DiY LED - Cree CXA3070

Thnx Scotch, it wouldn't be too hard to put something together like that. Not much has changed in the Reds/Blue arena lately but the whites sure have been on the move.

I have noticed some of the commercial panels are using cool white/red. Even though they have higher radiometric efficiency, I am not convinced that neutral or cool white is the ideal spectrum for flowering.
View attachment 3035878


To make matters even more in favor of 3000K than what is shown in that chart, we have the mid and top bin available to us whereas the we have only the low bin for 5000K. For example, the 3000K Z4 at 700mA is 46.16% efficient. The 5000K Z4 is 46.3% efficient but has a lower photosynthetic efficiency. So for those who are after the most "perfect" flowering spectrum currently available, I suggest mixing the CXA3070 3000K with Luxeon ES deep red and Luxeon ES deep blue (but there is nothing wrong with running 3000K on its own and it is so much easier). I am currently testing a panel with just that in mind and working on building some more if I ever get done trimming

View attachment 3035867

5 week update

Purple Chem


Blue Widow


Ace Of Spades

SupraSPL said:

Thnx Scotch, it wouldn't be too hard to put something together like that. Not much has changed in the Reds/Blue arena lately but the whites sure have been on the move.

I have noticed some of the commercial panels are using cool white/red. Even though they have higher radiometric efficiency, I am not convinced that neutral or cool white is the ideal spectrum for flowering.
View attachment 3035878


To make matters even more in favor of 3000K than what is shown in that chart, we have the mid and top bin available to us whereas the we have only the low bin for 5000K. For example, the 3000K Z4 at 700mA is 46.16% efficient. The 5000K Z4 is 46.3% efficient but has a lower photosynthetic efficiency. So for those who are after the most "perfect" flowering spectrum currently available, I suggest mixing the CXA3070 3000K with Luxeon ES deep red and Luxeon ES deep blue (but there is nothing wrong with running 3000K on its own and it is so much easier). I am currently testing a panel with just that in mind and working on building some more if I ever get done trimming

View attachment 3035867

Click to expand...

I wouldn't use anything lower then 80-90 CRI for growing plants, I believe its far more important to give the plants a better quality of light rather then an abundance. Once you saturate the canopy with a particular wattage of light other factors become more important for larger growth you could damage the plant giving it to much.

I myself would rather grow the best weed I can and fill the canopy with the most beautiful naturalistic light I can, I believe Bridgelux is unparalleled in their quality of product and manufacturing techniques to achieve, proberly the best looking light. The plants have to deal with this light everyday, we might as well spoil them

I use the HP power supplies because I want a true dimmable source, PWM is not dimming and it can be very useful when training...

Just because a transformer is larger does not mean you will have a higher voltage, or "fill the room with high voltage electricity", thats what a tesla coil is for You talk so much about electrical efficiency yet you cannot comprehend the basics of electronic circuits. tisk tisk

Let me know what you think of that Ace of Spades! Have 5 seeds in solo cups now and ten more to run through..

Thank you for the insight, what actually defines photosynthetic efficiency? amount of spectrum falling inside....? is or isnt worth the trade in spectrum to mix both since? And mono blues eh!? Man you and guod have both suggested em, cant ague wit'da'guru's. Thank you for the schooling.

I know this is a COB--> thread but how efficient would it be if we use cool/neutral white stars (xmls?) to mix the spectrum of 3000k (on a switch) What kind of efficiency are the best white stars? Or vice versa, run the 5k cobs/ww stars if efficiency trades that way- but highly doubtful in my newb expertise. HA! Im talking out my ass right abouttttt...

Guessing those mono blues are the most eff pucker out there though, best way to get my blue? Everything will be timed out (I WILL mimic the sun..) lol.

Be safe brother! I may take this to a PM if you prefer.

PICOGRAV said:

Just because a transformer is larger does not mean you will have a higher voltage, or "fill the room with high voltage electricity", thats what a tesla coil is for You talk so much about electrical efficiency yet you cannot comprehend the basics of electronic circuits. tisk tisk

Click to expand...

What I am referring to is designing a power supply that will drive several COBs in series to get closer to the mains voltage. I suspect that is the only way we can achieve 95% efficiency.

Forgive me Pico, but I am going to be direct with you. I politely asked you to inform us of your superior design and you have jabbed at me and attempted to insult me 3 times. I suspect your power supplies are inefficient and that you have not checked. If that is true, there is no point in growing with LED you might as well use CMH or HPS.

I also suspect that using high CRI LEDs is a less efficient approach than standard CRI. Building a high CRI white LED from a blue die requires more of conversion to red and deep red. Did you look at the graph I just posted? If you want deep red, it is more efficient to simply use red and deep red LEDs. I understand saturation and I do not think we are suffering the effects.

According to their own specifications, Bridgelux Vero is significantly less efficient and more expensive/W than the Cree CXA3070. Will it grow awesome bud, YES! Is it the best currently available, no.

You can get real, non PWM dimming using constant current drivers and still maintain a very high driver efficiency.

Scotch feel free to ask any questions on this thread. The Ace was awesome and is a very popular smoke. I ran 5 beans, got 3 females and kept 2 of them around. One has dark pink calyxs a very black cherry soda genotype.

Photosynthetic efficiency has two aspects. One is the fact that individual photons drive photosynthesis and since blue photons require more energy to produce, they are a less efficient way to get the job done than reds. Therefore we use as little blue as is required to get the job done. KNNA recommends 25-30% for vegging and 15-20% for flowering (PAR Watts) and I agree with his recommendation. Luckily for us, blue LEDs are the most efficient there are so, that makes up for it (Luxeon ES M4R is 55% efficient at 700mA). From my experience, adding deep blue to HPS increases trichomes, terpenes and maybe even cannabinoids (speculating). 3000K LED may benefit from some extra blue, especially the Vero.

Second is more of a speculation that blue and red photons are more efficient than green/yellow because the leaves absorb the blue and red/deep red much better. We used to think this would give LED an additional 25% advantage over HPS, but in practice it may not be quite that much after all. That said, I do believe there is some photosynthetic efficiency advantage over HPS and that adding the reds/deep reds makes buds denser and finish faster (speculating).

Regarding XML2, they are the most efficient non-COB whites, but they are not as efficient as COBs. I like the XML2 for vegging though because they can spread out their wattage better than COB and we dont need nearly as much wattage during veg.

Sounds like you have an awesome design in the works

Hello everybody... im a fellow led addict... i was wondering whos using the multichips from cree... and would they not blend in the spectrum better since the xte... xpe.. xml..xpg can all be customized by the manufacturer and theres is already some chips that have the 3000k and 4500 to 7000k up to 20000k... they run from 30 watts to 150watt i think the 100watt put out 10000 lumens @ 100 watt and 34- 38 vf/1400 ma / 3500ma... im sorry for hijacking the cob thread but all in all its still a multichip... btw great thread and so many usefull links for drivers totally feeling at home here... and also what do yu guys think of the constant current waterproof drivers from ebay... ive never had a problem with those... and also wanted to let people know that some ebay sellers have package deals.... which include the heatsink pre-drilled to fit a 10 watt up to 100 plus chip... they include multichip in any colour and have a driver either a meanwell or ebay waterproof drivers and i forgot they include a reflector with a lens of yur degree 5 to 120 anyways im also on the hunt for the same end result as everyone else... phat smelly dense buds!!!!

★★★kushed_out★★★

SupraSPL said:

What I am referring to is designing a power supply that will drive several COBs in series to get closer to the mains voltage. I suspect that is the only way we can achieve 95% efficiency.

Forgive me Pico, but I am going to be direct with you. I politely asked you to inform us of your superior design and you have jabbed at me and attempted to insult me 3 times. I suspect your power supplies are inefficient and that you have not checked. If that is true, there is no point in growing with LED you might as well use CMH or HPS.

I also suspect that using high CRI LEDs is a less efficient approach than standard CRI. Building a high CRI white LED from a blue die requires more of conversion to red and deep red. Did you look at the graph I just posted? If you want deep red, it is more efficient to simply use red and deep red LEDs. I understand saturation and I do not think we are suffering the effects.

According to their own specifications, Bridgelux Vero is significantly less efficient and more expensive/W than the Cree CXA3070. Will it grow awesome bud, YES! Is it the best currently available, no.

You can get real, non PWM dimming using constant current drivers and still maintain a very high driver efficiency.

Click to expand...

IMO. growing with LED is better then HPS or CMH regardless weather they are more efficient or not and you will never win that argument say thing LED arrays produce more measurable light output then HPS, because they don't, you can apply Photosynthetic efficiency curves to the calculations all you want, HPS is king of efficacy...

I use those HP power supply because I love vintage electronics and they way engineers thought back then, they really are something special and everytime I walk up to my power rack I smile thinking about my little weed laboratory

If you really want to make the best LED Array power supply, you would buy a Laminated Silicon Steel core, with the correct calculated inductance level to fully couple the windings at normal out put. Your AC mains coil should be wound around the core along with individual low voltage coils, one per LED you want to drive. This low voltage AC needs to be rectified into rough DC and then filtered with a large Capacitor and then can be sent through a single Bipolar voltage regulator with an adjustable limit.

The closer the system runs to peak output and the better matched the transformer windings are would be the only way to approach 95% efficiency, although the more you dim the light you will see huge losses into heat from the voltage regulator.

There are some more advanced ways to preform such a voltage modification but let me tell you they become far more complex, do you have a few hours to spare?

Peeps be going nasa down here.

Gonna have to do a test one day on efficiency of dimming. I'll run my light dimmed 50%...about 180w. If it grows a g/w I would say its efficient enough..

just a side note..the last inventronics driver I picked up is rated at 93.5% efficient. Good enough..

Thank you Pico that is good info. It sounds like I ought to stay the course with the 90% China cheapos.

DIYkindaGuy welcome to riu! So I looked over those chips in your pics and according to the ad they are only 71lm/W in a 6500K color temp which is less half the efficiency of the big name brand COBs. Also that 6" heatsink they provide would cook the COB in no time running at 1750mA yipes lol. So I think you can get a much better bang for your buck.

The XTE/XPE multichips are better, although if you are going to do a DIY might as well get the best

Yes the cree multichip are deff more efficient... then no name chinese... but yu gotta look at my point of view lets say yu get a top of the line cree cob with a meanwell driver... and yu get lets says 150l/w but for the same price of the quality yu can get two chinese cobs with chinese drivers pushing 60l/w now times that by two and your pretty much getting the same i guess... but at double the watts... now where i live my hydro is included but i do want the best just been using sub parr till i get all my data together then plan on crunching more numbers... and have yu noticed that the cobs have a higher vf then multichip... i wonder if that is a good or bad thing and how it all affects effeciency
★★★kushed_out★★★

Take a gander at this

Mega computing power went into this, but you can literally design your own panel AND see how it comes together in its' own graphs

http://www.bmlcustom.com/custom-led-strip/

SupraSPL said:

5 week update

Click to expand...

Solely COB?

Great results and pictures!

Thnx CB, this room is the first LED only I have run in a few years. The Blue Widow is directly under COB only but the other two are getting a mixture.

SupraSPL said:

Thnx CB, this room is the first LED only I have run in a few years. The Blue Widow is directly under COB only but the other two are getting a mixture.

Click to expand...

Sweet! If you would leave it there that would be awesome. I want to see only 3000k results

Yes I plan on leaving it. And the Ace of Spades in the middle is under 4 COBs also.

Cool link PF Ill have to look read into their design more.

Would 1x 5000k be alright during veg for a 1.5x2.7ft space?
Thinking of 2x 3000k, 1x 5000k and considering either using (veg) the 5000k solo or 1x 5000k 1x 3000k then switching on the other 3000k at flower for 2 plants.

Good idea or no? Definitely not an electrical engineer and all these driver specs are driving me a bit nutty also @@@

letitgrowson said:

Would 1x 5000k be alright during veg for a 1.5x2.7ft space?
Thinking of 2x 3000k, 1x 5000k and considering either using (veg) the 5000k solo or 1x 5000k 1x 3000k then switching on the other 3000k at flower for 2 plants.

Good idea or no? Definitely not an electrical engineer and all these driver specs are driving me a bit nutty also @@@

Click to expand...

You might want to check this thread out https://www.rollitup.org/led-other-lighting/733157-cree-cxa3050-100w-custom-led.html. He has about the same growing space as you. He is using two of cxa3050