0
\$\begingroup\$

I need a high pass filter for my measuring setup and I'am a total newbie to analog filters. I have weak hf pulses (0.5-1.2 MHz, 10 mV) superimposed on lf sine wave (8-40 kHz, up to 5 V), I'm interested in the pulses only and need the best resolution possible.

My first attempt is a 2nd order active Butterworth filter, cutoff 100 kHz, using NE5534 opamp, circuit below. I am not in the stage of using the signals intended or checking the stop band attenuation, just trying. 2nd order hp filter

  1. When using sweep as input signal (blue), I am measuring this on output (green). It looks like a resonance peak at approx. 90 kHz. When I ommit the filter parts (input on +), the amplification is the same for all frequencies; when I ommit the 10k resistors on the right side, I get nice filter with no peak. What causes this behaviour? enter image description here

  2. I am aware that the opamp is not suitable for MHz signals, but I am still not sure what bandwidth and V/us rate is sufficient for my application.

\$\endgroup\$
4
  • \$\begingroup\$ What are the supply voltages? Have you simulated the circuit? \$\endgroup\$
    – Tim Williams
    Jun 19 at 12:27
  • \$\begingroup\$ From ±5 V to ±15 V from my lab power supply, the only difference is the peak is cut off at the respective voltage. I did the simulation now and (if I used the ltspice correctly), my measurements are in agreement with the simulation with the peak appearing iff A=2 (resistors on the right have same value). So I cannot just calculate the filter parts based on the cutoff and attenuation and choose opamp and amplification independently? Is simulation always necessary? \$\endgroup\$
    – Nikd0
    Jun 19 at 12:40
  • \$\begingroup\$ There should be formulas or calculators available that are aware of amp GBW. The derivation is complex enough that, failing such a reference, simply simulating it and adjusting values to get the desired response, is common enough -- and easy enough to do for a 2nd order filter like this. You may also want to make adjustments around common component values (1, 2.2, 4.7, etc.) which isn't going to be analytically solvable, or not in much of a way that will help. \$\endgroup\$
    – Tim Williams
    Jun 19 at 18:27
  • \$\begingroup\$ You may find these references handy: ti.com.cn/cn/lit/an/sbaa236/sbaa236.pdf?ts=1687197854303 habr.com/en/articles/521478 \$\endgroup\$
    – Tim Williams
    Jun 19 at 18:29

0

Reset to default

Your Answer

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