vivkr
|
I have a front-end which has rather significant group delay variation when I perform an AC sim. The group delay varies from lower to higher frequencies within the band of interest by about 25% of the bit period.
I would logically expect that a PRBS simulation would show the eye closing in the X-direction by a similar amount. However, such a PRBS-sim shows negligible closure in the X-direction.
I have checked that I really do have a PRBS input, and that I have taken a large enough number of samples (5000 or so). Is my understanding of the relationship between group delay variation and eye closure incorrect ? Note: The model being used is linear.
Regards,
Vivek
|