LED Current Ratings

[Mosaic]

Glue is not what you want to use to attach the LED to the headsink. You need thermal paste/grease.

The glue will work as an insulator, keeping the heat IN THE led,

Thermal grease gets the heat AWAY from the led and into the heatsink.

RMMM, my friend, have a look here:
Thermal adhesive - Wikipedia, the free encyclopedia

 

[tom_pay]

Hi

Mr Audioguru, could you please explain this. I don't quite understand.

At a duty cycle of 50% the average current is half the peak current and the brightness is half.

Thanks

Tom

 

[RMMM]

RMMM, my friend, have a look here:
Thermal adhesive - Wikipedia, the free encyclopedia

Agreed. I should have been more clear.

Standard glue is no good in areas where a thermal glue or thermal paste and mechanical connections should be used.

Have you ever used any thermal glues for "high power" leds? Any tips for a decent product?

I would love to be able to skip the mechanical attachment.

 

[audioguru]

Mr Audioguru, could you please explain this. I don't quite understand.

With a 50% duty-cycle the LED is turned on for half the time and is turned off for half the time so its average power (and average heating) is half and it appears to be dimmed a little.

 

[tom_pay]

Ahh, Ok

So if i were to go to a 25% duty cycle, the heat would be a quarter. Is this how it works?

Would the light dim much?

Thanks

Tom

 

[audioguru]

Ahh, Ok

So if i were to go to a 25% duty cycle, the heat would be a quarter. Is this how it works?

Would the light dim much?

Yes the heating is reduced as much as the duty-cyle is reduced.
Your vision's response to brightness is not linear, it is logarithmic so you have a very wide range of sensitivity from starlight to sunlight. So 25% will not appear to be dimmed much. 1/10th power appears half as bright. You can see an LED glowing in the dark when its power is much less than 1/1000th.

 

[tom_pay]

Ok,

I knew that hearing was logarithmic, but vision I didn't. Well, you learn something new everyday

So I could even get it down to 15% and that would be 15/100ths of the heat. If I didn't mind it getting a little dimmer.

Another thing, should I put a capacitor across the LED? If so what value big, small or really small?


Thanks so much for your help

Tom

 

[RMMM]

Read up on PWM to see some waveforms and how duty-cycle affects the waveform. The amount of time that the wave is "high" is the amount of time that voltage, and therefore current, will be applied to the LED. (or whatever device/component you are feeding)

Also, you can get a bunch of schematics on creating and using PWM with LEDs. A simple 555 timer and a few caps and resistors is all you need. Throw in a potentiometer, and you can change the duty-cycle in real-time. That way you can test your LED to "see" what brightness level you can accept.

 

[cowana]

You do not want a capacitor across the LED when you PWM it - that will make it dimmer, rather than pulses of high brightness.

 

[RMMM]

You do not want a capacitor across the LED when you PWM it - that will make it dimmer, rather than pulses of high brightness.

If I caused that confusion, I apologize. I was referring to the bypass and timing caps for the 555.

 

[cowana]

You didn't

He asked if a cap should go across the LED in post 27 - you mentioned caps in post 28.

 

[RMMM]

You didn't

He asked if a cap should go across the LED in post 27 - you mentioned caps in post 28.

Ahh. I see. I missed that line in post 27.

 

[tom_pay]

Ok then,

No Caps across the LED.

Is it possible that when the LED is turned on, via PWM, the current will spike and go above the max current rating?

Tom

 

[RMMM]

No, LEDs dont function like motors or other inductive loads.

The PWM signal will average the current.

 

[tom_pay]

Ok,

So it's not going to hurt it.

Well, thank you everyone who has helped me, I really appreciate it.

Tom

 

[audioguru]

When LEDs are multiplexed they blink on for only a moment then the next one then the next one etc. So the LEDs have a low duty-cycle which causes their average current to be low and therefore they appear dimmed. So the current is increased until the average current is normal so the LEDs look normally bright. Then the peak current is much higher than the rated max continuous current. That is why the datasheet for most LEDs spec's a max allowed peak current at a certain max on-time.
My ordinary red LEDs are spec'd with a max allowed continuous current of 40mA and a max allowed peak current of 200mA for 100us.