MODELING AND BUDGETING OF TIMING JITTER AND NOISE - Advanced Signal Integrity for High-Speed Digital Design - page 578

100%

90%

Margin
ISI
Crosstalk
PS Noise
V ref Noise
Eq Resolution
Offset + Sensitivity
Offset + Sensitivity
Eq Resolution
V ref Noise
PS Noise
Crosstalk
ISI
Margin

80%

70%

60%

50%

40%

30%

20%

10%

0%

Single-ended Differential

Figure 13-20 Noise immunity comparison for Example 13-6.

ratio of 3.5:1, as opposed to 2.4:1 for the single-ended system. Figure 13-20 shows
that the singled-ended system has a minimum required swing that is 41.7% of
the actual swing, while the differential system requirements a minimum swing
that is only 28% of the actual. Viewed in this light, the differential system offers
improved noise performance over single-ended signaling [Dally and Poulton,
1998].

13.5 PEAK DISTORTION ANALYSIS METHODS

Since our primary focus in this topic is signal integrity, we now describe a method
for assessing the peak jitter and noise caused by the system interconnect. Recall
that the jitter and noise caused by crosstalk and ISI are deterministic, so that we
can budget them using worst-case values. The method that we use to obtain the
worst-case values is called peak distortion analysis (PDA) [Casper et al., 2002].

13.5.1 Superposition and the Pulse Response
We start by elaborating on the linear time invariance (LTI) concept that we intro-
duced in Chapter 12. In particular, we make use of the superposition property,
which says that if the system input can be expressed as the sum of multiple input
components,

x(t) = x i (t)

i