Amplifier Test Uncertainty And Reporting
Definition: the bounded doubt in a result caused by the test method, supply, load, instruments, probes, calibration, bandwidth, phase, timing, thermal drift, repeatability, processing and rounding.
Resolution Is Not Accuracy
A display may show tenths of a watt while the load resistance, probe ratio, current phase, temperature and run-to-run variation support only two or three significant figures. Preserve raw resolution, but report precision the complete chain can defend.
Typical Contributors
- terminal supply voltage accuracy and ripple
- load resistance, temperature coefficient and connection resistance
- voltage-probe gain, loading, common-mode rejection and bandwidth
- current-sensor ratio, offset, saturation, phase and insertion effect
- channel deskew and sampling synchronization
- filter, RMS, FFT, integration and endpoint method
- starting temperature, cooling and repeatability
Phase Matters
Small voltage-current timing errors can materially change real-power calculations on reactive loads. Include deskew and phase uncertainty rather than treating probe timing as exact.
Repeatability And Reproducibility
Repeatability describes repeated runs under the same setup. Reproducibility asks whether another bench, operator, instrument chain or day can obtain a compatible result. Report the number of runs, spread and all changed conditions.
Pass/Fail Near A Boundary
When a requirement is close to the measured result, use a guard band or explicitly state that the measurement cannot confidently resolve the decision. Do not pass a claim merely because the rounded display is on the preferred side of the limit.
Minimum Evidence Package
- test plan and amplifier revision
- bench diagram and photos
- instrument and calibration records
- raw voltage, current, distortion, temperature and supply files
- processing and calculation settings
- run table, variation, endpoint and first limit
- uncertainty or limitation statement
- final report and stable evidence identifiers
Comparison Rule
Compare two amplifiers or two revisions only when the load, supply, channel mode, waveform, duration, endpoint, cooling and measurement method are equivalent enough for the question. Otherwise describe the difference without ranking the watt numbers directly.