Jump to content
Cultivated Reef

LED Array for a Pico


neanderthalman

Recommended Posts

Take 3 of your LEDs.

 

If they don't burn up(they shouldn't), then you don't need resistors. If you want, put 27ohm resistors in line with the three LEDs.

 

Don't build an array without resistors without first setting up three of them to test to see if they burn out. If one of them burns out, then you need to include 27 ohm resistors with the LEDs, like in the LED wizard diagram.

 

Sorry for the late reply here you talk about 27 ohm resistor with 3 LEDs of the one i have and the wizzard in that site gives me 82ohms

led.jpg

 

withe the same LEDs any input please.

I dont wanna make any mistakes with this, thats why i am asking all this Questions.

thanks

Link to comment
  • Replies 872
  • Created
  • Last Reply
neanderthalman

That is a very good question to ask hyewiz - you had me stumped for a minute, 'cause I checked the math behind the 84ohm resistors and it's correct.

 

I had done the calculations assuming you were going to use the LEDs at their max voltage to get the most light out of the array. At 3.8V per LED, your total voltage across the three LEDs will be 11.4V. That leaves 0.6V across the resistor, which will have 25mA current. V = IR, or R = V/I. R = 0.6V/25mA = 24ohms. The next largest resistor is a 27ohm resistor. You can use any resistor between 27ohms and 100ohms or so, and the LEDs will still work. The smaller the resistor, the brighter the LEDs - part of the reason I didn't use any. Try putting 3.8V into the calculator - it should tell you 27ohms.

 

With a max forward voltage of 3.8V, you can probably run those LEDs at 4V - which is the case if you don't use resistors to lower the voltage. Think of it this way. Will the LEDs go out if they get 3.81V? no, 'cause if they did, the manufacturer would get a lot of returns on their product. What they do is make an LED that will work between 2.5V and 4V, and then tell you to use it between 3.0 and 3.8V. If that weren't true, I'd have three hundred burned out LEDs by now ;)

Link to comment

neanderthalman you amaze me with your knowledge, I have great respect to you, so i will go ahead with 48 LEDs (16x3) with 27ohm res. and i am going to try to find old printer PS.

 

thank you one more time.

Link to comment

Awesome work mans. I may have to finally order LED's for my 2.5G now, 300 LED's seem enough? They would be either the same as yours, or from besthongkong.

Link to comment
neanderthalman

Good luck with your project Lingwendil - 300 should be enough, though you'll be spreading them out a little more than I did. You'll probably wind up with them spaced similar to techdef's array. What are your plans for a power supply? Are you going to run with resistors or without?

 

Keep us posted on your project.

Link to comment

I plan to use a printer power supply, it has two leads and one common, 15v and 30v, I'll most likely use the 15v. I don't know yet if I'll use resistors... Probably only if I need them. I almost think I should use 350 or more LED's. The tank has a footprint of 6" by 12", so I have that much room to work with... I suck at math and such, so I need a way to figure out the spacing and such. What's a good CAD program? If a CAD program would be usable, I've never used one before... Another issue is the ratio of LED's to use. 3/2? 2/1? (white-to-blue of course). I've heard 50/50 is way too blue to look right.

Link to comment
neanderthalman

I didn't use any sort of CAD program to design my array...just drew it out on paper. I also make use of MS paint to draw schematics to post. Techdef used a CAD program as he was cutting the acrylic on an automated laser cutter.....still jealous, btw.....

 

I used 1 blue LED for every two white LEDs. I didn't put a whole lot of thought into it, I just figured that it would look pretty good. It's a nice crisp bluish-white. If you like the bluer look of a 20,000K bulb...then maybe 3whites for every two blues? It's kind of like asking whether you should get the 10K or 20K bulb...it's personal preference.

 

As for the spacing....if you have a 6" by 12" space to work with, and you put the LEDs spaced 3/8" apart like I did, you should have 15 leds x 31 LEDs, which is 465 Leds. Increasing that to 1/2" spacing gives you 11 LEDs by 23 LEDs, or 253.

 

7/16" doesn't work out to nice round numbers, so rounding down the number of LEDs, you can have 12 leds by 26 leds.....or 312.

 

With a 15V power supply, you can put the LEDs in series clusters of four at 3.75V across each, so there's no need for resistors. If you spaced them at 3/8", you'd have 116 "clusters" of four LEDs, or 464 LEDs, and one empty space. If you spaced them at 7/16", you'd be able to fit 78 clusters, for a total of 312 LEDs and no empty spaces. Lastly, a spacing of 1/2" would let you have 63 clusters of four, for a total of 252 LEDs and one empty space.

 

The easiest array to make will be the 7/16" spacing, as you'll be able to minimize your wiring by wiring them like the following diagram - very similar to how I reduced the wiring on my array. Note the direction change on the LEDs in the center column.

 

post-13532-1133551509_thumb.jpg

 

If you wire it this way, you'll only wind up with a grand total of eight wires in your entire project, which simplifies and neatens your array.

Link to comment

I am having a hard time picking my resistor sizes due to the fact the LED will give a different voltage based on how much current I give it. I can't decide until I test them apparently. It will either be 20mA @ 3.5V or 24mA @ 3.6V. I'm leaning toward the higher 3.6V mark, but if the diode runs w/ that resistor at 3.5V instead of 3.6V it will exceed the rated current of the LED. It's hard to figure this out.

 

I want to just do series of 6 w/ 24Vdc supply and a 100ohm resistor (should cause them to run at 3.6V and 24mA, but at 3.5V this would give 30mA)

 

hongkong LEDs

Link to comment
neanderthalman

The specifications of LEDs are always broad - there is no way to be certain exactly what the current will be. The only way to design a circuit using semiconductor devices is to base your calculations on the assumption that each LED is exactly like the typical values.

 

The calcualations, using the max forward current as a the target current, gives a 100ohm resistor. Set up six LEDs in series with a 100ohm resistor and put a 24V supply across them. If you're nervous, use a 150ohm resistor.

Link to comment

I'm all worried about the "True" lifespans of all of these LEDs. I've read in quite a few places now that manufacturers basically Lie about this spec. and a 100,000hr LED may last only 1000-10,000hrs... thats only maybe 1/3 of a year to 3 years for US!!!

Link to comment
neanderthalman

I've been wondering that too....mine don't seem quite as bright as they used to. I might just be accustomed to it now though.

 

I put my old PC hood on for a minute, the LEDs are still WAY brighter. In two weeks I'll be going home for christmas, I'll take some readings then.

Link to comment
I've been wondering that too....mine don't seem quite as bright as they used to. I might just be accustomed to it now though.

 

I put my old PC hood on for a minute, the LEDs are still WAY brighter. In two weeks I'll be going home for christmas, I'll take some readings then.

 

 

This article might be an interesting read about lumen maintenance http://www.lumileds.com/pdfs/AB07.PDF

Link to comment
neanderthalman

Thanks Gomer, that was an interesting read.

 

Some of their data seems.....odd....like greater than 100% output from their luxeons. I always take data offered by the company selling stuff with a grain of salt. Still, they quoted an independent study showing that typical white LEDs lose their luminance as time progresses.

 

I like Luxeons, but after designing my array to use them, when you take into account the size of the suckers and the heat sinks they need....I was able to pack just as much light into a small space with typical LEDs, and it cost me far less. If I was going to be spending the money to make this array with luxeons, I would have tossed that cash at a 70W HQI.

 

If the study quoted is accurate, then that means that after about 6000 hours, the intensity of my array will have dropped to 50%. That's over two years at 8 hours a day. 50% will also still be almost 6000 lux, a 15 times improvement over my old PC. I'm not too worried about it, but had I known about the loss of intensity when designing this thing, I may well have decided to bite the bullet and get luxeons or MH.

 

What I don't like about their brochure is they don't give any clear reasons as to why their LEDs last longer. They quote things that shouldn't conceivably make a difference in a short amount of time, or a difference at all, like oxidation of the leadframe material.

Link to comment

I apparently need to buy a lux meter before I complete my project so I can check on the output of the lights throughout time. I do realize that light output will go down as the LEDs heat up, because this drives their voltage/current down a bit. Also, operating near peak temperature will burn the LEDs out much faster, as neanderthalman will experience running at 4V. I've decided to "underdrive" my LEDs by running them at 80% output by limitting the current to 18mA instead of 25mA. This will make them run cooler, hopefully balancing out the degredation in lifespan caused by heat.

 

Now I must find a good cheap lux meter. I've seen em for about $50-60 that can detect either 10,000 or 20,000 lux. not sure which to go with yet.

Link to comment
neanderthalman

Go with the 20,000 lux model at least. I was getting almost 12,000 lux from my array when I first fired it up. I'm not too worried about temperature issues killing the LEDs early. With the ventilation fan, the LEDs run maybe five degrees or so above ambient. They CAN run at like sixty degrees above ambient.......

 

When you're planning your array Zachtos, remember that LEDs are a non-linear device. When you drop your current by 20%, you're probably going to drop your light output by a lot more, probably around 50%.

Link to comment

Well based on the spec sheet it says it should be 0.8 intensity. I will probably end up ordering smaller resistors down the road and soldering them in. I'll figure that out later. I did test my 24V 1.6A psupply, gives a fluctuation between 23.8-24.2V, the capacitors I ordered should smooth that out a bit but it wont hurt anything either. Ironically the LED array costs less then a 70W MH setup. And supposedly will give comparable lux and last 3-10 times longer. We shall see.

Link to comment
neanderthalman

that's the idea behind it, a cheap way to simulate MH lighting in a small space without producing a lot of heat. My array, at startup, was estimated to be about equivalent to a 150W or 175W. The only question is the lumen maintenance - it remains to be seen. As for the loss of intensity, I did some testing with my array and a variable voltage source. I found that the intensity dropped to 2/3 when I ran it at 10V instead of 12V.

 

Good idea on the caps - I don't think you'll have a problem with a 0.4V fluctuation divided over six LEDs, but the caps cost next to nothing, and as you say, they won't hurt anything.

Link to comment

Neanderthalman - bad news...

 

here we are a few weeks later and about 1/4 of my LEDs are still firing. WTF?

I guess either I got a bad batch, or I got my math wrong. Either way I'm pretty bummed. Next time I'll listen to you and skip the resistors...

 

hoping that theres just one fried in each series, then i might have enough leftovers to revive the array. I guess the thing to do would be to swap out for bigger resistors & try again?

 

myrrgh.

 

-jamie

Link to comment

LEDs fry because of too much current. That is why you need a way to make sure your current is under a threashold. Since the voltage drop isn't identical on each strand and hence a diffneret current, you might want to look at measuring each section and occasionally using a different resister where needed.

 

 

And a completely different question. I have a tank with a 10"square footprint (5g). I was looking at those 40,000mcd white LEDs you see on ebay. How many of them do you think would be a good number given your experience so far? I can easily run either 144 or 288 of them with 10% of them being the 8000mcd blues.

Link to comment
neanderthalman

That sucks about your array techdef. Having resistors shouldn't make them burn out, it should help prevent it. I'd go over the math again, and maybe use larger resistors. I found that most often, when the LEDs fail, they fail as an open circuit - saving the other LEDs in the string. Hopefully you can salvage most of them. I'm wondering if the problem might be your power supply. I blew most of my first array with a bad power supply. The other problem might be the heat inside your hood from the PC fixture. Does it get hot inside the hood? I think I remember ventilation fans in your design.

 

Gomer's right. LEDs fry because of the heat generated by too much current. If you can cool them effectively, you can run them at higher currents. There's a limit to it though. Too high of a current, and the temperature at the semiconductor itself will get too hot before the outer casing temperature rises at all. The opposite is true - you can't run as much current through them at higher temperatures.

 

I'd go with 288 in that space Gomer. I've got almost that many in a 6" square footprint, though they're only 12000mcd. 144 just doesn't seem like enough. You're using those luxeon current sources for this, right?

Link to comment

Correct. I have two 350mA and one 700mA. If 244, I can actually have them on different timers (say 9hr each with a 6 hr overlap) or something. I just wish the LEDs were cheaper LOL....I have a 2x13watt AH kit that can hold me over for now while I get the LR all settled in.

 

 

Sorry to further hijack your thread, but for the blue suplimentation, what wavelength should I look at. You can find "blue" LEDs from around 350-450nm.

Link to comment
neanderthalman

At the moment, this thread is all about other LED lighting projects...so...there's very little hijacking going on. I think it'd be a great idea to keep posting updates of your array in this thread, as it will keep all of the information in one place - similar to the wetworx nano-skimmer threads. Just for curiosity, what's the price tag like on those LED drivers? I know you already had them kicking around....

 

I went with 465nm blues, as they most closely approximate actinics. Actinics are in the 460nm range, and then another in the 420nm range. 460nm is more of a blue, while 420 is more of a purple. The 465nm give great color and fluorescence, IMO. The sunrise and sunset idea's good too - I'm planning the same with PWM.

 

I'm not familiar with how the LED drivers work exactly. I've got a question. Say you were running 144 LEDs on one 350mA current source, and your LEDs were wired in 18 strings of 8 LEDs - which is 360mA. The LEDs would be a touch underpowerd, but that's not the point. What happens if an LED burns out? Does the current in the other branches increase?

 

You'd have to lose a couple of LEDs, and therefore a couple of strings, but eventually won't the current source be forced to drive too much current through the remaining strings? Wouldn't that cause essentially a cascade that knocks out the rest of the array?

 

I guess you'd just have to be more vigilant in replacing LEDs as they burn out. As a side note - assume you were running the array I mentioned above, you'd be getting 9.8W of light (assuming 20mA @ 3.5V), and you lost a string. Woudn't more current would be pushed across the other LEDs, and assuming the voltage stayed the same, you'd still be getting the same light output from your array, 9.8W?

 

In fact.....the voltage across the other strings should go up slightly with the increased current. So would you get more power, and more light, when an LED burns out? Weird thoughts......

Link to comment

www.luxeonstar.com is where I got the drivers. There is hwoever a marked cheaper place if you buy in larger quantities.

 

iirc, when you burn an LED, you "melt" the semiconductor juntion so that it pretty much becomes a wire. So now the 8 LED string isn't dead but now runing with 7LEDs. Now, I am not completely sure how the 1 7string and 17 8strings compensate in voltage and current.

Link to comment

Archived

This topic is now archived and is closed to further replies.

  • Recommended Discussions


×
×
  • Create New...