DIY LEDs - The write-up

Status
Not open for further replies.
I should qualify this statement by disclosing that I made this up purely to satisfy my own interests. It has no basis other than that I made it up by comparing different LED builds that have been up for a few months or more.

That said, most people are in the range of 10 - 20 square inches per LED. Closer to 10 if you have a deeper tank (say, over 16") and/or want SPS, and closer to 20 if you keep easier corals or have a shallower tank. This is assuming highest-efficiency LEDs (Q5 bin XR-E or 100 bin cool white Rebels, mixed around 1:1 with high-bin royal blue and/or blue XR-E and Rebels.)

I haven't figured out exactly how optics fit into this equation, since though they make the downward light much more intense, which would suggest you can user fewer LEDs for the same PAR in the tank, they also make the beam narrower, which would suggest you need to keep a tighter spacing.
 
using stars

using stars

Is there really a reason to use the stars at all? Why not use a nice piece of thermal tape to hold the emitters directly to the nice huge over sized heat sinks we use. Once mounted, you really don't have to worry much about heat so long as you solder quickly. You can use 24 gauge wire to hook it all up as that will carry 3 amps just fine and we only need to carry 1amp max.

If you do this you can make a much tighter packing arrangement.

I DO NOT RECOMMEND THIS unless you can surface mount solder.

PS: the optics will still fit just fine. The heat sink becomes the mounting surface. You might have to shave off a mm of material to account for the stars missing thickness. Sandpaper is cheap.
 
Isn't the benefit of the stars that they are electrically isolated? It seems that directly mounting multiple LEDs to one metal heat sink is not a good idea... Of course this is all assumption as I have zero electrical engineering knowledge.
 
The emitters base solders onto the Stars. That is what gives them such good thermal conductivity to the star.

There are three contacts on the bottom of the emitters:
1 - Large surface for thermal conductivity
2 - LED +
3 - LED -

If you try to attach them directly to a piece of aluminum, the contacts on the bottom will short out.

Attaching with tape or even thermal epoxy is nowhere near as good as the surface mount solder as the heatsink.

Now - if you buy a thermal heat sink circuit board ( copper core ), then you could get away with soldering directly to the circuit board, but you would have to etch traces for the LED contacts.

Stu
 
<a href=showthread.php?s=&postid=15503556#post15503556 target=_blank>Originally posted</a> by stugray

Attaching with tape or even thermal epoxy is nowhere near as good as the surface mount solder as the heatsink.
Click to expand...

Thats interesting and it brings up a question for me. I was planning on using star mounted crees but i was going to attach them to the heatsink with an adhesive instead of the drilling and tapping method that i see so many people around do. Its just laziness really, i don't feel like drilling tapping and screwing all those holes when i don't have to, unless there is a reason that i should do it that way?
 
Skeptic_07 I too was thinking about expoxy, but I figured that in case I wanted to add some LED's later or take some away I should screw them on. Perhaps if you got a bad LED and need to replace it, much easier to loosen two screws and sever the wires than pry the star off and re-attach a new one, IMO. So I will be drilling and tapping all this week for my 72 stars!
 
Thats a good point Wesley, I hadn't thought of that. Although in that case, you would still have to pull out the iron to solder the wires which is a PITA in itself. Having to scrape one or two off with a razor isn't much more of a bother when you compare it to all the drilling a tapping. I was thinking more along the lines of them heating up too much because the star isn't directly touching the heat sink.
 
If you go the adhesive route, just make sure you use something designed for this situation (i.e. thermally conductive, and holds up under high heat conditions.) I wouldn't go at it with any old off the shelf epoxy, but there are adhesives out there with this environment in mind - i.e. arctic silver thermal epoxy, etc.
 
Thanks wille there is also some tape from 3m I was considering but the glue might be better because of the irregular shape of the stars
 
You have a good point, I guess its about even depending on how you look at it, maybe I will reconsider not doing the drilling aspect. It will definitely allow me to finish attaching the stars in about an hour or so in comparison to all that drilling...decisions decisions decisions.
 
Re: DMX control

Re: DMX control

<a href=showthread.php?s=&postid=15494331#post15494331 target=_blank>Originally posted</a> by drewread
Now to answer a few of the questions that were left hanging in the past 84 pages...

There was a DMX controlable LED driver that was suggested in the early going, and someone asked how it could be controled. There was an answer that explained what DMX is, but not really how it is used. So here I go.

DMX-512 is a protocol that is mainly used within the entertainment industry for control of all of the lighting equipment that you see used in your favourite rock shows, tv shows, feature films, etc. I know this because I happen to work as a lighting console programmer and lighting designer within the entertainment industry. I am in fact typing this from a feature film set in Toronto right now. I am sitting in front of a GrandMA (www.malighting.com) which is being used to control over 50 dimmer racks of conventional lighting and probably about 150 3 colour LED panels. Obviously this would be useable to control my tank lighting, however a $40K USD console would be a bit of overkill for this purpose. ;)

DMX protocol is also used for lighting large buildings in some cases, although it is not used for that as much as it is used in the entertainment side of things, where it is the standard.

There are obviously smaller and cheaper DMX-512 controllers on the market, some for the PC, however I personally think that none of them are either cheap enough, or simple enough to program a proper sequence of moonlight, sunrise, daylight and sunset for an aquarium.

I think that a better solution would be something along the lines of what lynxvs has done with his control and driver setup. We haven't heard from him in quite a while, but I think he has some very valuable information as to how his controller works and is wired that we could all use, so I am going to see if he can come an help us out. :D
Click to expand...



Thanks drewread for the vote of confidence, I have been following this thread without contributing for quite some time. I, like you would like to see more technical content in this thread, I mean how many pages are dedicated on whether to drill and tap the holes on the heatsink or epoxy them down! I would like to see more before and after results of LED lighting for one"¦ maybe compare the results from someone other then Soundwave. I am considering starting my own build thread for my design. Unfortunately the build is complex and requires special tools and software so I'm not sure it would help the average DIYer that is why I haven't done it yet "¦ I have been running my LED light for well over a year so I think I have worked out most of the bugs and it is a solid design. If I get more then one person to respond to this post I will consider starting a thread"¦ It takes a lot of work to put a good build thread together for something that generates zero interest. From what I have seen on this thread and others (Just my opinion and no reflection on any one person) any real technical content creates feedback from one or two people then gets basically ignored only to be asked again in the following pages.
 
i would too.....the ones i have contemplated doing would be smaller units that could be done w/ gang pots, but i have thought about what/how something on a larger scale would be done (when i lay in bed unable to sleep thinking of my "dream tank setup") LOL

the one i will probably end up doing in the near future will be a very small setup....all buckpucks & only (3) rows of 5-6 leds each row, so surise/sunset will be done by having (2) rows on 1 power supply & (1) row of 50/50 blue/white on another power supply & each timed differently.

i know larger setups of a similar configuration would benefit from pwm control, but since i have no experience in them & won't be doing any anytime soon, i haven't researched them
 
One thing related to Dimming & POTs that I forgot to mention.

There are 'serial addressable variable potentiometers with EEPROM'.

These devices have an internal variable resistor that is set via serial wires & they maintain the value set in EEPROM.

I used to use XICOR EEPOTS for robotics but they arent is business anymore.

MAXIM makes some like the DS3901: http://www.maxim-ic.com/quick_view2.cfm/qv_pk/4824/t/al

These could easily be used to DIM the buckpucks & operate from I2C ( so you can attach as many as you like on just two pins of a microcontroller ).

Now if Maxim would just make some high current I2C PWM drivers ;-)

Stu
 
My current nano build will have three banks of 8 LEDs each. Each bank will be controlled by a DIY driver built around On Semi's NCP3065 or NCP3066 HB LED driver. These chips can take a PWM signal for dimming.

Since it's looking like I'll be building a reef controller around an Arduino board for this tank, I'll use that to supply PWM signals for dimming. My plan is to have the "east" bank of LED begin to power up slowly in the morning, followed by the center and west banks, each delayed an hour or so. Reverse in the evening. This will give me gentle dimming to simulate the brightness curve real reefs get, as well as directional sunrise/sunset.

Once the system is running, I'll do a build thread, but I'm really a hack when it comes to this stuff, and I'm definitely learning as I go. When I finally get around to my next big build (~300g), I plan on applying the same concept. This nano is more or less a pilot project for this approach.

Stu, if you have a driver that wants a variable voltage (i.e. like the buckpucks) instead of PWM to dim, what would be the advantage of a programmable potentiometer like you're describing, vs. a PWM signal smoothed into a variable voltage by an RC filter?
 
DMX on the cheap

DMX on the cheap

http://www.behringer.com/EN/Products/LC2412.aspx

Here is a nice DMX control for $200.00

There are ones for around $60.00 that hook to your computer via usb so all of you who want to do the sunrise sunset thing could do it that way.

http://aa.usno.navy.mil/data/

This site give table for sunrise sunset for any location earth and allows you to calculate where the moon is as well. IF you really want to go to the end you can calc sunrise sunset, use red, yellow, orange leds to sim sunrise and sunset and than calc the phase of the moon and use a track of small white leds accross the tank to simulate the moon traveling across the sky and the intensity based on phase.

In addition to the DMX board you'd need to convert the DMX channel/intensity data to a voltage to run the buckpucks. None of this is really hard to do. I do caution you to use a look up table to do the sun rise sunset thing as the equations are a bit difficult to program and a look up table is much easier to program.

Have I done this no. Do I plan to no.
 
Status
Not open for further replies.

Similar threads

R
DIY AI-coded ESP32 based aquarium controller
Replies
0
Views
307
ronnyking
R
AquaticReefTV
DIY Cable management time lapse
Replies
1
Views
530
S|\/|S
S
AZBigJohn
New Build... I'm back
Replies
15
Views
446
Dustin07
D
B
Beam Deflection - 15 years and I finally broke down and wrote a how to.
Replies
16
Views
913
BeanAnimal
B
A
Fleece filter bypass problems (Innovitech X filter)
Replies
1
Views
35
Vinny Kreyling
Vinny Kreyling
Back
Top