A 6-Gb/s MIMO Crosstalk Cancellation Scheme for High-Speed I/Os

T. Oh and R. Harjani, "A 6-Gb/s MIMO Crosstalk Cancellation Scheme for High-Speed I/Os," in IEEE Journal of Solid-State Circuits, vol. 46, no. 8, pp. 1843-1856, Aug. 2011

"Transmitted signals with sharper transitions are more easily coupled because of the high-pass filter characteristic of the crosstalk channel"

"the FEXT pulse response is approximately the negative derivative of the channel pulse response".  This is the principle that allows the serdes.ChannelLoss system object to fabricate the crosstalk response.

"In our MIMO algorithm, the crosstalk signal energy is reutilized to reduce the jitter and increase the vertical eye opening while mitigating the ISI tails.

"The FEXT and its cancellation signal or primary and MIMO signal will be skewed even with an optimal value and degrades some of the benefits of both crosstalk cancellation and MIMO improvement discussed here."

RJA--> just like chord signaling, skew is a major detriment

Improved Jitter Distribution Tail-Fitting Algorithm for Decomposition of Random and Deterministic Jitter

 G. Soliman, "Improved Jitter Distribution Tail-Fitting Algorithm for Decomposition of Random and Deterministic Jitter," in IEEE Transactions on Electromagnetic Compatibility, vol. 62, no. 5, pp. 1852-1858, Oct. 2020

• In the abstract, it states that current approaches assume the tails of the PDF are Gaussian but this paper shows that this is inaccurate.  Nice!  I came to the same conclusion this summer when someone was trying to validate the SerDes Toolbox bathtubs with Q-scale rule of thumb.
• " not purely Gaussian but consists of a Gaussian multiplied by a term that is inversely proportional to TJ"

• "Observing (8) and (9), we note that the free factors and the mean terms that appear are directly related to the underlying DJ distribution unlike the parameters in the currently adopted Gaussian approximation of the tail."
• RJA--> The above equations show that the tails of the TJ are not purely Gaussian but have a factor that is proportional to the DJ.
• "Since the linear Q scale fitting technique is already implemented in measurement equipment, it might be attractive to use the jitter parameters estimated from this method as the starting point of a nonlinear fitting routine to (22) and (23) for more accurate jitter parameter estimates."
• I only read through section II and the conclusion.  I think that this paper is great and is something I should implement in the future.

3D Smith Chart

 A. A. Muller et al., "The 3D Smith Chart: From Theory to Experimental Reality," in IEEE Microwave Magazine, vol. 21, no. 11, pp. 22-35, Nov. 2020

• They extend the 2D Smith chart to 3D so that both negative and positive resistance circuits are able to be analyzed.
• Not sure how I could apply negative resistances to my work so I'm not sure if this is applicable.  Negative resistance can be created using active circuits.

Crosstalk Literature search

•  T. Oh and R. Harjani, "A 6-Gb/s MIMO Crosstalk Cancellation Scheme for High-Speed I/Os," in IEEE Journal of Solid-State Circuits, vol. 46, no. 8, pp. 1843-1856, Aug. 2011
• Classic circuit paper
•  J. Wang, C. Xu, S. Zhong, S. Bai, J. Lee and D. Kim, "Differential Via Designs for Crosstalk Reduction in High-Speed PCBs," 2020 IEEE International Symposium on Electromagnetic Compatibility & Signal/Power Integrity (EMCSI), Reno, NV, USA, 2020, pp. 145-149
• This paper is concerned with arranging the vias in such a way to reduce crosstalk without sacrificing the signal:ground ratio.
• H. Lin, B. Tseng and J. Yen, "Novel design with minimized far-end crosstalk on single-ended striplines," 2017 IEEE International Symposium on Electromagnetic Compatibility & Signal/Power Integrity (EMCSI), Washington, DC, 2017, pp. 261-266
• Reduce single ended crosstalk by placing it between two diff pairs?
• R. Enriquez, K. Xiao, B. Lee and M. Tlaxcalteco, "Differential symmetry principle for differential crosstalk cancellation," 2013 IEEE International Symposium on Electromagnetic Compatibility, Denver, CO, 2013, pp. 730-734, doi: 10.1109/ISEMC.2013.6670506.
• how to place via's to reduce differential crosstalk, Beomtaek Lee
• "If the symmetry is maintained between differential victim and its aggressor source, the differential crosstalk is minimized. In the case of coupling between two differential pairs, minimal differential crosstalk exists at some asymmetric physical positions, which provides the possibility and flexibility to design low-crosstalk differential vertical structures.
• 
  
• L. Pizano-Escalante, O. Longoria-Gandara, R. Parra-Michel and J. L. Naredo, "Crosstalk Cancellation on High-Speed Interconnects Through a MIMO Linear Precoding," in IEEE Transactions on Microwave Theory and Techniques, vol. 61, no. 11, pp. 3860-3871, Nov. 2013, doi: 10.1109/TMTT.2013.2283845.
• Crosstalk cancellation from a heavy signal processing point of view.  This is interesting since they do the cancellation in the transmitter with precoding.
• S. Yong, V. Khilkevich, X. Cai, C. Sui, B. Sen and J. Fan, "Comprehensive and Practical Way to Look at Far-End Crosstalk for Transmission Lines With Lossy Conductor and Dielectric," in IEEE Transactions on Electromagnetic Compatibility, vol. 62, no. 2, pp. 510-520, April 2020, doi: 10.1109/TEMC.2019.2902070.
• How to model FEXT accurately.
• W. Jiang, X. Cai, B. Sen and G. Wang, "Equation-Based Solutions to Coupled, Asymmetrical, Lossy, and Nonuniform Microstrip Lines for Tab-Routing Applications," in IEEE Transactions on Electromagnetic Compatibility, vol. 61, no. 2, pp. 548-557, April 2019, doi: 10.1109/TEMC.2018.2818695.
• Better microstrip equations