What's new
  • ICMag with help from Phlizon, Landrace Warden and The Vault is running a NEW contest for Christmas! You can check it here. Prizes are: full spectrum led light, seeds & forum premium access. Come join in!

Building New Panels - Sharing the Process

habeeb

follow your heart
ICMag Donor
Veteran
^ I'm confused..

so your doing ( 4 ) 10 inch wide x 18 inch long bars?

so your putting 3 in a row length wise, and 1 vertical

so like:

--- I
--- I
--- I

the I is one bar, the --- , means one bar
 

rrog

Active member
Veteran
Doing four (4) 10" x 18" panels. Each panel = 300W

The four panels will be arranged in a 2x2 pattern. 1 panel per quadrant. So when all 4 are combined, they cover a 4' x 4' square.

75W per square foot. Rives likes the overhead this offers, so for flower I get penetration. I have 9' tall ceilings, so I can go up, up, up if I want.
 

habeeb

follow your heart
ICMag Donor
Veteran
^ I see, makes sense now..

I still feel 3 longer bars would be better, better thermal management, and a little more cushion for spread, and one less bar to have to keep raising

the thing I see is, having less surface area ( your bar ) the less you can space LED, and more you need to pull the bar to get the corners, making for less light the higher you go to reach the corners.

second, if you want 4 panels, why not just do ( 2 ) bars at 36 inch length, which is the same size as ( 4 ) bars, and then you have two less lights to pull up?



this is your light though man, you build what you want, I am throwing out options
 

rrog

Active member
Veteran
I like the idea of 4 panels for a 4x4 footprint. That's 5 square feet of panel with 1200W covering 16 square feet of grow. For veg I can use 1 module. Then 2. Then 4 in flower. I can angle them, stagger them. Seems powerful and flexible. The 4 panels would be spaced apart and not contact each other.

I'm wondering if I buy these beautiful heat sinks if I can avoid drilling with these: http://www.rapidled.com/thermal-pads/. Sure would be nice to keep that smooth surface smooth.

That section of heat sink that Rives pointed out has 2160 square inches of exposed fin in a 10" x 12" hunk. 2.5" tall fins to cool 300W. And I guess we need to say that these are 10W not 2W so less efficient and even more heat. I could add fans.
 

PetFlora

Well-known member
ICMag Donor
Veteran
<b><h2 class="title icon">Idea to MAX Out nm Spectrums, Thereby Supercharging Growth

Know anything about Merkaba shape called Cosmic Energy?

Notice the shadow image the leds created in the center of the image below, taken by Guod (on another forum), who , due to language barrier, I'm not certain of his opinion, but he only did a single cluster. I'm thinking, a larger panel, made up of multiple clusters would pay dividends


2435590d1355054880t-diy-modular-led-light-mix1.jpg


To me, seeing them cast an actual sacred geometry image is very significant.

Proof to me, that there must be something to Sacred Geometry's relationship to pure energy.
Light is an expression of energy, and life itself.

I'm thinking how to enhance it. Perhaps a small power crystal (amplifier) in the center of each cluster... then,
duplicate the clusters, from the center out (encircling the first cluster, and so on). Panels can be as large as you want. Personally, I would follow Astir Grows panel size, and make 3+ small panels at least 3 ring wide

This may be the ultimate UFO/shape, whose cluster's cumulative energy effects might energetically super charge anything grown under it.

It certainly won't hurt, so why not? The remaining question is, which nm diodes, and where to position them to further amplify the energy

Anybody willing to test this? Pretty PLEASE
kiss.gif
I promise to only say nice things about you (especially if I get to Beta test)
smoking.gif
, like... Pioneer

</b></h2>
 

rrog

Active member
Veteran
PetFlora- this panel array that Rives originally turned me on to seems like your multi-cluster array
 

rives

Inveterate Tinkerer
Mentor
ICMag Donor
Veteran
You would have to have fans at those power levels, Rrog. Running it through the spreadsheet, a 10" x 18" would have be able to passively dissipate 110-130 watts (ideal vs acceptable). The mcpcb stars should be placed no closer than 1" apart, and wiring them at that distance might be a bit of a trick.
 

rrog

Active member
Veteran
Rives- you think I'm pushing it? Can I wire, then secure with the pads, or should I change the panel parameters (size or Watts)
 

rives

Inveterate Tinkerer
Mentor
ICMag Donor
Veteran
Rives- you think I'm pushing it? Can I wire, then secure with the pads, or should I change the panel parameters (size or Watts)

No, I think that it is easily doable. Yes, I would use a template that mimics the layout of the stars when attached to the sink and do the soldering, then attach the stars to the sink. It is very difficult to get the solder to flow properly when they are attached to the sink, particularly one of that size.
 

rrog

Active member
Veteran
OK. So I shouldn't push to a 22" sink or anything? Now's the time I suppose.

I'm trying to keep true to your 75W psf goal
 

rives

Inveterate Tinkerer
Mentor
ICMag Donor
Veteran
The bigger the sink, the more effective that it is going to be. You are going to have to use fans anyhow to get anywhere near the power levels that you want, but at least for me, it would be nice to avoid such high air flows that the fixture sounds like one of those wall-mount hand driers. My Lumigrow is so loud that I don't like using it. The hybrid fixture is much, much quieter even with a total of seven fans running (2 on each sink, one pushing and one pulling, and one in the control enclosure). The Volks is inaudible with a Noctua fan. I would suggest using multiple fans, so if one dies you still have some airflow until you see that you have a problem.
 

rives

Inveterate Tinkerer
Mentor
ICMag Donor
Veteran
Wouldn't umols^2 ft be a more useful measurement?

For the lighting it would be, but you would have to have the components pretty well selected to get there. For trying to determine the heatsink size, the drivers, etc, umols wouldn't be of much help.


Rrog, those seem to be very popular with a lot of people. I believe that the subsequent generation has just become commercially available also. You might want to check out these, too - http://www.philipslumileds.com/uploads/366/DS102-pdf. Futurelightingsolutions.com has some pretty good pricing on them but only sells them in tray quantities, which might work out well for you. http://www.futurelightingsolutions....w|1|1|,Ny:True,Ro:0,Nea:True,N:916-4294878041
 

rrog

Active member
Veteran
I'm ready to select components. I believe I have the LED assortment and ratio.

EDIT: As far as LED's I'm not qualified to judge what's better, but the Cree is popular, so maybe I'll try them.

So now that I've settled on the Cree XML, I'm on to drivers.

Looking at guvoo's power supply suggestion. Is this a good option for a newb like me?

Main Power-supply
NES-350-48
48V - 7.3A - 350W
http://www.meanwell.com/search/nes-350/default.htm

Led-Drivers
for every channel one of this
http://www.recom-lighting.com/upload.../RCD-48_02.pdf

So the power supply would provide a constant current, protecting the LEDs. I can then hook up 5 LED channels to one of those drivers as follows:

6 WW on dimmer

6 WW on constant current

6 NW on cc

6 NW on cc

6 CW on dimmer

The drivers have a 70W max, so in this case I'm running 60W each. Am I doing this right?
 
Last edited:

guvoo

Member
the recom RCD-48 are only going for 1200mA max.
the Cree XML at 10W needs more than this(up to 3000mA), so no way with this Idea.
 

rrog

Active member
Veteran
OK thanks for clarifying this for me guvoo. What is the process to determine the drivers at this point?
 

rives

Inveterate Tinkerer
Mentor
ICMag Donor
Veteran
the recom RCD-48 are only going for 1200mA max.
the Cree XML at 10W needs more than this(up to 3000mA), so no way with this Idea.

I was wondering where that part number/link for the 2 amp version came from, the link didn't work for me and I couldn't find it in any of the documentation that I was rooting through.

Driving the XML's at a true 10 watts looks like it is right at the outer reaches of the chip to me. In the spec sheet, it says that Cree recommends a 2000ma current maximum for best chip life. This would put you at about 6.5 watts per chip. It would be good to get some input from some people that are running these.

OK thanks for clarifying this for me guvoo. What is the process to determine the drivers at this point?
For individual drivers, you need to get a head count on the number of chips in each string and multiply that number by the forward voltage at the current that you want to run. Looking at the curves, it appears that the forward voltage at 2000ma is approximately 3.2 volts. So, if you want to run 10 diodes in a string, you would multiply (10 leds) x (3.2v) = 32 volts. You would then find a driver that has a minimum voltage a couple of volts below that, say 30v, and a maximum voltage that is a couple of volts higher, say 34v, and that will source the current level that you want (2000ma). So you would want a driver that had a minimum voltage output of 30v or less, a maximum voltage output of 34v or more, a current level of 2000ma, and a wattage output of at least 65 watts.
 
Last edited:

rrog

Active member
Veteran
Well if Cree says to run at 2000 mA current, then I'll just go ahead and say fine so I can keep this plan moving.

So 2000 mA gives me 6.5 watts per chip. So each driver needs to feed 39 watts? Assuming 6 LEDs per driver I guess.

Wait... this means my 300W panel just became a 195W panel because I dropped the current to 2000mA?
 

rives

Inveterate Tinkerer
Mentor
ICMag Donor
Veteran
Refresh your page, I think that I was still editing my last post when you posted.

Yes, now you need to add more LEDs back if you want to keep that same wattage.
 
Top