Building a Home Made LED

[guvoo]

Wow, I like those a lot! Nicely done. Have you had the opportunity to use them yet?

honored by the Master of DIY, hope i can handle this
i started a real world test in June.
seeds under one panel for the first 4 weeks. AF; 18/6 in soil
now there is one plant under 2 Panels(60W) and she is filling up my Growbox. plant size 45 x 50 cm height 50 cm.

 

[guvoo]

guvo, have you put a meter in series with the circuit to measure the current?

 

[repuk]

Hi!

144w for the 1.5" heat sink hits me as being a bit much. I would go with the 4" wide one, though your real power draw at the LEDs will probably be closer to 100w. It's going to take 2 of the Mean Well drivers, each maxing out at 60w per driver.
Check out the spec sheets and do the math.
http://www.rapidled.com/mean-well-lpc-60-1750-constant-current-driver/

I'm not going to buy the heatsinks for them, shipping would be too much. Gonna source the heatsinks localy or build my own using aluminum profiles.

I'm going to buy just the drivers and the LEDS plus other small parts from them.

As far as lenses go, I'm starting out with none. So that is a "we'll see". The spec sheet for the LEDs calls out a 120 degree angle, but the graph shows that they really start to drop off at about 90.

This also brings up the viability/usability of diffuse light.
IMHO, I'm thinking that we want to share the light from all of the leds, apposed to having 18 little spot lights. The cumulative advantages to having multiple HPSs' is well know.

Just say'in.

What about the X2-X3 "magnifier optics" HGL are using?

Can't wait to see pics of the build process/finished LED Lamp, measurements, etc tenthirty!!

 

[guvoo]

these are the lenses i use

http://export.rsdelivers.com/produc...0/twist-on-lens-for-osram-gd-led/7208933.aspx

 

[vukman]

Thank you so much for your sharing of this great and wondering knowledge....Big Thumbs up man!!!!!!!!

 

[vukman]

Hi!



I'm not going to buy the heatsinks for them, shipping would be too much. Gonna source the heatsinks localy or build my own using aluminum profiles.

I'm going to buy just the drivers and the LEDS plus other small parts from them.



What about the X2-X3 "magnifier optics" HGL are using?

Can't wait to see pics of the build process/finished LED Lamp, measurements, etc tenthirty!!

Hiya... I found a site called the 'Metal Supermarket'..

http://www.metalsupermarkets.com/MSC-home.aspx

^^^^^^^^^^^^^^^^^That is the Canadian site so you might have to change locations. I don't know, nor do I want to know..where anyone lives...(just safer that way).....

I know for the Canadian, it works via postal codes.. I'm guessing it's the same for US site...zip codes that is.....check it out...might be one close to you locally that might have heat sinks.. I know my local one has aluminium sheets but I have to call to see if they have 'finned' stock or know where to get a piece which would suit me needs. I know it's not necessary but would be preferable..

Good Luck

 

[rives]

View Image
View Image

these are the lenses i use

http://export.rsdelivers.com/produc...0/twist-on-lens-for-osram-gd-led/7208933.aspx

Has anyone seen these on this side of the pond? I haven't been able to find anything here wider than about 60 degrees, and these look like they would be close to ideal for my application.

Guvoo, I'm very impressed with your usage of the extrusions and the plates to build the fixture enclosure. Beautiful execution, it looks commercial-quality. Do your MCPCB stars have zeners across the power leads? I would be a bit concerned about the inductive kick from the fan motors on the led power circuit, I think.

 

[guvoo]

I would be a bit concerned about the inductive kick from the fan motors on the led power circuit, I think.
you see the diode between the Leds and the Fans?

 

[rives]

you see the diode between the Leds and the Fans?

Yes, I did. I'm kind of a belt & suspenders type of guy.

 

[guvoo]

and now the lady


a little bit younger

 

[BrnCow]

HeatSinks

http://www.heatsinkusa.com/10-080/

http://www.ledsupply.com/hs13137.php

 

[vukman]

Interesting read about the Luxeon Z chip........

Here is a quote from the article:
" The Luxeon Z is available in white as well as in a range of colors from 440 to 670 nm, using either AlGaInP or InGaN chips. It is unencapsulated, allowing the use of custom optics.
Lumileds says that the small form-factor of Luxeon Z will allow it be used to create specialized mono-color or multi-color arrays for existing and new luminaires. Up to 250 Luxeon Z LEDs can be mounted in one square inch, says the company.
"

http://ledsmagazine.com/news/9/9/6

enjoy

 

[tenthirty]

Here we are again boys and girls.
Please excuse the crappy pics form my point and shoot, but I was in a hurry.

First the tools of the trade.



A good, and I mean good soldering iron is a must. I've had this one for decades. Note the can of flux and roach clip.



A good meter is also a must. This one is at least 50 years old.

Now here we are with all of the leds mounted. I opted for the heatsink that was pre drilled because screws are better than glue for prototyping and changing out leds.



It's Alive!!!!



Note the criss cross pattern of the wiring. This is so the wire reach if a diode is removed.



Here are the 2 drivers. max output is 95w. Wish I had a kill a watt about now.



I'll have to say the assembly was a breeze with the exception of the 10w white leds dissipate heat so fast that it was really difficult to get a good solder flow. The red rebels were really easy to solder.

As far as Rapidled, the only thing that I can fault is the screw and washer set with the heat sink was short a few screws and washers, but I had replacements in stock, so no problem there.

The thing is definitely really bright. At about a foot the lateral coverage in the sweet spot should be about a foot wide. (Just what I was looking for!)

After an hour more or less the fixture is running at about 52c (125f) with no airflow in a room at 27c(82f). So with passive cooling, it should work just fine in 76f room with the standard room airflow.
Also due to the narrow profile of the heatsink we can put small fans above the luminaire and blow the warm air down on the plant, thus making up for the lack of IR in the light itself. (we call this plan B)
A 4" wide heatsink would not be as good for blow down as the 1.5" and the 4" didn't come in 36" lengths. (We'll call this rationalization)
I could call it design constraints, if I wanted to get all techie.

The size of the heatsink is right on the edge for the power and leds that are being used. You could not populate it with 18 10w leds and expect it to keep cool!!

So I spent another 30 bucks at the electronics store, but there are many extra parts left over. So there is about $230.00 in the project so far.
All that is left to buy is the mounting hardware and rope ratchets, so we'll call it at $250.00 complete at 95 power supply watts. You might say that this is pretty expensive for a DYI project,
and cheap Chinese leds can be had for much less already made.
We need to remember, that this was made with all top quality parts from the drivers on down. If a driver or led fails either can be sourced locally and replaced easily. (Try to unsolder a led from the PCB of a store bought led.)

I would say the assembly time is about 4 leisurely hour including testing time. (there were only 18 leds to solder not 200)

Coming up will be hanging, installation and the initial testing in the grow scenario.

 

[vukman]

TENTHIRTY!!!!!!!!!

Nice work man!! very nice although I'm gonna squeeze your shoes a bit and say you forgot one piece of equipment there with the individual pictures which I noticed on the table...The magnifying glass!!!!! ahahha...you old coot you!! LOL....Not to worry, I'm in the same boat and just joking around a bit.!! LOL

As for the workmanship and the care taken to tie the wires down and all that...very very nice.. I just hope I can do as well with mine when I finally get it up and going...ahhh...money..the love/hate thing in our lives...:-(
That is a very nice bit of info there about the temps and that passive will work just fine with it...very nice...also thank you for mentioning the issue with the 10W whites as well..a professional to the end...you can always tell a pro by the maintenance of his tools and the age of them....The ones of you out there in the trades and skilled work know what I'm talking about...My father was a machinist/tool&die maker..He still has his first vernier caliper which he got when he first came over from Europe back in '63.

Keep up the great work and don't forget the pics of the girls enjoying their new light source.!!

 

[tenthirty]

Vukman, thanks! the magnifying glass is just slightly helpful. I didn't show the usb microscope.
It's a must have at my age soldering itty bitty pieces.

Once the light is up and you can see what I'm working with you'll get a better understanding of where we are going with this.
If everything goes well at about 6 iterations, ought to give us "good enough for government work" results. In dollars and cents, this will cost about $1500.00, give or take.

The real fun will begin when we get to comparing spectrum's and plant response.
In my experience, nutrient manipulation has a very small influence on plant structure.
Changing ratios has very little effect on how the plant grows. (more like how much it grows)
I think light and VPD can have a much more profound affect on morphology than anything that you could do with nutrients.

A little light reading. These were posted on another forum and I'm shamelessly ripping them off.

http://plantphys.info/plant_physiology/light.shtml
http://plantphys.info/plant_biology/photopart.shtml

What I am most interested in is the porphyrin ring system and the antenna pigments to formulate experiments.

 

[repuk]

rives, vukman: I'm on the other side of the pond so shipping heatsinks is a no-no... thanks for all the info!

Tenthirty: great pics and unvaluable insights/tips!!!

I'll have to say the assembly was a breeze with the exception of the 10w white leds dissipate heat so fast that it was really difficult to get a good solder flow. The red rebels were really easy to solder.

What temp did you set your soldering station to (guess is 50W?)? That's a sign that there's good thermal transfer between the 10W and the heatsink... I know it sounds less logical, but guess it would be better to solder LEDs and wires first, then attach the LEDs to the heatsink. (I assume by the pics you first attached the LEDs to the heatsink then soldered the wiring)

I was wondering if driving them at 1750mA would be somehow wasting XM-L potential, but spec sheet it looks 2000mA being a sweet spot regarding luminous flux/life, so 1750mA little "step back" would help with heat and as heat means less lumens, would be a even sweetest spot. Was that your reasoning?

After an hour more or less the fixture is running at about 52c (125f) with no airflow in a room at 27c(82f). So with passive cooling, it should work just fine in 76f room with the standard room airflow.
Also due to the narrow profile of the heatsink we can put small fans above the luminaire and blow the warm air down on the plant, thus making up for the lack of IR in the light itself. (we call this plan B)
A 4" wide heatsink would not be as good for blow down as the 1.5" and the 4" didn't come in 36" lengths. (We'll call this rationalization)
I could call it design constraints, if I wanted to get all techie.

I already got a great heatsink for free, they're usually sold by lenght, was a leftover from my job's shop. Is 2,7in wide and 18,5in long. My flower cab is more square shaped (60x40cm) so I thought I'd better use two heatsinks like this one, with 9 XM-L each for better light distribution.

By your heatsink dimensions/watts and measured temps I hope putting just 9 XM-L in mine will render temps suitable for passive-only cooling.

The thing is definitely really bright. At about a foot the lateral coverage in the sweet spot should be about a foot wide. (Just what I was looking for!)

Don't you love it when a plan comes together? congrats!

Those are good news to me, as my cab is about 2 feet wide, using two heatsinks equally spaced, will get optimal light spread. Have yet to do some trigonometry to see if the 80 degree lenses will help or not.

keep up the good work and thanks for sharing!

 

[vukman]

Vukman, thanks! the magnifying glass is just slightly helpful. I didn't show the usb microscope.
It's a must have at my age soldering itty bitty pieces.

Once the light is up and you can see what I'm working with you'll get a better understanding of where we are going with this.
If everything goes well at about 6 iterations, ought to give us "good enough for government work" results. In dollars and cents, this will cost about $1500.00, give or take.

The real fun will begin when we get to comparing spectrum's and plant response.
In my experience, nutrient manipulation has a very small influence on plant structure.
Changing ratios has very little effect on how the plant grows. (more like how much it grows)
I think light and VPD can have a much more profound affect on morphology than anything that you could do with nutrients.

A little light reading. These were posted on another forum and I'm shamelessly ripping them off.

http://plantphys.info/plant_physiology/light.shtml
http://plantphys.info/plant_biology/photopart.shtml

What I am most interested in is the porphyrin ring system and the antenna pigments to formulate experiments.

Just too much to say thank you for so just a great big fat thank you for sharing.......especially with the reading info.. I'm always a sucker for some good ol' educational reading..

<add edit>...........amazing read!!!!!! just amazing which has led credence to something I've been talking about for a while now and that is the 680nm wave length...since the cat is out of the bag...once those chips come down in price....LED's will be a much more efficient flowering instrument.....as for now, we have to settle with the 660nm and the full spectrum whites.....which is not a bad trade off at all.....

All my Best ALWAYS!!!!!

 

[analogue]

Nice work mister .

Much better than just "close enough for guvmint work" (Had not heard that in over 20 years :lol: ).

I think that $3 a watt for a Diy led luminaire is about as good as it gets*.

Chinese companies can go right ahead and build theirs as cheap as they want - I don't care .

(edit: *assuming top of the line parts)

 

[PetFlora]

repuk, 2 of those in parallel should be tits. My starter/cloner tent is 2 X 4; the other foot has equipment in it anyway Hope you will keep me posted. Do you plan on starting a thread?

Anybody else unable to load the pics? The wheel just goes round and round

 

[tenthirty]

Repuk,

I'm to lazy to look, but I would bet that the soldering station is at least 50w.
Yes, it would be easier to solder them off of the HS, but I'm again lazy, and that would require a bunch of basic mockup and measuring. I was pretty sure that I could solder them on the HS.
What I had to do is solder the 10w diodes in layers. The fast recovery and power of the soldering station allows you to do that.
This isn't my first rodeo.

Now everything is a compromise isn't it. I don't want to reinvent the wheel, nor do I have the time. We have to remember this is prototyping.
I need something to work within the design parameters quickly and that is easily modifiable. This is not a build it and go kind of project.

The way I figured everything out is I started with the heat sink, did the calcs to determine the heat dissipation characteristics,
Read the datasheets and selected diodes that would be of a suitable wave length and finally selected the drivers based on the diodes.
Just like a walk in the park.
Ya, 1750ma is on the low side, but run what ya brung!
How long would it take to source and order all the parts and put together a reasonably good cc ps. (counterproductive)

By your heatsink dimensions/watts and measured temps I hope putting just 9 XM-L in mine will render temps suitable for passive-only cooling.

At nine XM-L's it's going to be close, maybe 8.