We could evaluate the performance of attribute control charts by normal approximation. The pn control chart, one of them, depends on sample sizes and defective fraction for its performance, and has an disadvantage that needs much samples to use the normal approximation. The usual pn control charts can not detect early shifts in process because it is taking the sampling intervals of fixed length over all times. We use, in this paper, VSI(variable sampling intervals) techniques to overcome this shortcoming. To analyze performance of pn control charts for FSI(fixed sampling intervals) and VSI(variable sampling interval), we compare ATS(average time to signals) of two charts, and evaluate the performance of each control chart by the sample sizes, process fraction defective and control limits that Ryan and Schwertman had proposed.