As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
A bump test asks one question and accepts one of two answers: expose the instrument to a concentration above its alarm point, and either it alarms or it does not. A calibration asks how much, by applying a certified concentration and recording how far the displayed value sits from it, after which an adjustment may or may not be made. The first proves the instrument is alive and takes a minute. The second quantifies how wrong it has become and needs certified gas, a regulator and a controlled flow. Both exist because a sensor can fail suddenly between calibrations and can also drift slowly in a way no pass or fail check would ever notice.
A go no go check and a measurement against a standard
The two procedures differ in their output before they differ in anything else. A bump test produces a boolean. The instrument either crossed its threshold and announced itself or it did not, and no number from the exercise is kept as a measurement. Because the output is binary, the test gas concentration only needs to be comfortably above the alarm point, and precision in the applied concentration buys nothing.
A calibration produces numbers. A certified concentration is applied, the displayed value is recorded, and the difference between the two is the quantity of interest. Everything about the procedure exists to make that difference attributable to the instrument rather than to the method: certified gas with a stated tolerance, a regulator that delivers a known flow, enough time for the reading to stabilise, and conditions within the reference range.
That difference in output drives everything downstream. A binary result can be produced in a doorway in sixty seconds. A numeric result needs equipment, a procedure and a record, which is why nobody does it daily and why the gap it leaves has to be covered by something cheaper. The dedicated page on why detectors need both covers the risk argument in more detail.
What gets adjusted and what only gets recorded
Two points define the behaviour of most sensors well enough to correct it. The zero is what the instrument reports when the measured quantity is absent, and the span is how much the reading moves per unit of the quantity. Setting the zero in clean air or certified zero gas removes an offset. Applying a known concentration and setting the span scales the response. Drift in either produces a wrong reading and the two drift for different reasons and at different rates.
The vocabulary matters because records depend on it. Strictly, calibration is the comparison and adjustment is the change, and an instrument can be calibrated with no adjustment at all if it is found within tolerance. The difference between calibration and adjustment is not pedantry: a record that merges them loses the as found value, which is the only evidence of how far the instrument had drifted while it was in service.
The same two point structure turns up far from gas work. The YR1035 four wire battery resistance meter states a four wire measurement method, a range up to two hundred ohms with a minimum resolution of a hundredth of a milliohm, and a hold key that enables a zeroing function for measuring internal resistance down to a fraction of a milliohm. That zeroing is an offset correction, removing lead and contact resistance before a very small measurement. It does nothing to the span, which is exactly the limitation a zero adjustment has on a gas instrument.
Why frequency differs between the two
Cost and consequence set the intervals. A bump test consumes a few seconds of gas from a small cylinder, needs no records beyond a line in a log, and can be done by whoever is about to use the instrument. The failure it catches is sudden and total, which is the failure most likely to appear between one day and the next. Daily, or before each period of use, is therefore both affordable and well matched to the risk.
A calibration consumes certified gas, equipment and time, and often takes the instrument out of service for part of a day. The failure it catches is gradual: a sensor losing sensitivity slowly enough that it still crosses a generous bump threshold while reporting readings that are quietly low. That kind of drift accumulates over weeks, so an interval of weeks to months matches it.
The two procedures also feed each other. A failed bump test escalates to a calibration, because the question immediately becomes how far out the instrument is rather than whether it responds. A failed calibration escalates to repair or replacement. Running one without the other leaves a gap in one direction or the other, and the gap is invisible until something happens in it.
The same split in instruments that are not gas detectors
Acoustics makes the structure unusually visible because both halves have formal names. The CRYSOUND CRY2831 sound level meter states conformity with IEC 61672-1 Class 2, a measurement range of twenty nine to one hundred and thirty eight decibels, a frequency range of twenty hertz to twelve and a half kilohertz, A, C and Z frequency weightings with fast and slow time weightings, IP54 protection and a calibration certificate supplied with the instrument. The certificate is the laboratory half. The field half is an acoustic calibrator applied to the microphone before and after a measurement session, which is a check of response rather than a measurement of error, and which plays the same role a bump test plays on a detector.
What that supplied certificate records, and the interval over which it is expected to hold, are Not stated in the listing bullets, so the document itself has to be read before it can be relied on. Accessories complicate both halves in the same way. The Pomona solar connector to four millimetre test lead set states that it connects test tools accepting four millimetre sheathed banana plugs to photovoltaic modules and systems for routine test and measurement work. A lead set sits in the measurement path and outside the scope of anything a certificate on the instrument body covers, which is why the zeroing step that removes lead resistance exists at all.
A certificate is a snapshot, not a guarantee
| Activity | Question it answers | What it changes in the instrument | What it leaves behind |
|---|---|---|---|
| Bump test | Does it still respond and alarm | Nothing | A pass or fail entry in a log |
| Calibration check | How far is the reading from a certified reference | Nothing | An as found value and a tolerance decision |
| Adjustment | Can the reading be brought back to the reference | Zero, span or both | An as left value |
| Certificate | What was true at one moment under stated conditions | Nothing | Traceability, conditions, as found and as left figures |
Read the last row carefully whenever a certificate arrives with an instrument. A document that records the reference, its traceability and both the found and left values is evidence you can reason about later. One that states only that a calibration took place is a claim about an event. The guidance on reading a calibration certificate works through which fields carry weight, and the related caution about automatic baseline calibration shows how an instrument can correct itself into an error when the assumption behind the routine does not hold.
Putting both into a routine you will actually follow
Tie the frequent check to an action and the infrequent one to a calendar. Before entry, before a shift, after a drop or a soaking: apply gas, confirm the alarm, log the result. That attaches the habit to something that already happens rather than to an intention. Keep the test gas somewhere the instrument lives, because a cylinder in a different building is a cylinder that never gets used.
For the slower procedure, set a reminder at the interval the manufacturer states and treat the as found figures as the real output. A pattern of instruments arriving slightly low tells you something about storage, environment or sensor age that no single certificate reveals. And keep the two words separate in every record you write, because a log that says calibrated when it means bumped is worse than no log: it reports a level of assurance that nobody actually obtained.
Frequently Asked Questions
Can a calibration replace a bump test?
Technically a calibration contains a response check, so an instrument that has just been calibrated has effectively been bumped. What it cannot replace is the frequency. Calibration needs certified gas, a regulator, a controlled flow and time, which is why it happens on an interval of weeks or months, and a sensor can fail the day after one. The cheap frequent check exists to cover the gap between the expensive infrequent ones.
Does every calibration involve an adjustment?
No, and conflating the two causes real problems with records. A calibration is strictly a comparison against a reference, which may well show the instrument is still within tolerance and needs nothing done to it. An adjustment is the separate act of changing the instrument so its reading matches the reference. A good record shows the value found before any change and the value left afterwards.
What is a zero adjustment actually doing?
Setting the output the instrument produces when the quantity being measured is absent, or as close to absent as the procedure can arrange. For gas that means clean air or a certified zero gas, and for an electrical measurement it means shorting the leads so the instrument can subtract the resistance of the leads and contacts. Both remove an offset and neither says anything about how the instrument behaves at the top of its range.
How often should a bump test be done?
Before each period of use is the common practice for instruments that protect people, and manufacturers of personal monitors typically recommend daily testing. The interval is a risk decision rather than a physical constant. An instrument kept in a clean cabinet and used monthly is in a different situation from one carried daily into an environment full of silicone sealant.
What makes a calibration certificate useful?
The detail behind the headline. A certificate that records the reference used and its traceability, the conditions, the values found before adjustment and the values left after it, is evidence. One that states only that the instrument was calibrated on a given occasion is an assertion. The as found figures are often the most informative part, because they show how much the instrument drifted over the interval.
Do these ideas apply outside gas detection?
Entirely. Any instrument with a reference inside it drifts, and the same split appears under different names. A sound level meter gets a field check with an acoustic calibrator before a session and a laboratory verification against a standard at longer intervals. A low resistance meter gets its leads zeroed before each measurement and its span checked against a reference separately. The vocabulary changes and the structure does not.