A robust and accurate algorithm for time measurements of periodic signals based on correlation techniques

Algorithms incorporated in the oscilloscopes allow automatic measurement of periodic waveforms. These algorithms are often based on IEC Standards. They are not only useful for built-in oscilloscope measurements, but also as stand-alone algorithms implemented, for example, on a computer. The conventional algorithms used for time measurement of periodic signals (e.g., period, positive and negative duty cycles and pulse widths) are based on crossings of a reference level. The crossings method for time measurement fails when there are spurious crossings around the reference level. We have developed a robust and accurate algorithm for time measurement of period, duty cycle, and pulse width, based on correlation techniques. The algorithm for determining the period is generic to any periodic waveform with at least one period. The algorithm for duty cycles and pulse widths is only valid for pulse-shaped waveforms. Using example waveforms, our algorithm is shown to produce more robust and accurate time measurements than the built-in algorithm used in a commercial oscilloscope. With modern-day computing power, our algorithm takes between 10 and 20 ms to perform measurements for a waveform of 1000 8-bit data-points on a Pentium III laptop.