采样理论（Sampling Theory） 和周期性信号采样 （Coherent sampling）

If a function x(t) contains no frequencies higher than Bhertz, it is completely determined by giving its ordinates at a series of points spaced 1/(2B) seconds apart.

换句话说就是，采样的频率F1和所传输信号的频率F2的关系如果满足，F1>=2F2，那么该传输信号可与等同于采样后生成的信号。

如果F1<F2，这种情况被称为Undersampling，这样会导致测试最后生成的波形的频率比F2小，

如果F1>> 2F2，这种情况被称为Oversampling，这样可以提高测试精度。

所以为了找到一个合适的采样频率和合适的采样次数，引人Coherent sampling（周期信号取样）

UTP is a time interval over which unique information is spread（用来确保采样的唯一性）

假设：

被测device是：8-bit 21MHz ADC Test

采样频率Fs是： 20MHz
被测试信号的频率Fin：5MHz
采样次数Msamples是：N ≧ 2的9次方 =512

带入公式的Ncycles=512/4=128

因为“This integer number should be chosen carefully. We have three possible types of integers, even, odd, and prime. Even is not a good idea since we would hit the same code every M samples, where M can be much less than N. Odd is a better idea since it takes longer to hit the same code.”，

所以我们将Ncycles改成129，然后计算出新的Fin=5.03MHz

http://en.wikipedia/wiki/Coherent_sampling