Study → Create Variant → Compare → Report
- Complete Study → Half Car · 4 DOF → Frequency, then select Compare a design change in Frequency. This preserves the Study definition and configures an identical Variant A without running.
- Alternatively open Compare → Half Car · 4 DOF → Frequency. Review the shared grid and explicit rear delay.
- Add physical parameter differences for Variant A. Unlisted parameters follow Baseline; Use Baseline removes an override.
- Choose Run Half-Car Frequency Compare. Inspect changed parameters, sampled peak changes, and Body, Suspension, Tires or Phase views.
- Export completed CSV, individual plot PNGs or the global Engineering Report. Export project preserves definitions; imported projects require a new run.
Two trusted Studies, one shared experiment
Both configurations use the unchanged Half-Car dynamic-stiffness solver, version 0.1.0. Baseline executes first, then Variant A. A comparison becomes available only after both succeed. All nine complex Study responses remain in completed evidence, including body heave, acceleration, pitch, both unsprung displacements, suspension travels and tire deflections.
The model, solver, logarithmic grid, point count, road type, explicit rear delay and front-road reference must match. Mismatches are rejected; there is no cross-grid interpolation. Variant A can change any of the twelve physical parameters, including geometry. Geometry differences preserve the shared delay.
Magnitude, phase and sampled peaks
Every plot uses the exact completed frequency samples. Baseline uses a solid line and Variant A a dashed line, with explicit legends. Displacement responses use m/m, body acceleration uses s⁻², and pitch uses rad/m. All phases reference the front road.
CSV phase differences use the shortest signed angle in [−180°, +180°]. Thus +179° to −179° is +2°. Exact half turns retain the sign of the raw difference. Exact zero response follows Study's zero-phase convention; phase there is a convention, not a resolved physical angle.
Peak tables compare the largest strict interior sampled maximum from each Study. A shift may compare different response features and is not a natural-frequency shift. Missing peaks have no delta. Relative magnitude change is (Variant A − Baseline) / |Baseline|, unavailable when the Baseline magnitude is at most 10⁻¹² in the stated unit. Absolute change remains available.
Evidence, freshness and exports
Edits mark completed evidence stale; exact definition restoration makes it current without rerunning. Changing the response view does not affect freshness. A failed rerun preserves the previous completed comparison; a failed first run publishes none. Journey marks this capability only after success.
CSV preserves all nine Baseline/Variant real, imaginary, magnitude and phase responses, magnitude deltas and wrapped phase deltas. PNG uses engineering plot data. PDF includes both full configurations, changed parameters, all sampled peak comparisons, eight overlays and provenance. Exports do not rerun physics. Project v1 contains definitions and metadata only.
Verification and limits
Comparison tests require exact Float64 equality against standalone Study on both sides, including all nine complex channels. Additional checks cover identical designs, one-parameter changes, symmetric mirroring, grid/excitation rejection, missing peaks, phase wrapping and Worker parity. The independent state-space, harmonic RK4 and modal checks from Frequency Study remain unchanged.
This is passive linear 4-DOF harmonic steady-state analysis. It complements transient road-event analysis. No experimental vehicle correlation, optimization or recommendation is claimed. Frequency Explore, sweeps, new physics and frequency video remain deferred.