Should you calibrate on the manufacturer's schedule or on a risk-based interval? Here is how to decide, justify it to auditors and avoid costly over-calibration.
A manufacturer’s calibration interval is a sensible starting point, not a rule. A risk-based interval is one you set yourself, from how the instrument is used, how much error your process can tolerate and how the instrument has actually performed. For most Australian manufacturers, labs and food and pharma sites, the strongest position is to start with the manufacturer’s advice and then move to a documented, risk-based schedule that you can defend to an auditor.
This guide explains the difference, shows where each approach fits and sets out the records you need so that your intervals hold up in an audit.
Manufacturer-based calibration follows the interval printed in the manual or on the quote: typically 12 months for many instruments, sometimes 6 or 24. It is simple to run. You diarise a date, send the instrument and file the certificate.
Risk-based calibration sets the interval from evidence about your instrument in your application. The question changes from “what did the maker suggest?” to “how long can this instrument stay within the tolerance my process needs, and how do I know?”
The two are not opposites. The manufacturer’s figure usually becomes the first data point in the risk assessment. NIST says it does not require or recommend any set recalibration interval for measuring instruments in general, which tells you where the responsibility sits: with the user of the instrument.
A manufacturer writes one interval for every customer. That customer might be a school lab using a balance twice a week, or a packing line weighing product every few seconds in a hot, dusty room. The same instrument carries very different risk in those two places.
The manufacturer also does not know your tolerance. A balance that is more than good enough for dispensing 50 kg batches may be unfit for weighing 10 mg of an active ingredient. Fitness for purpose is defined by your process, not by the data sheet.
OIML D 10, which is also published as ILAC-G24, is explicit that each laboratory chooses which methods to apply based on its own needs and risk assessments. The initial interval can take manufacturer advice into account, but it is then reviewed against your own history.
Published guidance on setting intervals keeps returning to the same questions. A presentation on selecting calibration intervals lists them plainly: the type of instrument, the manufacturer’s interval, how extensive and severe the use is, the environment, and whether the instrument tends to wear or drift. OIML D 10 adds the required measurement uncertainty and the effectiveness of intermediate checks.
Turned into a working checklist:
Take a 3 kg bench balance on a packing line, calibrated annually on the manufacturer’s advice. Over three years, the as-found errors at each calibration sit at a small fraction of the tolerance your product specification allows. A weekly check with a certified check weight has never failed.
That history is the basis for a risk assessment that may support a longer full-calibration interval, with the weekly check kept in place. Now change one detail: the same balance moves to a washdown area and begins to show check failures every few months. The same method now shortens the interval. The method is the same, only the evidence differs.
The numbers above are illustrative, not a recommendation for your instrument. Your own tolerance and history decide the outcome.
NATA’s General Accreditation Guidance is written to be read alongside the informative annex of its General Accreditation Criteria on equipment assurance, in-house calibration and equipment verification. It gives guidance for calibration and checking intervals rather than a single mandatory number. Its companion calibration reference equipment table frames its intervals as guidance for establishing calibration and checking intervals for reference equipment. Accredited facilities are expected to justify their intervals, not simply adopt a default.
The pharmaceutical pharmacopoeias have moved in the same direction. The review Metrology of weighing in the pharmaceutical industry notes that regulations are usually vague on how instruments should be calibrated and tested, and give no specific time interval for the tests other than in general terms. Its section on a risk-based approach is where the argument for application-specific frequencies is made. USP General Chapter <41> Balances introduces the relationship between calibration and performance checks within a life-cycle approach, and the concept of a risk-based approach to the frequency of both.
One practical consequence: the habit of a mandatory daily check on every electronic balance is a legacy of mechanical instruments. A documented risk assessment may support less frequent manual sensitivity tests where the balance has built-in adjustment. That is a decision for your quality unit, backed by data, not a shortcut.
Some sectors set stricter expectations. CASA’s guidance on calibration of inspection and test equipment covers approved maintenance organisations, which must control tools and test equipment that require calibration. If you work under a regulator’s manual, follow its requirement first and apply risk-based thinking inside it.
These terms answer different questions, and mixing them up causes most of the confusion.
| Question | Manufacturer interval | Risk-based interval | Accredited (NATA) calibration |
|---|---|---|---|
| What does it decide? | When the maker suggests recalibrating | When your evidence says you should recalibrate | Who performs the calibration and how competence is shown |
| Where does it come from? | Manual or data sheet | Your tolerance, drift, usage and environment | ISO/IEC 17025 competence assessed by NATA |
| What do you need to keep? | Certificate and date | Risk assessment, as-found data, check records | Certificate with traceability and scope |
| Main weakness | Ignores your application | Weak if undocumented | Does not set your interval for you |
A risk-based interval tells you when. Accreditation tells you how well the work is done. You can, and in regulated work usually should, have both. NATA’s guide for users of calibration labs explains that scopes of accreditation describe the parameters a lab can measure, the ranges and the smallest measurement uncertainty it can achieve. Check the scope matches your instrument and the uncertainty you need before you book.
If the difference between traceable and accredited certificates is unclear, our guide to traceable vs accredited calibration documents and our NIST vs NATA certificate checklist go through what to look for on the paper.
Many teams count an internal adjustment as their calibration. A built-in mass can correct sensitivity at a single point, and that is useful. It does not show how the balance performs across its range, at low load, or in repeatability, and it does not produce traceable evidence against external standards.
A sound risk-based plan therefore keeps three layers apart:
Our article on calibration vs adjustment explains why auditors ask to see the as-found data and not only the “after” result.
You do not need to overhaul everything at once. Work through it in order.
The most common mistake, as the NIST and OIML material makes clear, is treating the maker’s number as a requirement. The second most common is extending intervals to save money without any check data behind the decision.
An auditor reviewing a risk-based schedule is not asking whether your intervals are long or short. They are asking whether you can show your reasoning.
If an instrument is found out of tolerance, you need to assess the impact on product measured since the last good result. That is why the interval matters: a longer gap means a bigger look-back. Our pharma calibration checklist covers out-of-tolerance steps, and our guide to calibration certificate checks lists what to confirm on each certificate.
Keeping calibration records clean is a discipline in its own right. A technician’s checklist for measurement videoscopes shows how the same principles of repeatable readings and recorded checks apply to another instrument type.
Risk-based is not always the better answer. Stay with the manufacturer’s interval, or shorten it, when:
An undocumented “we think it is fine” extension is the weakest position of all. A documented annual calibration is better than a poorly argued 24-month one.
When we calibrate instruments across manufacturing, food, pharma and aerospace sites, we see the same pattern. The sites with the fewest surprises are not the ones calibrating most often. They are the ones that know their instruments: they keep the as-found data, they run simple checks between visits and they can explain their interval in two minutes.
We would rather help you build that evidence than sell you more visits than you need. Because we are NATA accredited and ISO 9001 certified, and we service multiple equipment brands regardless of where you bought them, we can look at a mixed fleet and help you sort which instruments warrant tighter control. A balance that drifts or a logger that sits near its limit shows up in the as-found results, and that is the conversation worth having.
One piece of advice we repeat: treat the first interval as a hypothesis. Calibrate, check, review and then decide.
— Kaz
A risk-based schedule is only as good as the data your provider returns. Ask for:
If you are reviewing intervals across a mixed fleet, our NATA calibration, traceable calibration, trade verification and maintenance services can be done on-site or in our workshop, with 24/7 emergency support. You can see how the process works before you request a quote. If you need check weights, thermometers or loggers for your intermediate checks, browse our range of measuring equipment.
Yes, if you can justify it. The manufacturer’s figure is an initial interval, and OIML D 10 expects you to confirm or change it using your own history. You need stable as-found results, documented intermediate checks and a written rationale. Without those, keep the manufacturer’s interval.
Generally yes, when it is documented and reviewed. Auditors want to see the method, the data and evidence that you act when results drift. Your customer or regulator may set specific requirements, so check these first.
No. An internal adjustment corrects sensitivity at one point using a built-in mass. It can support a reduced frequency of manual checks, but it is not a full-range calibration against traceable external standards.
Not always, as accreditation is not a legal requirement in every industry. NATA accreditation is, however, the main way in Australia to show competence to ISO/IEC 17025, and many pharmaceutical, food and manufacturing customers expect it. Confirm what your customers and regulators require.
A drop or overload, a repair, a move to a harsher environment, a failed intermediate check or a reading that looks wrong. Each of these should be written into your procedure as a trigger.
