Mathematical Correction
Mathematical procedures remove the parasitic effects of test fixtures and cables from measured data to isolate the performance of the device under test. Standardized de-embedding protocols allow engineers to obtain accurate S-parameters of a board trace or component. This process stops at the reference plane defined by the user during the calibration phase.
Fixture Characterization
Precision measurement requires a known model of the launch structures and test leads. Calibration standards such as thru and reflect are measured to build an error matrix. This matrix is then mathematically subtracted from the raw measurement of the total system.
Accuracy depends on the repeatability of the probe contact and the stability of the test environment. Small variations in how the probe touches the pad can introduce errors that the math cannot fix. Proper fixture design minimizes these errors before the software correction is even applied.
Reference Extension
The location of the measurement plane determines the validity of the final data. Moving the plane into the substrate allows for the evaluation of internal traces without the influence of connectors. This shift relies on the assumption of a uniform transmission line environment.
Properly executed corrections provide a clear view of the inherent electrical behavior of the circuit.
