Wikipedia gives the following circuit as a negative capacitance circuit:

enter image description here

I was curious if I would be able to create a low-pass resistor-negative-capacitance filter, analogous to an RC low-pass filter, using the above circuit, which would have a positive phase shift above the cutoff frequency.

This is what I tried:


simulate this circuit – Schematic created using CircuitLab

Running a frequency response simulation showed promising results:

enter image description here

enter image description here

Obviously, at a certain point, the op-amp's performance begins to flag, so I am not surprised that the phase shift continues to rise above 90° when the frequency is above 1 MHz.

However, I was very disappointed to see the time-domain response. My low-pass filter unfortunately is also a relaxation oscillator. Typically, an op-amp relaxation oscillator has the capacitor connected between the inverting input and ground, but obviously this topology works as a relaxation oscillator as well.

enter image description here

My questions are:

  1. Is there a (not too complicated) way to suppress the relaxation oscillations, or is this negative capacitance circuit useless for making an R/"C" low-pass filter?

  2. Is there an alternative way to make a low pass filter with a positive (90°) phase shift in the region where there is 20 dB / decade attenuation?

This question is related, but unfortunately, the OP only did a frequency response, and didn't understand why the circuit did not work in practice. A scope or a time domain simulation would have shown the relaxation oscillations.

  • \$\begingroup\$ 1) the arrangement itself is a relaxation oscillator having an oscillation frequency of \$f\approx1/(2\ln(3) \ RC)\$ where R is 100k (there are 3) and C is 1n. Without any input the circuit should oscillate. Maybe you can start with different resistance values (just to disturb the R ratio). 2) Did you consider negative resistance instead? \$\endgroup\$ Commented Jan 23 at 16:28

1 Answer 1


Your circuit is not a lowpass filter. Why not? Because you must not trust a small-signal simulation in the frequency domain only!

In the case under discussion the rising phase and - at the same time - the falling amplitude (gain magnitude) are a clear indication of instability.

Note that such an ac analysis (small-signal analysis in the frequency domain) assumes:

1.) That the power supplies are ideal (constant) and available without any switching transients.

2.) That there are absolutely no external influences (noise an/or other fluctuations).

3.) Mechanical analogy: Under such ideal conditions even a small ball could be positioned on the top of a larger ball.

4.) As a consequence, each ac analysis gives correct and realistice results only for stable systems

5.) Even a visual inspection of your circuit reveals that for frequencies above a certain frequency limit the positive feedback will govern over the negative feedback.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.