how to estimate the noise caused by pcb

[xuexucheng]

I can use the method in the great books to lower the noise caused by PCB. however, I often design analog and digital mixed system. When I have finished the design, my boss will ask me how low the noise of the system. What I can say is just I don't konw. Yes, I can use the method to lower the noise, but it is very hard to give the quantity of noise caused by PCB. Do you have the noise estimation method or some examples or some other books or papers? Software simulation, I think, can also not give the answer, because the system have analog parts and digital parts and ADC parts. I can use the PSPICE and Cadence PCB simulation software.
This question is the most important question for me. ^_^

 

[steinar96]

If i remember correctly noise originating from the circuit itself is composed of at least
Johnson noise (thermal noise)
Johnson

Shot noise.
Shot noise - Wikipedia, the free encyclopedia

Flicker noise
Flicker noise - Wikipedia, the free encyclopedia

There are other types of noise related to certain types of devices but i believe they are hard to calculate and analyze.

You can use PSPICE to see how external noise affects your circuit with the noise simulation.

 

[FvM]

Strictly spoken, a PCB causes no noise. Apparentling, you're referring to the role of PCB design in propagation of interferences from digital to analog circuit parts.

Obviously, you can't analyze it, without knowing which interference sources and which pathes for their propagation exist in your design. E.g. interfering currents propagating through the ground plane, power supply decoupling and filtering, capacitive and inductive coupling. Without knowing about possible pathes, e.g. Cadence signal integrity tools won't help much. Cadence sales will possibly tell different, but it's mainly a matter of experience, empirical tests and measurement.

 

[xuexucheng]

oh
well, How can I do?
My boss need this!

Strictly spoken, a PCB causes no noise. Apparentling, you're referring to the role of PCB design in propagation of interferences from digital to analog circuit parts.

Obviously, you can't analyze it, without knowing which interference sources and which pathes for their propagation exist in your design. E.g. interfering currents propagating through the ground plane, power supply decoupling and filtering, capacitive and inductive coupling. Without knowing about possible pathes, e.g. Cadence signal integrity tools won't help much. Cadence sales will possibly tell different, but it's mainly a matter of experience, empirical tests and measurement.

 

[xuexucheng]

Thank you very much, FvM, you are a great person!

but it's mainly a matter of experience, empirical tests and measurement.
Could you give some detail explanation.
such as how to measurement, how to do empirical tests.
Thank you!

 

[FvM]

You didn't exactly specify what kind of "noise" your referring to. Because you addressed mixed signal designs, I understood mainly crosstalk of digital to analog signals, that appears e.g. as a noise floor respectively insufficient signal-to-noise ratio. But also violation of EMC rules by conducted and radiated interferences may be a problem. All kinds of "noise" can be measured. Identifying their source and effectively reducing their spread is possibly more complicated. That's where a designers experience comes into play.