A gas detector only protects anyone if it responds to gas and alarms at the right level, and there are two very different checks that confirm those two things. A bump test is a quick functional confirmation that the sensor sees gas and the alarm fires. A calibration is a deeper procedure that adjusts the zero and the span so the reading is actually accurate, not merely responsive. Confusing the two is dangerous, because a detector can pass a bump test while reading twenty percent low, which is exactly the error a calibration exists to correct. Knowing when each check is enough, and how to run each correctly with the right gases and flow, is core competence for anyone responsible for fixed or portable gas detection.
Calibrate a Gas Detector in one line: A bump test applies a known gas for a short time to confirm the sensor responds and the alarm activates, without adjusting anything, and it is done frequently as a go or no-go check. A calibration is a full procedure that first sets the zero on clean zero air and then sets the span by applying a certified span gas of known concentration at the correct flow, adjusting the reading to match. A detector that passes a bump test can still be reading inaccurately, which is why span calibration and documented drift tracking are still required on a schedule.
A bump test answers one question: if gas reaches this sensor, will the detector notice and alarm? You apply a gas above the alarm threshold, watch the reading rise, and confirm the alarm activates within an expected response time, then you remove the gas and confirm it clears. You do not adjust the instrument during a bump test; you are checking function, not accuracy. Because it is fast and non-invasive, a bump test is done far more often than a calibration, often daily on portables before entry into a hazardous area, and it is the right check when the question is simply whether the detector is alive and its alarm path works.
A calibration answers a different question: is the reading correct? It has two steps that must happen in order. First you zero the detector on clean air that contains none of the target gas, establishing the bottom of the scale so that a genuinely clean atmosphere reads zero and does not float or alarm falsely. Then you span the detector by applying a certified gas of accurately known concentration and adjusting the reading until it matches that concentration, which sets the slope of the response. Zero then span is the mandatory sequence, because the span adjustment assumes the zero is already correct, and spanning against a drifted zero produces a reading that is wrong across the whole range.
The reason both checks exist is that they catch different failures. A bump test catches a dead sensor, a blocked inlet, a failed alarm relay, or a detector that has drifted so far it no longer responds meaningfully. A calibration catches the slower, quieter failure where the sensor still responds but its sensitivity has drifted, so it reads low or high while still appearing healthy. A common and defensible practice is to bump test frequently and calibrate on a defined interval or whenever a bump test fails, so the fast check guards against sudden failure while the periodic calibration guards against gradual drift. Neither one substitutes for the other.
The span gas has to match the sensor and sit at a sensible point on its scale. For a combustible sensor measuring lower explosive limit, the span gas is typically a defined percentage of LEL of a specified fuel such as methane, and it matters that the gas is the fuel the sensor is set up for, because a sensor calibrated on one fuel responds differently to another. For a toxic sensor such as hydrogen sulfide, the span gas is a certified concentration in the range the alarms live in. The cylinder carries a certificate of analysis stating the exact concentration and its uncertainty, and you calibrate to that stated value, not to the round number printed on the label, because the certified value is what the standard traces back to.
Flow is the detail that quietly ruins calibrations. Many sensors, particularly those that rely on diffusion, need gas delivered at a specific flow rate so the sensor sees a representative sample, and too little or too much flow gives a reading that does not reflect the true response. You deliver the gas through a fixed-flow regulator or a demand-flow regulator matched to the detector, and you apply it through the calibration cap or adapter that seals the gas onto the sensor and directs the whole flow across it. Using the wrong flow, or letting gas leak around a poorly seated cap, produces a span that looks fine on the day and reads wrong in service. Letting the reading fully stabilize before adjusting is equally important, because a sensor still rising has not reached its true response.
The order of operations at the detector is zero first on zero air, then span on the certified gas, then a confirming re-check. Zero air is not just any ambient air; it is air verified to be free of the target gas and of interferents, supplied from a cylinder or a scrubber when the local atmosphere cannot be trusted. Zeroing on contaminated air pushes the whole scale off before you even span. After spanning, applying the gas once more to confirm the detector now reads the certified value, and that the alarms trip at the right points, closes the loop and turns the calibration into a demonstrated result rather than an assumed one.
What you record during calibration is as valuable as the adjustment itself. The most useful number is the as-found reading, meaning what the detector displayed on the certified gas before you adjusted anything, because comparing as-found to the certified value quantifies how far the sensor drifted since the last calibration. A sensor drifting a little each interval is aging normally; a sensor drifting a lot, or drifting faster than it used to, is heading for failure and should be scheduled for replacement before it fails in service. Recording as-found and as-left values at every calibration turns a maintenance chore into a data trail that predicts sensor end of life.
Response time deserves the same attention. During the bump test or calibration you can note how long the detector takes to reach a defined fraction of the applied concentration, and a sensor that is getting slower is losing the sensitivity that makes it useful in a real release. A detector that still reaches the right final reading but takes far longer to get there is a warning that the sensing element or the sample path is degrading. Documenting response time alongside the reading captures a failure mode that a simple pass or fail check misses entirely, and it is often the earliest sign that a sensor is on its way out.
Fixed gas detectors are almost always wired into a control system, which makes their calibration history and health a natural thing to trend rather than to bury in paper. A cloud SCADA platform such as Merobix can hold the detector's live reading, its alarm state, and any fault or maintenance signal the head reports, so an operator sees at a glance which detectors are in alarm, which are in fault, and which are overdue for calibration. Trending the readings also exposes slow zero drift between formal calibrations, and logging each calibration event against the tag builds a per-detector record of as-found drift and response time that supports both the maintenance schedule and any audit of the detection system. For portables, the same discipline applies at the docking station, where bump and calibration results are captured and rolled up rather than trusted to memory.
No. A bump test applies gas briefly to confirm the sensor responds and the alarm fires, but it does not adjust anything, so a detector can pass a bump test while still reading inaccurately. A calibration goes further by zeroing on clean air and then spanning on a certified gas to correct the reading itself. The safe practice is to bump test frequently as a go or no-go check and calibrate on a schedule or whenever a bump test fails.
The span adjustment sets the slope of the sensor's response and assumes the zero is already correct, so spanning against a drifted zero throws off the reading across the whole range. You establish the bottom of the scale first on verified zero air, confirming a clean atmosphere reads zero, and only then apply the certified span gas to set the accurate reading at concentration. Doing it in the reverse order produces a calibration that is wrong everywhere.
Use a certified span gas appropriate to the sensor and sitting in the range where the alarms live, such as a defined percentage of LEL of the correct fuel for a combustible sensor or a certified toxic concentration for a sensor like hydrogen sulfide. Always calibrate to the exact value on the cylinder's certificate of analysis rather than the round number on the label, since the certified value is what the standard traces back to. The span gas fuel must also match what the combustible sensor is configured for.
Safety & engineering notice. This article is general educational information, not site-specific engineering, safety, or legal advice, and it does not reflect any particular facility. Standards and regulations (for example OSHA, API, IEC, ISO, NFPA, NIST, and NERC CIP requirements) change and vary by edition, jurisdiction, and application. SCADA and remote monitoring cannot verify physical isolation, atmosphere, lockout/tagout, permit status, or a safe go/no-go decision. Qualified personnel must perform site-specific engineering, hazard analysis, and safety review, and confirm current requirements with the authority having jurisdiction, before acting.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.