Hi puestadeluna,
Welcome to Talk Stats.
In #1 if n is sample size, when you take different samples, n may be different.
For #2, you could get one sample with sd very close to ppopulation sd just by chance. In general, sample sd will approximate the population sd as sample size increases.
I don't quite understand #3, if you have more than one population then you need to specify where the samples come from.