EEVblog® Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: hsynylcn on April 28, 2025, 08:26:08 pm

Title: Design of a 4th Order Sallen-Key Bessel Low Pass Filter for Seismometer Data - C
Post by: hsynylcn on April 28, 2025, 08:26:08 pm
I am building a device that converts analog data from a seismometer at school to digital. I need to filter the analog signals before giving them to the ADC. For the low-pass filter I chose a Sallen-Key Bessel type filter because it has low phase delay and group delay. The low pass filter I want to do is to cut above 50 Hz. I made this circuit in FilterPro application. Do you think it is correct? Is there any advice you can give?[attach=1]
Title: Re: Design of a 4th Order Sallen-Key Bessel Low Pass Filter for Seismometer Data - C
Post by: Benta on April 28, 2025, 09:05:29 pm
You're mixing up filter topology and filter response.
Sallen-Key is a certain circuit configuration and has nothing to do with phase respone or group delay.
I see you've chosen a Bessel response which is optimised for linear phase response (= constant group delay), so a very good choice if that's the main design parameter.
But I suggest you choose a MFB filter topology instead, which has superior performance.
See https://www.ti.com/lit/pdf/sloa049 (https://www.ti.com/lit/pdf/sloa049) for more info (especially p.14/15).

Title: Re: Design of a 4th Order Sallen-Key Bessel Low Pass Filter for Seismometer Data - C
Post by: Kleinstein on April 29, 2025, 04:47:09 pm
In most cases one would today use some oversampling from the ADC and do the final low pass filtering in the digital domain. So the analog filter would mainly be there as AA filter for the ADC and the exact charcteristics is not that critical. It also depends on the ADC - a SD ADC may need less filtering than a sampling type ADC. Seismic data usually also don't contain much higher frequency background, though maybe mains hum, that could be in the transition region.