0
\$\begingroup\$

if i'm using Sampling frequency less than Nyquist Sampling frequency, Aliasing will occur which will scramble the whole signal, so one solution is to use anti-aliasing filter

How can i calculate it's Bandwidth, also is it a band-pass filter ?

\$\endgroup\$
1
\$\begingroup\$

Aliasing occurs if there are signals present that are greater than half the frequency of the sampling frequency. Lower picture shows aliasing. Upper picture is OK: -

enter image description here

Using a filter to eradicate those frequencies is impossible because you would need a brick wall filter so you have to decide on how much attenuation you have to provide in the filter to give an acceptably low level of aliasing.

It's a low pass filter that reduces aliasing.

So decide on the attenuation required for any aliased signals and design a filter based on (typically) multiple sections of Sallen Key filters together forming (typically) a butterworth (maximally flat) low pass filter.

\$\endgroup\$
4
  • \$\begingroup\$ So if using fs lower than double the fm frequency it's impossible to recover the signal, the only way is to reduce this effect? \$\endgroup\$
    – JenuRudan
    May 23 '16 at 12:44
  • 1
    \$\begingroup\$ Correct @JenuRudan \$\endgroup\$
    – Andy aka
    May 23 '16 at 14:14
  • 1
    \$\begingroup\$ ... unless you are intentionally doing bandpass sampling. fs must be at least double the bandwidth you are trying to read. It's just that because overwhelmingly most people want to sample from DC, that we usually say that fs has to be >twice the highest input frequency. You can use a bandpass filter, around a narrowband signal, and sample at much less than its centre frequency, as long as you sample at >2x its bandwidth. It's called 'High IF' sampling. \$\endgroup\$
    – Neil_UK
    May 23 '16 at 17:12
  • \$\begingroup\$ @Neil_UK good point - in effect you can regard "aliasing artefacts" as being pushed down below the lower bandwidth limit of your desired signal i.e. they can be rejected without loss of desired signal integrity. \$\endgroup\$
    – Andy aka
    May 23 '16 at 17:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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