Lighting Bilbo's Brief Guide to Choosing Your LED.

It's been a while since I stopped by, busy getting new stuff sorted ;)

There are numerous studies done, hundreds, if not thousands, if not hundreds of thousands, and many studies show one result, and many others supposedly under the same or very similar conditions show entirely different results.

So light drives photosynthesis, we all know this, however optimising the quality of the light drives it better, BUT, some wavelengths are efficient, some are inefficient and overall, very few experiments get the same results time after time. Refer to earlier post that shows that one professor can achieve (and we have seen the results personally) 2 slightly different results from 2 bays using exactly the same stock, regimes etc. side by side...

The same scientists say that they want this wavelength or that wavelength adding or subtracting from a light and still they get the same results, I guess as has been said elsewhere and as I myself have said numerous times, plants are not ready to give up all their secrets. For the time being, and running 9 bays as we do of reasonably open research plants, we test different spectra regularly, we have the capability to produce millions of different hues and still the majority red/blue spectra seems to come out on top ever so slightly...by using less power in the wavelengths that plants overall seem to be using less of. That's all we can say...white, overall does a pretty good job but not usually better than the optimised spectrum for example we run in our phosphor mix, but this in turn doesn't do quite so well as the red/blue/phosphor mix all together...I'm looking forward to sharing some of this information as soon as I can.

Science certainly isn't demanding us to stop making red/blue lights though, far from it, in our case they are simply adding more wavelength to optimise against the 450/660nm base we already have.

As for laser, getting an even beam out of laser is difficult, (I come from an entertainment background and have been using lasers and LED for a very long time)...laser is very difficult to diffuse properly as it is designed to give a powerful focused beam, and all the tech is pointed this way...it's also worth noting that when you see multiple laser beams, it is actually one beam moving very quickly, and/or being split. Xagor is right in that it can be quite efficient, but it's really very difficult to control, both thermally, optically and spectrographically, and that is where the cost comes in...again it's another use of a phosphor dope. The terminology "laser light" is often very misused, from memory and without resorting to spending hours wandering around the tinterweb, the current crop of white lasers are doped from blue, creating a "quasi" laser beam suitable for use in headlights, much the same way as most LED is now made from blue, and doped.

With a little luck it might not be weeks before I stop by again!
 
What is your testing like?
100% artifical light, >500umol?

Yes spectrum can be improved compared to white, e.g. adding some deep reds to it.
But...
What do you think who comes out ahead in this situation:
flowering, specialized r/b spectrum 400umol/s/m^2 vs 600umol/s/m^2 warm white spectrum, both same power consumption, growing weed?

In hubbas links there is a video testing BML Spydr, beeing one of the best companies in the market. A Cob setup costs much less and has about 40-50% more photon output per watt. Heavy stuff horti companies should consider.
For R&D you should get a cxb3590 best bin, slap it onto a computer heatsink and a driver, run it at 1A. Measure wattage. Put it over your light wand and you will think the wand is broken :)
 
Last edited:
A Cob setup costs much less and has about 40-50% more photon output per watt. It´s sick. Heavy stuff horti companies should consider.

Glad I purchased GroGlo COB lights then!

:woohoo:
 
I dont think its actually cobs that are comparable to cree or bridgelux cobs. More like an array of various diodes.
Ok, or I should be more specific: not cobs in general but cobs of latest generation cree CXB series or bridgelux Vero V2 series.

CXB3590 CB bin 3000K: at 0,7A about 2,86 umol/J, 58% efficiency, 190lm/W.
Run with a meanwell HLG driver thats about 0,94% efficient and add 10% reflector losses it will be ahead of
BML Spydr in the video with 1,73umol/J....
 
Last edited:
CXB3590 CB bin 3000K: at 0,7A about 2,86 umol/J, 58% efficiency, 190lm/W.
Run with a meanwell HLG driver thats about 0,94% efficient and add 10% reflector losses it will be ahead of
BML Spydr in the video with 1,73umol/J....

I am still learning about Cree and Bridgelux COB (circuit on Board). You are talking about 3100K LEDS right?
Are they more efficient than a 3100k Ceramic Metal Halide (CMH) rated at 1.91 цmol/J

I am interested in COB LEDs, can we compare the cost per watt of COB vs CMH. Is that a fair comparison, just for a start point... They are both 3100K but spectrums are a little different, maybe it matters?

$500 for a 315 watt CMH light. Plug and play.
$500 for a 300 watt COB LED: $240 for COBs, $100 for driver, $160 for other parts like 6 heatsinks and fans, reflectors, wires, power cables, mounting frame. Plus the time for construction of a COB light.

LEDs are theoretically more reliable....LOL, not if I build it. My first DIY LED will be as crappy as my first Cannabis grow.
"In theory, there is no difference between theory and practice. But, in practice, there is."

https://www.autoflower.org/posts/915020/
 
Last edited:
LEDs are theoretically more reliable....LOL, not if I build it. My first DIY LED will be as crappy as my first Cannabis grow.
"In theory, there is no difference between theory and practice. But, in practice, there is."

He he he ...

True story: ... just yesterday, my daughter asked me what my favourite animal is.
I said, "the bumblebee".
She looked at me confused and asked why.

I said. "because according to scientists, it can't fly ... and yet every day, it does the theoretically impossible."

The humble bumblebee ... Nature's one finger salute to scientific theory and the 'unconscious incompetents' [scientists].
Showcasing the practical miracle of intelligent design ... and a reminder that we do not know, what we do not know.
Nature still holds all the secrets ...
 
You are talking about 3100K LEDS right?
There is a 2700K, 3000K and 3500K version, but not 3100K.

See attached for your comparison. Best Driver for this is meanwell hlg-185h-c1400.
At 1,4A cobs run with about 50W for the diagramm. For 240$ you get about 5 pieces of those, so about 250W.
That 50C means at 50°C which is about what you get when you use active cooling.
Vf Forward voltage.
dissW -> dissipated wattage.
% Efficiency.

The humble bumblebee ... Nature's one finger salute to scientific theory and the 'unconscious incompetents' [scientists].
Sry, but thats an old myth. Of course it can fly according to science.
It could not if its wings were fixed like those of an airplane...thats where the myth comes from.
 

Attachments

  • CXB3590 3K CB.png
    CXB3590 3K CB.png
    39.3 KB · Views: 152
Sry, but thats an old myth. Of course it can fly according to science.
It could not if its wings were fixed like those of an airplane...thats where the myth comes from.

A bit like the scientific theoretical myth of not being able to travel faster than the speed of light ?

:pighug:

he he he ... all good fun ...
 
There is a 2700K, 3000K and 3500K version, but not 3100K.

Thanks for catching my mistake. You are right, 3000k.
And thank you Xagor, for pointing me towards COB LEDS, they are interesting!

Would you estimate the cost for parts to make a 300w COB LED at 1.93 цmol/J on the table you shared..
Or we could go with your estimate for a 200w unit and add 50% for simplicity.
Was I too high or too low?
 
Last edited:
4x 3590 about 200$
1x HLG-185H-C1400B about 65$
4x CPU Heatsink e.g. Alpine (I think alpine 64 fits for the holder), includes fan, about 40$
4x Cob Holders e.g. "ideal" 12$
thermal paste e.g. artic silver 5$
1x 5V or adjustable 5V-12V power source for fans 5$-10$
some connectors, wire, maybe 5$
Some aluminium parts to connect the heatsinks and screws from local shops, maybe 20$
so about 360$ maybe with shipping about 420$
for: 200W led power with 94% driver efficiency its about 212W + maybe 5W for the fans, so about 217W draw.
Leds at about 2,52 umol/W.
Including total draw and about 10% losses Id guestimate its about 2,1umol/J for the system.

In the vid of your link they mention blackdog plat xl-u at about 1,19umol/W and it costs 1850$ at 750W draw.
You need roughly about 2x of these cob fixtures to equal the light output, save 300W wasted energy $ (heat) and save about a 900$ starting investion.
edit: I forgot to mention: you should have a drilling machine and a drill for threads (like 5$-10$) to be able to screw the holders to the heatsink and for mounting the alu sheets to the heatsinks Ghettostyle would be using thermal glue instead of thermal paste, but id not recommend it for these otherwise high quality parts.
Maybe sounds complicated, but its all quite easy. With holders no soldering is required.
 
Last edited:
Back
Top