How to Commission a Fixed LEL Gas Detector

Back to Blog

Confirm What You Need Before You Start

Commissioning a combustible point needs the paperwork before the tools. Have the gas-detection layout drawing or mapping study that placed this point, the detector datasheet with its warm-up time and calibration method, the certified span gas cylinder and regulator specified for this sensor, the calibration cap or flow housing that fits the head, and the cause-and-effect matrix that says what this point is supposed to trigger. A combustible point reports in percentage of the lower explosive limit, so you also want to understand the setpoints in front of you - the background on that is in the guide to gas detector alarm setpoints in %LEL.

On the physical side you need a stable regulated flow of span gas at the rate the manufacturer specifies, clean air or a zero-gas source for the baseline, and access to the control system or a second person on the radio who can confirm what the point is doing at the panel. Do not substitute a cylinder of a different gas or concentration for convenience; the sensor is characterized to a specific span gas, and a mismatched cylinder produces a reading that looks like a pass but is not. If you are unclear whether the sensor is catalytic bead or infrared, check the datasheet, because an infrared point cannot see hydrogen and that changes what a valid commissioning gas is.

Verify the Mounting Location Against the Hazard

Before you touch calibration, confirm the point is where the mapping study put it, because a perfectly calibrated detector in the wrong place protects nothing. Combustible placement follows gas physics: a detector for a gas heavier than air belongs low, near where a dense cloud would pool, and a detector for a lighter-than-air gas such as methane belongs high, near the ceiling or roofline where it would rise and collect. Check the mounting height and horizontal position against the drawing, and confirm the point actually faces the credible leak sources it was placed to watch - flanges, seals, compressor packing, or a specific vessel.

Look at the local geometry the drawing cannot fully capture. Confirm nothing has been installed since the study that blocks the sensor from the air it is meant to sample, that the head is not tucked behind a beam or inside dead air where a real cloud would never reach it, and that the calibration cap port is accessible for the technician who will bump it for years to come. If the as-built location differs from the study in any way that matters, that is a finding to raise with the gas-detection designer before commissioning, not a thing to quietly accept. The selection logic behind why combustible and toxic points sit at different heights is covered in the guide to choosing combustible versus toxic gas detection.

Power Up, Zero, and Apply the First Span

Energize the point and let it complete its warm-up before you trust any reading. A catalytic-bead sensor in particular needs its filaments to reach stable operating temperature, and the datasheet gives the stabilization time - reading it too early gives a drifting baseline that you will chase in circles. Confirm the point is communicating with the control system and that the tag is live and correctly labeled with its gas and %LEL scale.

With the point stable, establish the zero in known clean air. If the ambient air at the location is genuinely clean you can zero on it, but where any background hydrocarbon may be present, apply a certified zero gas so the baseline is true. A false zero shifts every subsequent reading. Once zeroed, fit the calibration cap and apply the certified span gas at the specified flow, watch the reading climb, and let it settle. The stabilized value should match the cylinder concentration in %LEL. If it reads low, the sensor may be aging, contaminated, or the flow may be wrong; if it will not respond at all, suspect a poisoned catalytic element or, for an infrared point, the wrong gas entirely. Record the as-found response, adjust the span per the manufacturer's method only if the point supports it, and record the as-left value. The distinction between this span adjustment and a routine field check is spelled out in the procedure for a gas detector bump test versus a full calibration.

Prove the Alarm Chain End to End

A calibrated sensor that does not actually trip anything is a silent failure waiting to happen, so the last commissioning step is to drive the point through its setpoints and confirm every downstream response. Apply enough span gas to push the reading past the low alarm setpoint and confirm the control system annunciates it, then past the high alarm setpoint and confirm the high alarm and any executive action - a ventilation start, a damper, a shutdown, or a beacon and sonalert - actually fires. This is the only time you will deliberately take the point to alarm on purpose, so make it count and confirm each rung of the cause-and-effect matrix.

Where the point contributes to a voted arrangement rather than acting alone, confirm its role in that vote, because a point commissioned as if it stood alone can behave differently inside a two-out-of-three scheme. The logic behind those arrangements is described in the guide to gas detector voting and coincidence. Coordinate every trip test with the control room and any affected operations first, since a real shutdown or ventilation start has consequences beyond the detector you are testing.

Verify the Result and Return to Service

Commissioning is complete only when the record proves it. You should end with a documented clean-air zero, an as-found and as-left span response that matches the certified gas, and a confirmed alarm chain showing the low alarm, high alarm, and each executive action operated. Restore the point to its normal state, confirm any bypass or inhibit you used for testing is removed, and verify the live tag reads a plausible background value at the panel.

The commissioning record becomes the baseline the point is judged against for the rest of its life. When the point's %LEL value and its health and fault status are trended in a monitoring platform such as Merobix, a slow decline in span response at the next bump, or a sensor that starts reporting fault at a remote site, shows up against that commissioning baseline rather than being discovered at an annual visit. The first span you recorded is the number every later check is compared to.

Frequently Asked Questions