The target gas reacts at an electrode in an electrolyte and produces a small current proportional to concentration.

Cells are specific to a gas or a small group of gases.

A carbon monoxide cell and a hydrogen sulphide cell are different products rather than different settings.

The output is linear enough over the working range to be read directly.

Response is fast for most target gases but not instantaneous.

Temperature and humidity affect the response and are compensated for in the instrument.

They are chemically consumed by the reactions they measure and by ambient exposure, whether or not gas is ever detected.

Life is stated in months or years by the manufacturer.

A cell at the end of its life may still respond, but not accurately and not quickly.

That degradation is invisible without testing.

Calibration with a known test gas is therefore a recurring duty rather than an installation task.

Budgeting for cell replacement is part of owning a fixed system rather than an unexpected cost.

Deficiency kills without warning; enrichment dramatically increases fire risk.

Oxygen deficiency in confined spaces, cellars and tanks gives no sensory cue at all.

Inert gas systems, nitrogen supplies and fermentation all displace oxygen.

Enrichment from a leaking oxygen supply makes materials burn that normally would not.

Both alarms are set on the same instrument, in opposite directions.

Oxygen monitoring is often the first detection installed in a confined space programme.

By gas density and airflow rather than by servicing convenience.

Gases heavier than air collect at low level, in pits and in basements.

Lighter gases collect at high level and under roof structures.

Ventilation moves everything and can carry gas away from a detector entirely.

A detector positioned at head height for ease of servicing may never see the gas it was installed for.

Placement is a design decision informed by the release scenarios, not a mounting preference.

An electrochemical cell responding to gases other than its target, to varying degrees.

Manufacturers publish cross-sensitivity tables for each cell type.

An unexplained reading is often a real response to the wrong gas.

That can produce false alarms or, more seriously, mask a genuine reading.

Knowing what else is present in the atmosphere is part of specifying the cell.

Where interference is likely, a different detection technology may be more appropriate.

On a schedule set by the manufacturer and the local regime - and it is the duty most often allowed to slip.

Calibration applies a known test gas and confirms the reading and the alarms.

A bump test confirms response without full recalibration and is done more frequently.

Records of calibration are normally required and are what an inspector asks for.

Test gas cylinders themselves have expiry dates that are frequently overlooked.

An uncalibrated detector provides false assurance, which is worse than no detector at all.

A well-designed system reports the fault rather than reading zero and looking healthy.

Fault reporting distinguishes a failed sensor from a clean atmosphere.

That distinction is essential, because both would otherwise show no gas.

Controllers indicate sensor fault separately from alarm conditions.

End-of-life warnings are provided by many instruments ahead of actual failure.

A system without fault reporting cannot be relied upon between calibrations.