BEST DIY LED fixture ever?
- Thread starter ChadTheSpike
- Start date
ChadTheSpike
Reef Engineer
Thanks y'all! (and the ghetto rig thread from last year that pops up from time-to-time is truly awesome!! If this were in a fish room instead of my living room, it would be a fully different story!!)
Taking a break from the tedious and more mundane wiring of LEDs (summary pictures later), I did a little bench testing of my GLCD. Since this image, I have modified and added some finishing flare to the images and cleaned up the wiring quite a bit (though I will be completely redoing the wiring when I mount the screen to the top).
Oh, and I will replace the array of small red LEDs with some really bright ones :spin3:
Taking a break from the tedious and more mundane wiring of LEDs (summary pictures later), I did a little bench testing of my GLCD. Since this image, I have modified and added some finishing flare to the images and cleaned up the wiring quite a bit (though I will be completely redoing the wiring when I mount the screen to the top).
Oh, and I will replace the array of small red LEDs with some really bright ones :spin3:
Last edited:
ChadTheSpike
Reef Engineer
Thanks!
LOL, I don't know that my wife would agree with you!!
Those are kind of neat. I am not immediately sure, it appears the answer would be yes, but I would need to look into it further. Cool find!
I don't mind. These numbers don't include the hundreds of hours I have doing R&D, shipping, or the random components her or there that I needed to test a concept before putting it to use.
LEDs, optics, and heat transferring stickers: ~$350
Microcontroller and peripheral parts (RTC, protoshield, GLCD, buttons, wire, etc): ~$150
Power supplies, driver parts, fans, plugs, switches, etc: ~$150
Heat sinks: $11 (not a typo... thanks to the recycling bin at a buddy's office!)
Misc. hardware (fasteners, eye bolts, trim, grommets, etc.): ~$100
Fixture material and CNC time: ~$350
ChadTheSpike
Reef Engineer
Yes, I do.
The code is completely written and works (at least on a very small LED array mounted to a breadboard).
It really isn't more difficult to write to 10-12 channels than it is to 2 channels if you use a "for" command.
The code is completely written and works (at least on a very small LED array mounted to a breadboard).
It really isn't more difficult to write to 10-12 channels than it is to 2 channels if you use a "for" command.
wrxreefer823
New member
Absolutely amazing. When will they be for sale? 
ange062
New member
Wow man, this is very impressive! I think you have taken too many engineering classes and have too much time on your hands though, lol! 
To answer your question, yes you take my award for best DIY LED fixture ever!
And I actually did read through all the details, not just glancing at the pics
To answer your question, yes you take my award for best DIY LED fixture ever!
And I actually did read through all the details, not just glancing at the pics
ReefRescue
New member
that's a great built will be looking forward to the finished product...
ChadTheSpike
Reef Engineer
Thanks for the comments, y'all!!
I don't think I will sell them, I like projects, but I am ready for my next one!
A little bit more progress... 7.5 hours of running wires and soldering (you might say I am a little anal about it...). Still not doing powered checks, but I will be soon.
Part way into the tedious process of wiring.
Wiring done. Soldering started. After I finished soldering, I scrubbed everything with alcohol, then checked every LED for continuity, grounds, and verified it was wired correctly. After I bolt down the heat sinks, getting back to make a repair and trouble shoot will be more difficult.
Heat sinks in place after all the strings are wired, soldered and tested:
Close up of sinks in place, still need to be bolted down. When I put the terminals in, the wires will be cleaned up pretty nicely.
Slowly, but surely, it is coming together!
Next up: Finish bolting the heat sinks down, wire the strings to the drivers, and do powered checks.
I don't think I will sell them, I like projects, but I am ready for my next one!
A little bit more progress... 7.5 hours of running wires and soldering (you might say I am a little anal about it...). Still not doing powered checks, but I will be soon.
Part way into the tedious process of wiring.
Wiring done. Soldering started. After I finished soldering, I scrubbed everything with alcohol, then checked every LED for continuity, grounds, and verified it was wired correctly. After I bolt down the heat sinks, getting back to make a repair and trouble shoot will be more difficult.
Heat sinks in place after all the strings are wired, soldered and tested:
Close up of sinks in place, still need to be bolted down. When I put the terminals in, the wires will be cleaned up pretty nicely.
Slowly, but surely, it is coming together!
Next up: Finish bolting the heat sinks down, wire the strings to the drivers, and do powered checks.
Sean the Nailer
New member
Very nice lighting project. Good job on both the fixture as well as the tutorial.
I have a number of questions that I'd like to ask, while at the same time NONE of them are intended to detract nor diminish the accomplishments that you've made. I don't have my own reef aquarium yet, since I have been studying it/them and their requirements for quite a while. Mostly do to funds and abilities, but that's another story.
Whenever I make a project, I always look at it as a 'prototype', no matter how many "do-over's" and "re-makes" I've accomplished. Call me OCD.
FWIW, here goes:
A) Could the spaces between the LED's on each heatsink be filled with more LED's that are mounted on 45° angles? Each on alternating between 'left' and 'right'?
--The point for this, to be simulating the angular direction of 'sunrise' and 'sunset'.
B) If so, could then your "Gaussian" programming be increased to, say, 66? --This would then enable you to program in a more defining daylight cycle, would it not?
C) Taking this eleventeen steps further, would it be possible to then double that "Gaussian cycle" and program in the lunar cycle as well?
--Simply to create a 'night cycle' as well, using individual LED's instead of a bank of them. Depending on the parameters set.
The whole point of the above questions, is to try to simulate the actual Light Intensity and annual cycle, so that the reward would be possibly a 'coral spawning' event.
That, and also the effect of a varying and subtle 'night-light' type thing in the house. It is our intent to build an aquarium which is approximately 7 1/2' long, and about 2 feet wide by roughly 3 to 3 1/2 feet deep. The layout of the rocky corals as well as the lighting of the unit has been a somewhat constant interest for me, for a few years now.
None of these questions OR points are intended to detract from the accomplishments made. I am only asking with the intent of 'going further' and/or to find out if I am confusing myself into a corner.
Considering the location intended, as well as the final effect desired, I am wanting to do this as 'right' as I can. This thread is definitely showing me that my thoughts are on the right track so far. But where do the rails end? :reading:
I just got to this site, as suggested by another. This thread inspired me to make this first post. Thanks for the opportunity.
I have a number of questions that I'd like to ask, while at the same time NONE of them are intended to detract nor diminish the accomplishments that you've made. I don't have my own reef aquarium yet, since I have been studying it/them and their requirements for quite a while. Mostly do to funds and abilities, but that's another story.
Whenever I make a project, I always look at it as a 'prototype', no matter how many "do-over's" and "re-makes" I've accomplished. Call me OCD.
FWIW, here goes:
A) Could the spaces between the LED's on each heatsink be filled with more LED's that are mounted on 45° angles? Each on alternating between 'left' and 'right'?
--The point for this, to be simulating the angular direction of 'sunrise' and 'sunset'.
B) If so, could then your "Gaussian" programming be increased to, say, 66? --This would then enable you to program in a more defining daylight cycle, would it not?
C) Taking this eleventeen steps further, would it be possible to then double that "Gaussian cycle" and program in the lunar cycle as well?
--Simply to create a 'night cycle' as well, using individual LED's instead of a bank of them. Depending on the parameters set.
The whole point of the above questions, is to try to simulate the actual Light Intensity and annual cycle, so that the reward would be possibly a 'coral spawning' event.
That, and also the effect of a varying and subtle 'night-light' type thing in the house. It is our intent to build an aquarium which is approximately 7 1/2' long, and about 2 feet wide by roughly 3 to 3 1/2 feet deep. The layout of the rocky corals as well as the lighting of the unit has been a somewhat constant interest for me, for a few years now.
None of these questions OR points are intended to detract from the accomplishments made. I am only asking with the intent of 'going further' and/or to find out if I am confusing myself into a corner.
Considering the location intended, as well as the final effect desired, I am wanting to do this as 'right' as I can. This thread is definitely showing me that my thoughts are on the right track so far. But where do the rails end? :reading:
I just got to this site, as suggested by another. This thread inspired me to make this first post. Thanks for the opportunity.
ChadTheSpike
Reef Engineer
Thank you for the comments everyone!! I appreciate them all greatly 
Also, this is a good time to point out that I have made many, many, many design decisions over the course of this project. Ask away with questions, if I have thought about it I will give the why/why not with the direction I went. And if I haven't thought about it, I want to. It is never too late to make a design change for a great reason.
...believe me, I understand.
I haven't run calcs on the heat transfer (funny, I just had that conversation this morning with one of my co-workers) because I am confident that the heat removal capacity is well beyond the heat generated by the LEDs. That being said, there is quite a bit more space on the heat sinks. So additional LEDs could be mounted.
In my original sketches (I will post one if can find a scanner or take a decent pic with my phone), the LEDs were mounted on underside of an arch that extended beyond the length of the tank itself. (which is actually how I landed on 11 channels... (180* - 15*)/15 = 11)
Due to the size and general klunky looking arched fixture, I flattened the arch, but maintained the angles on the LEDs. Later on when having manufacturability discussions with my CNC guy, I decided that maintaining the angles wasn't worth the cost so I pointed the LEDs down with as wide angle of optics as I could find and shortened the fixture to be the length of the tank.
I also considered mounting the LEDs with the bottom of the fixture open and being able to manually put the angle on the heat sinks. I discarded this idea since I wanted all of the electronics mounted within the cabinet.
At the time of this decision I realized that one of my main design goals (casting shadows in both directions depending on time of the day) was at risk. However, designing things is a give and take balance. In this case I wasn't willing to pay double the manufacturing costs on the fixture to ensure I achieved my "cool factor" design goal.
But all that being said, sure, it could be done, and there is space for it. Also, I thought the other day that if I had made the bottom of the fixture out of clear acrylic, it may have solved the problem (would need to test light lost through 1/4 acrylic at various angles with a meter to make a decision there... more work to be done).
The number of channels could be increased to anything. I'm not quite sure what you are getting at here. I chose 33 because based on playing with the numbers that is the number it took to "push" the bell shape 99% out of the visible range. By keeping the same number of LED channels and raising the number of Gaussian channels, it takes up programming space without benefit.
Yes, it is possible. And can be accomplished entirely via programming. I am considering it as a future software upgrade.
That would be awesome!
Your project sounds like a fun project too!
Coming from a fellow thinker, start setting decisions now and work to turn those decisions into reality, then design around the "rock" desicions. Otherwise, it becomes a never-ending iteration problem.
Also, this is a good time to point out that I have made many, many, many design decisions over the course of this project. Ask away with questions, if I have thought about it I will give the why/why not with the direction I went. And if I haven't thought about it, I want to. It is never too late to make a design change for a great reason.
Very nice lighting project. Good job on both the fixture as well as the tutorial.
I have a number of questions that I'd like to ask, while at the same time NONE of them are intended to detract nor diminish the accomplishments that you've made. I don't have my own reef aquarium yet, since I have been studying it/them and their requirements for quite a while. Mostly do to funds and abilities, but that's another story.
Whenever I make a project, I always look at it as a 'prototype', no matter how many "do-over's" and "re-makes" I've accomplished. Call me OCD.
...believe me, I understand.
I haven't run calcs on the heat transfer (funny, I just had that conversation this morning with one of my co-workers) because I am confident that the heat removal capacity is well beyond the heat generated by the LEDs. That being said, there is quite a bit more space on the heat sinks. So additional LEDs could be mounted.
In my original sketches (I will post one if can find a scanner or take a decent pic with my phone), the LEDs were mounted on underside of an arch that extended beyond the length of the tank itself. (which is actually how I landed on 11 channels... (180* - 15*)/15 = 11)
Due to the size and general klunky looking arched fixture, I flattened the arch, but maintained the angles on the LEDs. Later on when having manufacturability discussions with my CNC guy, I decided that maintaining the angles wasn't worth the cost so I pointed the LEDs down with as wide angle of optics as I could find and shortened the fixture to be the length of the tank.
I also considered mounting the LEDs with the bottom of the fixture open and being able to manually put the angle on the heat sinks. I discarded this idea since I wanted all of the electronics mounted within the cabinet.
At the time of this decision I realized that one of my main design goals (casting shadows in both directions depending on time of the day) was at risk. However, designing things is a give and take balance. In this case I wasn't willing to pay double the manufacturing costs on the fixture to ensure I achieved my "cool factor" design goal.
But all that being said, sure, it could be done, and there is space for it. Also, I thought the other day that if I had made the bottom of the fixture out of clear acrylic, it may have solved the problem (would need to test light lost through 1/4 acrylic at various angles with a meter to make a decision there... more work to be done).
The number of channels could be increased to anything. I'm not quite sure what you are getting at here. I chose 33 because based on playing with the numbers that is the number it took to "push" the bell shape 99% out of the visible range. By keeping the same number of LED channels and raising the number of Gaussian channels, it takes up programming space without benefit.
Yes, it is possible. And can be accomplished entirely via programming. I am considering it as a future software upgrade.
That would be awesome!
Your project sounds like a fun project too!
Coming from a fellow thinker, start setting decisions now and work to turn those decisions into reality, then design around the "rock" desicions. Otherwise, it becomes a never-ending iteration problem.
Sean the Nailer
New member
.... and so it starts.
Time to start taking a chainsaw to the living room. Um,, I guess that it would be best to wait until "She" is out of the house. I can see it now: "But Honey, he told me to,,,,,"
Seriously though, and thanks for your replies, my "Gaussian" statements were considering that you have your lighting cycle set up for 11 banks. I was inferring that be increased to at least 33 banks. This being also based on the 'angular' LED's being individually controlled, not just as a bank. Ergo, you could have one come on (angular), then a second, then say 4, all leading up exponentially until your 'daylight' sequence is in effect. Then, once your daylight sequence extinguishes, the 'dusk' portion would then commence, until the darkness/night rendition begins.
i'm not sure if my thoughts are conveying the way they are in my head. Probably because of all the voices overcrowding them, as well. (HA) Point is, with the improvements of LED's in the past 10 years, as well as PLC abilities, I see the only holdback being material access and time to actually doing the deed. Maybe I go too far, I dunno.
:reading:
Time to start taking a chainsaw to the living room. Um,, I guess that it would be best to wait until "She" is out of the house. I can see it now: "But Honey, he told me to,,,,,"
Seriously though, and thanks for your replies, my "Gaussian" statements were considering that you have your lighting cycle set up for 11 banks. I was inferring that be increased to at least 33 banks. This being also based on the 'angular' LED's being individually controlled, not just as a bank. Ergo, you could have one come on (angular), then a second, then say 4, all leading up exponentially until your 'daylight' sequence is in effect. Then, once your daylight sequence extinguishes, the 'dusk' portion would then commence, until the darkness/night rendition begins.
i'm not sure if my thoughts are conveying the way they are in my head. Probably because of all the voices overcrowding them, as well. (HA) Point is, with the improvements of LED's in the past 10 years, as well as PLC abilities, I see the only holdback being material access and time to actually doing the deed. Maybe I go too far, I dunno.
:reading:
Here is the best night lighting i have seen with leds. I have a feeling yours will soon be the best in my book.
http://www.youtube.com/watch?v=MWU7HgxnJNY&feature=youtube_gdata_player
Sent from my iPhone using Tapatalk
http://www.youtube.com/watch?v=MWU7HgxnJNY&feature=youtube_gdata_player
Sent from my iPhone using Tapatalk
Don't suppose you cold dig up that site, those heat sinks look way too versatile for me to not have a stockpile sitting in my garage, LOL
Thanks, in advance!
Similar threads
