A gas detector is only as trustworthy as its last calibration. Over time every sensor drifts, so its zero point and its sensitivity slowly wander away from truth, and calibration is how they are pulled back. A full zero and span calibration sets both ends of the measurement: the baseline in clean gas and the reading against a known concentration. This guide walks through the procedure, contrasts it with a quick bump test, and covers the flow, timing, and record-keeping that make the result reliable.
Zero and span calibration in one line: To zero and span calibrate a gas detector, you first apply zero gas, clean air or nitrogen, to set the baseline reading to zero, then apply a certified span gas of a known concentration and adjust the detector until it reads that value, correcting for drift in sensitivity. This full calibration differs from a bump test, which only confirms the detector responds without adjusting it. Calibration frequency, gas flow rate, and settling time all affect accuracy.
Calibration has two steps because a detector's reading depends on two things: where its zero sits and how steep its response is. The zero step comes first. Clean gas known to be free of the target, either instrument air or nitrogen depending on the sensor, is applied to the detector so that its reading can be set to true zero. This removes any baseline drift, the slow offset that would otherwise make the detector read a small false value in clean air or, worse, hide a genuine low concentration.
The span step sets the gain. A certified span gas, a cylinder holding a precisely known concentration of the target gas, is applied to the detector, and its reading is adjusted until it matches the cylinder's stated value. This corrects sensitivity drift, the tendency for a sensor to respond more weakly or strongly than it should as it ages. With the zero fixing the low end and the span fixing a known upper point, the detector's whole scale is anchored to real, traceable values.
The two steps must be done in that order, because the span adjustment assumes a correct zero. Spanning a detector whose baseline is off would bake the zero error into the gain calculation. Doing zero first, then span, gives a clean two-point correction. Some workflows repeat the pair or verify the zero again after spanning to confirm the adjustments held.
A bump test and a full calibration are often confused but serve different purposes. A bump test is a quick functional check: a known gas is briefly applied and the technician confirms the detector responds and alarms, a simple pass or fail. It answers the question is this detector still working, and it is fast enough to do frequently. Crucially, a bump test does not adjust the detector; it only verifies it reacts.
A full zero and span calibration goes further by actually correcting the reading. It measures how far the detector has drifted and adjusts the zero and the gain so that its output once again matches known values across the scale. This is more involved, needs both zero gas and certified span gas, and is done less often than bump testing. The usual practice is to bump-test regularly to catch failures early and to calibrate on a longer interval, or whenever a bump test reveals the reading has drifted out of tolerance.
The distinction matters for confidence in the numbers. Passing a bump test tells you the detector responds, but not that it responds accurately. Only calibration guarantees that a reading of, say, 30% LEL truly corresponds to 30% LEL. A sensible maintenance program uses both: frequent bump tests as an availability check and periodic calibrations as the accuracy correction, with a bad bump test triggering an early calibration or a sensor swap.
The mechanics of applying gas matter as much as the sequence. The calibration gas must reach the sensor at the right flow rate through a proper calibration adapter or cup, so the sensor sees the gas the way it would see the atmosphere, without over-pressurizing it or starving it. Then the reading needs time to settle: the detector must be given enough exposure to reach a stable value before the zero or span is set, because reading it too soon captures a value that is still climbing and yields a bad calibration.
Calibration frequency depends on the sensor type, the environment, and the site's policy, and there is no single universal interval that fits every detector. Sensors in harsh or contaminated conditions drift faster and are calibrated more often, while stable installations can go longer. The right approach is to set an interval, track drift through bump tests and as-found readings, and shorten the interval if a sensor is repeatedly found out of tolerance.
Recording the results turns calibration into evidence. Good practice logs the as-found value, what the detector read against the gas before any adjustment, and the as-left value, what it read after calibration. The as-found number reveals how much the sensor drifted and whether it might have been misreading in service, while the as-left number confirms it was returned to accuracy. Capturing these in a maintenance record, with calibration-due dates that flag detectors as they come up, lets a platform like Merobix keep many detectors on schedule and preserve the drift history that shows which sensors are aging out.
A bump test is a quick pass-or-fail check that a detector responds to gas and alarms; it does not adjust anything. A full zero and span calibration measures the detector's drift and actually corrects its zero and gain so readings match known values. Bump tests are done frequently to catch failures, while calibrations are done less often to restore accuracy.
The zero step uses gas known to be free of the target, typically clean instrument air or nitrogen, to set the baseline. The span step uses a certified cylinder of the target gas at a precisely known concentration to set the sensitivity. Using a traceable, in-date span cylinder is essential, because the calibration is only as accurate as the gas standard behind it.
The as-found reading, taken before any adjustment, shows how far the detector had drifted and therefore how accurate it was while in service. The as-left reading confirms the detector was returned to correct accuracy after calibration. Keeping both builds a drift history that reveals which sensors are aging and supports auditing the reliability of the gas detection system over time.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.