diy version of low profile led propagation fixture for small area?

Randomblame

Randomblame

Well-Known Member
Because 470/520μMol/s is a bit low for flowering C. you should consider to buy a third XP-G3 stripe and a HLG-80H-C700B(129v/700mA). This would increase the total light output to Ø 780μMol/s/0,28m² (- 10% wall losses).

Driver and stripes fit's perfect!

2,83v x 45 = 127,35v x 0,7A = 89,15w x 54% eff. = 48,15PAR/w and 41w heat.
Lighting calculations:
48,15PAR/w x 4,52μMol/j(for 4.000°k spectrum) = 217,5μMol/s PPF : 0,28m²(3sft) = 777μMol/s/0,28m² - 10% wall losses = 700μMol PPFD
Heatsink calculations:
For 41w of heat you need at least a heatsink with 4510cm²(41x 110cm²), better take 5000cm².

A 3mm thick aluminium sheet with 88x 28cm has 2464cm² x2(because we talk about total surface area) = 4928cm². And the sheet should be okay to perfectly fit in your 3' x 1' area.
 

Attachments

bri77

bri77

Well-Known Member
Randomblame said:
Because 470/520μMol/s is a bit low for flowering C. you should consider to buy a third XP-G3 stripe and a HLG-80H-C700B(129v/700mA). This would increase the total light output to Ø 780μMol/s/0,28m² (- 10% wall losses).

Driver and stripes fit's perfect!

2,83v x 45 = 127,35v x 0,7A = 89,15w x 54% eff. = 48,15PAR/w and 41w heat.
Lighting calculations:
48,15PAR/w x 4,52μMol/j(for 4.000°k spectrum) = 217,5μMol/s PPF : 0,28m²(3sft) = 777μMol/s/0,28m² - 10% wall losses = 700μMol PPFD
Heatsink calculations:
For 41w of heat you need at least a heatsink with 4510cm²(41x 110cm²), better take 5000cm².

A 3mm thick aluminium sheet with 88x 28cm has 2464cm² x2(because we talk about total surface area) = 4928cm². And the sheet should be okay to perfectly fit in your 3' x 1' area.
Click to expand...
I only want to veg under this light, so I should be fine right?. Im really interested in your ppf calculations.I have all the inputs except 54% efficiency and uMol/j for 4k spectrum, I cant see them on the datasheet?

And the 110-120cm² surface per watt for passive cooling, is that a general rule of thumb I can use for all leds?
 
Airwalker16

Airwalker16

Well-Known Member
bri77 said:
I only want to veg under this light, so I should be fine right?. Im really interested in your ppf calculations.I have all the inputs except 54% efficiency and uMol/j for 4k spectrum, I cant see them on the datasheet?

And the 110-120cm² surface per watt for passive cooling, is that a general rule of thumb I can use for all leds?
Click to expand...
I think so. Don't quote me though.
 
Randomblame

Randomblame

Well-Known Member
bri77 said:
I only want to veg under this light, so I should be fine right?. Im really interested in your ppf calculations.I have all the inputs except 54% efficiency and uMol/j for 4k spectrum, I cant see them on the datasheet?

And the 110-120cm² surface per watt for passive cooling, is that a general rule of thumb I can use for all leds?
Click to expand...

For veging 2 stripes are more than enough, you are right.
μMol/j per kelvin spectrum/color please take from the images and efficiency is only estimated from lumens/w, 175lm/w ~ 52-55% depending on LER and other factors.
There are a few threads in LED section about LED- and COB-efficiency. Please search a bit, because it's a bit more than I can remember currently, Lol!

Edit!
Have a look here, my friend!
https://www.rollitup.org/t/math-behind.868988/
 

Attachments

Last edited:
bri77

bri77

Well-Known Member
Thanks very much for you help RB. Thanks for the link, very informative. I read that thread and happy to know someone else has worked it out treat as a black box.
I'm eager now to get my hands on some components and start trying things out.

Actually one question about that maths thread. The efficiency figures (umole/j) for each cob seem huge, they're theoretical maximums right? Is there a rule of thumb we can used to estimate what we might get in practice. I'm guessing we might get 50% of the max maybe? .
Man, going diy for leds is a nobraner isn't it? You can implement the latest technology with minimal risk way before commercial companies.
 
JorgeGonzales

JorgeGonzales

Well-Known Member
@bri77 it's a good rule of thumb for all LEDs. It gets a little trickier if you build a cob light with individual radial or pin heatsinks, but it works out right in the ballpark for bare extruded aluminum heatsinks like they sell at heatsinkusa.com.

When I say trickier, I mean pin and radial heatsinks tend to be better at their job, so you might oversize if you went by that rule of thumb. You'll also see manufacturers spec heatsinks by C/W, which is how many degrees celcius per watt of dissipated heat. Lets say your ambient temperature was 25C, and you wanted to keep your LED under 50C. A 2C/W heatsink would be good for a 50% efficient COB running at 25W ie 12.5W of dissipated heat. It's not quite as linear as that, since heatsinks work better the hotter they get because natural convection increases, but with a napkin and some overkill, it's pretty easy.

Also, LED spacing matters, the farther away from the heatsource, the less effective that chunk of heatsink will be, if that makes sense. The most cooling goes on closest to the source. See here for a good conversion chart for a single LED:
http://www.aavid.com/thermal-tools/length-correct

Of course the opposite is also true, the more evenly spread out your heatsources are, like in a strip, the better a heatsink cools. As you can see it's as much artful overkill as science.
 
bri77

bri77

Well-Known Member
"A 2C/W heatsink would be good for a 50% efficient COB running at 25W ie 12.5W of dissipated heat."

The other 12.5w of heat just stays in the fixture and slowly radiates away? We can increase the heatsink c/w but quickly hit diminishing returns. We can also improve things by removing heat from the sync faster.

So in your example, by maxing out our useful heatsink and providing good airflow, we might get another 8 watts of heat into the sink .Fixture temps would then be at 4.5w * 2C = 9c + 25c.= 34C
 
JorgeGonzales

JorgeGonzales

Well-Known Member
bri77 said:
"A 2C/W heatsink would be good for a 50% efficient COB running at 25W ie 12.5W of dissipated heat."

The other 12.5w of heat just stays in the fixture and slowly radiates away? We can increase the heatsink c/w but quickly hit diminishing returns. We can also improve things by removing heat from the sync faster.

So in your example, by maxing out our useful heatsink and providing good airflow, we might get another 8 watts of heat into the sink .Fixture temps would then be at 4.5w * 2C = 9c + 25c.= 34C
Click to expand...
The other 12.5W is light :)
 
mauricem00

mauricem00

Well-Known Member
Airwalker16 said:
And are you saying that you're powering cobs with these $3 parts instead of using meanwell drivers and the like? Is it safe?
Click to expand...
these cheap transformerless power supplies have been used for decades and are found in many consumer electronics.it is as safe as using a resistor to limit current to a led but by using reactant instead of resistance no power is dissipated by the current limiter.it is actually safer than conventional drivers because the power transistors those drivers use can fuse and cause a short or even blow up. like the solid state detonators use in automotive airbags. I would suggest using motor run capacitor instead of the motor start capacitors shown because motor run capacitor are designed for continuous operation and motor start capacitors are not.I am a T5 grower but have a background in electronics and like playing with LEDs.
 
bri77

bri77

Well-Known Member
mauricem00 said:
these cheap transformerless power supplies have been used for decades and are found in many consumer electronics.it is as safe as using a resistor to limit current to a led but by using reactant instead of resistance no power is dissipated by the current limiter.it is actually safer than conventional drivers because the power transistors those drivers use can fuse and cause a short or even blow up. like the solid state detonators use in automotive airbags. I would suggest using motor run capacitor instead of the motor start capacitors shown because motor run capacitor are designed for continuous operation and motor start capacitors are not.I am a T5 grower but have a background in electronics and like playing with LEDs.
Click to expand...
Could you put a shopping list together and explain to a beginner how to make it? I know thats not a trival thing to ask , Ill understand if you pass.
 
mauricem00

mauricem00

Well-Known Member
bri77 said:
Could you put a shopping list together and explain to a beginner how to make it? I know thats not a trival thing to ask , Ill understand if you pass.
Click to expand...
what is the voltage and current of your led? what is you AC line voltage? subtract led voltage from line voltage and divide that by LED current. that will give you the value of reactance you need for the series capacitor. use this formula 1/ reactance * line frequency* 2pie to get the value you need. example. im running 20 3 watt leds in a string (64 volts) of a 120 volt 60 cycle line at 300 ma so 120 minus 64 (56 volts) divided by 300ma gives me a reactance of 187 ohms. times 2 pie (6.283185307) times 60 cps gives me 70497.33915 divide 1.0 by this number gives me .000014185 farads or 14.185 micro farads.a 15uf capacitor will work fine if a 14uf can not be found connect this capacitor in series on one side of your AC line feeding a bridge rectifier and place an electrolytic capacitor rated at 200 volts DC or more across the DC output of your bridge rectifier and connect your led to that DC output as well.use 400 volt polypropylene capacitors like motor run capacitors or speaker crossover capacitors and your all set ( I know but my computer keyboard does not have a symbol for pie)
 
You must log in or register to reply here.
Top