Overheating LEDs ?... warning to fellow Noobs.

[kramer5150]

Nice relaxing day off work today...
I was playing around with my Dorcy light and measured 3.18V across the Nichia LED, (after the 4 ohm limiting resistor). Before the resistor, voltage measures a healty 3.8V. Hmm

The dork-in-me thinks Ooh.. extra .7V... gotta tap into that. Bust out the soldering iron and jump the resistor for 3.8 at the LED. After, 10-15 seconds the emitted light starts turning noticeably more blue. Fearing a meltdown I shut it down, and felt the LED with my nose it was warm but not hot. I reverted the circuit back and all is well now. No harm done.:thumbsup:

What happens to LEDs when they overheat? Lets say you feed it too much voltage? Are these little things that sensitive where ~.5V makes that much of a difference?

Thanks!!

 

[Marduke]

The extra current is probably what did it.

 

[mwaldron]

Many LED-based lights (especially the nice bright ones that we all love here at CPF) are being run EXTREMELY close to and in some cases beyond their maximum ratings.

Jumpering over the current-limiting resistor leaves the emitter to go hog-wild and will almost certainly result in .

As you noticed, when LED's are overdriven they change color, and dim significantly as they near death.

It may not have felt hot to your touch, but remember your touch is through the lens built into the emitter.

 

[Phaetos]

And what possessed you to test heat with your nose?

 

[kramer5150]

OK thanks for the replies. So... many LEDs when used in an unregulated circuit will automatically try and draw whatever current is available, and thermally self-destruct, correct? So, at the minimum they NEED at least a resistor to limit current am I correct on that too?

I've been itching to drop a bare Seoul emitter into this Dorcy 3AAA.

My understanding is that AAA cells naturally have a 500 mAh limit, but I should STILL use the 4 ohm resistor to play it safe... correct? Is there a better resistor value for this?

Theres a lot of room on the PCB, and I can expose the +/- PCB layers with my dremel. Not much room for a heatsink though. Can I get away with just thermal-epoxying the emitter to the PCBA? Or should I design a small heatsink too?

Thanks again!! I am new and really appreciate the help:thumbsup:

 

[kramer5150]

And what possessed you to test heat with your nose?

clueless noob... no better way to check for temps. I wasn't about to use my tounge.

 

[chimo]

What happens to LEDs when they overheat? Lets say you feed it too much voltage? Are these little things that sensitive where ~.5V makes that much of a difference?

Thanks!!

This has been gone over many, many times on this board, why not once more.

Yes, they are very sensitive to voltage, that's why it is best practice to drive them with a current source. As you change the current, the voltage across teh LED changes. A property of LEDs is that the V-I curve is very steep. That is to say, a very small change in voltage can result in a large change in current.

Lots of good information can be found with a search here, or check out this website for lots of good info (including results of over-driving).

 

[chimo]

My understanding is that AAA cells naturally have a 500 mAh limit,

I would not make that assumption, especially with NiMH cells. Try shorting one (momentarily) on the 10A scale of a DMM.

 

[Marduke]

OK thanks for the replies. So... many LEDs when used in an unregulated circuit will automatically try and draw whatever current is available, and thermally self-destruct, correct? So, at the minimum they NEED at least a resistor to limit current am I correct on that too?

I've been itching to drop a bare Seoul emitter into this Dorcy 3AAA.


My understanding is that AAA cells naturally have a 500 mAh limit, but I should STILL use the 4 ohm resistor to play it safe... correct? Is there a better resistor value for this?

Theres a lot of room on the PCB, and I can expose the +/- PCB layers with my dremel. Not much room for a heatsink though. Can I get away with just thermal-epoxying the emitter to the PCBA? Or should I design a small heatsink too?

Thanks again!! I am new and really appreciate the help:thumbsup:

If you're planning on just replacing the emitter, the PCB already has the circuitry to limit the current and voltage.

I can understand the nose thing. The nose and lips are two parts of the skin which are most sensitive to subtle amounts of touch and heat.

 

[Lighthouse one]

I have several lights that I run straight power to a Seoul Led. Some of the lights have the led on a star that was the base for a previous Lumileds...
I also have some homemade lights with 3 c cells direct to a Seoul led. In each case you should measure the tailcap current to be sure you aren't overdriving the Led. If the seoul is well heatsinked 1 amp or a bit more is ok.
My one light with a seoul- direct drive with a 3.7 volt lion 18650 runs 12 hours of slowly diminishing power. Even when it's down to 10 ma draw-it still gives enough light!
A seoul led will turn blue when not heatsinked properly- and live to run another day- if you cut the power and don't overdo it.

 

[oronocova]

OK thanks for the replies
...
I've been itching to drop a bare Seoul emitter into this Dorcy 3AAA
....

I have that same light and I did the same upgrade in almost the same way you described. I took off the old emitter and soldered in a ssc. Made a big difference. Then I went a little further and instead of replacing the stock resistor I just soldered the other joint closed and left the other resistor. I believe that makes it a parallel circuit with 4ohm on one side and 0ohm on the other which still gives you some resistance. I'm no expert so I'm not sure how long it will last that way. At any rate, I have been using it off and on to walk the dog for about a month and it is still going strong.

 

[Mr Happy]

OK thanks for the replies. So... many LEDs when used in an unregulated circuit will automatically try and draw whatever current is available, and thermally self-destruct, correct? So, at the minimum they NEED at least a resistor to limit current am I correct on that too?

My understanding is that AAA cells naturally have a 500 mAh limit, but I should STILL use the 4 ohm resistor to play it safe... correct? Is there a better resistor value for this?

An incandescent lamp has a resistance at zero volts of about 1/10th or 1/20th of the resistance at normal operating voltage. As you increase the applied voltage, the resistance increases to give some degree of current regulation (a fact that has sometimes been used in makeshift battery chargers).

An LED is quite different. At zero volts it has a very high resistance and will pass very little current at all. When you increase the voltage nothing happens until you reach a certain threshold voltage, when the resistance drops suddenly, the current surges, and the LED lights strongly. Small increases in voltage beyond this point can produce large increases in current. Secondly, the increased current heats up the LED, and the resistance decreases with increasing temperature. So you have a positive feedback loop leading to more current, more heat, lower resistance and yet more current.

Consequently, to safely drive an LED almost always requires a series resistor or some other form of current regulation.

Regarding current limiting on AAA cells, I have some NiMH cells that are delivering 0.75 A without breaking a sweat. It's usually the little button cells that have small value limiting currents.

 

[AndyTiedye]

I believe that makes it a parallel circuit with 4ohm on one side and 0ohm on the other which still gives you some resistance.

Nope. 4 ohms in parallel with a dead short is a dead short. You are probably relying on the resistance of the contacts and the wiring and the internal resistance of the battery.

 

[kramer5150]

I have that same light and I did the same upgrade in almost the same way you described. I took off the old emitter and soldered in a ssc. Made a big difference. Then I went a little further and instead of replacing the stock resistor I just soldered the other joint closed and left the other resistor. I believe that makes it a parallel circuit with 4ohm on one side and 0ohm on the other which still gives you some resistance. I'm no expert so I'm not sure how long it will last that way. At any rate, I have been using it off and on to walk the dog for about a month and it is still going strong.

Thats good to know, thanks. I think you are still jumping across the resistor though, just going about it in a different way on the PCB. At any rate, I think I will still use the resistor just to play it safe.

THANKS!! everyone who replied. Going to place my DX order tomorrow.