Anyone know what algorithm the Spice AC Noise Analysis uses?


Is it some spectral modeling synthesis? I.e. that it estimates the main signal using peak detection and subtracts those from the signal in order to get the noise?



Noise analysis in Spice (Berkeley Spice) is done by summing up the power spectral density from every noise source in the circuit. There are a couple caveats.

  1. The circuit is assumed to be linear. In other words the circuit is first solved for a specific DC operating point then each of the components' equivalent linear noise sources are modeled as thermal noise sources. Each of the components internal resistance values produce thermal noise whose frequency spectrum is shaped by the LINEAR response of the circuit.
  2. Each source is assumed to be uncorrelated with any other. That means that the power spectral density of the resulting noise is the sum of the power spectral density of each source individually. This is not necessarily the case when you have matched pairs of transistors that are closely coupled and share phonic coupling.
  3. The 1/f noise is a complex parameter model. 1/f noise is difficult to model but there have been several standard models developed. If you stay away from low frequency in the analysis you can avoid dealing with this issue.

See google for more info on noise models such as Noise Sources in Bulk CMOS.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.