Building LED lights from facts, no theories

[flyingimam]

Jackpot!!!
I'm subscribed, totally!
keep the precious info comin, there are people lime me reading in silence

I'm a total n00b in regards to LED lighting n generally lighting, but I'm learning and since i probably wont own a place of my own for another 2 years, i may as well start the grow operation when there are some improvements made with a ton of education that i will accumulate till then

[knna]

Or another one.
Match the spectrum to
420-490 30%
490-620 10%
620-700 60%
What kind of soft do you use to build spectra graphs, may I have a link?

Check the 4R3+CW+RB on post #8. It has almost the same distribution, except on the red range where it only reach up to 670nm. In order to get a less concentrated output on the red range and increase its bandwith up to 700nm (it should cover slighty after it) you can use 2R3 and 2 650-660nm peaked red leds, as the LedEngin ones. They are less efficients than the M bin of red Crees, but no so much you cant use them if you want to cover the longer part of the red range.

You should be aware than when choosing LEDs, you need to consider carefully not only models/brands, but bins of each one. Usally LEDs are binned according three parameters: luminoux flux, color and Vf (forward voltage).

For same model, often there are 3-4 color bins that cover the range. For example, red Crees are avalaible on R2 (620-625nm dominant wavelenght), R3 (625-630nm) and R4 (630-635nm) color bins, although they give up serving the R4 on the high power models (its only avalaible for the medium power XL4550 currently). There are aswell 3 color bin for the Royal Blues and more than 20 for coolwhite.

Luminosity bins determines LEDs efficiency. For same color bin, there are different light fluxes avalaible. As for same SPD, higher photometric flux (lm) means higher energy efficiency, and often top bin is more than double than bottom bin, almost any LED model have a very wide range of efficiencies.

We always should try to get top bins, but they are more expensive and not always avalaible.

Cree is currently the better manufacturer for InGan leds (Royal blue, Blue and white). They are a step forward the competence. They are serving currently up the bin group 15 for Royal Blues (>450mW@350mA) and up to bin Q5 on coolwhites (107-114lm@350mA) for the XR-E model.

There are some other manufacturers that are using the Cree EZBright1000 chip on their models, as Seoul Semiconductor, Edixeon and Kingbright. They are a cheaper alternative, but construction quality is often far from Cree ones. Osram has been licenced the EZBright too but its still unavalaible.

For the red side, Cree does good LEDs, but difference is smaller than for InGans. Indeed, Osram does more efficient red leds, but unfortunatelly Osram not allow to select a single bin as Cree or Lumileds.

So when selecting what LED model to use, its impossible to generalize: best choice would depend on what you can get on your situation, involving how many LEDs are you going to buy, where you live, if your distributor allows to select concrete bins, etc.

Although prices for manufacturers is very competitive and there is little differences, when an individual look for a given model/bin, differences may be very large.

So one thing is the ideal LEDs to use, and other what you can actually obtain, and price.

So it would be better to have a general idea of what models are valid for us, seek for them and select the best avalaible for you.

In the USA and Canada, for example, ordering from Future Electronics (official distributor of Lumileds) allows you to select individual bins and get decent prices still buying low quantities. They arnt the best LEDs avalaible, but probably they have the best price/perfomance for a northamerican.

And the other good source is Digikey, that distributes Osram at very competitive prices in NorthAmerica. Not bin selection allowed, but excelent average price/perfomance. For example, the red Argus (Golden Dragon with lenses, from 20º to 160º beam angles) is 1.64$ buying 200.

So i suggest you first look for what leds (models/bins) you have avalaible, and then select the best choice between them. I can help you on the selection, but i dont think it would be worthwhile to list the best model/bins possible if you cant access them.

I built those spectrums using OppenOffice (its a free suite equivalent (but better, IMHO) than MS Office. But ive uploaded the Bulb Analizer tool on Excel format too. Ive liked it on the Perfect LED grow light thread. I uploaded too the sheet for the Cree some post later, aswell as how to calculate the true radiometric output.

In order to build the spectra, ive used the Output percentual wavelenght column of the sheet. By multiplying it by the true radiometric output of the LED you get the normalized SPD so you can sum different LEDs SPDs and build spectra of any combination.

[glutek]

hi
first I have to mention that i didn't manage to read all the news on 'perfect grow' so forgive me if I ask/say smth that's not new..

knna, could you please gimmie a hint where should I buy small numbers of LEDs in EU? do you live here actually, I noticed you removed Spanish flag from the profile?

btw, I wouldn't say cannabis is little demanding about light quality (otherwise we could grow under MV ie.), better to say it's just resistant as a weed and grows quite good under different conditions. just a better word.
wheat is a monocot so not a best comparison. still good, as most important photoreceptors are so old origined that are found ubiquitus, and crys are found in likely all animals, including mammals, as humans, for sure
I'd rather say cannabis will benefit from more complex spectrums. but in this topic I mostly wanted to ask for a hint at the moment, dont really know if I want to talk about quality here. thanx a lot in advance

another question.. do you think(have you experimented) that in vivo it doesn't make a critical difference to have equal quantity of photons at, say, 10 or 20 nm away from 660 and other peaks? as we know from mcree and inada it doesn't affect photosynthesis rates much, but what about the overall effect?photomorphogenesis, plant architecture?

third..what's more efficient for FR, blacklight incands or blacklight fluos?

[glutek]

third..what's more efficient for FR, blacklight incands or blacklight fluos?

[glutek]

if possible, please, can you also include the data about where to order FR diodes too? for some reason I'd like to include them in project

[knna]

hi
first I have to mention that i didn't manage to read all the news on 'perfect grow' so forgive me if I ask/say smth that's not new..

knna, could you please gimmie a hint where should I buy small numbers of LEDs in EU? do you live here actually, I noticed you removed Spanish flag from the profile?

In EU there is little good choices for getting LED directly from distributor on small quantities. Just LED shops, that charges large margin profits and often not inform about bin being on sale. LED tech, LED1, dotlight...

For little leds, i use dealextreme. Its from honk kong, but they send you the items (less than 100â?¬) without any problem. Competitive prices with free shipping. But they are more especialized on top end white leds, although sometimes they have some decents of colors.

Im currently doing a group buy of leds and constant current drivers (from mains). At manufacturer's price, top bins (SSC P4 red 635nm [email protected], Crees Royal blue (group 15) and white (Q4-Q5)). 1.94, 3.3 and 2.4â?¬ respectively. CC drivers less than 10â?¬. If you are interested, send me a PM.

btw, I wouldn't say cannabis is little demanding about light quality (otherwise we could grow under MV ie.), better to say it's just resistant as a weed and grows quite good under different conditions. just a better word.

When i say is little demanding, its on the sense it requires little (if any) of any given wavelenght. Its a different thing if it performs better using certain spectral distributions.

Cannabis had been grown under MV lights. They were displaced by HPS and MH (that are enhanced MV bulbs, BTW) due mostly the superior energy efficiency. More photons per watt=more yield per watt.

Spectrally, MH are superior to HPSs, but we already know that except very special conditions, at equal wattages HPSs produce more, just because they emit more photons per watt. Some people choose to use MHs at a cost of lower yield because they prefer the quality, but cannabis production is mostly determined by photons absorbed by plants (given there is no other limitant factors, that shouldnt on a well designed grow room).

The main parameter to know perfomance growing cannabis of a given bulb is how much photons per watt burned it emits. And HPS rule here. Until this year it has been impossible to get same photons per watt using LEDs, and spectral advantage only compensate partially that.

wheat is a monocot so not a best comparison. still good, as most important photoreceptors are so old origined that are found ubiquitus, and crys are found in likely all animals, including mammals, as humans, for sure
I'd rather say cannabis will benefit from more complex spectrums. but in this topic I mostly wanted to ask for a hint at the moment, dont really know if I want to talk about quality here. thanx a lot in advance

Yep, i know wheat and cannabis are very different. I just use it as example of a very well known plant that is almost light quality insensitive.

another question.. do you think(have you experimented) that in vivo it doesn't make a critical difference to have equal quantity of photons at, say, 10 or 20 nm away from 660 and other peaks? as we know from mcree and inada it doesn't affect photosynthesis rates much, but what about the overall effect?photomorphogenesis, plant architecture?

Yes, i have experienced that. I think is good to use both 645 and 660nm red light in order to avoid electron flow imparements between Photosystems I and II that may reduce quantum yield. But it happens mainly at high irradiances. When using medium-low irradiances (below 500 uE/m2), there is little difference on using 660 or 635nm.

For that reason, i choose to distribute the light evenly (to avoid high puntual irradiances) and use LEDs that emits the most photons possible per watt burned. Currently, 635nm ones. Its sad that most radiometricaly efficients AlInGaP leds emits at 654nm, which is almost perfect for us (maximum quantum yield (10.3 photons per O2) recorded was using a 657nm peaked led) but manufacturers dont do it (i dont know if LedEngin uses AlInGaP chips, but i doubt it, as effciencies are about 21-22% and its possible to get over 40% with known technologies. Maybe its just a patents issue).

Photomorphogenesis is mostly driven by blue light (cryptochromes) and red/far red relationship (phytochromes). From 620 to 700nm is red light, so there is little difference on using a peak 20nm ahead. Practical differences are negligible.

[knna]

third..what's more efficient for FR, blacklight incands or blacklight fluos?

By far, an incandescent. And you dont need a blacklight one. You can use it, but it use high K to get some UV emission, and it results on a way reduced FR emission. I think people uses incand blacklight to avoid getting a filter, due they emit very little on the visible range. But im not sure at all that they filter all visible light.

A filtered standard incand is the cheaper and more efficient way of adding FR.

FR leds arnt very efficient, but decent enough. And for experiments with 1-2 small plants, you would need little. I believe LEDEngin has released 740nm leds recently. Roithner-laser has them too.

[glutek]

thanks again. I'm good at light in physiology but not a specialist on light sources. And electron flow imparements was smth I didn't thought about before I know what sensors drive photomorphogenesis, and wanted to confirm what I thought, that in vivo they react quite good also to somewhat shifted wavelenghts, thn for your opinion. Anyway, imo, of course, photons quantity are very important, but quality counts indirectly via driving morphogenesis. So better to say cannabis is not 'almost insensitive', it probably just less sensitive/needs higher ratio of red to green and blue than more shade loving plants. But it's only the issue of formulating words for what you of course already know, otherwise you would propose pure red, while proposing some blue and green to treat stretching, stomatal and other responses
Well, a propos green light, I guess it could be possible to reduce GL if adding FR, becuse leaves probably would grow larger and thinner under FR supplementation.
Nice you want to help with buying..but I can't send you a PM

[knna]

It seems there is no PM service at this board. So i give you the link to the group buy: http://www.cannabiscafe.net/foros/sh...d.php?t=125174

Im knnabinoide there. Its a spanish forum, but it has a english users subforum (used little, BTW).

[oldmac]

knna,

Sorry I've been away for a few months, or I would have chimmed in sooner.

First, thank you so much for your thread here and postings on the other LED thread. I've spent days catching up on my reading here, and I have gotten a whole lot of good info from your postings, plus a bunch of confirmations to things I've seen and done.

I started experimenting about 2 years ago, and in the last year just said screw it, let me apply the technology NOW; build some DIY stuff and also try what's out there: I bought two Procyons last year and two SmartLamps this past August.

Last year I had a failed experiment for a bloom bulb set-up, a combo of red Cree's and (another mfg) whites. What I found was the actual spectograph of the white sucked, it was a bunch different then the mfgs graph. The blue die used was way to the left of ideal blue and the coating (phosphor ?) responce really lacked any red. Oh, it's light output was white...to the human eye, but lacked blue and red for a plant. The real dissapointment was, I had solved other physical problems (I was growing in a rotating garden and needed 360 degree light output, not LEDs strong point!) using an octagon aluminium piece mounting the LEDs inside the glass tube used for HPS.

I stripped the white LED PCBs from the alum octagon (don't ever use heat sink epoxy- gease and mechanical fasteners are better) and replaced them with 2' T5s driven to VHO, 40watts a tube. 8 times 40w gives 320watts of T5 and 8 strips of 2 watt red Cree for 160 watts, the resulting 480watt hy-bred exceeds my wildest expectations. It out grows a 600watt HPS by a good amount. IMHO LED/T5 together is better then either on its own.

I'm just about to build a 4 light 4' T5 VHO (340watts) to supplement a TI SmartLamp (300watts) to get enough "umph" to cover a 4'x4' aero/fog grow tray. We'll see how well it does soon.

I have some questions for you, hope you'll be around awhile. There have been so many good LED experimenters on these boards, both DIY like physicsnole, redline, veggi and others and the guys willing to take the plunge and buy then document thier grow experiences, led by the likes of snsstealth and now others. While some here just want to do pure LED, I think many will start to embrace the hy-bred concept, untill we can get ideal LEDs manufactured.