..seeking for what offer the best perfomance/price for our application and of the easier assembly so anybody can build his own array. Im seeking too for solderless ways of assembly aswell for reflective finishing of the arrays.
MrNebulized said:
Really i dont know any
That info is dispersed. Probably the best source to learn about it is the Candlepower forums. Look for the "fixed lighting", so large arrays are in it. About different LEDs perfomance and how to run them, you can learn all there. But users there are little into the coloured LEDs and high wattage arrays, as its mostly a flaslight's hobbyist forum.Im trying to compile such info on located places, so anybody wanting to do their own array can learn how to.
MrNebulized said:
Be aware that is refered to percentage of PAR watts, not to the number of LEDs or power consuption. And anyway, that guideline is based on my own experience, but i still dont know if its the best distribution. I hope to have narrowed it after some monts more.
But i believe the best distribution is going to be between:
10-25% blue
5-25% green
55-70% red
Again, talking about PAR watts. Translating it to number of LEDs depends of what LEDs are going to be used and at what current.
Jay_Druthers said:
Thanks, Jay, but it have the same problem. Response of the sensor vary a lot depending of wavelength. I need a sensor that gives similar current along a large bandwith, or measuring error will be too large to use with narrow wl emission LEDs.
Peace,
knna
MrNebulized
Member
Oh yeah, I understand that PAR watts is what we actually shoot for. I am speaking directly from a electrical stand point. My space is tiny 2 x 4, so I figure two 100 watters should be quite sufficient, especially since I plan on supplementing with a couple T5s(Hybrid lighting seems the best at this point). I know that the 400 watt CMH that I am using puts out guesstimatingly 130-140 PAR watts. So I do understand that there is quite a difference between watts and PAR watts. I just figure that if I have 100 watt foot print over 2 X 2 a long with a couple T5s it should give me even more PAR watts. Is there a conversion for LED watts to PAR watts? I have read a lot of your posts but never have I seen a conversion.
So I do know Procyon is full of it as well as UFO when they put the equivalent of. PAR watts is the game we will be playing, but electrically that is the system I am looking for.
So I do know Procyon is full of it as well as UFO when they put the equivalent of. PAR watts is the game we will be playing, but electrically that is the system I am looking for.
OperationIvy
New member
knna,
would you be comfortable sharing the spanish forum with us? I can translate well enough for my own enjoyment and I would like to follow some of the work you are doing in spain. also, Could you explain what it is that you meant about the moderation of this forum? Thank you for your efforts.
would you be comfortable sharing the spanish forum with us? I can translate well enough for my own enjoyment and I would like to follow some of the work you are doing in spain. also, Could you explain what it is that you meant about the moderation of this forum? Thank you for your efforts.
billyjojimbob11
Member
Hi guys,
I'm thinking this for flowering...
21 x XREWHT-L1-0000-009E7 2600-3700k Warm white top P4 bin 80+ lumens per watt (1 x 21w) spot light @ 350ma
(covers all colors from 420-750nm) with excellent red and blue PAR efficiency white!
42 x (2 x 21 spot) watt highest efficency top bin OSRAM red spot light.
(Knna boid, what is most efficient red top bin model #)?
NO BLUE, as blue is covered plenty with the Warm white
I think this will be hard to beat for flowering at this time.
These warm whites are more efficient than warm CFL's in PAR.
I'm thinking this for flowering...
21 x XREWHT-L1-0000-009E7 2600-3700k Warm white top P4 bin 80+ lumens per watt (1 x 21w) spot light @ 350ma
(covers all colors from 420-750nm) with excellent red and blue PAR efficiency white!
42 x (2 x 21 spot) watt highest efficency top bin OSRAM red spot light.
(Knna boid, what is most efficient red top bin model #)?
NO BLUE, as blue is covered plenty with the Warm white
I think this will be hard to beat for flowering at this time.
These warm whites are more efficient than warm CFL's in PAR.
OperationIvy
New member
billyjo,
i was also looking at the warm white cree's. I think i read somewhere that the cool whites are more efficient but i like the PAR spectrum of the warms. As a side note I was looking at the neutral whites displayed by the green line in the image below. If you were going to combine reds past 650nm but wanted to get a "better", lower nm blue spectrum then I think the neutral white is the best choice. they have a much better blue than the warm and slighter better than the cools. if you're going to combine them with the reds i think your going to get your best PAR coverage
these are the xr-e bins. let me know what you think. im going to photoshop this diagram so you can see the reds on the graph with whites. might give a better overall impression of the PAR coverage. if i am trucking down the wrong path please feel free to let me know.
i was also looking at the warm white cree's. I think i read somewhere that the cool whites are more efficient but i like the PAR spectrum of the warms. As a side note I was looking at the neutral whites displayed by the green line in the image below. If you were going to combine reds past 650nm but wanted to get a "better", lower nm blue spectrum then I think the neutral white is the best choice. they have a much better blue than the warm and slighter better than the cools. if you're going to combine them with the reds i think your going to get your best PAR coverage
these are the xr-e bins. let me know what you think. im going to photoshop this diagram so you can see the reds on the graph with whites. might give a better overall impression of the PAR coverage. if i am trucking down the wrong path please feel free to let me know.
One of my arrays I built last year uses Cool White P4 CREE XREs and red CREE XRs, works very well, I think the coolwhite + red combination is the best one to go for from my experiences.
billyjojimbob11
Member
Ivy
I think your on the right track, depending on what you are doing
If you are building an all in one array(veg and flower), then yes... But my goal is to create the most efficient flowering array using leds on the market with the most PAR, mostly red. I will use a different, mostly blue array for vegging. I have found that the colors of blue in a "flowering array" is the least of our concerns. Blue, as Knna pointed out, is much more efficient than red due the material emitting the light and its efficiency. My goal is to have just enough blue light, but not to much, with the majority of my colors going to yellow, amber, red, and far red, and some green, all covered quite well within the 2600k cree XRE, bin P4. If you look at the spectral distribution of an HPS, we see that blue accounts for very little energy, actually it has nearly as much green (nearly useless for PAR) light. And again, blue is not what I am after.
I am assuming that I will get plenty of this blue light with the white leds since there is still a fairly large InGan spike, I can always add blue leds if i dont have enough, and blue leds are much more efficient than If I were to give up the broader yellow-red spectrum that I will be getting with Warm White. After all, white leds are phospered covered blue leds to begin with.
One main difference between MH and HPS is far more 580-600nm light, the other being more blue light. This is what I am after...Read this abstract:
http://www3.interscience.wiley.com/journal/119928937/abstract?CRETRY=1&SRETRY=0
That gap, along with more efficient red leds is the only thing keeping Leds from taking the HPS market.
BTW, Knna, I think I will go with OPTEKS, as they have the most efficient red leds I could find @ 53lm/watt @ 350ma
http://www.optekinc.com/viewparts.aspx?categoryID=48
OVTL01LGAR
Of course with an angle of 135°, we will need a lense to narrow.
I think your on the right track, depending on what you are doing
If you are building an all in one array(veg and flower), then yes... But my goal is to create the most efficient flowering array using leds on the market with the most PAR, mostly red. I will use a different, mostly blue array for vegging. I have found that the colors of blue in a "flowering array" is the least of our concerns. Blue, as Knna pointed out, is much more efficient than red due the material emitting the light and its efficiency. My goal is to have just enough blue light, but not to much, with the majority of my colors going to yellow, amber, red, and far red, and some green, all covered quite well within the 2600k cree XRE, bin P4. If you look at the spectral distribution of an HPS, we see that blue accounts for very little energy, actually it has nearly as much green (nearly useless for PAR) light. And again, blue is not what I am after.
I am assuming that I will get plenty of this blue light with the white leds since there is still a fairly large InGan spike, I can always add blue leds if i dont have enough, and blue leds are much more efficient than If I were to give up the broader yellow-red spectrum that I will be getting with Warm White. After all, white leds are phospered covered blue leds to begin with.
One main difference between MH and HPS is far more 580-600nm light, the other being more blue light. This is what I am after...Read this abstract:
http://www3.interscience.wiley.com/journal/119928937/abstract?CRETRY=1&SRETRY=0
That gap, along with more efficient red leds is the only thing keeping Leds from taking the HPS market.
BTW, Knna, I think I will go with OPTEKS, as they have the most efficient red leds I could find @ 53lm/watt @ 350ma
http://www.optekinc.com/viewparts.aspx?categoryID=48
OVTL01LGAR
Of course with an angle of 135°, we will need a lense to narrow.
billyjojimbob11
Member
BTW
Red efficiency by brand: @ 350ma 1 watters
Cree 40lm
Seole 48lm
Phillips 40lm
Osram 49lm
Optek 53lm
Red efficiency by brand: @ 350ma 1 watters
Cree 40lm
Seole 48lm
Phillips 40lm
Osram 49lm
Optek 53lm
