CO-Pocket-2027 & Temperature Calibration

HOW IT WORKS  /  NOVEMBER 2026

By James Wiebe, CEO, Radiant Instruments

2 November 2026

A pilot walks out of a freezing hangar, climbs into a sunlit cabin, and later turns on the heater. Your carbon monoxide monitor is the quiet passenger through all of it, and through every degree the number on its screen has to mean what it says. Temperature calibration is how we make sure it does.

Where this started

In our early design research, we put our sensors in freezers, refrigerators and a heated test chamber. The cool and cold were quiet: the detector stayed steady. Heat was a different story. In the chamber, every unit’s output rose with the temperature, and not in a straight line. With no CO anywhere near them, the units began to report CO anyway. We logged the chamber against each unit’s own internal thermometer, over many units and many tests. Sensors from the same batch followed different curves: one moved slowly, another climbed hard. We re-calibrated the units that bothered us most and ran them again, and they traced almost the same curve. Minutes after leaving the chamber, they returned to baseline, every time.

This heat response is a known behavior of this kind of sensor, not unique to ours, as the next section shows. It was not random and it was not damage. It was repeatable and individual, a thermal fingerprint, and anything you can measure you can calibrate. Every unit now gets its own temperature calibration on top of its gas calibration, and each finished unit is checked hot before it ships. CO-Pocket-2027 carries that lesson.

What others say

We also read what other manufacturers tell their customers. Published manuals ask users to start in fresh air at room temperature. Some detectors show zero until a reading passes several ppm. Some warn that small readings can appear after a sudden temperature change while compensation catches up. Others state that they compensate for temperature and calibrate each unit with gas. None of that is wrong; it is the same physics everyone faces. It told us the problem is universal, and it showed us what we wanted to do differently: measure each instrument’s behavior instead of raising thresholds until the detector goes quiet.

Hot, cool, and below freezing

Cold was benign, but we tested it thoroughly anyway, because cold is when cabin heaters run, and exhaust and cabin-heat systems are the usual CO sources in aircraft. A detector has to be steady exactly when it matters most. CO-Pocket-2027 is specified for −5°C to 42°C (23°F to 108°F), and one flight can cross much of that range: a sun-soaked ramp, then a cold cruise. Because the sensor moves with heat, that is where the calibration effort goes.

How every CO-Pocket-2027 is calibrated and verified

1. Gas. Each unit is zeroed in clean air and then calibrated against known 100 ppm CO.

2. Temperature. Each unit is warmed in a chamber to 106.5°F (41.4°C) while it records how its own sensor’s baseline moves. It keeps a six-point curve: its room-temperature baseline, points along the heating path, and its fully warmed baseline.

3. Settled, then stored. The warm end is recorded only after both the unit’s temperature and its sensor output have held steady for five minutes. The curve is then written into the unit.

4. Finished-product verification. Every assembled instrument then gets a separate final hot test. A unit that does not meet our acceptance criterion goes back for temperature recalibration and is tested again.

In use, the instrument reads its own internal temperature every second and applies its own curve as it warms and cools. The DIE temperature at the bottom of the screen is that same internal temperature.

What we learned

A steady temperature is not a steady sensor. A sensor’s baseline can keep moving after the temperature stops changing, so the factory waits for both.

Every sensor has its own thermal fingerprint. That is why each curve is stored in its own unit, not shared across the product.

Even the thermometer needs calibrating. The temperature chip inside each unit has its own offset; units on the same bench can report temperatures a few degrees apart. So each unit’s temperature scale is anchored to a known bench temperature and the chamber, not taken on trust.

Calibrate the transition, not just the endpoints. A unit that is right at room temperature and right when warm can still be wrong in between, so the curve follows the heating path.

Using it well

Start in clean air and let the three-minute warm-up finish. During warm-up the unit sets its baseline, absorbing small sensor drift of up to 20 ppm. That also means CO present at startup can be absorbed into the baseline, which is why clean air matters.

Do not restart to clear a reading. A restart takes a new baseline from whatever air the unit is in. Treat an unexpected reading as real: ventilate and follow your aircraft checklist. A detector supplements inspection and maintenance and never replaces them.

Conclusion

A CO monitor earns trust by being right when conditions change, not only when they are easy. That is why every CO-Pocket-2027 is calibrated for temperature one unit at a time, and why every finished unit is checked hot before it ships.

The customer experiences the transition—not just the two calibration points. That is what we calibrate and test for.