A correctly sized combustible gas detector starts with three numbers: the target gas's Lower Explosive Limit in %vol, the alarm setpoint expressed in %LEL (commonly 10–25% LEL), and the gas's relative vapor density versus air, which dictates vertical mounting height [S1].
Selection is then driven by the sensing principle that survives the environment: catalytic bead for general hydrocarbon service, NDIR for hydrocarbon streams that poison catalysts, and semiconductor or electrochemical only where the duty is narrow and cost-driven [S1][S5].
LEL, UEL, and the %LEL Alarm Language
Combustible gases ignite only inside a flammable range bracketed by the Lower Explosive Limit and Upper Explosive Limit; the LEL figures most often used for methane (~5% vol), propane (~2.1% vol), and hydrogen (~4% vol) are the reference points against which detector output is normalized as %LEL, a unit that lets one controller scale methane, propane, and ethylene on the same input [S1].
Because ignition is possible anywhere between LEL and UEL, alarm thresholds are set well below the LEL: 10% LEL is a common low alarm and 20–25% LEL a common high alarm in many industrial code frameworks, with the higher trip reserved for evacuation or shutdown logic [S1]. On hydrogen, the UEL is unusually wide (~75% vol) and the gas is roughly 14 times lighter than air, so detectors must be ceiling- or roof-mounted rather than placed at breathing zone [S1].
Matching Sensor Technology to the Gas and Environment
Catalytic bead (pellistor) sensors oxidize gas on a heated catalyst and read the heat of combustion as a resistance change; they are the workhorse for general %LEL measurement of hydrocarbons in air but require oxygen, draw more power than semiconductor sensors, and can be permanently damaged by silicone compounds, lead, and certain sulfur-bearing chemicals [S1][S5].
NDIR (non-dispersive infrared) sensors measure hydrocarbon absorption at a gas-specific IR wavelength, so they are immune to the catalyst poisons that disable pellistors, run for years without recalibration, and are the preferred choice for methane, propane, and butane in wet, dirty, or oxygen-deficient streams where a catalytic bead would starve or foul [S1][S4]. Semiconductor (metal-oxide) sensors detect combustible gases by resistance change on a heated oxide; they are low cost, robust, and used widely in residential and light-commercial methane alarms, but they are cross-sensitive to humidity, temperature, and other reducing gases, so they should not be specified where selectivity or ppm-level toxic monitoring is required [S3][S5]. Electrochemical and PID sensors address adjacent duties (toxic CO/H2S, VOC vapor) and are not the primary pick for %LEL combustible service, though they are common in the multi-gas carts that workers carry into confined spaces [S4][S5].
Placement and Mounting Height by Gas Density
Because leaked gas migrates, detector height must track vapor density, not convenience: methane (relative density ~0.55, lighter than air) mounts near the ceiling; propane (~1.5, heavier) and butane (~2.0) mount within ~30 cm of the floor or in low pockets; hydrogen (~0.07) goes to the highest point in the room, often near roof decks or exhaust stacks; ethylene (~0.97) is close to air and is best placed at breathing zone with an airflow study [S1].
The same density logic also explains why so-called "two-in-one" alarms marketed for both natural gas and propane routinely fail in the field: a single sensor cannot sit simultaneously at the ceiling and in a floor pit, so a dual-gas residential unit either under-detects propane (mounted high) or under-detects methane (mounted low) [S2].
Fixed vs. Portable, and Single-Gas vs. Multi-Gas
Fixed detectors are hardwired (often 24 VDC with battery backup) and tied into ventilation, shutoff, and fire-panel logic; portable units run on rechargeable or disposable cells and serve workers entering confined spaces or doing leak surveys [S3][S4].
Single-gas units are simpler, lighter, and cheaper, and they make sense where the hazard is well-defined (an LPG cylinder store, a CO boiler room, an H2S pump skid); multi-gas carts typically combine LEL, O2, CO, and H2S in one body and are the standard for utility crews, refinery turnaround teams, and anyone entering a Permit-Required Confined Space, because the four-gas combination is the historical minimum mandated for entry in most U.S. and EU confined-space rulesets [S3][S5]. A typical four-gas portable runs 12–24 hours on a charge, weighs 200–500 g, and is paired with a calibration cap and a calibration-gas cylinder of known concentration (commonly 50% LEL methane, 25 ppm H2S, 100 ppm CO, 20.9% O2) [S5].
Selection Criteria, Certifications, and Common Specs
For industrial fixed units, the spec sheet should lock down: target gas and range (0–100% LEL is the most common, with 0–10,000 ppm LEL or 0–100% vol options for high-concentration or process measurement), output (4–20 mA analog, HART, Modbus RTU, or Foundation Fieldbus), response time T90, ingress protection (IP66/IP67 typical for wash-down areas), and hazardous-area approval (ATEX 2014/34/EU category 2/3 or IECEx for Zone 1/Zone 2), plus operating temperature typically -40 °C to +70 °C for outdoor chemical or refinery service [S3][S4].
For residential methane or LPG alarms, look for a minimum 85 dB audible output at 1 m, end-of-life signaling, and a listed certification (UL 1484 for residential gas detectors in the U.S., EN 50194 in the EU for domestic combustible gas detectors), and avoid units that bundle methane and CO on a single sensor element, since the two chemistries do not share a sensing mechanism [S2][S3].
Comparison: Sensor Technologies Against 4 Decision Criteria
Across the four technologies most often quoted on combustible-gas datasheets, the trade-off for a specifier is cost versus selectivity versus poisoning resistance versus oxygen dependence. Catalytic bead wins on cost and is the default for general hydrocarbon %LEL service; NDIR wins on poisoning resistance and zero oxygen dependence; semiconductor wins on upfront price and is dominant in residential methane alarms; electrochemical falls out of the comparison for %LEL because it targets ppm-level toxic gases rather than flammable concentrations [S1][S3][S4][S5].
Failure Modes and What Not to Specify
Do not specify a catalytic-bead sensor in a silicone-rich environment (silicone sealants, some lubricants) or where H2S, leaded gasoline vapors, or phosphate esters are present: the bead can be permanently poisoned and will read low or zero, which is the worst possible failure mode for a safety instrument [S1]. Do not specify a semiconductor sensor in a cold-storage room below ~0 °C or in a process where the target gas is not methane; cross-sensitivity to alcohols, solvents, and humidity can cause nuisance trips [S3]. Do not mount a propane detector at ceiling height, do not mount a methane detector at floor level, and do not rely on a "two-in-one" residential alarm to cover both gases in a single enclosure [S1][S2].
Specifiers should also budget for bump testing on each shift and full calibration on a documented interval (90 days is a common industrial default), since no LEL sensor holds accuracy over years without a reference-gas check, and the standard for calibration-gas concentration on LEL channels is typically 50% of full-scale LEL in air [S4][S5].
For a related breakdown of how these sensors get packaged into multi-gas carts and fixed controllers, see the Multi-Gas Detector Sizing map; for a closer look at sensor-vs.-mounting trade-offs, the Combustible Gas Detector Selection note extends the criteria above with form-factor guidance, and the 2026 sourcing map lines vendors against the same duty classes. The standard reference language for these instruments sits in the combustible gas detector encyclopedia entry.
Spec-level background on the components involved: linear guide, and crossed roller guide.