Hello guys, In a PNP current mirror does somebody know why the Noise Voltage at my Load decreases with the increase of the emmiter resistors value?

 

A resistor Johnson (thermal) RMS noise is proportional to the square-root of its resistance.

 

A resistor Johnson (thermal) RMS noise is proportional to the square-root of its resistance.

Yes, but in this case the noise voltage at my load decreases, i assume that it might be the Emmiter Resistor at the transistor if it´s bigger, there is less mismatch between both emmiters resistance and therefore the noise gain decreases, or myabe it coud be something related to less dissipated power by the transistor (smaller Vce) and ir produces less noise to the load. But I dont know

 

By current mirror I take it you mean this configuration where R2 is the load? If that is the case I am not sure what you mean by an emitter resistor:

 

Yes, but in this case the noise voltage at my load decreases, i assume that it might be the Emmiter Resistor at the transistor if it´s bigger, there is less mismatch between both emmiters resistance and therefore the noise gain decreases, or myabe it coud be something related to less dissipated power by the transistor (smaller Vce) and ir produces less noise to the load. But I dont know

yes but with resistors in the emmiters (with the same value)

 

By current mirror I take it you mean this configuration where R2 is the load? If that is the case I am not sure what you mean by an emitter resistor:

View attachment 286963

yes but with resistores at both emmiters

 

yes but with resistores at both emmiters

In other words this configuration where you are increasing both R3 and R4?

 

This might help.

 

In other words this configuration where you are increasing both R3 and R4?

View attachment 286965

exactly

 

exactly

what are you measuring the noise with a scope? Sorry for all the questions, just trying to understand the test setup.

 

what are you measuring the noise with a scope? Sorry for all the questions, just trying to understand the test setup.

it´s amplified many many times, and then it goes to a signal analyser

 

What’s the amplification. Define ‘many many times’

do you have any capacitors in the circuit?

 

it´s amplified many many times, and then it goes to a signal analyser

My theory would be it is current noise. If you increase both emitter resistors the programmed current goes down and the current noise measured in microvolts per volt across the load goes down, but this is a squared term here because not only does the current go down through the load but so does the voltage, thus it goes down due to both less current and less voltage across the load. Thermal noise is not a function of voltage or current so it would in theory remain the same.

 

My theory would be it is current noise.

My that I assume you are referring to shot noise?

 

My that I assume you are referring to shot noise?

Possibly, I got it from this link among others:

https://eepower.com/resistor-guide/resistor-fundamentals/resistor-noise/#:~:text=The vibrations of the electrons,contributor to noise for resistors.

It 'appears' they are the same thing, here is another link I thought was pretty good, gives a formula for it:

https://www.ap.com/technical-librar... is that,is highly dependent upon temperature.

 

My that I assume you are referring to shot noise?

I am starting to think they are not the same. From this quote:



And it appears to be based on the 'construction type' of the resistor.

Shot noise appears to be proportional to the square root of the current as opposed to current noise which is proportional to the current. Maybe someone with more knowledge than me can step in here with better information.

 

1/f noise is different. Current (Schottky) noise and Voltage (Johnson) noise are independent of frequency.
1/f noise is much more difficult to quantify mathematically. Each device will have its "1/f noise corner" specified on the datasheet, which is the frequency below which 1/f noise dominates. Above that frequency noise voltage/current is flat, below that frequency it rises at 6dB/octave as frequency falls.
Have you read Doug Self's "Small Signal Audio design"? If not I can scan the relevant section for you.
Whatever you do, don't type "Doug Self Small Signal Audio pdf" into Google, because that might find you an illegal copy.

 

1/f noise is different. Current (Schottky) noise and Voltage (Johnson) noise are independent of frequency.
1/f noise is much more difficult to quantify mathematically. Each device will have its "1/f noise corner" specified on the datasheet, which is the frequency below which 1/f noise dominates. Above that frequency noise voltage/current is flat, below that frequency it rises at 6dB/octave as frequency falls.
Have you read Doug Self's "Small Signal Audio design"? If not I can scan the relevant section for you.
Whatever you do, don't type "Doug Self Small Signal Audio pdf" into Google, because that might find you an illegal copy.

the paper you send me had some interesting things, thanks.