Pushing LEDs Beyond Ratings: Can My 25ma LED Survive?

[Topher925]

How far can you push most LEDs beyond their recommended rating before they burn out? I currently have a 25ma, 34k mcd LED that I am planning on putting 50ma through for 10ms durations for 30ms periods. The LED is inside of a aluminum block to help dissipate heat. Do you think and LED like this could survive this kind of operation for <100 hours?

 

[MATLABdude]

How far can you push most LEDs beyond their recommended rating before they burn out? I currently have a 25ma, 34k mcd LED that I am planning on putting 50ma through for 10ms durations for 30ms periods. The LED is inside of a aluminum block to help dissipate heat. Do you think and LED like this could survive this kind of operation for <100 hours?

Can you just spec out a brighter LED, or run with less current? You could just leave it on for 3 days and see if it works for that 100 hour spec. I'd bet it'd last less than the stated MTBF, however.

 

[Topher925]

No one makes a brighter LED at this wave length that still has a decent response time. I'll put one LED through testing and if it dies then I've still got 9 more.

 

[berkeman]

How far can you push most LEDs beyond their recommended rating before they burn out? I currently have a 25ma, 34k mcd LED that I am planning on putting 50ma through for 10ms durations for 30ms periods. The LED is inside of a aluminum block to help dissipate heat. Do you think and LED like this could survive this kind of operation for <100 hours?

I googled LED life current power, and got lots of usefful hits. Here's one:

http://www.philipslumileds.com/pdfs/WP12.pdf

.

 

[Redbelly98]

You could just leave it on for 3 days and see if it works for that 100 hour spec.

Good suggestion. Since the OP is a grad student, 33-hour days should not be a problem

 

[MATLABdude]

Good suggestion. Since the OP is a grad student, 33-hour days should not be a problem

I, too, am a grad student, and could also use 33 hour days!

(Miscalculated 24 x 4 as 72 instead of 96)

EDIT: D'Oh! 24 x 3 as 96 instead of 72

 

[uart]

How far can you push most LEDs beyond their recommended rating before they burn out? I currently have a 25ma, 34k mcd LED that I am planning on putting 50ma through for 10ms durations for 30ms periods. The LED is inside of a aluminum block to help dissipate heat. Do you think and LED like this could survive this kind of operation for <100 hours?

Is 25mA a peak rating or an average rating? If it's an average current rating then there should be no problems. LED's usually don't mind being driven with a pulsed current waveform. It's not uncommon to drive them in pulsed mode with peak currents 2 to 3 times the average rated current.

BTW. Is there any reason why you're using such a low frequency (approx 33Hz) for the pulses? Typically 100 to 1000 Hz would be more usual for LED pulse mode operation.

 

[Topher925]

I let one LED run all night at 50ma. I'll check and see if its still alive when I get to campus. 25ma is the rating for its max current, an average current isn't given on the data sheet.

I'm running it at 33Hz because the decay time of the phosphor its exciting is around 4ms and time is required for it to settle. I may end up using a sinusoidal wave form if I can get a decent wave form out of it.

 

[uart]

25ma is the rating for its max current

Then most likely it's the absolute maximum rating of the continuous current. This is not usually the same thing as the maximum peak current for pulsed operation. Tpyically the data sheet for a LED would contain data something like the snippet shown below.

Absolute Maximum Rating (Ta = 25C)

PARAMETER                                MAXIMUM RATING         UNITS

DC Forward Current                       30                       mA

Reverse Voltage (IR = 100mA)             5                        V

Peak Pulse Forward Current               100                      mA

Avg. Forward Current (Pulse Operation)   30                       mA

 

[Topher925]

The data sheet actually says "Absolute Maximum Rating" which I'm going to assume can either mean DC current or average pulsed current. The data sheet also has data going up to 80ma. I think as long as I stay below 80ma with my current pulse routine it should be fine.

 

[uart]

Yes keep the peak below 80mA and the average below 25mA and it should be fine.

 

[Phrak]

Typically, any silicon device has a max duty vs. frequency curve, whether it's published or not. I've move often seen this for collector or drain current. The curve depends on the bond-outs and die's ability to dissipate heat before it gets hit with the next pulse.

In general, the larger the die the lower the frequency for any given duty.