Calibration is the least glamorous part of owning a breath analyzer and the one that decides whether the device is an asset or a liability. An uncalibrated instrument does not stop working or display an error — it quietly returns numbers that are wrong, with exactly the same confidence as when it was right. Here is what calibration involves, how often it is needed, and what it costs you to skip it.
Calibration versus an accuracy check
These two terms get used interchangeably and mean different things.
An accuracy check — sometimes called a verification — exposes the device to a standard of known alcohol concentration and compares the reading against the expected value. It tells you whether the instrument is still within tolerance. Nothing is altered.
A calibration is the adjustment itself: the device's response is corrected so its output matches the known standard again. Every calibration should begin and end with a check, so you have a documented before-and-after.
For a compliance programme, the check is what you schedule frequently and the calibration is what the check triggers. Many organisations run a monthly or quarterly verification and only recalibrate when a device drifts toward the edge of tolerance or hits its scheduled interval.
Why sensors drift
No sensor is stable forever, and the two common technologies age differently.
Semiconductor sensors rely on a heated metal-oxide element whose surface chemistry changes with use. The heater ages, the surface accumulates contamination, and ambient temperature and humidity influence the response. Drift is comparatively fast, which is why semiconductor devices typically carry recalibration intervals of around 3 to 6 months.
Fuel cell sensors lose electrode activity gradually as they oxidise alcohol, so their drift is tied to the number of tests performed as much as to elapsed time. The decline is slow and predictable, and recalibration intervals are correspondingly longer — commonly 6 to 12 months depending on the model and how hard it is worked.
Heavy use compresses both timelines. A device screening several hundred people a week at a factory gate will need attention sooner than the same model used a few times a month.
Typical intervals — and why the manufacturer's figure wins
A widely used rule of thumb for professional devices is every 6 months or after a set number of tests, whichever comes first. Treat that as a starting point, not gospel. The interval that matters is the one in your specific device's documentation, because it reflects that model's sensor, its expected duty cycle and the tolerance it is specified to hold.
Set a calendar reminder against each device the day it enters service, and track test counts if your model reports them. Programmes fail here far more often than they fail at the sensor.
How the adjustment is performed
Calibration requires a reference sample of known ethanol concentration. Two approaches are standard.
A wet bath simulator bubbles air through a solution of known alcohol concentration held at a controlled temperature — usually body temperature — producing a vapour sample that mimics human breath. It is accurate and long established, and it requires careful temperature control to be meaningful.
A dry gas standard supplies a certified ethanol-in-nitrogen mixture from a pressurised cylinder through a regulator. It is faster, more portable and needs no temperature bath, which makes it the practical choice for field and multi-site programmes. Cylinders have expiry dates and must be within them to be valid.
Whichever is used, the device is allowed to warm up and stabilise, the standard is introduced, and the instrument's output is adjusted to match. The reference standard itself must be traceable to a recognised measurement authority — an uncertified gas mixture calibrates nothing, it just moves your error somewhere else.
Documentation is half the value
For enforcement, transport compliance and workplace disciplinary use, the calibration record is not paperwork — it is the foundation the result stands on. A challenge to a positive test will go after the instrument's history before it goes after anything else.
Keep, for every device: its serial number, the date of each check and calibration, the concentration and lot or certificate number of the standard used, the readings before and after adjustment, and who performed the work. Store it where you can produce it months later without a search.
What skipping it actually costs
A drifted analyzer fails in both directions, and both are expensive.
Reading high, it flags sober people. Every false positive costs a confirmatory test, a delayed shift and, if you act on the screening result alone, a grievance you will lose.
Reading low, it clears people who should have been stopped. This is the failure nobody notices until an incident happens, and it removes the entire reason the programme exists.
Either way, the moment a calibration lapse is discovered, every result that device produced since its last valid calibration becomes arguable. One missed interval can undermine months of records.
Building it into the programme
Treat calibration as a scheduled operation with a named owner, not as maintenance that happens when someone remembers. Assign each device to a person, log every check, keep at least one spare device in rotation so a unit going out for calibration never leaves a gate uncovered, and retire sensors that no longer hold adjustment rather than nursing them along.
AlcoBreath supports calibration for the devices we supply, including guidance on the correct interval and standard for each model. If you are running a multi-device programme and want help setting up a rotation and record format that will survive an audit, get in touch and we will work through it with you.