Subject: Re: Determining bjt noise parameters for Spice models?
Date: Thu, 12 Dec 2002 22:15:58 +0100
Organization: Wanadoo, l'internet avec France Telecom
References: <3DF6537E.B0FC9DBE@worldonline.fr> <email@example.com>
NNTP-Posting-Date: 12 Dec 2002 21:15:58 GMT
X-Mailer: Mozilla 4.79 [en] (X11; U; Linux 2.4.20 i686)
Jim, Thank you. I have a few follow-up questions...
Jim Thompson wrote:
> AF and KF only affect low frequency (1/f) noise performance.
> Bipolar transistor equations for noise.........
> Noise is calculated assuming a 1.0-hertz bandwidth, using the
> following spectral power densities (per unit bandwidth):
> parasitic resistances thermal noise (AC terms on the left of the =
> sign, DC terms on the right).....
> RC: Ic^2 = 4·k·T/(RC/area)
I know that the termal noise of a resistance is that: Ic^2=4*k*T/Rc, what is
the "area" ?
> RB: Ib^2 = 4·k·T/RB
> RE: Ie^2 = 4·k·T/(RE/area)
Are RC, RB, RE the same as the RC, RB, RE specified in the bjt Spice model?
Do you mean that those noise currents are "flowing out" of B,C,E of the
> base and collector currents shot and flicker noise
> IB: Ib^2 = 2·q·Ib + KF·Ib^AF / FREQUENCY
> IC: Ic^2 = 2·q·Ic
To which resistances should these currents be applied? For instance, AoE
talks about the shot current across RE..
I've tried to do some simulation in Spice. For a bjt, it calculates some
output square noise voltage values called "RB", "RC", "RE", "IBSN", "IC",
"IBFN" ; I have not yet succeded in matching the calculated values of the
thermal (to RB, RC, RE) and shot noise to those (to ISBN, IC). Can you help?
> ...Jim Thompson