Estimation of data-dependent jitter using single pulse analysis method in high-speed interconnects

Abstract

This paper presents a simple and precise method to estimate data-dependent jitter (DDJ) in high-speed interconnects. Single pulse analysis method, which is based on analyzing the single bit pulse response as the worst case pattern, is proposed for simple and fast estimation of data-dependent jitter. For the accuracy of the estimation, frequency-dependent model of high-speed data channel is applied in this work. The digital data stream on the transfer channel can be broken down into many single bit pulses. If the data channel has simple characteristics, it would be regarded as a quasi-linear system. This means that the entire output waveform can be reconstructed by the responses of each single bit pulse, and it can be easily noticed that a single bit pulse is the worst case pattern of the signal which determines data-dependent jitter. Because of its simplicity, the single pulse analysis method gives quite fast estimation of data-dependent jitter. In this paper, we model simple microstrip lines depending on the frequency and estimate the data-dependent jitter by analyzing the worst case pattern of the data; single pulse analysis method. And we verify the estimated data-dependent jitter with the jitter measurements. Finally, we expand the method to differential signaling and show some applications with single pulse analysis method.