Building LED lights from facts, no theories

[oldmac]

veggi,

Sorry to hear you've been ill.
Hope things get better for you....that's what I keep hoping for myself.

[knna]

Hey, oldmac, thanks for sttoping by here.

I cant agree more whit your post. Ive used CFLs (twin T5s) plus red leds on my main cab. Until now, its been by far the best compromise of price/perfomance. It avoid any spectral lacking and its an excelent way of growing SOG or SCROG style.

But last CREE whites are pretty efficients, and their spectra is decent, similar to halophosphor fluorescents. With some far red emision too. With efficiencies over 90 lm/w at 350mA they offer more light per watt burned than any fluorescent. But price is still higher for just an small perfomance enhancement. But at least its possible to use them together with reds and blues (or just reds).

I wonder why you mounted the octagon inside the cooltube. I think it would be better to mount the exhaust directly to the octagon and aircool it form inside (if the diameter is large enough, add some aluminium sheets inside to improve cooling perfomance) and remove the glass which is blocking at least 10% of the light.

[RackitMan]

What color temperature T5s did you use?

Also, I thought overdriving to be less efficient power wise (watts/lumen), is it not?

[oldmac]

What color temperature T5s did you use?

Also, I thought overdriving to be less efficient power wise (watts/lumen), is it not?

Hey RackitMan,

I used full spectrum, tri-poshphor T5s, mostly because I was gun shy of "warm whites" from my LED experiment. I wanted to make sure I covered the blue side of things to off set the red Cree's. The light set I'm building right now is going to use 4'- 2,700K T5s, sold or made by Sunblaster. Finally a T5 with less then 3000K! (They also make a 2' T5 and for CFL fans a 200watt CFL in 2,700K.) The TI SmartLamp seems to have a more balanced (?) light spectrum output, so I'm pushing the "warm" side for flowering with the T5 2,700K bulbs.

A T5 at HO levels has about 95-100 lumens/watt at VHO levels abt 115-117 lumens/watt or more. Almost twice the lumens for 65-75% (depending on size) additional power input....Id say it's more efficient. Also takes fewer bulbs to cover an area, I can use 4 VHO driven tubes to cover the area of 8 HO.

[oldmac]

It avoid any spectral lacking and its an excelent way of growing SOG or SCROG style.

I wonder why you mounted the octagon inside the cooltube. I think it would be better to mount the exhaust directly to the octagon and aircool it form inside (if the diameter is large enough, add some aluminium sheets inside to improve cooling perfomance) and remove the glass which is blocking at least 10% of the light.

I came to the conclusion mnay years ago that for indoor growing, single cola plants (many of them) less then 2' was the way to go, 18" or so is ideal.

As for using the cooltube, it was just a matter of necessity....to protect the LED PCBs and T5s from dripping nutrient and water. Plus it is easier to clean a glass 6" tube then 8 individual T5 tubes not to mention the 8- 20" metal PCBs. I clean the glass at least once a day!

It sounds sexy to say I grow in a rotating garden, but it is the most labor intensive system invented by man so far, and it is not a "clean" operation. Only thing worst I can think of is being a dirt farmer. BUT it has had one advantage... more then twice the growing output with the same amount of light input.

The question I have for you kanna is can you quantify light output in terms of micromoles of photon energy for T5s? I have to admit I didn't know a micromole from a gopher until I read some scientific papers posted by physicnole; but it seemed to make the most sence when talking about plant growth. I saw you mention that a 250HPS had about 100 micromoles, that would put SmartLamp at 310 claim well within a possible 600w HPS.

I like how the SmartLamp preforms; in a grape vine propagation experiment, it covers and grows equal to a 600HPS, but with a high light level loving plant like MJ..."a photon whore" (I love that) it suffers what most "brick design" hi power LEDs suffer from, coverage area. Bring the light close enough to take advantage of its higher output, (and you can because there is no heat) and the area covered is just too small. Raise the light to get a wider footprint and suffer with not enough "umph" for those photon whores.

I just happened to luck into my first T5/LED hy-bred light. Its results and what I'm hearing from others, seem to confirm that currently it's the way to go. When I first started with LEDs, I was comming from the "I'm going to save at least 50% of my electric bill". I even got hung up with trying to compare grows to a 600w HPS, but frankly now even if I use the same amount of electric power, if the results in both weight and quality are superior, that is what counts. My new hy-bred uses about 640watts of light, so I hope it can compare favoribly to a 600watt HPS. If not I really am going backwards here.

[RackitMan]

Did you overdrive your T5s (how did you do it?) or did you purchase them overdriven?

Also saw an older fluorescent light spectrograph, though I imagine it has not changed very much. Very interesting is that the warm white actually has more blue than the cool white. What makes the cool white look bluer is more green and yellow. This is backward to almost everything posted here, but check it out for yourself. (Link is currently not working for me: www.weedfarmer.com/cannabis/lights.php ) The cool white has much more red light.

[knna]

I came to the conclusion mnay years ago that for indoor growing, single cola plants (many of them) less then 2' was the way to go, 18" or so is ideal.

Fully agree. I grow SOG style, 15-20" tall plants at harvest time. Floros and LEDs (without narrow optics) don go further than that.

I prefer to grow two stacked tables (one over the other) of 15" plants than one with 30" ones. Higher bud to total dry matter ratio with shorter plants, apart of the reduced (almost none) veg time.

Anyway, the arrays im designing now are intended to be used into canopy so im going to try to grow taller plants that produces good buds along all the height.

As for using the cooltube, it was just a matter of necessity....to protect the LED PCBs and T5s from dripping nutrient and water. Plus it is easier to clean a glass 6" tube then 8 individual T5 tubes not to mention the 8- 20" metal PCBs. I clean the glass at least once a day!

It sounds sexy to say I grow in a rotating garden, but it is the most labor intensive system invented by man so far, and it is not a "clean" operation. Only thing worst I can think of is being a dirt farmer. BUT it has had one advantage... more then twice the growing output with the same amount of light input.

Yep, in that sense vertical grows works better, although the problem with them is to get an even nutrient distribution for plants at different heights.

The question I have for you kanna is can you quantify light output in terms of micromoles of photon energy for T5s? I have to admit I didn't know a micromole from a gopher until I read some scientific papers posted by physicnole; but it seemed to make the most sence when talking about plant growth.

Yes, i can do it. You can do it aswell. Ive uploaded the spreadsheet to do it easily. The spreadsheet uploaded on The Garden's Cure already have floros SPD's from phillips digitalized.

54w T5 HO have about 27.5% energy efficiency working at 35ºC, for about 76 uE in PAR (~1.29 uE/watt, already counting ballast losses). It drops to 24% and 66 uE when running at 25ºC.

Losses at reflector are about 25%. So 76*0,75=57 uE avalaible for plants. Close to 1 uE/W.

Clearly uE (micromols of photons per second, abreviated) is the relevant figure when talking about plant's lighting. Any calculation or stimation must be based on it.

I saw you mention that a 250HPS had about 100 micromoles, that would put SmartLamp at 310 claim well within a possible 600w HPS.

I said that if the Procyon uses the top bins for CREE it may emit more than 200 uE so it should compete with a 250W HPS. If it uses normal bins, it likely gets half way to do it.

A good 250w HPS gives about 385 uE. If working with reflector, its about 290 uE avalaible for plants. Pretty close to Smarlamp output, which should outperform slighty it (the other smartlamp model states 265 uE, and that should be on par with the 250w HPS). Slighty improved photons absorbance from the LEDs spectra (near 10% based on NASA experiments) and similar average quantum yield (photosynthesis for photon absorbed).

I like how the SmartLamp preforms; in a grape vine propagation experiment, it covers and grows equal to a 600HPS, but with a high light level loving plant like MJ..."a photon whore" (I love that) it suffers what most "brick design" hi power LEDs suffer from, coverage area. Bring the light close enough to take advantage of its higher output, (and you can because there is no heat) and the area covered is just too small. Raise the light to get a wider footprint and suffer with not enough "umph" for those photon whores.

The critical factor is to give the adecuate light density (uE/m2).

Most HPS grows range from 500 uE/m2 (at 400w/m2, or roughly 35 w/sqft) to 1000 uE/m2 (70 w/sqft). Its clearly more than required, as those high light densities are provided by means of achieving enough photosynthesis on lower areas. Probably 300-400 uE/m2 are enough if growing short plants. And when growing taller plants, adding LED lighting to lower areas directly instead than from top seems the way to use way lower average light densities than those used by HPS grows without losing much yield but increasing strongly productivity (g/W).

I just happened to luck into my first T5/LED hy-bred light. Its results and what I'm hearing from others, seem to confirm that currently it's the way to go. When I first started with LEDs, I was comming from the "I'm going to save at least 50% of my electric bill". I even got hung up with trying to compare grows to a 600w HPS, but frankly now even if I use the same amount of electric power, if the results in both weight and quality are superior, that is what counts. My new hy-bred uses about 640watts of light, so I hope it can compare favoribly to a 600watt HPS. If not I really am going backwards here.

You have choosed a really ambitius target. 600w HPS are the kings about uE emission per watt burned. And they have a very decent spectra for cannabis. Perhaps easy to beat, but for little. Quantum yield for well tuned LED lamps should be slighty higher, but not much more of 10%.

So to beat a unreflectorized 600w HPS is really difficult today. Its way easier to beat a reflectorized one, wich cost a 25% of the light, and easier too if it has a glass barrier, which blocks an aditional 10% of light. You need to give at least 80% of the photons the HPS emits to be able to compete with it.

With current LEDs efficiency, you need to use top bins runned at 350mA max to be able to achieve it.

I think current state of art of LED are still insufficient to compete with 600w on circular gardens. Its possible to beat it by little, but at way higher cost, impossible to take back for electric savings.

Currently i think LEDs are good to give supplementary lighting to low areas of HPS grows and for very small grows. Large wattage HPS are still unbeatable on price/perfomance, IMO.

Today CREE has reported a lab result of 161 lm/W of a white high power LED. Surely it will take about two years to see it commercially avalaible, but at those LEDs efficiencies beat large HPS is going to be much easier.

[oldmac]

Did you overdrive your T5s (how did you do it?) or did you purchase them overdriven?

Also saw an older fluorescent light spectrograph, though I imagine it has not changed very much. Very interesting is that the warm white actually has more blue than the cool white. What makes the cool white look bluer is more green and yellow. This is backward to almost everything posted here, but check it out for yourself. (Link is currently not working for me: www.weedfarmer.com/cannabis/lights.php ) The cool white has much more red light.

I used before and I'm using now a fluorescent ballast made by IceCap inc. They are capable of driving T12, T8, T5 to VHO levels and can be used to power up non-ballast cfls like TT-105s.
As a matter of fact, this time I bought one of their T5 "retro kits" that comes with the ballast, end caps (water proof ends), stand offs, reflectors and wiring harness. I'll be building a frame out of aluminum channel, that will also hold the Ti SmartLamp.

I failed to mention before that the only downside is bulb life is shorten.

I think the trade off is good. For me the biggest up side is the ability to light a 4'x4' tray, with a square light footprint. And hopefully get an added bonus from the LEDs.

[knna]

Overdriving shorten tubes life strongly. Up to 1/10. And it reduces energy efficiency too, although little (about 15%).

I think is better (and especially, cheaper in the long run) to put twice the tubes than to run them twice the power. If you do it because space limitations, go to Reflex tubes. You can mount them almost touching themselves because they have an internal reflector with a 160º window. If there is no space for that, yep, overdrive is the solution, but its an expensive practice on the long run.

I wouldnt advice to do it except as the last option.

RackitMan, you are right on the money. Coolwhites (~4000-4500K) often are the tone which emits more between 500 and 600nm, the range in what we are less interested. Ive always used fluorescents to grow and i avoid 840. I use 865, 830 and 827.

But daylight tubes (typically 6000-6500K) emits about 40% of light on blue (simplifying the 400-500nm range) while warm whites less than 20%.

[oldmac]

knna,

I was trying to digest some of your earlier post.

As to my use of 4-4' T5 VHO, it is not the cheapest way out, but gave me a few pluses, like a square light footprint to match my 4'x4' trays and still have room for the Ti SmartLamp mounted in the middle. Something that 8 tubes would make difficult. As to shortened tube life, everything I've read indicates about 50% less usefull life. The 2700K T5 tubes are $9 each, I can afford to change them periodically.

Actually, from a cost standpoint, I would have done better with 2- 200watt 2700K CFLs ($71/ea) and a cheap bright star fixture ($35/ea)....but felt the light coverage better using 4' tubes.

I currently use regular 4' T5s over my ez-cloner set-up, actually two (2) 120s cloners side by side. I currently have a 5 bulb set up that I may increase to 7. I have one 4' in the middle, without reflector that is a super actinic bulb. Talk about blue output....it's all blue. The other 4 tubes are 6500K. I use the single actinic tube for the first 3-5 days, untill the stems sprout, then they get the rest of the light.

In my old set-up for the rotating garden, I clone with Rockwool plugs lit by 5mm LEDs mixed 3 to 1 blues to reds, about 30 watts worth.
My transplants from plugs to 4" delta blocks happens on a 2'x8' shelf that uses HGL DIY 14watt 5mm boards.....16 of them. For the last year and a half the little 5mm LEDs have done their thing...remarkably well. All of this prior to hi-power LEDs, but they actually worked.