How I Set Up the Food Probe Accuracy Test Template with Customers in Pilla

Understanding What's Required of You

Probe accuracy testing isn't explicitly named in law, but it sits squarely inside your obligations under Regulation (EC) 852/2004. The regulation requires food business operators to maintain effective temperature monitoring and keep records that prove their controls work. A probe that reads 2°C off doesn't just give you a bad number. It undermines every temperature record you've taken with it: fridge checks, cooked food checks, cooling logs, delivery temperatures. Your EHO won't treat a drifted probe as a minor issue. They'll treat it as a gap in your monitoring system.

The standard method is a two-point test at 0°C and 100°C. These are physical constants: water freezes at 0°C and boils at 100°C at sea level. Testing at both ends confirms accuracy across the full range you'll use in a kitchen. A probe can read correctly at one point and drift at the other, which is why a single-point check isn't enough.

The ±1°C tolerance is the accepted standard for food safety purposes. Tighter than that is difficult to achieve in a working kitchen. Wider than that and your other temperature checks become unreliable. If your probe reads 1.5°C high, that means a fridge you think is at 5°C is actually at 6.5°C, which is getting close to the legal maximum of 8°C.

Your EHO expects to see weekly probe accuracy records at a minimum. In busy kitchens with heavy probe use, more often is better. They'll also want to see what you did when a probe failed: did you adjust it, note the offset, or replace it? A test result showing a 2.5°C drift with no follow-up action recorded is worse than not testing at all, because it shows you knew there was a problem and carried on.

You should also test after dropping a probe, after any physical damage, after battery replacement on digital probes, and whenever readings seem inconsistent. Probes get knocked around in kitchens. They fall off prep benches, get thrown into washing-up bowls, sit in pockets with keys. The sensor at the tip is the most delicate part and it doesn't take much to shift its calibration.

Setting It Up as a Work Activity

I've built a probe accuracy test template in Pilla covering probe identification, the ice water test reading, the boiling water test reading, the pass/fail result, and a notes field for recording issues and corrective actions. It gives you a structured starting point that captures what your EHO wants to see.

Create a weekly recurring work activity and assign it to the person responsible for your temperature monitoring equipment. Tag it with "Food Safety Checks" so it groups with your fridge checks, cooked food checks, and other monitoring tasks. Poppi can then track them as a set.

1. Probe identifier

Record which probe you're testing. If you have multiple probes, each one needs its own test.

Why it matters:

Most kitchens use more than one probe. Without clear identification, you can't track which probes are accurate and which are drifting. If a probe fails the test, you need to know which one to pull out of service. And if your EHO asks to see the calibration history for the probe you're currently using, you need records that match.

What good answers look like:

• "Probe 1 (blue handle, digital)"
• "Kitchen probe, ThermaLite"
• "Probe A, raw foods"
• "Bar probe, serial #4821"

Use whatever identification makes sense for your operation, but be consistent. The name in Pilla should match a label on the physical probe. If you label your probes with coloured tape or numbered stickers, reference those.

Common mistakes (and how to avoid them):

"The probe" tells you nothing if you have three of them. Be specific.

Changing the naming between tests makes it impossible to track history. Pick a convention and stick with it. When new staff fill in the check, they need to know which probe is which without guessing.

Not labelling the physical probes. If the probes in the drawer all look identical, people will grab whichever is closest and assign it whatever name comes to mind. Label the probes themselves, not just the template.

Best practices to follow:

• Label each probe with a number, colour, or identifier that matches your template
• If probes are dedicated to specific tasks (raw food, cooked food, fridge checks), include that in the name
• Keep a simple register of your probes showing serial numbers, purchase dates, and when each was last tested
• When you replace a probe, update the template to reflect the new unit

2. Ice water test (0°C)

This records the temperature your probe reads when placed in a properly prepared ice water bath.

Why it matters:

The ice water test gives you a known reference point at the bottom of your working range. If your probe reads correctly at 0°C, you can have confidence in your fridge temperature checks and cold storage monitoring. A probe that reads high at 0°C will understate how cold your fridge actually is, or overstate it, depending on the direction of the drift.

How to prepare the ice bath correctly:

This is where most people get it wrong. A container with a few ice cubes floating in water won't give you 0°C. The water between the cubes can sit at 2 to 4°C, and your test becomes meaningless.

1. Crush the ice. Use crushed ice, not cubes. Crushed ice has more surface contact with water and chills it more evenly
2. Pack a deep container at least three-quarters full with crushed ice
3. Add just enough cold water to fill the gaps between ice pieces. The ice should still be packed tightly, not floating
4. Stir the mixture and wait one to two minutes for the temperature to stabilise
5. Keep adding ice as it melts. If the ice drops below the waterline, your temperature rises

Insert the probe into the centre of the ice slush, making sure the sensor tip is fully surrounded. Wait 30 to 60 seconds for the reading to stabilise, then record it.

What good answers look like:

A number between -1°C and +1°C. Record the actual reading, not the target. If your probe reads 0.3°C, write 0.3°C. If it reads -0.8°C, write -0.8°C.

Common mistakes (and how to avoid them):

Too much water. If ice is floating rather than packed, the water temperature is above 0°C and the test is invalid. You're testing the probe against the wrong reference point.

Using ice cubes instead of crushed ice. Large cubes leave big gaps of warmer water. Crush the ice first. A Lewis bag and a rolling pin works if you don't have an ice machine that produces crushed ice.

Not waiting for the reading to stabilise. The probe needs time to reach the temperature of the ice water. Pulling it out after ten seconds gives you an inaccurate reading, usually higher than the actual temperature.

Testing the probe against the side of the container. The container wall is warmer than the ice slush. Keep the probe tip in the centre, away from all surfaces.

Best practices to follow:

• Prepare the ice bath before you get the probe, so it has time to stabilise
• Use a deep container so the probe tip is fully submerged in ice slush
• Record the reading to one decimal place if your probe displays it
• If the reading is outside ±1°C, note it and move on to the boiling water test before deciding on action

3. Boiling water test (100°C)

This records the temperature your probe reads when placed in actively boiling water.

Why it matters:

The boiling water test gives you a known reference point at the top of your working range. This end matters for cooked food checks, where you're verifying that food has reached 75°C at the core. A probe that reads low at 100°C could be telling you food is safe when it's not.

How to prepare the boiling water test:

1. Boil fresh water in a kettle or pan. Bring it to a rolling boil, not just simmering
2. Test immediately. Water cools quickly once it stops boiling. Keep it on the heat if using a pan, or test within seconds of the kettle clicking off
3. Insert the probe into the water, keeping the sensor tip away from the sides and bottom of the container. The bottom of a pan can be hotter than 100°C, and the sides transfer heat differently
4. Wait for the reading to stabilise, then record it

One thing to know: water boils at 100°C at sea level. At higher altitudes, the boiling point drops because of lower air pressure. In most of the UK this makes no practical difference. But if you're at a highland venue above 500m, your boiling point might be 98 to 99°C, so factor that in before failing a probe that reads 99°C.

What good answers look like:

A number between 99°C and 101°C. Record the actual reading. If your probe shows 100.4°C, write 100.4°C.

Common mistakes (and how to avoid them):

Testing water that's stopped boiling. Once you take the kettle off the heat, the temperature starts dropping immediately. By the time you pour it into a container and insert the probe, it could already be 95°C. Test in the kettle or keep the pan on the hob at a rolling boil.

Touching the bottom or sides of the pan with the probe tip. The pan surfaces can be hotter than the water, giving you a falsely high reading.

Not being careful with the boiling water. This is obvious, but burns from probe accuracy testing do happen, usually because someone is rushing. Take your time. Handle the probe carefully and keep your hands clear.

Best practices to follow:

• Keep the water at a rolling boil during the test
• Hold the probe in the centre of the water, away from all surfaces
• Record the reading to one decimal place if available
• If testing multiple probes, reboil between tests rather than using the same water that's been cooling

4. Test result

After completing both tests, record whether the probe passed or failed.

Why it matters:

This creates a clear record of whether your probe is fit for use. Over time, it shows whether a probe is stable or gradually drifting. A probe that passes consistently is reliable. A probe that passes one month and fails the next needs attention, even if you manage to calibrate it back.

What good answers look like:

Pass: Both readings within ±1°C. The probe is accurate and can continue to be used for temperature monitoring.

Fail: One or both readings outside ±1°C. The probe needs calibration, an offset note, or replacement before being used for food safety checks.

Be honest with this. I've seen records where a probe read 2.3°C high at 100°C and someone still recorded a pass. That's not a pass. If it's outside ±1°C at either test point, it's a fail.

How to answer this for yourself:

If both readings are within tolerance, it's a pass. If either reading is outside ±1°C, it's a fail. There's no grey area.

The harder question is what to do after a fail. You have three options:

Adjust the probe. Some probe thermometers have a calibration screw or button. Insert the probe in ice water, wait for it to stabilise, and adjust until it reads 0°C. Then verify at 100°C to check the adjustment holds at both ends. Not all probes have this feature.

Note the offset. If a probe consistently reads a fixed amount high or low, you can continue using it if you apply a mental adjustment. A probe that reads +1.5°C high means a reading of 76.5°C is actually 75°C. This is a temporary measure. Replace the probe when you can.

Replace the probe. If the probe is more than 2°C off at either point, gives inconsistent readings, or has visible damage, replace it. Probe thermometers cost between five and twenty pounds. Compared to the cost of a food safety incident or a poor hygiene rating, that's nothing.

Common mistakes (and how to avoid them):

Recording a pass when readings are borderline outside tolerance. ±1°C means ±1°C. A reading of 1.2°C off is a fail.

Failing the test but continuing to use the probe without doing anything about it. A fail without corrective action is worse than not testing at all, because it shows you identified the problem and ignored it. Your EHO will pick up on that immediately.

Not testing at both points. A probe can be accurate at 0°C and drift at 100°C. Both tests are required for a valid result.

Best practices to follow:

• Record the result immediately after testing
• If the probe fails, record what action you're taking in the notes field
• Keep replacement probes in stock so a failed probe doesn't leave you without equipment
• Track pass/fail history for each probe to spot ones that are deteriorating

5. Notes

Use this field to record any observations, issues, corrective actions, or follow-up needed.

Why it matters:

The pass/fail result tells you the outcome. The notes tell the story. If a probe failed, what exactly was wrong? What did you do about it? If it passed but you noticed something, this is where you flag it. Your EHO will read the notes when they review your records, and "Probe replaced, new unit tested and passed" tells them far more than a blank field next to a fail result.

What good answers look like:

For a straightforward pass:

• "Both readings within tolerance. No issues."
• "0°C test: 0.2°C. 100°C test: 99.8°C. Probe in good condition."

For a pass with observations:

• "Readings within tolerance but probe is responding slower than usual, taking over 60 seconds to stabilise. Will monitor next week and replace if it gets worse."
• "Pass. Battery indicator showing low. Replaced batteries after test."

For a fail:

• "Ice water test: 0.4°C (pass). Boiling water test: 102.3°C (fail). Probe reads 2.3°C high at top end. Adjusted calibration screw, retested at 100.1°C. Passed on retest."
• "Both readings outside tolerance. 0°C test: -2.1°C. 100°C test: 97.8°C. Probe withdrawn from service, replacement ordered. Using backup probe (Probe 2) until new unit arrives, Probe 2 tested and passed."
• "Failed both tests. Visible corrosion on probe tip. Discarded. New probe purchased same day, serial #4922, tested and passed."

Common mistakes (and how to avoid them):

Leaving notes blank on a fail. A fail with no notes is a red flag for your EHO. It looks like the problem was identified and nothing was done.

Vague notes like "probe not working." What specifically was wrong? How far off were the readings? What did you do about it?

Not recording the corrective action. "Failed" on its own is incomplete. "Failed, adjusted calibration, retested and passed" closes the loop.

Not recording when replacement probes were tested. If you replace a failed probe, test the new one and record the result. Don't assume a new probe is accurate out of the box.

Best practices to follow:

• Write something for every test, even a simple "No issues" for a clean pass
• Include both readings in the notes so there's a full record even if the summary just says pass or fail
• Name the corrective action and when it was taken
• If you need to follow up (e.g., replacement arriving next week), note that and check back

Automate the Follow-Up with Poppi

We're still finalising the best automation setup for food safety checks. Once that's ready, this section will show you how to use Poppi to chase overdue tests and flag failed probes so you can act on them before your next EHO visit.