enter image description here

Many data-acquisiton boards are recommended to use anti-aliasing filters between the analog inputs and the ADC. Here what I understand from anti-aliasing: If one has a DC like signal and the freq. of interests of the signal is not more than 10Hz one can use an RC low pass filter with a 20Hz cutoff freq.

But what if the signal is a pulse? I want to make an RC aliasing filter for pulse frequencies changing between minimum 3Hz up to 500Hz. My sampling freq is 4kHz. In this case, one might use an RC filter with a cutoff freq of 1kHz but this could effect the rising and falling edge shapes significantly. I still want the pulse shape as sharp good as possible.

Is there a rule of thumb for an RC antialasing filter if the input is a pulse in daq systems?


1 Answer 1


Do you know about the Fourier transformation ? If not, look here.

The Fourier transformation relates signals described in the time domain (for example: a pulse) to their frequency spectrum, i.e. the frequencies they contain.

If your pulse signal has frequencies components above a certain frequency which you filter out (with your anti-aliasing filter) then you change the signal. Indeed steep edges in the signal result in high frequency components.

An ideal pulse or squarewave signal with rise and fall times of almost zero will have frequency components up to infinitely high frequencies !

You will have to figure out what you need, what are the rise and fall times of your pulse. If these are for example 1 ms (0.001 second) then the bandwith of this signal is in the order of 1 kHz (1 / 1ms) so a 4 kHz filter will not affect this signal much.

If the rise and fall times of your pulse are 100 ns you need a 10 MHz bandwidth.

  • \$\begingroup\$ amazing answer. i didnt know we need to set it due to rising edge time for pulses. \$\endgroup\$
    – floppy380
    Nov 24, 2016 at 10:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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