Re: MOS11 noise implementation in Verilog-A

Geoffrey,

     Frequency would be used in NQS models. But I think we decided to
realize NQS models with lumped elements or with a Laplace element. This
is consistent with your approach for noise models and eliminates the
need to access frequency.

      I think this is fine for what we require today for compact models.

Ilya

Geoffrey.Coram wrote:

>Reposted for Sri:
>
>
>
>
>>Just a side note on $freq. $freq i am assuming would be synonymous to
>>$abstime in transient. I am assuming $abstime would return 0 in a AC
>>analysis. What would be good to have a generic function - something like
>>$domainval or $xval or something like that which would return either the
>>time/freq/voltage depending on the analysis being performed.
>>
>>"Geoffrey.Coram" wrote:
>>
>>
>>
>>>Ilya -
>>>I recall a discussion (don't recall the context) about writing
>>>noise PSDs -- and thus the need to allow access to the small-signal
>>>analysis frequency at which the noise analysis is performed.
>>>
>>>The main problem is in specifying what this "$freq" should do in
>>>other analyses: should it only be allowed in noise expressions?
>>>Are there other cases that one might think one wanted $freq in
>>>a transient analysis?
>>>
>>>The "j" would be another odd beast; you'd like to restrict its
>>>use to noise analysis, but I can imagine people wanting it for
>>>complex filters or for constructing a signal with a quadrature
>>>component.
>>>
>>>-Geoffrey
>>>
>>>
Received on Fri May 21 06:44:30 2004