DIY initial thoughts

So having spent many hours pouring over the DIY threads I have got the bug.

I am so glad I didn't rush off and buy an LED unit (though I know its probably cheaper that way) I love soldering and fixing electrical stuff and this kind of diy is right up my alley though the technical aspect behind it may well be beyond my understanding.

I would like to try an experiment and see if I can build a panel or two or three depending on 1w/3w/5w/
to do one plant only.

I'd like to use the following spectrum in the hope that it would do (ok) for veg and flower

3-NW 8-WW 1-CW 3-RED(630) 1-Blue(420)

The problem at present is deciding power/watts I need for one plant in a 2ft by 2ft area?

I am guessing about 80 real watts if I do 2 panels of 16x3 watt leds.?
That's 128 dollars in LEDs already!
while I know its a pricey game at present are there any other good sellers apart from ledgroupbuy and rapidled?

Anyway before I get to deep into it in first post, at present I am asking:

1:Suggested watts for 2x2 area one plant?
2:Would you be inclined to go 1w/3w/5w? (maybe even the cree xm-l 10w!!!) ok maybe not
3: Spectrum looks fine? Will the UV be an issue in veg? If doing 2 panels was considering an Ir on the second panel

[FONT=Verdana, Helvetica, Arial, sans-serif]Since you are diying and not just looking to purchase an already made fixture:

Wattage of chips - I wouldn't use 1 watters unless you have a very limited grow area height or just want to grow very short plants. 3 watters can be kept pretty close too and will have decent penetration for a plant under 3 foot. Higher wattage chips will be able to penetrate further and grow taller plants. Of course, side lighting could provide a similar effect.

Your guess of 80 real watts seems off - 16 x 3 = 48 x 2 panels = 96 watts total however, I'd guess that you will be driving them about 60% so 96 (.6) = 57.6 so lets round to 58 watts for the two panels. You'll have to make the choice at what current you will run them but 80/96 = .833 and I doubt you will be driving them at that high..but maybe you are.. It can become difficult to provide adequate heat management for LEDs being driven hard. Heat will affect spectrum and longevity. Do NOT cheap on the heatsinks.

Recommended watts: Varies with quality of fixture, chips, coverage, etc but for general purposes let's say 35 watts/ft2 so 4 x 35 = 140 watts. If you are using high quality chips, drivers, good heat management (and I'm going to guess that you are since you mention ledgroupbuy) you can get away with 30 watts/ft2 so you need 4 x 30 = 120 watts If you are more interested in efficiency than absolute quality then maybe 25 watts/ft2 with 3W LEDs IF you have the coverage.

Spectrum - definitely not a nailed down thing - look at the spectrum provided for your brand/bin - I would probably let the WW chips provide the 630 and add some 660nm chips...although a good WW has 660nm in it too. The CW may provide enough blue for you. Since you are buying quality brand/bin LEDs you can get some very good info from the spec sheets to make your design decision.

Coverage - LEDs grow well what is directly under them and a few inches to the side/end - maybe 2 - 6. It does drop off quickly on the sides/ends. Make the fixture to fit your space whether one or two. I would be looking to cover about 22 x 22 to maybe as small as 20 x 20 with one or two panels or you could do a series of strip lights.. There are many ways to skin a cat - just keep the coverage in mind.

other good sellers are stevesleds.com Philipps luxeon es chips are quality..especially for deep red and royal blue...but generally high quality like Cree.

Some drivers are dimmable and are easily dimmed with just the addition of a pot - some require more work than that so look at it closely. This can help you adjust your spectrum for maximum efficiency for various stages of growth...or not....it does cost so it's your decision whether you want that flexibility.

If heat is a major driver moving you to LEDs, then be sure to leave long leads for your driver/power supply so you can put it/them outside your grow. As you might imagine they are definitely a source of heat in a grow area.

[/FONT]

Check out my thread. Lots of good diy led 411

PetFlora said:

Check out my thread. Lots of good diy led 411

Click to expand...

I think he is looking to build his own LED panel Pet....

Good thread though.

trunche0n said:

I'd like to use the following spectrum in the hope that it would do (ok) for veg and flower

3-NW 8-WW 1-CW 3-RED(630) 1-Blue(420)

Click to expand...

too much blue, ditch the blue diode and add some 660nm deep red.

The problem at present is deciding power/watts I need for one plant in a 2ft by 2ft area?

I am guessing about 80 real watts if I do 2 panels of 16x3 watt leds.?

Click to expand...

There are doods finishing single plants under 28 watts on here. I would do 2 plants under 80 watts personally. Check out HiloReign's thread.

That's 128 dollars in LEDs already!
while I know its a pricey game at present are there any other good sellers apart from ledgroupbuy and rapidled?

Click to expand...

Yep.

www.cutter.com.au

www.mouser.com

www.farnell.com

www.rs-online.com

Remember, DATA SHEETS ARE YOUR FRIEND!!

1:Suggested watts for 2x2 area one plant?

Click to expand...

40 if you're a ninja

2:Would you be inclined to go 1w/3w/5w? (maybe even the cree xm-l 10w!!!) ok maybe not

Click to expand...

Nothing wrong with XML, XT-E is nicer though IMHO. ~3W chips are good but remember this is a ballpark figure - the wattage will depend on how hard you drive them (amperage otherwise known as current). Remember Power(in watts) = Voltage(V) X Current(A). Voltage tends to go up the harder you drive them. Data sheets will have a current vs voltage graph. You really must learn to understand data sheets as they will be your bread and butter.

3: Spectrum looks fine? Will the UV be an issue in veg? If doing 2 panels was considering an Ir on the second panel

Click to expand...

As above, you need way more deep red, plenty of red, and some blue (you already have plenty if not too much). The whites are a good choice IF YOU BUY BROAD SPECTRUM'D WHITES. Another reason to look at data sheets - there's a graph for this.

Forget the idea of a veg/flower panel. Its a myth that for some fucking reason people cannot shake. You don't need a seperate panel for veg.

IR = infrared = Heat. Nothing special, just heat the same from putting your hand above a flame. I'm assuming you mean far red at around 730nm (the bit of the spectrum just before Near-infrared, which is in turn next to infrared). There are all kinds of processes associated with FR and you should familiarise yourself with them before even thinking of putting them on your panel. For a start look up the phytochrome cycle http://en.wikipedia.org/wiki/Phytochrome

I think i covered everything you ask. For a hands-on guide to actually assembling a neat little panel check bassclef's slightly dated but still absolutely awesome guide. https://www.rollitup.org/led-other-lighting/580314-super-easy-diy-led-micro.html

Thanks a million for the detailed replies! Cant beleive everyones so damn helpful. Thought i'd get ignored for a basic info thread. cheers

The spectrum is what is beating me at the moment - I read (and apparently didnt understand!) alot of threads here and thought my ration would be fine.

Is there at least a target spectrum of light to achieve?

maybe that way I couldsort something better out ratio wise.

I can read the data sheets ok but its knowing how much is needed of each temp is getting me ...

Go and look around this subforum at peoples grow journals - look at what light they are using and the spectrum and make up your mind.But here's a "classic" chlorophyll graph, for what its worth: http://www.google.co.uk/imgres?imgurl=http://www.sd84.k12.id.us/shs/departments/science/yost/Biology/5%20-%20Energy%20&%20Metabolism/chlorophyll_absorption_graph.jpg&imgrefurl=http://www.sd84.k12.id.us/SHS/Departments/Science/Morgan/Biology/Unit%205-%20Energy%20&%20Cell%20Resp/graph%20results.htm&h=279&w=563&sz=19&tbnid=sQ5xEPKjarhZBM:&tbnh=59&tbnw=120&zoom=1&usg=__Ot_dmAiuum5KkSPWIwIm22grFTo=&docid=9DK0U5Z_-CAvPM&sa=X&ei=wFtkUeOxLYLH0QXJhYHgBg&ved=0CDgQ9QEwAQ&dur=212

Looking at that it would seem the best would be lots of blue right on the peaks, and lots of 660nm red where theres a massive peak. however, we know from experience that too much blue has detrimental effects, in part because of the energy carried by blue photons (to my understanding) damaging chlorophyll. To really really oversimplify, here are some thoughts:

- The general concensus is that is that while looking at a chlorophyll graph and just matching your LED with the highest peaks is helpful to an extent, the best and most healthy growth is achieved with a wide spectral coverage. This is especially true with lots of power, when the area is saturated with light.

- you are balancing spectral preference with efficiency of the LEDs.This is best illustrated by the move towards warm white LEDs - they have the benefit of a wide band of different colors (to make white), often with a nice smooth peak in the red range, and have great efficiency - MUCH greater than if you were to try and replicate their wide spectral coverage with single-colour LEDs (which normally only cover a ~5nm spectral band).

- But, we are balancing here with the strengths of LEDs. Deep red 660nm for example, is not particularly visible to the human eye compared to blue or green for example, so it is not really seen in the otherwise-awesome white LED. This is one situation where supplement with single colour LED is very appropriate.


Personally I aim for around 10-15% blue (including the blue given out by the whites).
Again, personally, I like to have around 40% at deep red coz weed loves it.
The rest is made up broadly around the Red (630nm~) part of the spectrum through warm whites.

when an ledis rated 6000k to 6500k how the hell does one guarantee it is either, or and in between?
Is this where "bins" come into the equation?
sorry, i am learning.....

edit - nevermind the Q. i get the whole bins thing now...

ok so i joked about it.... but I don't know.... I am considering the xml's.
- they look from my rather extensively limited knowledge to have quite a broad color spectrum.
I can buy 18x10w for 150 dollars. I can sort of justify that cash wise although I know they are less efficient/hot -

In respect to each other of the three whites ;
the Cool White only seems to offer a bump over the WW and NW at 440 which is the chlorophyll A peak
the Neutral White only offers a limited spectrum of 450 - 525 where it is stronger than the WW/CW but that's kinda chlorophyll B territory so still very useful.
The Warm White takes over and has the most power from 525 up, it looks to cover the chlorophyll B 620 peak but because I have no idea how to interpret the data if its not obvious I cant tell does it have enough juice in the 650 area?

cree xml data sheet

As i said above, 660nm supplement is good. As you rightly interpret, there is not much ~660nm in those WW's.

Remember 2 things:

the harder you drive LED the less efficient they are. Those XML might be rated for something crazy but practically its probably not smart to actually run them that hard.

The heatsink is gonna be probably a bigger expense than the LEDs. You found one to cool 150 watts yet?

Looking at the data sheet I figure 1.8a-2a could be nice sweet spot. I will definitely be going ghetto on the heatsink assembly. I am thinking of using a sheet of steel to mount all LEDs to the underside. On top I will be attaching CPU copper heatsinks as many needed. I have to figure the ratio first then the amount then I can figure the square inch of heatsink needed.
If I were to go 1:1:1 cw.nw.ww 10w. How many 3w would supplement?

Sorry that was meant to read how many 3w red 660 to each 3 10w (ww. Nw. Cw.) ?

Do keep in mind that if your LEDs aren't properly heat sinked that you will affect the spectrum and the longevity that you paid for. Steel isn't a good heat sink. Do you have a metal recycler nearby? Maybe you can get a cheap piece of aluminum....if not buy less XML LED...and pay for the sheet of aluminum. Here's a link: http://www.engineeringtoolbox.com/thermal-conductivity-d_429.html Or maybe you have access to aluminum tubing....look at stevesleds and how he is using them for heat sinking.
I would lean the ratio of whites away from CW and toward WW. It doesn't take a lot of blue. As I said earlier, spectrum isn't nailed down; people have different views yet on what is best. Consider putting the 660s on a dimmable driver. You need a separate driver for the 660s from the XMls because of the current differences anyway so leverage that with dimmability for the 660s and you can play with the spectrum over time...the needs change from veg to late flower anyway.

Thanks. Aluminium isn't a problem- if it is I'll mount the LEDsdirect tto the CPU coolers or use tubing. Yes driver is an issue. I think the cheapest I can do it without cutting corners is to go with an inventronics driver that I know can run 16 XML chips at 1800ma. Then a little meanwell driver for what other 3w spectrums are needed at 700ma. There's capacity there for 16 LEDs again in that little driver.
So bearing in mind you said to stear clear a bit of so much the cool white I suppose 3cw 8ww 5nw in the 10watt XML on the inventronics driver. Does that seem OK or could I stand to loose another cw in favour of another.
As for the 3w xte i dont think there is cool for 16 x 660nm so maybe I could mix it up more but I am already far into guessing land as it is.....
I know it would be easier to go with a cidly but I do love to solder!

My honest opinions on what you said so far:

1. you are thinking wayyyyy to little about cooling, esp. since junction temperature (between the LED and the heatsink) directly affects the brightness of your LEDs. The hotter your LED are the shitter they will run. While you haven't described your cooling arrangements in detail, what you have said sounds totally inadequate. Check out SDS' DIY LED (making-of thread) (trial thread) for an idea of what the epitome of cooling looks like (highly thermally conductive copper PCB, over-size heatsink, manifold, active cooling with backup and an emergency shut-off in case of fan failure, etc....this is the "extreme" i guess, but junction temperatures are really the key to having a pimp-ass baller setup instead of some shitty overpriced christmas lights.

2. The XML are overpowered. ESPECIALLY in conjunction with #1. I'm trying to avoid just being like "use this and that LED" but it is a jungle out there, i know, so i'm going to give you my recommendation. they are just that, my own opinion, and as JCM said, nothing is nailed down in the LED game yet. There's also absolutely notihng wrong with XML, i just don't think theyre the most suited for your situation)

3. I would use Cree XT-E Warm whites (you should be able to get R2 or R3 bin pretty easy) for your WW LED. They run at more "normal" currents (600-1100ma) which also opens up potentially having your deep reds on the same driver.

I didn't see which deep reds you were discussing earlier. People here really like Luxeon Rebels, though personally i have never used them. I use Ledengin LZ-1s driven at 1050ma in conjunction with my XT-E (driven at 950ma).

4. In regard to your distribution of whites at the moment, I would definitely lose 2 of your NW for more WW or even extra 630nm Red supplement (Cree warm whites like XML and XTE tend to peak a little before 630nm - around 605-610nm is their peak). personally i'd lose almost all the NW and CW on the logic that they are all putting out the same wavelength blue. maybe go 8 WW, 2 CW, or - 5 WW, 2 CW, 3 630nm Red. This is definitely not like "DO THIS SPECTRUM MAGGOT" coz i am just trying to illustrate the small proportion of blue you actually need (The above doesn't even take into account Deep reds, which i personally have a propensity for compared to most growers - i put A LOT in the mix.)

Sincere thanks for the tough love, sound advice and great links! you guys astound me.
I appreciate my lack of knowledge together with my presumptions may be a bit trying so thanks for hanging in there!

As for cooling - Sorry I have totally glossed over it - But to me cooling isn't a big issue in the sense I know it has to be done and it costs...
I currently used a Watercooled PC - a few years before I thermo-electrically cooled my cpu's with peltiers so I am used to mucking about with different cooling methods.
When I started overclocking/advanced cooling years ago heatpipes were only in their infancy regarding their application to VGA / Ram /MOSFET and CPU coolers but now they are common place and the technology has come on a lot.

Once I have figured my ratio of led's I can see how many watts I need to dissapate... Then I will put forward my ghetto yet useful cooling idea with more facts -
As until I know my watts and heat output there's no point deciding on what amount to use and how many boards to make.

It seems a shame to ditch the xml idea but maybe seeing as I am learning I would be better of not pioneering and veer more towards the tried and tested as I dont have the knowledge to really hash out the data on the spectrums - not to mention the different penetration using some 3w and some 10w might be a problem.

Anyway, time for some further reading and investigate the 3w option

Take a look at stevesleds.com again. He has drivers of various sizes, prices and amperages which would drive XTEs, XMLs, or Luxeon ES. I'm using his drivers on my small diy fixtures and so far so good. They require a separate power supply and are easily dimmable with an inexpensive pot.

As far as LEDs, I also quite like the spectrum of the XMLs and went with them for my diy strips. Basically, I have 14 XMLs - mix of WW and CW - on about 60 total inches of aluminum tubing with one of steve's pressurized fans. I'm driving my strips with a 2100ma inventronics driver. The fixtures run very cool. I'd give the edge to Cree on whites and the edge to Philipps on deep red and blue but generally both are quality. The XTEs are definitely an alternative and attractive option as well.

I also agree with Snot on losing the NWs. Now that I've said that, some people do really like them. 8WW and 2CW sounds reasonable to me. I would probably emphasize the 660nm (deep red) over additional 630 as the XML WWs have 630...but additional 630 is good too. You could run the xpe red 630s and the 660nm philips luxeon es at the same current and therefore on the same driver. Steve's single driver at 700ma will drive about 10 of them. Dimmability will allow you to customize your spectrum for your grow. I would definitely like to have it on the 660 if I were you...either that or a switch to turn it off. You don't need 660 during veg but during flower, it makes a difference. I'ts even nice to be able to dim the white fixtures as you don't need as much light during veg so you can save a few watts for a couple of weeks If there's a magic spectrum, I don't know it yet but I guess I'll keep working on it. My little diy fixtures for a spinner are aimed more at the extremes - blue for veg and deep red for flower - so I add a color that I can dim and control as I think that it's good for the grow. Guess given some time, I'll get it fairly dialed in. Ya' know if you don't have the money for the separate driver and 660nm chips, then just go with the WW, CW mix. Since you are diying, you could always leave some space for the chips on your heatsink which you can add later or add a separate small strip of red/deep red when more money drops from the sky. Design choices are really rather fun to contemplate as there are so many ways to approach any "problem".

Oh, and if you stick to this kinda' basic quality we've been discussing and heat sink well with good coverage of your grow area, you WILL make a fixture FAR superior to a cidly. Slam Dunk.

Oooooo kay. phew - I think if I read anymore my head will explode - I will go with 3w leds..... mostly for consistency of using same wattage for all spectrums

I would like to have 48 leds on at any time and a total of 64 leds.

I will use three drivers each running 16 leds at 700ma (48 leds covered)

The plan is to actually have 64 leds on the panel (or two panels of 32 not sure yet) and in both veg and flower there will be a bank of 16 used to boost the ratio for the spectrum need at the time.

So in saying that here is my proposal for red orientated flower mode and more blue/6000k orienated veg mode. I have ommitted the use of NW as it seems redundant for growing IF using xt-e crees why not just use the WW as it covers the NW and some. (as you all suggested previously but at least now I understand why as I understand the spectral power distribution charts now... ) I have also omitted the use of blue leds in the hope the CW does the job. There is no problem in adding a few if you think it would be advisable.

So the whole point of this plan is that out of 48 LEDs I have 16 of which I can use to sway the spectrum for flowering or veg.

45% Warm White xt-e - 25% Red 660 Luxeon - 17.5% Red 630 - 12.5% Cold White xt-e ( In my Flowering mode led no. = 22WW : 12(660) : 8(630) : 6CW )

Now the plan is to adapt that ratio to suit for veg by only changing 16 leds - as this way its just a case of pluging my ldd driver into a different bank of leds - no soldering or mucjing about - so to push the 6000k cold white and remove as much useless red in veg I have come up with the following:

45% Warm White xt-e - 5% Red 630 - 50% Cold White xt-e ( In my vegging mode led no. = 22WW : 2(630) : 24CW )

is this looking a bit more realistic a plan to my previous thoughts or have I gone a little crazy on the adaptation for veg mode using too many CW?
On reflection I repeated myself a few times in this post - sorry!

NOW YOURE GETTING IT!

This is starting to look like a project you have put some thought and cleverness into.

The advice below is with the caveat that personally i don't switch spectrums for veg/flower. However, lots of people here that i respect swear by it, so my mind is not entirely closed to the idea.

My thoughts on your current design would be as you said, a little much CW on the "veg mode", don't underestimate the importance of red in veg. Additionally, i like to have All spectrums that will be used in flower to be present in veg - My theory is that this allows the plant to adapt its' chlorophyll production to these wavelengths.

With that said i would re-shuffle to have some 660nm deep red present in "veg mode", possibly a tiny bit more 630nm red.

edit: also you could connect your driver to a simple 3-way switch and not have to unplug stuff. I like the idea of switching the driver around rather than just buying extras and turning them on and off, though.