This paper deals with the problem of estimation of an helicopter micro-Doppler signature using sub-sampling frequencies. One seeks a frequency range such that the events that impair the blade period estimation for a helicopter have minimum probability. Using a stochastic model based on a Markov chain to describe the probability of spike detection, the complete solution is given and the analysis demonstrate that higher frequencies in the sub-sampling range, do not necessarily imply lower loss of information regarding the blade period estimation.