OPA4192: Bainter notch filter real design Q

Hello Lurii,

I copied your Bainter 50 Hz band-stop filter into our TINA-TI simulator. You can see the results obtained from the simulation below. The filter topology appears to produce an accurate notch frequency, but poor notch depth - at least with the particular filter parameters slected for this case. It appears to achieve only 13 dB of attenuation at 50 Hz. There may be some filter settings you could set differently such as the -3 dB bandwidth and the stop-band bandwidth, to achieve better notch depth.

There is an older TI filter synthesis program called FilterPro. It uses a couple of different band-stop typologies than Webench Filter Designer's Bainter topology. I synthesized a simple 2nd-order, 50 Hz notch filter using its Multiple-Feedback (MFB) topology. It is simply an MFB band-pass topology that inverts the input signal polarity and then sums it with the original bandwidth input signal. The result is the band-pass response that gets subtracted from the original signal bandwidth and a band-reject response is the result.

You can see the results here:

The notch depth at 50 Hz is on the order of -33 dB. Note that this notch depth  is achieved with exact components and tolerance variations will affect the actual depth.

Regards, Thomas

Precision Amplifiers Applications Engineering

HI Lurii,

Increasing the number of samples in the simulation provides higher resolution. I always use a number like 500, or higher, and 500 was the number I used to get the -13 dB attenuation with the Bainter circuit. Since the greater notch depth is ~- 30 dB with 2000 samples, then the circuit is capable of that performance. But keep in mind that notch depth is very, very narrow and will only be achieved with exact component values. The notch depth will quickly degrade if wide tolerance components are used.

Regards, Thomas
Precision Amplifiers Applications Engineering