Ask any instrument technician about gas detection and you will hear the same complaint: these are some of the most temperamental devices on site. Detectors drift, fail quietly, and sometimes read perfectly normal while being completely blind to gas. None of that shows up on the design drawing. It shows up two years into operation, usually during an audit or, worse, during a real leak.
This guide covers the failure modes that come up again and again in the field for fixed combustible (LEL) gas detectors, and the practical habits that keep them reliable. It is written for facility, maintenance and MEP engineers who inherit these systems rather than design them from scratch.
Catalytic Bead vs Infrared: Know What You Are Maintaining
Most fixed combustible gas detectors use one of two sensing principles, and they fail in very different ways.
- Catalytic bead (pellistor). Gas burns on a heated bead and the change in resistance is measured. It responds to a wide range of flammable gases, including hydrogen, and it is relatively cheap. But it needs oxygen to work, and certain contaminants can permanently kill it.
- Infrared (NDIR). Gas absorbs infrared light at specific wavelengths. It cannot be poisoned, does not need oxygen, and usually drifts less. The trade-offs are a higher purchase price and one important blind spot: it cannot detect hydrogen, because hydrogen does not absorb infrared.
Before you plan maintenance, check which type is installed at each point. The inspection routine for a catalytic bead head is not the same as for an IR head.
The Silent Failure: Catalytic Bead Poisoning
Poisoning is the failure that worries experienced technicians most, because it gives no warning. Compounds such as silicones, lead compounds and some sulfur compounds coat the bead and stop it from reacting with gas. The detector does not raise a fault. It just reads low, or reads zero, while gas is present.
The sources are often everyday maintenance products rather than process chemicals:
- Silicone lubricant and release sprays
- Silicone sealants cured near the detector
- Some cleaning products and personal-care products that contain silicones
The typical story goes like this: a new spray product arrives on site, operators mention that the detectors “feel slow” during bump tests, and nobody connects the two until several heads are found dead. Once a bead is poisoned by silicone, recalibration will not bring it back. Raising the span to compensate only hides a sensor that is failing.
What to do: keep silicone products away from catalytic detectors, treat a slow bump-test response as an early warning rather than a nuisance, and consider IR heads in areas where silicone exposure cannot be avoided.
Bump Test vs Calibration
These two terms get mixed up constantly, and the difference matters.
- Bump test: expose the detector to a known gas and confirm it responds and alarms. It proves the sensor can still see gas. It adjusts nothing.
- Calibration: set the zero in clean air and adjust the span against a certified gas concentration so the reading is accurate.
A detector can hold perfect calibration on paper and still fail a bump test if it has been poisoned since. Follow the manufacturer’s intervals and your site procedure for both. For flammable gas detectors, IEC 60079-29-2 is the usual reference for selection, installation, use and maintenance.
Mounting Mistakes That Make Good Detectors Useless
- Ignoring gas density. Methane and natural gas are lighter than air, so detectors go high, near the ceiling or roof ridge. Propane, butane and petrol vapour are heavier than air, so detectors go low, near floors, pits and trenches.
- Ignoring air movement. Density is only half the story. Supply air, exhaust fans and open doors can carry a plume straight past a detector that is at the “correct” height. Look at how air actually moves in the room before fixing positions.
- No plan for access. A detector 8 m up under a roof is easy to install once and painful to test every quarter. Where heads are hard to reach, specify remote calibration tubing or a calibration port brought down to working level, so routine checks do not need a scissor lift.
- Wrong orientation and missing guards. Most heads are designed to point downward. Leave off the splash or weather guard, or let the sinter clog with dust, washdown spray or insect nests, and gas simply cannot reach the sensor.
Wiring and Power Problems
- Shield grounded at both ends. For 4–20 mA and similar instrument loops, the common practice is to earth the cable shield at one end only, normally the control panel end. Earthing both ends can create a ground loop that shows up as a wandering zero and random low-level readings.
- Frequent power cycling. Sensors need time to warm up and stabilise after power is restored. Zero or span calibrations done before that point will not pass, or will pass and then drift. Check the manufacturer’s warm-up time and avoid switching detectors off during routine shutdowns unless you have to.
Calibration Pitfalls
- Expired test gas. Calibration gas has a shelf life. A surprising number of “failed” calibrations come down to an out-of-date cylinder. Check the date before you start.
- The wrong gas without a correction factor. A detector calibrated for methane will read differently on propane. If you calibrate with a different gas, use the manufacturer’s response or correction factor.
- Poor adapter fit or wrong flow rate. A loose calibration cup lets ambient air dilute the test gas. The detector then reads low and gets set with too much gain. Use the correct adapter and the flow rate the manufacturer specifies.
- Calibrating a dying sensor. If the span adjustment needed keeps growing, the sensor is reaching end of life. Replace it rather than turning the gain up again.
A Practical Field Habit: Keep a Known-Good Sensor
When a detector misbehaves, the quickest diagnosis is to swap in a sensor cell you know is good. If the fault clears, the old cell was the problem. If it stays, look at the transmitter, the wiring or the panel input. It saves hours of guessing, and it is worth keeping one spare cell per detector model on site.
Simple Maintenance Routine
- Keep a register of every detector: location, gas, sensor type, installation date and sensor replacement date.
- Bump test at the frequency set by the manufacturer and your site procedure, and record the response, not just pass or fail.
- Calibrate at the specified interval, using in-date gas and the correct adapter.
- Inspect guards and sinters for dirt, water and insects at every visit.
- Check that alarms reach the control panel or BMS end to end, not just at the detector.
- Review any head that needed a large span adjustment and plan its replacement.
Gas Detectors vs Liquid Leak Sensors
Combustible gas detectors watch the air for vapour. Liquid hydrocarbon leak sensors, such as sensing cables and point sensors in drip trays and bunds, detect fuel or oil on the floor before it spreads. Many fuel rooms and generator rooms need both. For the liquid side, see our Hydrocarbon Leak Sensor guide.
