Real measurements allow the engineer to proactively improve the design, saving time and money.
Shrinking timelines, budget constraints and the overwhelming need for rapid innovation challenge even the most battle-weary and experienced engineer. But this is common in most industries. However, the ubiquity of electronics in modern life (imagine a world without laptops and cell phones) simply underlines this fact for those working in the Electronics Design Automation (EDA) industry.
For engineers wanting to learn from others and the industry as a whole, the best way to approach design is through methodologies and successes that will enhance overall productivity. They quickly learn that at every stage of the circuit design flow (capture, simulation, layout, validation and fabrication), there is room to improve practices and to spend less time and money in design.
Engineers have for sometime had circuit simulation as a tool to not only help drive better analysis of circuit decisions, but also to use that analysis earlier in the design flow. The idea is that if more time is spent improving the printed circuit board (PCB) design using simulation, time can be saved later through fewer prototype iterations and respins.
Enabling simulation to be a true best practice, however, requires us to better bridge the gap between the world it emulates and the virtual interface. This article will show the true advantage in simulation, particularly as we use ever-increasingly complex models through real measurements.
The Need for Simulation
The backbone of electronics continues to be the PCB because it interfaces to various signals and external systems. Despite its varied use throughout the past 80 years, the PCB continues to require a high level of activity through iterative design attempts. The fact is that errors continue to be identified and, for many engineers, this results in the tedium of multiple prototype stages. As an engineer once mentioned, it is “offensive to his sanity” when having to continuously iterate on a design. These errors can be the result of a non-appropriate topology, or oversights in design decisions. Identifying these issues at the prototype stage means a return to the drawing board to correct the problem; this unfortunately leads to a major time and budget sink. Engineers need to avoid these iterations. Identifying errors before silicon meets copper translates into a quicker and more productive route to final product.
Simulation Program with Integrated Circuit Emphasis (SPICE) allowed for the modeling of circuit devices to help provide more in-depth understanding of board-level (and IC) designs. Parasitics of the design can be realized and accounted for, providing a tool to better equip engineers when making appropriate design decisions.
It is important to mention that simulation, to a great degree, has been misrepresented in the past. It does not completely replace the need for the physical prototype. Its role is to investigate behavior of design decisions and to remove the common cascade of errors to the prototype that force the multiple iterative stages between prototyping and manufacturing. The simulation concept is meant to shave away time in the design cycle associated with error-related prototype iterations.
Improving Simulation with Real Measurements
Circuit simulation is increasingly used technique in design. However, simulation is fundamentally a mathematical model of device behavior and is inherently limited by that math. Prototype validation still uncovers flaws that were not seen in the mathematical models of simulation.
Therefore, the ultimate limitation for circuit simulation depends on how comprehensive the models are. Models for transistors, switched-mode power supplies and MOSFETs are complex and sophisticated and emulate circuit behavior quite accurately within defined regions of operation. In fact, SPICE models have been noted to sometimes better realize parasitics that will appear in the prototyping stages than bread boarding and other ‘traditional’ design approaches. To improve simulation, we need to improve any remaining idealized or simple models.
When evaluating the simulation environment, it is important to note that design factors such as noise, cross talk or even the complex nature of a signal (perhaps a heartbeat or mechanical signal) are not traditionally included as a stimulus in simulation. This is because the source that usually validates a schematic is based on a simple curve, such as a sine wave or square wave. This can be considered as the most prevalent source of idealization on the individual engineer’s desktop.
If this source model can be improved to better represent a real signal, then we can increase the effectiveness of simulation. The idea is therefore to try to plug a real source into a schematic and simulate it based on a real signal. Real measurements have typically not played a standard role at such an early stage of design. The first instance of real signals being introduced to the design flow has typically been during prototype validation. However, through an earlier introduction, we can facilitate a far improved simulation model of a stimulus.
The Virtual Prototype
The integration of simulation and real stimuli in a systematic design approach is termed Virtual Prototyping — bridging the gap between emulated behavior and reality to build a more effective computer-based model. This is seen in multiple industries including mechanical and construction engineering.
An electronic system, for example, must often interface to a mechanical force. A mechanical stimulus can introduce a repetitive noisy element to the input of a designed PCB. If not accounted for at this time, then it will not be identified and resolved until prototype validation —when it is implemented into a system.
Introduction of this signal as a simulation stimulus will help visualize the effect of that noise before the prototype stage. Filtering and signal conditioning for this element of the design can be built before the first iteration.
Again, biomedical signals, noisy inputs, periodic functions and nontraditional stimuli can be introduced earlier in design for a more effective approach. Modern measurement hardware is able to acquire real signals and to save it to an ASCII text format. This can be done both by manually transferring it (via disk or thumb drive) to a PC, or more easily, by using the PC directly through a General Purpose Interface Bus (GPIB) to acquire and to save measurements. Simulation software can then load the text-based measurements as a stimulus for simulation.
A Shift in Design Practices
Making the transition of design analysis to an earlier stage of the design flow requires a shift in the thinking of the electronics design community. With as many as 28% of small companies requiring as many as three respins of designs, as noted in a report by the Aberdeen Group in 2007, it is apparent that there is still room to improve and to address issues earlier in the design flow.
In the past few years, innovation has created Design for Manufacturing (DfM) and Design for Assembly (DfA). The intent is to analyze a design earlier in order to understand the potential for fabrication and manufacturing issues. Integration of the PCB design flow with manufacturing tools has created an EDA-industry-aided infrastructure to shift the focus of appropriate fabrication information, data and techniques earlier. We are now able to aid manufacturing by shifting analysis even earlier through a more effective use of simulation. Combining real measurements with simulation provides an effective platform to reduce errors even earlier.
Building the Virtual Prototype into the Design Approach
The idea is that for a successful approach to design, the virtual prototype should be employed as a third and final validation stage prior to prototyping. There are three stages of simulation-aided validation: component evaluation, iterative simulation and analyses and virtual prototyping.
The first stage of simulation use is based on the evaluation of components — judging the performance of transistors, operational amplifiers and other devices in small modular circuits to assess their use. Datasheets are still a trusted and highly available source; however, the convenience of using the simulation models provided by most chip and device manufacturers can greatly enhance design productivity. Also, the flexibility of modern ECAD tools means that there can be easy experimentation of “what if” scenarios by examining the effect of different parts.
Shifting to an interactive and iterative simulation approach enables a far more in-depth view of design. Building small modular circuits and analyzing behavior provides graphical feedback on the topology. Analyses ranging from transient to frequency sweeps to noise figure and sensitivity are commonplace in SPICE simulation tools and, as such, help characterize overall behavior. For example, few would be able to comfortably validate a filter without the added analysis of bode plots (frequency sweep) as validation.
At this point in the design flow, the overall behavior is well known and the specifications from an idealized view of circuit performance are complete. Real measurements and the virtual prototype approach can provide the seal of approval on a design. Introduction of the nuance of real signals can provide the added assurance of how a design will perform in the real world. Though virtual prototyping does not remove the need for a physical prototype and its validation, it does allow for more errors to be identified and resolved earlier in the design flow.
Facilitating Integration between Simulation and Real Measurements
As the need for more innovative design practices grows, it is fair to say that integration between traditionally separate stages is necessary. For real measurements to have a greater role in PCB design and simulation, greater integration between design and test systems is required for the acquisition of real data to be seamlessly provided to simulation inputs.
As technology can be better used to acquire and to link the stages of the design flow, so the divide between simulated measurements and real will be bridged.
The easier that this process can become, the more likely it is that engineers will use the overall benefits of improved simulation in the design flow. As such methods become standard throughout the industry, so will the productivity and effectiveness of design.
Conclusion
Technology has provided us with increased computing capabilities, the inclusion of real measurements into simulation has the ability to evolve SPICE and circuit simulation for improved productivity in design. The virtual prototype is simulation, the next generation. By validating design decisions with real effects, engineers are provided with the impetus to proactively improve design at conceptive stages rather than after latter validation. The result is saved time, saved budget and, perhaps for the engineer, some saved sanity. PCD&F
Bhavesh Mistry is product marketing engineer with National Instruments Electronic Workbench Group and can be reached at This email address is being protected from spambots. You need JavaScript enabled to view it..
