SPICE vs ERC: simulation and schematic checks

Use ERC for structural diagnostics

ERC can identify suspect pin interactions and connections based on the symbol and project rules. It does not compute whether an amplifier has the requested gain or a power stage behaves correctly under a load transient.

Use simulation for measurable behavior

Define stimuli, models and acceptance ranges before running SPICE. Inspect operating points and relevant corners. Simulation can still produce a persuasive result for the wrong physical device if the model mapping or assumptions are incorrect.

Connect both to the actual design

Ensure the simulated netlist corresponds to the intended schematic. Preserve model versions and testbench files. Review non-modeled requirements, part ratings and assembly mapping rather than treating one successful analysis as a complete release decision.

Practical checklist

1. Run structural checks.
2. Verify model pin assignments.
3. Define numerical observations.
4. Test relevant operating conditions.
5. Review unmodeled requirements.

A concrete example

Validate a generated sensor amplifier.

Evidence to accept

ERC checks structure and simulation checks required transfer behavior under stated conditions.

Evidence that is insufficient

The simulator converges, so the board is claimed correct without checking the model or footprint.

Common questions

Primary references and further reading

Use the matching device documentation and project configuration when applying this checklist. These references explain the underlying tools and evaluation approaches.

• KiCad schematic editor documentation
• ngspice documentation