Simulation challenges in designing high speed serial links

Signal speeds of high speed serial links double almost every generation and with these increasing speeds come a wide range of new modeling and simulation challenges. Modeling challenges involve making sure that models are passive, stable and causal. Frequency-domain models, such as scattering parame...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chada, A. R., Mutnury, B., Wallace, D., Winterberg, D., Minchuan Wang, Scogna, A. C.
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Signal speeds of high speed serial links double almost every generation and with these increasing speeds come a wide range of new modeling and simulation challenges. Modeling challenges involve making sure that models are passive, stable and causal. Frequency-domain models, such as scattering parameter models that have measurement noise or limited bandwidth or incorrectly performed interpolation or extrapolation operations, may exhibit non-causality and non-passivity in time domain. Simulating millions of bits in timedomain to measure the interface merit in terms of bit error rate (BER) is CPU and memory intensive. This challenge has given way to new simulation algorithms and methodologies. The challenge here is that no two simulation approaches result in the same answer. The difference between approaches is aggravated at high frequencies and with inclusion of effects like crosstalk and transmitter and receiver equalization. In this paper, the results from various simulation approaches are contrasted against each other and also against measurements to understand their inherent assumptions along with their impact in designing high speed SerDes.
ISSN:0569-5503
2377-5726
DOI:10.1109/ECTC.2012.6248821