Anti-Aliasing filters must be pretty important, since most data acquisition systems have them. But, what are they? How do they work? It turns out that they are pretty simple and easy to make if you need to.
Very well presented and explained. Although the filter commits to the requirements, it's not complying to basic rules of design. When designing various parts you should give the most expensive component a fixed commercially available value, then design the rest to fit as much as possible under more commercially available components. In this case, you should have picked the capacitor first, then calculate the resistor. It's easier to find many resistors than capacitor values on real life design.
You don't. The fewest possible number of cycles is two, but that's the ideal case. In practice, you need more than two. All the data acquisition systems I've used sample at more than 2x the maximum frequency. I think the Agilent (used to be HP) system sampled at 2.56x.
If fs=10kHz and the input signal has fmax=5kHz, then the cutoff frec. cannot be less than 5kHz, because then you 'filter' the 5kHz signal. An anti-aliasing filter tries to delete the high frec. signals in order to allow the computer read the signal we want. It is usually used in AD Converters. Anyway, fs is usually 10-20 times higher than the signal frec. and also 10 times higher than the cutoff frec. Sorry for the English... ;)
