Hello, in the transistor BJT pi hybrid model, sometimes they consider a resistor at the emitter (re), which is equal to 1/gm. Does this resistor produces thermal noise? If it is a fictional resistor it shoudnt produce noise, but is it a real resistor? gm=Ic /Vt re=1/gm
If somebody knows some things about it please help me out. Thank you.

 

Real resistors do produce a bit of noise, but usually it does not matter in most applications.
Effective resistors in models may or not produce noise, depending on the model and it's purpose.

 

Real resistors do produce a bit of noise, but usually it does not matter in most applications.

Johnson noise is real in ohmic conductors (resistors), and does matter in many applications (at least in mine). Don't dismiss it so easily.

OTOH, the hybrid transistor models are small-signal models used to approximate the function of a transistor under AC conditions. The "small-r" parameters are not physical ohmic resistors but the result of dv/di on a particular node at a particular operating point. To the best of my understanding, no Johnson noise is produced by these "small-r"s. (I'll be quickly corrected here if I am wrong.)

Generally, shot noise is the dominant noise source across PN junctions, and is inversely dependent on the square root of current across the junction.

 

https://eepower.com/resistor-guide/resistor-fundamentals/resistor-noise/


https://forum.allaboutcircuits.com/threads/super-moon-shine.100322/post-899995

a first-order resistor-capacitor filter on each diff input pair to reduce external noise from being part of the measurement. The needed bandwidth is very low. Those are just Mylar capacitors (10 nf 100 nf) and the resistors (1k) are low noise metal film on the analog side.

 

Base-spreading resistance (Rbb) is a real resistance in a transistor and does contribute to noise.
http://www.dicks-website.eu/low_noise_amp_part4/part4.html

 

Base-spreading resistance (Rbb) is a real resistance in a transistor and does contribute to noise.
http://www.dicks-website.eu/low_noise_amp_part4/part4.html

But how is Rbb determined from a datasheet? It's not the small-r parameter of the hybrid-pi TS was referring to.

 

But how is Rbb determined from a datasheet? It's not the small-r parameter of the hybrid-pi TS was referring to.

It's rarely included in the datasheet, as the author states "So, the Rbb is an important noise parameter, but almost no transistor datasheet does mention the Rbb value.", but when looking for a transistor for a moving-coil cartridge preamp, or something else with a really low source resistance and tiny output voltage, it the number you need to know.

 

It's rarely included in the datasheet, as the author states "So, the Rbb is an important noise parameter, but almost no transistor datasheet does mention the Rbb value.", but when looking for a transistor for a moving-coil cartridge preamp, or something else with a really low source resistance and tiny output voltage, it the number you need to know.

Funny how such an important number would be neglected by the manufacturers.

 

Funny how such an important number would be neglected by the manufacturers.

Often low-noise transistors are discovered by chance, being designed for some other purpose. A classic is 2N4403 which was a very ordinary switching transistor until it was found to be super low noise. A lot of the Ferranti (Zetex) medium power switching transistors are low noise, because they were also designed with high gain.

 

Often low-noise transistors are discovered by chance, being designed for some other purpose. A classic is 2N4403 which was a very ordinary switching transistor until it was found to be super low noise. A lot of the Ferranti (Zetex) medium power switching transistors are low noise, because they were also designed with high gain.

I've got my misgivings.

If they're not publishing Rbb, or the resulting noise figure, then they are not guarantying it or even testing for it.

How can one confidently design-in in a "discovered" low-noise part and expect that it will continue to perform as well into the future?

 

I've got my misgivings.

If they're not publishing Rbb, or the resulting noise figure, then they are not guarantying it or even testing for it.

How can one confidently design-in in a "discovered" low-noise part and expect that it will continue to perform as well into the future?

That's always a risk one takes, but the noise characteristics of such transistors as 2N4403 have been well known and stable for many years, so one can assume (though perhaps not safely assume) that they will remain so.
Have you ever tried to buy a "low noise" transistor? There are plenty that claim to be "low noise". Some of those also don't specify the noise in the datasheet. Some specify a noise figure at a single source impedance and collector current. Finding a "contours of noise figure" graph is a rarity! And when you do, you tend to have found a transistor that is low noise at high input impedance and low collector current. Finding one that is low noise for low source impedances tends to be guesswork. Look for one that has lots of gain and a low Vce(sat) and it is likely that it is low noise.

 

I've got my misgivings.

If they're not publishing Rbb, or the resulting noise figure, then they are not guarantying it or even testing for it.

How can one confidently design-in in a "discovered" low-noise part and expect that it will continue to perform as well into the future?

Usually if you buy "ultra-low" noise transistors for special applications, they give you the r_bb , the only problem is that in the usual models they always discard r_e, beacause r_bb is much larger than r_e. But in this "ultra-low" noise transistors r_bb and r_e are not so far apart, because r_bb is very small. i've already discovered the answer to my questio, rather better to share it with the interested.

 

For many applications some analog noise does not matte at all, and for some it does not matter much.
Parameters that are not given on the data sheet are usually neither promised nor even tested for. So there is another reason for designers and engineers to fully understand and read device data sheets.