Burning out LEDs...

[Pogue]

Hi all,

So I'm working away on my DIY led lights.

My plan is to have 3 channels.
#1 - 12 x 3W Royal Blue
#2 - 12 x 3w Whites (20000k & 15000K)
#3 - 4 x 3w UV
2 x 3w Red
2 x 3w Green
4 x 3w Royal Blue (just to fill it out).

My Channel #1 is working fine...nice and bright.
I haven't wired Channel#2 yet.

Channel #3 is blowing a RB bulb every time I plug it in, and I can't figure out why.

By my math:
The UV, Green, and RB can run 3-3.6 vdc.
The Red can run at 2-2.4 vDC.
So that's:
10x3.6=36.0
2x2.4=4.8
==40.8

I'm running my psu at 39.4vdc (lowest it can go).
(Direct powering it right now. I'm having problems with my arduino build so I cant send pwm to my ldd1000s)

I'm honestly surprised that its the RB blowing and not the Red.
Anyone have any ideas? I'm gonna get some resitors and see if that can help.

 

[mcgyvr]

If you aren't limiting the current then you WILL have problems. Stop doing that.

 

[km1das]

Are you running the LED's in series or parallel? Are you running them off of the LDD-1000H? I think we need a little more information, also, if they're in backwards they'll blow out every time, what are you using for LED's? Most companies mark the cathode (negative side), but I know that at least Cree marks the anode, making it a lot easier to accidentally install backwards. Thanks.

 

[Cymonous]

I would say the resistors will probably help as you more than likely have too much current running to the LEDs. Voltage is not the big factor here, but current is.

 

[oreo57]

Direct powering it right now. I'm having problems with my arduino build so I cant send pwm to my ldd1000s

Why? LDD's run fine (on full ) w/ no PWM signal..Just hook them to each string (I'm also assuming you have 4)
IF I understand what you re implying..

OUT OF curiosity.. is the blowing one heat sinked properly??? For some reason I picture you just dangling it to make the V(f)
Overheating will cause thermal runaway and blow the LED..

Oh and without the current limiting of the Meanwell..

My Channel #1 is working fine...nice and bright

I'm not too sure this would last too long either..

 

[fiestazs90150]

out of curiosity why green leds?

 

[Pogue]

Thanks for all your replys.
Ill see if I can answer the questions properly from my phone.

I am running 12 leds in series. ive checked several times and my + - wiring is correct. As is each bead on their stars.

I have no resistors in series.

I have a 39-48v (pot) 7.5amp power supply that I am directly connecting for testing.
I will be connecting 6 ldd1000 to the psu. One for each channel.
(3 channels per pannel, 2 pannels).

I will try connecting one of the ldd1000s to power the leds.

So from what im hearing the current is the problem.
I was always under the impression that it shouldnt matter how much current can be supplied, as the ciruit will only draw what it can consume.
Edit: this is assume that im limiting my voltage. Which I am.

(Ex. My channel #1,12 RB, used a current of 880ma) (that is a little high for a 3w led).

I have the leds stars screwed down onto a massive heatsink using thermal compound.

Ive added the green just because ive noticed a few "full spectrum" reef lights include them.

 

[Pogue]

In case you're wondering. This is not a solid block. It has fins. But it is heavy duty for sure.

 

[ravi197699]

I would run the red and greens on one channel dont mix the UV's in that I am runnign 12 of the red, green and cayans and they are all runnning on one channel and 700ma LDD drivers

 

[oreo57]

LED's are kind of funny.. Yes w/ constant voltage can run them
and to my understanding as the LED heats it will actually increase in resistance..thus regulating itself..
UP TO A POINT.. and that is the catch..
As the resistence drops the I draw increases.. Then the Vf drops.. then it heats some more.. then the Vf and R drops more and more current flows until "pop"....

Pretty simple huh...(well it may not be "exactly" like this but the results are the same.. )

The electrical resistance of silicon increases with temperature up to about 160 Celsius, then starts decreasing. When resistance decreases, it allows more current to enter through the overheated areas, in turn causing yet more compounding heat, eventually leading to total failure of the device and hence the term Thermal Runaway.

With adequate thermal management and the use of a constant current driver, an LED will not exceed its maximum operating temperature and will never approach Thermal Runaway.

Combine that w/ LED's having a very fast I increase at small V changes.. not exactly ideal..

http://www.maximintegrated.com/app-notes/index.mvp/id/3256

And efficiency increases as you drop I...
Well not any really good reason to run them at Max instead of just adding a few more LED's............

 

[km1das]

Well, for now I would say trying hooking up the LDD 1000 and see if that fixes it. Also, I would personally run each of the different colours/LED type on it's own channel, there are just too many differences between LEDs to say for sure that something weird isn't happening with them. Also consider your power supply, how well regulated is it and how certain are you that it's running at just 39.4V? It's mounted to Metal so be VERY careful of shorts, the thermal compound could even make it worse depending on what type it is and although the heatsinks of LED's are supposed to be neutral, don't assume that it is. If you can measure the current draw on that channel before the LED blows that could be helpful information. Also, if you're running the 12 3W RB led's all in series off of the 39.4 volt supply, a current draw of 880mA still isn't running them at a full 3 watts. Anyway, I wish you luck with this, the constant current regulators will probably fix it all for you though, so give that a shot. Thanks.

 

[Pogue]

Good info oreo57, thanks.

Also consider your power supply, how well regulated is it and how certain are you that it's running at just 39.4V? It's mounted to Metal so be VERY careful of shorts, the thermal compound could even make it worse depending on what type it is and although the heatsinks of LED's are supposed to be neutral, don't assume that it is. If you can measure the current draw on that channel before the LED blows that could be helpful information. Also, if you're running the 12 3W RB led's all in series off of the 39.4 volt supply, a current draw of 880mA still isn't running them at a full 3 watts. Anyway, I wish you luck with this, the constant current regulators will probably fix it all for you though, so give that a shot. Thanks.

PSU is set to 39.4v, set and monitored with my multimeter.
For the RB, I was running it at 43v initially. I dialed down the PSU to test the other channels.

I'm going to be hooking up an LDD and i'll test it. I'll find out the current and post the info.

Thanks again everyone.

 

[chill86]

I would have problems with the red and the green light on my uv channel. When I want uv, I want uv and uv alone.

 

[Pogue]

:headwallyamb you LEDs! Stop changing the RULES!!!!:headwally:

I've tested running all 3 channels with the LDD using different PSU voltages.
More and more confused!!!grrrrrr!!!!!

Results of my tests:

(sorry, don't know how to attach this better)

(In all the tests in above, only 4-6 LEDs lit up for Channels 2 & 3. Oddly it was a couple at the start and a couple at the end. The LEDs in the middle did not light. However, if I test the LED with a battery they all light up)

What I see from this is that mixing LEDs in series doesn't work well.
But I fail to see why that is. Especially with the whites which have the same vdc and mA specs.

I mean I've mixed 2 red, 1 white, and 1 blue together powered by a 12vdc, 800ma driver for my fuge and it works fine.

Weird stuff!!!

 

[km1das]

Well, based on the table, it looks like everything is fine more or less. I am a bit confused as to why channel 1 has the changes in the voltage and current drawn as input was increased, with an LED driver that shouldn't happen. Also, with channel 3 it is a little unusual for the voltage to have increased for the 2 higher PSU settings. With the Constant Current driver that shouldn't happen, so there may be something else going on there. Also, Which LED's weren't lighting up? Also, channels 3 and 4 are not drawing as much as they should be, so I'm not sure if there is something else wrong, but it appears so. Also, are the red LEDs 3W as well? If not then you'll probably run into some problems in the future anyway. Trust me, when it comes to electronics, especially LEDs, you can easily spend weeks trying to find a problem that turns out to be something super simple like one small short. Good Luck. Thanks.

 

[oreo57]

:headwallyamb you LEDs! Stop changing the RULES!!!!:headwally:

One thing to test.. W my digital VOM if I put it on Ohms you can get a small glow from the LED's when you go between the + and -..
sooo test to make sure you aren't grounding any to the heat sink..Put one probe on + then neg to block and aluminum pad side.. Reverse procedure.. see if any light up.
I've found, for me, it is rather easy to "oversolder" the wire so it is actually contacting the Al pad and of course the heat sink. I usually have to clean up an LED or 2.. Of course using 18ga wire makes this a bit easier to have happen..

As to testing, my understanding it is way more acurate to put a series resistor in, check v drop and multiply it by the change in voltage, rather than try to check mA directly..

That is, connect a 0.1 ohm resistor
in series with the LED, then turn the unit on, then measure voltage across
the resistor and use ohms law to calculate the current.

Somewhere it is the recommended procedure from Meanwell as well..apparently safer and more accurate..
More an FYI...

 

[Pogue]

Well thank you very much oreo57.

The more I thought about circuits, it made zero sense that leds 1,2 & 11,12 would light but not the ones in the middle. It had to be a short.

Sure enough, a simple continuity check found the problem spots.

I had tightened a couple screws down a bit too tight and they warped the nylon washers enough to create the short.
Easy fix, just doubled the washers at those locations.

Problem solved.

:thumbsup:Thank you all.:thumbsup:

Now I have another question regarding voltages.
So here's my test results at PSU set to 45vdc:


As you can see Channels 2, 3 are running a bit hot.
But the Channel 1 RBs are only running around 2.5w. Should I push the voltage up so Channel1 is running at 3w and then add resistors to channels 2&3 to bring them back down to around 3w?

thanks

 

[oreo57]

Well thank you very much oreo57.

The more I thought about circuits, it made zero sense that leds 1,2 & 11,12 would light but not the ones in the middle. It had to be a short.

Sure enough, a simple continuity check found the problem spots.

I had tightened a couple screws down a bit too tight and they warped the nylon washers enough to create the short.
Easy fix, just doubled the washers at those locations.

Problem solved.

:thumbsup:Thank you all.:thumbsup:

Now I have another question regarding voltages.
So here's my test results at PSU set to 45vdc:
View attachment 274960

As you can see Channels 2, 3 are running a bit hot.
But the Channel 1 RBs are only running around 2.5w. Should I push the voltage up so Channel1 is running at 3w and then add resistors to channels 2&3 to bring them back down to around 3w?

thanks

Well since you insist on your constant voltage setup and are asking for opinions.. I guess my "opinion" would be to bring the 2 down to your 3W leaving the blue as it is.. OR bring them both to the blue level currently.

My "opinion" is to never max out the LED's regardless ... efficiency decreases and problems increase..I also don't down hill ski or sky dive.. so there..

 

[Sulcata102]

interesting