Do LEDs generate audible noise?

[rock soderstrom]

Maybe a stupid question, but I wonder if LEDs produce noise and if so to what extent?

An example application would be the cathode biasing of a tube with LEDs or as operating status indicator in B+ PSUs?

What do you guys think?

 

[volker]

Maybe a stupid question, but I wonder if LEDs produce noise and if so to what extent?

An example application would be the cathode biasing of a tube with LEDs or as operating status indicator in B+ PSUs?

What do you guys think?

Yes they do, this is a relevant post that also has been referenced here in the past. If I remember correctly the takeaway was that red LEDs are the quietest parts.

Biasing: noisier than the equivalent resistor, see Merlin's later post (corrected)
Indicator: is shorted by the supply capacitors

 

[rock soderstrom]

Yes they do, this is a relevant post that also has been referenced here in the past. If I remember correctly the takeaway was that red LEDs are the quietest parts.

Biasing: less noise than the equivalent resistor
Indicator: is shorted by the supply capacitors

Thank you for the informative link and the clear answers!

 

[JohnRoberts]

Um no, the thread title is asking about "audible" noise. I have never heard audible noise from a LED.

JR

 

[Tubetec]

Is it possible under certain circumstances the cathode voltage could get driven low to the point the diode ceases to conduct ?
Could it effect how short term overloads are handled as a result ?

 

[saint gillis]

I've experienced on some transistor guitar pedals (fuzz-wah, this kind of things), that adding a led would add noise, and that I needed to filter the psu feeding the led iirc

 

[Michael Tibes]

I've experienced on some transistor guitar pedals (fuzz-wah, this kind of things), that adding a led would add noise, and that I needed to filter the psu feeding the led iirc

Hard to imagine, do you have an example? I could imagine that adding an LED as indicator leads to a change in the circuit due to the current drawn by the LED. In my opinion that does not mean that the LED is noisy, but the circuit gets noisy due to the added current consumption - which might change bias points or something alike in the worst case. But - in a Monthy Python spirit - always expect the unexpected ;-)

Michael

 

[saint gillis]

It's been a while, so it's not fresh in my head. But I think when putting a true bypass with a led on a classic wah-wah and a DC power adapter connector, the led was bringing an audible noise, and filtering the psu feeding the wah circuit with an RC filter (1K2 with 47µ) resolved the problem, it seems a massive filter but it was also efficient to filter the noise from bad smps power adapters and droped the voltage only by ~0,45v as the current consumed is not big. The led was powered directly fom the adapter, separately from the filtering for the audio circuit.

 

[JohnRoberts]

An LED is a semiconductor junction so will have electrical noise. In my experience they do not make any "acoustic" noise.

JR

 

[Doug_]

LEDS are often used in the constant current sinks in transistor long tail pairs. Blue are the quietest, dig around there are academic papers on the topic.

 

[JohnRoberts]

LEDS are often used in the constant current sinks in transistor long tail pairs. Blue are the quietest, dig around there are academic papers on the topic.

are they quieter for "acoustic" noise, or electrical noise?

JR

 

[Tubetec]

I presumed Rock meant audible noise within an electronic circuit , I never stuck a led in my ear to listen to it ,did you ?

 

[JohnRoberts]

"audible"​

adjective​

1. That is heard or that can be heard.
2. Capable of being heard; loud enough to be heard; actually heard.
3. Able to be heard.

Of course all semiconductor junctions have electrical noise. Electrical noise can be "heard" when added to an audio path signal, and monitored through a playback system.

YES I am being pedantic.

JR

 

[Chrisfromthepast]

Im constantly reassured that this place is in fact inhabited by sound guys.

 

[Tubetec]

Well lets say were using a red led to keep the cathode of a tube at +1.8v , in the absence of signal thats all very well ,
but wont the resistance of the led vary dramatically depending on the drive voltage seen at the grid ?

 

[merlin]

An example application would be the cathode biasing of a tube with LEDs or as operating status indicator in B+ PSUs?

Yes, LEDs generate a similar noise current to a preamp triode, so LED biasing will typically add around 3dB to the noise figure of a gain stage (much noisier than a simple bias resistor). You can fix this by bypassing the LED bias with a capacitor.

In the PSU the noise doesn't matter.

 

[merlin]

Well lets say were using a red led to keep the cathode of a tube at +1.8v , in the absence of signal thats all very well , but wont the resistance of the led vary dramatically depending on the drive voltage seen at the grid ?

Yes, but the LED resistance is so much smaller than the tube internal cathode resistance that it doesn't matter. Any 'LED distortion' is buried under the tube distortion.

 

[rock soderstrom]

Yes, LEDs generate a similar noise current to a preamp triode, so LED biasing will typically add around 3dB to the noise figure of a gain stage (much noisier than a simple bias resistor).

Thanks for your feedback. I always thought that LEDs noise is less than resistors. Of course, I only read that on the Internet, but it was often cited as a reason for using LEDs for cathode bias instead of a resistor. Interesting.

 

[merlin]

Noise current in a diode is given by Schottky's theorem whereas resistors generate Johnson noise. Suppose we have 1.8V bias at 1mA:
A 1.8k resistor would generate 428pA of Johnson noise current.
A diode would generate 2.5nA shot noise current (six times greater).

 

[volker]

Thanks for your feedback. I always thought that LEDs noise is less than resistors. Of course, I only read that on the Internet, but it was often cited as a reason for using LEDs for cathode bias instead of a resistor. Interesting.

Me too, I'll correct my first answer.