LeCroy offers spread spectrum clock tracking (SSCTrack) which extracts and displays the instantaneous frequency variation of the clock as a function of time. This is an exclusive feature of LeCroy Serial Data Analyzer II (SDA II) where it is available as a math function. Figure 2 (page 2) shows and example of SSCTrack being used to analyze the spread spectrum clocking of a SATA lone bit compliance test pattern. Zoom trace, F5, shows an expanded view of the test pattern.

SSCTrack is based on a uniform sampling of the frequency of the waveform shown in Figure 1. The frequency at each measurement position is determined using a number of points appropriate to achieve the bandwidth specified by the user. Spacing between the frequency measurement positions is set using the 'Decimate by' control on the menu. The purpose of setting “Decimate By” is to minimize the amount of data processed and also to reduce process noise. Large decimation factors may reduce the ability to track the frequency, keep the decimation low enough to avoid having a problem.

The resulting series of measurements are low pass filtered and plotted as a track function which is time synchronous with the source waveform.

Figure 1:

How SSCTrack measures frequency

Figure 2:

Spread Spectrum Clock Analysis for a SATA lone bit pattern signal using SSCTrack

The SSCTrack tab shown at the bottom of Figure 2 in the dialog box shows the setup for the SSCTrack function. The user enters the mean frequency of the serial pattern. This can be obtained, for a data signal, as one half the Bit Rate parameter (P1 in the figure). The ‘Bandwidth’ entry sets the width of the measurement gate. However, there is a limit on how many samples the SSCTrack can use to determine local frequency, which corresponds to a minimum bandwidth at a given sample rate; so there is a chance that the bandwidth of the FM demodulation will be higher than requested, if a very low bandwidth is requested. Therefore, the SSCTrack component includes a low pass filter to enforce a particular bandwidth limit. The ‘LPF Passband width’ and ‘LPF Transition width’ setup the characteristics of the low pass filter applied to the track.

The ‘Scale” entry sets the vertical scale of the track function. As can be seen in Figure 2 the SATA signal has a peak to peak frequency deviation of 7.4 MHz and a frequency of 32 kHz. Math trace F2 shows the track of the signals clock frequency versus time. The FFT trace, F4, shows the frequency spectrum of the signal about the nominal fundamental frequency (half the bit rate) of 1.49 GHz. It shows the same clock rate variation in the frequency domain.

The SSCTrack function is not limited to any one serial standard. Figure 3 shows the same analysis applied to a PCIE G2 compliance test pattern. The SSCTrack function (trace F1) showing a peak to peak deviation of just below 25 MHz (P3) at a frequency of 31.7 kHz (P2).

Figure 3:

SSCTrack showing the frequency variation in a 5 Gbps PCIE G2 signal.

Because the alternate polarity ‘bursts’ of this PCIE test pattern can cancel each other out in the frequency measurement process, we apply the absolute math operator to the signal to make it unipolar prior to inputting it into the SSCTrack function. This is done using the dual math feature of the LeCroy oscilloscope. A fortunate side effect of making the signal unipolar is that the "Mean Frequency" we need to specify in the dialog is the mean bit rate, which can be determined directly by the BitRate parameter (P1) in the figure. This can be seen on the Math setup dialog box at the bottom of the screen image.

As you can see the SSCTrack function is a powerful tool for analyzing the frequency characteristics of spread spectrum clocking.