• LED pendant build

    I still remember in my early days of fish keeping as a teenager, walking into a LFS (local fish store) asking for "brighter lights". I was handed a couple dual 40w fluorescent fixtures for my 55g (quite expensive too I remember). I went home, installed it, and was disappointed that the light was not only not bright enough, it was lacking that (what I now know to be referred to as) "shimmer". This was before I knew of anything called metal halide. Since then (almost 20 years ago), I've learned a lot and that "shimmer" was something I could not do without. After experimenting with various combinations (CF/MH VHO/MH, T5/MH) I finally settled on using MH (metal halide) solely as my only lighting in all my various setups through my on and off years of reefkeeping.
    Any combination seems to just diminish that "shimmer" effect. But, lighting has always been an area for which I felt aesthetically limited. Commercial fixtures were bulky and cumbersome. HQI units seem to come closest to my personal ideal. However, with this type of powerful lighting also comes associated disadvantages. Halides generate a great deal of heat, which in turn heat up the display below (not to mention the many burns suffered from carelessness while working in the tanks). This can really be problematic in hot climates, or when small water volumes are concerned. In addition, one soon starts amassing boxes and boxes of old bulbs as years pass.

    Enter LED (light-emitting diode) technology.
    Until fairly recently, LEDs were not used in any type of real commercial applications to actually provide usable light. The most common applications for these tiny emitters were inside computers and other devices which need some type of low power indicator lighting. In the past ten years, advances have allowed this technology to finally be used in practical lighting applications.
    LEDs have many advantages over other types of lighting:
    • Efficiency: Low energy usage. They use a fraction of the energy used in the most efficient form of lighting to produce the same output (when compared to what we have available today).
    • Efficiency: Low heat production. More energy is turned into light output than any other form of lighting currently available.
    • Specific spectral output. LEDs produce very specific spectrums, allowing users to control end color easily by mixing different types of LEDs. They also do not have spectral shifts when they age, unlike metal halides.
    • Longevity: For example, the LEDs used in my build will keep 70% of output (lumen maintenance) within 11 years when used 12 hours a day. This means no bulb replacement for at least 10 years!
    • Eco friendly: In addition to using less energy, there are no toxic materials to add to the landfill (unlike fluorescents and halides).
    • Small package size: Due to the small package sizes of these light emitters, a great deal of design possibilities are now available to light fixture designers, including us reefkeepers. I will attempt to share just one of these ideas in application in the following pages.

    I've always admired the clean look of Japanese spotlit tanks, but those fixtures were never available in the States as far as I know. With the flexibility and now maturity of LEDs, I finally can make an attempt to create my ideal type of lighting.

    First prototype unit was built and underwent testing.
    Majority of readings were conducted out of water for baseline, since it provided for a more controllable test environment. A few readings were then taken in my display for practical purposes.

    First, some views of the first prototype:


    Bottom: notice the absence of exposed wires.

    Side: low profile

    Close up of optics:

    Some pictures taken during PAR measurments:

    Some in tank measurements: (flash was used to expose the LCD display on the Quantum Meter)

    FTS (full tank shot) with array >4 feet from tank bottom

    And finally, the PAR measurements:

    In tank measurements are not consistent I believe due to the difficulty in consistent sensor placement. However, it does show that little if any light is lost through water of this depth. One reading read more than when measured outside water.

    After the first prototype was built and testing completed, I set out to build a few more for myself and some locals.

    I realized that the first build was over-engineered in the cooling department and only half of the original heatsink was needed, so I designed the Gen2 with that in mind. The advantage is that the pendent is half the size of the original while putting out identical output. This time, I attempted to document some of the steps leading to the Gen2's completion.

    Parts list:
    The following is a list of items needed to build this unit:
    • Heatsink:
    The size I used for the latest build was ≈6" X 8.5", with fins running on the longside. I purchased some 17" stock and cut them down to size.

    • Aluminum flat bar used for fan mount and connectors:
    I purchased these from my local HomeDepot.
    I'm sure they can be found online for less.

    • 0.25 inch self-tapping stainless screws (sizing is something like 3/38 thread) This was a major source of headache for me, because I had originally found some screws at a local hardware store which worked beautifully, but could not figure out the correct scale measurements. It would have been nice if everything was metric.

    • Nylon washers for "M2" screw
    I also purchased the aluminum channel and mounting hardware from them.
    I highly recommend them.

    • Cree Royal Blue and Q4 cool white LEDs
    Q5 cool whites can be used, but are more expensive. After extensive testing, I did not find them necessary. Since they are very slightly brighter than the Q4s when run at the same current, using the same 50/50 white/blue ratio will yield a slightly whiter look. I must also reiterate, these "royal blues" from Cree are very different than what most may be used to when it comes to LED blue. Most LEDs come in blues which peak at 470nm and are very low output. These are not particularly useful for photosynthesis. They also do not excite the pigments in corals which cause them to glow under the actinic lighting we are so accustomed to seeing. The "royal blues" have a spectral output in the 450-460nm range, and at relatively high output. These blues are also used in horticulture LED grow lights to stimulate terrestrial plant photosynthesis.

    • Meanwell non-dimmable drivers
    Dimmable drivers can also be used, but I do not think manual dimming is useful (more on that later).
    (Driven at 700mA)

    • Optics.
    I experimented with 40˚, 60˚, and 80˚ optics and settled on using 40˚ optics for my application.
    These are one of the most important aspects and variables in LED reef lighting (more on that later).
    Both of which I've had good experiences.
    As far as pricing is concerned, I think it they are about the same. However, I like how Rapidled sells packages now which contain almost everything you need to complete an array.

    • Cooling fans
    These are just case fans from Newegg.com.
    80mm fans for the Gen2 build, 120mm for the original prototype.
    • Power supplies for cooling fans
    I had some 5V cell phone chargers laying around which I used. However, anything =5V or =12V or less can be used.
    Some switchable power supplies like this from www.summitsource.com can also be used.
    • DB9 connectors male/female
    Radio Shack. I know there are cheaper places online, I just haven't gotten around to finding them yet...

    • Wires for LED
    I used a 24 gauge solid copper wire roll I had from a previous project. I think it was purchased from RadioShack.
    tip on that: the wire shielding needs to be strong, as fishing the wires in and out and around the heatsink can cause problems with the jacket if you're not careful or are not using quality wires.

    That's it! I think I've spent more time experimenting than actual building. But all the experience gained from trial and error definitely makes the actual build much easier.
    Total cost of materials for one build was a bit over $300.

    Here are some pics of the second build... I did not have a free hand, so it was difficult for me to document as well as I wanted.. but it does show how much work is involved with prepping the heatsink (my least favorite part).

    Cutting main stock to size:

    Completed cut:

    Jagged edges to be finished:

    A template was created to facilitate the building of multiple units:

    Drilling the 72 holes needed to mount LEDs for each unit.

    Mount holes complete:

    Polishing surface for better LED/heatsink mating:

    Fan mounts are cut and refinished:

    Mounting begins:

    Mounting complete and soldering begins:

    Backside wiring:
    This was done so that no exposed wires show; for safety, as well as aesthetic reasons.

    Finished and hung on rails over display:

    The orginal prototype on the left and the Gen2 on the right:

    I mounted all pendents on wheels so they can be moved if needed:

    Another FTS... All exposures were lowered to prevent washout, hence the seemingly dark surroundings:

    Side shot with center actinic (blues) on only. Exposure was short to preserve clarity. It is much brighter in reality.

    Top down actinic

    No post processing was made. The reds are really that red. These have been over the display for a month at the time of this writing.

    I have read from many early DIY'ers using Cree emitters that it seems to put out more light than expected. I think my PAR readings reinforce that. As such, I am slowly acclimating the display to these lights with very short cycles. All three pendents overlap no more than 3 hours a day, and they are staggered so that there is light on the display for 11.5 hours a day.

    So far, all corals seem to be responding favorably with new tips, and in the case of the plating montiporas, more folds are developing, and surface is becoming more textured. Some acroporas in direct light under the center pendent are lightening in response to the more intense light.

    As far as costs are concerned, I find it a moot point when you consider the costs of MH bulb replacement. In my case, I replace bulbs every six months as I find that output decreases dramatically by then and spectrum changes. I notice because my corals almost stop growing. At 3x250, that's at least $210 per year. Nevermind the shock the corals have to go through each time I change bulbs.

    I often get questions from reefers interested in color manipulation through dimming. I do not think manual dimming of LEDs to control end color is efficient, since it involves reducing output on one or more colors to achieve end result. This approach means some emitters are not running at full potential. A better approach would be to adjust the ratio of the different LEDs. This ensures all emitters are running at 100% of the current supplied.

    Simulating sunrise/sunset is the best use for dimming, but requires microprocessor control. There are a few reefers working on that now.

    Optics are perhaps the most important aspect in the use of LEDs, as they channel and focus scattered light into a beam. The focusing also concentrates the light and increases intensity dramatically (see PAR table above). This data paper from Cree shows the effects of 60˚ optics on an XR-E LED. Taking advantage of optics available to us will open up a whole new frontier not explored before in reef lighting. Fixture mounting height is no longer limited. Using narrow optics will allow mounting even on the highest ceilings, opening up possibilities of dramatic gallery style presentations of our reefs. It also allows unfettered access to our reefs from the top, whether it is for maintenance or photography. Manipulating color intensity and specific highlighting can also be achieved through use of different optics for different colors.

    My display used to increase anywhere for 2-4˚F during the day when the halides are on, causing unavoidable temp fluctuations. Now, there is no fluctuation. Zero.
    LEDs will surely provide relief for those currently blasting their tank with 400w halides and running massive chillers to cope with heat. Chillers may soon go the way of the Dodo bird.

    Coral color:
    Coral coloration is an area which I hesitate to comment on, since it is a very subjective subject. What may be appealing to one may be a turn-off for another. However, I have always believed that coral coloration has more to do with water chemistry than lighting alone. I certainly do not have any complaints regarding coral colors, as they are brighter and appear more colorful to my eyes than what I was able to achieve using metal halides alone.

    Notice the small acropora frag in the center:

    Same acropora species as above picture, next to each other (1st and 2nd from right). The one on the left is newly mounted and darker. One on right changed color when placed under more direct light. It becomes even lighter as shown in the picture above.

    Mother colony in lower light area:

    LED intensity clearly enough for color manipulation...

    LED technology is here and ready now. Costs are no higher than a comparable MH system when considered over 3 years (fixture/ballast and bulb replacement costs). In addition to using only a fraction of the energy of any other reef lighting available to us, these emitters have a projected lifetime of 11 years running around 12 hours a day before light output decreases to 70% of original output, which means you can say farewell to routine bulb changes! They contain no toxic materials, unlike fluorescents and MH, which contain mercury and other toxic materials.

    With the recent Orbitec patent lawsuit against PFO, many reefers became discouraged and thought that advancement in LED reef lighting would be stifled. Perhaps as a response, many took on the task of designing and building their own systems. Although we are only getting started, the results have so far been revolutionary. Hopefully, as more hobbyists adopt this technology, commercial markets will adapt to allow end users to customize and build their rig with universal parts rather than fully-built proprietary units.
    I encourage anyone with the DIY ability and inspiration to try this. You will be pleasantly surprised.

    Added by Marc: Check out his great video:

    Added by Robert:
    Here's an updated Video from June for comparison:

    December 2010:
    Comments 53 Comments
    1. melev's Avatar
      melev -
      Is that your next project?
    1. chucelli's Avatar
      chucelli -
      Quote Originally Posted by melev View Post
      Is that your next project?
      unfortunately not in the near future.
      Our current house can't accommodate anything bigger than a 75 in that space. I'm not sure I want to go through the effort to upgrade to the same size tank. Also currently spending a lot of time building fixtures for people.
      Rimless will have to wait...
    1. balistikb's Avatar
      balistikb -
      How would I calculate how many of these fixtures I would need on top of a 96Lx24dx29h tank?
      This DYI article is great btw....
    1. chucelli's Avatar
      chucelli -
      If you are planning on this exact design (24 tightly grouped emitters), I would just double what I have over my 4' 75g. Depth wise, at 2', it is still within the coverage area for this type of design. However, depending on your specific application, you have countless other options...
      The advantage (and disadvantage for some) of LED technology is the countless different options the end user has when designing a system.
      I personally would love to have a very long tank like yours (for tangs), but only light parts of it, creating areas of shade and bright islands...
    1. balistikb's Avatar
      balistikb -
      If I wanted to light my whole tank, would you still suggest I go with your design?Also, do you make them for others? If so how much?The reading on all the forums are quite confusing and I am not sure which way is up. this write up is great.
    1. chucelli's Avatar
      chucelli -
      it would depend on how you want it lit. 6 units identical to what is featured in this article would be more than enough for your tank dimensions.
      There are other designs possible, which can be used in combination with pendent style lighting to create the custom look you want.
      So far, I have only been working with local hobbyists and LFS around my area. Cost wise, it is close to higher end commercial units, except that you can customize to suit your needs perfectly.
    1. Michael's Avatar
      Michael -
      if i mount this unit inside the canopy about 10 inches of the top of the water line which degree of the lenses should be used? tank demotions are 25"H, 24"W, 72"L.
    1. Michael's Avatar
      Michael -
      can you also explain how you wire from one led to the next?
    1. chucelli's Avatar
      chucelli -
      IME, with any type of mounting within 12" of the water surface, wide angle optics, or no optics should be used. Many times, no optics results in a look most are used to (dispersed lighting), but IMO much light is wasted with no optics.

      These LEDs are wired in series, which means DC positive goes into the positive of the first LED in the array and the negative of that LED goes to the next negative of the next LED and so forth and so on. The last LED will have a negative out which goes back to the DC. This is just one of the many articles online which explains the different ways of wiring an LED array.
    1. bac's Avatar
      bac -
      great write up , I like the clean look of your lights
    1. PastorKDS's Avatar
      PastorKDS -
      Thanks for a GREAT article and excellent sourcing! I appreciate also your response on the optics if used in a canopy. I have a 150 - 48x24x30" which optics would you suggest for a 10-12" over water mount, I'm assuming 60 or 80?

      Would you still go with 3 arrays at the lower height, with 2 - 24 light arrays, or with 2 - 36 light arrays (to avoid the bracing shadows). Again, thanks for one of the clearest and most helpful articles I have seen on LED usage!
    1. Michael's Avatar
      Michael -
      I am trying to find out what type of wheels you used inside that railing?
    1. chucelli's Avatar
      chucelli -
      I would recommend 80˚ optics for mounting 12" or less above water level. 60˚ if your tank is over 24" deep.
      As for 24-36, that really depends on how much light you really want. 36 will give you more coverage with more intensity as well depending on optics used. For 30" wide (front to back) tank, I would probably go with two 36 for more coverage.

      the wheels inside the railing can be found at McMasters under "channel trolly".
    1. Leafeater's Avatar
      Leafeater -
      Have you tried the thermal adhesive and not liked it? Is that why you tap and drill all those holes?
    1. chucelli's Avatar
      chucelli -
      I tap and drill to facilitate future upgrades or maintenance.
      This has already come in handy a few times when users have a change of heart (different LED ratios/colors) after units have been built.
      Also, for industrial professional solutions, I want to make sure that these units can be worked on by anyone with basic electrical knowledge. I know of too many Solaris units that have a few burnt out LEDs just sitting in someone's storage because it is difficult to acquire parts for them, or not being able to find anyone with enough know-how to service them.
    1. Alaska_Phil's Avatar
      Alaska_Phil -
      Very cool set-up! I've got an empty 29 gal I've been dying to set up again. How many LED units would you recommend for that tank? 30"x13"x18"

      I was thinking 24 in 3 rows or 8, about a foot above the tank with 40 degree optics since it's a tall narrow tank. Does that sound about right?
    1. chucelli's Avatar
      chucelli -
      3 rows of 8 sounds like a good plan for your footprint.
      Although at 1' above the tank, 40˚ optics may be too narrow to provide the needed coverage. You may have to experiment a bit to find the perfect optics solution for your setup.
    1. Arvind's Avatar
      Arvind -
      Hi Chucelli, Nice article. What is the size of this tank?
    1. Arvind's Avatar
      Arvind -
      Quote Originally Posted by Arvind View Post
      Hi Chucelli, Nice article. What is the size of this tank?
      Ok, I got it. It is 48". thanks!
    1. PBlueKan's Avatar
      PBlueKan -
      Sorry I may have missed it, but what did you use to power the Drivers themselves and also, how many leds did you link to each diver?