DIY led grow

Positivity

Well-Known Member
So the 3590's will still be 5k along with the 660's?
If I ever flower with it...yes. Always wanted to flower with a cool white and red mix, something I've never tried before.

I'm impressed. It looks like something you'd buy online for a ton of money!

It's got a place in my top three favorite DIY LED light fixtures.
Thanks AP. That's why I go through all the trouble of building them now. I still pay a similar amount of money but in the end I end up with all the features I want. With all the parts readily available you can build damn near anything

Keep up the amazing Job !
(it reminds me the "Jet" somewhat !!! :-P.. )

Too much blue ,Pos ...
Even for a veg light ...
For my taste at least ..
You 'll get to see the results .
:wink:
(not that it won't work ....But ...)

Cheers.
:peace:

Thanks sds...it's the stealth bomber..lol.

Too much blue? I'll find out shortly...zero node distance? Oh well, I had the bjb connectors and all my small stars so I figured I'd give them a home. Can always put white in there or something...xpl!? It's only a small 12w driver running them..on a switch too so I don't need to use them. If I get weird growth I'll just turn them off..
 

Positivity

Well-Known Member
A few notes I took down..


After 1 hour running. Ambient 32*c . Cxa 3590 thermocouple attached with kapton to Tj on cxa. Temp readings fluctuate between 41*c to 48*c, about a 7*c fluctuation could be due to single piece kapton tape used and radiant heat affecting reading. High readings taken.

Main fan off. Led 1 fan off. Led 2 fan on high

Inner case temp. 62*c. Led 1 temp. 81*c. Led 2 temp. 62*c

Notes. fan temp sensor giving warning buzz at 70*c. Adjustable warning temp. Only sounds when it hits 70*c. Does not sound above 70*c. Only will hear alarm at 70*c.


Main fan off. Led fan 1 on high. Led fan 2 on high

Inner case temp. 62*c. Led 1 temp. 65c. Led 2 temp. 62*c.


Main fan on high. Led 1 fan on high. Led fan 2 on high. Cooling down from 80*c fan off test in about 10min

Inner case temp. 38*c. Led temp 1. 48*c. Led temp 2. 47*c


Notes.

This CPU cooler will passive cool at 82*c. Likely to rise as ambient rises. Thermal switch won't kick in until 85*c -90*c. Will lower that to maybe 70*c in the future. 80*c makes for a very hot light.

6*c rise in case with all fans on high at 1a dissipating 200w.

CPU coolers provide roughly a 10*c rise over ambient case at 1a on a cxa 3590 dissipating 88w.

Close in line with other tests showing a 20*c rise at 150w load from this particular CPU cooler.

Case runs at ambient at 500ma, cxas about +5*c. Where it will be spending most of its time..
 
Last edited:

churchhaze

Well-Known Member
Also, keep in mind that warm white is not the same thing as cool white + red even if color temperature comes out to be the same thing.

The warm white's warmness mostly comes from the large amount of yellows, ambers, and oranges. No amount of 660nm red will replace amber or yellow or orange..

My guess is that cool white + red will not do as well flowering as warm white, and that the 660nm will be more of a benefit to veg than flower. Of course, since I haven't tried all the combinations, It's only a guess based off how i think plants work. Also of course, had I thought 660nm + cool white was the best combination, I would have tried that first myself, so my guess is implied that I didn't think it was the best combo.
 

churchhaze

Well-Known Member
There could be some gotchas here. With humans, if you mix green light with red light, the human brain will perceive yellow light. Since plants have a very similar form of color vision using phytochromes, it's likely there are combinations of wavelengths that are perceived identical as a different single wavelength.

For example, the right ratio of 730nm + 660nm may be perceived by the top layer of a plant as 600nm if the %Pfr converges at the same place. Of course that's an oversimplification, as plant vision is "3d", and the bottom layers see different colors than the top. In the case of the 600nm single wavelength source, all layers of canopy would converge at similar %Pfr, but the dual combination of 730nm + 660nm would appear different at different levels of canopy, based on how much 660nm was already filtered out.
 
Last edited:

Positivity

Well-Known Member
There could be some gotchas here. WIth humans, if you mix green light with red light, the human brain will perceive yellow light. Since plants have a very similar form of color vision using phytochromes, it's likely there are combinations of wavelengths that are perceived identical as a different single wavelength.

For example, 730nm + 660nm may be perceived by the top layer of a plant as 600nm if the %Pfr converges at the same place. Of course that's an oversimplification, as plant vision is "3d", and the bottom layers see different colors than the top. In the case of the 600nm single wavelength source, all layers of canopy would converge at similar %Pfr, but the dual combination of 730nm + 660nm would appear different at different levels of canopy, based on how much 660nm was already filtered out.

Apache tech always struck me as an example of photons are photons.

Look at how bright white the spectrum is compared to most other lights.

The only thing to be concerned about it seems it's the morphological effects that some nms bring

Hard to beat the simplicity of a warm white cob though...

One more year of spectrum manipulation should be enough to sort this out well enough
 
Last edited:

DonPetro

Well-Known Member
Apache tech always struck me as an example of photons are photons.

Look at how bright white the spectrum is compared to most other lights.

The only thing to be concerned about it seems it's the morphological effects that some nms bring

Hard to beat the simplicity of a warn white cob though...

One more year of spectrum manipulation should be enough to sort this out well enough
Thanks to you for being a pioneer in the quest for DIY Spectral Knowledge! I was theorizing about the same type of spectrum but you will likely beat me to having one built. How many reds do you plan to use with the COBs?
 

Positivity

Well-Known Member
Thanks to you for being a pioneer in the quest for DIY Spectral Knowledge! I was theorizing about the same type of spectrum but you will likely beat me to having one built. How many reds do you plan to use with the COBs?
Your welcome don. Just got sick of the mess of spectrums and secret sauces out there..

From this point on any cob supplementation will be minimal, 20w or less in a 2' x 4'.

If I use this new light it will be just 4 LEDs at 1.8w each. Probably leave the 430nm in the middle. Maybe two lights set up with that color combination.

It will be a good and semi final comparison against all 3000k..

So far the 3000k with Oslon hyper red has had the most impressive growth. Just a bit less red next time..dependent on how my current 3k run finishes
 

churchhaze

Well-Known Member
Usage of the correct units is my first filter for bullshit math. :eyesmoke::bigjoint::joint:

Like if someone is trying to measure efficiency and there are units, I know their calculation doesn't find efficiency. Efficiency is unitless since the units cancel out. (power out / power in) It should also be a number less than 1... :P

I only mentioned it because lots of folks are watching and get confused easily with w's a's and l/w etc.
 

Positivity

Well-Known Member
Usage of the correct units is my first filter for bullshit math. :eyesmoke::bigjoint::joint:

Like if someone is trying to measure efficiency and there are units, I know their calculation doesn't find efficiency. Efficiency is unitless since the units cancel out. (power out / power in) It should also be a number less than 1... :P
lol...back when I first started posting and was learning things I made a lot of those type mistakes. I got blasted for it pretty quickly..haha. I have a thick skin and a bit of determination though
 

Positivity

Well-Known Member
Had some issues this time around. Think I input too much calcium with my lime mix on the last soil amendment. Put in a bunch of worm castings and I should of known that was enough.

Think the magnesium was being locked out on top of a also hot mix. Put just a little neem and crab meal but not long enough of a compost time.

Anyway, after a few weeks of rehab their back on track.

All 3000k stretches less, which is welcome. But I can see why the reds stretched them out the way they did. The 3000k has a little of that hps stretch to it all on its own.

Also...my heatsink runs quite a bit cooler pushing more watts using just cxas. The reds and 365nm contributed more heat than I gave them credit for.

Just after another Epsom foliar...get in what I can before the buds start setting. Also removed the middle reflectors trying to lessen the middle par hotspot most lights give out.


image.jpg image.jpg image.jpg image.jpg image.jpg
 
Last edited:

PSUAGRO.

Well-Known Member
Also, keep in mind that warm white is not the same thing as cool white + red even if color temperature comes out to be the same thing.

The warm white's warmness mostly comes from the large amount of yellows, ambers, and oranges. No amount of 660nm red will replace amber or yellow or orange..

My guess is that cool white + red will not do as well flowering as warm white, and that the 660nm will be more of a benefit to veg than flower. Of course, since I haven't tried all the combinations, It's only a guess based off how i think plants work. Also of course, had I thought 660nm + cool white was the best combination, I would have tried that first myself, so my guess is implied that I didn't think it was the best combo.
https://www.icmag.com/ic/showthread.php?t=215147 got over 2 gpw using CW/R (evo v1)........if you believe the #s:wink: This is before apache/a51 similar specd panels
 

churchhaze

Well-Known Member
hmmm... maybe I should replace my 700mA drivers with 500mA drivers for my vero 18 and replace the lost power with xp-e photo reds if GG gets good results with his new build.

I'll admit that I'm a copier. I look for good results and copy.

https://www.icmag.com/ic/showthread.php?t=215147 got over 2 gpw using CW/R (evo v1)........if you believe the #s:wink: This is before apache/a51 similar specd panels
 

Positivity

Well-Known Member
https://www.icmag.com/ic/showthread.php?t=215147 got over 2 gpw using CW/R (evo v1)........if you believe the #s:wink: This is before apache/a51 similar specd panels

I remember that grow from way back. That was a high density sog..wasn't it? I wouldn't doubt it if it produced that much...all tops !?

hmmm... maybe I should replace my 700mA drivers with 500mA drivers for my vero 18 and replace the lost power with xp-e photo reds if GG gets good results with his new build.

I'll admit that I'm a copier. I look for good results and copy.

I wouldn't call it copying...more research..lol...:wink:.
 

Greengenes707

Well-Known Member
I am noticing that the cxa's are out stretching the AT's. They are untrained...but I still think there is more stretch from the 3K cxa than the 2800K spectrum blend of the AT. Personally it's the amber and yellow in the 3000K that is similar to hps that I think gives it that growth. AT's are cool white and lack in that area comparatively...but my light will be neutral white so should be closer to the 3000K growth we have been seeing. All in theory of course. We will see soon.

EDIT:
I also have the 80cri 4K's...more red than the graph in the data sheet.
 
Last edited:
Top