High performance turnable notch filter

[Wenye]

gic notch filter

Does anyone has idea for design a turnable notch filter in audio frequency?

The goal is:

1. the notch filter can be precision turn by control voltage, or can be turn digitally.

2. center frequency can be turn from 20Hz to 100KHz;

3. stop band att. will be equial to or over 60dB, stop band bandwidth will be less than 5% of center frequency.

4. addtional THD+N will be less than -120dB.

Is there any ideas or advice?

Or, where can I find the reference designs or papers?

Thank you in advance!
Wenye

 

[sinatra]

You can probably do this using an Antoniou GIC.
Just making a series RLC network to ground.
The inductor is simulated by 2 operational amplifiers (GIC).
You can adjust either the inductor (adjust internal resistor of the GIC) or adjust the resistor from the RLC circuit, to get to the desired frequency.
S.

 

[moda]

Look at:

**broken link removed**


Regards

Moda

 

[flatulent]

Your -120 dB distortion and noise specification is very severe and may not be met with any form of active filter using operational amplifiers. It will be very difficult to get this level of distortion with any active devices in the signal path.

You will probably have to use some sort of bridge circuit with transformers in and out.

 

[Wenye]

You can probably do this using an Antoniou GIC.
Just making a series RLC network to ground.
The inductor is simulated by 2 operational amplifiers (GIC).
You can adjust either the inductor (adjust internal resistor of the GIC) or adjust the resistor from the RLC circuit, to get to the desired frequency.
S.

Hi Sinatra,

Considering FET as a VCR for turn the center frequency. I need to make some test for the THD+N performance....

Thank you
Wenye

Added after 12 minutes:

Look at:

h**p://focus.ti.com/lit/ds/slas226b/slas226b.pdf


Regards

Moda

Hi Moda,

My project is audio analyzer, I need outstanding analog singal conditing circuit performance for ADC to reach the goal of system performance.

TAS3001 is an wonderful signal processing ic for digital audio purpose, it doesn't suite for audio analyzer...

Thank you

Added after 14 minutes:

Your -120 dB distortion and noise specification is very severe and may not be met with any form of active filter using operational amplifiers. It will be very difficult to get this level of distortion with any active devices in the signal path.

You will probably have to use some sort of bridge circuit with transformers in and out.

Hi Flatulent,

My project is an high performance audio analyzer.

I found the opamp ad797 today from analog device Inc. It looks that it can offer -120dB THD+N from 20-20KHz.

I think that if I design a notch filter with around 40dB or more gain for harmonic, attenua the Fs in over 60dB and compensate it digitally in display, then reach 120dB DR is possible even with a 12bit ADC.

Is it right?

Wenye

 

[sinatra]

Hi Wenie.
I agree with flatulent, this 120dB will be difficult to get using active devices.
It maybe possible but it will not be easy.
I have another idea that could help to solve your problem.
I vote again for the RLC series network to ground. But this time I propose something different for the synthezis. Instead of using Antoniou GIC you couls use switch capacitor technique and keeping unit gain ( no need for opamps or active devices).

You should build an convensional air core inductor to get rid of any kind of distortion.

And I believe, for the capacitor (or the resistor), a good solution would be to use a switched capacitor. A good quality capacitor, polystirene for example.

And maybe simple inverter switchs like 4066 would do the job. Or if something better is needed maybe mosfets like BS170, BS250, could do it....

We would have to search for a good swtched capacitor (unit gain) configuration to realize this transfer function. (notch)........

The opamp you found could be used as buffer......

And now a question....What would be the configuration to implement?????

S.

 

[Wenye]

Hi Sinatra,

Could you explain about what is "Antoniou GIC" ? Or do you have any papers about it?

Another question, if the switched capacitor is emploied, how do you think about the clock feed-through problem?


Thank you
Wenie