Filter attenuation on pulse signal

I designed a 7-th elliptic filter sampled at 150 MHz, with 80 dB attenuation, 0.5 dB ripple and cutoff at 0.6 MHz

[bnb,anb]= ellip(7,0.5,80,0.6/75);

Then I generated a Cos/Cos2 pulse 0.80 us rise time and 3.4 us width at 50%. Then modulated it with 2 MHz sinusoid, and I applied the filter to it.

I checked attenuation at peak value of filtered modulated pulse and it resulted to be 41 dB instead of the expected 80 dB for which I designed the filter.

Where is the error?

Thanks in advance for support

Silvio

[Star Strider]

It is difficult to follow your code, so I cannot determine what the problem is with it.

However analysing the filter with freqz:

[bnb,anb]= ellip(7,0.5,80,0.6/75);

figure
freqz(bnb, anb, 2^14, 150e6)

demonstrates that it has the required 80 dB attenuation.

That leads me to believe that there is likely a problem with your analysis of it, not the filter design itself.

[Silvio zanchet]

Thanks for the answer. I've no doubt about the attenuation of the filter. I'm just wondering why filtering the pulse modulated at 2 MHz, which should be attenuated by 80 dB, its amplitude after filtering results to be attenuated only by 41 dB.

Maybe the transient response of the filter?

Hoping now is more clear.

Regards

Silvio

[Silvio zanchet]

Sorry, I just realized I attached the wrong file.

Silvio