My homemade 120W LED lamp - feedback wanted!

JMD

Well-Known Member
I've been wanting to make my own LED lamp for a while, so.. now was the time!

The lamp consist of 12x 10W LEDs, of which 8 are red and 4 are blue. I went for some of the cheapest I could, and an attempt to test if they are worth anything or not, and if growing with LEDs is something I want to continue using.
I'm posting my work in order to get some feedback - so comments and suggestions are more than welcome!

2013-07-18 19.59.19.jpg 2013-07-18 19.59.29.jpg

The LEDs have been arranged with 2 red and 1 blue in series, with two of these in parallel for each "channel". They are mounted on a 300x120x20mm heatsink, with thermal grease underneath.
For driving the LEDs, I'm using two cheap boost converters bought in eBay. The LEDs will then be driven with a constant voltage instead of a constant current - the downside of this, is that the current through the LEDs are now dependent on temperature. If the temperature rises (fan failure or other), then there is a big risk of thermal runaway. On the other hand, this way of running the LEDs is MUCH cheaper.

The lamp will be supplemented by a few other lamps (CFL and 2700k LED), to broaden the overall spectrum for the plants.
Total cost of the project is roughly $65.
 
I've been wanting to make my own LED lamp for a while, so.. now was the time!

The LEDs have been arranged with 2 red and 1 blue in series, with two of these in parallel for each "channel". They are mounted on a 300x120x20mm heatsink, with thermal grease underneath.
For driving the LEDs, I'm using two cheap boost converters bought in eBay. The LEDs will then be driven with a constant voltage instead of a constant current - the downside of this, is that the current through the LEDs are now dependent on temperature. If the temperature rises (fan failure or other), then there is a big risk of thermal runaway. On the other hand, this way of running the LEDs is MUCH cheaper.


... Killer feature

go for some basics!
 
nice..... and for 65 bucks hell yeah
if you build another add a couple white leds in place of a couple of the blues
5700k if you can find one or 6500k would be good too just to add a broader spectrum of light
how are the 10 watt leds? I have been using 1 and 3's but have been intrigued :)

give us a pic with that beast running :bigjoint:
 
I'm gonna play Devil's advocate and say I think they're way too close to be running under constant voltage. And is that a gap running underneath the SMD? Bad choice IMO. You're going to have a hot strip running down the middle of that panel and your environment is gonna need to be perfect to avoid fluctuations. That's tough. And I hope those are quality die making up those SMDs cause they will only last as long as their weakest die. And once rows of lights go off 10 and 20 watters suck for growing. It's like the intensity just gets wiped away. Oh and how long have you let that run for?

But it is interesting JMD. Good Luck with it.
 
Too much blue imho, at least for a flowering light. But it has plenty of red, so who knows?

I will even it out a bit with 2x 20W 2700k CFL. But yea, time will tell how it works :)



nice..... and for 65 bucks hell yeah
if you build another add a couple white leds in place of a couple of the blues
5700k if you can find one or 6500k would be good too just to add a broader spectrum of light
how are the 10 watt leds? I have been using 1 and 3's but have been intrigued :)

give us a pic with that beast running :bigjoint:

I will definitely add a broader spectrum in my next build. I'm still in the process of learning which spectra works the best, and at the lowest cost.
My camera isn't that great, so a picture of the "beast" running at full power wasn't possible - but a low power picture is right here:

2013-07-17 22.27.54.jpg



I'm gonna play Devil's advocate and say I think they're way too close to be running under constant voltage. And is that a gap running underneath the SMD? Bad choice IMO. You're going to have a hot strip running down the middle of that panel and your environment is gonna need to be perfect to avoid fluctuations. That's tough. And I hope those are quality die making up those SMDs cause they will only last as long as their weakest die. And once rows of lights go off 10 and 20 watters suck for growing. It's like the intensity just gets wiped away. Oh and how long have you let that run for?

But it is interesting JMD. Good Luck with it.

Which gap are you referring to?

I've done a full power test for about 2 hours, and the current was stable within +/- 50 mA (measured with a Fluke 289 multimeter). Heatsink temperature was ~36ºC while the LEDs were ~48ºC - measured with a Fluke IR thermometer.
I've adjusted the boost driver's voltage to the point where I had 1800 mA in each channel - which is 900 mA in each string of 3 LEDs. They are rated at about 1050 mA, so it should be plenty room for variations.

I completely understand your skepticism - and I agree with you for the most of it. The quality of these cheap Chinese LEDs is no match for Cree or other high quality LEDs - but that's not the point! The point here is to make a really cheap setup and see how well, and if, it works. I share this process of testing with you guys, so we can all learn from it.

I might add a bi-metal temperature switch, to insure that the temperature doesn't go too high.



hey what are all the supplies needed for this project?

Blue LEDs: http://www.ebay.com/itm/360679653246
Red LEDs: http://www.ebay.com/itm/360680648327
Boost converter: http://www.ebay.com/itm/300933289543
Heatsink: http://www.ebay.com/itm/390602467555

Again, this is meant to be a really cheap setup - as you can see, it really is :)
 
Hey JMD. I was talking about that line that runs down the heat-sink and that looks like there is a slot where the LEDs are screwed on. Just like those heat-sinks from that one company whose name escapes me atm. And can you post temp readings at different hours of the day? I would love to see the readings at 6 hours and 12 hours. That's more real world in my book and I would love to see something like that since I want to get rid of my SMD's spotlight casing and start using more custom heat-sinks. Like yours :)!
 
I think in general, you want to run them at about a max of 70% current capacity.

I might have to adjust it later - but until my current grow in done, and this new lamp installed, I won't know for sure. I will keep this thread updated!



Hey JMD. I was talking about that line that runs down the heat-sink and that looks like there is a slot where the LEDs are screwed on. Just like those heat-sinks from that one company whose name escapes me atm. And can you post temp readings at different hours of the day? I would love to see the readings at 6 hours and 12 hours. That's more real world in my book and I would love to see something like that since I want to get rid of my SMD's spotlight casing and start using more custom heat-sinks. Like yours :)!

Ah, that line is merely a scratch/mark I made in the paint when measuring where to put the LEDs. It looks worse than it is, because of the contrast between the dark paint and the shiny aluminium.
The measurements I did were done at an ambient temperature of roughly 25ºC (summer here) - it will only be marginal warmer in my grow cabinet.

LED temperature is no where near alarming, and a 10ºC increase in room temperature will result in a 10ºC LED temperature (not taking increased current into account - I will compensate for this once the lamp is installed.
 
For driving the LEDs, I'm using two cheap boost converters bought in eBay. The LEDs will then be driven with a constant voltage instead of a constant current -

Great idea, can you post the link to the led drivers thanks.
 
I must agree, switch out the blue`s with some white`s preferable 6500K

will provide enough in the blue spectrum plus it will cover the full spectrum, cept much red, but that you got plenty of as I see it

what kind of Red`s are they ? 630nm or 660nm ?
 
I must agree, switch out the blue`s with some white`s preferable 6500K

will provide enough in the blue spectrum plus it will cover the full spectrum, cept much red, but that you got plenty of as I see it

what kind of Red`s are they ? 630nm or 660nm ?

I just ordered 4x 10W 6500k LEDs - I can swap them directly with the blue LEDs.

The red LEDs are 630nm, because the 660nm is quite a lot more expensive. The ones in my current setup is 630nm as well, and they work fine.
 
kool, Im sure you will be happy with the white`s and you can always make a smaller fixture with the blues and more 6500K for Veg if your happy with the way they work for you

seems like the white LEDs is the new black this year, so to speak

sure red and blue have a hugh impact of the grow of MJ, and most plants I guess, but seem`s like they need more then that to thrive, even some yellow, and som also talk about the green have a effect

the white`s can take care of that and since 6500K is high in the blue ranges and plant dont need as much of that as the red (1/3) you can do with em, wile you hit em hard with a lot of Red, specially in flowering
 
kool, Im sure you will be happy with the white`s and you can always make a smaller fixture with the blues and more 6500K for Veg if your happy with the way they work for you

seems like the white LEDs is the new black this year, so to speak

sure red and blue have a hugh impact of the grow of MJ, and most plants I guess, but seem`s like they need more then that to thrive, even some yellow, and som also talk about the green have a effect

the white`s can take care of that and since 6500K is high in the blue ranges and plant dont need as much of that as the red (1/3) you can do with em, wile you hit em hard with a lot of Red, specially in flowering

So currently it will be:
8x 10W 630nm red
4x 10W 6500k white
Supplemented by 2x20W CFL.

Would it be too much to add 100W 630nm red? I ordered one a long time ago, and it has yet to arrive.
 
What about taking out 2 0r 4 630nm and add some Warm Whites that have some output in the 630+ region? You really do want some 660 and some FR/IR. At least I would :).
 
that could maybe work as a flowering supplement to get the % of Red upin flowering

sure 12/12 tricker the hormones to start flowering but a higher amount of Red`s also have some effect here

maybe if you could make a switch of some kind and then turn it on with the 12/12 or even better 2-3 weeks in to 12/12 after the stretch it could work Im sure

you could also look in to inferred +700nm (seems like 720-740nm do the shit) a single or two 3W doing the last part of flowering will also have some effect

and if you want to do more research try to google "the Emerson effect"
 
lol franjan you beat me by .... erm .. 0 minutes :D but yes look in to 660nm and the Emerson effect

Edit: tho look at Area 51

http://area51lighting.com/the160.html

seem`s like one of the most legit companies and in front of most and there fixtures only have natural white (5000K) and Red in 630nm
 
What about taking out 2 0r 4 630nm and add some Warm Whites that have some output in the 630+ region? You really do want some 660 and some FR/IR. At least I would :).

My next lamp could be 1x 630nm + 1x 660nm + 1x 6500k for each string. Got a lot of testing to do!



that could maybe work as a flowering supplement to get the % of Red upin flowering

sure 12/12 tricker the hormones to start flowering but a higher amount of Red`s also have some effect here

maybe if you could make a switch of some kind and then turn it on with the 12/12 or even better 2-3 weeks in to 12/12 after the stretch it could work Im sure

you could also look in to inferred +700nm (seems like 720-740nm do the shit) a single or two 3W doing the last part of flowering will also have some effect

and if you want to do more research try to google "the Emerson effect"

I plan to only grow autos, so it will be 18/6 or 20/4 every day. I will look into the infrared LEDs and see what I can come up with.

What is your opinion about using UV? Got a 10W lying around here somewhere.
 
Back
Top