DIY LEDs - The write-up

[AlgaeNator]

AC 12V 1W LED's

AC 12V 1W LED's

These are the AC/DC? 12V 1W LED's I installed (do Google search) using the above 12V AC Transformer. Not sure how they pull it off, they do have a mini driver disc that Obviously does some of what KKress mentioned in order for the LED's to work on 12V AC. But it appears that you do indeed LOSE Quite a bit of Luminocity going with 12V AC 1W emitters. (or they are using older emitters)

But they do work, and put out a goodly amount of white crisp light, enough light to pretty much light up a 20-30ft tree all the way up. 70Lumens aint that great though as ONE P7 puts out 900lms at 3.6 Watts.

Intermatic LL0851-2 Malibu 2-Pack Low Voltage 1-Watt LED Metal Flood Lights, Textured Black


Found on amazon.com

 

[AlgaeNator]

Speaking of AC/LED's ....I just stumbled on this interesting Article about the "Ariche A4 from Seoul SemiConductor) SSC.

You can view the PDF on this product here: (add www) element-14.com/community/docs/DOC-14316

Operating temps are -30 to 85 C = Max is 176 F

Havent gotten all the details down about cost, total lms output or availabilty, as I just found this article, I will see what I can research and dig up about the A4, and I will post anything I find about the above items as they are pretty critical details.

Anyone else is welcome to chime in too about these if you know anything....

The A4's really shows some promise at first blush for home and Hobbyist's

See.. Quotes below from "ledsmagazine.com" ( search this site or just google "Ariche A4" as I cannot provide direct link on RC)

Seems interesting info, esp. about beating out DC efficiencies too, but the color spectrum and color is limited to WHITE, and WARM White at that...it is still in the yellowish (3000k) for our needs, but hopefully with all new technologies, this could also change over time...to different colors upto 10K.

If this is all true, you could just power these in mutiple Array's with simple straight AC 120V current from the outlet, using the "inexpensive diode bridge" to connect each array?

Not sure about cost either, they say "inexpensive" but as far as I know the A4 is not out yet (will look around)


Not sure about heat either come to think of it which is pretty important, though I think they run much cooler than the DC counter parts?


Here is a module or two with SEVERAL AC A4 Emitters installed.

What exactly is the "50Vrms per package" (see below)? It appears this has to do with the coversion of the sign Waves of the AC current.. (KKress feel free to chime in here.. heheh)


Trying to figure out how many lms this thing actually puts out to get a handle on brightness, but they dont say how many Amps it draws (that I have found yet)

IF you powered this say at 120V and ___Amps... that gives ____W. Then x 72 lms/W gives you total lms../emitter?

But they dont tell how many Amps it will draw?

Maybe you guys can figure out how many Amps these draw... that would give us the total lms of light they put out.

From ledsmagazine.com

"Korean LED maker Seoul Semiconductor has introduced the Acriche 4, a new version of its alternating-current (AC) LED product, in major markets around the world.
The company claims that the A4, which will be mass produced and available this quarter, offers “better luminous efficiency” than direct-current (DC) LEDs.

The A4 has a luminous efficiency of 75lm/W, and Seoul says it is capable of replacing warm-white incandescent and compact fluorescent light sources in many consumer and commercial applications.

According to Seoul, white LEDs are divided into two major markets: products with a high color rendering index (CRI) greater than 85, which is close to natural light, and products with a normal CRI of 70 to 80.

Acriche A4
High CRI products will often be significantly less bright than their lower CRI counterparts, but Seoul claims to have solved this problem
The A4 devices have a color temperature of 3000K and efficiency of 75 lm/W, while also offering a high CRI of 85.

This, says the company, is the first time that AC LEDs have achieved better luminous efficiency than comparable DC LEDs. “With the mass production of the Acriche A4 series, we will be able to provide both high quality and normal use markets with light sources that have exceptional performance and lower prices. With the A4 series, we want to spread the name ‘Acriche’ far and wide as the world’s best LED brand.”

Seoul lists several “significant advantages” of the Acriche A4 series, including the possibilty to use Acriche products on AC power freely between 100V and 230V without additional conversion. Only a simple, inexpensive diode bridge is required.

However, the A4 differs from other Acriche products (such as A2 and A3 series) in that it uses 50Vrms per package. A total voltage of 110V-220V can be achieved by connecting multiple devices in whatever manner is needed.

“Whether used in light bulbs or light structures with arrays of LEDs, the most suitable design is guaranteed due to the degree of freedom of placement and configuration of the A4,” says Seoul."

 

[AlgaeNator]

Found some tech Data on the AN4224 Module

Found some tech Data on the AN4224 Module

From SSC: (add www) Link..... "acriche.com/en/product/prd/acriche.asp"

"Acriche is the world’s first semiconductor light source that operates directly from AC power without a converter. This makes it suitable for many residential and commercial lighting applications where the main source of available power is AC. A design with Acriche product minimizes component count and board space. Acriche (pronounced ‘Ak rich) comes from the Greek “Acro” which means “the top” and the French “riche” meaning richness. "

This module is detailed below..

AN4214

My main questions still remain, are how many Tlms will this module produce at at 120V? @ 8W is it 400lms? Am I calculating the lms/W correctly based on the table below? If so, then that seems way less than the 900lms that a 3W DC P7 emitter will put out using only 4 emitters on one module?

Here is the P7: 900lms at 3W (3.6Fv @ 700mA)

Here are the spec's on the AN4214 Module

 

[kcress]

Hmmm interesting stuff. AC LEDs.

Simply put 50VRMS means common 50V AC.

The AN4214 at 120V will be driving the LEDs at 120V x 80mA = 9.6W

Since these aren't even out yet the price is ??? First do-hickies are gong to be over-priced for the bleeding edge weenies I'm afraid.

 

[der_wille_zur_macht]

I think we talked about those LEDs a few pages back. While I suppose they're interesting, IMHO it all boils down to one thing: efficiency.

If we aren't doing these LED rigs with absolutely the highest efficiency possible, we might as well not even bother. Even among the commonly available XR-E bins and commonly used drivers, the difference in overall efficiency can be 30 - 40%. At that point, it's not *really* that much better than a carefully designed MH or T5 rig, and it'll cost way more upfront.

AlgaeNator, I was at a similar point to where you seem to be several months ago - trying to find a loophole that would let me build LED rigs more cheaply. I didn't find anything - it really seemed that the current approach used in these threads is pretty much the best. I don't want to stifle creativity though, because surely there's a better way out there somewhere. The one thing I'm doing differently than most people is building my own (DC, constant current, dimmable) drivers, since it's gonna save me a bit of money, get me a slightly more efficient end package, and it's a fun challenge. The driver design I've settled on costs about $6 - $8 each, and can run 8 LEDs, so it has a significant price advantage when you're building large arrays.

Regarding the specifics you're digging up - think about efficiency, and how it alters both upfront cost, and longterm cost. The LED landscaping modules you posted above are likely using low-binned LEDs, the wrong color, and/or poor driver designs (I've opened similar units up, and you guessed right - there's basically a little tiny driver on the back of the LED(s) in each unit.) So, you're paying an overhead to put a driver on each LED to go that route - vs. using a buckpuck or DIY driver that can power many LEDs at a time. Buckpucks are a standard way to do this, but they're pretty poor from a price/performance ratio for really large builds, since they can only light 6 LEDs each.

Those AC LEDs may seem attractive because of the ability to do a build without drivers, but look at efficiency. In the last image you posted, the AN4214 consumes 8W (thought as kcress pointed out it's really 9.6W by the math). It puts out 400lm. That's around 40 - 50 lumens/watt, depending on who's watt value you use.

By contrast, a Q5 XR-E or a 100-bin Rebel will put out around 90 lumens/watt at the currents we typically drive them. So, around twice as efficient. So to get the same output with the AC LEDs, you'd be consuming twice the electricity.

Even if the AC LEDs you posted were very cheap, you'd have to seriously consider longterm cost - what it really boils down to is lumens per watt, from that perspective. And, at 40 lumens/watt, you aren't much better off than MH or T5 (perhaps slightly worse off than the best of these technologies). You'd still get some other advantages of LED lighting, but you wouldn't win the longterm cost battle. Plus, since these LEDs aren't available in blue or royal blue, you'd likely still have to use "traditional" LEDs for those colors (rebels or XR-E). So that would reduce any benefit of using the AC LEDs in the first place. If you were going the standard route for the blues, might as well do it for the whole array.

Again though, I'm trying to play devil's advocate, so please don't let this stifle your creativity.

 

[AlgaeNator]

Der (i will just call you "der" ok? lo)

No worries, I didnt realize I was going to go so deep into this AC thing, and didnt want to Hijack this thread either.... however, it does seem relative and interesting enough to talk about and submit.

As KKress and I discussed it though, its a ways off for us anyways, until they:

1. Come out
2. Get color temps for our needs
3. Can determine pricing to compare costs
4. Can determine accurately lms/W compared to DC emitters

As far as costs of using them, NOT having to purchase DC drivers is going to bring price WAY down to build our rigs, and also if the price of the LED Modules are really offered at "lower prices" as they claim this will also dramatically cut down initial costs of the emitters. (lower prices is a broad term, compared to what?? 40$/bulb home lights?)

Well as far as efficiency, goes I understand and totally agree with you, they CLAIM that the A4 SURPASSES DC LED's efficiency at 72lms/W? But the sheets dont prove this out in lms/W.

Actually KKress's total watts was incorrect, its actually 110V x 80mA 8.8W (table bears this out too) 8.8W x 72lms/W = 633.6lms is what the module should produce yet they show it only at 400lms? Not sure why that is, maybe they are figuring running it at a lower fV or not or how they come up with the 400lms total output of this module.

It may also be this is the lms/EMITTER and I think there are 8 emitters on this module. So the MODULE'S total lms output would be 3200lms?

I guess the key here is "comparable DC LED emitters" they claim in the article. The A4 is darned close though. Because, the P7 puts out 900lms at approx 3W of power consumed. I suppose the P7 may not a be a "comparable DC LED emitter"? I think we figured that P7 was putting out 225lms/emitter @ 3W. Since there are 4 emitters you divide 900lms/4 to get this figure.

Based on that, divide the 225lms/emitter by 3W/emitter consumed = 75lms/W.

Here was the TITLE that caught my eye in the link in the first place.

This, says the company, is the first time that AC LEDs have achieved better luminous efficiency than comparable DC LEDs.


The A4 devices have a color temperature of 3000K and efficiency of 75 lm/W, while also offering a high CRI of 85.

 

[kcress]

Nice summation der_wille_zur_macht!

AlgaeNator; I gave the number you would see here in the USA. If you look closely at the data sheet the voltage they give is stated to only mean the nominal country's voltage. Our country's is more typically 120V not 110V.

In that case the the 9.6W is more likely the consumption. It could also mean that that added 1W doesn't even go towards more light. No telling with so little data.


It's all a mute point though as der points out you are still going to need to have blue ones, and the colors offered, - even the rumored ones - are lousy for displays.

 

[wmilas]

I've been following this thread for months. I think I'm getting to the point where its time to speak up.

I have what is considered normally to be a MH "type" of tank because of the size. Its 10 feet long, 2.5 feet high, 3 feet wide. I light it currently with 2 Sfigiloi Stealth 4 foot (12 bulb each) t5's. I don't use MH because I don't particularly like the light color options and for mounting reasons.

I'm going to make a small LED setup for my macro algae tank to get a feel for things. However I'm still playing with the idea of taking my main tank over to LED. With the release of the new Cree XP-G's it is starting to enter the realm of possible.

Using 40 degree optics the old calculation was 1 led per 10 sq inches which would be around 432 leds. This could be lowered by using 20 degree optics along the edges. Using the new XP-G's this could be lowered again most likely.

I currently have my two large (and heavy!) t5 pendants mounted on light rails I constructed out of sliding door rails. There are two sets of rails. I'm trying to figure out how I would lay out two LED "pendants". How I would arrange the heatsinks and what the LED layout on each sink would be. I've drawn out a bunch of ideas but I was hoping for a fresh perspective.

The easiest way to do it is to use 8 or 10 inch sinks and run them length ways in strips. Each strip being 4.5 or so feet long. So perhaps 3 8 inch wide strips that are 4.5 feet long. This would comprise one pendant.

Or I could use shorter strips and run them say 30 inches long in the other direction.

This stuff is heavy though and I'm not sure what is actually practical.

Help

 

[AlgaeNator]

To correct my post above about how many lms this module should put out (I wanted to edit it but had to leave for an hour) I realized that each emitter us pulling 1W ea. and there are 8 emitters on the module, so hence the 8W draw. So, for the module lms, its 1W x 8 emitters x 75lms/W that's 600lms this module should put out NOT 400lms as the table states. I dont know where they are coming up with that 400lms figure? and claiming 75lms/W? Der is right thats actually like 40-50lms/W if it is the case.

As far as efficiency of the emitters vs DC emitters, I couldn't agree with you more. it comes down to lms/W/emitter. It is how I did my research to pick the P7 emitter.

From that one P7 emitter you would get 900lms. at 3.6-7V @ 700mA (approx 3W)

The SSC C bin P7 emitter is one of the brightest emitters on the market today and has a total lms output of 900lms at approx 3W used.. Which at first blush sounds incredible for one emitter, until you look closely at the emitter itself, you see it actually contains 4 tiny LED diode/chips/ emitters inside one emitter housing, So its really 225lms/emitter/ 3W = 75lms/W, which is comparable to other emitters out there like you mentioned like the CREE HO emitters.

Amazingly now, SSC claim for the FIRST TIME that they are getting 75lms/W with the A4, AC emitter as well. If true, this means that the A4 AC emitters are at LEAST Equal to, if not GREATER in efficiency than their HO P7 DC emitter counterparts?

In this article about the A4, they CLAIM its 75lms/W output for this emitter "SURPASSING comparable DC emitters efficiency" which means that soon we will have AC powered HO LED emitters that require no expensive drivers, and can be daisy chained using "inexpensive diode bridges" to connect several emitters into an array.

Right now how much will be $aved is an unknown since we dont know prices and they havent come out yet, and also they dont have the color temps we want.

BTW on the chart, it lists the Operating temps, they seem similar to DC at at around -30 to 80C (about 180F max).

 

[Chemistry Guy]

I'm going to admit that I didn't read the entire thread(s) to see if this was addressed, but I just read an interesting write-up on polishing aluminum heat sinks. It's my understanding that the current belief is that we should be polishing our heat sinks to a mirror finish for the best thermal contact. Well, these guys disagree...

They're CPU overclockers who took some knowledge from a scientific research lab that was developing a high efficiency heat sink for their laser crystal (similar research I'll be doing soon ). They concluded that a mirror polish actually hurts your heat transfer because you end up having a thin film of thermal paste separating the sink from the device with no aluminum actually contacting the device. If the sink is slightly coarse (they say 600-1000 grit) then the micro peaks in the aluminum will contact the device, and the paste will fill in the micro valleys.

http://forums.overclockersclub.com/index.php?showtopic=74393

 

[AlgaeNator]

KKress,

Just trying to point out that the 400lms on table was based on 110V* (I did notice on table when posting it that it says + - 10%). So the Lms with your math will ALSO go up higher at 120V than the table was listing them at 400lms at 110V

Yes i also know there is a wiggle room with real live Voltages we can expect in the US (hence the 10%), at 120V vs 110V. And you are also right we dont have enough data yet to confirm or deny the claims, what I find odd is their own tables dont support the claim, unless I am doing my math wrong?

Btw.... When I said you were "wrong" I meant that 400lms the TABLE is getting was based on using the 8W figure. not the 9.6W real life Wattage we can expect that you posted. actually it really doesnt matter how many volts are used (110 or 120) the real question is the lms/W formula they are claiming these emitters can put out, since THAT formula can be applied across ANY voltage and shows the REAL Effeciency of the emitters.


So say at 120V and 9.6W the lms will and should be slightly higher around 720lms/module at 120V. (based on the 75lms/watt claim)

So the question I guess still remains is how they are getting 400lms at 8W? When based on the 75lms/W claim and 8W used, it should be ... around 600lms? Again maybe under promising over delivering, but the math should still add up on the tables? Im sure there is an explaination for this, Im open to suggestions on it. Maybe the Key word is "upto 75lms/W"?

What is interesting to note as well, is the amount and CONFIGURATION of the LED emitters on the Module.. There are 8 emitters on the ONE module that appears to be no larger than the size of a quarter (or so)

 

[der_wille_zur_macht]

Algaenator, you're raising lots of interesting points. I only have about 23 seconds online right now so I can't answer them all, but a few thoughts:

1) SSC uses repackaged Cree chips in some of their DC LEDs, hence the end results are similar. The Cree models are usually easier to get ahold of for hobbyists, though.

2) The cost for drivers on DC LEDs doesn't *have* to be that high - if you didn't mind dealing with surface mount components, you could make 90% efficient drivers capable of powering a dozen LEDs for $6 or $8 or so. That's like 60 cents an LED, not really bad at all. Plus, again, it's important to consider the difference between upfront cost and ongoing cost. I'd rather spend 50 cents extra per LED upfront, if I knew it was getting me 10% efficiency in the longterm.

Also, IMHO you can ignore the creative configurations (4 emitters on a single die, multiple chips on a star, etc) unless you are specifically designing for them, which is a challenge. I've built three LED fixtures for nanos that used 3-up rebels, and it works at that small scale. But at a larger scale, it turns out to be a problem - you have too much of your output coming from too small of an area, and it's hard to avoid the "columns of light" effect. Extrapolating that up, if you had 8 emitters on a single star, you'd have very few stars spaced far apart to get a desired light output, and you'd certainly end up with patchy coverage.

Just some thoughts. . . I'll put more effort into the discussion when I get some free time.

 

[AlgaeNator]

Der,
Thanks for responding. boy you type fast! ahaha.. (as do I)

Im about to build my fist rig shortly, and obviously no matter what is going on with AC emitters, I realized after talking with Kcress a few days ago, that I will have to go DC setup as that's what's available NOW, and DC emitters have the color temps we all want/need (maybe minus 03s). I dont expect AC emitters color temps to change any time soon (probably for a year). The way I plan on building, I will be able to SWAP out or change my RIG to AC if it becomes more efficient to do so. But it seems the only thing that will realistically happen is that AC Emitters will EQUAL the current output in lms of DC emitters, not necessarily PASS them.

That said, these AC modules are going to be HUGE in the Home Market when they hit, as they will finally be able to create equal amounts of lms, and color temp of LED light in standard fixtures as CF, and Incandescent bulbs do now. The real question is how much they will cost and what "lower prices" actually mean. There is no doubt this is the reason for now, why they are are at first focusing first on the "natural 3000k" light temps, and even the "whiter" light emitters in this temp. range for this is what MOST people want in the home.

You are also right the cost is not THAT bad to power these DC LED's if you can handle surface mounted open drivers. There are also several new affordable contained drivers have come out since I started following this thread about a year ago that many are using. There are many affordable drivers now which will do the trick. DX has one that can power about 36 DC 3W LED's at 700mA for around 30$ if you dont want a protective case. Though like you say the DC current must be stable and to scrimp on the driver to save $$$ at the price of efficiency of the emitters in the long run doesnt make sense.

One upside of sticking with LV DC is that we will have LOW Voltage in our fixtures from the drivers to the lights, which is much safer, around the emitters, I would assume, than running 120V AC from emitter to emitter, but even then it would be no different or dangerous than wiring up NORMAL halogen/incandescent bulbs over your tank if you had to. The amp draw is still pretty low so wiring requirements wont change with the move to AC.

What could get really interesting is if somehow they can create 75lms/W emitters which can be powered by LV 12V AC, and thus be low Voltage AC. (maybe this is coming but the AC loss seems now to be too much to overcome) Then simple transformers pluged into a wall such as the Malibu 100W one above, could power hundreds of HO LED's from one power LV power source/transformer. You could also dim them with a simple dimmer switch attached to a WALL plate or other source around the plug and timers and dimmers/controllers could be cheaply added (they already exist) on the LV transformers themselves without the need for various pots around the light. One could even used the GE Wireless lights exterior lighting control/base with dimmer option on the plug and remote, to dim the whole show remotely.

 

[AlgaeNator]

Your comments about lighting/emitter placement are also valid I am sure, it will be interesting to see the coverage spread of light of these modules once installed. If one was to NOT use any optics on the LED's I would assume you could space them out a little further with lesser spotting effects? Granted you will lose some penetration going optic-less, but gaining spread to the sides for better blending?

For this reason (correct me if I am wrong thinking here anyone) I plan on NOT using any optics with my setup, in order to allow the light to spread out, from the emitters un focused. IMHO my tank (24" deep) should have PLENTY of PAR at the sand without Lenses, and would rather have it spread out side to side than focused down. As you mentioned the LED's tend to focus the light beam/ lms directionally ANYWAY even without lenses. What I do like about the lenses is the protection factor being over SW. So Im a bit unsure which way I will go, those could be added later anways as long as my spacing is pretty close to the norm I should be fine? I will be using thermal tape to hold my emitters in place, so the design and emitter placement can be changed if I need to.

Also, the space issue you mentioned, I think this bodes well for the P7 as you have a LOT of light coming off the one emitter at 900lms/emitter...with the 4diodes/emitters placed tightly on the single chip yu are combining the power of 4 3W emitters in the space of ONE 3W emitter. I am wondering with that many lms (900) from each emittter, that maybe I would have to use LESS of them than the standard 3W output single emitters to achieve the same results depending on blending I guess... The standard being 4 rows of 6 emitters on each side for a 48" tank? I guess the only way to find out is to build the thing, I would be glad to report back on how it all shakes/works out.

My tank is 48" 100GA Bowfront.

Here is my plan, if I am wrong about this plan, anyone feel free to correct/instruct me on my plans, esp if anyone has build a rig using the CBin P7s Im VERY interested to see how you ended up configuring/mixing/spacing the emitters with standard 3W-5W emitters.

My plan is to go with like 16 ea (8/side blended alternated) P7 C bin emitters for my White's, and 16ea CREE 5W (700mA) for BLues, and then 16ea 320mA Cree's for my RB (run them a bit lower). How many of each in the end, TBD. I want to experiment to see what is pleasant to MY eyes, esp using the P7s' I might want and need more blue "tinge" to the white lights. And see which combination and blend works the best for brightness and spread.

 

[AlgaeNator]

question about 400-420nm LED's

question about 400-420nm LED's

I am a believer that we are missing some of the UV spectrum you get with HQI, and Halides when using LED's and I think we need to consider using some 420nm range LED's in our fixtures. It will also give you that TRUE 03 Pleasant Actinic more Violet look vs the sometimes CLOWNISH looking bluish 440nm (RBs) that tend to "POP" only the Oranges and Yellows making your tank look like a BLACK LIGHT POSTER!! hehehe....Mixing in the 03s will give a very nice blended Pop IMHO.

For the UV reason (not only aesthetically) I would like to add some 03 (420nm) LED's to my rig and am curious about any feedback on the possibility of safely installing 400-420nm emitters into the fixture. There are very few if any True 03 (actinic) HO LED emitters that I could find that operate or even PEAK in the 420nm range (Tru03). Most are either in the 440nm range or well below 400-410nm.

I asked one MFG about this in China and they said the True 03 is difficult to make in HO Emitters, So for now its in the 400-410nm ranges.

If anyone has a link to any other HO 420nm emitters, please post it here! thanks!

I found one TRUE 03 (420nm not HO) 5mm LED emitter in my research (hard to find) that I plan on adding as a part of a moonlight strip and or blending/integrating it in such a way with the others, so it will also to add some 420nm LED light to my setup. These are probably not bright enough as they are not HO to effect the tank THAT much visually, but still will probably do the job giving UV to the tank.

I figure if I insert enough of them it might, but it would require a LOT of them to proablably even influence the RB,B of the HO emitters to the human eye.... I am also thinking to tie these into the BLUES to run as part of the dusk lighting/moonlights and to add the UV.

5mm found here

(add www) superbrightleds.com/specs/v1015_specs.htm

I also found some 1W HO 200lm emitters that could effect the overall look with the RB and B's, running, these nm ranges are min 400nm and peak at 410nm. And am wondering If you guys think it would be safe to add them to the mix? The Question I have to the RC community, Is how low in the Upper 300s- lower400s nm range can we go to stay in the "safe" range without having to filter it? I have researched this several years ago and I think I am right in thinking that a 400nm HO Emitter would be fine/safe nm range for our tanks. and even in the high 380s and 90s.... is still safe for human eyes and fish (not instant damage) Anyone please correct me if I am wrong on this. I think the "evil" sterlization UV is much lower in the 220nm ranges.

400-410nm 200m/W emitter found here.. Using bin 4

(add www) superbrightleds.com/pdfs/UV-1WS.pdf

Using Bin #4 it peaks at 410nm. I figure if you drive it at its max voltages (i know it will shorten the life)would it then will tend to operate in the higher safer nm ranges, or does it even matter?

Of course you have to be careful with UV emitters not look into them for extended periods of time etc.. but I think 400nm -410nm should be safe for fish and humans?

Interested in any backfeed on this topic!

 

[kcress]

wmilas; Since you are going to 'proto' first, you have a great opportunity to see what you can do towards minimizing heatsink mass. Large numbers of LEDs are going to require fans to move the bulk heat. Finned heat sinks lose their necessity around forced air because it's a potent heat transfer multiplier.

You should consider perforated aluminum or square tubing either scheme with forced air efficiently applied. Build one and see if it works. If it does move to multiple units over you MD tank.


Chemistry Guy; Yes super smooth is bad for heat transfer. Don't bother.

 

[stugray]

I posted this back on page 13, but didnt get a lot of comments.

It seems that it would be perfect for what wmilas wants to do.

These "bulbs" dont take up much more space than a t5, they are the same size as a VHO. I can buy this heatsink material at my local hardware store.

These have now been running for almost a month at the LFS with no fans.

The air temperature in the middle of the tubes gets to about 108 F.

I havent experimented again yet, but I am sure I can get higher density of LEDs if you used fans and ran the XP-Gs.





Here's my latest build:

I took ~1" X 1" aluminum channel & mounted 3W CREEs to it every 8 " using the drill & tap method with nylon screws & thermal grease.

They are 84" long and were built to go over a fish display tank where each tank is 8" wide.
Each little fishy tank gets one 3W LED.

Here's a pic of the 3 strings lit:

Unlit:

End detail #1:

End detail #2:

Drivers:

For the splash covers I used the tubes for covering VHO fluorescent tubes.
They are $4 at HD in 84" lengths.

I used the same nylon screws that I used to hold down the stars as "standoffs" to keep the tubes from touching the emitters.

Each ELN-60-48 drives 12 LEDs


For those VERY observant types: one tube has a brown wire running up it.
That is a thermocouple used to check the temp inside the tube.
Ambient temps were 85 & after running for a few hours the temp inside the tube got to 100.
The aluminum at one end was warm to the touch.

I dont think this setup requires fans, but using some very small CPU fans would increase emitter lifetime.

Stu

 

[AlgaeNator]

Nice solid build there stu,

Any pics of the thing installed and running, be interesting to see how well each emitter lights up each little 8" fish holding tank...

 

[wmilas]

I saw your build but it didn't really click until you just mentioned it again. Quite frankly thats a better setup than anything I've come up with so far. What I am imagining is a frame that holds those 1" aluminum channel stock runs. I could then cut the tubes (if I wanted to use them at all) along one side and bend them apart and "clip" them on the channel. Although frankly it might be easier to just put one large acrylic splash shield on the whole pendant by dropping it off the steel frame by say, two inches.

How does the 1"x1" aluminum stock do for heat dissipation? Your numbers look really good and with active cooling it definitely looks like it would work. I'm a bit amazed though that it would be "this easy". The thermal mass of that stock is much lower than the traditional heat sink material that most people are using. It would sure make things much easier if I could use it though as I would make it so each run would be 1 or two ELN-60-48's. Would make the whole setup VERY modular.

If I'm aiming for approximately 432 leds for the whole setup, thats 36 ELN's, or 18 per pendant. With some judicious placement I might be able to knock it down to 16 or even 14 per pendant thus yielding 336 to 384 LED's.

Because of my tank depth I'll have to use optics, probably 40 degrees. That would demand that I use a drop down sheet like splash guard.

Still this looks very doable. The hardest part is trying to figure out how and where to put the ELN's. The cleanest way would to be to mount them directly ontop of the channels where they are used. I'm nut sure about the safety implications of this though.

 

[kcress]

Stu thanks for dredging that back up!

Was that thermocouple attached to the aluminum or in the air near the aluminum? I'm curious because it makes a pretty big difference.


wmilas; Stu's configuration is really quite good. The tubing protects the LEDs and it would constrain the air flow to around the LED and on all sides of the aluminum channels.

While other methods would work too, none would be as bullet proof with regards to dead air spaces or eddies showing up in assorted locations over large areas of LEDs.

It also lends itself to a nice linear modular unit. Put it here - put it there.

The key is a good fan that can work with the high restriction presented by a long 'clogged pipe'.