EN 12830: the temperature recorder standard almost nobody mentions

What it is

A standard for the recorder, not the fridge

The first edition, EN 12830:1999, covered air temperature recorders for chilled, frozen and quick-frozen food and ice cream. The 2018 edition, published in the UK as BS EN 12830:2018, was completely revised by CEN/TC 423. Its scope is any temperature-sensitive goods between −80 °C and +85 °C, and its examples include pharmaceuticals, blood, organs and biologicals as well as food.

It applies to the whole recording system, wired, wireless or cloud-hosted, and the 2018 revision added data protection, software tests and rules on sensor location. It covers recorders only: a thermometer that shows a reading without recording it falls under EN 13485.

The first requirement

Monitoring that does not depend on the control system

Both editions open their requirements the same way. As manufacturers' published compliance records reproduce it (clause 4.1 in 1999, 5.1 in 2018), the means of measurement the recorder uses must be independent of any temperature measurement used to control the refrigerating system.

Medicines guidance says the same in its own words. The HPRA's IA-G0011, section 3.1, says recording probes should be independent of controlling probes. WHO TRS 961 Annex 9, clause 4.5.1, asks for control sensors independent of the monitoring system. EU GDP 3.3 treats equipment that controls and equipment that monitors as two things to calibrate, but does not say they must be separate.

The reason is a single point of failure. If the controller's sensor drifts, moves or sits in cold air off the evaporator, the controller reads 5 °C while the product warms, and a monitor sharing that sensor reads 5 °C too. The RPS (2018, archived but not withdrawn) describes a vaccine fridge whose display read normally while an independent thermometer showed 0 °C; every vaccine was destroyed.

What it covers

Accuracy, recording, data and the environment

In summary, from the published scope and manufacturers' clause-by-clause records (the full text is licensed):

• Accuracy classes: 0.2, 0.5, 1 and 2, named for the maximum permissible error in °C (class 0.2, ±0.2 °C, was added in 2018). Recorders for chilled and frozen food should meet at least class 0.5 between −25 °C and +7 °C (per manufacturers' published compliance records).
• Recording: temperature, date, time and time zone, with the recording interval, duration and storage capacity defined by the manufacturer, limits on timing error, and a maximum response time.
• Data: protected against manipulation, an audit trail, access by user and role, readable copies, data kept and readable for at least a year, and nothing lost in a power cut.
• Environment: a stated climatic range for each part of the system and a minimum enclosure rating by location, from IP20 in a heated room to IP55 inside a cold enclosure.
• Documentation and marking: battery life, versions and combinations of sensors and cables, software documentation, and the marking set out in clause 8.

Where it is law

Quick-frozen food, in store and on the road

Commission Regulation (EC) No 37/2005 requires the means of transport, warehousing and storage of quick-frozen foodstuffs to have recording instruments monitoring the air temperature at frequent, regular intervals. Article 2(2) says that from 1 January 2006 those instruments must comply with EN 12830, EN 13485 and EN 13486, and that the food operator must keep the documents that show they conform. Records are kept for at least a year. It remains law in Great Britain, enforced in England through the Quick-frozen Foodstuffs (England) Regulations 2007. Retail display cabinets and local distribution need only a visible thermometer.

For international road and rail transport, the ATP agreement (Annex 2, Appendix 1, as amended in 2012) says the equipment carrying quick-frozen food must have a recorder that complies with EN 12830 and is verified to EN 13486 by an accredited body. Food detail is in our food guidance.

Where it is voluntary

Medicines, blood and laboratories

The only UK health guidance we found that cites it is NHS SPS: its sensor guidance asks for data loggers compliant with BS EN 12830. EU GDP asks for calibration at defined, risk-based intervals and sets no accuracy or traceability figure. The HPRA, WHO and NHS SPS go further and give a figure: probes accurate to ±0.5 °C, calibrated against traceable references (and, for WHO, a reading at least six times an hour). None of them names EN 12830 except SPS. We found no blood or laboratory guidance that names it.

So in a pharmacy, blood bank or lab, EN 12830 is a shortcut to evidence about the recorder: accuracy, data protection and an audit trail tested against one published method. It does not replace mapping, calibration or the validation your computerised system still needs.

The companion standards

EN 13485 for thermometers, EN 13486 for verification

EN 13485 does for thermometers what EN 12830 does for recorders; its 2023 edition widened it from food to all temperature-sensitive goods. EN 13486 is the periodic check: how to verify that a recorder or thermometer still meets its class. The 2023 edition added class 0.2, a functional test for cloud-hosted recorders and guidance on uncertainty.

Compliance when a recorder leaves the factory says nothing about the probe three years later. That is the job of periodic verification or calibration.

An annual UKAS-accredited calibration across the working range meets the periodic verification EN 13486 describes, in our view (EN 13486 is licensed; we paraphrase). That is our position, not a statement by UKAS or by the standard. Our UKAS laboratory (No. 10723) calibrates temperature from −90 to +135 °C in a dry block, for probes, and from −35 to +40 °C in an air chamber, for whole loggers and sensors, so check that the range you use sits inside it.

What to ask your supplier

Five questions before you buy, or at renewal

A compliant recorder comes with a manufacturer's declaration and a test report. Ask for both, and ask the questions below. (Ours: the Sematics recorders we supply are designed to meet EN 12830:2018, accuracy class 0.2, so specified to ±0.2 °C or better, as the Node+ datasheets state.)

• Is your recorder EN 12830 compliant, to which edition, and in which accuracy class over our range?
• Is it declared for storage, for transport or both, and which parts of the system did the tests cover, including the cloud service?
• Is the monitoring sensor independent of our fridge's control sensor?
• How is it verified to EN 13486 or calibrated, how often, and at how many points?
• What happens to the data in a power cut or a network outage, and for how long is it kept?

Worked example

The controller said 5 °C

This is an illustrative example. A hospital pharmacy cold room runs at +2 °C to +8 °C. Its controller sensor sits in the return air near the evaporator, and the controller display has read 5 °C all week.

On Thursday night an independent monitoring probe at the mapped warm spot, by the door, crosses 8 °C and alerts the on-call pharmacist. The controller still shows 5 °C. An engineer finds a torn door seal, with the control sensor reading cold air off the coil. The monitoring record shows when the rise began and how far it went, so each product can be assessed. Had the only reading come from the controller, nobody would have known.