Introduction
Simulation software provides invaluable assistance to PCB designers verifying circuit performance before physical prototyping. However, the maxim "garbage in, garbage out" applies - inaccurate inputs produce meaningless outputs. Setting up simulations properly requires awareness of key sources of error to avoid. This article examines common pitfalls using simulation tools and strategies to avoid wasting time and effort on faulty analyses.
Know Your Circuit Conditions
Attempting to simulate a circuit without understanding operating conditions and design constraints will lead to unrealistic results. Key factors to define upfront are:
- Input signal characteristics - frequency, amplitude, rise/fall times, etc.
- Expected loading - resistive, capacitive, inductive
- Output current/power requirements
- Voltage levels - logic levels, power rails, signal swings
- Frequency range of interest
- Environmental conditions - temperature, moisture
- Applicable regulatory standards
Trying to simulate without pinning down the above will yield generalized results unrelated to actual performance. Clarify the specific use case and operating context.
Pay Attention to Model Fidelity
Most simulators rely on component models to emulate real-world behavior. Inaccurate models lead to false results. Areas to check closely are:
- Passives - Parasitic characteristics like ESR, ESL, and tolerance need proper modeling at frequency of interest.
- Active devices - Non-ideal properties not captured by ideal transistor models. Verify model valid over operating conditions.
- Board effects - Trace impedance, couplings, and stray capacitance affect high frequency performance.
- Connectors, cables - Necessary for system-level simulations including external interfaces.
- Radiated emissions - Very difficult to model accurately so focus on relative comparisons.
In general, only rely on modeled results up to ~10% of a component's specified bandwidth or frequency limit. Review modeling guides carefully.
Double Check Your Netlist
Most simulators rely on a netlist generated from the schematic to connect components together and define connectivity. Any errors or mismatches with the actual circuit will lead to incorrect results. Areas susceptible to netlist problems:
- Unconnected pins or nodes
- Missing components or pins
- Undefined component pinouts
- Incorrect pin assignments
- Unmatched hierarchies between schematic and netlist
Run basic electrical rule and connectivity checks before simulation to identify netlist issues. Review any warnings or errors.
Validate Model Libraries
Simulators include model libraries with pre-defined parts covering common components. However, default models may not match vendor datasheet specifications leading to inaccurate results if used without validation. Recommendations are:
- Review all model parameters against datasheets
- Replace generic models with manufacturer models when possible
- Create custom models from datasheets for specialized components
- Verify models over full range of operating conditions
- Update models as new device data becomes available
Maintaining an validated component model library is essential for reliable results.
Match Simulation Goals to Solver Capabilities
Modern simulators offer a variety of analysis options from DC operating point to complex multidimensional solvers. Selecting the inappropriate solver wastes resources. Match analysis goals to available solvers:
- DC, small signal AC - Basic nonlinear DC and AC solvers
- Transient - Time domain waveform analysis. Needs appropriate timesteps.
- AC/RF - High frequency S-parameter and impedance analysis
- Thermal - Steady state and transient thermal modeling
- Channel/Stripline - High speed signal propagation analysis
- EMI/EMC - Emissions modeling. Complex problem.
- IBIS - Modeling interfaces between components
Consult simulator documentation to select the right virtual experiments.
Verify Manufacturability and Constructability
A circuit may simulate successfully but still run into issues when actually manufactured and built up. Factors like manufacturability, assembly methods, and enclosure/chassis effects are often overlooked. Double check:
- Tolerances - Do values exist for easy purchase or assembly?
- Footprint and pinout - Suitable for available packages?
- Layout constraints - Simulate effects of board layout.
- Enclosure - Model effects of chassis and components.
- Thermal - Heating and airflow in enclosure.
- Testability - Can all nodes be probed?
- Reliability - FMEA study for high MTBF.
While not always modeled, thinking through these issues identifies potential pitfalls before prototyping.
Correlate and Validate with Measurements
The gold standard is correlating simulated results across multiple levels of modeling fidelity with empirical measurements from the actual built up circuit. This helps build confidence in the simulator while revealing modeling limitations. Suggested validation steps are:
- Simulation 1 - Simplified linear model
- Simulation 2 - Detailed nonlinear model
- Simulation 3 - Full 3D model including layout
- Prototyping - Build up representative circuit
- Measurement - Characterize with expected instruments
- Correlation - Compare models and measurements
- Refinement - Improve model based on differences
This incremental process gets the most value from the simulation toolset.
Summary
Leveraging EDA simulations allows PCB designers to identify and fix issues before costly prototypes are built. But care must be taken to configure simulations properly and match models to the real system to avoid "garbage in, garbage out". Follow the guidelines in this article to maximize the utility of virtual prototyping while minimizing false results. What best practices do you follow when simulating PCBs?
Frequently Asked Questions
Q: What are some common simulation "gotchas" to avoid?
A: Not accounting for component tolerances, improper test fixture modeling, exceeding valid frequency range of models, and insufficient simulation timesteps.
Q: How can I check a netlist matches the schematic?
A: Cross-probe nets between schematic and netlist in the EDA tool. Review netlist in text editor checking component names match schematic. Check for dangling nodes.
Q: What modeling improvements come with more advanced simulators?
A: Finite-element modeling of conductors, full 3D electromagnetic field solvers, thermal modeling, vibroacoustic models, and support for analog behavioral modeling.
Q: What measurements best validate a simulation?
A: Frequency sweeps, transient responses, eye diagrams, impedance measurements based on simulation type. Environmental stress screening also helps correlating MTBF predictions.
Q: How can I determine if a model library is accurate?
A: Review model parameter data sheets in detail and compare against manufacturer datasheets over full operating range. Build simple test circuits focused on verifying model performance.
No comments:
Post a Comment