CIFF Delta-Sigma Modulator Design

[prcken]

Hi, I am doing a 3rd-order 2-bit CIFF DSM, I used Scherier's delta-sigma ToolBox to do the system design, ideally i can get 130dB SQNR with OSR=128. but i found the b1 and c1 coefficients are small so that the capacitors for the 1st stage integrator are very large when doing thermal noise budgeting. for this reason i increased the OSR to 256, this should lead to 149dB SQNR, ideally. but didn't really solve the big cap problem.
Now forget about the cap value first. I did Simulink simulation, i can get more than 20-bit ENOB even with OSR=128. But when I implement it in Cadence with ideal switches, caps and veriloga opamps, the ENOB is less than 15 bits with OSR=256. please see the comparison below.

I used the coefficients got from the matlab, why this two results are not consistent?
Any body has any idea about this?

Thanks!

 

[prcken]

alright, i found the problem, in some version of Matlab (eg.2009, R2010a), the hann function is w=.5*(1-cos(2*pi*(0:n-1)/n)); ds_hann in the toolbox also using this one.
but in some version of Matlab (eg. R2010b, R2013a), the hann function is using gencoswin function.
all because of the hann function is changed in newer version of matlab.
any comments?