LED Design

Man,

I read the whole thing. You are rocking the details!

Love the work and thanks for posting along the way.

Looks like you have it licked and almost there

My public edgumukasion made me squint abit at the specs.

Very Cool.
 
Son of Adam;737704 wrote: Man,

I read the whole thing. You are rocking the details!

Love the work and thanks for posting along the way.

Looks like you have it licked and almost there

My public edgumukasion made me squint abit at the specs.

Very Cool.

Thanks, considering I changed the entire design about 50 posts in... :)
 
Here's a test-fit of one of the mounting bars, drilled for 8-32 screws on the ends, and drilled for 6-32 screws in the middle.

The screws on the ends go through the bars down into the acrylic supports on the sides, which are drilled and tapped for 8-32 threads.

I had to use special "wire gauge" drill bits to give the screws enough room to pass through the bars - I used a #16 for the 8-32 screw holes, and a #25 for the 6-32 screw holes.


The bar fits perfectly, with a little wiggle room.

Now that I've done one bar, I have 11 more to go... :)

9ivQT.jpg
>
9ivQTl.jpg
alt="" /></a>
 
Finished drilling the holes in the mounting bars - 12 bars, 4 holes in each - that's a lot of drilling...


Here's two of the bars, test fit with the heatsink in between.

Screws will go through the sides of the bars into the heatsink, which is resting in between the two bars.


SbHgd.jpg
>
SbHgdl.jpg
alt="" /></a>


Here's a shot with the top on.

The top of the fan housing is just about level with the top of the cover.

The hole in the cover is a little larger than I originally designed for, but that's OK.

R1cVj.jpg
>
R1cVjl.jpg
alt="" /></a>
 
You want to Build me one to??


Amazing Detail which has been said already! I also agree this is VERY professional looking and I bet the Results will be Awesome!
 
PFCDeitz;737971 wrote: You want to Build me one too??


Amazing Detail which has been said already! I also agree this is VERY professional looking and I bet the Results will be Awesome!

I might!

The most tricky part is the acrylic box - I did the layout and design in Sketchup, but since I don't have a CNC Router (yet!) I had the MRC guys make the box for me.


The design has a lot of potential, and is semi-modular since you can get the same LED forms in 50W, 100W, 10,000K / 20,000K, 455nm Royal Blue - and even LEDs with combinations of the above, all in one chip.


Here's me laying out the mounting points for the LED on the bottom of the heatsink.

The piece of paper is a 1:1 print out from the Sketchup design, and all I have to do is use the center-punch tool on X marks the spot.

MLAni.jpg
>
MLAnil.jpg
alt="" /></a>
 
GiulianoM;737657 wrote: Cutting jig on the miter saw.

The piece of aluminum is sandwiched between strips of wood, then clamped to the fence.

naCq0.jpg
>
naCq0l.jpg
alt="" /></a>


The cutting blade - a 10" Diablo 80 Tooth (not 40) - spraying a little WD40 on the blade before the cut makes it much easier to cut through the aluminum - the lubrication is essential for metals.

vSBeD.jpg
>
vSBeDl.jpg
alt="" /></a>


Here's the mounting bars in place for a test fit.

They'll be marked and drilled for holes next, along with the heatsinks.

pQxi5.jpg
>
pQxi5l.jpg
alt="" /></a>

Edit:

Yes, there would be far too much current for 2 50W LEDs.

You'd need 3.4-3.5A to drive 2 LEDs at 1700-1750mA, at 30-36V DC.


You'll want a 120W, a HLG-120H-36(B) - 36V max, 3.4A constant current.

http://www.meanwell.com/search/HLG-120H/default.htm">http://www.meanwell.com/search/HLG-120H/default.htm</a>


The ones I got are 185W, driving 3 LEDs.

They're only 50W if driven at 30V, technically it's 60W at 36V.

It's the constant current range that's important, the voltage range is secondary (but still important).[/QUOTE]

Please let us know how this works out when your finished testing. I ordered enough material to perform some testing myself on my system.

I decided to go with the meanwell HLG-240H-36B driver so I can drive (4) of the 50 watters in parallel. I already received my Molex and heatsink/fans; I should be getting my Leds and Driver real soon. I'm real interested in you sharing your results, including the spectrum testing your doing with the 50watters.

After the preliminary testing we did at Jon's house a couple of weeks ago, I'm convinced these 50 watters are the way to go. My only concern is the overall spectrum. I truly believe the issues with all these LED fixtures that people are buying all over the country is the spectrum is not "as advertised". With the right combination of LEDs, I believe you can achieve the proper spectrum, good par values and achieve good steady growth in a SPS tank.
 
So do you know anyone with a PAR meter so you can check the levels when you are done. hehehehehehehe :)
 
I haven't measured the full spectrum yet, but I suspect I may supplement the light with some 400nm UV leds. Can be had on eBay for about $5 per 1w.
 
Marking the holes for the sides - these will be drilled out with a #30 bit, and tapped for 6-32 threads.

The center points between the two circles are exactly 2.000" apart (yes, to 3 places).


Digital calipers are such useful measuring tools.

wGxtT.jpg
>
wGxtTl.jpg
alt="" /></a>
 
Here's one of the heatsinks (fan removed), attached to the mounting bars with a LED chip underneath.


g8ZYH.jpg
>
g8ZYHl.jpg
alt="" /></a>

Here's the LED chip undernath - those tiny dots in the middle are the LEDs themselves. This is one of the 20,000K ones I originally got - it's a bit tarnished but otherwise fine for a test fit.

vp26a.jpg
>
vp26al.jpg
alt="" /></a>


And here's a shot with the plastic reflector in place.

The center is whited out because of the camera flash reflecting on the reflector.

The reflector is held in place by the pressure of the aluminum heatsink mounting above it. The lens would also go in between the reflector and the box bottom. There's a lip routed in the acrylic to hold the reflector in place.

Z9s91.jpg
>
Z9s91l.jpg
alt="" /></a>
 
Keep us posted on the spectral chart. I have some KZ Fiji Purple's ready to go as a potential supplement if needed (big spike in 430 - 440 nm - for chlorophyll a). Also, got them to enhance reds.

Fiji Purple
http://atlantareefclub.org/forums/attachment.php?attachmentid=35329&d=1329446572" alt="" />
<fieldset class="gc-fieldset">
<legend> Attached files </legend> [IMG]http://atlantareefclub.org/boards/data/uploads/attachments/739704=35329-giesemann fiji purple.jpg>
739704=35329-giesemann fiji purple.jpg
class="gc-images" title="giesemann fiji purple.jpg[/IMG] style="max-width:300px" /></a> </fieldset>
 
Yep, been getting some spectrum diffraction in early tests, the light splitting into a wide color band.

However there's too much light, and the sensor is whiting out in places.

Going to make an opaque tube to channel the light through from a farther distance, maybe reduce the intensity via distance.
 
Prepping another heatsink for drilling / tapping.

The heatsinks come with white thermal paste on the bottom - a little Goo Gone gets rid of it in minutes.

Before
6iAzL.jpg
>
6iAzLl.jpg
alt="" /></a>

After
TRZLU.jpg
>
TRZLUl.jpg
alt="" /></a>

Removing the fan, you see the heatsink retaining clip which needs to be removed.

It's just set in place, easy to pull out.

MydBQ.jpg
>
MydBQl.jpg
alt="" /></a>

00VQb.jpg
>
00VQbl.jpg
alt="" /></a>


Also an interesting note about the fan - they're a standard size fan, 90/92mm in size. They can be replaced with a $5 case fan if needed.
 
I was finishing up drilling and tapping the 8 holes I needed in the heatsink, and on tapping the 8th hole, what happens?

I snap the 6-32 tap RIGHT off at the surface!


4nCOS.jpg
>
4nCOSl.jpg
alt="" /></a>


... I'm going out to buy another heatsink.. and another tap.
 
Hate it when that happens!!!!! I always buy at least a couple of taps. If I buy one, I know I will break it. If I buy two, I never break it. It is like insurance. :)
 
BLARGH!

Just broke ANOTHER tap!

6-32 taps, the things snap like twigs...


Ruined two heatsinks this way, the taps snap off right at the surface...
 
Back
Top