Written by: Localyn · Published: August 2026 · Last reviewed: 22 October 2026

How wireless fridge sensors work in restaurants

A wireless fridge sensor measures a temperature, sends that reading on, and can tell a named person when it leaves the range you set. It does not decide whether the food is safe, and it does not replace a food safety management system. This guide explains the path from the probe to the alert, what the number does and does not mean, and how a UK kitchen should place, check and answer sensors so the record stands up at inspection.

The path from sensor to a decision

Every useful system has the same jobs, whatever the brand. A sensing element measures temperature at a point you chose. A radio sends that measurement out of the cabinet. Something stores the time and the value against a named fridge. A rule compares the value with a limit from your safe method. A person who can act receives the exception, looks at the food and the equipment, and writes down what they did. If any step is missing, you have a gadget, not a control.

The sensing element is often a digital probe or a thermistor in a food-safe housing, sometimes in a small bottle of glycol or water so it behaves more like food than like air. The transmitter may sit inside the fridge, or outside it with a cable in. Metal boxes block radio, which is why a sensor buried in a full walk-in can go quiet the week after installation. A gateway, the site Wi-Fi, or another radio link carries the reading to a service that keeps the history and sends the alert. Where Localyn monitoring is switched on, that product path is described in how Localyn temperature monitoring works. The checks below apply whichever system you use.

Write the chain in the safe method in ordinary words: which unit, which device, what limit, who is alerted, what they do, and what happens if no reading arrives. Staff should be able to say that without a vendor manual. Buying the hardware before those sentences exist is how kitchens end up with dashboards nobody trusts.

Air temperature is not food temperature

Most fridge sensors measure air, or a bottle sitting in the air stream. Air moves quickly. Open the door for a delivery and the air reading can jump several degrees, then fall after the door shuts. The ham on the shelf changes more slowly. A single air spike is not, by itself, proof the food is unsafe. It is a prompt to look.

A glycol or water bottle warms and cools more like a product, so brief door openings disturb it less. That is a placement choice, not a legal requirement. Say which kind of reading you take. Do not treat an air sensor, the fridge display, and a probe in food as if they must always match. They answer different questions. When they disagree, find out why before you "correct" one of them to match the others.

In England, Wales and Northern Ireland, food that must be kept cold is generally held at 8°C or below. That is a legal maximum, not a target to sit on. Safer Food, Better Business tells caterers to work colder, commonly 5°C or below. Frozen food should be at −18°C or colder. Set sensor limits to the working temperatures in the safe method. An alarm only at 8°C stays quiet while a fridge sits at 7°C all week. The fridge temperature log guide covers how those limits are used in day-to-day records.

How the reading leaves the fridge

"Wireless" covers several designs. Some sensors join the kitchen Wi-Fi. Some use a short-range link to a hub that must stay in range. Some use another radio to a gateway, which then uses the broadband connection. The differences that matter are battery life, whether the signal escapes a metal cabinet, and what fails when the internet drops. Ask the supplier to show each failure on installation day.

Battery. A sensor that dies quietly is worse than no sensor. You need a low-battery warning early enough to change it, and a named person who does that job.
Radio and metal. If the transmitter is inside the cabinet, confirm the gateway still hears it with the door shut and the unit full. Re-test after you move shelves.
The network. Know whether readings queue on the device when the gateway cannot reach the service, or whether they simply stop. A store of unread readings is not an alert.
The person. An email to a closed office is not a response. Name a first contact on shift and a second if the first does not act.

Power cuts need a sentence in the same procedure. When the power returns, the fridge may still be warm. The sensor shows that only if it has battery or power, and only if someone looks. A graph back at 5°C by morning does not tell you how warm the cream was at 2am unless the history is fine enough to show it. Walk the food.

Thresholds, delays and door openings

A threshold copied from a default is why alerts get muted. If every breakfast delivery pages the manager, the manager will silence the phone. Set the limit from the safe method, then set a delay that ignores a short door opening but does not hide a door left ajar. Many kitchens use about 15 to 30 minutes for an air sensor. That is a house rule, not a figure in the regulations. Write it down. Review it after a fortnight and record why you changed it. Do not keep nudging the limit up until the alerts stop.

A reading that hits 9°C for four minutes and is back at 4°C is often a door. A reading that climbs from 4°C to 7°C over two days may never alarm, and it is still a failing seal, a blocked condenser, or overloading. Someone needs a weekly look at the trend, unit by unit. A fixed air sensor also does not tell you the centre of a joint or the temperature of soup in a bain-marie. Keep probes for those checks. The guide to comparing monitoring systems sets out where a sensor helps and where a person with a probe is still the control.

Where to put the sensor, and how to check it

Place it where the reading represents the food you are protecting. Air off the evaporator is colder than the rest of the cabinet. The top of the door is often warmer. Neither is a fair picture of the middle shelf unless you chose the worst spot on purpose. Fix the probe so it cannot fall into a door pocket or sit under a tray. Label the device and the fridge with the same name the team uses. "Unit 3" on a dashboard and "the dessert fridge" on the diary is how actions get written against the wrong cabinet. One sensor does not cover two fridges. After a deep clean, check the probe went back to the same place.

Compare the sensor with a food probe that has passed the ice-water and boiling-water checks, commonly within about 1°C of 0°C and of 100°C. The probe calibration guide and the probe calibration record cover that habit. Leave both in the same part of the fridge until they settle. Write the date, both readings, and what you did if they disagreed by more than your method allows. If the sensor reads 3°C colder than the probe, a dashboard showing 5°C may be an 8°C fridge. Adjust, replace, or take the unit off the list until it is fixed. Recheck after a battery change or a move. A factory certificate is the start of the story, not a reason to skip the check in your kitchen.

When the data stops

The dangerous fault is silence that looks like calm. Some systems draw the last reading forward. Some leave a gap. Some send "no data" and some do not. Your procedure should make a missing reading as visible as a high one. At opening, someone confirms that every unit has a reading from this morning, not merely that no alarm went off. If the data has stopped, treat the fridge as unchecked: read the display or probe the food, write that reading down, and raise a fault. A fridge temperature log sheet is a sound fallback for that day. Do not type pretend readings into the gap so the graph looks continuous.

Three failures deserve their own lines. The sensor is wrong, compared with a trusted probe. The sensor cannot communicate. The person who was alerted does not act. Test the third one. Send a trial alert during service and time how long a written decision takes. If the answer is "we would have seen it in the morning", you do not yet have a night-time control.

What environmental health officers check

Officers do not require wireless sensors. Paper readings are acceptable when they are real, timely and reviewed. If you rely on sensors as the temperature record, that history is what they will ask to see: the unit, the time, the reading, the limit, and the action when the limit was missed. They may compare the screen with the fridge in front of them. A login that only the owner knows, on a laptop at home, is not an accessible record during the visit.

Is every fridge and freezer that holds food on the system, under the name staff use?
Do the limits match the safe method, including a colder working target where you set one?
When the graph went high, who was told, what happened to the food, and where is that written?
When was the sensor last compared with a checked probe?
What do you do on a day with no readings, and is that manual check filed with the history?
Can you show several weeks? A history that expires after seven days will not answer a question about last month.

They will still open the door. Overloaded shelves, raw meat above ready-to-eat food, and a split seal are visible whether or not the air is at 4°C. Corrective actions are the bridge between the alert and the decision. Our corrective action guide is the standard those notes should meet. Sensors do not clean a slicer, train a starter, or complete the 4-weekly review. A head-office dashboard that nobody at the site can explain will not satisfy an officer standing in that kitchen.

Worked example

A cafe walk-in alerts at 6:10am: air temperature 9°C, held above 8°C for 35 minutes. The delay was 30 minutes, so a short delivery would not have paged anyone. The opener finds the door not latched. The sensor is in a glycol bottle on the middle shelf. They shut the door and probe a bottle of milk and a tray of ham with the checked probe: 6°C and 5°C. The food has not followed the air all the way up. They write "door ajar after delivery, food probed 6°C and 5°C, door shut, Sam". At 6:50am the sensor is 4°C. No food is thrown away. The note is on the corrective action, not only in a text message.

The following Wednesday there is no alert and no reading after 4pm on Tuesday. The opener notices because the opening check says "confirm today's reading exists". They write the display temperature, 5°C, on the paper log, swap the battery, and record the gap rather than backfilling it. Friday's comparison is sensor 4°C, probe 4°C. The 4-weekly review mentions both events: a door that does not latch, and a battery changed because the silence was visible. The sensor did the job it is capable of. It still needed a person.

Common mistakes

Treating a door spike as spoiled food, or a flat graph as proof nobody opened the door.
Alarming only at the legal maximum while the safe method says 5°C.
Muting alerts because door openings were never given a delay.
A sensor put back in the wrong fridge after cleaning, still named as the walk-in.
No check that readings are arriving. Silence looks like a quiet night.
No comparison with a probe you have checked in ice water.
Typing normal readings over a gap so the report looks complete.

Where this comes from

Temperature limits here follow Food Standards Agency Safer Food, Better Business advice for caterers and the cold-holding duties in the food hygiene regulations for England, Wales and Northern Ireland. Scottish businesses should follow Food Standards Scotland and CookSafe where they use it. Wireless sensors are not an FSA product and they are not mandatory. They are one way to collect the readings your own safe method already asks for. The method, the person, and the corrective action remain yours. If a supplier's default delay or limit disagrees with that method, the method wins, and the change should be written down.

Fridge sensor readings shown on a temperature report — A sensor history is inspection evidence only when the exception and the action sit with it.

Frequently asked questions

Do wireless fridge sensors need Wi-Fi?

Not always. Some join the kitchen Wi-Fi. Others use a radio link to a gateway, and the gateway uses the broadband connection. Ask what happens to readings and alerts when that link is down, and write the fallback into the safe method.

Can a sensor replace the daily fridge check?

It can supply the temperature history and an alert. Staff still need to confirm readings are arriving, inspect the fridge, respond when the limit is missed, and record the action. A sensor does not clean the unit or separate raw and ready-to-eat food.

How often should a wireless sensor record?

Often enough to show a drift and to support the delay on alerts. A reading every few minutes is common. Very long gaps miss a door left open until the food has already warmed.

What temperature should the alarm use?

Use the working limit in your safe method. In England, Wales and Northern Ireland the legal maximum for chilled food is generally 8°C, and SFBB tells caterers to aim colder, often 5°C or below. Frozen food is kept at −18°C or colder. A delay can ignore brief door openings. Do not raise the limit just to reduce alerts.

What should we do if the sensor stops sending?

Treat the fridge as unchecked. Read the display or probe the food, write that reading down, and repair or replace the sensor. Do not invent readings for the missing period. Keep the manual entry with the usual history.

Fridge temperature logs Freezer temperature logs Fridge temperature log sheet EHO readiness quiz Bakeries Chilling module

Sources

Last reviewed 22 October 2026. These links are the pages this guide was checked against. They are not an endorsement of Localyn, and they do not make a kitchen compliant.

Template, tool and hub

Put this into practice with Localyn

Localyn gives cafes, restaurants and takeaways a digital SFBB diary, temperature and cleaning records, allergen management and inspection-ready reports — all in one place.

View the demo Create free account All guides