An enhanced BSIM modeling framework for selfheating aware circuit design

Abstract

Abstract. This work proposes a modeling framework to enhance the industry-standard BSIM4 MOSFET models with capabilities for coupled electro-thermal simulations. An automated simulation environment extracts thermal information from model data as provided by the semiconductor foundry. The standard BSIM4 model is enhanced with a Verilog-A based wrapper module, adding thermal nodes which can be connected to a thermal-equivalent RC network. The proposed framework allows a fully automated extraction process based on the netlist of the top-level design and the model library. A numerical analysis tool is used to control the extraction flow and to obtain all required parameters. The framework is used to model self-heating effects on a fully integrated class A/AB power amplifier (PA) designed in a standard 65 nm CMOS process. The PA is driven with +30 dBm output power, leading to an average temperature rise of approximately 40 • C over ambient temperature.

Advanced design and characterization methodologies for memory-aware CMOS power-amplifier implementation

“…1 and 3. A detailed introduction into the automation and model verification is given in Schleyer et al (2014). The algorithm processes the AM/AM and AM/PM time-domain trajectories, and applies calibration with chirp-signals (Wolf et al, 2009).…”

“…2 Modelling, parameter extraction and design flow (Schleyer et al, 2014), provide more insights into the SystemC/AMS time-domain model (timed data flow model) which is linked to the MATLAB controller. The advantage is that model parameters can be directly used from the measurement setup, such as balun specifications, and from the design files, such as the impedance matching network values (Spectre small signal results, Sparameter touchstone files, csv data extracted from measurements, circuit simulations, or system specifications).…”

“…7 for a stacked-cascade single stage amplifier. Details on the transistor to transistor interaction due to thermal coupling is discussed in Schleyer et al (2014). …”