LED Array for a Pico

[Gomer]

Will do The beauty of so many strings is that if an LED does go, the current will only jump from 19.4 to 21.2mA if it ends up an open circuit. If I use the 700mA driver, I'd have to run 36 strings so 19.4 to 20 isn't much eaither. I suppose I could probably loose a couple of strings before it is stressfull on any LED, so I have a nice safe leeway.

[Hal]

What an absolutely fascinating topic! I know almost NOTHING about LEDs and must confess here that I'm not a nano reefer, but am captivated by this lighting subject.

 

And I know so little about electronics, but just have to ask a couple of extremely basic things.

 

1) Exactly how are you gluing these LEDs into the acrylic array. Neanderthal..I believe you said you used removable silicone. But exactly how do you squeeze the acrylic onto the hole and attach the LED without getting the silicone all over the wires? What is "removable" silicone and where do you get it?

 

2) All your pictures show very professional looking designs and wiring arrangments. How exactly do fabricate a wiring harness (connector?) that holds up to the saltwater environment.

 

Thanks.

 

Hal

[neanderthalman]

The holes in the acrylic were drilled to only be about ten mil larger than the LEDs themselves, so the fit is already fairly close. To clue the LEDs, I just put a dab of the silicone in the hole, and pushed the LED through. There's a small ridge that keeps the LED from going all the way through the acrylic.

 

The removable silicone is sold as "Draft Stop". The intention for the product is to seal all of your windows for the winter, then remove the seal in the spring. It peels off of almost anything without damaging the paint or finish.

 

I also didn't really make up much of a wire harness at all. The power cable is soldered directly to the array, and the other end of the cable is soldered inside the power supply. I didn't make up any connectors for this project. I gave little thought to the effects of saltwater on the wiring, and I have yet to see any noticable oxidation.

 

If I were to make up a connector that had to stand up to an environment like this, I would be purchasing one intended for industrial locations. Many industrial connectors are sealed against outside contamination, so the saltwater can't penetrate to the metal contacts.

 

HTH

[zachtos]

I'm worried about the saltwater rusting my array too. I dont plan to silicone any LEDs in place either, I dont want any contamination on the lens because that would skew the light emmited. I thought of siliconing all of the wires but that would cause a major problem for future maintenance. So I settle on just a simple acrylic cover between the water and the lens's in hopes that little or no salt creep will get onto the lens's. I will be wiring my array during my timeoff these next 2 weeks.

 

Zachtos's array

[jmt]

Will 9 9000mcd white LEDs be enough to light a pico (~1ga.)?

 

Thanks.

 

-jmt

[Gomer]

JMT, I'd say that you would need a lot more than that!

[neanderthalman]

The metal on LEDs won't rust in a saltwater environment, as the leads and solder don't contain any iron. The leads are aluminum, and aluminum DOES oxidize. This oxidization, while chemically similar to rusting of iron, produces a tough outer coating on the aluminum, as opposed to porous flaky rust. This layer of aluminum oxide will protect the leads from further oxidation. Aluminum is actually MORE reactive than iron, but takes a lot longer to be destroyed because of the protective nature of the oxide.

 

Similarly, the lead and tin in surface of the solder will also oxidize, and I beleive will enjoy the same benefits as aluminum.

 

Copper wires are also subject to fairly destructive corrosion, so it'd be best to try to use aluminum wiring. You could just seal up any exposed copper inside liquid electrical tape or silicone, but it's easier to just use aluminum. I also don't like bringing copper near my aquarium, for obvious reasons.

 

JMT, I've got 267 over a 1.6 - you're probably going to want more than 9 LEDs.

[Hal]

Hey Neader..

 

Thanks for the reply. Sorry I didn't see it sooner. Thought the bbs software would send me a notification. I must not have something configured right.

 

Anyway....

 

Sorry for still not understanding the glue up. Are you pushing the led itself through the silcon from the back side? If so doesn't that mess up the face of the led? Or do you somehow clean it off after setting up? If your pushing the led into the silcon from the front, don't the leads get coated with the silicone and make it difficult to solder?

 

The "ridge" that prevents the light from going all the way through...Is it on the led or the acrylic?

 

The holes in the acrylic were drilled to only be about ten mil larger than the LEDs themselves, so the fit is already fairly close. To clue the LEDs, I just put a dab of the silicone in the hole, and pushed the LED through. There's a small ridge that keeps the LED from going all the way through the acrylic.

 

The removable silicone is sold as "Draft Stop". The intention for the product is to seal all of your windows for the winter, then remove the seal in the spring. It peels off of almost anything without damaging the paint or finish.

 

I also didn't really make up much of a wire harness at all. The power cable is soldered directly to the array, and the other end of the cable is soldered inside the power supply. I didn't make up any connectors for this project. I gave little thought to the effects of saltwater on the wiring, and I have yet to see any noticable oxidation.

 

If I were to make up a connector that had to stand up to an environment like this, I would be purchasing one intended for industrial locations. Many industrial connectors are sealed against outside contamination, so the saltwater can't penetrate to the metal contacts.

 

HTH

 

Thanks..

 

And Merry Christmas everyone.

[neanderthalman]

The ridge is on the LED itself, and I am pushing the LEDs through from the backside. The silicone does get all over the lens of the LEDs, but it's removable. Once it's dry, it can be pulled off with your fingers.

 

I haven't received an email notification since the site upgraded the forum software....so I'm not surprised.

[Hal]

Thanks NM,

 

Now I've got it . Appreciate your patience.

 

And thanks for putting my mind to ease on missing email notices. I had gone back and updated my preferences, etc. and still didn't get notified you had answered.

 

For Christmas my wife gave me a Sylvania DOT it light made up of three led bulbs connected to three double AA batteries. Amazing how bright it is.

 

Thanks again and take care.

Hal

[Gomer]

Anyone order from Best HongKong on ebay?

 

How long did it take for you to recieve your LEDS? I ordered the 17th and got shipping conformation on the 20th. My Tracking number they sent me is still not a valid number. Also had no reply from their email system.

[neanderthalman]

I have not ordered from besthongkong, but I have heard of a number of satisfied customers. I ordered from ProfessorLED, based in Hong Kong, and shipping took the better part of two weeks. I'd anticipate slower shipping due to christmas traffic this time of year.

[Carinya]

With a 12v supply and bulbs with a min volts recommendation of 2.8 and a maximum of 4.0, would it be a problem to run them in series of 4? So 3v over each? I was thinking this would likely be a little dimmer than running them at 4v but it might help guard against blowing bulbs due to volt overages. Am I missing something or would this be OK?

 

TIA,

C

[Gomer]

keep in mind that the voltage per led is a voltage drop across it. Some people think it is safe to run with no resistor or current regulator for systems where the sum of the drop is less than the supplied. You can supply all the voltage you want to an LED and it will drop the same (unless you under supply). The issue on frying LEDs is current. LEDs in theory have zero resistance so they have infinate current. THe real world introduced resistances throught he wires so that is why you don't ahve true infinate current. Even a small resister is likely to be substantially larger than the led/wire string you have set up. My advice since most people are using a LOT of LEDs where the price adds up, is to have some sort of current control in place.

 

(I'm pretty sure the above is more or less accurate)

 

 

Here is a little calculator for a simple resistor based current regular

http://ourworld.compuserve.com/homepages/Bill_Bowden/led.htm

[neanderthalman]

^^ What?

 

Carinya, yes. You can take four LEDs with those specifications, wire them in series across a 12V source without a resistor. They will be dimmer than if you used three LEDs in series. This is the wiring configuration of my moonlights.

 

LEDs do not have zero resistance. At low voltages, they have a VERY high resistance, some would say it approaches infinite. At higher voltages, the resistance drops dramatically, and approaches zero. Between these areas of nearly infinite resistance and nearly zero resistance, lies an area referred to as the 'knee' of the IV characteristics.

 

In this 'knee' area, the resistance is not zero, and is not infinite, but is a finite number. What is NOT certain, is exactly what the effective resistance of the LED is. The resistance will change depending on the voltage drop across the LED. The specifications listed for LEDs give the voltage range of the 'knee'. In this case, 2.8 to 4V. Above 4V, the resistance quickly drops to nearly zero, and below 2.8V, the resistance quickly rises to incredibly high values.

 

If you re-read the whole thread Gomer, you'll see I've already proven this concept with my array. It's also the same concept that allows 5V LEDs to be driven by logic-level signals without a resistor.

 

Edit - Always wanted to quote myself. This is from another LED thread back in October. It's a better explanation of what I was talking about.

 

Nate, it isn't wise to make any absolute comments, 'cause some bastard (like me) is going to come along and prove you wrong. A wise man always leaves himself some wiggle room. When you say that something is ABSOLUTELY FALSE, or can't be done, or is impossible, well...it makes it far too easy to be proven wrong.

 

Resistors are only needed in an LED circuit when the source voltage is greater than the operating voltage of the LED. I am, for the moment, going to discuss only a circuit with a single LED. The concept remains the same for circuits with an array of LEDs.

 

The IV curve of a typical Si diode is given by the equation I = Is[e(qV/kT)-1]. This equation varies slightly for an LED, but the principle, again, is the same. Now, if you were to take the derivative of that equation, it will give you the slope of that function - the inverse of which is the resistance. I am not going to fill this post with the calculations, but you will find that for extremely low voltages, the resistance is incredibly high, and for extremely high voltges, the resistance is incredibly low.

 

Now, I ask the question, how does the resistance of the diode go from infinite to zero. There is a region, in the case of a Si diode, between 0.6V and 0.7V, where the resistance of the diode is not zero, and is not infinite, but is a finite value. That voltage region is where you are aiming for when you match the voltage of the power supply.

 

In the case of a blue or white LED with typical characteristics of 20mA @ 3.5V, the resistance according to my good friend Ohm is equal to V/I, or in this case, 175ohms. Funny...that's neither zero or infinite, and must lie in that region I spoke of earlier. Again, this region of curvature on the IV graph of a diode is where you are aiming. If you connect a 3.5V LED to a 3.5V voltage supply, it will behave as if it were a 175ohm resistor. LEDs can be modeled as voltage dependent resistors.

 

Also, Nate, your design theory is inefficient at best. Designing an array to use a specific power supply is difficult. A bettter approach is to look at your needs. You determine your light needs, which determine your LED needs, which determines your electrical needs, which determines your power supply needs. To apply this concept to the LEDs, first you decide how much light you want, or in this case, how many LEDs you want. Then, you determine how best to arrange them.

 

Based on how you've arranged your LEDs, you know what voltage they require and how much current they will draw at that voltage. If you haven't matched that to a typical voltage you then calculate what resistance to use, assuming the use of the next higher common power supply voltage. Now that you know the voltage and current requirements of your lights, you can start looking for an appropriate power supply. Power supplies are easy to come by in a wide variety of specifications. Also, a transformer isn't what you need to find around the house Nate, as LEDs run on DC power - you need an AC-DC adapter.

 

Lastly, I hope you like pudding. My tank is lit entirely by 267 blue and white LEDs, and I don't have any resistors whatsoever. You can view the full thread here. Neanderthalman's LED Array for a Pico. The proof is in the pudding, you CAN run LEDs without resistors IF you design your array properly.

 

In the case of Mr. Piranha's LEDs, if he wires all three of them in series, they'll require a 10.5V power supply to get 20mA of current through them. If he connects a 9V power supply, like his battery, then he will get less than 20mA of current. Running less current through an LED will not damage it, but it won't be as bright. In the case of moonlighting...that shouldn't be an issue.

[Carinya]

Thanks very much, to both of you. Neanderthalman, sorry to be a nudge but one aspect of my question is unanswered. If I wire them in series of 4 instead of 3, they'll be dimmer but will this in any way help guard against the bulbs blowing out?

 

I mean, would I be getting anything in return for dealing w/dimmer lights? Because if not, then I may as well do 3-bulb series.

 

Thanks again,

C

[neanderthalman]

Yes, the chances of your LEDs burning out from overvoltage are reduced. In the earlier pages of this thread, I had a small 50 LED array that I built. I hooked it up to a 12V power supply, but it turned out to be closer to 13.5V. That extra 1.5V burned out most of the LEDs because I was already running them at their maximum voltage. In fact, I was pushing them a little beyond the maximum - but the point remains the same.

 

If you run with only three LEDs, and you use a 12V power supply, you'll be running those LEDs at their maximum voltage. If the power supply isn't all that accurate, and gives even a small overvoltage, you run the risk of losing your LEDs.

 

If you run with four LEDs, your LEDs will be a bit dimmer, but you don't have to worry about the quality of the power supply so much. So long as the output of the 12V supply is less than 16V, you'll be fine. It'd be a big jump for a 12V supply to output 16V.

 

Just food for thought, even though the individual LEDs will be dimmer, you'll have more of them. Overall, you'll get a little less light - but not a whole lot less. You can breathe a little easier knowing that it's not as likely to go out on you.

 

Another tip, when LEDs burn out, I've noticed that they nearly always become an open circuit. The other LEDs in the string that are out should be just fine - the weakest LED sacrificed itself before the others were damaged. You can identify and replace the single burned out LED, without replacing the whole string.

 

Carinya, what are your planning on using for a power supply? Are you making moonlights, or is this going to be a larger daylight array?

 

Oh, just for interest - I'm setting up a second tank, and I got a 150W 20K HQI for it. It's modded into a regent halogen fixture - There was a huge thread on it on here a while ago. I mounted the MH over my 1.6 that usually has the LEDs, just to get a comparison. I was a little disappointed, as I saw no improvement in color or intensity.

 

After a few seconds, I was incredibly happy, as I realized that it meant that the LEDs were similar to a 150W MH. Unfortunately, I am away at school, without access to a digital camera or my lux meter. I'll try to remember to pick up the meter over reading week in February.

 

To the naked eye, I could tell no difference between the 150W 20K MH and my 50W LED array.

[Llamaguy]

if the power supply is too high for 3 LEDs and not strong enough for 4, you need a resistor. http://led.linear1.org/led.wiz

[Carinya]

Thank you so much. This makes sense to me but having no experience with electrical work, I wasn't sure about trusting my own common sense.

 

Carinya, what are your planning on using for a power supply? Are you making moonlights, or is this going to be a larger daylight array?

 

This project is to light the 4g mantis pico in my sig. When I made the tank I never finished the lights. It has been sitting under a regular desklamp with no corals. The mantis doesn't mind. But I am re-working my tanks to create a seahorse tank for macro & softies & I want to put my LPS in with the mantis. So obviously I need some lights in there. I had always planned it to be LEDs. I've had the supplies for a while, just not the time.

 

Thanks again for the help. I'm taking pics as I go & will post them when it's running. Might be a while though...

 

Edit: Oh and I was planning on buying an appropriately-sized wall wart power supply, probably from radio shack or similar.

 

also, llamaguy, per the manufacturer the bulbs are rated for a range of volts (2.8 to 4.0) and the power supply will have them at the low end (3.0).

[neanderthalman]

How many LEDs? I'm thinking that a typical wallwart won't be able to produce enough current for the size of an array that I'd imagine over a 4gal.

 

I'd recommend a computer power supply - great clean power, and you can pull one from an old scrap computer, probably for free. Open the case and cut off the bundle of wires real close, except for one yellow wire and one black wire, and the green wire if it's not an ancient power supply. The yellow is the +12V and the black is the common.

 

To turn on the power supply, take the green wire and connect it to the case. There's a perfect spot for it too. If you look at the connector where the power cord from the wall plugs into the power supply, you'll see a short green or green/yellow wire that is bolted to the inside of the case. Loosen the nut, slip the other green wire underneath, and tighten. Only slightly more involved than a wallwart, but with much better results.

[zachtos]

allelectronics.com sells a decent variety of powersupplies. I have tried four different supplies from their site, they are all less then $12 and accurate +/- 0.2V so far.

 

My own LED array is on delay due to holiday shipping

[Carinya]

How many LEDs?

 

Initially, 180. I know this is less density than you used. However, you did post (somewhere here on this long thread) that you though perhaps folks could get by with fewer than you used & still have a very bright/workable setup. So that's reason #1, just to try it & see. My pico sits right next to my 18g, which has a 150w MH, so I'm hoping I can make a reasonable visual comparison on the brightness. Reason #2 is that I think it'd be nice to have a sunrise/sunset effect and if 180 bulbs don't do the trick I'll add another set of bulbs to the mix & have them on a separate timer. If (and I figure WHEN) this happens I'll also have an opportunity to re-visit my blue to white ratio & fine-tune it. I want the tank pretty blue, but I may have gone too far. We'll see.

 

So on my 180 bulb set, it's 45 clusters of 4 bulbs each. Per your earlier post to me on a separate thread & per checking the maufacturer's specs, I am assuming each cluster will draw 20ma. Thus, I need a 900ma, 12v power source, which is doable I think.

 

That said, you DO have me thinking about a computer power supply option. My husband might be able to get one at work. I would guess the computer power supplies come in variable volts & amps...and that if he came home with one it'd be labeled? I guess this whole computer power supply thing just sounds more iffy to me because if he comes home with one (or even two), what if it's either not labeled clearly or what if it doesn't match my array specs? Are you sure it wouldn't be better for me to get 2 wall warts that match up exactly to what I need?

 

Appreciate your continued advice.

C

[neanderthalman]

If you're only going with 180, then your current draw will only be 900mA. 1000mA wallwarts are fairly easy to come by, and that's what I'd use. You want the power supply current rating to be just slightly more than you expect to draw from it, when dealing with wallwarts. If you use too large of a power supply on too small of an array, the voltage output can increase to compensate. This doesn't happen on most wallwarts, but I've seen it before. Computer power supplies contain more sophisticated circuitry that helps keep the voltage output stable at 12V.

 

I was expecting to hear somewhere over four hundred LEDs - where you'd have over 2A of current to supply. I've yet to see a wallwart that big, though there were some in-line supplies on allelectronics.com, as posted by zachtos.

 

Computer power supplies are standard. They output several different voltages on different colored wires. Red wires are +5V, Yellow are +12, and then there's also a -5V and a -12V in the mix somewhere too. Looks like you're not going to draw enough current to warrant using one though. I'd go with a 1000mA wallwart.

 

If you want to expand your array later, you can either use a separate power supply (sunrise and sunset) or just replace your wallwart with a larger supply.

[zachtos]

allelectronics is the best electronics store I've found online in relation to price and variety and reliability. digikey is great but expensive. If you find any better places let me know please.

[sammydee]

Never mind.