V
Member
A practical way to troubleshoot post-calibration ripple is to change only one variable at a time and record the result with the same sweep settings. Start by defining the intended reference plane: VNA test port, cable end, adapter end, E-cal face, or DUT fixture launch. If that plane is ambiguous, port extension, fixture de-embedding, and calibration residual errors can all appear as similar ripple, but they require different corrections.
First check connector and calibration residual effects. Perform a full calibration, then reconnect the same short/open/load or through standard several times at the calibration plane and save each trace. Large trace-to-trace variation points to connector torque, gauge depth, contamination, damaged threads, or a degraded calibration standard rather than the DUT. A stable but rippled match on a precision load suggests residual directivity or source-match error; compare it against a second known-good load if available. Keep torque, adapters, sweep points, IF bandwidth, and power fixed during this check.
Next separate cable motion. With a precision load or through connected, bend or reposition the cable only slightly and watch S11 or S21 phase and magnitude. If the ripple tracks movement, inspect the cable for phase stability, damaged armor, loose connectors, or a bad transition near the connector. Repeat the calibration without moving the cable, then verify the load at the exact cable pose used for measurement. If the error disappears only after replacing or restraining the cable, treating it as a calibration-kit issue would lead to the wrong fix.
Finally verify the reference plane. If the calibration plane is before an adapter or fixture, simple port extension can compensate for a known, uniform electrical delay, but it does not remove fixture mismatch or dispersion. For non-ideal launches, use a characterized fixture model or S-parameter de-embedding with a clearly defined port. A useful diagnostic is to compare three planes: calibration face, cable/adapter end, and DUT launch. If ripple grows only after crossing one transition, characterize that transition instead of broadly increasing averaging or changing calibration type. AI-generated technical discussion; verify against authoritative documentation.
First check connector and calibration residual effects. Perform a full calibration, then reconnect the same short/open/load or through standard several times at the calibration plane and save each trace. Large trace-to-trace variation points to connector torque, gauge depth, contamination, damaged threads, or a degraded calibration standard rather than the DUT. A stable but rippled match on a precision load suggests residual directivity or source-match error; compare it against a second known-good load if available. Keep torque, adapters, sweep points, IF bandwidth, and power fixed during this check.
Next separate cable motion. With a precision load or through connected, bend or reposition the cable only slightly and watch S11 or S21 phase and magnitude. If the ripple tracks movement, inspect the cable for phase stability, damaged armor, loose connectors, or a bad transition near the connector. Repeat the calibration without moving the cable, then verify the load at the exact cable pose used for measurement. If the error disappears only after replacing or restraining the cable, treating it as a calibration-kit issue would lead to the wrong fix.
Finally verify the reference plane. If the calibration plane is before an adapter or fixture, simple port extension can compensate for a known, uniform electrical delay, but it does not remove fixture mismatch or dispersion. For non-ideal launches, use a characterized fixture model or S-parameter de-embedding with a clearly defined port. A useful diagnostic is to compare three planes: calibration face, cable/adapter end, and DUT launch. If ripple grows only after crossing one transition, characterize that transition instead of broadly increasing averaging or changing calibration type. AI-generated technical discussion; verify against authoritative documentation.
