Problem with a RC Notch Filter in a circuit

[aryajur]

RC Notch Filter Doubt

Hello,
In the circuit attached is a simple RC Notch Filter. On analysis it can be shown that Vo goes to 0 for a particular frequency. But on simulating this in SPICE Vo does not come to be 0.
Also thinking Qualitatively, it is just a parallel connection of a High Pass and a Low Pass Filter and each considered separately never have their Output voltages = 0 except at the extremeties of infinity ( For Low Pass) and 0 ( for high pass) so it should not go to zero. What is wrong with this, why aren't things matching up?

 

[flatulent]

Re: RC Notch Filter Doubt

This is a classic circuit called the 'twin T' and the notch works in real hardware for many decades of time past to present. I would suspect something wrong with your simulation.

 

[VSWR]

Re: RC Notch Filter Doubt

What do you mean with "simulating this in SPICE Vo does not come to be 0"? See attached plot in which the notch is 60 dB deep (510 freq points used in the AC sweep).

 

[Frikkie]

Re: RC Notch Filter Doubt

Agreed.

If youre simulatar has the option make sure you fed it the
real values in for each component. On some of my
simulators I got diffrend results.
That is why I prefer to see before passing a cicuit on.

 

[aryajur]

Re: RC Notch Filter Doubt

Suppose you don't put a load resistor to the notch and then simulate it. The Vo terminal goes to 0 for a frequency of w = 1/(2*R*C) that means that it should show an attenuation of -infinity. But this does not come to be there in spice.

 

[unkarc]

Re: RC Notch Filter Doubt

Suppose you don't put a load resistor to the notch and then simulate it. The Vo terminal goes to 0 for a frequency of w = 1/(2*R*C) that means that it should show an attenuation of -infinity. But this does not come to be there in spice.

Only theoratically goes the Vo terminal to 0 for a frequency of ω=1/2*R*C !!!
In practice the situation is much like (or even worse) as the simulation shows in VSWR's post.
A circuit simulator dos not like such things as "infinity" or "dividing by zero" etc cases because mathematically it may collapse. There are built-in precautions the user of the software may not know of, for instance, unterminated points are grounded automatically by 10Mohm resistors etc etc, simply to avoid situations leading to hang up for the software.

I hope this helps.

unkarc