Scope
This is a performance reproducibility protocol, not a safety standard or medical statement. One unit is a sample-only result. A model-level mark requires at least three retail units from different lots and a duplicate measurement at an accredited facility.
Record before power-on
Record brand, exact model and SKU, serial, hardware and firmware, purchase channel and date, lot, supply voltage and frequency, modes, dimmer, stated wavelengths, the vendor irradiance quotation with distance, statistic and instrument, active emitter area, case dimensions, label image, front image, and provenance.
Environment and stabilization
Use a matte-black surround and record the dark background. Signal must be at least one hundred times background. Record temperature and humidity; the protocol window is twenty-one to twenty-four degrees Celsius. Trace the center at zero, five, ten, and fifteen minutes. Take three readings spaced sixty seconds apart with spread no greater than two percent. Drift above three percent invalidates the run or enters uncertainty.
Instrument
Record portable spectroradiometer serial, range, calibration statement and date. Keep the manufacturer accuracy statement separate. Record aperture, cosine response, integration time, dark correction, range, saturation and raw spectral power distribution. Fixed integrals are red six hundred to seven hundred nanometers and near-infrared seven hundred eighty to one thousand fifty nanometers. Take three readings per point; above two percent coefficient of variation, take ten.
Distance
Measure from the optical emitting plane to the detector diffuser plane along the normal, within one millimeter. Publish mean, minimum, and maximum. For curved wearables, map local gap on a matte-black headform.
Panels
Measure planes at fifteen and thirty centimeters, the exact vendor distance, and one recorded use distance. Use at least a seven-by-seven grid inside the active rectangle plus four outside points. If adjacent points differ by more than fifteen percent, use an eleven-by-eleven grid. Grid convergence must be within five percent. Report center, arithmetic mean, area-weighted mean, median, minimum, maximum, tenth and ninetieth percentiles, spatial variation and uniformity.
Masks
Fasten the mask on a matte-black anthropometric headform without flattening it. Use at least twelve zones and exclude eye zones from the mean. Record local gap and normal at every zone. The primary quantity is the surface reading in wearing geometry. Repeat mounting and use two headform sizes.
Caps, beds, and blankets
For caps, use a bald headform with recorded circumference and tension and at least sixteen zones. For beds, use a mannequin in operating position and at least a seven-by-five grid per surface. Record the center drift over a full session.
Uncertainty
List instrument contribution, repeatability, distance positioning from measured offsets, spatial sampling, warm-up drift, spectral truncation, unit-to-unit variation and wearable geometry. With one unit, label sample only. Design targets before validation are ten to fifteen percent for a point, twelve to eighteen percent for a panel mean, and fifteen to twenty-five percent for wearables. A portable-to-reference difference above fifteen percent or control drift above five percent blocks a model mark.
Cross-calibration
Repeat Bestqool BQ60 center values at three, six, and twelve inches; reproduce one current MitoPRO report; duplicate one serial at an ISO 17025 facility with a forty-nine-point map at fifteen and thirty centimeters and a sphere total as a separate quantity; record a transfer source before and after the campaign; review calibration annually.
Result card
Publish sample, mode, geometry, environment, instrument, spectrum, spatial coordinates and three raw readings per point, summary results, time trace, uncertainty table, scope, CSV, JSON, and any signed report. Dose arithmetic is shown only when a duration is declared. A vendor irradiance number without distance remains unqualified.