Hello,

I have an active Twin-T notch filter built up with the notch frequency being 10kHz. It works splendidly by itself when exciting the input terminal using a function generator.

The problem arises when I connect an inverting amplifier stage to the output and drive the circuit using the function generator at the amplifier's input. With this new configuration, the notch filter is providing no attenuation whatsoever.

I assume that there is some change in effective impedance of the Twin-T components close to the input such that they are no longer matched with the rest of the components. Although I would expect that the inverting amplifier op-amp output impedance is low and it can drive both the amplifier and the filter.

Reconfiguring the inverting amplifier as a buffer produces desirable effects from the notch filter (although it seems that the notch frequency attenuation is lower than that seen without this stage).

Is a 100 ohm open-loop impedance of the op-amp at 10kHz significant?

 

What is the opamp?

 

I suppose, of course, that your filter "output" is the output terminal of OA1.
More than that, the opamp´s output impedance of 100ohms cannot be the source of the problems because this value is drastically reduced due to negative feedback.

However: Did you check the DC operating point of the opamps - in particular, opamp OA2 and the additional output amplifier.
What are the supply voltage(s) ?

EDIT: Do you realize that the dc voltage of 2.5 volts at the input of OA3 appears at the output multiplied with "11" (if the supply voltage allows 27.5 volts)?

 

@crutschow - It is OPA170. Datasheet here.

@LvW - Correct, the "output" is taken at the output terminal of OA1. To clarify, I suspected that the output impedance of OA3 is affecting the notch filter's behavior, not that of OA1.

All opamps are supplied with 5V and GND. The DC operating voltage of OA1 is 2.5V and for OA2 it is what you would expect from the voltage divider at it's input terminal.

Regarding the amplification of the 2.5V reference, it is not shown, but V1 has a DC offset of 2.5V which keeps the DC output at 2.5V.

Again, the notch filter works as it should when driving the terminal where R1 and C2 join with a function generator and taking the output at OA1. It is only when adding the OA3 amplifier stage that it fails to work.

 

Regarding the amplification of the 2.5V reference, it is not shown, but V1 has a DC offset of 2.5V which keeps the DC output at 2.5V.

That is an important information.

Again, the notch filter works as it should when driving the terminal where R1 and C2 join with a function generator and taking the output at OA1. It is only when adding the OA3 amplifier stage that it fails to work.

Has the used function generator also a dc offset of 2.5 Volts?

 

My simulations do not show the problems you see. But I show only about a -5db notch in the response at about 7kHz both with and without the input opamp. Is that all the attenuation you want?

 

Update: Well... this is embarrassing! Something must have been connected incorrectly and caused the notch stage to be bypassed. I figured I'd try increasing R6's value to 100k in case 10k was too low relative to R1. This led to attenuation at the notch frequency. I then replaced it with 10k and now it works...

@crutschow - That is very strange. I have played around with various components to produce different notch frequencies and I guarantee you that I get much more than -5dB (roughly -40dB).

Thank you for your feedback, it helped to confirm that the amplifier should have no effect whatsoever!

 

Okay, my turn to be embarrassed. I had the bottom of C1 connected to the top of R3 instead of the bottom. Correcting that, the notch is at 10.5kHz and it's depth is about -31.5dB. I got that value for both the AC (linear small-signal) simulation and the Transient (large-signal) simulation.