State Variable Notch Filter Depth

[sameerdhiman]

Hi,

Can we practically achieve -60dB or more deep notch with state variable filter topology ?

Is the state variable notch depth also depends up on close matching of components like in Twin-T ?

 

[BradtheRad]

I have the Active Filter Cookbook by Don Lancaster. It discusses state variable, biquad, etc.

It has graphs showing performance curves. Most give curves up to 40 db only.

Looks like 60 db will require a more elaborate circuit.

The book recommends close component matching in order to achieve greater Q and a more pronounced center peak / notch.

---------- Post added at 10:58 ---------- Previous post was at 10:39 ----------

It talks about a stronger type of notch filter called the Cauer or elliptical filter. It's derived from a state variable type.

The response curves show increasing dropoff rates for second-order, third-order, fourth order.

The fourth order response curve does go down to 60 db at the bottom of the notch.

 

[sameerdhiman]

Active Filter Cookbook : Do you have soft copy ?

Well, currently I am able to achieve -45 dB deep notch without matched components (Resistors=5% Carbon and Caps=20% Polyester). The intended purpose of this filter is to extend sound card based spectrum analyzer's range.

 

[zorro]

Hi Sameer,

you should have two adjustments (i.e. two preset resistors) in order to adjust perfectly the zeros.
You can see that there are 2 variables to adjust: alternatively, frequency and depth of the notch, or real and imaginary parts of the zero pair.
If well adjusted and if the circuit components are stable, you can get a very deep notch. But its permanence is limited mainly by component stability.
An option is to place more than a filter in cascade: two filters rejecting 30 dB at the same frequency give 60 dB; if the notch frequencies are something different, you should have more than 30 dB individually in orter to have 60 dB total.
regards

Z

 

[FvM]

Is the state variable notch depth also depends up on close matching of components like in Twin-T ?

With a suitable circuit, the notch depth does not depend on the matching of components. It's mainly limited by the finite OP GBW.

 

[sameerdhiman]

Thank to everybody.

In Sallen-Key's Twin-T topology I used NE5534 and manually matched capacitors/resistors within 1% and utilized multi-turn pots for tunning. I am able to get -64 dB rejection. NE5534 is configured as voltage follower to give pass band unity gain.

After receiving parts I'll put separate OP for feedback and will see how much narrow and deep notch I can achieve.