| 摘要 |
A method of estimating random jitter from measured samples of a transmitted data signal includes a first step of obtaining a plurality of measurements (e.g., pulse widths) for a plurality of selected signal edges within a transmitted data stream, where the data stream comprises a repeating data pattern having a known bit length and known number of rising edges, and wherein the time difference between adjacent measurements is determined by an event count increment equal to an integer multiple of the known number of rising edges. A time interval error value is then computed for each measured signal edge. Time interval error values are then transformed into corresponding TIE frequency components (via, for example, an FFT) for selected of the measured signal edges, wherein the TIE frequency components are representative of both noise floor as well as multiple distinct frequency peaks. Noise floor is separated from the multiple distinct frequency peaks representative of periodic jitter by replacing each TIE frequency component greater than a predetermined value with a lower predetermined replacement value. The power of the noise floor may be computed to provide an estimate of random jitter variance, from which the standard deviation may be calculated.
|
