Ever wonder how companies build some of the best-known pieces of technology and massively popular products? Of course there are engineering challenges to overcome spanning hardware and software, but simulations are an important tool for pre-build and post-build qualification. A simulation-driven design process aims to bring simulation capabilities into the front-end of a design and ensure a new design can operate properly.
What process would be involved in simulation-driven design in advanced electronics? An electronic product is more complex than just its circuitry, which means simulation-driven design must happen at multiple levels to provide useful insight into the operational reliability of a product. Simulations give designers the ability to look at multiple levels in a system while only focusing on certain performance aspects. Taking a domain-specific approach provides focus on specific performance characteristics and design factors that are not approachable from a prototype and testing process.
Your Organization’s Simulation-Driven Design Process
Simulation-driven design and engineering is more common than one would think and it is an indispensable part of developing many complex systems. There are many goals in simulation-driven design, but one of the most important is the qualification of a new design and its intended prototype before volume production. If prototype builds can be eliminated, then costs are saved and the path to market is accelerated. An up-front investment in simulation and analysis software are needed, but the payoff in terms of competitiveness and cost savings can be substantial.
What to Simulate
The specific domains to simulate and analyze for an electronic product are outlined below. Many companies today operate at the systems level, and their development activities span from chips up to entire assemblies and systems. The list of simulation activities below looks at all levels of a product, from its individual chips/packages up to the entire assembly.
Assembly (thermal and electrical)
Focuses on the PCBA and its enclosure
Multiple thermal and mechanical reliability simulations
The process for simulation-driven design requires inspecting each of these areas with numerical solvers. These areas are addressed individually at different points in the design phase, but they can be put through an iterative simulate-analyze-design-repeat process as outlined below.
Simulation-driven design is an iterative process because, oftentimes, the design being evaluated or investigated is part of an experiment. You don’t always know how a particular circuit or assembly will operate in practice, so simulations become the only way to iteratively optimize a portion of a design and reach performance targets. The optimization concept in simulation-driven design is a central theme in the overall process.
The simulation-driven design process is outlined broadly below. The process begins with a finished design that needs to be qualified and evaluated in a numerical solver. Typically, there will be some performance target to examine and qualify in the simulation, which will be informed by some broader performance metric in the final product. Engineering requirements gathering in the initial portions of the design phase are often referenced here to illustrate performance targets and as a basis for comparison.
The simulation-driven design process is formulated under a simple iterative concept as outlined below. RF engineers and high speed digital engineers are most likely familiar with this process, but the same ideas apply at the system level. The process should be applied in each of the areas in the above table to better understand system behavior and reliability. Eventually, through an iterative process, a systems designer can find an optimal design that balances performance, cost, reliability, and manufacturability.
Logic and Schematics
For an electronic system, the schematics are the first place to start with simulation-driven design. The schematics are where circuits are constructed and later transferred to a PCB or assembly for placement and physical layout. Because schematics show the electrical connections that will appear in the final assembly, they are the ideal place to start implementing a simulation-driven design process specifically targeting logic.
The major types of simulations used in schematics at the system level are:
- SPICE simulations for circuits (usually analog circuits)
- Logic simulations for complex digital circuits
- Signal integrity simulations with IBIS models
- Signal integrity simulations with network parameters (usually S-parameters)
Because schematic simulations with SPICE or IBIS modeling are electrical simulations, electrical models are required to better understand how certain components will operate in a larger system. To learn more, read our resource on S-parameter extraction from DUTs and channel measurements. This important part of front-end design and engineering will give a systems designer the S-parameter models they need to simulate the behavior of important components used in your system.
Consider the below simulation example. Operating at the system level requires each component to have an electrical model to describe signal propagation. Effectively, these models map inputs to outputs.
System-level phase noise simulation along an interconnect performed in Cadence AWR.
Electrical models in front-end simulations must be determined from a component-level simulation or directly from measurements. For many RF components (amplifiers, filters, etc.), these models are supplied by vendors or they can be measured with some test equipment.
There is no single list of simulations to perform on a PCB to qualify performance and/or reliability. Different devices will be deployed in different environments, and the performance or reliability needs of each sub-system on a PCB may not need to be qualified at deep levels individually. On the PCB, there are three broad areas where the simulated electrical performance should be evaluated:
- Whether specific sub-systems in the design will meet operating needs
- How electrical performance in the layout deviates from logic simulation results
- Identify which portions of the layout need more detailed simulation to understand behavior
Due to the complex layout of PCBs, these systems are simulated using numerical field solvers. With a field solver, signal behavior in the design can be modeled in the presence of a complex set of nearby components. Due to the presence of components, other conductors, specific materials, and other physical items in the assembly, the signal behavior in a PCB layout will deviate from the ideal behavior being simulated in the schematics.
In simulation-driven design, it’s useful to divide up the PCB into specific portions that are simulated individually. This provides a more targeted approach that allows specific design choices to be evaluated and qualified. Specific subsystems that are normally simulated are the PDN, high-speed channel and connector interfaces, RF elements like antennas, and transmission line designs to verify losses/impedance.
Chips and Packaging
In addition to the PCB, there are some individual components that are often simulated as part of systems analysis. Semiconductors and their packaging are another area where a design should be evaluated thermally and electrically. Semiconductor components are normally simulated with their own set of analyses during their design phase prior to signoff. The packaging (for SoCs/modules) and the placement of these components in a PCB also need to be simulated for signal integrity, thermal behavior, and mechanical behavior.
Example thermal simulation results for a component and its packaging.
What about thermal simulations involving the chip/package in the PCB? Because a PCB is an electrical system, and because all electrical systems generate some heat, the thermal characteristics of a PCB can’t be examined in isolation. Instead, the interplay between thermal and electrical behavior in the design design must be
These simulations are performed at the end of the design phase the entire assembly design has been completed and is ready for full evaluation. These simulations often center around reliability and thermal behavior, or they could involve EMI/RF qualification. This is summarized in the table below.
In the Design phase of a simulation-driven design process, the overall system does not typically go through extensive redesigns unless there is a catastrophic failure. For example, the potential for a particular component to fail (e.g., solder ball fracture) due to thermal or mechanical shock would be one reason for a design change. Such a change could be swapping the failing component for an alternative package.
Instead of redesigning every aspect of the PCB, some strategies like assembly-level or enclosure-level design changes might be more appropriate. This is because they could involve lower costs, or they could reduce the redesign time required to correct any reliability problems. For example, when any cooling measures are found to be inadequate, the enclosure might be changed to allow more airflow or conductive cooling through the chassis. Implementing a reliability-based change on the PCB or component itself in this case would simply be too expensive.
CFD simulations are used to track airflow throughout an enclosure and predict equilibrium temperature in the assembly.
Because these simulations can be very complex, the right simulation toolset is needed. These systems are often simulated as multiphysics problems so that coupling between physical behavior in different domains can be understood. Although these tools can be complex, EDA vendor ecosystems can help users quickly move between design tools and simulators for an ideal simulation-driven design process.
Design teams working on advanced electronic products can get to market faster with higher quality by leveraging the complete set of system analysis tools from Cadence. Only Cadence offers a comprehensive set of circuit, IC, and PCB design tools for any application and any level of complexity. To learn more about interconnect simulation in high-speed systems, watch our webinar, Understanding Simulation Analysis Parameters for Serial Link Systems in SystemSI.