Designer’s Guide Community

LeonQQ

Junior Member

Offline

Posts: 17
LA

Eldo noise simulation problem
Apr 01^st, 2009, 4:04pm

Hello everyone, I am characterizing a subthreshold PMOS transistor thermal noise density in AMI0.5um technology using ICStudio Eldo simulator, the simulator using BSIM3V3 for both flicker and thermal noise.

I found that at the same drain current level my simulated thermal noise density is much lower than the theory value (say 12nV/sqrt(Hz) in compare with 34nV/sqrt(Hz)), looks like too good to be true

The theoretical calculation is based on Tsividis's book v^2 = 4kT*(1/2n)*(U/nI), where U is the thermal voltage, I is the drain current, n is the subthreshold slope factor (typical value of 0.7)

I don't know how accurate the equation models the subthreshold thermal noise, but several papers' measured input referred noise density is close to the theoretical value based on this equation.

Or the simulator is using a very bad model thus giving a wrong simualtion value? I am not sure and struggling on this for days. Thank you for all the valuable inputs.

BTW, the gm = n*I/U agrees with simulation very well with the n value of 0.7

Back to top

IP Logged

Geoffrey_Coram Senior Fellow Offline Posts: 1999 Massachusetts, USA	Re: Eldo noise simulation problem Reply #1 - Apr 2^nd, 2009, 5:05am What NOIMOD are you using? NOIMOD=1,3,5,6 use 8/3 kT (gm+gds+gmbs) (and 5,6 have another factor if vds < vdsat) NOIMOD=2,4 use 4kT ueff qinv / Leff^2 I don't recognize the equation you cite; I thought Tsividis used the qinv expression.
Back to top	If at first you do succeed, STOP, raise your standards, and stop wasting your time. IP Logged

LeonQQ Junior Member Offline Posts: 17 LA	Re: Eldo noise simulation problem Reply #2 - Apr 2^nd, 2009, 10:29am Thank you Geoffrey for the input. NOIMOD = 2 for my simulation. The equation is adapted from eq. 8 in Harrison JSSC 2003 (A Low-power Low-Noise CMOS Amplifier for Neural Recording Applications) He cited Tsividis's book.
Back to top	IP Logged

LeonQQ Junior Member Offline Posts: 17 LA	Re: Eldo noise simulation problem Reply #3 - Apr 2^nd, 2009, 10:34am Hi Geoffrey, I also checked the BSIM3V3 manual and found qinv is too complicated for hand calculation in the equation 4kT ueff qinv / Leff^2 Do you know any approximated expression for qinv which is good for hand calculation? Thank you again.
Back to top	IP Logged

Geoffrey_Coram Senior Fellow Offline Posts: 1999 Massachusetts, USA	Re: Eldo noise simulation problem Reply #4 - Apr 3^rd, 2009, 3:47am The BSIM3v3.3 manual, chapter 8, equation (8.9), cites the source of the qinv noise equation as T.P. Tsividis, Operation and Modeling of the MOS Transistor, McGraw-Hill, New York, 1987. Sorry, I don't have an equation suitable for hand calculation.
Back to top	If at first you do succeed, STOP, raise your standards, and stop wasting your time. IP Logged