delta9nxs

since my last visit here i have been all over every site i could find regarding leds and have come to the conclusion that it would cost me a fortune to replace my 4000k hps with leds. probably close to $10,000 us by the time i finished. the people who manufacture led "grow lights" are making claims that are not supported by users experience. the people building their own arrays are mostly in the experimental stages, with much debate about meeting the plants action spectrum with different frequencies. there exists the possibility that i could spend this money and not even get the same yield i get now. especially since i grow 1.5 meter trees with average yield in the 9-10 oz area. my lights are hung vertically in between groups of four plants. typically i run them about 8-10 inches from the foliage and rotate the plants daily. i'm getting excellent foliage penetration with the hps's. have you done any penetration experiments? at this point it appears leds are more suited to flat style grows like sog or scrog. again, thanks for your help! in the meantime i will be following this thread closely. later, delta9nxs

scottcanada

I have a medical permit to grow and I'm in Canada. Found this forum and it has been very helpful so far. My question is if any of you could comment on the setup offered by Theorem Innovation called the SmartBar: Theorem Innovation Inc. - TI SmartBar - Presentation

The only reason I think this unit (or a similar home made version) might work is that the heat is removed with cold water. My current setup already has water cooled air conditioners so I have cold water lines available. But these systems cost $2000 US each so I am hoping to be able to build my own for like $1000 each if possible. Can anyone comment on the above referenced unit and if a setup like this is likely to compete with 1000 watt HPS? Here is a link to review of it but I can only find this one:

http://forum.grasscity.com/grow-jour...-take-1-a.html

WeeD22MaN

Great thread!!

WeedMan

__________________

choader

delta raises an excellent question about penetration. Since light decays according to an inverse square relationship based on distance (with the assumption the light comes from a single point) AND all of an HIDs lumens come from what is essentially a single point versus and LED setup where there are many points of light source with lower lumens each, I would assume penetration is much better with an HID.

However, with LEDs you can:
1. get the light closer
2. use them all around the box instead of just hung above
3. put them underneath or between plants

Not sure how this translates to penetration comparison. I'd love to hear the answer.

Your work is always stellar knna. Love it

Choader

__________________
SS Bubblegum & DP Mazar Small Cab - Complete
Choader's 600W HPS Grow - Complete
Small Cab Grow - Complete

READ ME!! Posting Guidelines
Male and Female Pre-Flowers

knna

"Penetration" is a concept coined for HID lights. It has little sense when using multiple sources of lights, as with a LED setups.

For HID lamps, limited to be placed at some distance over plants due to heat, all light is emitted from a single point of light, and it must reach the bottom areas of the grow. As light emitted from the light source spreads out, light density goes dropping: same amount of light (photons) distributed along a larger volume. For an unreflectorized HID, it drops at a rate of the square of the distance (known as the square law), because the perimeter surface of a sphere with the lamp as center is 4*(pi)*r^2.

For reflectorized HIDs, light density with distance drops stteper, about 1/3 of the square of distance on average. But it still drops fast.

What matter for plants is the light density (amount of photons falling on the surface of a leaf of a given surface area). If you want it has good figures at some distance of the light source, its a must it emits the light with strong intensity (light intensity, photometrically measured in cd, radiometrically in w/sr (sr=sterorradian)). The higher the intensity of the light source, the higher the "penetration": ability to keep light density high enough at a given distance from the bulb.

But keep in mind all this reasoning is based on the fact of a single point of light source, from what all photons spreads out along the grow space.

Things are very different when using multiple sources of light. Considered individually, the light density produced by each LED alone goes dropping as farer from it, because same amount of light spread out on a larger volume. But when considered all together, as light density from one LEDs drops, it raises from the LED next to it. For a flat LED lamp, it works like with fluorescents, keeping light density very even at good levels along all the space below the lamp, but that still goes reducing as we go farer from it.

Here is when arise the main advantage of LEDs compared to other ways of lighting: they are solid and cool, so can be placed with no problems at the sides of plants, at any height. Indeed, they can be placed between plants (called by NASA intracanopy lighting).

This way, we can keep light density even along all the height of the plants, opposite to other ways of lighting. In this context, penetration means nothing. The problem is other: how to distribute LEDs to achieve a even light density, as it would depend of the LED used: how is its beam angle (light distribution) and intensity.

If using a flat LED lamp over the canopy, we have the same problem as with fluorescents: limeted penetration into the canopy, advicing to use SCROG or LST styles of growing. In this case, we can increase the penetration of the lamp by using LEDs with narrow angles of emission, that produces way higher light intensity than wider ones.

But, always that its possible, the best way to grow with LEDs is by using them too with side lighting or IC lighting. On that configuration is where LEDs really worth.

After this analysis, raises to consideration the best use currently for LEDs on large grows: bottom areas supplementary lighting. Ill go into it on next post

PS: i havent forgotten the pics for the tutorial

__________________


Bulb Analyzer Tool

knna

HID lighting has one great advantage for indoor growing and a one large disadvantage: very good energy efficiency and very bad light distribution, respectivelly.

In order to get a good penetration from a HID, it must emits with a very strong intensity. Reflectors, on the other hand, often concentrates most of the light just below the lamp. This produces a very uneven light distribution, both on the x-y axis (horizontally) and z axis (vertically)

Horizontally, on the center of the grow area there is a spot with very high light density, while on the edges, light density is way lower. Some types of reflector, with double parabola, as the adjust a wings, try to reduce this problem. Setups with several lamps minimize this problem, too.

But the larger problem with HID is on the z axis. Thus, the importance of the penetration concept when using HIDs. If you want the light density is high enough on bottom areas, you need to use very high light densities at the top canopy level.

This configuration is strongly inefficient. Plants use very bad the light when the light density is too high, wasting lots of them. HID growers are using a minimun of 500 micromols of photons (uE) for sq meter (45 uE/sq ft), often 800 uE/m2 (72 uE/sq ft) and many growers still way higher.

While using light densities on the sweet spot of plants we can use 250uE/m2 (22uE/sq ft) and get excelent buds.

Vertical SOG using unreflectorized HIDs has proven this long ago. Ive checked it too using LEDs.

So HID growers are throwing massive amount of light that almost not increase nothing upper production in order to get some production on the lower areas. This practice is a severe waste of valuable light and strongly reduce the productivity (g/W) of the whole grow.

Here is where LEDs comes to the rescue We can free our HID lights of the need of provide light to the bottom areas, and use them where they are really unbeatable currently: for upper areas.

(Side note: it is possible too to enhance HPS lamps by placing some blue LEDs next to them, in order to use cheap regular HPSs instead of expensive blue enhanced ones, that anyway arnt good at all emitting the little blue they emits).

How to implement this suggestion (using the HID just for the top)?

First off, using less lamps. We are going to strongly reduce the top lighting, because it isnt really necessary, just by avoiding to waste light.

It would depend of how much light are you using, but reduction from 25 to 50% of the watts installed should be considered initially.

The new reduced number lamps has to cover all the grow space, so you will need to raise them. Now we wont need such strong intensities like before, we can use open reflectors, with wide light distribution pattern, and lamps way higher than most HID growers are used. See pics of very large grows and check what I want to mean.

With this procedure, we have achieved a way even lighting of the canopy using lower light densities, but that still are more than enough to produce huge buds on the upper canopy.

After than, we need to supplement the lower areas with LEDs. For this purpose, not so much light is required as with the upper areas.

I often think on cubic feets instead of sq ft. I let the upper cb ft for the HIDs (or slighty more, depending of the power of the lamps used and how crowded is the canopy) and reinforce the light of the bottom areas, at a rate about 15-20 uE/cb ft. For this purpose, we use mostly red LEDs, that are the cheaper ones, with a little help of blue ones (sometimes, white ones). Good red LEDs are emitting currently about 1.5uE/w when not running too hard, so it means about 10-15 watts for cubic feet. A single 600W out of the top alows us to supplement 40-60cb ft of bottom areas Usually, way less than that is required to match the previous yield.

Of course, we can end using same amount of watts, but producing a larger yield, due the way better light use due the better light distribution. NASA has obtained increases of yield up to 35% just by this process on plants similar to cannabis. Its more than 1/3 more harvest per watt

__________________


Bulb Analyzer Tool

knna

Although they are doing some of the better LED lamps avalaible currently, they are clearly not cost effective.

The 300W TI Smartlamp emits just 183 uE. Its a really low figure that, dont matter how good is the spectrum, cant outperform a 400W HID emitting 650uE (about 500uE if reflectorized). And, if it have to compete with a 250W without energy savings, how to recover the initial excess investment? 1500$ of gap price is still too large.

About the water cooling, Ive written about it on the heatsink's post. The water cooled TI lamp rely exclusively on the water cooling. It runs LEDs very hard (thus, with low efficiency) without any heatsinking. A pump failure or res water becoming too hot (it happen) and your expensive LED lamp is f***d.

Doing it DIY is really easy. You just need a copper pipe and attach LEDs to it. But I would recomend, still when using water cooling, to use some passive heatsinking on the opposite side of the copper pipe.

As the copper pipe is curved, I suggest to mount first LEDs on a narrow flat aluminium sheet (1cm wide) and then attach it to the copper pipe (with standard pieces for supporting pipes and thermal paste). The cost of the system would be just of the copper pipe, LEDs and driver. A 200W unit should cost about 400$ (500$ as much) using top end LEDs not running as hard as the TI lamp does (meaning more light emitted for same watts burned).

__________________


Bulb Analyzer Tool

scottcanada

Thank you KNNA for your valuable input. I would very much like to proceed with a water cooled DIY setup. Just so you know, my water supply will be coming from a tap in the washroom of an apartment. This way a pump will never break because the tap will always be on. Wasting water is not a concern right now as I feed 500 litres of water every 3 days so much of it could be recovered. The water temp will always be cold if I just run my cold tap constantly (please comment on ideal water temp is using copper pipe heatsink...is 10 - 15 C an okay range?). I need to run everything at 220 volt power if possible as my sub panel is all wired for my current 1000 watt HPS lights at 220. Given these conditions, if I want to build a test model that is likely to compete with a 1000 watt HPS then can you suggest which would be the best LED bulbs for this (given my temps, voltage supply, heatsink, 12 hours per day of operating, and specifically a focus on HPS/flowering light). There has been some debate about models like the SmartLamp and SmartBar because they are a "one light fits all" and not specifically designed to match either an metal halide or HPS spectrum. I am flowering cannabis plants where I would like penetration of about 3 feet, but I can grow sativas and train them to grow flat or use a SCROG method. Just need the same yield with less than 50% of power and willing to spend $1000 or so on a test model. Hoping I can use your guide and your input on a good design. Lets assume that we will NOT have a pump failure and the water temp will never go above 20 degrees, and may be as low as 10 in the winter. Water is never recycled back into the pipe so the heat transferred to the water never gets added to twice. Which LEDs would you suggest for this application given my power/temps, how many in total, and what proportion of red/orange/blue? What diameter of pipe and what length? My grow areas are 4x4 feet so I assume each grow area would be best to have 2 long tubes similar to the smart bar application. The HPS bulbs I currently use (and would like to mimic spectrum of) is EYE Hortilux Super HPS: EYE Hortilux Super HPS (or whatever would be best for flowering legal cannabis - I currently get best results with Hortilux bulbs but they cost $150 each and need to be replaced every 6 months). [mod note: edited] Thank you kindly in advance.

scottcanada

OK so I have been thinking about this more. The tubing I currently use is 1/2 inch, so I would probably want to use 1/2 inch copper pipe for my heat sink. Since I want good light distribution, and good contact of the aluminum to the pipe, how about I use a sheet of aluminum closer to 2 cm wide so I can wrap it much of the way around the pipe. The aluminum could be secured with straps after each LED so that it basically wraps around the pipe giving the best contact and heat transfer. Would I want to use a thermal adhesive or kapton tape here? I'm thinking the pipe should be about 3 feet long so the aluminum sheet would be about 3 feet x 1 inch (or about 1m x 2 cm).

next question: would it be possible to mount the LEDs directly to the copper pipe? you mentioned that a "thermal conductive adhesive" might work. Could you provide a link to an example of one of these products for sale online? Which would be better for my purpose - a long strip of aluminum wrapped around the pipe with LEDs mounted to it, or mounting the LEDs directly to the copper using a thermal adhesive?

What kind of drivers would I need? I would like to start looking up prices but didn't see any specific product names. One thing that would be really helpful would be a link to international sellers that can provide the products we would need like the specific LEDs and the specific drivers. You mentioned that you were making some drivers tailer to this application - do you have a website to buy from? How do I get the money to you and the product to me?

Lastly, when you suggest using passive heatsinking on the opposite side of the copper, do you mean some sort of insulation? Would it be a good idea to have this entirely surrounding the copper pipe everything except where the LEDs are actually mounted?

Please clarify what you mean by forced cooling versus heatsinking. Are they essentially the same thing? I assume the water cooled copper tube is a form of heatsink that could also be classified as forced cooling. Please let me know if am understanding the terms. Thank you for this wonderful insight. If you have any other articles or a website or company please post a link. Thanks

knna

Lots of questions, scottcanada, so Ill try to answer them one by one.

-220V? There is no problem with this. Its just a matter of choosing a Power Supply/driver designed for that line voltage. I usually work with universal voltage drivers, that accept both 120V and 220V and 50/60Hz. About the LEDs strings, there is no any difference at all. My line is 230V/50Hz and I have the electrical design optimized for this, but the difference is very little.

-Components suppliers. Finding the right ones is the toughest part when you want to build a DIY LED lamp. Once sourced the components, building it is really easy.

We have a problem with this because most LED's shops are more intended to flashlights hobbyist and such, that requires a low count of LEDs. So profit margins of this shops are abussive, of 300% on average. Aditionally, they arnt focused on the particular LEDs better suited for growing lamps.

Im building right now a online shop specialized on the best LEDs for growing lamps. Its already designed and now Im adding the contents, so I hope it will be working in two weeks as most. What I try to offer is something that currently is not avalaible anywhere: top bins of each LED model, with light emission well caracterized, both in lm and uE, so you know what you are getting (most LED shops says you the LED model on sale, but not its actual bining, which is esential). As for our application we need a high number of LEDs, profit margin may be kept very low.

As example, right now (price are dropping constantly) Im selling locally the Osram Golden Dragon Plus (up to 1A constant current) :
Red (639nm peak) at 1.25€ (~1.63$) (bin HZ, 39-45lm@400mA, 2.15V)
Red (639nm peak) at 1.5€ (1.95$) (bin JX, 45-51lm@400mA, 2.17V aprox 1.2uE@400mA (0.868W), 1.51uE@600mA (1.4W), measured at true operating conditions, not lab conditions) (it includes 16% taxes).

Coolwhite at 2.45€ (3.19$) (bin KZ (97-112lm@350mA) and 6Q (aprox 6200K)

Royal Blue at 3.45 (4.48$) min 350mW@350mA.

I have too top bins of Cree, but Ive seen people is having some problems mounting them manually, so I dont suggest them for DIY now. Same for Edixeon ones (interesting because they have 660nm and 730nm power LEDs). Thats why I dedicated the tutorial to easier to mount LEDs, as the Osram Golden Dragon, or Phillips K2

I dont lose selling at those prices, while same bins (on the little sites where you can get them) cost at least 2$ more each, 3-4$ on many sites. Ive put these prices of the colors most used for growing just you can compare with any other site. Prices are changing constantly, as with computers.

Apart of LED shops, an interesting option when you are going to buy a large amount of LEDs in USA and Canada are Digikey (for Osram components) and Future Electronics (for Phillips Lumileds). But you need to order a minimun of 500-800 pcs of each color to get access to the good prices.

Drivers are toughest to find yet. Easy to source are the Osram Quictronic for LEDs and the Phillips Xitanium. Prices of the simpler and lower power versions starting at 40$. Meanwell have a good range of drivers too, somewhat cheaper.

I sell high power LED drivers adjustable on current output (that are the problem of those cited: dimmable ones starts on 70$ each) for less than 30$. Im not trying to do spam here, its just Ive started to sell these components after seeing all LED shops gets abussive profit margins on them, doing very difficult to an individual to build a cost effective LED lamp. My only intention is to give access at decent price to these components, I wouldnt mind to link any other people doing so, if they exist. On my local forum there is another guy working on the LED industry that is selling them at good prices and he links me as I link him. I really would like more people sell them without ripping off uninformed costumers, as unfortunatelly, is very common on many LED shops (by selling 2 years old components at today prices; of course they didnt state the bin on sale).

-Thermal adhesives. They are used mostly for computers modders. The most known and easy to find is the 2 part Arctic Silver Adhesive (beware, what you find on most computer shops is the one part AS thermal paste, that is not an adhesive). Their thermal transference is very low compared to aluminium, so its a must to keep adhesive layer very thin, about 0.1mm thick.

Some people DIY their own thermal conductive adhesive, by adding small particles of silver or aluminium to a standard 2 part epoxy adhesive. But I dont reccomend that practice. Arctic Silver cost about 10$ and you can glue some hundred of LEDs with it.

Another choice I havent explained on the tutorial (at the end, there are lots of different ways of mount LEDs, and I prefer to explain the way Ive seen is easiest for plain people: electronic skilled ones sure that knows how to deal with it) is to use double adhesive sided copper tape. It adheres perfectly to kapton tape and then you can attach the LED on the adhesive and later solder it to the copper. Later Ill add a post explaining this way of mounting. But I dont know what is easier to source, if adhesive copper tape or adhesive pads. Anyway, I have both

(long post, Ill continue on next)

__________________


Bulb Analyzer Tool