Do you have a link to the screw-in COB you're using? How powerful? How distant do you have to keep it from the plant?
I'd like a link to whatever is being recommended, at least an example. The 2-3 I looked at in the past were down around 80 LPW. That was less than T5HO.
I use
thanks for the info about the lights. i grow my plant on the floor indoors at night and during inclement weather. otherwise, they're outside. after flowering begins, they stay outside.
That page says 9.5w & 450 lumens. That's only 47 LPW. CFLs are in the 60s. Cree PAR38 (Home Depot) spot or flood is 84 LPW. The A19 lightbulb replacement is low 90s with the gummy rubber film removed removed from the translucent globe. SMD 5730 corncob lightbulbs are 105'ish (clear plastic globe).
"Chip on board" and "lumens per watt." COBs appear to be the next generation. Higher LPW, lower labor costs to assemble because it's one chip with dozens of diodes on it. Not individual diodes mounted on a large white backboard. (Downside to me is that it's more light from a smaller point. I lean towards less light from more sources. They're probably great for large grow operations. For a handful of plants like I do, I like having the lights closer, spread around the plant, more coverage and diffusion.).
That's part of the reason I went for new drivers.... Out of the box they go at about 100ma, and after I modify 'em they run at slightly over 400ma.... I don't have a light meter but I'd imagine that after market, they push at least in the 90's LPW wise.... I have no complaints! Right now 2 Cobs are out producing 6 cfls 3-42 watt 5600k and 3-26 watt 2600k in veg... I believe that they'd out perform cfls in flower too.
http://www.ebay.co.uk/itm/High-Power-50W-COB-LED-Light-Bulbs-Spotlight-E27-White-Warm-White-/331146501394?pt=UK_Light_Bulbs&var=&hash=item4d19de9d12
Thanks for the links. Of all the links, that one looks the most promising. 100 LPW, presumably 110-115 with the plastic diffusion dome removed.
How hot does CXA1304 run? There must be more to it than connecting a driver. (Heatsink requirement?). Is there a thread showing how?
Found a slight variation of the "Cree" screw in corn style bulbs using 3014 SMD instead of 5730s, they claim a substantially higher lumen per watt, but I know that the 3014s have much lower intensity so it's still probably a trade off. But still, 120-150 l/w is impressive, if they're not lying.
Maybe @mauricemm can give his opinion. He turned me onto the SMD 5730 corncob lightbulbs. I don't see much info about SMD 3014 except that it's 0.01w while 5730 is 0.5w. I agree the 100-120 lm/w sounds good.
In most cases it's all we have to go on. Have you seen the $40 PublicLabs DIY spectrometer? (<<link). I'm wondering if there's a way we could use a standard reference point so we could use one to measure umoles or something more meaningful than lumens/lux. I have one and play with it occasionally. It's calibrated using a CFL, looking for the spectral spike for mercury. They claim a 5nm accuracy. The "spectral workbench" software has the ability to save/export data. I haven't looked at the data it produces, but wonder if it could be processed in a way to normalize everyone's measurements using info from their own reference point. For example, the sun at noon? Using that measurement the device's sensitivity would be known.
Yeah, but you added almost 50% more light. It's unclear if more lumens did the job, or more efficient targeting of spectrum. It would be more convincing if you reduced the 3000 white by the amount of monochromatic red/blue watts you add. (Which gets back to how we don't have a lot to compare such things. Just comparing cool or warm whites by lumens per watt is about as good as it gets unless we can devise a standard way to measure spectrum power distribution.).