DIY LEDs - The write-up

[der_wille_zur_macht]

if seperating the whites & blues (as you mention below), & running each string at 700ma, you would need a 24v power supply that put out at least 1.5amps....like THIS one (so 2 of those) & then you'd need a 3rd one of those that would run the 50/50 string of white/blues & you could also run the fans for the heatsink on it as well.

If you're running 700mA on each string, you'll want to inflate these numbers to have a safe margin - most DC power supplies really start to heat up (read:shorten their lifespan!) above 70 - 80% of their rated output. So, add 20 - 30% for safety margin. In other words, if you're doing two 700mA strings on one power supply, you'll want more like 2A capacity, not 1.5A.

And since you mentioned heatsink fans, this is probably obvious, but consider the voltage. Most PC fans are 12v, and 99% of these LED builds are using 24v supplies. So if you want one fan, you'll need a voltage regulator to knock the 24v down to 12v. Or, if you use two fans, just wire them in series so that each sees 12v. Even better would be three fans, in series, such that each saw 8v. For most fans, this will result in vastly quieter operation without a huge airflow hit, so the the three fans run on 24v will probably be quieter and higher airflow than two.

if you decide to do as i stated above & do 2 buckpucks of whites & 2 or blues, then you just need a gang-pot like THIS....one for the whites & one for the blues

a "gang pot" has multiple inputs & outputs depending on how many "gangs" it has. that one is a dual-gang, so it will control 2 inputs/outputs w/ just one knob, so if you had 1 for the whites, it would dim both buckpucks at once....same for the blues

There's no need to have a gang pot for controlling multiple buckpucks that you want dimmed the same - you can use a single pot wired to several of them. Just use one buckpuck's REF pin to supply +5v to the pot, and connect the other side of the pot to the CTRL pin of all the "similar" buckpucks (i.e. all the ones you want to control the same.) One buckpuck's REF pin can provide enough juice to control 4 buckpucks, plus or minus (5mA each, the REF pin can supply 20mA.) If you wanted to control more than 4 pucks off a single pot, you could just use an external 5V supply, as long as you grounded it to LED- as shown in the buckpuck datasheet. That way, you could control a virtually unlimited number of puckpucks on a single pot, limited only by the capacity of your external 5V supply.

 

[widmer]

Even better would be three fans, in series, such that each saw 8v. For most fans, this will result in vastly quieter operation without a huge airflow hit, so the the three fans run on 24v will probably be quieter and higher airflow than two.

And affordable :strange:. Yesterday I bought 3x 60 mm fans (i needed this size for my LED project) rated at 18 cfm/ 24 dba on ebay, total cost after shipping $12.48. And they're not coming from Hong Kong. Hopefully after going in series on 24v as per der_wille's recommendation, I'll still be getting ~40 cfm all together at something close to silent operation...

 

[james3370]

There's no need to have a gang pot for controlling multiple buckpucks that you want dimmed the same - you can use a single pot wired to several of them. Just use one buckpuck's REF pin to supply +5v to the pot, and connect the other side of the pot to the CTRL pin of all the "similar" buckpucks (i.e. all the ones you want to control the same.) One buckpuck's REF pin can provide enough juice to control 4 buckpucks, plus or minus (5mA each, the REF pin can supply 20mA.) If you wanted to control more than 4 pucks off a single pot, you could just use an external 5V supply, as long as you grounded it to LED- as shown in the buckpuck datasheet. That way, you could control a virtually unlimited number of puckpucks on a single pot, limited only by the capacity of your external 5V supply.

interesting......i was thinking about 3 or 4 strings & was having a hard time finding a 4-gang pot that was a decent price:worried2:

 

[MontyJack]

Hi Folks,
I need help with my setup. I have 12 MW ELN-60-48D drivers and I would love to dim them with my Profilux. The problem is I will put 6 of them on one dimming channel. One driver has 3.3mA, so together they have 19.8mA. Now I find out that than I would have ~10mA to much on the dimming channel. Is there any way to amplify the signal (0-10V, 10mA max) coming out of the controller up to somewhere around 20-40mA. Just the amps not the volts! I'm asking this way, because I'm not in electronics, but handy enough to solder something, if you give me a hind.

Thanks in advance.

Monty

 

[Cramz07]

I just got through reading this whole thing and still don't have a complete grasp but i think i'm heading in the right track with what i want to do specific to my wants....My goal is to have a led light working by the end of may so i've got a while to do a little research and really get a good hand on this. My only real questions as of right now is how many buckpucks max. can you run off of one power supply lets say 24v. and also i know the question is in this forum for sure but just to make sure you can run the royal blues at 1000mA and you need to run the whites at 700mA to be efficient or should they both be at 700mA?

Sorry if i am being repetitive but i want to make sure so i can start ordering this in when i get a good grasp on this...

Thanks

 

[der_wille_zur_macht]

My only real questions as of right now is how many buckpucks max. can you run off of one power supply lets say 24v.

When running buckpuck drivers, there are two parameters you need to keep in mind. The voltage a driver will need, and the current it will pull. The voltage is dependent on the number of LEDs you use, and their forward voltage. Typically, people are driving 6 LEDs in series per buckpuck (their max, according to the datasheet). This results in a forward voltage around 22v. The buckpucks themselves need about 2v of headroom, so a 24v power supply is absolutely perfect for this application.

The second consideration for a buckpuck driver is the current it will draw. Typically, you can assume this will be the current on the output side (i.e. the drive current) plus a small margin. If you are using a 700mA buckpuck, you want a power supply that can do 700mA, plus 20 - 30% margin. Thus, the power supply can have a max current above ~1A and you'll be fine.

So, for a single buckpuck driving 6 LEDs at 700mA, you need a 24v supply capable of ~1A or more.

What about multiple drivers on the same DC power supply? Typically, they are wired in parallel (i.e. all the + pins on the drivers to the + pin on the power supply, and all the - pins on the drivers to the - pin on the supply.) When wiring loads in parallel, the voltage requirement stays the same, but the current requirement is the sum of the requirements for each individual load.

What this means is you need ~1A per 700mA buckpuck on a 24v supply. So, if you wanted to run 6 buckpucks on the same supply, you'd need a supply running at 24v capable of 6A or more.

mpja.com has really nice 24v power supplies in the 6 - 8A range for around $15 - $20, which makes this all very easy.

and also i know the question is in this forum for sure but just to make sure you can run the royal blues at 1000mA and you need to run the whites at 700mA to be efficient or should they both be at 700mA?

Your questions are delicate because of the wording you used. You need to decide what your most important goal is - efficiency, or absolute output. You can't have both. The blue, royal blue, and cool white XR-E LEDs can all be run at 1000mA max. However, they are all most efficient at lower currents (350mA is what their efficiency numbers are typically quoted at.) Many people choose 700mA because it's a reasonable combination of intensity, efficiency, and safety. Running these LEDs at 1000mA results in more output, but at a lower efficiency rate - and also more heat, etc. IMHO, if you think you need 1000mA drive current, you might be better off driving at 700mA and using 20% more LEDs. That'll result in a fixture that's more expensive upfront, but less expensive in the long run, cooler, and more efficient.

 

[Cramz07]

Ok so the buckpucks ran in parallel, the led's themselves, however, must be in a seres to prevent blowing the buckpuck correct? and the led's total amps must be within, with head room, to the total amps put out by the buckpuck and then same with the buckpuck to the power supply.

And with how much mA to run the led's at i'd rather go with max. efficiency rather than max. output and add led's like you said.....Now, for max efficiency i should run both at 700mA?

 

[der_wille_zur_macht]

Ok so the buckpucks ran in parallel, the led's themselves, however, must be in a seres to prevent blowing the buckpuck correct?

Yes, correct. Multiple buckpucks on the same PS would be wired to the PS in parallel. But, LEDs are always wired to the output side of the driver in series. It's not so much to keep from blowing the driver, more just that it's the way it's designed to work. (Technically, you can put multiple series strings of LEDs in parallel on the output side of the driver, but there are lots of good reasons to not do this.)

and the led's total amps must be within, with head room, to the total amps put out by the buckpuck and then same with the buckpuck to the power supply.

The amps running through the LEDs will be dictated by the driver - that's it's purpose in life. So the only thing you need to do when matching the driver to the LEDs is choose the right drive current.

But your second statement is correct. Regardless of how many drivers you are running on one PS, the sum of the current used by each driver must be less than the capacity of the PS, with a 20 - 30% margin.

And with how much mA to run the led's at i'd rather go with max. efficiency rather than max. output and add led's like you said.....Now, for max efficiency i should run both at 700mA?

Maximum efficiency will be more like 350mA, but at that low of a current, absolute output is pretty low. The decision can't be made in a vacuum - you need to take into consideration the dimensions of the tank, the height the LEDs will be mounted, any optics you're using, etc. Many people choose 700mA because it's a good all around compromise, but it's not really the best in any one parameter.

 

[Cramz07]

ok good. I think i've got it....throwing around voltage, amps, milliamps, and all these different terms throws me off but i'm pretty sure my understanding of this is confirmed. Thanks!

One more quick question to satisfy my curiosity....have they put a par. rating out there that is needed by corals like they have with the "general" rule of thumb with watts per gallon....

 

[der_wille_zur_macht]

One more quick question to satisfy my curiosity....have they put a par. rating out there that is needed by corals like they have with the "general" rule of thumb with watts per gallon....

That's pretty much impossible to do - different corals will do just fine over HUGE PAR ranges, so it's tough, other than to try to get something in the same ballpark as other lighting technologies. Luckily, we seem to be getting far enough along with this LED stuff to have pretty good approximations, and at least a few people have done PAR tests that confirm that the output is reasonable.

Though, with any lighting technology, there are a million variables that can alter output. With LEDs, it's a slightly different set of variables. For instance, we don't have to worry about reflector design, but we do have the option of using optics.

 

[Cramz07]

so how much is too much...is that even an answerable question through the trial and error that the people diy'ing it know yet

 

[der_wille_zur_macht]

If you dig in the various LED threads, there are data to suggest that some of the more intense LED rigs are definitely "too much" - there have been a few reports of mass bleaching events.

In the end, it's like any other lighting technology. If you put a 400w MH on a 15g nano, even the most light-hungry SPS would probably not do well. So, decide how much intensity you want, and plan the rig around that. If you have a reasonably deep tank and you stick around 10 - 15 square inches of tank surface per LED, run at 700mA with wide (or no) optics, you'll be around the same intensity as some of the more intense MH or T5 rigs commonly in use. And as plenty of results have verified, that's about as intense as you want to be.

 

[Danielrvb]

Ok, So I have finally finished going through this after several days of reading (seriously). I suppose it is now time to start to contribute a little. One thing I gone back and forth with is the drivers to use. Initially I was thinking the meanwell approach for simplicity.. However now I am starting to lean back to buckpuck method.
I am hoping to mimic der's diy drivers to get costs down a bit (though I still have to go through that thread). I am not sure how much I am concerned with dimming them as price is a little more important to me then having sunrise/sunset so just the cheapest is good for me.

And this brought me to another idea. driving the leds parallel is something that was discussed early on and quick discouraged, And I see the reason behind it. but my question is this. What about the option of saying wiring them parrallel at 700ma, but on each line putting a 1amp fuse wired in, so that if one line goes out and forwards the power to the second line, it will blow the fuse and kill all power? a thought for some feed back.


Second area I would like to see some decrease in standards :eek2: (just what everyone wants to hear) Is the area of heatsinks. it seems very few people are having heating issues but it is still getting suggested to go over kill based on making the leds last longer. Now this is good, however I truly dont expect to be happy with my current LED diy 10 years from now. I would think at best I would be looking at 3 years before the technology is so much better I am wanting to upgrade.

For that same matter, I have had tanks for going on 9 years now, and I dont think I have had the same set of lights over a tank for more than 2 years.

Just some more ideas to start bouncing around and make this thread even longer and harder to go through for the people who still have not found it. :beer:

 

[kcress]

driving the leds parallel is something that was discussed early on and quick discouraged, And I see the reason behind it. but my question is this. What about the option of saying wiring them parrallel at 700ma, but on each line putting a 1amp fuse wired in, so that if one line goes out and forwards the power to the second line, it will blow the fuse and kill all power?

You have it wrong.. If you put a 1A fuse inline with a current regulator regulating to 700mA it would NEVER blow.

Now if you put in two strings of LEDs in parallel with Ireg at 700mA you would want 400mA fuses in each string so the two 350mA strings are protected. Then if one string opens the other will too.

BTW in all your studies here did you read my "cheap LED" tutorial? Sounds like the ticket in your case.

Second area I would like to see some decrease in standards :eek2: (just what everyone wants to hear) Is the area of heatsinks. it seems very few people are having heating issues but it is still getting suggested to go over kill based on making the leds last longer.

You are absolutely correct. As soon as a fan is dragged to the party big official heat sinks are certainly not required. Neither the weight, the fins, nor the mass. You do have to make sure the 'wind' visits all the LED mounts however.

 

[Danielrvb]

I dont believe I saw that. There were a few later nights were I started getting tired and started skimming through stuff that seemed to have been posted several times previously so maybe I skimmed over it.
but truly I have gone through each page, sometimes getting distracted and moving to other stuff and going through other threads (partially why it took me so long)

As far as the fuse, I must have worded it wrong. I think your saying what I meant would be an acceptable approach.

ie If you were to get a 1400ma driver, running two parrallel strings at 700ma. From what I gathered when you were discouraging people from doing this in the early parts of this thread, the concern is if one led blows. 1400ma would then start running through the other string, cause the entire string to blow because of one led.

So, what I am thinking is with the led's (I am looking at the common standard of the Cree XR-E RB, and now for whites the Cree XP-G) being maxed out at 1000ma, protecting both strings with a 1 amp fuse would prevent the entire 1.4 amps from being diverted to a single string and over powering them. Then worse case scenario, the max that would go through anything would only be the leds max.

these numbers are just quick ideas, I really have not gone to far into looking into it. a quick google search yeilded 1 amp fuses (did not look for any other size) and also 1400 ma drivers (again, did not look further) So Ideally I was hoping there would be potential for me to simplify further and have far fewer drivers by getting much bigger ones then parallel wiring the strings and protecting each string with a 25 cent fuse.


Ok, just search through threads you have started and found you write up, I will go read it now and it looks to be a long read, wtg. I was also wondering about simply doing resistors, but before I comment on that I will go read.

For others that did not see it before, here is the link http://reefcentral.com/forums/showthread.php?t=1751598

 

[der_wille_zur_macht]

Daniel, fusing the parallel strings would prevent failure, but it leads to some other (potential) issues:

1) If the LEDs in one string have a different total forward voltage at a given current than the LEDs in the other string, they won't balance out well. This would be especially true if you mixed different colors/types of LEDs on the same driver. I'd want to carefully "bin" the LEDs I was using (set up a test station where you could drive a single LED for a few seconds to record it's voltage at a given current) to avoid this.

2) As you get more LEDs on a driver, you start to lose control resolution. Maybe this isn't an issue on a very large tank, but on a smaller tank, if you had drivers doing 12, or 24 LEDs each (for example) you quickly lose resolution to the point that it would be hard to implement the sort of control people are starting to show interest in. For an extreme example, I have a nano rig with 16 LEDs run at very low current. This is two of my DIY drivers, 8 LEDs each. A driver capable of doing all 16 wouldn't even let me dim blue and white separately.

From the sounds of your posts, neither of these would be huge stumbling blocks for you, but I wanted to point them out in case others were following along.

 

[der_wille_zur_macht]

Oh, and another thought. (warning: beware the rambling fool!)

Now this is good, however I truly dont expect to be happy with my current LED diy 10 years from now. I would think at best I would be looking at 3 years before the technology is so much better I am wanting to upgrade.

For that same matter, I have had tanks for going on 9 years now, and I dont think I have had the same set of lights over a tank for more than 2 years.

It's exactly this sort of situation that I'm hoping to avoid by using LEDs. I don't want to feel like my rig is total garbage in 3 or 4 years, like I have with most of the lighting I've used in the past. That's why I've been playing with nano-sized prototypes for 5 or 6 months while dream about doing it on my new big tank. Here are my thoughts with respect to longevity:

1) Besides extending life, running the LEDs cooler also increases efficiency (output per watt consumed) at a given current. So, even if I do want to replace LEDs in a few years, they will have been more efficient in the meantime.

2) The very modular nature of LEDs means that it'll be easy to "upgrade" part of the system (drivers, power supplies, LEDs, optics) if something vastly better comes out.

3) And while the XP-G is arguably way better than the XR-E, I'm wondering when we'll hit a "wall" of new LEDs being over the top for our purposes. You can get 1000w MH lamps, but most would agree that 250w lamps are sufficient for 90% of reef tanks. Similarly, I'm guessing that at a certain point we'll discover that the newest and most intense LEDs are simply too much light from a single point. I'm sure we'll come up with a creative way around that (optics?) but in the end, I'm satisfied with the performance we're able to get now. There's an argument that newer LEDs could be run at lower currents for the same output but more efficiency, but honestly I think we'll hit a ceiling there, too - the current LEDs are already so vastly more efficient than any other type of lighting that further increases will result in a lower absolute difference in operating cost. My 360g tank might cost $700/yr (in electricity cost) to light with MH, but with current LEDs the electrical cost will be like $250/yr. If LEDs come out that are 10% more efficient, that's only like a cup of coffee a month in savings.

It's interesting that the LED culture mirrors electronics in general. There's always going to be something better in the near future. So, you can buy what's there now, or wait for something better. You need to make a decision based on features available at some point, instead of just waiting for the better thing purely because it's better. I still use an "antique" original iPod with "only" 4gb of memory, because that's all the music I care to have at one time. Similarly, I've reached a point where the current LED technology strikes me as good enough.

All that said, if the Cree AB-C comes out in two years and it's vastly better than XR-Es or whatever I build my rig with, I can always buy a dozen and sub them in at the "hot spots" in the tank where I want more light. That's the beauty of LEDs. You get vastly more flexibility and upgrade potential than other forms of lighting. Again, that's why I'm going this route with my big tank. It's going to be cheap and efficient to operate, it's an overall system I know I'll be happy with long term, AND it's something I'll be able to constantly tinker with long term. Just the potential to write different software routines to fade the lights on and off in different patterns will certainly keep me busy far longer than the "joy" of trying out a new MH lamp would.

 

[Danielrvb]

I see your point, it would be nice to have a light and not worry about it. My thoughts are with yours mostly. But were I differ is I hope in a few years its just that. The "high Power" leds are out of our range, and so we are using the top of our range, and running them at much lower power levels, thus making the fixtures much more efficient then now, also hopefully they are much cheaper.

The day you can build a very efficient, controllable, and cheap LED fixture is when I would want to start looking at replacing again. But as of now, I dont really want to front that kind of money for something that will be much more within my price range in a few years. Thus doing an economy style build until then.

another thought I have had for people to think of regarding the ideas were starting to head towards with the abilities to customize lighting in very exact places is this.

I wonder about the ability to get the lighting to come from the front of your tank as nearly as possible. My thoughts with this is the idea of people who have top view only tanks. Corals present their colors to the light. So when your looking into a tank, your actually not getting to see your coral in its most colorful area. but if you could get light to also shine largely from the front, maybe you could get the viewing angle of a standard tank to have very colorful corals.

I know with far lower powered LEDs you can get them to be underwater. Are their any for our level? would it be feasible to have a large square, just like a power head, sitting at the front corners of your tank half way down in the water shining light at the front of the coral?

hmm...

 

[Santoki]

wow, well put DWZM.
That is exactly my thought whenever people talk about future technology.
I've always thought reef lighting fixtures are always one of the ugliest part of an aquarium setup. They are traditionally big and bulky, and always seemed like it wanted to be seen as part of the setup rather than a tool to highlight the setup, like in art galleries... you want to see the painting lit, not the fixture that's lighting it.
I was never fully satisfied with any of the available options until these LEDs were available and I was able to build my own. I can really say I am sufficiently satisfied with the quality, output, and aesthetic look of my own units that I won't be searching anytime in the near future. Before this, I had experimented with every commercial and DIY reef lighting available to the average hobbiest, but never satisfied, leaving my wife to wonder whether I had some incurable disease which caused me to constantly tinker with my light setup.
I think if I ever decide to upgrade to the latest greatest emitters, it really is as simple as swapping out the LEDs (provided they are mounted to similar stars or boards).
Same goes for power supply upgrades. I don't think I can think of one single con with going full LED except maybe the slightly higher initial investment... even that becomes a moot point in a couple years time...
-R

 

[der_wille_zur_macht]

I see your point, it would be nice to have a light and not worry about it. My thoughts are with yours mostly. But were I differ is I hope in a few years its just that. The "high Power" leds are out of our range, and so we are using the top of our range, and running them at much lower power levels, thus making the fixtures much more efficient then now, also hopefully they are much cheaper.

The day you can build a very efficient, controllable, and cheap LED fixture is when I would want to start looking at replacing again. But as of now, I dont really want to front that kind of money for something that will be much more within my price range in a few years. Thus doing an economy style build until then.

Makes total sense. I'm sure better LEDs will be out soon and I'll be eating my words. :lol:

That said, I DO have to eat what I've said more than once in the last few months - it doesn't quite seem like LEDs are following quite the same price/performance development curve as you might hope from Moore's law. Newer, more efficient and more powerful LEDs are coming out, but they're not getting cheaper as fast as I'd hope. The XP-G has been out for a few months now, but it's expensive, and the XR-E and others that have been around for a year or two haven't really changed in price by more than a few percent at best.

another thought I have had for people to think of regarding the ideas were starting to head towards with the abilities to customize lighting in very exact places is this.

I'm 100% on board with you. To date, the majority of LED builds have basically been imitations of T5 or MH builds. Even light, distributed vertically and symmetrically above the tank. Static, other than perhaps gentle dimming in the morning and evening. We're missing on a TON of potential here!

I wonder about the ability to get the lighting to come from the front of your tank as nearly as possible. My thoughts with this is the idea of people who have top view only tanks. Corals present their colors to the light. So when your looking into a tank, your actually not getting to see your coral in its most colorful area. but if you could get light to also shine largely from the front, maybe you could get the viewing angle of a standard tank to have very colorful corals.

Again, totally on board. Most of my experimental nano fixtures have been in repositionable fixtures (clip-on lamps with goosenecks, etc.) It's been fascinating to say "hey, I wonder what that coral would look like if the light was slanted a bit," then reach up to the fixture, slant it a bit, and WOAH! the coral looks TOTALLY different!

To me, that raises two possibilities:

1) "Interesting" static placement of LEDs - like you're mentioning, having them angled from the front, or something else creative.

2) "Interesting" dynamic positioning of LEDs - I've dove on the same reefs at different times of day over several days' time. At 8 in the morning, the reef looks TOTALLY different than at 5 PM. It's not just an intensity thing, it's an angle and color thing. LEDs will let us duplicate that dynamic pattern in our tanks. The tools are THERE, right in front of us. We just have to grab them and use them!

I know with far lower powered LEDs you can get them to be underwater. Are their any for our level? would it be feasible to have a large square, just like a power head, sitting at the front corners of your tank half way down in the water shining light at the front of the coral?

hmm...

That's one thing that I don't think we're at yet, but hopefully will be, soon. Though, it raises a bunch of functional challenges, even if cooling wouldn't be an issue. You'd have to hide the box to keep it from looking weird. You'd have to clean the lens regularly. etc.

I can really say I am sufficiently satisfied with the quality, output, and aesthetic look of my own units that I won't be searching anytime in the near future.

I'd be totally satisfied with your build, too - it's one of my favorites. Good work! I appreciate what you mean about the lighting fixture being less obvious, and I think you've really pulled that off. This tank pulls it off well too, despite not using LEDs:

I can't wait for people to start doing things like that on reef tanks with LEDs.