Modeling and Co-simulation of Signal Distribution and Power Delivery in Packaged Digital Systems

Modeling;Co-simulation;Power delivery

Mandrekar, Rohan Uday ; Electrical and Computer Engineering

University:Georgia Institute of Technology

Department:Electrical and Computer Engineering

关键词: Modeling;    Co-simulation;    Power delivery;   

Others  :  https://smartech.gatech.edu/bitstream/1853/10459/1/mandrekar_rohan_u_200605_phd.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

The pursuit for higher performance at a lower cost is driving rapid progress in the field of packaged digital systems. As the complexity of interconnects and packages increases, and the rise and fall time of the signal decreases, the electromagnetic effects in distributed passive structures become an important factor in determining the system performance. Hence there is a need to accurately simulate these parasitic electromagnetic effects thatare observed in the signal distribution network (SDN) and the power delivery network (PDN) of an electronic system. The accurate simulation of high-speed systems requires information on the high frequency transient currents that are injected into the power distribution network causing simultaneous switching noise. Existing techniques for determining these transient currents are not sufficiently accurate. Furthermore existing transient simulation techniques suffer from two major drawbacks: 1) they are not scalable and hence cannot be applied to large sized systems, and 2) the time domain simulations violate causality. This dissertation addresses the above-mentioned problems in the domain of high-speed packaging. It proposes a new technique to accurately extract the transient switching noise currents in high-speed digital systems. The extracted switching noise currents can be used in both the frequency domain and the time domain to accurately simulate simultaneous switching noise. The dissertation also proposes a methodology for the transient co-simulation of the SDN and the PDN in high-speed digital systems. The methodology enforces causality on the transient simulation and can be scaled to perform large sized simulations. The validity of the proposed techniques has been demonstrated by their application on a variety of real-world test cases.