Signal Processing Q

1
Signal Processing QRs --- Edge equalization
of NRZ and duobinary signals

• Lei Zhang, PhD Candidate
• Electronics Department
• Carleton University

2
Why Edge Equalization

• High speed data transmission needs zero ISI
• Conventional SSF channel equalization can hardly
  achieve zero ISI at high data rate
• A recent research1 introduces a new approach to
  reduce ISI by changing the edge of a differential
  signal, which is so called as pulse width
  modulation (PWM) in the paper. However, we
  consider this as edge equalization since it
  changes the edge of a differential signal, not
  the duty cycle of the pulses

3
A Comparison of SSF and Edge Equalization---
channel pulse response
4
A Comparison of SSF and Edge Equalization(cont.)
5
A Comparison of SSF and Edge Equalization(cont.)
6
A Comparison of SSF and Edge Equalization(cont.)
7
Impulse Response of Duobinary Filter
Nyquist interval 1.5UI
Filter
8
Duobinary Pulse Response with Edge Equalization
9
Duobinary Pulse Response with Edge Equalization
10
Conclusion

• The above edge equalization approach achieves
  nearly zero ISI in both NRZ and duobinary signal
  transmission, with the price of dramatically
  increased amplitude attenuation
• Apply the above edge equalization approach to a
  duobinary signal can eliminate the need for an
  FSF. However, this edge equalization approach
  cannot replace duobinary coding.
• Such an edge equalization approach puts more
  loads on the driver design.

11
Reference

• 1 J.H.R. Schrader et al., Wireline
  Equalization using Pulse-Width Modulation, in
  IEEE Custom Integrated Circuits Conference
  (CICC), pp. 591-598, 2006.