+ Post New Thread
Results 1 to 3 of 3
  1. #1
    Full Member level 3
    Points: 2,026, Level: 10

    Join Date
    Jul 2007
    Posts
    170
    Helped
    8 / 8
    Points
    2,026
    Level
    10

    step response of cic filter

    Hi Sir/Madam,

    I had done an evaluation using
    1. CIC filter with decimator factor of 128
    2. CIC filter with decimator factor of 32 and 2 half band filter.

    I am seeing the time needed for the filter to get stable is longer for option 2 above.

    What i did is i run a step response for filter option 1 and filter option 2.

    I attached the foils for your better understanding.

    Can you let me know whether this is make sense? Thanks.

    •   AltAdvertisment

        
       

  2. #2
    Super Moderator
    Points: 255,159, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    44,463
    Helped
    13543 / 13543
    Points
    255,159
    Level
    100

    cic filter step response

    I am seeing the time needed for the filter to get stable is longer for option 2 above.
    It's the group delay of the FIR filter, according to the respective number of taps. Everything is normal operation. A suitable choice of filter characteristic depends on your application, which you didn't tell about.



    •   AltAdvertisment

        
       

  3. #3
    Full Member level 3
    Points: 2,026, Level: 10

    Join Date
    Jul 2007
    Posts
    170
    Helped
    8 / 8
    Points
    2,026
    Level
    10

    is a cic filter stable

    Hi FvM,

    I am using this filter to reduce the frequency sampling rate.

    Do you think i can use different filter structure to reduce the group delay of the FIR filter?

    For example, let's talk about a simple low pass 4 tap FIR filter and use direct form structure to implement this filter.

    Is there any different if i use a serial multiplication and parallel multiplication? that mean for serial multiplication, i need to run 4 multiplication on 4 clock delay before i get a new output. And for parallel multiplication, i run 4 multiplication on each clock and output a new value on each clock.

    Thanks.



--[[ ]]--