The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Modeling >> Semiconductor Devices >> Cyclostationary device noise
https://designers-guide.org/forum/YaBB.pl?num=1406605126

Message started by nrk1 on Jul 28^th, 2014, 8:38pm

Title: Cyclostationary device noise
Post by nrk1 on Jul 28^th, 2014, 8:38pm

How is cyclostationary device noise modeled in circuit simulators(Spectre in particular)? i.e. The shot noise of a BJT is 2qIc. Ic is periodically varying in a circuit. Is the output noise spectrum calculated as [average of 2qIc(t) over one cycle] times the relevant (LPTV) transfer function to the output?

Similarly, the MOS transistor's drain current noise is constant*|Qi| where Qi is the channel charge. Again, is it the average of |Qi| over one cycle that is used for noise calculations? What happens to the gate current noise in this calculation?

Thanks

Title: Re: Cyclostationary device noise
Post by Ken Kundert on Jul 31^st, 2014, 12:51pm

No. Shot noise is modeled as 2qI_c, meaning that noise is modulated by the instantaneous junction current, as you would expect.

-Ken

Title: Re: Cyclostationary device noise
Post by nrk1 on Aug 7^th, 2014, 7:08pm

I understand that the noise is modulated by the instantaneous current. But how is the spectrum modeled? In a time-invariant case, the spectrum of the collector current noise is 2qIc and at some other point it would be 2qIc*|H(f)|^2 and everything is independent of time. In the cyclostationary case, how is the time dependence eventually removed? (The final noise spectrum is independent of time, right). I was guessing that it is the average(over one period) of 2qIc(tau) or 2qIc(tau*|H(f,tau)|^2) where Ic and H are periodic in tau.

Title: Re: Cyclostationary device noise
Post by Ken Kundert on Aug 11^th, 2014, 12:05pm

It is more like the latter. The true cyclostationary noise on the output is computed internally by PNoise, and then the noise is averaged over one cycle and output.

-Ken