Reciprocity
Reciprocal vs. Non-Reciprocal Devices
| Device | Reciprocal? | S12 = S21? | Mechanism |
|---|---|---|---|
| Passive filter | Yes | Yes | Linear dielectric |
| Transmission line | Yes | Yes | Linear conductor |
| Antenna (passive) | Yes | TX pattern = RX pattern | Maxwell's equations |
| Amplifier (LNA, PA) | No | No (gain ≠ isolation) | Active transistor |
| Circulator | No | No (S21 ≠ S12) | Magnetized ferrite |
| Isolator | No | No (forward ≠ reverse) | Magnetized ferrite |
∮ (E1×H2 − E2×H1) · dS = 0
For any closed surface in a reciprocal medium
S-matrix consequence:
[S] = [S]T (symmetric) for reciprocal networks
Sij = Sji for all i, j
Antenna reciprocity:
GTX(θ,φ) = GRX(θ,φ) at every angle
Frequently Asked Questions
Antenna measurement impact?
TX pattern = RX pattern (guaranteed). Measure any antenna by receiving, not transmitting. Eliminates high-power source requirements. Works for all passive antennas. Active antennas with integrated amps may not be reciprocal.
Why are ferrite devices non-reciprocal?
Magnetized ferrite has asymmetric permeability tensor: CW vs. CCW waves see different properties. Violates Lorentz conditions. Enables circulators (S21 ≠ S12) and isolators. Only passive non-reciprocal devices.
S-matrix symmetry in practice?
S12 = S21 for all passive components (filters, cables, dividers). Active devices: S21 = gain, S12 = isolation (different by design). Verifying symmetry in simulation catches modeling errors. In VNA calibration, reciprocity of the thru standard is assumed: any measured asymmetry in S12 vs. S21 of a passive thru connector indicates a calibration problem, not a physical asymmetry, making reciprocity a built-in calibration diagnostic.