Thermal-Imaging Calibration Drift in Student-Built Electronics Cooling Benches for Engineering Education: A Thermocouple-Referenced Evaluation Linking Emissivity Misassignment to Junction-Temperature Safety Margins

Student-built electronics cooling benches are common in undergraduate engineering laboratories, yet the thermal cameras used to characterise them are rarely recalibrated against contact sensors, and the emissivity values entered into their firmware are seldom audited. The framework is calibrated against a public reanalysis record for Lagos, Nigeria (NASA POWER MERRA-2, 2000–2023, 312 monthly observations), which yields a mean air temperature of 26.414 °C and a warming trend of 0.1553 °C per decade, establishing the ambient boundary condition against which bench measurements are interpreted. A formal error-propagation model is derived, showing that a fixed emissivity offset produces a multiplicative bias in indicated temperature that scales with the fourth power of absolute temperature and inversely with the assumed emissivity. The model is used to construct a safety-margin map that identifies the combinations of emissivity error and junction-to-ambient rise at which a nominal 20 °C margin is exhausted. The analysis indicates that plausible student-entered emissivities (0.85–0.95) on low-emissivity surfaces can consume between one-fifth and one-half of a typical safety margin before any thermocouple discrepancy is considered. The contribution is a transferable calibration protocol and a quantitative bound on the conditions under which thermal-imaging results from student benches can be trusted for junction-temperature assessment.