PRIMARY LED lit Custom glass tank
- Thread starter zachtos
- Start date
zachtos
Active member
well some bad news. 6 of my LEDs died already which caused 6 arrays of 10 to go out. after investigating the cause, it has been determined that heat caused the LED array to reduce the internal resistance of the LEDs causing too much current to go into them. some strings were getting as high as 32mA when fully hot. they start at 25mA cool and then climb. the LEDs are rated for 30mA so who knows how much damage has been done. I changed out the resistor in the array to a much larger one nearly twice its size (150ohm instead of 91.5ohm). this gives a current between 22mA and 24mA when hot. much better on the LEDs, but I'm betting more will die.
I also changed out several UV bulbs and their resistors also, due to the same problem. The UVs are very confusing though, but I dont care, I'm just happy they work since you cant tell if they get more or less intense visually anyways.
I also picked up some frags:
2-kenya tree
1-ricordia
1-green mushroom
1-"hydro" neon green SPS
I wont bother light acclimation since the LUX is now severly reduced by lowering the current. It has dropped to nearly 10,000-12,000 LUX at the surface as opposed to the 20,000-18,000 LUX acheived before... a 60% decrease in light output apparently. I will get you more detailed numbers when I'm not so tired this weekend. I still think the LEDs are plenty powerful enough for this small tank, but am very discouraged for their use in larger tanks now. I do have an insane amount of corraline growth over the last week though. I cant really believe it can grow that fast! nearly my entire top portion of the foam wall is covered in speckles. My macro has grown grown nearly 30% more this week. This may all be slowing down though due to the lowered LUX numbers to help protect my lights. I'm just hoping more LEDs dont burn out.
I also changed out several UV bulbs and their resistors also, due to the same problem. The UVs are very confusing though, but I dont care, I'm just happy they work since you cant tell if they get more or less intense visually anyways.
I also picked up some frags:
2-kenya tree
1-ricordia
1-green mushroom
1-"hydro" neon green SPS
I wont bother light acclimation since the LUX is now severly reduced by lowering the current. It has dropped to nearly 10,000-12,000 LUX at the surface as opposed to the 20,000-18,000 LUX acheived before... a 60% decrease in light output apparently. I will get you more detailed numbers when I'm not so tired this weekend. I still think the LEDs are plenty powerful enough for this small tank, but am very discouraged for their use in larger tanks now. I do have an insane amount of corraline growth over the last week though. I cant really believe it can grow that fast! nearly my entire top portion of the foam wall is covered in speckles. My macro has grown grown nearly 30% more this week. This may all be slowing down though due to the lowered LUX numbers to help protect my lights. I'm just hoping more LEDs dont burn out.
badpacket
Premium Member
Question, when you took the original LED readings, are you sure they weren't so high because you were pumping max i through them?
Also, don't know much about these specific LED's, however once you get into higher power ones, the manf. usually specifies a heat sink for each LED to wick away heat build-up and the subsequent increase in i.
Have you considered a PCB with +/- traces running the length of the board? Doing it like that would do multiple things.
1. Allow you to run the LED's in parallel at a voltage that might be better controlled when thermal i increases.
2. Allow the LED leads to conduct heat onto the traces, maybe reducing heat and increase in i.
3. No need for holes to be drilled, just solder and point.
4. Easy replacement of dead LEDs.
You could also consider finding the sweet spot on the Luxeon LEDs and have fewer, more powerful LEDs ( 3W/ea), with 30' angle?
http://www.lumileds.com/products/family.cfm?familyId=2
Not being an electrical engineer, from my lrking on several electronics boards, I'm sure there is some way of setting up some sort of inverse voltage control to counter thermal i increases. Might be worth asking around.
Also, don't know much about these specific LED's, however once you get into higher power ones, the manf. usually specifies a heat sink for each LED to wick away heat build-up and the subsequent increase in i.
Have you considered a PCB with +/- traces running the length of the board? Doing it like that would do multiple things.
1. Allow you to run the LED's in parallel at a voltage that might be better controlled when thermal i increases.
2. Allow the LED leads to conduct heat onto the traces, maybe reducing heat and increase in i.
3. No need for holes to be drilled, just solder and point.
4. Easy replacement of dead LEDs.
You could also consider finding the sweet spot on the Luxeon LEDs and have fewer, more powerful LEDs ( 3W/ea), with 30' angle?
http://www.lumileds.com/products/family.cfm?familyId=2
Not being an electrical engineer, from my lrking on several electronics boards, I'm sure there is some way of setting up some sort of inverse voltage control to counter thermal i increases. Might be worth asking around.
zachtos
Active member
yes, I pointed out that less currrent means less LUX output. they were apparently running higher voltage then I measured initially with the multimeter.
I will get LUX readings again this weekend with the new safer current ratings and post them to compare. My guess is that they are still much more powerful then PCs but are not as good as 250W MH, they may be competitive to a 150W HQI MH though. I sent some of my extra LEDs away for a spectrograph reading, those results should be here in a few weeks courtesy of a fellow reef addict.
I considered printed circuit boards, but that would require too much money to get one printed, and the LEDs wont stay at a perfect 90deg angle on the board since the leads can bend. The soldering was the easy part. Bending/trimming and placing each LED in the correct order was the time consuming part. running LEDs without a resistor is VERY bad, a SMALL change in voltage gives a LARGE change in current. For example: my white LEDs measured about 3.4 Vdc at 28mA and about 3.43 at 32mA, 3.35mA at 22mA... now 30mA will kill the LED quickly. very touchy, so I still suggest using resistors and designing to run at 20mA instead of 22mA, the current will creep up w/ heat. I could make another 380 LED array in 8hours or less next time.
I will get LUX readings again this weekend with the new safer current ratings and post them to compare. My guess is that they are still much more powerful then PCs but are not as good as 250W MH, they may be competitive to a 150W HQI MH though. I sent some of my extra LEDs away for a spectrograph reading, those results should be here in a few weeks courtesy of a fellow reef addict.
I considered printed circuit boards, but that would require too much money to get one printed, and the LEDs wont stay at a perfect 90deg angle on the board since the leads can bend. The soldering was the easy part. Bending/trimming and placing each LED in the correct order was the time consuming part. running LEDs without a resistor is VERY bad, a SMALL change in voltage gives a LARGE change in current. For example: my white LEDs measured about 3.4 Vdc at 28mA and about 3.43 at 32mA, 3.35mA at 22mA... now 30mA will kill the LED quickly. very touchy, so I still suggest using resistors and designing to run at 20mA instead of 22mA, the current will creep up w/ heat. I could make another 380 LED array in 8hours or less next time.
zachtos, I want to thank you for all the work you have put into this project, and for sharing with us, step by step. I am sure you have answered alot of questions for many people, as well as, peeked more intrest into DIY aquariums and lighting. You have done a wonderful job.
fppf
New member
You can run LED's without resistors. But you don't use a voltage supply you use a current supply. The voltage drop across the LED's will always be the same given any current. The reason its changing on you is not because of current but because of heat. If you could control the junction temp and change the current you would not see any voltage change across the LED. Basic Diode and PN junction theaory. Thats how a lot of these temp sensors work.
Also custom PCB's may not be as bad as you think. I do custom PCB's often for home projects. $60 for a basic small board. It all depends on how big the board is.
And they make clips for LEDs and there are ways to install LED's on PCB's so there square. Pick and place machines do it all day long.
Also custom PCB's may not be as bad as you think. I do custom PCB's often for home projects. $60 for a basic small board. It all depends on how big the board is.
And they make clips for LEDs and there are ways to install LED's on PCB's so there square. Pick and place machines do it all day long.
zachtos
Active member
Lowering heat will help keep the current draw down, but I can only keep them so cool. I'm considering cutting holes in the hood to mount fans that draw air across the array instead... It will just look ugly though.
current drivers are $$$. I wanted to keep the project under $200, not spend 500 some dollars to get the same results. I'll sacrifice some headache for some money.
I'm hoping that by throwing in the 150ohm resistors instead of the 91.5ohm resistors that I will lower the overall current entering each branch because the resistors are "limitting" the current that is allowed to enter the branches. This will hopefully push them back under 25mA and farther from the unsafe 30mA range but does severly lower my LUX output apparently.
current drivers are $$$. I wanted to keep the project under $200, not spend 500 some dollars to get the same results. I'll sacrifice some headache for some money.
I'm hoping that by throwing in the 150ohm resistors instead of the 91.5ohm resistors that I will lower the overall current entering each branch because the resistors are "limitting" the current that is allowed to enter the branches. This will hopefully push them back under 25mA and farther from the unsafe 30mA range but does severly lower my LUX output apparently.
FiberLux
New member
Zachtos, I'd like to echo the comments from Reef4Fun. 
Anytime you have an exciting new technology, its takes a few people to jump in, take the heat and make some mistakes, so all of us can reap the rewards later on.
You are definitely headed in the right direction and I applaud your efforts. Keep it going and you will get it right!
Anytime you have an exciting new technology, its takes a few people to jump in, take the heat and make some mistakes, so all of us can reap the rewards later on.
You are definitely headed in the right direction and I applaud your efforts. Keep it going and you will get it right!
zachtos
Active member
thanks for the encouragement, I can only hope no more LEDs burn out on me with the new resistor size shown. It took 2 hours to change out all 45.
Does it make me a monster that I crushed and destroyed my emerald crab? I caught him eating polyps a few days ago despite his supplemental seaweed diet. And then again caught him eating a kenya tree so I caught him and mashed him good. No return for that demon!:mad2:
Does it make me a monster that I crushed and destroyed my emerald crab? I caught him eating polyps a few days ago despite his supplemental seaweed diet. And then again caught him eating a kenya tree so I caught him and mashed him good. No return for that demon!:mad2:
miatawnt2b
New member
By stringing enough LED's together, you can run directly off of 120V house voltage. I've been playing with an LED christmas light string... here how they wired it.
There is nothing special about the LED's on these light strings, they are standard LED's
First off, the lights are a string of 60 LED's with the string split into 2 halves of 30 LED's each wired in series. The first 30 LED's is wired to run from the first half of the AC wave, and the second half is wired opposite series to run from the bottom half of the wave.
I wanted to make some moonlights, so I ripped the string apart. I divided 30 into 120, and came up with each LED using 4 volts. A bit much for LED's that I've used in the past, but math doesn't lie, right?... this was my first mistake. I wired 3 of the LED's in series, applied 12VDC and within a min, the LED's cooked. DUHHH I thought, since each string of 30 was only running off of half the AC wave, the delta voltage is really 2V going to each LED which was a bit lower than I am used to for LED's, but it's a safe bet since specs on most mini LED's are roughly 2.5-3.2V.
So you can do something very similar here without transformers and such, just wire 30 3V LED's in series and connect them to 120V or run 60 total reversing the LED polarity for the back half of the string. Since the MAX Reverse voltage of an LED is usually about twice the forward voltage spec, you won't blow teh LED string with -120V part of the AC wave.
Great option for a 90Gal array and you don't need to worry about huge currents when using 12VDC. Since each LED is only running at a delta 2V, you also shouldn't have overheating problems which cause current spikes killing LED's.
-J
There is nothing special about the LED's on these light strings, they are standard LED's
First off, the lights are a string of 60 LED's with the string split into 2 halves of 30 LED's each wired in series. The first 30 LED's is wired to run from the first half of the AC wave, and the second half is wired opposite series to run from the bottom half of the wave.
I wanted to make some moonlights, so I ripped the string apart. I divided 30 into 120, and came up with each LED using 4 volts. A bit much for LED's that I've used in the past, but math doesn't lie, right?... this was my first mistake. I wired 3 of the LED's in series, applied 12VDC and within a min, the LED's cooked. DUHHH I thought, since each string of 30 was only running off of half the AC wave, the delta voltage is really 2V going to each LED which was a bit lower than I am used to for LED's, but it's a safe bet since specs on most mini LED's are roughly 2.5-3.2V.
So you can do something very similar here without transformers and such, just wire 30 3V LED's in series and connect them to 120V or run 60 total reversing the LED polarity for the back half of the string. Since the MAX Reverse voltage of an LED is usually about twice the forward voltage spec, you won't blow teh LED string with -120V part of the AC wave.
Great option for a 90Gal array and you don't need to worry about huge currents when using 12VDC. Since each LED is only running at a delta 2V, you also shouldn't have overheating problems which cause current spikes killing LED's.
-J
fppf
New member
Flashing an LED on and off wont hurt it. However when the voltage reverses if it exceeds the reverse voltage threshold and causes the diode to avalanch that will hurt it. If the LED's don't have a high enough reverse voltage you can just add a diode that does. Just remember that those LED's can switch that fast and they really do flash at 60 times a second, or 120 times a second if you run a full wave rectifier.
zachtos
Active member
sounds too expensive for me to gamble $50 in LEDs on. One more strand of LEDs has burn out. I'm hoping it was already close to death and that more will not follow... otherwise this is going to turn into a short project.
On a better note, all of my new corals are open and erect, the ricordia, mushroom, GSP, Kenya and hyrdra SPS all look good.
On a better note, all of my new corals are open and erect, the ricordia, mushroom, GSP, Kenya and hyrdra SPS all look good.
badpacket
Premium Member
Well yeah, however I know one place on eBay sells the LED's with the resistor already soldered on.
Great effort though, we may still be a year or more off from realizing this economically, but you definately get props.
BTW, check out sparkfun and some of the elec. forums. sparkfun has pretty good prices and if you follow some of the other boards there are several smaller houses which give great free/low-cost intro orders for small qty orders to get their name around. I bookmarked this one for some reason, not sure if they are cheapest or not.
http://www.pad2pad.com/
miatawnt2b
New member
I'm sure that phillips has done quite a bit of research on these LED strings considering they are an actual UL listed product. I simply took the string apart and drew its schematic.
As I said earlier, every LED I have seen has a reverse voltage that is higher than its forward... but FPPF is right, if you're worried about it, just put a cheap rectifier diode on there (A 1n4004 is good for 1A at 600 piv)
J
As I said earlier, every LED I have seen has a reverse voltage that is higher than its forward... but FPPF is right, if you're worried about it, just put a cheap rectifier diode on there (A 1n4004 is good for 1A at 600 piv)
J
fppf
New member
Actually most people don't know this but a DC shock is worse than an AC shock. A DC shock will just make all your muscles clentch hard and not let up. That makes it so you have no chance to let go. AC at least pulses so it does not make you clentch as hard. And with saltwater around its not that hard to get shocked with 24V DC. Also remember its not the voltage that kills its the current. It only takes a few mA to kill you as long as there is enough voltage to push it.
