2
$\begingroup$

I am testing the frequency response as a filter as below, what should be the sample rate of the chirp signals, someone told me it should be at least 2 times of the input signals frequency, someone said it should be at least 10 times? Why? enter image description here

Improve this question
$\endgroup$
4
  • $\begingroup$ 1 is the filter continuos time filter or discrete time filter ? 2 Is the block signal sampling representing a sampling circuit ? 3 If the filter is a discrete time signal, what sampling frequency is it designed for ? $\endgroup$
    – AJN
    Commented Sep 29, 2021 at 18:51
  • $\begingroup$ A1: It is a discrete time filter running in my software A2: It is a signal processing module running at specified frequency, I think it could be treated as a sampling circuit. A3: what do you mean? The bandwidth of the filter? $\endgroup$
    – LHX
    Commented Sep 30, 2021 at 2:42
  • $\begingroup$ When designing a discrete-time filter, the coefficients of the filter is a function of the sampling rate at which the filter works. So if you have the discrete-time filter coefficients already designed, then the sampling rate at which it works is already known. The signal sampling block should supply the samples at that same rate. $\endgroup$
    – AJN
    Commented Sep 30, 2021 at 11:06
  • $\begingroup$ Please add more details and context for the question. The answer would depend on whether the chirp is the thing which is already decided / fixed or if the filter is the thing which is already decided / fixed. $\endgroup$
    – AJN
    Commented Sep 30, 2021 at 11:08

1 Answer 1

Reset to default
4
$\begingroup$

2 times is the minimal sampling rate to ensure the signal can be possibly measured. This is called the Nyquist frequency. It shows that if your sampling rate is twice the frequency of the signal, you measure at least 2 different values per period. If you would only measure 1 value, a lower frequent signal or even a constant value could fit your measurement.

10, or even 20 times is what you should use if you desire to filter the signal (or for that fact, control it). This allows to fit your response so smooth, continuous-time filters can be properly applied without to many sampling issues.

Improve this answer
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.