The SSU Noise Model

Adopting the values K_f=4500 and K_w=2.75 resulted in TDEV that lay below the mask. This is shown in figure 8.1. It turns out that the beginning of the MTIE ramp at $\tau=10s$ is the region at which the clock is most in danger of producing noise that exceeds the mask. This region is intermediate in that it is affected by both K_w and K_f to a reasonable degree. Thus either could be lowered to ensure mask compliance. It was decided that lowering K_w was more appropriate since a smaller proportional change would be required. it can be seen in figure 8.1 that the model produces TDEV data of the same form as that measured on a real life add-drop multiplexer (figure 6.4). It can be seen that the white phase noise that is high pass filtered by the PLL is present at short observation intervals as indicated by the $-\frac{1}{2}$ gradient and that approximate flicker phase noise is dominant at long observation intervals corresponding to low frequencies.

 

Figure 8.1:   TDEV produced by SSU model satisfying wander mask

 

Figure 8.2:   MTIE produced by SSU model satisfying wander mask