Designer’s Guide Community

gk

New Member

Offline

Posts: 2

input referred noise in timedomain pnoise sim
Aug 09^th, 2009, 10:50am

Hi,

I am trying to simulate noise a Switched Capacitor Integrator based on http://www.designers-guide.org/Analysis/sc-filters.pdf . For including sampled aliased noise, I use noise_type as time domain. At the end of the sim, I can see the output referred noise, but the simulator refuses to plot the input referred noise. Is there any way to either plot the input referred noise or any other way of calculating my SNR.

There is a similar post, but here wa no answer given for this.
http://www.designers-guide.org/Forum/YaBB.pl?num=1186710158

Thanks,
gk

Back to top

IP Logged

Forum Administrator YaBB Administrator Offline Posts: 145	Re: input referred noise in timedomain pnoise sim Reply #1 - Aug 10^th, 2009, 9:59am Use a voltage source or current source as the input source. Do not use a port. -Ken
Back to top	IP Logged

gk

New Member

Offline

Posts: 2

Re: input referred noise in timedomain pnoise sim
Reply #2 - Aug 10^th, 2009, 10:10am

Hi Ken,

Thanks for your reply. I am already using a voltage source as input.

Also, I simulated a simple R-switch and C to find the integrated noise and it gives me the right (and same) value for different R's, but as soon as I put a transistor, I get different values of integrated noise/PSD for different transistor widths. Isnt it just supposed to be kT/C and independent of R??

Thanks,
gk

Back to top

IP Logged

analog_ali

New Member

Offline

Posts: 3

Re: input referred noise in timedomain pnoise sim
Reply #3 - Oct 2^nd, 2009, 6:26pm

Hi,

In order to get the thermal noise power independent of "R", maximum number of sidebands (msb) in the sim. setup has to be changed together with "R". (As R decreases, msb should increase to take care of the aliasing.)

But did you eventually have any success in simulating the input-referred noise of the integrator? I have tried to do the same simulation, but have not been able to confirm the 2kT/C result for a single-ended integrator. I have tried both of the closed loop configurations discussed in the delta-sigma paper by Ken Kundert (Figures 10 & 11). None of the approaches lead to 2kT/C. I could only obtain this result when I reset the integrator every cycle (i.e. when the integrator effectively becomes a SC gain stage)

Any ideas why this is the case??

Thanks,

Ali

Back to top

IP Logged