Continue to Site

# simple transistor noise calculation and simulation

#### maria98

##### Newbie level 5
Hi,
i am looking for simple noise calculation of the transistor and cross verification through cadence simulation. I went through many equations, derivations, but did not find any simple example, which has a simple transistor with its calculation and verification. If any one know, please share. thanks.

Hi,

the way you ask .. lets me assume (that´s all I can do without you giving more details) you did not read or did not understand the documents.

It´s like asking: what´s the easiest way to build a formula 1 car.

There are so much parameters to your question, that a simple answer like "V_noise = n * applied_voltage" is not possible.

* what type of transistor you use, like BJT, MOSFET...
* transistor parameters
* what frequency bandwidth you are interested in
* what output measure you expect like V_pp, I_RMS ...
* circuit, voltages, currents, nodes ....
* and many more...

If there was such a simple answer .... why would one use the difficult ones described in the documents?

Klaus

Yes, you are right. Thanks

Simple to do with a simulator (Simetrix) :

Regards, Dana.

#### Attachments

• Noise-in-Transistor-Circuits-P-J-Baxandall.pdf
2.4 MB · Views: 54
Hi,

the Simetrix results ... Is this uV pp or RMS?
I guess they would differ at least by a factor of 5.

Klaus

RMS

Knight

Here is output noise at collector and device alone noise.

The prior table is accumulated noise over the frequency range simulated. The plot is "per-root-Hertz"

Regards, Dana.

Last edited:
What type of noise you talk about ?? Johnsson ? Flicker ? Shot Noise ? Random Walk ? Pop-Corn ? Pink ??
Noise is a huge chapter.

Random walk, Pink are external sources unless stipulated in Spice, and I see for BJT
spice has no such parameter for either, just one for flicker. PopCorn not "normally"
speced for a BJT, or for that matter most OpAmps.

In the prior sims the BJT does have a flicker parameter, and I am in process of confirming
Johnson is also evaluated in sim....

Regards, Dana.

What you need to read is "Low-Noise Electronic System Design" by Motchenbacher and J. A. Connelly. There are some worked-out examples that you can enter into any simulator of your choosing. My suggestion is that if you want to corroborate your results obtained by hand with those of the simulator, do not use a commercial transistor model, but rather simulate the transistor using a small signal pi or T model. This way you'll be able to follow more closely if your analysis is correct or not. Commercial transistor models take into account a lot more stuff than the one you would normally take into account in a hand written problem.