LED Replacement for T5 - Help with DIY

Ok you tech guys, my funds are tight right this second but I would like to start gathering input on what I think would be a good first led project.
I have this shelf where I have been starting my clones/seedlings which is 12"D x 18"H x 36" L
I currently have 3 of the 2ft Sunblaster T5's but would like something that lights the full length of this shelf and produces a bit more squat/tighter node spaced seedlings.

What ideas do you guys have to throw in. COB, Epiwhatevers? Looking for ideas on something efficient and relatively economical.

That might be a good one to stay with. Just saying

Sent from my SM-G900V using Rollitup mobile app

Maybe do something like this:

https://www.rollitup.org/t/diy-70-watt-hundred-dollar-veg-lamp.870199/#post-11576271

Your lamp is 75 watts right?

Vero 18 at 4000K is 145 LPW and 20 watts at .7A

3 of those would give you a 60 watt lamp putting off more light than your current setup. Twice the light actually. Plants could get as close as 2-3- inches. Since it's passively cooled it would sit at the same height and have the same profile as your current lamp.

For driver substitute with one HLG-60H-C700. That would give all the parts 5 year warranties. Price would be under $150

Rahz said:

Maybe do something like this:

https://www.rollitup.org/t/diy-70-watt-hundred-dollar-veg-lamp.870199/#post-11576271

Your lamp is 75 watts right?

Vero 18 at 4000K is 145 LPW and 20 watts at .7A

3 of those would give you a 60 watt lamp putting off more light than your current setup. Twice the light actually. Plants could get as close as 2-3- inches. Since it's passively cooled it would sit at the same height and have the same profile as your current lamp.

For driver substitute with one HLG-60H-C700. That would give all the parts 5 year warranties. Price would be under $150

Click to expand...

Thanks, this is what I was looking for.
So all three or the Cob's can run on that single driver?
What source do you recommend for the led and driver?

All three on that driver in series.

Ebay for the heatsinks
Digikey for the LED. Grab some Pico EZmate connectors if you don't want to solder.
I've bought Meanwell drivers from Powergatellc.com and Arrow, but Mouser and a couple other places have them.

Rahz said:

All three on that driver in series.

Ebay for the heatsinks
Digikey for the LED. Grab some Pico EZmate connectors if you don't want to solder.
I've bought Meanwell drivers from Powergatellc.com and Arrow, but Mouser and a couple other places have them.

Click to expand...

Thanks for taking the time to help on this Rahz!
I need to go post hunting and find the one about pairing led's to drivers so I can get a bit better grasp on all of that
Drilling, soldering, etc. is no problem here just need to learn the parts compatibility and how to read all the led specs, bins, etc

Would something like these work as well mounted to a long heat sink?

http://www.rapidled.com/cree-xp-g2-5w-neutral-white-led-with-40-60-or-80-degree-lens/

Rahz, would it be acceptable to mount the three Vero 18's on one of these 36" heatsinks?

http://www.ebay.com/itm/LED-DIY-Light-Aluminum-Heatsink-12-24-36-48-/270831824949?pt=LH_DefaultDomain_0&var=&hash=item3f0ed5b035

I think that would probably work fine. Check out heatsinkusa for more options.

Without having a quantum sensor, is there a way to determine the led requirement and arrangement to supply 250 μmol/m2/s at a distance of say 4-6" above the canopy for an area 1'x3'?

If you know the umol (there's a chart floating around somewhere),

watts * efficiency / sqft * 10.7 * umol/J

Rahz said:

If you know the umol (there's a chart floating around somewhere),

watts * efficiency / sqft * 10.7 * umol/J

Click to expand...

This doesn't seem correct.

OneHitDone said:

Without having a quantum sensor, is there a way to determine the led requirement and arrangement to supply 250 μmol/m2/s at a distance of say 4-6" above the canopy for an area 1'x3'?

Click to expand...

You can get a light with the output of ~70 µmol/s which would make an average of 250 µmol/m^2/s.

How does one attain the µmol/s output of different light sources without a quantum sensor?
Say a 2ft T5?

alesh said:

This doesn't seem correct.

Click to expand...

If the formula isn't correct I would very much like to see it corrected

OneHitDone said:

How does one attain the µmol/s output of different light sources without a quantum sensor?
Say a 2ft T5?

Click to expand...

http://rollitup.org/t/math-behind.868988/

Rahz said:

If the formula isn't correct I would very much like to see it corrected

Click to expand...

dissipation [W] * efficiency * QER [µmol/J]
-----------------------------------------------------
area [m^2]


380-780 nm range


400-700nm range

Vero 10's, the last version at test [350mA] put out about 16.5 umol/s each....it varies 3000k is 16.43, and 4000k is 16.65 and 5000k is 17.01 umol/s......
But these are the old models unfortunately.....and the new Vero line have increased photon output 2-4%?

Mrflux
http://rollitup.org/t/cree-cxa-analysis.743645/#post-9761101


Personally I think you could get away with 2 Vero's/per sq. ft....driven anywhere from 200-300mA's....

In US we are working with canopies sized in ft², so we have to convert to m² to get PPFD. To do that we multiply by 10.7. So as a formula it would look like:

dissipation [W] * % efficiency * QER [umol/s/PAR W] * 10.7
--------------------------------------------------------------------------
area [ft²]


So for the OPs example, canopy is 1'X3' so if you used CXA3070 @ 1.4A X 2 = 104W dissipation @ 43.3%:


104W * .433 efficiency * 4.66umol/s/PAR W *10.7 = 2244
2244/3ft² = 748 PPFD averaged

or

104W * .433 efficiency = 45 PAR W (minus reflector/lens losses)
45 PAR W/ 3ft² = 15 PAR W/ft²
15 PAR W * 4.66 umol/s/PAR W = 70umol/s/ft²
70 *10.7 = 749 PPFD averaged

or

We could use Aleshs' formula if we convert from ft² to m² first:
3ft²/10.7 = .28m²
104W * .433 efficiency = 45 PAR W (minus reflector lens losses)
45 PAR W * 4.66umol/s/PAR W = 210umol/s
210/.28 = 750 PPFD averaged

OneHitDone said:

Rahz, would it be acceptable to mount the three Vero 18's on one of these 36" heatsinks?

http://www.ebay.com/itm/LED-DIY-Light-Aluminum-Heatsink-12-24-36-48-/270831824949?pt=LH_DefaultDomain_0&var=&hash=item3f0ed5b035

Click to expand...

I would order from heatainkUSA. Same price as above maybe cheaper. Don't skimp on the heatsink, if you get some curved, warped one from eBay you will notice the difference and be disappointed
Take a look here
http://www.heatsinkusa.com/4-230-wide-extruded-aluminum-heatsink/

Notice the eBay seller only leaves out one dimesnsion ....base thickness, mucho important, Buyer beware!

OneHitDone said:

How does one attain the µmol/s output of different light sources without a quantum sensor?
Say a 2ft T5?

Click to expand...

-The best fluoros are about 30% efficient, the SPD is very spikey/poorly distributed and maybe worst of all, they are very heavily dependent on their reflector, even when close to the canopy.
GE 5000K fluoro:


So those characteristics are horticulturally unfriendly. That said, if you already have them might as well use them (or sell them) but if you are considering investing in them, Veros run hard or even generic COBs might actually be cheaper, especially considering that quality electronic ballasts for fluoros are probably the same price as quality LED drivers. Cheap fluoro ballasts lead to increased bulb failure rates and lower system efficiency (more heat).

So if you have an 90lm/W fluoro 24" T5 (24W bulb dissipation) and a 290lm/W LER (LER varies a lot with fluoro) that is 31% efficient. 24W * .25 efficiency = 7.44 PAR W. Assuming 40% reflector loss, 4.5 PAR W exit the fixture. Assuming 4.25 umol/s/PAR W that is 4.5 * 4.25 = 19 umol/s. So a 24" fluoro covering a 24X6" space (1ft²) 205 PPFD. That is good intensity for vegging.

Smash the T5 and use 47% efficient baby killers. (total of 40W)