Strange pz-analysis

[Stefan Joeres]

Hi altogether,

I just tried to implement a 4th order lowpass filter function in cadence
like this :

V(vout) <+ laplace_zp(V(vin),{-1,0,-1,0,-1,0,-1,0}
{0.718,0.396,0.583,0.133,0.583,-0.133,0.718,-0.396});


So - I'd expect the pole zero analysis to plot me exactly the same ...
but ...
Poles (Hz)
Real Imaginary Qfactor
1 9.27873e-02 +/- 2.11676e-02 **RHP -5.12846e-01
2 1.14273e-01 +/- 6.30254e-02 **RHP -5.71005e-01
Zeros (Hz)
at V(outp,0)/V4
Real Imaginary Qfactor
1 -1.59121e-01 +/- 3.35491e-05 5.00000e-01
2 -1.59189e-01 +/- 3.35671e-05 5.00000e-01

The ac simulation didn't plot the expected curve either ... am I missing
something important ?

Why ??

Regards,

Stefan

 

[Andrew Beckett]

On Mon, 23 Jan 2006 16:56:13 +0100, Stefan Joeres <joeres@ias.rwth-aachen.de>
wrote:

Hi altogether,

I just tried to implement a 4th order lowpass filter function in cadence
like this :

V(vout) <+ laplace_zp(V(vin),{-1,0,-1,0,-1,0,-1,0}
{0.718,0.396,0.583,0.133,0.583,-0.133,0.718,-0.396});


So - I'd expect the pole zero analysis to plot me exactly the same ...
but ...
Poles (Hz)
Real Imaginary Qfactor
1 9.27873e-02 +/- 2.11676e-02 **RHP -5.12846e-01
2 1.14273e-01 +/- 6.30254e-02 **RHP -5.71005e-01
Zeros (Hz)
at V(outp,0)/V4
Real Imaginary Qfactor
1 -1.59121e-01 +/- 3.35491e-05 5.00000e-01
2 -1.59189e-01 +/- 3.35671e-05 5.00000e-01

The ac simulation didn't plot the expected curve either ... am I missing
something important ?

Why ??

Regards,

Stefan

It doesn't surprise me that the pz analysis doesn't extract the poles and zeros
of a behavioural laplace function - in essence this is a frequency dependent
component, and it will only be evaluated at a particular frequency. I think I
tried using this approach to see whether pz was working properly during the beta
phase, and R&D told me something similar...

However, I'd expect the ac response to be OK. I've not checked your example to
see if there's a good reason why not (I'm a bit pushed for time at the moment to
do the experiments).

Regards,

Andrew.

 

[fogh]

Andrew Beckett wrote:

On Mon, 23 Jan 2006 16:56:13 +0100, Stefan Joeres <joeres@ias.rwth-aachen.de
wrote:


Hi altogether,

I just tried to implement a 4th order lowpass filter function in cadence
like this :

V(vout) <+ laplace_zp(V(vin),{-1,0,-1,0,-1,0,-1,0}
{0.718,0.396,0.583,0.133,0.583,-0.133,0.718,-0.396});


So - I'd expect the pole zero analysis to plot me exactly the same ...
but ...
Poles (Hz)
Real Imaginary Qfactor
1 9.27873e-02 +/- 2.11676e-02 **RHP -5.12846e-01
2 1.14273e-01 +/- 6.30254e-02 **RHP -5.71005e-01
Zeros (Hz)
at V(outp,0)/V4
Real Imaginary Qfactor
1 -1.59121e-01 +/- 3.35491e-05 5.00000e-01
2 -1.59189e-01 +/- 3.35671e-05 5.00000e-01

The ac simulation didn't plot the expected curve either ... am I missing
something important ?

Why ??

Regards,

Stefan


It doesn't surprise me that the pz analysis doesn't extract the poles and zeros
of a behavioural laplace function - in essence this is a frequency dependent
component, and it will only be evaluated at a particular frequency. I think I
tried using this approach to see whether pz was working properly during the beta
phase, and R&D told me something similar...

However, I'd expect the ac response to be OK. I've not checked your example to
see if there's a good reason why not (I'm a bit pushed for time at the moment to
do the experiments).

Andrew,
this large signal explanation is not all that clear to me. Both ac and pz are
small signal, aren t they ? Can you please elaborate ?