Gaius
Active Member
May grab these bad boys for my build since it appears the angle is a bit wider, which should work well for the COB spacing I chose. Still really hoping to find something that is multi-faceted like PetFlora recommended though.
DiY LED - Cree CXA3070
- Thread starter SupraSPL
- Start date
SupraSPL
Well-Known Member
Gaius I agree multi faceted does seem like it would be better. I might just make some diy flat white reflectors for now.
mtnstream at 66% power (2A) the CXA3070 3000K is 33% efficient. There would be a savings up front buying less emitters but it also means buying more drivers and more heatsinks to get the same job done (photons). On top of that there is the additional electrical cost so it is a tough call to find the value point because it depends on what the price of electricity does.
Hyroot, what are you currently vegging with?
mtnstream at 66% power (2A) the CXA3070 3000K is 33% efficient. There would be a savings up front buying less emitters but it also means buying more drivers and more heatsinks to get the same job done (photons). On top of that there is the additional electrical cost so it is a tough call to find the value point because it depends on what the price of electricity does.
Hyroot, what are you currently vegging with?
PetFlora
Well-Known Member
SupraSPL
Well-Known Member
While I was laboring on the heatsinks I was wondering how much power we can put into each heatsink and still get a decent Tj with active cooling. I was able to cool 100W with .5W of fan power to a stable Tj of 37C. Ambient was only 17C so it was not a perfect test because the grow room ambient will be up to 24C, but I think it will stay below Tj 50C in the grow room.
These were running at 730-750mA. I had a 120mm fan blowing into the sink at 5V. So if we bumped up the drivers to 1000mA they would dissipate 150W. Then increase fan speed and/or use 2 fans should keep Tj reasonable and lamp efficiency would be 39% Parts cost would be about $275 so that is $1.83/W and with a pair we could replace 600HPS.
These were running at 730-750mA. I had a 120mm fan blowing into the sink at 5V. So if we bumped up the drivers to 1000mA they would dissipate 150W. Then increase fan speed and/or use 2 fans should keep Tj reasonable and lamp efficiency would be 39% Parts cost would be about $275 so that is $1.83/W and with a pair we could replace 600HPS.
hyroot
Well-Known Member
I have an inda Gro 420 in one room. 1 a51 sgs 160 in the other veg room. The a51 will go back into the flower room when plants are ready. I just want permanent veg lights for the led veg room for plants on the led side of the flower room. I plan on having at least 2 panels in the veg. 4x4.
Mellodrama
Well-Known Member
Wow, only 5V on a PC fan. That's barely spinning, right? I've had a few that wouldn't turn at 5V. If you don't mind me asking, how'd you get 5V? Do you have a pile of wall warts to pick from?
Also, can we revisit Tj for a sec? How are you measuring? Are you aiming that infrared thermometer you mentioned way back when at some specific place on the COB? I notice the COB's have a test point, but that's for a thermocouple, right?
Also, can we revisit Tj for a sec? How are you measuring? Are you aiming that infrared thermometer you mentioned way back when at some specific place on the COB? I notice the COB's have a test point, but that's for a thermocouple, right?
SupraSPL
Well-Known Member
You are correct on all counts. I have an adjustable 300mA DC power supply and some fans will not start at all on the lowest settings 1.5V 3V and 4.5V. At 5 V my .2A 120mm fan is turning slowly and makes no sound at all. It was drawing 100mA at 5V.
I do have a pile of wall warts that I collected over the years from old chargers or junk boxes at thrift shops. They all test 50-60% efficient. I ordered a cheapo 12V 2A switching power supply to see if I can do better.
The way I estimated Tj was by measuring vF at ambient temp and then measuring again once it stabilizes. Tj increases by 1C for every drop of .1vF so you just subtract the difference and add the ambient temp.
I also check the heatsink temp out of curiosity to see how uniform it is and how it differs from the Tj. I aim the infrared in between the fins at the base plate. For reference, with ambient at 17C and fan at 5V, the heatsink was 22C with a Tj of 37C. With no fan at all and heatsink on its side, the heatsink temp stabilized at 36C and Tj was 76C. I forgot to check the dot on the COB with the infrared but will be doing more testing.
I do have a pile of wall warts that I collected over the years from old chargers or junk boxes at thrift shops. They all test 50-60% efficient. I ordered a cheapo 12V 2A switching power supply to see if I can do better.
The way I estimated Tj was by measuring vF at ambient temp and then measuring again once it stabilizes. Tj increases by 1C for every drop of .1vF so you just subtract the difference and add the ambient temp.
I also check the heatsink temp out of curiosity to see how uniform it is and how it differs from the Tj. I aim the infrared in between the fins at the base plate. For reference, with ambient at 17C and fan at 5V, the heatsink was 22C with a Tj of 37C. With no fan at all and heatsink on its side, the heatsink temp stabilized at 36C and Tj was 76C. I forgot to check the dot on the COB with the infrared but will be doing more testing.
Hi Supra, I'm new here and i've read this thread from top to bottom in an instant. there's some really good technical info here to learn from.
I have a few suggestions, if I may
I've been thinking about the exact same setup for a few weeks now and just found your thread. - what do you think about a vertical grow in the shape of a wall? it may take a little more room but the light spread will be more even and it will be easier to control how much light each leaf gets, what do you think?
I think reflectors will reduce your efficiency, my thought was to find a better way to arrange the plant or the light so that the light will be used more effeciently.
- about the fans, if you have a lot of them you can concider using a computer fan controller which you can find at almost any online computer store. all they need is a simple power adapter and you can hook up as many as you like,
with speed control and temprature control in some models.
i'm planning to use one - I'm waiting for the CXA3590 to come out and run them on 1000ma, with active cooling. still concidering which ones to buy though.
which reminds me, I got two questions too
-why use the 3000k leds with the low blue %, and not mix both 3000k and 5000k with dimmers so you could find the perfect red/blue ratio?
-is the red spectrum on the 3000k deep enough into the 660nm or is that not necessarily needed? i though to build my setup only with 5000k because they are more effecient than the 3000k and maybe to add red leds for a richer spectrum and one 3000k led for all the mid range spectrum.
anyways, your setup looks amazing and I can't wait see how everything adds up to a full grow. good luck!
^^^you got it! I use a 6 speed fan controller with a molex to molex ac-dc adapter and 3-pin lines for fans.
This way, you can manage all fan wiring cleanly.
Vertical railing with wheel rollers (think filing cabinet) for heat sink blocks then you could slide and reposition
array as plant grows. In this concept, you would want to make certain you position wall panels from eyes.
Do not look directly at cob chips, they are like a weld torch light that will burn the retina.
This way, you can manage all fan wiring cleanly.
Vertical railing with wheel rollers (think filing cabinet) for heat sink blocks then you could slide and reposition
array as plant grows. In this concept, you would want to make certain you position wall panels from eyes.
Do not look directly at cob chips, they are like a weld torch light that will burn the retina.
SupraSPL
Well-Known Member
Welcome to RIU! Good suggestions and questionsHi Supra, I'm new here and i've read this thread from top to bottom in an instant. there's some really good technical info here to learn from.
I have a few suggestions, if I may
I tried vertical 600HPS for a few runs and decided that I did not prefer it because the back side of the plants do not get lit and I do not rotate (too much work, sticky stinky and snapping branches). Also, the plants leaned in toward the light so it was almost as if it were an overhead angle into the canopy anyway. Finally as you mentioned vertical can take up more space to get the same job done.
LED is great for side lighting. I try to shape my canopy into a dome by bending topping and tying down. The LED modules cover all of it with no rotation required. I just clean out the underbrush that is shaded out.
I agree it is ideal if you can keep your modules in tight enough to the canopy that you do not need reflectors. They add complexity, cost and bulk to the modules. But if there are still stray photons escaping out the sides, the reflectors can redirect some of them to be useful.
If I understand correctly the fan controller requires a 12V input? That is where I am falling short, finding an efficient conversion from 120V to 12V. Once I get that part knocked Id like to check the efficiency of the fan controller. It might actually save power by adjusting for the perfect fan speed.
That would be a helluva light! 80W 37% efficient (at 3000K).
I have not been a fan of the 5000K because it peaks in the yellow just like HPS. Although the 3000K might be a bit short on blue, we could add a Luxeon M or Luxeon ES deep blue if necessary. They could run in series with the COB if you run in the 700mA range, the only question is will it reach and mix well into the canopy.
KNNA explained that you can run without deep red just fine. We used to build RWB modules that way before there were any efficient deep reds and they worked great. He also explained that some growers ran into trouble by running deep reds without reds. So when I build RWB modules I aim for 50/50 red to deep red. Based on that I believe that 3000K will do just fine on its own but that it would be improved by adding deep reds. Just like the blues, you can add deep reds in series with the COB is you are running in the 700mA range. Or you could have a blue/red string separate from your COB string.
Thanks and everything I wrote is conjecture and anecdotal so please don't let my ideas discourage you from experimenting with your build. Good luck!
Abiqua
Well-Known Member
Here is a neat little fan voltage control circuit for $5-10 in parts.
Uses a LM1088Ti adj chip [which has a lower voltage dropout than the LM3xx's] and kicks the fans on @ 5v and then you set with a potentiometer.
http://forums.bit-tech.net/showthread.php?t=219273
---also has a couple of great circuits for PWM controllers and dual timers.
If you wired in a little $3 voltage meter, these would be perfect.
Here is another great guide discussing wall warts. 5-6 pages. Comprehensive. Also a pre-filter/regulator build using a LM317 [but the above could be used as well] for wall warts is discussed
http://www.dxing.info/equipment/wall_warts_bryant.dx
Doubt any of this is over your head, but just some ideas for squeezing that last little bit out of those fans.
Uses a LM1088Ti adj chip [which has a lower voltage dropout than the LM3xx's] and kicks the fans on @ 5v and then you set with a potentiometer.
http://forums.bit-tech.net/showthread.php?t=219273
---also has a couple of great circuits for PWM controllers and dual timers.
If you wired in a little $3 voltage meter, these would be perfect.
Here is another great guide discussing wall warts. 5-6 pages. Comprehensive. Also a pre-filter/regulator build using a LM317 [but the above could be used as well] for wall warts is discussed
http://www.dxing.info/equipment/wall_warts_bryant.dx
Doubt any of this is over your head, but just some ideas for squeezing that last little bit out of those fans.
about that, if price is not something you're worried about - you can buy a high quality (80+) efficiency computer power supply and hook it up to that. (the way these controlers were supposed to be used
)that way you can add many controllers but it really depends on the size of your grow and your cooling requirements - but it should hold way more fans than you can handle.
btw, certain fans are more efficient than others. but thats a whole other story...
regarding the color temps - i've found this chart that I tried to follow. with it in mind I imagined I need high blue content and high deep red conent with some red in the middle.
View attachment 3002483
do you think this chart is acurate in any way?
I was shooting for a very high 450 and very high 660 with a bit of red in the middle - one 3k maybe.
if you could refer me to color temps discussions you found useful I'd be very greatful.
SupraSPL
Well-Known Member
Thnx abiqua I will check out those links,
Kakamika, I am on the case trying to find a small 12Vpower supply, 80% or better efficiency for a good price. I will need 2, one for each flower room. Each one will be powering 6-12 140mm fans.
Regarding flowering spectrum, the theory I subscribe to is 15-20% blue (PAR Watts) half red, half deep red and the rest white. That is how I build my RWB modules but the 3000K is low on blue, low on deep red and high on white. Not perfect but should be better than HPS spectrum.
WDIK, they do get hot and if I understand correctly keeping them as cool will improve reliability. I mounted mine on a board so they get cooled by convection at the least.
Kakamika, I am on the case trying to find a small 12Vpower supply, 80% or better efficiency for a good price. I will need 2, one for each flower room. Each one will be powering 6-12 140mm fans.
Regarding flowering spectrum, the theory I subscribe to is 15-20% blue (PAR Watts) half red, half deep red and the rest white. That is how I build my RWB modules but the 3000K is low on blue, low on deep red and high on white. Not perfect but should be better than HPS spectrum.
WDIK, they do get hot and if I understand correctly keeping them as cool will improve reliability. I mounted mine on a board so they get cooled by convection at the least.
Mellodrama
Well-Known Member
If the drivers are warm on the outside, they're hotter inside. Don't put them in a box. Heat is the enemy of all electronic devices. Hanging vertically on a wall, probably fairly close to the ground where it's a bit cooler, would probably be OK. Hanging vertically on the wall with a small fan blowing across them would be better. If your grow area isn't too hot, maybe you could kill two birds with one stone and place the drivers downstream of a fan that's jostling the plants?