Every four-gas instrument displays its channels in the same order: oxygen first, then combustible gas as %LEL, then hydrogen sulphide, then carbon monoxide. Operators learn this as a sequence to recite. Fewer know why the sequence runs that way, and the reason matters because one of the four readings is only trustworthy once another has been checked.
Ordered by time to incapacitation
The sequence reflects how quickly each hazard removes a person’s ability to escape.
Oxygen kills in minutes and does so without warning. At 10% by volume, judgement is impaired before the person notices anything wrong. At 6%, collapse is immediate. There is no smell, no irritation, no gradual onset to react to.
Combustible gas at alarm concentration is an explosion risk rather than a physiological one, and the transition from safe to ignitable is instantaneous.
Hydrogen sulphide is the fastest-acting toxic gas in the set. Above roughly 100 ppm it paralyses the olfactory nerve within a breath or two, so the characteristic rotten-egg smell disappears at exactly the concentration where it would have warned you. Loss of consciousness can follow in a single breath at higher concentrations.
Carbon monoxide binds to haemoglobin and accumulates. It is lethal, and it takes longer to do it. A person at 200 ppm has a headache and a few hours.
Oxygen and LEL come first because the consequence is immediate. H2S comes before CO for the same reason it is feared: the knockdown is fast enough that there is no second chance to read the instrument.
The dependency that catches people out
There is a technical reason layered on top of the physiological one, and it is the part most often missed.
A catalytic bead sensor measures combustible gas by burning it. Burning requires oxygen. In an atmosphere below roughly 10% oxygen, the bead cannot combust the sample properly and the %LEL reading falls toward zero.
An operator who tests LEL first in an oxygen-deficient sump sees a low combustible reading and concludes the atmosphere is safe. It is not: it is oxygen-deficient, and the combustible reading is meaningless because the sensor had nothing to burn with.
Reading oxygen first establishes whether the LEL channel is operating in a valid regime. That is why the sequence is not merely a convention about danger, it is a dependency.
Electrochemical toxic sensors are less affected but not immune: very low oxygen shifts their baseline, and readings taken in that condition should be treated as indicative at best.
Three heights, not one
Stratification is the second thing the sequence does not tell you.
Methane and hydrogen are lighter than air and collect at the top of a space. Hydrogen sulphide, carbon dioxide and most solvent vapours are heavier and collect at the bottom. A single reading taken at breathing height can miss either.
The practical routine is to sample the top, the middle and the bottom of the space before entry, and to sample again after any ventilation, because moving air redistributes what was stratified.
Waiting for the reading, not just taking it
A sensor takes time to respond. The datasheet figure is T90, the time to reach 90% of a step change, typically between 20 and 120 seconds depending on the cell.
With a pumped instrument, add the transport time through the sample line. A 10 metre hose at a typical pump rate adds 30 to 60 seconds before the gas even reaches the sensor. Readings taken 5 seconds after starting the pump are showing the atmosphere in the hose, not the atmosphere in the space.
The routine that works: run the pump for the full transport time plus the T90 time, at each height, before trusting the number.
What the instrument cannot tell you
A four-gas instrument covers four hazards. It does not cover the fifth.
A space that previously held sewage sludge may also carry ammonia. A tank that held solvent may carry a VOC at concentrations the LEL channel reports as a small percentage. A space purged with nitrogen has an oxygen problem and nothing else, and the other three channels will read zero and look reassuring.
The gas list comes from the space’s history, not from the instrument’s channel count. Before entry, someone should be able to say what was in the space and what it might have become.
A sequence worth keeping
To summarise the practical routine:
- Test from outside the space, before anyone enters.
- Read oxygen first. If it is outside 19.5 to 23.5%, stop: the LEL reading that follows will not be reliable.
- Test at the top, middle and bottom.
- Allow the full pump transport time plus T90 at each height.
- Ventilate if needed, then test again, since ventilation changes the distribution.
- Re-test on entry and at intervals, because a space that was safe an hour ago is not a fixed condition.
Response times, ranges and pump flow rates differ between instruments. Confirm against the datasheet for the model in use; this page describes the testing logic, not a specific instrument’s performance.