Hi, I need to extract a signal from an input signal contaminated by noise and interferences. The input signal is sampled at 100MHz. The signal I want resides from 0Hz to 500KHz only, and noise and interferers are everywhere. In order to extract the narrow band signal in 100MHz sampling rate, I am thinking of LPF and then downlsampling but I don't know how to decide the number of stages and the filtering corner. Any suggestion on the most effective and efficient approach? cfy30
extract narrow band signal from high sampling rate input
Started by ●July 19, 2009
Reply by ●July 19, 20092009-07-19
On Jul 19, 7:31�am, "cfy30" <cf...@yahoo.com> wrote:> Hi, > > I need to extract a signal from an input signal contaminated by noise and > interferences. > > The input signal is sampled at 100MHz. The signal I want resides from 0Hz > to 500KHz only, and noise and interferers are everywhere. In order to > extract the narrow band signal in 100MHz sampling rate, I am thinking of > LPF and then downlsampling but I don't know how to decide the number of > stages and the filtering corner. Any suggestion on the most effective and > efficient approach? > > cfy30For such high decimation ratios it is common to start with a CIC filter. The CIC filter is followed by an FIR filter which can filter the signal to the final oversampling rate you need. You can find many references and papers out there to understand the structure of the CIC filter. Moti
Reply by ●July 19, 20092009-07-19
Moti Litochevski wrote:> On Jul 19, 7:31 am, "cfy30" <cf...@yahoo.com> wrote: >> Hi, >> >> I need to extract a signal from an input signal contaminated by noise and >> interferences. >> >> The input signal is sampled at 100MHz. The signal I want resides from 0Hz >> to 500KHz only, and noise and interferers are everywhere. In order to >> extract the narrow band signal in 100MHz sampling rate, I am thinking of >> LPF and then downlsampling but I don't know how to decide the number of >> stages and the filtering corner. Any suggestion on the most effective and >> efficient approach? >> >> cfy30 > > For such high decimation ratios it is common to start with a CIC > filter. The CIC filter is followed by an FIR filter which can filter > the signal to the final oversampling rate you need. You can find many > references and papers out there to understand the structure of the CIC > filter.See my sig. You don't directly choose the number of stages for your filter, you determine its necessary specification and then use as many stages as needed to meet it. Using a CIC filter as part of low-pass chain will reduce the computational effort, as will multi-rate decimation. The specs themselves grow out of the problem statement. As long as they are met, how they are met is a secondary consideration. Cookbook (handbook) approaches save time and thought. They are useful when repeating what others have done. When they are the only tool one can bring to bear, no innovation is possible. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
