Temperature probe placement in cold rooms, fridges and display cases

Most refrigerated units already show a temperature, and most of those readings are taken in the wrong place for food safety. The controller's probe sits where the refrigeration engineer needed it for control: in the air entering or leaving the evaporator. The food sits somewhere else, often in the warmest part of the unit. A HACCP record needs the temperature the food sees.

This guide covers where to put an independent probe in each kind of unit, when to put it in a buffer, how to stop defrosts and door openings raising false alarms, and how to check the probe. The HACCP critical limits guide covers the limits and records the probe feeds.

Find the warmest point first

The rule is simple: monitor the warmest place where food is kept. A probe there gives the worst case, so if it stays inside the limit, the rest of the unit does too.

EU law describes those places for quick-frozen food. Directive 92/2/EEC, the official sampling method, takes samples "representative of the warmest points". In a cold store, that is "near the doors (upper and lower levels), near the centre … and near to the air return of the cooling unit". In a display cabinet, it is three locations representative of the warmest points.

Measured data agrees. A 2007 survey of retail display and catering cabinets (Evans, Scarcelli and Swain) found the warmest packs here:

The warmest point in a specific room can only be found by measuring it. A temperature mapping study puts temporary loggers in a grid and records them through loading, defrost and door openings. The WHO method for storage areas places loggers every 5 to 10 m, at three heights in rooms up to 3.6 m, and records cold rooms for 24 to 72 hours. Permanent probes then go where the map says. The cold room monitoring guide covers this step.

The HACCP temperature monitoring guide applies these positions to the fridges, freezers and hot holding of a kitchen or shop, and the food processing guide to blast chillers and dispatch bays.

Probe positions, unit by unit

For open display cases of quick-frozen food, Regulation (EC) No 37/2005, Article 3, adds a specific rule. The maximum load line must be clearly marked, and the thermometer "shall indicate the temperature at the air return side at the level of that mark". ISO 23953-2 governs the display instrument fitted to a cabinet. It allows a maximum error of 2 K and states that the "measured air temperature is not identical with foodstuff temperature". An independent probe in the warmest product position fills that gap.

Write the position down for each probe, with a photograph. A probe that moves 30 cm during a stock rotation reads a different place, and a replacement fitted a year later needs the same position to keep the record comparable.

Air or product: buffered probes

Air temperature changes in seconds. A full shelf of food changes in hours. A door left open for a delivery can raise the air near the door by several degrees while the product rises by a fraction of one. An air probe at the warmest point therefore reports real air excursions that mean little to the food, and an alarm on that probe either fires constantly or needs a long delay.

A buffered probe solves this. The probe sits inside a small volume of liquid or solid that responds like food. The US CDC lists glycol, glass beads, sand and Teflon as buffers, and says that temperatures "measured by a buffered probe match vaccine temperature more closely". NIST's study of vaccine refrigerators (NIST IR 7900) found that air and wall sensors produced "false positive" alarms, while sensors in glycol bottles stayed within the 2 °C to 8 °C range through door openings that caused large spikes in the air.

Use both where it matters:

• Air probe: shows the unit's behaviour, catches a door left open, finds a failing fan. Alarm with a long delay, or use it for information only.
• Buffered probe: represents the food. This is the HACCP record and carries the main alarm.

Say in the plan which probe reads what. The same 8 °C limit means something different on a buffered probe than on an air probe.

Size the buffer to the food. A small bottle follows a shelf of retail packs; a large one lags further and can hide a real excursion until it is well under way. Record the air and the buffered probe side by side through a door opening and a defrost before you fix the alarm settings.

Defrost cycles and alarm delays

Every chiller and freezer with frost on its evaporator defrosts on a schedule. Heaters, hot gas or a stopped compressor let the coil warm, and the air in the unit rises for the length of the defrost and the pull-down after it. This is normal, and UK law accepts it: the guidance on temperature control legislation names "defrosting of equipment" as a reason food may briefly be above its limit.

Controllers already handle this with delays. Their factory defaults show the range:

An independent system needs the same logic. There are three ways to give it:

1. A delay longer than the defrost. If defrosts last up to 30 minutes, an alarm delay of 45 minutes on the air probe never fires on a normal defrost. A buffered probe barely moves during a defrost and can use a shorter delay.
2. The defrost signal from the controller. Read the controller's defrost state over Modbus and block the alarm while it is set, plus a recovery time. The refrigeration controller guide shows where that state lives.
3. A schedule. If defrosts run at fixed times, restrict the alarm to the hours between them. This breaks when someone changes the defrost schedule, so it is the weakest option.

A door alarm is a separate problem with a separate answer. A door switch, or a fast rise on the air probe near the door, tells staff to close it. The food alarm on the buffered probe tells them whether the food is at risk.

Two cautions. First, a long delay hides a slow failure for that long, so set it from how fast the product warms, not from how often the alarm annoys people. Second, Food Standards Scotland's RetailSafe guidance tells staff not to take a manual reading "immediately after the door/lid has been open" or during a defrost. An automated record contains those readings, so whoever reviews it must know when the defrosts ran.

Reading interval

Regulation (EC) No 37/2005 asks for quick-frozen food air temperatures to be recorded "at frequent and regular intervals", without a number. WHO guidance for pharmaceutical storage records every 1 to 15 minutes.

A TES probe temperature sensor runs for up to ten years at 15-minute reporting, by the datasheet's estimate at 23 °C with full wireless coverage. A shorter interval, a colder site or a weaker radio signal shortens that. A buffered probe changes slowly, so a 15-minute interval loses little on it.

Radio, cables and enclosures in cold rooms

A cold room is a box of steel-faced insulated panels, and steel blocks radio. A wireless sensor inside a closed cold room may not reach a gateway outside. The usual answer is to mount the sensor outside the room, on the wall next to the door, and run the probe cable in through the door seal or a sealed gland. The TES waterproof probes come on 1.5 m and 2.5 m cables for this. The wireless link budget calculator estimates the margin for a path that has to cross a panel or a door.

Check the enclosure's rating against the room. The TES enclosure is IP67 and operates from −30 °C to 55 °C, so in a blast freezer below −30 °C it goes outside and only the probe goes in. Wash-down areas need the enclosure and the probe rated for water jets; the IP rating checker explains the codes.

Calibration checks

A fixed probe needs the same checks as a hand thermometer:

• Ice point: crushed ice and a little water, stirred, at 0 °C. The FSA's Safer Food Better Business pack accepts −1 °C to 1 °C. FSAI asks for −0.5 °C to 0.5 °C and probe accuracy of ±0.5 °C.
• Boiling point: 99 °C to 101 °C (FSA) or 99.5 °C to 100.5 °C (FSAI), for probes used on hot food.
• In place: a calibrated reference thermometer beside the fixed probe at the operating temperature, after both have settled.

FSAI asks for a check at least twice a year. The FSA pack gives monthly as an example. EN 13486, the standard for verifying cold chain thermometers and recorders, leaves the interval to the user and asks for a check after any repair, sensor change or incident. Record the result per probe, with the reference used. Use a reference thermometer with a current calibration certificate, so the record is traceable.

Commissioning checklist

1. Each probe is at a position from the mapping study or the table above, photographed and labelled.
2. The plan says whether each probe reads air or a buffer.
3. Each probe agrees with a reference at the operating temperature, within the plan's tolerance.
4. The alarm delay has been tested against a real defrost and a normal delivery.
5. A warm probe raises an alarm that reaches the named person, out of hours.
6. A sensor with its battery removed raises a no-data alarm.
7. The reading interval, the time source and the record retention are set and written down.