It’s one of the most common questions we get asked, and the honest answer is unsatisfying: it depends.
That’s not a dodge. Calibration intervals aren’t a single number you can look up — they’re a judgement based on what the instrument does, how hard it works, and what happens if it’s wrong. A pressure transmitter on a critical safety loop and a temperature gauge on a non-critical utility line might sit in the same plant, but they don’t belong on the same schedule.
Here’s what actually determines the answer.
First, what “drift” really means
Instruments don’t fail all at once. They drift.
A transmitter that read accurately on the day it was installed will, over months of thermal cycling, vibration, and process exposure, gradually start reporting values that are slightly off. Nothing alarms. Nothing looks broken. The reading just quietly stops being true.
That’s what makes calibration different from most maintenance work. You’re not fixing something that’s visibly wrong — you’re checking whether something that looks fine still is. And because drift is gradual, the cost of getting the interval wrong is rarely dramatic. It’s a slow accumulation of decisions made on bad data.
The four factors that set your interval
1. What the instrument is measuring, and how critical it is
An instrument tied to a safety interlock, a regulatory reporting requirement, or a quality-critical process step carries more consequence than one providing general indication. The higher the consequence of an incorrect reading, the shorter the sensible interval.
2. How hard the instrument works
Duty cycle matters. An instrument operating continuously at the top of its range, exposed to temperature swings, vibration, or corrosive media, will drift faster than the same device in mild, steady conditions. Two identical transmitters in the same plant can genuinely need different schedules.
3. What the manufacturer recommends
Manufacturer documentation is the starting point, not the final word. It’s based on typical conditions, and your conditions may be harsher or gentler than typical. Treat it as a baseline you then adjust.
4. What your industry or client requires
Some sectors have calibration requirements written into regulation, certification schemes, or client contracts. Where those apply, they set a floor — you can calibrate more often, but not less. Worth checking what applies to you before setting anything else.
The trap of calibrating too often
Most conversations about calibration focus on the risk of doing it too rarely. The opposite is also a real cost.
Every calibration takes time, may require taking a loop out of service, and consumes technician hours. Calibrating a low-consequence instrument on an aggressive schedule doesn’t make your plant safer — it just moves budget away from the instruments where tighter intervals genuinely matter.
The goal isn’t maximum frequency. It’s the right frequency for each instrument.
Let your own data set the schedule
Here’s the part that’s often skipped: your calibration history is the best guide you have.
If an instrument comes back in tolerance at every single calibration, year after year, that’s evidence the interval could be extended. If one regularly comes back out of tolerance, that’s evidence it should be shortened — or that something about its installation or environment needs attention.
This is why calibration records matter beyond compliance. A well-kept history turns interval-setting from guesswork into something you can actually justify, to an auditor or to yourself.
Starting from scratch with no history? Begin conservatively, record everything, and adjust as the pattern emerges.
What good documentation should include
Whoever carries out your calibration, you should end up with records showing:
- Which instrument was calibrated, identified unambiguously
- When it was done, and by whom
- The as-found readings — what the instrument was reading before adjustment
- The as-left readings — what it read after
- Whether it was found within tolerance
- What reference equipment was used
The as-found reading is the one people most often overlook, and it’s the most useful. It tells you how far the instrument had drifted, which is exactly the information you need to decide whether the interval was right.
A practical way to start
If you don’t currently have a structured programme:
- List your instruments and mark which ones carry real consequence if they read wrong.
- Group them by criticality and duty, rather than treating every device the same.
- Set conservative starting intervals for the critical group, using manufacturer guidance as a baseline.
- Record as-found readings consistently from the first cycle onward.
- Review after two or three cycles and adjust intervals based on what the data shows.
It’s less work than it sounds, and it replaces a recurring guess with something defensible.
