Distinguish the impedance being discussed
Characteristic impedance describes the relationship between traveling voltage and current waves on a transmission line. It is different from the DC resistance of its copper trace and from the input impedance of a complete loaded circuit. A resistance measurement across a trace does not verify its characteristic impedance.
For differential structures, specify whether the requirement is differential, single-ended or another defined mode. State the target and tolerance using the terminology expected by the design and manufacturing teams. Do not assume that a familiar nominal value applies to every interface.
Collect the geometry and material inputs
Identify microstrip, stripline or coplanar construction, the reference conductors and their continuity. Record finished width, spacing, copper thickness, dielectric thicknesses and the material-property assumptions. Solder mask and nearby copper may be relevant to the model.
Analytical calculators can help compare initial options, but their formulas have geometry and material assumptions. A solver can model a more specific cross-section when configured correctly. Neither tool removes the need to check units, input definitions and the manufactured geometry.
Review manufacturing variation and adjustments
Ask the fabricator to review the proposed construction before freezing the routing. Finished copper geometry can differ from nominal artwork. Agree which adjustments may be made to reach the target and which require approval because of pad, spacing or coupling constraints.
Retain the proposed and approved values in a stackup or impedance table. A change of laminate grade, resin content, dielectric thickness or copper option can change the calculation. Review the effect rather than reusing a calculator result from an earlier revision.
Specify an acceptance method
Where testing is required, define coupon structures, test method, tolerance and reporting. Time-domain reflectometry and frequency-domain measurements answer related but different questions and need appropriate fixtures and interpretation. An isolated number without the test basis is weak acceptance evidence.
Use the measured construction and model to investigate discrepancies. Coupon results do not automatically qualify connectors, vias or other transitions in the product channel. Keep fabrication impedance acceptance separate from the complete interface or system validation plan.