Epistar/BridgeLux/noname: A red LED Under The Microscope - Real and Electronically

I just test shopped for water cooling equipment: -this is what I would require and is selecting the cheapest items possible.
The price seams a little costly, but I guess it all depends on what your running now to keep cool,
if your using electricity on a AC unit then then power saving will help the pay back time.


WaterCooling.jpg
 
And I've just run some test on the 4 chip mounted on the black 1U alu tube,
and to my surprise the box was not that good comparing to the cpu cooler with has a much larger surface area even if its more compact.


[TABLE="width: 500"]
[TR]
[TD]
10_DIY_20.jpg
[/TD]
[TD]
10_DIY_21.jpg
[/TD]
[/TR]
[TR]
[TD]Using screws from the computer screw box - with build in spacer.[/TD]
[TD]The center holes maybe not the best position.[/TD]
[/TR]
[/TABLE]
10_DIY_22.jpg

10_DIY_23.jpg


After 7 minutes the temperature stabilizing at around 36°C/96.8°F

Before I run the test I thought that I easily could fit more chips on the tube, but now I'm not so sure,
maybe more fan holes would do the trick, or else it's simply the fact that blowing air on many tiny fins
is much better than a smooth surface.
 
Excellent posts Rasser. Thank you so much for all this info and insight! Seems I'm unable to LIKE at the moment but I assure you I do :). I'm surprised the 4x10 Reds didn't run hotter on that CPU cooler, which is great news. I guess the efficiency of reds and not having to go through any phosphor converting process keeps them running cooler than a similar white LED. And while I like your DIY water cooling, that's pretty extreme for me at the moment. One little leak that you don't notice and you can kiss your LEDs goodbye. I could never forgive myself if I screw up that big. Especially since I want to overdrive the LEDs ultimately, so commercial products for now. Still it will probably save a person a ton of money going with DIY water cooling. Right now it seems if I want to go with water cooling for some kind of panel, I have to find the coolers and build around them, which I don't think is the best way to design a light, but the right parts for the right price just might be out there. Still there's no reason for me to buy a LED spotlight again after this thread :), so thanks again Rasser, you little LED revolutionary you :).
 
A salesman at the Cidly factory that manufactures the Apollo LED lights tells me that they use EpiLED. I believe him as he is a friend way before he is a salesman.
 
FranJan said:
Excellent posts Rasser. Thank you so much for all this info and insight! Seems I'm unable to LIKE at the moment but I assure you I do :). I'm surprised the 4x10 Reds didn't run hotter on that CPU cooler, which is great news. I guess the efficiency of reds and not having to go through any phosphor converting process keeps them running cooler than a similar white LED. And while I like your DIY water cooling, that's pretty extreme for me at the moment. One little leak that you don't notice and you can kiss your LEDs goodbye. I could never forgive myself if I screw up that big. Especially since I want to overdrive the LEDs ultimately, so commercial products for now. Still it will probably save a person a ton of money going with DIY water cooling. Right now it seems if I want to go with water cooling for some kind of panel, I have to find the coolers and build around them, which I don't think is the best way to design a light, but the right parts for the right price just might be out there. Still there's no reason for me to buy a LED spotlight again after this thread :), so thanks again Rasser, you little LED revolutionary you :).
Click to expand...

Thank you , yes water cooling is not something you fit to a UFO or an Apollo I think, if you start from scratch and plan accordingly then
maybe one can justified spending that much on cooling.

All modern CPU have overheat protection, can't see why a bi-metal cutoff switch at 50-60°C won't work.
or are you thinking about the water and the chip?

I'm thinking about the cpu cooler's great performance, it's surface area vs. the tubes:
the cpu cooler size is (seen from the side) 7x3cm x 38 Fins x 2 sides - 7x3= 21cm2 * 38 * 2 = 1596 cm2 or 1,6 m2 - 17.2 ft²
and there is a strong air flow blowing down on all the fins.

Sure why not :bigjoint:

[video=youtube;40jK4WeCl_g]http://www.youtube.com/watch?v=40jK4WeCl_g[/video]
 
Along with the heatsinks I also bought 10 red and 10 blue Osram LED's since they cost about 2$ and seams to be able handle 2W,
to play with, and see if i can 1. avoid re-flow soldering and 2. compare the output to the no-name LED's PAR value for watts used.

Incredible efficiency 44% at 100mA !

A composition images of the data from the data sheet and a shot through the microscope of the two very tiny chips:
Osram_Golden_Dragon_450nm.jpg


Osram_Golden_Dragon_660nm.jpg
 
I just tried to measure the temperature on the back side of the Osram hyper red LED while it was floating without a sink, i stopped at 55°C 131°F @ 250mA
I thought that it would run way hotter than that, okay there is a bit of heat transfer to the sensor via the thermal paste and maybe to the test wires,
but it nice to get a feeling of the level, so its not like it will burn off instantly without being mounted to a heatsink when running below 250mA.

[TABLE="width: 500"]
[TR]
[TD]
Osram_Golden_Dragon_450nm_freerunning_01.jpg
[/TD]
[TD]
Osram_Golden_Dragon_450nm_freerunning_02.jpg
[/TD]
[/TR]
[TR]
[TD][/TD]
[TD][/TD]
[/TR]
[/TABLE]
There was about 500-700 uMol(full sunshine is 1600-1900) an inch from the LED when measuring with the PAR meter, while running at 250mA.
 
wonderful work rasser......so when are you going to market your own led panel?....ha but i'm telling you if you could copy rhinogrow led panel but switch out their driver for a dimmable one and run all osrams= blue/red/cool white with a heavy red emphasis ....id buy one:-)
 
Why not CREE (or equivalent) Neutral Whites with a few 660s on a separate switch for mid flower- harvest?
 
^^^Don't forget to get some FR action going too.

And +1 to Neutral Whites, but PetFlora, no 630s? :o LOL
 
FranJan said:
^^^Don't forget to get some FR action going too.

And +1 to Neutral Whites, but PetFlora, no 630s :o. LOL
Click to expand...

Not sure I understand. If you are asking a question "No 630s?" my answer is the Neutral White Crees cover them really well, and probably 660, too. So maybe the bump should be 680-740 on a separate plug/switch/timer
 
Rasser said:
What a piece of shit,,,
10W_UnderTheMicroscope_11.jpg
Click to expand...


The UK-Ebay dealer of the chips said that it might need some more voltage before all the LEDs would light up,
and thats normally ~true when operating with very low voltage, but I was not, the chip was drawing 250mA,
and all other chips worked right from the start.

So I put the LED back on the test bench connected to my LAP psu, and i increasing the voltage until the amp meter was reading 300mA,
and again only a single diode was alight in the 10W chip, but then i increased it to 700mA, then two more diodes turned on,
lowering the voltage again so the chip drew 300mA now all thee diodes was alight, I then chocked the chip with pulses of 1000mA and got everything working.

Now all the diodes in the chip is lighting equally, all the way down to 10mA.

So that was great I was sure that the chip was short circuit internally, but it's properly just skipped some factory burn in process.
 
PSUAGRO. said:
wonderful work rasser......so when are you going to market your own led panel?....ha but i'm telling you if you could copy rhinogrow led panel but switch out their driver for a dimmable one and run all osrams= blue/red/cool white with a heavy red emphasis ....id buy one:-)
Click to expand...

Thanks, I'm just having fun learning and teaching, and in DIY LED grow lights all my yearlong LED fetish,
interest in electronics-Light-science-energy savings, cannabis consumption and online community have finally rolled into one,
so I'm not looking to make money from this.

I Looked at rhinogrows led panel for 5 sec. they looked old like my 120W or maybe worse(reflow soldering) what am I missing ?
 
The rhino grows looks eerily similar to the Haight Solid States which perform pretty terribly in comparison to other LED grow lights. At least Rhino got rid of the frosted plastic in front of the LEDs.

Congrats on getting your red array working Rasser. I'm very tempted to throw a few of those between my white arrays.
 
Rasser said:
Hi, If we are talking about this panel, then it's hard to see anything,
but if it's reflow soldered, then changing a single faulty LED is a nightmare,
and it uses no lenses by the looks of it.


http://shop.rhinogrow.com/Grow-Lights_c2.htm

For 650$ I think one of these at 600$ will beat the shit out of it.


What am I missing ?
Click to expand...
Aren't all panels using reflow soldering? I know hans' panel does, and plantphotonics' as well. I wasn't aware of any other alternatives to stick SMD LEDs to a thermal interface/MCPCB.
As for the lens, the LEDs are already focused to 80° so lenses should not be necessary. Plus a lens is never 100% transparent, sometimes as much as 15% of the light is lost in the lens.
What you're missing though is that it uses the most efficient LEDs on the market so I guess it should be much better than the Apollo. I've personally tested none of those but from the technical specs the rhinogrow seems definitely a killer. I would also add that the LEDs are spread over a much larger surface with a massive heatsink which seem to allow passive cooling which I think pretty neat.

As for the price, of course it is expensive. But it's made from quality components and the assembly seems great so it kinda justifies the price. I think it is can be put in comparison with plantphotonics and hans panels (both 3 using very high quality components) which are in the same kind of prices.
 
PetFlora said:
Not sure I understand. If you are asking a question "No 630s?" my answer is the Neutral White Crees cover them really well, and probably 660, too. So maybe the bump should be 680-740 on a separate plug/switch/timer
Click to expand...

Hey PetFlora I'm just playing with you a bit since I've seen you be a big advocate for 630nm in some conversations. I'm more of a 660nm person so your 660nm suggestion made me chuckle a bit, so I decided to have a little fun, that's all. And neutral white or not, I would have to see the data sheet on the LED if I was depending on a white LED to cover or assist with specific spectrums, especially if it's a phosphor based white diode which can be all over the RGB (and Y) spectrum it seems. Something like a Neutral White Cree XR-E is terrible for 630-660nm, though I would assume there are better Cree models to choose from. Peace and Happy Growing PF!
 
The rhino grow controls current regulation to the various different diodes via resistors on the PCB correct? How much of a difference does controlling current with the driver make versus controlling it at the diode with resistors? I've seen many people using resistors instead of constant current drivers. Any chance you could break down the pros and cons in layman terms between the two for me Rasser?
 
You must log in or register to reply here.
Back
Top