Active notch filter behaves like a low-pass filter

Question

I'm trying to build an active notch filter centered at 50 Hz. I'm trying to wire it using a breadboard but I'm facing some problems with the output signal.

Before wiring the circuit, I tried to simulate it with LTspice and it worked as you would expect (great attenuation around the central frequency, almost nothing outside the attenuation band). But when I try to actually build it, I register strange behaviour of the output signal.

Not only the input signal (a 50 Hz sinusoide) gets a weak attenuation, but the higher the frequency of the input signal, the stronger the attenuation becomes. If I didn't know that it was a notch filter, I would have said it was the frequency response of a low-pass filter. Even if the central frequency was wrong, I would expect at least the same behaviour with a thin range of filtered frequencies (classical v-shape frequency response).

It's almost like if there is a lowpass filter inside this notch layout which is predominant and forces the low-pass behaviour.

Here is the circuit I simulated:

And this is the result:

Some details about my circuit on the breadboard:

1. The breadboard is a hobbyist breadboard
2. The op-amp I'm using is a MCP6002 supplied with Vdd = 5V - Vss =0 V (the input signal is always between these two values)
3. The input signal is supplied by the function generator of a PicoScope (same of the simulation)
4. The output signal is measured by a PicoScope
5. The used capacitors are electrolytic capacitors
6. The output signal is not distorted, it's just slightly attenuated at 50 Hz
7. The higher the input sine frequency, the stronger the attenuation (low-pass behaviour)
8. I tried a different layout for the notch filter, but same results: simulation OK (less aggressive attenuation) but low-pass behaviour on the breadboard

Honestly, it's the first time I experienced something like this. Do you have any suggestions?

---

Ok, I think the chosen layout is not best option.

I tried to simulate another version of notch filter. It is one of the most common online:

This time it looks like the input values are inside the desired range.

I'll try to wire it up and see what happens.

---

I also noticed that the elecotrolytic capacitors are not ok for this second circuit because the voltage on C4 and C3 is not always positive or negative. Looks like I need to buy ceramic capacitors.

---

Ok, I bought some ceramic capacitors and i tried to wire up the circuit above using a single supply configuration (0-5V) with a MCP6002. I'm experiencing the same issues I saw in the other configuration: increasing the frequency of the input sinusoid (which has an offset of 1.5V and an amplitude of 500mV), the output becomes weaker and weaker. The input signal is generated with the function generator of my PicoScope.

On the LTspice simulation the frequency response is perfect, with a V-shape centered on 50Hz. In the real circuit the frequency response shows a low-pass behaviour and at this point I can't figure out what's the problem...

Maybe is the single supply, but on the LTspice simulation there isn't one single negative value on the various nodes on the circuit (with respect to ground).

I debugged multiple times the circuit but it seems ok.

Have you ever experienced something like this?

Answer 1

The notch filter circuit is supposed to have a positive and negative supply. With your single positive supply, you wrongly biased the signal input positive 1V but instead R3 should be biased positive 1.25V so that the opamp works properly. The notch filter reduces the level of 50Hz a lot.