Juntion temp of LED Experiment

[JimboJones21]

I'm trying to find the juntion temp of an LED. I found a method where you run a small current through the led and measure it's forward voltage, then increase the current till the led heats up, then turn the current back down to the inital current and measure the forward voltage again. Then the difference between the 2 forward voltages divided by 2mV/C will give you the Juntion to Ambient temperature. The test says you you have to take the measurement in under 10ms when you turn the heating current down. So what I've done is hooked up 2 power supplies is parallel with diodes, one power supply at 10mA and the other at 300mA. I take a Vf measurement at 10mA, turn on the 2nd supply and leave it for 30min, then turn it off and take the Vf measurement again. I'm useing a data logger so my measurements are taken in less than 10ms, but my results seem to be very high. Has anyone else done this and could give me some help on performing the experiment? I've found some info on 'structure function', but this seems very complicated?

 

[itsmoked]

I'm seriously confused.. How can you run 300mA thru an LED without totally destroying it?

 

[IRstuff]

Why shouldn't it be high? If you've really run 300mA through a single LED, then you run over 1W into a VERY TINY device.

TTFN

 

[JimboJones21]

Check out these LEDs. Their test currents are 350mA.

http://www.luxeon.com/

I just wanted to know if anyone has done this test method. If so, could they tell me if what I've done is correct?

 

[IRstuff]

and did you get 185ºC like they claim?

Your approach is basically OK, however, I would do a direct thermal characterization of the diode T-V curve before accepting your results.

Any mV/ºC number has to be taken with a grain of salt. It usually applies to a relatively limited temperature range and is also partially dependent on the forward current and series resistance. You should put the device in an oven and run the temperature up in 10ºC steps or so, up to 185ºC, or so, and measure the forward drop.

Then, when you run your op current junction temp test, you don't have to extrapolation from essentially room temp data.

TTFN

 

[Marke]

Hi JimboJones21

2mV/C is a good rule of thumb for Si based devices, but I am not sure that this also applies to LEDs where the junction is made up of a different material. Where did this figure come from??

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[zeitghost]

We used a piece of kit to do this sort of thing with Mospower devices back in the 80s as a check on die attatch.

It seemed to give reasonable results, but I suspect that your 2mV/C value needs checking... it was about -2.7mV/C for the devices we were testing.