ISI CHANNEL COMPARISON RECORD
Purpose: compare channels or settings while holding the other test conditions fixed.
Protocol/mode/rate, pattern and pattern length:
Transmitter and receiver configuration; correction/equalization:
Channel/coupon/connector identity, reference planes and fixtures:
Observation point, bandwidth, clock recovery and acquisition depth:
Baseline result; candidate result; return-to-baseline result:
Illustrative sampled model: y[n]=0.80x[n]+0.25x[n-1]+0.10x[n-2], with x=+1/-1. Current +1
gives 1.15 after +1,+1 and 0.45 after -1,-1; nearest levels are +/-0.45. This is a toy
model, not a measured eye.
Bits counted N, errors, duration, synchronization and pre/post correction:
Illustrative zero-error bound: independent stationary errors, one-sided 95% upper bound
approximately 2.996/N. N=3e12 gives about 1e-12; at 1e9 counted bits/s this takes 3,000 s
(50 min). Correlation or bursts can invalidate the model.
Operating corners, acceptance criterion, next controlled experiment:

Yenra | Updated September 10, 2026
Article and explanation: https://yenra.com/electr/isi.html

Source references (use exact equipment/material documents for actual work):
https://www.tek.com/en/documents/application-note/bertscope-bit-error-rate-testers-jitter-map-under-hood
https://www.tek.com/en/documents/whitepaper/advanced-serdes-debug-bert