For a linear wideband amplifier used in digital serdes applications, I'm curious to know if there is a way to predict the NRZ signal amplitude being fed into the wideband amplifier that will cause nonlinearity and how a VNA can help obtain that information. Imagine an asserted wideband amplifier connected to a pattern generator and a scope and start with very small amplitude swing and captured a normalized eye, where it is safely assumed that the wideband amplifier is completely linear. Now, slowly increasing the amplitude of the pattern generator, at some point the normalized eye will start to change. It would be very beneficial, I think, if VNAs could help characterize this type of performance.
Any ideas in Dr. Joel's book?