On Mo, 22.08.2005, 13:09, claudio.girardi@... wrote:
> Hallo,
Hi!
> while trying to do a noise analysis with the enclosed schematic,
> I have found some strange results regarding the noise factor; indeed
> the circuit is a bit "unconventional" since the transistor is used in
> the "reverse active" region, but the simulator computes a negative
> noise factor at certain frequencies !
> I think the problems come from
> the computation of the noise spectal densities, where the involved
> currents (Ib and, most of all Ic) are used without taking the absolute
> values, so when the BJT is in the reverse active region Ic is negative
> and a negative power spectral density is computed (!). Or at least
> adding a couple of fabs() for Ic and Ib around the noise correlation
> matrix computation code seemed to help in obtaining more meaningful
> results. A similar problem is present with the JFET also.
Nice bug you found. Indeed.
I am about to apply an appropriate patch to cvs just now... I hope
this fixes the problem.
Cheers, Stefan.
