Re: Frequency Domain and Noise Domain

Alain, Amro,

As a suggestion, there is an excellent Forum on Ken Kundert's Designer's 
Guide website (www.designers-guide.org/Forum) with a lot of traffic on 
both design/simulation related issues, as well as language design. There 
is even a specific VHDL-AMS section on the forum with all VHDL-AMS related 
discussions:

        http://www.designers-guide.org/Forum/YaBB.pl?board=vhdlams

By the way, I do think the discussion of the last weeks has been 
insightful for language discussions as well: if a feature is missing, is 
that a tool vendor's problem, or should the standardization community also 
be involved. Sometimes a solution chosen as elegant and efficient from a 
language design point of view might fail to connect with the user's view 
on that particular solution. That's not something that an ordinary user is 
able (or willing) to analyse and distinguish -- the user has a problem and 
it needs to be solved. If it turns out to be a vendor/tool related issue, 
we can still ask the discussion to be moved to, for instance, the 
Designer's Guide forum.

Regards,
Marq Kole



owner-vhdl-ams@eda.org wrote on 20-03-2006 08:56:19:

> Dear Amro,
> 
> John Shields correctly pointed out that the IEEE 1076.1 mailing list is 
> dedicated to issues related to the development of the 1076.1 standard. 
> It is not the place to discuss the pros and cons of tool 
implementations.
> 
> I therefore firmly invite you to discuss tool issues directly with 
> people or vendors and please do not put the vhdl-ams@eda.org mailing 
> list in copy anymore. Maybe there is a need to have a separate, 
> non-IEEE, mailing list to discuss such issues. Any proposal to set-up 
> such a "VHDL-AMS users mailing list" is welcome.
> 
> Best regards,
> Alain Vachoux
> P1076.1 chair
> 
> Amr Turk wrote:
> > Dear Ernst,
> > I understand what you want to point out. But Let's get into what 
happens 
> > in such a case. I made a resistor, in VHDL-AMS. Then I decided to use 
> > this model in any spice simulator like Eldo for instance. If I 
simulated 
> > the time and frequency domain. The resistor would give the output 
> > expected. But If I run noise simulation, due to that there is no noise 

> > domain in VHDL-AMS the resistor modelling in VHDL-AMS is not accurate 
> > and will give ideal output, while if proper modelling used, there 
should 
> > be noise.
> > 
> > I know that the above case is so simple, but it can be generalized on 
> > more sophisticated systems.
> > 
> > Thank you,
> > Amro Tork
> > 
> > On 3/17/06, *Ernst Christen * <Ernst.Christen@synopsys.com 
> > <mailto:Ernst.Christen@synopsys.com>> wrote:
> > 
> >     Amro,
> > 
> >     Your question seems to be about the VHDL-AMS language definition
> >     related to the concept of
> >     a domain.
> > 
> >     Quiescent domain, time domain and frequency domains are very 
general
> >     concepts. These
> >     domains are, as you observed, supported by the language 
definition.
> >     There can be several
> >     different simulation types in each domain, and simulation types 
may
> >     span domains. For
> >     example, IEEE Std 1076.1-1999 defines in 12.6 a simulation to 
start
> >     in the Quiescent
> >     domain and then enter the time domain. In SPICE, this would
> >     correspond to transient
> >     analysis (which in SPICE is preceded by an operating point
> >     analysis). Similarly, it
> >     defines in 12.8 two simulation types that both start in the
> >     Quiescent domain, then enter
> >     the frequency domain: small-signal frequency domain calculation 
and
> >     noise calculation,
> >     both of which correspond to SPICE analyses. Other simulation types
> >     are possible in each of
> >     the domains.
> > 
> >     Conversely, noise is a phenomenon of physical circuits. As such, 
it
> >     can be modeled and
> >     simulated. I have already discussed the IEEE Std 1076.1-1999 noise
> >     calculation, which
> >     occurs essentially in the frequency domain. However, noise
> >     calculation can also be done in
> >     the time domain, although such a simulation is not as well
> >     established as a frequency
> >     domain (small-signal) noise calculation. Each type of noise 
analysis
> >     requires a different
> >     way of modeling the noise effects. We conclude that noise is not a
> >     domain, but something
> >     orthogonal to it.
> > 
> >     Thanks.
> >     Ernst Christen
> > 
> >     On Wed, 15 Mar 2006 05:05:10 -0800, Amr Turk wrote:
> >      > Dear all,
> >      > I'm working on a VHDL-AMS noise source, that can be used in 
ADMS.
> >     As known that ADMS
> >     doesn't support noise sources in VHDL-AMS tell now. So I - with 
the
> >     help of Joachim Haase-
> >     created a sin source that generates noise on the original signal. 
I
> >     have posted it before.
> >     But the problem now, is I want this source to interact with the
> >     simulator according to the
> >     domain of analysis.
> >      > I mean that when working in the Transient Domain, No noise 
should
> >     appear.But when
> >     working in the Noise_Trans, noise appear. And also, when working 
in
> >     AC domain, no noise
> >     effects should appear. but in Noise domain, Noise effects should 
appear.
> >      >
> >      > The problem is that VHDL-AMS has only three types of domains :
> >     QUIESCENT_DOMAIN,
> >     TIME_DOMAIN, FREQUENCY_DOMAIN.
> >      >
> >      > So, I don't know how to differentiate between frequency domain
> >     and noise domain for
> >     example.
> >      >
> >      > Thank you for your support,
> >      > Amro Tork
> > 
> > 
> > 
> > 
> >