LED Temperature Coefficients When Used as Bias Sources

A resource for technical documentation. Datasheets, application notes, instruction manuals, books and links to resources are found in the Document Library.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

LED Temperature Coefficients When Used as Bias Sources

Post by mediatechnology »

LEDs are often used as bias sources in audio and the question of LED tempco often comes up. Most of the published work regarding actual LED tempco relates to high-power illumination, not bias current sources.

The following paper, "Junction temperature in light-emitting diodes assessed by different methods," characterizes LED tempco for various LED colors and uses the measurement of delta-Vf (the change in forward voltage resulting from tempco) to indirectly determine junction temperature. As bias sources, we benefit for the work of S. Chhajed, Y. Xi, Th. Gessmann, J.-Q. Xi, J. M. Shah, J. K. Kim, and E. F. Schubert of the Rensselaer Polytechnic Institute.

https://www.ka-electronics.com/Images/p ... Tempco.pdf

They found:
"The coefficient dVf /dT is equal to – 2.3 mV/K for the UV LED sample (? = 375 nm), – 5 mV/K for the blue LED sample (? = 455 nm), – 3.8 mV/K for the green LED sample (? = 530 nm), and – 1.5 mV/K for the red LED sample (? = 605 nm)."
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LED Temperature Coefficients When Used as Bias Sources

Post by mediatechnology »

E. Fred Schubert, one of the co-authors of the paper cited in the previous post kindly sent me this paper where they study delta-Vf vs delta-T for GaN UV LEDs. They found the relationship to be quite linear:

https://www.ka-electronics.com/Images/p ... Tempco.pdf
bcarso
Posts: 2
Joined: Tue Jul 22, 2008 2:33 pm

Re: LED Temperature Coefficients When Used as Bias Sources

Post by bcarso »

It's interesting that they found red to be such a low tempco. It must have been a different red (not "standard red") than I've used.

It would be costly to use UV LEDs for biasing! But how cool would it be when you are in anot-too-well-lit room and you see things fluorescing when brought near a board?
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LED Temperature Coefficients When Used as Bias Sources

Post by mediatechnology »

Brad - their "red" sample was 605 nm which should be orange. I had always read that it was ~ -2mV for "standard" red around 650 nm.

There's a chart here that's interesting:

https://www.oksolar.com/led/led_color_chart.htm

And chromaticity:

https://www.oksolar.com/led/led_tech-ch ... _chart.htm
bcarso
Posts: 2
Joined: Tue Jul 22, 2008 2:33 pm

Re: LED Temperature Coefficients When Used as Bias Sources

Post by bcarso »

Good point---that is not what I would call red at 605.

Yes, the original ref I recall was in a PMI app note, where it was asserted that the red LED (and at the time the only red there was I think) was a close match to a 2N3906 Vbe tempco, given the particular bias conditions they showed fot the 3906 use as a current source. Or was it a 3904?

I remember when I had to do an ultra-cheap level shift circuit that nonetheless had to work over the automotive temperature range to stay within a silly spec out of Ford. I got roundly criticized for daring to do this common-base stage with LED bias, and I was scrambling to find any support from the literature.

The major critic, who was almost universally hated but remained influential despite this, summoned up some ancient Mil material on the unreliability of LEDs. Of course they were being evaluated as light sources. I also had a hella time getting the purchasing agent to understand what was needed, and her to get the stream of reps to understand, what the requirement was. No I don't care how bright and wonderful your parts are...
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: LED Temperature Coefficients When Used as Bias Sources

Post by JR. »

Using parts outside their app note targets, is a little like using prescription drugs for diseases they aren't approved for. The manufacturer may change their process and never tell you as long as they continue to meet their published specs.

Sounds a little scary to me for mass production, but you know how that goes...

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LED Temperature Coefficients When Used as Bias Sources

Post by mediatechnology »

Using parts outside their app note targets, is a little like using prescription drugs for diseases they aren't approved for. The manufacturer may change their process and never tell you as long as they continue to meet their published specs.
Maybe so, but other than the bulk resistance effects - and I haven't correlated it - it would appear that Vf, wavelength, chemistry and tempco are all interrelated.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: LED Temperature Coefficients When Used as Bias Sources

Post by JR. »


Maybe so, but other than the bulk resistance effects - and I haven't correlated it - it would appear that Vf, wavelength, chemistry and tempco are all interrelated.
I'm talking out of my ass here... Lets see.. the wavelength has something to do with the electron orbit energy potenial so it seems X frequency would be associated with a given chemistry. I don't know enough about what the tempco mechanism is driven by to speculate. If it is intimately linked to wavelength, then that (wavelength) would be an adequate spec to hold purchasing to.

It won't stop the LED salesman du jour from trying to temp purchasing with some different piece.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LED Temperature Coefficients When Used as Bias Sources

Post by mediatechnology »

Salesman to PA: Yes red is good so "super red" is better.

Engineer to PA: But what about the wavelength and tempco?

PA to Engineer: Huh?
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: LED Temperature Coefficients When Used as Bias Sources

Post by mediatechnology »

Yes, the original ref I recall was in a PMI app note, where it was asserted that the red LED (and at the time the only red there was I think) was a close match to a 2N3906 Vbe tempco, given the particular bias conditions they showed for the 3906 used as a current source. Or was it a 3904?
It may have been a 2N2907 in this SSM2220 datasheet:

Image

Jung also used an LED as bias in this Line Amp here circa 1975:

Image
Post Reply