How to Read a Red Light Panel Irradiance Test Report (And Spot a Weak One)

RedVance builds red light panels, so this is written from the side that produces these documents. The framing throughout is deliberately generic: report formats vary between labs and factories, and the point is to give you a reading method that works on any of them, not to describe one house style.

Strip away formatting and a credible report contains four blocks. If one is missing entirely, that absence is the finding:

Notice that only the last block contains the number everyone looks at. The first three decide whether that number means anything at all.

The instrument, and its calibration date

This is the single most skipped line in the entire document, and the most revealing. Optical meters drift with time and use, which is why they are periodically recalibrated against a traceable reference. An instrument named without a calibration date leaves every number in the report unverifiable, because a drifted meter produces readings that look perfectly consistent and are simply wrong.

Instrument type matters just as much. A spectrometer or spectroradiometer resolves output by wavelength and is calibrated for the narrow bands red and near-infrared LEDs emit. A solar power meter is not, and on a dual-wavelength panel it can read both channels together as one inflated figure. The mechanics of that particular distortion are in our irradiance claims guide.

Does the report name your device?

A report describes the unit that was on the bench. Check the model number against the product you are buying, and look for a serial or batch identifier. A report for a different model in the same range covers nothing, which is the same gap that appears constantly with compliance paperwork, as covered in our certifications guide.

Then ask the follow-up that most buyers never reach: does this report correspond to current production, or to a sample from an earlier batch? A report dated two years before your order describes hardware that may have changed components since.

• Distance. Irradiance falls sharply with distance, so a figure without a stated distance is not a measurement. Look for a realistic treatment distance such as 15cm (6 in), ideally with a second figure at 30cm (12 in) so you can see the fall-off.
• Thermal state. LED output declines as the panel heats. A reading taken at switch-on reads higher than the same panel after a full session. A credible report states the warm-up time, and the reasoning behind that is in our thermal drift guide.
• Ambient temperature. Rarely stated, and it matters: a panel measured in a cool lab behaves differently from one running in a warm treatment room.
• Active channels. The subtle one, explained below.

Output is never uniform across a panel, so a single number always hides a distribution. Three ways to report the same measurement session:

• Peak. The brightest single point, almost always the centre. The most flattering and least representative figure.
• Average. Mean across sampled points, closer to what a body actually receives.
• Map. A grid of readings showing how output varies across the treatment area.

A grid looks something like this, in mW/cm²:

Read a map for its spread, not its highest value. A tight cluster indicates even coverage. A grid where the centre towers over the corners describes a hotspot, and the uneven “polka dot” pattern that produces is covered in our coverage guide. A report giving only a peak figure, with no average and no map, is telling you about one square centimetre of a panel you intend to treat your whole body with.

1. No instrument named. Nothing else in the document can be assessed.
2. No calibration date. The instrument may be accurate; you cannot know.
3. No distance stated. The headline figure is unanchored.
4. No thermal state or warm-up time. Suggests a switch-on reading.
5. Peak only, no average or map. One point standing in for a whole panel.
6. Model or serial does not match your device. The report covers something else.
7. Solar meter used on a multi-wavelength panel. Wrong instrument class for the job.
8. Lab branding without an identifiable lab. A logo is not accreditation; check the lab exists and what it is accredited for.

None of these individually proves bad faith. A supplier may simply have a thin internal document rather than something to hide. What matters is the response when you ask: a factory that measures its own output can usually produce the missing details on request, which is exactly the engineering-depth test in our factory audit checklist.

Even a rigorous report has boundaries, and knowing them prevents over-trusting a good document:

• Unit-to-unit variation. One tested unit is not every unit. Ask what tolerance applies across production and whether each unit gets an output check before shipment.
• Performance over time. A report is a snapshot at one date. Output declines with operating hours, per our L70 and LM-80 guide.
• Wavelength accuracy. Irradiance and spectral output are different measurements. Peak wavelengths and tolerances live in a spectrum report, covered in our spectrum report guide.
• Everything non-optical. EMF, flicker, electrical safety and durability sit outside this document entirely.
• Whether the panel suits your use. A high figure at 15cm is irrelevant if your application needs coverage at 30cm.

Reading reports is one part of the wider verification discipline in our spec validation guide and the category-wide pattern described in our spec trust gap overview. Definitions for every term used here are in our glossary, and checking a physical sample against its paperwork is in our sample evaluation guide.

What a manufacturer can do about it: state the instrument and its calibration date, identify the specific model and unit tested, give the measurement distance in both 15cm (6 in) and 30cm (12 in), state the warm-up period and ambient temperature, confirm all channels were running simultaneously, and report peak alongside average with a distribution grid rather than a single flattering figure. None of this is difficult for a factory that measures its own panels. The completeness of the document is itself a signal.