The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> High Q SC biquad filter architecture question
https://designers-guide.org/forum/YaBB.pl?num=1184741812

Message started by skythunder on Jul 17th, 2007, 11:56pm

Title: High Q SC biquad filter architecture question
Post by skythunder on Jul 17th, 2007, 11:56pm

Hi,all,I am a new comer for help.
Here is a bandpass filter with a higH Q biquad architecture, and this is from Allen's notes.

Obviously, this biquad contains a charge amplifier(the first stage) and a noninverting SC integrator,plus a switched capacitor and a capacitor as feedback into input.

I think this filter has no regular gain. In the first stage, the gain is the ratio of C0 to C1 (C1 is much larger than C0);
In the second stage, the gain of the integrator is related with the ratio of input signal frequency to the sample clock frequency.

There is a magnitude response vesus frequency ratio of integrator in Allen's book, my signal of interest (120K) is less than one tenth of the clock(1.5M), so there is a little gain which is more than 1 a little bit.
In total, there is no gain since the first stage attenuate the signal too much, while the two feedback capacitors can't help increase the gain too much.

Therefore, my question is : can this architecture be used as basic block in a cascade architecture to be a high Q bandpass filter ? I think it attenuate signals too much!
I hope I can get some different opinions . Thanks in advance!

By the way, the possible drawbacks of my opamps : about 0.6 mV DC offset, 9 MHz UBW with 10pF capacitive load. I am not sure how much bad effects can these drawbacks bring ?(DC gain is 85dB)

Title: Re: High Q SC biquad filter architecture question
Post by skythunder on Jul 18th, 2007, 2:50am

Let me explain my question more clearly.
My spec of the SC bandpass filter is : center frequency is 120K, desired passband  width is 10K,
so my bandpass Q factor is 120K/10K=12, which is rather high value,
and then lead to that the pole Q in the s-plane is very close to the j axis,  so the real part of the pole Q is very small.

However,  the damping factor of the bandpass filter is the reciprocal of the twice real part of the Pole Q,which means the damping is serious. Therefore, I got much smaller output than the input signal magnitude.

The above discussion is my own explaination of why my bandpass filter get attenuated output in the passband.

BTW, in chapter 9 of  allen's book , there is an example on designing 5th order butterworth LPF using cascaded biquad. Allen also said that the output of each stage is smaller than input because of the effect of sin(x)/x ,  so I guess my opinion may make some sense.

Anyway, I hope there are some different opinions , especially indicating my errors, I will appreciate that very much !

The Designer's Guide Community Forum » Powered by YaBB 2.2.2!
YaBB © 2000-2008. All Rights Reserved.