What happens if I exceed max forward current?

I'm working on a project that uses 5mm through hole red LEDs. I understand that increasing the forward current (IF) could increase the brightness, but I also understand that there is an upper limit for IF. I checked the data sheet I have and it says that the IF is 50mA, what if I push the forward current a bit higher, maybe somewhere between 51-60mA, would it do any good the LED brightness? Or it will blow out the LED? If it does blow out the LED, would it be right away or it will take a while? Is there a safe operating area where I could drive the current higher and not blow out the LED?

Thanks! : )

A 5mm LED can have thermal resistance as high as 200 C/W. That means that if you drive it at half a watt, the LED inside the plastic blob will be 100C hotter than the little legs. This temperature quickly degrades the plastic blob, making a black covering over your LED. Plastic blackens more quickly when hotter, so without any other problem, you'll have a dim LED after a few hundred hours.

If you have the little twig-thin leads going into the LED package (Plastic blob), I suggest you stay below 40 mA or use power LEDs. Thicker leads help get heat out of the LED.

If the LED die exceeds 150C, it is going to fail very soon. White LEDs give the pretty 'blue flash' failure mode. High temperature is often preceded by an angry bluish light as the phosphor overheats and acts differently. Red LEDs are less obvious; you usually just get a sudden burned smell after failure.

The datasheet, for some reason, assumes that you have soldered the tiny LED leads to a massive copper plate. This would keep the LED working more nicely at a high current. This is somewhat different from the usual installation, however.

In other words, it will die sooner, with more pain, but you can probably get away with it for a few dozen hours.

AnAppleSnail said:

A 5mm LED can have thermal resistance as high as 200 C/W. That means that if you drive it at half a watt, the LED inside the plastic blob will be 100C hotter than the little legs. This temperature quickly degrades the plastic blob, making a black covering over your LED. Plastic blackens more quickly when hotter, so without any other problem, you'll have a dim LED after a few hundred hours.

If you have the little twig-thin leads going into the LED package (Plastic blob), I suggest you stay below 40 mA or use power LEDs. Thicker leads help get heat out of the LED.

If the LED die exceeds 150C, it is going to fail very soon. White LEDs give the pretty 'blue flash' failure mode. High temperature is often preceded by an angry bluish light as the phosphor overheats and acts differently. Red LEDs are less obvious; you usually just get a sudden burned smell after failure.

Click to expand...

Is there a way to tell the thermal resistance from the data sheet or from the graph? I have several LEDs from SuperBrightLEDs, and this is one of them:
http://www.superbrightleds.com/more...ing-angle-12,000-mcd/283/#/tab/Specifications
Do power dissipation and the graph of forward current vs ambient temperature tell me which range I should be in? If so, how? (Sorry about this, I'm still new to LEDs)

AnAppleSnail said:

The datasheet, for some reason, assumes that you have soldered the tiny LED leads to a massive copper plate. This would keep the LED working more nicely at a high current. This is somewhat different from the usual installation, however.

In other words, it will die sooner, with more pain, but you can probably get away with it for a few dozen hours.

Click to expand...

By the way, I'm soldering the LEDs on a PCB so I'm worried if it will become too hot and need a heat sink eventually. I'm assuming what you're saying is that it might last for some hours but it will burn out very quickly?

Thanks! : )

I'm working on a project that uses 5mm through hole red LEDs. I understand that increasing the forward current (IF) could increase the brightness, but I also understand that there is an upper limit for IF. I checked the data sheet I have and it says that the IF is 50mA, what if I push the forward current a bit higher, maybe somewhere between 51-60mA, would it do any good the LED brightness?

Click to expand...

You might increase the brightness at a hugely proportional loss. This depends upon the LED maker and tech behind the making of the p-n junction wafer.

So the answer is in reality a 'no' until you can solve for those issues.

That red LED on Super Bright LEDs is poor. Intensity is relatively low at the tight beam angle of 8 Deg. Cree's C503B series red LEDs can hit over 20,000 mcd at only 20ma and a beam angle of 15 Deg so it is making many times the light of the LED in your link. They can run at 50ma as well. For continuous use, I'd keep it under 40ma. Best of all they are under 20 cents a pop.

I used to buy LEDs from Super Bright LEDs years ago, but they are behind and charge too much. $1 for a white 5mm LED? I can get the latest Cree 40,000 mcd white 5mm for 25 cents at Mouser.

polarpan said:

Is there a way to tell the thermal resistance from the data sheet or from the graph? I have several LEDs from SuperBrightLEDs, and this is one of them:
http://www.superbrightleds.com/more...ing-angle-12,000-mcd/283/#/tab/Specifications
Do power dissipation and the graph of forward current vs ambient temperature tell me which range I should be in? If so, how? (Sorry about this, I'm still new to LEDs)



By the way, I'm soldering the LEDs on a PCB so I'm worried if it will become too hot and need a heat sink eventually. I'm assuming what you're saying is that it might last for some hours but it will burn out very quickly?

Thanks! : )

Click to expand...

Keep the power low - This is the type with the typical 200C/W rating. At 0.125W, the LED die is 25C hotter than the leads. The leads are 0.5mm wires, very poor heat removal devices. Especially if you trim them flush to the PCB. Big copper traces and a decent lead length between the LED and the PCB help. But generally, keep the power low. Underdrive this type of LED.

LEDPunisher said:

You might increase the brightness at a hugely proportional loss. This depends upon the LED maker and tech behind the making of the p-n junction wafer.

So the answer is in reality a 'no' until you can solve for those issues.

Click to expand...

I'm assuming what you're saying is that exceeding the max brightness would greatly reduce the LED life, please correct me if I'm wrong! Thanks!

JohnR66 said:

That red LED on Super Bright LEDs is poor. Intensity is relatively low at the tight beam angle of 8 Deg. Cree's C503B series red LEDs can hit over 20,000 mcd at only 20ma and a beam angle of 15 Deg so it is making many times the light of the LED in your link. They can run at 50ma as well. For continuous use, I'd keep it under 40ma. Best of all they are under 20 cents a pop.

I used to buy LEDs from Super Bright LEDs years ago, but they are behind and charge too much. $1 for a white 5mm LED? I can get the latest Cree 40,000 mcd white 5mm for 25 cents at Mouser.

Click to expand...

Regarding luminous intensity, I read from some site that says something like "2000 mcd at 30 degree puts as much light as 8000 mcd at 15 degree", is that correct? Is there a direct conversion?

I checked ou Cree before but it seems like most of them require me to purchase in bulk, which I couldn't afford to do it now because I am still playing with different LEDs and see which one works out best. Is there any site that you would suggest me to try?

I am testing with 5mm white LEDs too, one from SuperBrightLEDs and another from KingBright, both do not seem bright to me. I am thinking of trying Cree for white LEDs now.

Thanks for the input! : )

AnAppleSnail said:

Keep the power low - This is the type with the typical 200C/W rating. At 0.125W, the LED die is 25C hotter than the leads. The leads are 0.5mm wires, very poor heat removal devices. Especially if you trim them flush to the PCB. Big copper traces and a decent lead length between the LED and the PCB help. But generally, keep the power low. Underdrive this type of LED.

Click to expand...

Okay this makes sense! but I'm a bit confused with the "lead length between LED and PCB" you mentioned, could you explain it a little bit?
Thanks again!

|=D long lead, a bit more pin to radiate heat
short lead, nowhere to radiate heat

The | is your fiberglass waferboard with trace bits of copper in it - No heat acceptance to speak of.


Candela is intensity (lux, cd). If you integrate intensity over area, you get output (lumens).The "degrees" is usually "Full-width to half intensity," or something. A cone with a doubled angle has 4x the area. Think of it in a square:

o - 15 degree optic

oo
oo - 30 degree optic (Twice as wide, twice as tall, 4x as broad)

However, we don't see lumens, we see intensity of light. The full moon shines a hundred trillion lumens on the Earth at night... It's about 4 lux (4 lumens per square meter) over 1 trillion square meters. So your "equal output" light will look rather dimmer with the more-spread output.

I did not leave space between LEDs and PCB because the housing for the board is very limited. I guess my only option is to increase the size of copper trace.

Thanks AnAppleSnail! This is very helpful! : )

If you look at the LED very carefully, you will see one lead has a cup containing the LED chip, the other lead just floats and has a tiny bond wire to connect to the LED chip. This lead with the cup is the one you want to heatsink. To use the water bucket with a hole analogy, extending the area of the copper trace is a way to increase the mass of copper attached to the LED lead, kind of like increasing how much water your bucket can hold by using a larger bucket. The larger bucket still drains water at the same rate as the smaller one, but with a larger bucket, you get more time before the bucket fills up.

You could find some solid copper wire or shape a chunk of copper and solder it onto the copper trace to give more thermal mass, in addition to increasing the area of the copper trace itself. Just be careful to attach it in such a way as to not short out other traces, and attach it with as much contact with the trace as possible. It's a hacky solution, but may help if you find yourself stuck in a corner.

Thanks for the explanation! And I didn't know you could solder copper wire to copper trace to act as a heat sink. Thanks again! : )