Tunable notch Filters for audio frequency

Hi!

Can anyone help me on how to limit the attenuation in the rejection band of the notch filter?

Thanks

Tunable Notch LTC1067.asc

Hi Philip,

thanks for your kindly answer. I'm testing this and up to know this is what I was needing.

BTW. Do you have any suggestion for a VCO to feed the LTC1067 clock between 30kHz and 50kHz.

Thanks again, Julian.

VCO LTC6990.asc

VCO LTC6990.asc

Thanks, Philip. You really helped a lot.

The problem now is the dual supply. Do you see it possible to use 2x LTC3388-3 for achieving +/-5V from a single 9V supply and then use two LDO for powering the system with +/-4.5V for audio application?

The LTC1067 notch filter can be operated from a single 9V battery and then ac couple the output. Use a 5V LDO for the the LTC6990 VCO circuit.

A switching regulator is not recommended for a low noise signal circuit.

Yes, that's why my question. But how to redesign the circuit and also I should have a single supply for the LTC6240

Tunable Notch LTC1067 Single 9V.asc

Hi Philip, thanks again! This really helped.

BTW: I see you are creating the notch subtracting a BPF from the original signal and you are adjusting the gain of the original signal in the rest.  Isn't this possible to adjust the gain of the BPF instead? I'm asking this because with the actual design, I'm changing the gain outside the notch zone.

Thanks again, Julian

Yes, adjusting  the BPF gain will adjust the notch attenuation at fnotch

Note: The notch summing resistors must be equal if the BPF gain is adjusted. 

Hi Philip, at the end I have assembled the board with the filter but something isn't working.

The Oscillator is working and is being regulated in the required frequency range. The output is a square of 0-5V amplitude and the V+ of the LTC1067 is +9V with respect to V-.

Unfortunately, the notch isn't working and I found the output of the bandpass filter is null. 

The simulation has shown that INV_A should be DC AGND and this is fine but BPA has DC at 300mV below AGND instead of signal centered in AGND. This is for any oscillator/signal frequency, for example, oscillator at 51kHz.

Do you have any suggestions?

Thanks in advance, Julian

Hi Julian.

If the LTC1067 is operating with a single 9V supply then AC couple the 0-5V clock signal.

The clock input levels for 9V supply are Hi >=7V and Lo<=2.5 biased at V+/2.

Good point! I´m sorry for my dumb question. Now it is well working with everything at 5V across a regulator.

I will try lather to amplify the oscillator with a transistor to make this work at 9V

Thanks