What is the effect of connector repeatability on measurement uncertainty at high frequencies?

What is the effect of connector repeatability on measurement uncertainty at high frequencies? Connector repeatability is the variation in measured S-parameters when a connector is disconnected and reconnected, and it becomes a significant uncertainty contributor at high frequencies: (1) What causes connector non-repeatability: pin depth variation: each connection results in a slightly different center pin contact position. At DC-6 GHz: this has negligible effect. At 40 GHz: a 0.01 mm pin position change causes approximately 0.05 dB and 1-2° change. At 110 GHz: the same 0.01 mm change causes approximately 0.15 dB and 5-7° change. Outer conductor alignment: the outer conductor shells may not align perfectly each time, changing the TEM mode boundary. Contact resistance: the contact resistance between the mating surfaces varies with pressure, cleanliness, and surface condition. Connector wear: after many connections (> 500-1000), the contact surfaces degrade, increasing non-repeatability. (2) Repeatability specifications: N-type (DC-18 GHz): S11 ≈ 0.01 dB, S21 ≈ 0.01 dB, < 0.5° phase at 18 GHz. SMA (DC-18 GHz): S11 ≈ 0.01-0.02 dB, S21 ≈ 0.01-0.02 dB at 18 GHz. 3.5 mm (DC-34 GHz): S11 ≈ 0.01-0.03 dB, S21 ≈ 0.01-0.02 dB at 26.5 GHz. 2.4 mm (DC-50 GHz): S11 ≈ 0.02-0.05 dB, S21 ≈ 0.01-0.03 dB at 50 GHz. 1.0 mm (DC-110 GHz): S11 ≈ 0.05-0.15 dB, S21 ≈ 0.02-0.05 dB at 110 GHz. These values assume proper torque, clean connectors, and connectors in good condition. (3) Measuring connector repeatability (Type A evaluation): connect the DUT (or a thru). Measure S-parameters. Disconnect and reconnect the DUT. Measure again. Repeat 10 times. Calculate the standard deviation of each S-parameter. This standard deviation is the Type A uncertainty contribution from connector repeatability. At 40 GHz with 2.4 mm connectors: typical S21 repeatability: ±0.02-0.03 dB (1σ). (4) Impact on measurement budget: for a measurement with 4 connector interfaces (VNA port 1 → cable → DUT input, DUT output → cable → VNA port 2): total connector repeatability contribution: u_conn = √(4 × u_single²) = 2 × u_single. At 40 GHz: u_conn ≈ 2 × 0.03 = 0.06 dB. This is significant compared to VNA residual errors (0.03-0.05 dB) and may dominate the uncertainty budget.