Currently for all the CPU tests, we use a simple channel model (RC filter) generated using a simple time constant. This model is too simple to really verify the CDR's performance. It works sufficiently for functionality. The current CPU model uses PWL based generation. Maybe there is a way to encode the measured channel values into a PWL filter? Otherwise, we can look into using two time constant filters to create precursor ISI.
Currently for all the CPU tests, we use a simple channel model (RC filter) generated using a simple time constant. This model is too simple to really verify the CDR's performance. It works sufficiently for functionality. The current CPU model uses PWL based generation. Maybe there is a way to encode the measured channel values into a PWL filter? Otherwise, we can look into using two time constant filters to create precursor ISI.