Why do my leds blink on this power supply?

[vanillasnake21]

I have 3 CREE leds wired over my aquarium, my old supply broker yesterday, so I found another one which was never used from my Netgear modem, it's supposed to deliver, 12v at 1a which is exactly what I need but when I plug my LED driver into them the whole thing blinks at about 1 flash per 2 seconds. They don't blink on and off, they get dim, then they get very bright. I'm sure it's not my wiring or the driver, I used another power supply and they were on fine. I thought it might be driver so I bypassed the driver and kept them on for about 10 seconds, they still blinked. When I measure the voltage from this supply it reads a constant 14.85 with no load, but when I plug the driver back up, during the jump the voltage goes up by about 1.5v. Why? Can I fix this somehow? I have a bunch of capacitors lying around, which I can use. I really need to get these light up as soon as possible as my plants will start dying off quickly.

 

[chopnhack]

I have 3 CREE leds wired over my aquarium, my old supply broker yesterday, so I found another one which was never used from my Netgear modem, it's supposed to deliver, 12v at 1a which is exactly what I need but when I plug my LED driver into them the whole thing blinks at about 1 flash per 2 seconds. They don't blink on and off, they get dim, then they get very bright. I'm sure it's not my wiring or the driver, I used another power supply and they were on fine. I thought it might be driver so I bypassed the driver and kept them on for about 10 seconds, they still blinked. When I measure the voltage from this supply it reads a constant 14.85 with no load, but when I plug the driver back up, during the jump the voltage goes up by about 1.5v. Why? Can I fix this somehow? I have a bunch of capacitors lying around, which I can use. I really need to get these light up as soon as possible as my plants will start dying off quickly.

First question that comes to mind is why did your supply die in the first place? Is it possible the LED driver has become faulty and is drawing too much current thus killing your supply?

Are you able to take a amperage reading while the circuit is powered to see what kind of draw is occuring?
Do you have a schematic or can you draw one up so that we can assist in troubleshooting?
Do you have a schematic or a link for the driver?

 

[vanillasnake21]

hey @chopnhack, thanks for quick reply, so in order: power supply died because I've soldered and resoldered it many times since internal cables came loose, I picked it up this morning and pulled on it by accident, it exploded like a firecracker I guess the two mains cables shorted, I resoldered it back up, and it exploded again, I don't want to touch it anymore. The led driver is a home made 317 with the right number of resistors. But I really don't think it's that, I plugged in another supply and it was drawing just fine, at least by the brightness of the LEDs. I shorted that one also after about an hour, all my power supplies are kind of in shambles, I have to spend more time making sure the cables don't touch. So now I'm down to this last, blinking, supply and I have to get it to work soon. Unfortunately I can't take amp reading, I have a really cheap radioshack mult, it only goes up to 200ma. I don't have a schematic, but I actually came here on help with the wiring and had great help, I have all the LEDs wired in series. The driver "schematic" is here: http://www.instructables.com/id/Super-simple-high-power-LED-driver/

I really do think it's the power supply though, because like I said I tried it without the driver, I plugged the leds in directly for a short time and they were still dimming/brightenning. What's surprising is that it's very consistent almost like on a timing circuit. I've read up, and I think it could be the discharge rate of the capacitors of that supply where they don't even out the voltage so I'm getting a consistent timing at every discharge, but that's just my guess. I would expect this supply to be more or less decent since it's meant for a modem, it's not a knockoff either it's an official Netgear supply with their logo on it, fairly well built too. But the very first thing that alreted me was the 14.98v reading from a power supply that's rated for 12v, I've seen supplies provide 0.3-0.4v more, but 3 volts higher then rating, a little fishy.

 

[davenn]

But the very first thing that alreted me was the 14.98v reading from a power supply that's rated for 12v, I've seen supplies provide 0.3-0.4v more, but 3 volts higher then rating, a little fishy.

not if it's an unregulated supply, that sort of output would be about right

its quite probable that you replacement 12V 1A supply is struggling to supply the required current
either that or your LM317 driver is failing

get a plugpack that can supply 12V at at least 1.5 - 2A

 

[KJ6EAD]

A definitive but potentially destructive test would be to connect the circuitry to a 12V battery with an abundance of reserve amperage. Either the lights will work as intended and the current consumption and heat generation could be measured or something will fail catastrophically.

 

[vanillasnake21]

Ok, I found a little 600ma adapter that I tested with again. The lights do light on and stay on just fine, with the driver. So it's not the cabling or driver. Thy're dim, as it's a 1.5A led @davenn, but the LM317 is rated at 1.5A, won't plugging in 2A destroy it? It's also running pretty hot as it is, when I was driving 5 leds @1A, even with a heatsink I had to have partially submerged in the cool aquarium water. In the meantime I got a CFL so I'm ok for now. But I did look up unregulated power supplies and it seems to be spot on. Is there any easy way that I can "regulate" it? I have a bunch of mosfets, transistors, and capacitors lying around will it take a lot of effort?

 

[BobK]

Did you ever notice that you can plug a 7W (58mA) night light into the same socket that you plug a 1500W (12.5A) heater?

Plugging something into a power supply that can supply more current than needed is not a problem. The device will use only the current that it needs, your power supply should generally be able to supply about 1.5X the current needed for best reliability.

Bob

 

[chopnhack]

Aside from the excellent advice already received, you stated some problems that have to do with assembly methods.

power supply died because I've soldered and resoldered it many times since internal cables came loose, I picked it up this morning and pulled on it by accident, it exploded like a firecracker I guess the two mains cables shorted, I resoldered it back up, and it exploded again

At the very least, you may want to pickup some shrink tubing to isolate the legs of the 317 (if you followed the instructables way of building).

Can you post a picture of your setup?
There may be some other suggestions that we can make, like connectors, binding posts, etc. there are lots of ways to inexpensively make your project safer and more durable!

 

[vanillasnake21]

Yea, I'll post some pictures, but don't expect to get anything out of them. I desoldered a few of the wires here and there, and even when they are soldered I just rely on bending them in the right shape so they don't touch, so it's confusing what's going where. To make it even more confusing they do generally blink when the wires touch like if the filter outflow from the aquarium is hitting one of the wires, but this time they weren't touching I made sure, the whole thing was on my lap while I was soldering it. @BobK thanks, I wasn't sure if that was for the current or voltage, so basically if I don't draw more then I should then I could use my laptop power brick which puts out 19v at 7A? I just have to make sure I get the ristors right?

Edit: So I used the laptop supply just like BobK suggested, and the remainder of my leds went out completely, like they won't even light up with a multimiter. I'm not sure what went wrong, I plugged in the numbers right, 19v out, running 2 leds at 3.3 each thats 19-6.6 = 12.4, then I needed the 317 to do 1.5A so I divided by 1.5 and got 8.2ohm, so i wired 9 1ohm resistors, and the whole thing went up in flames.

edit: sigh, nope, they're all dead now. All 6 of my leds. While were on the topic, is it safe to have a CFL floating an inch above water? I made sure there are splashes, but can the moisture crack the glass or something, my other aquarium have cfl I dropped the whole hood in the aquarium a few times but they didn't explode or anything. Since this thread is dead I'll post up pics anyways, and myable I can get general suggestions, on what to do better in my system, how to wire better etc.

the last one above is a co2 diffuser that I built using a brushless motor and an arduino, I had it on for about a week now non stop and it seems to be fine, but is there any possibility I can electrocute my fish like that? Will it wear away the motor eventually from spinning underwater? I'm running it at 5v.

Also, do you see the heatsink in the 3rd image that the 317 sits on? I usually dangle the bottom of it in the tank since it overheats by a lot, sometimes it submerges completely with the chip for a few hours. It actually runs much cooler fully submerged, but again I was wondering if I can electrocute my fish like that and if it will cause it to wear away?

 

[BobK]

Arghh! I am sorry that you lost your LEDs, but I did not advise you to do the calculation and do what you did.

The 317 circuit for a constant current driver does not involve the input voltage, so I don't know why you needed to change anything based on the input voltage. How did you wire it? If you had wired 9 1Ω resistors in series as the resistor in the LM317 constant current driver, you should have gotten 1.25/9 = 139mA which would not have destroyed the LEDs. Did you wire them in parallel? This would give a calculated current of over 10A! which it could not supply anyway, but it might have briefly supplied enough to kill them.

The current calculation for the LM317 in constant current mode is:

R = 1.25 / I

Setting I to 1A, the R should be 1.25Ω. This does not involve the input input voltage, but putting 1A through the LM317 at a drop of 12V is 12W which would have killed the LM317 without a heat sink.

Bob

 

[vanillasnake21]

Arghh! I am sorry that you lost your LEDs, but I did not advise you to do the calculation and do what you did.

The 317 circuit for a constant current driver does not involve the input voltage, so I don't know why you needed to change anything based on the input voltage. How did you wire it? If you had wired 9 1Ω resistors in series as the resistor in the LM317 constant current driver, you should have gotten 1.25/9 = 139mA which would not have destroyed the LEDs. Did you wire them in parallel? This would give a calculated current of over 10A! which it could not supply anyway, but it might have briefly supplied enough to kill them.

The current calculation for the LM317 in constant current mode is:

R = 1.25 / I

Setting I to 1A, the R should be 1.25Ω. This does not involve the input input voltage, but putting 1A through the LM317 at a drop of 12V is 12W which would have killed the LM317 without a heat sink.

Bob

so I've been using the wrong forumla this whole time, that's great. I'm not even sure where I got that one from. I went back to the page of the driver and you're completely rithgt, it's ref voltage 1.25v/ desired amps = resistance. But I did wire 9 in series, I didn't touch it, you can see how i wired it in the 3rd picture. This blows, I've been using this formula for like a week to change the brighness of the lights, and it seemed to be working...

 

[BobK]

I think it was "working" because the current was limited by what you power supply could provide. Once you upped the voltage and current capability of the power supply, the flaw was revealed. And yes, the picture does show them in series. Something must have shorted to blow out your LEDs.

Bob

 

[vanillasnake21]

I think it was "working" because the current was limited by what you power supply could provide. Once you upped the voltage and current capability of the power supply, the flaw was revealed. And yes, the picture does show them in series. Something must have shorted to blow out your LEDs.

Bob

well I appreaciate your help either way, I think you helped me fix the parallel issue intially, and you saw it through after it's demise. If only I had a decent station with a normal soldering iron not a gun like I have now this would have been so much easier :/

 

[davenn]

Thy're dim, as it's a 1.5A led @davenn, but the LM317 is rated at 1.5A, won't plugging in 2A destroy it? It's also running pretty hot as it is, when I was driving 5 leds @1A, even with a heatsink

that version of a LM317 isn't really capable of 1.5A ... that's an absolute maximum ... no wonder it's running hot

use a TO3 cased metal version and still heatsink it well

 

[chopnhack]

Some suggestions:

Use perfboard for all your components and then run insulated wire from that to the actual LED units. This way you can keep all your higher voltage wiring, including your mains plug or plug pack outside of the tank. You will have less moisture on your electrical parts too! You can mount the perfboard in a small box or put standoffs in the corners to raise up the connections on the bottom side of the board. For any other connections, either use some electrical tape or shrink tube to isolate and insulate them from each other. Your heat sink does not have to be mounted so low. I would suggest placing the bottom of the heat sink to the bottom edge of the plastic package. That way the heatsink will not short out on the other legs.

 

[KJ6EAD]

A common failure mode for LM317 current regulators that bears mentioning: if the resistor leg of the circuit gets shorted, the regulator will attempt to drive infinite current into the load so the load will see the supply voltage minus about 3 volts.