Looking for a specific DC CC driver for LEDs

yeah, the light emitting side gets darn hot on those at close range, they say LEDs are cool to the touch but not with that many photons in such a small space! and even with the current generations of "3watt" chips that is really a myth, even those could scorch paper and will melt plastic at close range under the right conditions.

In this case, its probably around 30W of radiant energy, getting swallowed by a slab of metal. Of course it will get warm :fun4:
 
In this case, its probably around 30W of radiant energy, getting swallowed by a slab of metal. Of course it will get warm :fun4:

30% loss w/ that led is pretty optimistic. Probably south of 50%.. So 50w of heat if not a lot more. Basing this on 34 V @ 3A.... 102W
 
Was going for 30% efficient (which is then absorbed by my old ATA hard drive and back to heat). The rest is all direct heat :)
 
The LED itself won't put out radiant heat (IR). The semiconductor substrate carrying the current will get hot as a function of the diode drop, and that's all in the metal. If you keep the metal back cool, you can run any LED.

100W will burn without heat sinking. I use a little plate and very fast fan to keep it cool during test.

The real system (2000W) will use these babies in a chilled liquid cooled configuration. More on that later :)

<a href="http://s1062.photobucket.com/user/karimwassef/media/393F1028-EFD8-43BB-9515-447EAF9FED18_zpsto6ybiwa.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/393F1028-EFD8-43BB-9515-447EAF9FED18_zpsto6ybiwa.jpg" border="0" alt=" photo 393F1028-EFD8-43BB-9515-447EAF9FED18_zpsto6ybiwa.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/F5E5ED68-9D60-4C22-B6A3-AF957A459046_zpsbjfknmhg.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/F5E5ED68-9D60-4C22-B6A3-AF957A459046_zpsbjfknmhg.jpg" border="0" alt=" photo F5E5ED68-9D60-4C22-B6A3-AF957A459046_zpsbjfknmhg.jpg"/></a>
 
No IR, but still radiant energy, just at a smaller wavelength.

Initial testing shows about 93% efficiency (error margin is still not good as I haven't accounted for all the losses offboard) from a 38->34V 3A run. The low rent test fixture. Fan helps :)

0eb32030f5c8f2c9b22ac7f219df0a9a.jpg
 
Weight, and something that will block light and not melt. Didn't feel like drilling and tapping for this quick setup
 
Indeed, not IR but near UV, gets really hot at close range, you'd burn your hand holding it close under a 100w chip for more than a few seconds. This sounds silly but it will easily melt an acrylic splash guard as soon as the acrylic gets a little dirty and starts heating up. photons are photons and once absorbed by a solid surrface they turn into heat.
 
Weight, and something that will block light and not melt. Didn't feel like drilling and tapping for this quick setup
LO after you corrected me on the efficiency, and I finally understood what you meant I then pictured your LED "sandwich"..

I got a kick out of it..picture is as I pictured it..
 
LO after you corrected me on the efficiency, and I finally understood what you meant I then pictured your LED "sandwich"..



I got a kick out of it..picture is as I pictured it..


Should've led with the picture, would have reduced confusion :)

I have a very expensive heater! I'm also extra cautious about looking at the output of blue LEDs at high power levels. The studies are mixed, but I've seen the effects. http://www.cree.com/~/media/Files/C...p/XLamp Application Notes/XLamp_EyeSafety.pdf
 
Ok. I'm finally getting around to it... well, some of it.

I got the cheap LED ($5)

http://www.ebay.com/itm/271974328722

hooked up to an old cheap LED driver (AC in, no PWM or dimming) I already had from an older project ($25). Here's one comparable to it - the one I have is too old to find.

http://www.amazon.com/High-Power-Light-Output-30V-36V/dp/B00DI3N04E

I use a very cheap CPU fan/heatsink that comes with thermal epoxy for $10 delivered.

http://www.amazon.com/gp/product/B004PLXV8S

The purpose is to establish a baseline. I'm going to measure current, voltage, and PAR with this setup.

For now, I'm going to let it run for 24hrs (call it a short burn-in), just to make sure I'm getting real results for a real system.

First, this cheap power supply doesn't turn on immediately. It pulses on and off for a while. The colder the LED or PS, the longer it takes. I assume this is a cheap current limit circuit or a cheap constant current sense circuit.

Here are the pictures so far:

<a href="http://s1062.photobucket.com/user/karimwassef/media/4DAA8154-E33F-4BA0-B07F-5B5E5659B545_zps4tdzsi1e.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/4DAA8154-E33F-4BA0-B07F-5B5E5659B545_zps4tdzsi1e.jpg" border="0" alt=" photo 4DAA8154-E33F-4BA0-B07F-5B5E5659B545_zps4tdzsi1e.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/AF075EE7-C7FA-4C15-A7E4-EDFBA0BBACDB_zpsubsstamv.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/AF075EE7-C7FA-4C15-A7E4-EDFBA0BBACDB_zpsubsstamv.jpg" border="0" alt=" photo AF075EE7-C7FA-4C15-A7E4-EDFBA0BBACDB_zpsubsstamv.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/59F75E57-D63B-41BB-B69C-A15713BD4D06_zpsr8sjjgoh.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/59F75E57-D63B-41BB-B69C-A15713BD4D06_zpsr8sjjgoh.jpg" border="0" alt=" photo 59F75E57-D63B-41BB-B69C-A15713BD4D06_zpsr8sjjgoh.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/5E7161A2-885A-428C-9E77-8B771AD9C44B_zpsmwudzt77.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/5E7161A2-885A-428C-9E77-8B771AD9C44B_zpsmwudzt77.jpg" border="0" alt=" photo 5E7161A2-885A-428C-9E77-8B771AD9C44B_zpsmwudzt77.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/0e11fd2c-30e7-4593-83a4-2e54f0a36825_zps9xqkae1g.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0e11fd2c-30e7-4593-83a4-2e54f0a36825_zps9xqkae1g.jpg" border="0" alt=" photo 0e11fd2c-30e7-4593-83a4-2e54f0a36825_zps9xqkae1g.jpg"/></a>

I couldn't look directly to see the pixels when lit, but I could see it in the reflection from my tank's glass. The camera didn't catch it though.

<a href="http://s1062.photobucket.com/user/karimwassef/media/5DD1B881-26B9-4860-9454-A90B22D5A1B8_zpsydzvzm6u.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/5DD1B881-26B9-4860-9454-A90B22D5A1B8_zpsydzvzm6u.jpg" border="0" alt=" photo 5DD1B881-26B9-4860-9454-A90B22D5A1B8_zpsydzvzm6u.jpg"/></a>

and because I hate wasting light, I put it up over my quarantine/chaeto growout tank while running burn-in.

<a href="http://s1062.photobucket.com/user/karimwassef/media/79A116B7-E601-46D7-9CCD-9B5705CC60ED_zpst0fhfan7.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/79A116B7-E601-46D7-9CCD-9B5705CC60ED_zpst0fhfan7.jpg" border="0" alt=" photo 79A116B7-E601-46D7-9CCD-9B5705CC60ED_zpst0fhfan7.jpg"/></a>

And the video

<iframe width="560" height="315" src="https://www.youtube.com/embed/PKyFhVVyoAI" frameborder="0" allowfullscreen></iframe>
 
Must be hitting voltage limit. I can't put a current probe and get it to stop flashing. I checked just voltage with a short direct connection and it levels off at 39.8V

I'm assuming it's got a Vo Max set at 40V with a reset... Hence the flashing.

That is a lot more V than I was expecting.
 
High voltage/low efficiency bin dioeds that no one else wants end up in these cheapos, looks like you maybe got a lot of 4+ V diodes, judging from how the strings don't power down even close to evenly across the board. that string that takes the longest to power off is also getting the most current. it appears the binning was not very consistent on this chip, or possibly the mixed batches of LEDs from two different runs of diodes?

Got a 48v adjustable supply handy you could hook up and dial down below 40V, working back up, and see what you get?
 
The same scenario happened with my cheapo 100W RB model. Binning is off so its a 34Vf total LED, though my strings are more balanced.
 
Still hopping bridgelux, Cree, or someone, comes out with a royal blue with good binning in royal blue COB array. High quality and high power blue COBs are lacking in the availability right now, I expect that the horticulture industry will start to change that soon though!
 
I made the video to show the uneven distribution that I could only catch when I turn it off.

Most of the multichips I find are Bridgelux - but they may be excess stock or out of spec.

But given the difference in cost - does it really matter?
 
technically, it's an oxidization layer... rust is specific to iron :D

I scrape it off, heat and solder on the tin layer underneath and it's all good.

For industrial use, the chips I get are considered scrap. But for a DIY, I think it's a good deal.
 
technically, it's an oxidization layer... rust is specific to iron :D

I scrape it off, heat and solder on the tin layer underneath and it's all good.

For industrial use, the chips I get are considered scrap. But for a DIY, I think it's a good deal.

Fine I suppose, the main possible issue is just in longevity when pushed near the max rated current of the chips, consider that those are probably only 350mA chips and when pushed up to 3amps that one or two lower voltage strings are getting perhaps over their rated current and may fail prematurely. With adequate cooling it may never be an issue, just something to be aware of.
 
Back
Top