DiY LED - Cree CXA3070

akaki

Well-Known Member
It's true that the high CRI has more far red, on the other hand most of this >700nm light is (imo) wasted energy.

The Vero 4000K vegs very well when the intensity is high enough. For lower intensity the 5000K is a better choice. For flowering the Vero 3500K and 4000K both are superb.

Btw I'm curious why no one has ever tried the CXA 3500K for flowering.
my next project is with 2 veros 29 4000k 80cri and 2700k 90 cri i was curious if i need more blue...
 

SupraSPL

Well-Known Member
Guod pointed out something similar to this in another thread. I agree with Mr Flux, the high CRI wastes more in the infrared area. Looking at this chart I am sold on the 3000K. I also agree with speedyganja, if you want to improve the spectrum you could add deep reds on a separate string. I have tried it both ways and they all work great, but 3000K straight up is much simpler and it works great on its own.
CXA spectral flux mod.png
 

speedyganga

Well-Known Member
How much oslon do you thing would be good to supplement my 3000K ? I was thinking of 3 oslon (make 6W at 700mA) for a cxa at 1.4A it should add a nice red curve on the end. Also I would put the red only at the end to induce nice fat colas.
 

SupraSPL

Well-Known Member
These CXAs are running at 700mA, so there is 50W of 3000K COB. I added this supplemental string mostly to use up what I have on hand but also to try and improve the spectrum without harming efficiency too much. Cree XPE 630nm red, Luxeon ES 660nm deep red, Oslon SSL80 660nm hyper red and Luxeon ES 450nm royal blue. The colors are on their own string @ 650mA.

So your suggestion would work great.
DSC07301a.jpg
 

epicfail

Well-Known Member
It's true that the high CRI has more far red, on the other hand most of this >700nm light is (imo) wasted energy.
Guod pointed out something similar to this in another thread. I agree with Mr Flux, the high CRI wastes more in the infrared area. Looking at this chart I am sold on the 3000K. I also agree with speedyganja, if you want to improve the spectrum you could add deep reds on a separate string. I have tried it both ways and they all work great, but 3000K straight up is much simpler and it works great on its own.
View attachment 3174031
Thank you both for the info,

I definitely have learned a lot from the people in this forum, there is a true wealth of knowledge available here to those willing to learn. if it wasn't for people like yourselves there would be a lot more people who would still be slaves to the HID vendors. I know people who only started growing since I showed them this thread and bbspills thread because using HID was never going to be an option.
 

churchhaze

Well-Known Member
Totally agree with this. It was that chart that had me sold on the 3000K too. (mainly against the high CRI models)

The amount of power lost between 500nm-650nm is unacceptable for a small gain in far-red. If it turns out more far-red is desirable (on at all times), we can use 730nm leds.

Guod pointed out something similar to this in another thread. I agree with Mr Flux, the high CRI wastes more in the infrared area. Looking at this chart I am sold on the 3000K. I also agree with speedyganja, if you want to improve the spectrum you could add deep reds on a separate string. I have tried it both ways and they all work great, but 3000K straight up is much simpler and it works great on its own.
View attachment 3174031
 

Bueno Time

Well-Known Member
Thank you both for the info,

I definitely have learned a lot from the people in this forum, there is a true wealth of knowledge available here to those willing to learn. if it wasn't for people like yourselves there would be a lot more people who would still be slaves to the HID vendors. I know people who only started growing since I showed them this thread and bbspills thread because using HID was never going to be an option.
These are really awesome devices for sure. You can now buy a single COB and run a tiny micro cab as small as 1 sq ft and pack in the light intensity of a big HPS and only be running ~50w so its not going to be hard to cool.
 

uzerneims

Well-Known Member
These are really awesome devices for sure. You can now buy a single COB and run a tiny micro cab as small as 1 sq ft and pack in the light intensity of a big HPS and only be running ~50w so its not going to be hard to cool.
Nice that is what i need, because of stealth issues igot some CFL's and when someone comes over i must to turn off fan, so... My plant almost died... This could be a solution - LED's
 

churchhaze

Well-Known Member
If you have an arduino or micro of any kind laying around, you could make a very crude cheapo DIY light meter from a photo resistor connected to one of the analog inputs through a voltage divider. The photoresistor makes up one resistor in the voltage divider, while a standard ordinary resistor makes up the other half of a voltage divider.

With a 4x 7segment LED display, "a number" can be displayed from 0-9999. (or 0000-ffff if you can read hex)

Obviously it sees some wavelengths better than others, and wouldn't be calibrated to anything, so the numbers would be absolutely meaningless to anyone else, but it would tell you which areas are brighter than others.

(I did something similar myself just to try out the a/d converters on my arduino, and to write/test simple drivers for a 4x7seg led setup.)

 
Last edited:

churchhaze

Well-Known Member
That'd probably be a lot more accurate. If you wanted to save money, however, you could actually attach the photoresistor voltage divider to the multimeter and read out the voltage as brightness too.

Obviously it would just give you a general idea of what's brighter than what rather than absolute readings.

http://www.amazon.com/50-100K-Photoresistor-Resistors-Light-Dependent-Resistance/dp/B00HUHC9D2/ref=sr_1_3?s=industrial&ie=UTF8&qid=1402441794&sr=1-3

Thinking of getting one of these,you can just connect it to a multimeter.

http://www.apogeeinstruments.com/sq-120-electric-calibration-quantum-sensor/
 
Top