Measurement Types & Analyses¶
Every measurement type and analysis RFlect supports, the input it expects, what it computes, and the MCP tool that drives it. RFlect is fully deterministic. Every value here is computed from your data, with no LLM or API key involved.
Scan types (primary measurement modes)¶
| Scan type | Input | What RFlect computes | Primary MCP tools |
|---|---|---|---|
| Active TRP | WTL chamber .txt (transmitting DUT) |
Total Radiated Power, H/V power split, EIRP, 2D/3D power patterns | import_antenna_file, analyze_pattern, get_gain_statistics, bulk_process_active |
| Passive gain | WTL chamber HPOL + VPOL .txt pair |
Total/H/V gain, efficiency, directivity, HPBW, sidelobes, polarization | import_passive_pair, bulk_process_passive, get_gain_statistics, compare_polarizations |
| VNA / S-parameters | Copper Mountain / generic VNA .csv, Touchstone .s2p |
S11, VSWR, return loss, impedance bandwidth, resonances | analyze_s11 |
Analyses (the Tools menu + MCP surface)¶
| Analysis | Input | What it produces | MCP tool |
|---|---|---|---|
| Bulk passive processing | Folder of HPOL/VPOL pairs | Batch gain/efficiency/pattern outputs | bulk_process_passive |
| Bulk active processing | Folder of active TRP files | Batch TRP/pattern outputs | bulk_process_active |
| Polarization analysis | Passive HPOL/VPOL | Axial ratio, tilt, sense, XPD, H/V balance | compare_polarizations |
| Antenna comparison | 2+ loaded measurements | Per-frequency peak gain/TRP overlay, deltas vs reference, best-per-frequency | compare_antennas |
| Group delay | Transmission phase sweep (S21/S12) | Group delay vs frequency, peak-to-peak, in-band flatness, distance error | analyze_group_delay |
| System Fidelity Factor (UWB) | S2VNA .csv / .s2p pair |
SFF, transfer function, impulse response | calculate_sff_from_files, analyze_uwb_channel |
| Impedance bandwidth | S11 sweep | Return loss, VSWR, fractional bandwidth | analyze_s11, get_impedance_bandwidth |
| Horizon / maritime statistics | Active/passive pattern | Horizon-plane gain stats, coverage | get_horizon_statistics |
| Frequency extrapolation | Loaded measurement | Estimated metrics at an unmeasured frequency | extrapolate_to_frequency |
| MIMO / diversity | Envelope correlation (ECC) | Vaughan-Andersen diversity gain, 2×2 capacity, isolation rating | analyze_mimo_diversity |
| Link budget / range | Measured gains + path-loss model | Max range, link margin, fade margin (Rayleigh/Rician) | estimate_link_budget |
| Calibration drift | TRP-Cal run archive | Drift vs prior epochs, setup-group mismatch flags | cal_drift_ingest, cal_drift_compare, cal_drift_report |
| Active chamber calibration | Power + gain-standard + HPOL/VPOL reference files | TRP cal file + summary; auto-recorded to drift history | generate_active_cal |
| CST FFS conversion | HPOL/VPOL data | CST-format far-field source | convert_to_cst |
| Combined / branded report | Any loaded measurements | DOCX with plots, gain tables, deterministic or agent-authored prose | generate_report, preview_report |
| Folder orchestration | A folder of any of the above | One-call auto-detect → import → analyze → optional report | process_folder |
| Multi-angle iperf validation | Two bench iperf sessions (installed + reference) at matched azimuths | Per-angle throughput delta, worst/best angle, polar plots | analyze_iperf_angle_sweep |
Where the math lives¶
All analyses are pure, deterministic functions in plot_antenna/ (calculations.py, analysis_engine.py, uwb_analysis.py, cal_drift.py, file_utils.py). The MCP tools in rflect-mcp/tools/ are thin wrappers that validate inputs, call these functions, and return structured results. They never make network or LLM calls.
See the Tools Reference for signatures and return shapes, and the User Guide for the methodology behind each measurement type.