DIY LEDs - The write-up

[hllywd]

Does anyone have the exact measurements of height of the optics (when mounted to the star) from the surface they are mounted on? I am building my frame and am incorporating a tempered glass or acrylic shield and want to make sure I leave enough clearance for the optics.

I epoxied mine on yesterday, they're just under 3/4". Pads or screwing them on may make a difference. I'm planning to put my shield at 3/4" with the optic almost on the shield.
Tim

 

[MikePetro]

My personal opinion and just a feeling is that you will have still too much LEDs=light and then = algae. But as I said"¦ just a feeling"¦

Thanks Monty. Yea, I have kept Africans for many years, I love them, I stay away from the mbuna now though, too hard to maintain a community tank with those aggressive little suckers.

I am hoping it wont be too much light. I figure I will start at 700ma and then back off to 350ma if it is too much intensity. My fear was not having enough spread, even without optics. I plan on building one panel first just so I can do some real-time testing before I complete the whole build. I ordered all the LEDs this morning but just one driver. I figure I will play with the single panel before figuring out my ultimate driver requirements.

I even thought of using an IDEC SmartRelay with an analoug output to give me a 0-10v signal that would gradually get brighter, peaking at mid afternoon, and then gradually get darker fading to only blues for a while. With the smartrelay I can program most any dim cycle.

BTW, does anyone know of a line of DIN-Rail mount drivers?

Mike

 

[kcress]

Mike, I think you should consider using U channel instead of 1/2 a ton of heat sinks. Your linear layout caters perfectly to U-channel. They will allow you to slightly adjust all your placements after-the-fact. That will let you more easily compensate for your restrictions. They'd also save you, probably, $300.

 

[MikePetro]

Mike, I think you should consider using U channel instead of 1/2 a ton of heat sinks. Your linear layout caters perfectly to U-channel. They will allow you to slightly adjust all your placements after-the-fact. That will let you more easily compensate for your restrictions. They'd also save you, probably, $300.

Do you have a supplier you can recommend for me to look at? I can get the heatsinks I mentioned for free out of the scrap bin at work.

 

[MontyJack]

Do you have a supplier you can recommend for me to look at? I can get the heatsinks I mentioned for free out of the scrap bin at work.

Since you can hang a ton of aluminum above your tank, there would be no discussion with you about fans and thermal concerns!! :lolspin::lol2:

 

[MikePetro]

Since you can hang a ton of aluminum above your tank, there would be no discussion with you about fans and thermal concerns!! :lolspin::lol2:

And I should get good life out of the LEDs with almost any current. :dance:

 

[hllywd]

First light!!! Oh how pretty!!!!:bounce1::bounce1::bounce1:

 

[widmer]

Do you have a supplier you can recommend for me to look at? I can get the heatsinks I mentioned for free out of the scrap bin at work.

True, because you are in fact planning on driving them at 700 or even 350 mA, and because you do have them in a linear arrangement, you are an ideal candidate for the aluminum U-channel.

Any big box hardware store has it. A six foot long peice that is one inch wide with 1/2 or 3/4 inch tall walls is normally in the range of $7 at the big box. And you would only need three of them, so you would be looking at ~$25. Even though your heatsinks are free, you might want to look into the aluminum U-channel because it would amount to less than 5 pounds over your tank rather than 50 pounds with the heat sinks. It would look a lot less intrusive as well, if you don't use a canopy. Just something to think about..

 

[MikePetro]

True, because you are in fact planning on driving them at 700 or even 350 mA, and because you do have them in a linear arrangement, you are an ideal candidate for the aluminum U-channel.

OK, I am interested.

• Do these U-channels provide enough sinking to use without a fan? Say 20 stars stretched across 6 feet?

• Are you guys mounting the stars like this |_| on the bottom and then allowing ventilation across the top?

• After doing some searches I have found the material as thick as 1/8" and in a variety of widths. How do I calculate what size and thickness I would need for my application? For example Speedy had some 2" square and 1/8" thick. Can I get by with less.

 

[widmer]

• Do these U-channels provide enough sinking to use without a fan? Say 20 stars stretched across 6 feet?

I would guess that it would be just fine if you're maxing out at 700 mA and they're spaced 3" apart, but I'm sure someone with more experience will chime in.

• Are you guys mounting the stars like this |_| on the bottom and then allowing ventilation across the top?

Yes

• After doing some searches I have found the material as thick as 1/8" and in a variety of widths. How do I calculate what size and thickness I would need for my application? For example Speedy had some 2" square and 1/8" thick. Can I get by with less.

I wouldn't go any thinner than 1/8". Think of it this way - if it were as thin as paper, heat would already be lifted off by air before it travels more than a couple millimeters from the LED star. But this doesn't mean thin is more efficient, it's just that thinner material will not allow the heat to be drawn out as far across it to be conducted away, essentially creating a super hot spot at the star. This is why heat sinks have thick beds.

The caveat is that based on your spacing of the LEDs, it is probably not worthwhile to spend the money to go much thicker than 1/8".

When you say that Speedy has 2" square, hopefully you mean U-tubing. There is a such thing as square tubing, but it is actually a lot less useful than U-tubing unless you're pushing air through the square tubing.

All you technically need to mount your stars is 1" (or maybe even 3/4" depending on where you bought your stars). But you can imagine all of the extra surface area that comes with having 2" U-tubing rather than 1" U-tubing. So if there isn't much price difference, I would go with the 2". But maybe someone will chime in to say that you're just fine with 1" and no fans @ 700 mA...

 

[hllywd]

First light!!! Oh how pretty!!!!:bounce1::bounce1::bounce1:

A little more info... I set up my Meanwell Ds temporarily with a variable voltage wallwart. Pretty happy so far, I set the driver to .73A for both the blue and white with the dimmer signal @ 9V. My Kill-A-Watt says 71W at that level, with dimming set @ 1.5V (the lowest setting) it reads 18W.

At full power 4 of these will be around 280W vs approximately 700W for my MHs and PCs... I have one of the four wired so far and it's not complete so not mounted. However I did try it over the tank and it outshines the existing lights by a lot!

Probably more light than I can use and at least a 60% reduction in electric usage... WOW!

 

[MikePetro]

Design thoughts and Emitter protection

Design thoughts and Emitter protection

OK, you guys gave me food for thought on the whole U-Channel idea.

Imagine:

• A long rectangular frame, the length of your tank (or any length), built out of 1/4" inch aluminum angle.

• Multiple frames to cover the width of your tank, as many as needed.

• A pvc wire trough running in between the individual frames, think Panduit or similar.

• Individual emitter modules made onto U-Channel heatsinks that could be placed inside the angle frame, anywhere along its length.

• All wiring connections get made in the wiring trough. Either by simple wirenuts or it might be an ideal application for some sort of buss bar.

• Individual emitter modules could not only be placed anywhere laterally, but the bottoms could also be shimmed (screws on each corner would be ideal) to allow angular adjustment to focus directly on that special piece of coral.

• Individual emitter modules could be moved or replaced easily to accommodate a change in placement/chromacity/intensity needs, or even a failed emitter.

• Easily scalable design.

First, I need to design the emitter modules.

• While testing would certainly be mandatory, I am envisioning starting with a 2" square piece of 1/8" to 3/8" thick U-Channel. Thermal requirements of specific emitters could theoretically be compensated for by the thickness.

• Ideally, each emitter module should have stand-alone circuit protection.

OK, shoot holes in my concept PLEASE so I can figure out the pitfalls.



I am trying to figure out a way to protect the string from a single emitter module failure.

If I go by the assumption of the emitters operating at 500ma or less (for my own personal application) would the following device work? Lets say using the the 1206L050/15 to give me the most voltage variability.

http://www.littelfuse.com/data/en/Data_Sheets/Littelfuse_PTC_1206L.pdf

It trips in .1 sec, would that be fast enough?

 

[MikePetro]

BTW, if U-channel didnt have enough thermal dissipation, individual square heatsinks could be used just as easily. Has anyone figured out what a single XPG running at 700ma requires for a heatsink?

Fans would also help but I would prefer not to use them.

 

[widmer]

The idea of allowing the angle of the LED to be adjustable is something I really got into as well. Here's my build thread:

http://www.reefcentral.com/forums/showthread.php?t=1784873

My LEDs are for the most part just adjustable one one axis, and each one is certainly not individually adjustable. I had thought about figuring something out so that they are each individually adjustable, but resided to the fact that it was too much work, especially for a guy like me with very limited power tools.

If you think you can swing it easily enough, I would say go for it. The only thing I can forsee is that you may find the actual noticeable change in light from adjusting any single LED's pitch might inevitably not justify the work you put into it.

Also, I'm not exactly sure what the product is that you linked to... Some sort of fuse? Either way, it is important to note that after pages and pages of discussion about putting LEDs in parallel as opposed to in series is just not feasible.

Aside from the very few who are building their own circuitry, 99% of the people on this board are using either the Meanwell ELN or LPC series of power supplies, or the buck puck series of power supplies to power their LEDs because these end up being the most well-proven, cost-effective ways of driving our LEDs.

 

[MikePetro]

I personally dont need the pitch adjustment, but by putting four screws into the bottom of each module I though it might be a nice feature for you coral guys.

If a single emitter fails in series, then the others are safe?

 

[widmer]

Using meanwells or buckpucks, yes they should be safe (someone correct me if I'm wrong).

But it should also be said that you should simply not count on any of the Cree LEDs ever failing, except maybe after several years of heavy use. I think the concept of LEDs failing is mostly the blame of crappy off-brand manufacturers quality control issues, and improper wiring/driving of the LEDs, both instances demonstrated by the two-thousand-dollar Solaris fixtures before their manufacturer was put under by a copyright lawsuit.

 

[laverda]

OK, you guys gave me food for thought on the whole U-Channel idea.

Imagine:

• A long rectangular frame, the length of your tank (or any length), built out of 1/4" inch aluminum angle.

• Multiple frames to cover the width of your tank, as many as needed.

• A pvc wire trough running in between the individual frames, think Panduit or similar.

• Individual emitter modules made onto U-Channel heatsinks that could be placed inside the angle frame, anywhere along its length.

• All wiring connections get made in the wiring trough. Either by simple wirenuts or it might be an ideal application for some sort of buss bar.

• Individual emitter modules could not only be placed anywhere laterally, but the bottoms could also be shimmed (screws on each corner would be ideal) to allow angular adjustment to focus directly on that special piece of coral.

• Individual emitter modules could be moved or replaced easily to accommodate a change in placement/chromacity/intensity needs, or even a failed emitter.

• Easily scalable design.

First, I need to design the emitter modules.

• While testing would certainly be mandatory, I am envisioning starting with a 2" square piece of 1/8" to 3/8" thick U-Channel. Thermal requirements of specific emitters could theoretically be compensated for by the thickness.

• Ideally, each emitter module should have stand-alone circuit protection.

OK, shoot holes in my concept PLEASE so I can figure out the pitfalls.

The only problem I see is it sounds very complex and makes the wiring more complicated then it needs to be. There really is not that much wiring and wire nuts or a bus bar are not the best way to go about it. A buss bar is fine for each grouping of LEDs.
I am building a test rig now using a 8' for U channel for my 240. It will have 26-39 RB XPEs and is intended to hopefully replace two 46" 54 watt atinic T5s. It is really simple and should be about all you would need for your tank.

 

[bstohrer]

OK, I am interested.

• Do these U-channels provide enough sinking to use without a fan? Say 20 stars stretched across 6 feet?

• Are you guys mounting the stars like this |_| on the bottom and then allowing ventilation across the top?

• After doing some searches I have found the material as thick as 1/8" and in a variety of widths. How do I calculate what size and thickness I would need for my application? For example Speedy had some 2" square and 1/8" thick. Can I get by with less.

Mike,

I' using u-channels (1" X 0.5" X 0.125") for my build. I'll have 25 XR-Es and XP-Gs spaced out along 6 feet. I built a 3 foot test rig using a 1" X 1" X 0.125" U with 12 XR-Es held on with screws and with thermal grease. With no airflow in an open room at 70 F, driven at 950 mA, the temp. at the back of one star stabilized at 140 F. With very gentle airflow (small fan 15 feet away) temp only reached about 100 F.

I'm going to go with fans in my canopy becauae I think that temps will get well above 140 F with no airflow in my semi-enclosed canopy.

Hope that helps.

Bob

 

[der_wille_zur_macht]

Using meanwells or buckpucks, yes they should be safe (someone correct me if I'm wrong).

More or less correct IME. If an LED fails open, the string just turns off. If it fails short, the driver regulates voltage down to the point where the current remains the same. The only real problem in a pure series arrangement should be if you get so many failures that the Vf of the string drops below the minimum output of the driver, but I really can't see that happening. And if it did, you'd already have lost most of your LEDs anyways.

Parallel and matrix arrangements, on the other hand, require much more careful thought about fool-proofing. In a "several series strings in parallel" arrangement, if one string has one LED fail open, the other strings get I/n extra current, where I is the overall current and n is the number of strings left. In a simple arrangement where you have two strings, this means that the still-running string gets TWICE it's nominal current, which can clearly be dangerous! Even if it doesn't insta-fry the LEDs, it's now running WAY above the temperature you designed it for. Hence fuses are important in these arrangements!

Mike, the only issue I see with your proposed design is that series-wired LEDs arrays aren't really ideal for that type of flexibility, and non-series arrays are tough(er) to design for the type of performance we're used to from series arrays.

Maybe at some point in the future when thermal requirements are way less, drive/control circuits are simpler and cheaper, and HP LEDs have become as cheap and easy to use as conventional LEDs are today, someone will implement a LEGO-style reef lighting array where you can just place LED bricks on a framework and everything will "just work" regardless of number, color, etc. of LEDs you add.

 

[laverda]

How do I test the forward voltage of individual LEDs?