Analog FastSpice RF delivers noise analysis for RF circuits

Berkeley Design Automation Inc. has announced AFS RF (Analog FastSpice radio frequency), which Chief Operating Officer Paul Estrada calls the industry’s first true Spice-accurate noise-analysis tool for RF circuits. AFS RF accurately analyzes nanometer-scale device noise impact for all types of prelayout and postlayout circuits, ensuring early insight into its impact on performance, power, and area.

Before the emergence of AFS RF, designers had to use limited-spectrum RF tools that can only approximate device noise impact on RF circuits, Estrada explains. Such approximations are increasingly inaccurate with decreasing process geometries, often becoming grossly inaccurate in nanometer-scale circuits. Circuits with sharp transitions, such as switched-capacitor filters, charge pumps, and dividers; high-frequency circuits, such as RF front-end blocks; and oscillators are especially sensitive to these inaccuracies. Without accurate analysis, designers must include expensive design margin or risk missing specifications in silicon.

Using the industry’s first full-spectrum device-noise-analysis engine, Analog FastSpice RF provides true Spice accuracy for every run. For complex circuits, it is five to 10 times faster than traditional RF tools that can only approximate device-noise effects. AFS RF features the DNA (device noise-analysis) Advisor to characterize DNA requirements, high-capacity periodic-steady-state analysis for greater than 100,000-element postlayout circuits, full-spectrum periodic-noise analysis with true Spice accuracy, full-spectrum total oscillator-device-noise analysis capability with phase and amplitude noise, and harmonic balance for fast single-tone analysis of moderately nonlinear circuits.

“We have been using the AFS platform for the last two years for full-circuit transceiver verification and more recently transient noise analysis of our analog/RF blocks,” says AlanWong, head of IC design at Toumaz Technology. “AFS RF delivers full-spectrum periodic-noise analysis and does not trade off accuracy for performance. AFS RF allows us to analyze our prelayout and postlayout RF blocks, delivering true Spice accuracy and a five- to10-times speedup over traditional RF-analysis tools.”

“Full-spectrum periodic-noise analysis is critical for accurate characterization of device noise in nanometer analog/RF designs,” says Boris Murmann, PhD, assistant professor in the department of electrical engineering at Stanford University.

“Without accurate device-noise analysis, designers need to add significant margin to ensure performance. This [addition] can be expensive. For example, adding just 0.5 bit of margin—that is, 3-dB SNR (signal-to-noise ratio) in a noise-limited circuit will double the required power,” he says.

Berkeley Design Automation Inc.

Toumaz Technology

Stanford University