Electronic Products & Technology

Device modeling: the first place where EDA meets measurement

Over the years modern measurement has evolved to accommodate the changing needs of designers. Today, a new evolution is underway, prompted by the increasingly complex nature of modern defense- and commercial communications-based designs.

September 7, 2012  Staff

That evolution aims to marry both electronic design automation (EDA) tools and processes with measurement for system level design and verification.

Figure 1: Using an integrated design, simulation and test methodology to evaluate bit-error-rate in a typical RF transceiver enables designers to uncover design flaws earlier, while decreasing cycle time. (Click on image to enlarge.)

Doing so promises to enable designers to more quickly and accurately bring a system from “idea” to “product.”

One area where measurement and EDA has long been successful in working together is device modeling. Since designs are built from device models, their accuracy is paramount. Models, however, are only as good as the measured data used to extract them, making measurement critical to gathering accurate and meaningful device data for modeling.

Today, dedicated EDA software controls measurement of device data for modeling and optimizes the extraction of compact model device parameters based on a series of specific measurements. This technique extends beyond transistor-level modeling to modeling in general (e.g., S-parameters for high-frequency linear devices and X-parameters for nonlinear devices).

web-author pk lamModeling systems now produce higher-level “black box” models that capture the complete nonlinear behavior of an entire subsystem and are used by high-level system simulators, such as Agilent’s SystemVue or Advanced Design System, for design and validation. {nomultithumb}

By unifying baseband algorithm development (DSP) together with accurate RF/analog design, system architects can now more quickly identify and address architecture design challenges in modern communication systems.

The same integration that worked successfully for device modeling is now needed to close the gaps between design, validation and final manufacturing test. The need stems from the sheer speed with which the communications industry moves (e.g., existing standards continually change and new standards are constantly in development) and the complexity of modern communications systems – both of which translate into new challenges confronting designers.

Here, traditional linear-based design methodologies fall short. On both the design and validation sides of a given flow, multi-discipline teams tend to work in relative isolation with limited interaction across modeling and signal domains.

Analog/RF teams rarely interact with their baseband DSP or embedded systems counterparts during the design or hardware validation phases, leaving much uncertainty until final hardware integration and testing.

Furthermore, designers use many different design languages and point tools with poor to no interaction with the real world. The lack of common methods and test benches that span the many disparate tools and multi-discipline domains means there are simply too many places where design flaws can hide and performance issues can be masked. In addition, there is no way to truly validate and optimize design performance without expensive over design.

In this complicated design environment, connecting design with test is the only true way to quickly and accurately turn an idea into a product. Using an integrated design, simulation and test methodology would allow designers to virtualize measurement science early in the design phase to bring laboratory-quality measurement to the design creation process and simultaneously analyze virtual models of a system with actual hardware measurements (Figure 1). Additionally, they could bring their design intellectual property (IP) into the measurement space for more specialized and accurate hardware validation.

The benefits to be reaped by this approach are many. Design flaws can be captured early and corrected in a time- and cost-efficient manner, reducing the overall cost of development and improving development cycle time. The result is increased design margins, more accurate simulations, improved cross-domain team collaboration, and the ability to demonstrate design performance to customers earlier.

Manufacturing test engineering time is also reduced since designers can use design IP from research and development to generate and analyze the test vectors and waveforms used in final hardware and system test. Furthermore, the integrated methodology affords designers the flexibility to quickly interpret and adapt to complex evolving requirements.

With such benefits, an integrated design, simulation and test methodology offers the ideal answer to the challenge of quickly turning today’s communications system ideas into tomorrow’s products.

P.K. Lam manages the Agilent EEsof business in Eastern Canada.

Print this page

Related Stories

Leave a Reply

Your email address will not be published. Required fields are marked *