Thermometer Calibration Checklist

  • Decide if you need to calibrate

  • Use lots and lots of ice for the ice check point, then use some more

  • Use an insulated container for the ice check point

  • Stir while checking

  • Don’t place the thermometer on the bottom of the vessel

  • Put the thermometer inside the kettle for the boiling water check, but watch for burns

  • Do both hot and cold checks

  • Take a photo of the results and write it down

  • Number your thermometers

  • Have a plan for thermometer failure

  • Validate your cooking method as a backup but beware of par-thawed goods

  • Give yourself some safety margin

  • Don’t forget to calibrate oven probes

  • Don’t forget to calibrate temperature recorders and air temperature recorders

  • Check calibration on infrared thermometers

Checking the Thermometer – Probe

Under the template food control plan rulebook, thermometers must be checked at least every 12 weeks to make sure that they are providing accurate temperature readings, or whenever there is reason to think the thermometer is not working correctly. Some people check more often and I know some verifiers are quite uncomfortable with the 12 week timeline – three months is a very l0ng time to be using an inaccurate thermometer.

Is Calibration Important?

Yes, absolutely, it is important. If you think your chicken is 75°C, and its really 60°C, you and others have some troubling times ahead.

This bit blows my mind though, and I am not the only one. One part of MPI treat it as vitally important, and another department seems not to.

In the original template food control plan guidance there was a really good form for recording calibration checks. In later editions of the MPI guidance, this form was removed. Verifiers, evaluators and consultants were left scratching our heads. We all know that calibration is important, so why did MPI remove this form? It becomes difficult to explain to the users of the system that this activity is really important when the people designing the system put so little focus on it.

On the other hand, if you are unfortunate enough to have gone through an ordeal with MPI Compliance, you will know they do not share this attitude about calibration. They are very, very focussed on this important task and are very highly skilled on the topic. You will be an expert on calibration by the time they are done with you.

Do you need to check calibration?

The first thing to ask is whether you need to calibrate? The answer is usually yes, but not always. For instance, I had one client who had a thermometer, but it was only for quality and production purposes, not food safety. They needed to heat up brown rice syrup to make it pourable, and melt chocolate. In both instances, there are no food safety implications and they needed little precision. They were about to get in trouble in an audit for not calibrating it regularly when I pointed out that it wasn’t important from a food safety perspective. I use one myself with an alarm as a level sensor, once again, no food safety implications.

If their thermometer failed it could make these processes a little more difficult for them. In this instance, quickly checking calibration in boiling water when in doubt might be a good idea (its faster than ice point check), but keeping records might be skipped.

How to do the ice point check?

This check must be done if the thermometer is used for checking cold foods. Personally, I always do this check no matter the purpose of the probe.

Below are some instructions I found somewhere. Personally, I find them a bit impractical.

  • 1. Half fill a glass with broken or shaved ice – you can scrape some ice from the side of a freezer.

  • 2. Add a small amount of water until it is visible at the bottom of the glass.

  • 3. Insert the thermometer into the mixture, leave until the temperature display is steady.

  • 4. Do not let the thermometer touch the sides or bottom of the glass.

  • 5. Record the result in the table below. If the result is outside the range, write down the action taken in the table.

Here is my revised instructions. This overcomes the problem of heat loss/gain.

  • 1. Go to Countdown and buy an insulated coffee cup (Countdown do the cheapest I can find at time of writing https://www.countdown.co.nz/shop/productdetails?stockcode=674236) – they cost $6 and will save you time. While you are there, buy an ice tray. Take it back and make a lot of ice. The more ice the better. You can also use a thermos flask, but it is more difficult to get the ice into the narrow opening. The important bit is that it is insulated.
  • 2. Give your thermometers a name, number of code so you can recognise them. Most kitchens have several thermometers.
  • 3. Fill your insulated cup with ice, as much as you can fit in.
  • 4. Add some water. Stick your probes in now to start to cool them. They can take a while to cool down.
  • 5. Leave it a while. Make a coffee so you have boiling water ready for the next step.
  • 6. Start to stir. This will achieve 2 things – the water will cool down, the heat transfer between the water and the probe will speed up.
  • 7. Don’t aim for perfection, aim for stability. If the temperature is still falling, then keep going. If the temp has stabilised, and it is within 1°C of 0°C, then you have reached your goal. For instance, if it reads 0.2°C and you can’t get it any further, then you have done your job.
  • 8. Record the reading, but make sure you add the decimals. If it is 0.1°C, write 0.1°C, not 0°C – no rounding! If a verifier comes along and sees a column of 0’s and column of 100’s, but can see you thermometer has 1 decimal place, they will know you have just made the numbers up (badly) the day before the audit.

How to do the boiling point check?

Some instructions state that the boiling point check should only be done if the thermometer is used for checking hot foods. Personally I always do this check for 2 reasons. Firstly, it gives a scientifically more valid result than just an ice point check, and secondly, its quick and easy.

  • 1. Boil unsalted water in a pot.
  • 2. Once boiling, insert thermometer and leave it until the temperature display is steady. If I am trying to be as accurate as I can, I leave the lid up on the kettle so that it runs continuously. If the water is actively boiling then you know you are getting 100C. Just be careful because steam burns.
  • 3. Do not let the thermometer touch the sides or bottom of the kettle as this may give a false reading.
  • 4. Record the reading, but, again, make sure you add the decimals.

It may be that your device cannot cope with 100°C and you will not be able to perform a boiling point check. For instance, the 1Buttton system will not cope with 100°C in water and neither will a lot of temperature recorders – especially if they are designed for recording air temperature rather than water temperature. Refer to manufacturers instructions in this circumstance.

Number or Name Your Thermometers

Most operations will have several thermometers, often more than they realise. Sometimes clients will produce a thermometer calibration record which will show regular and diligent calibration of a single thermometer, but they might have 3 of them lying around. One of them might be in a draw and only used if the main one can’t be found. It doesn’t matter if a thermometer isn’t used often. If it is used, or could be used, it should be checked for calibration.

Always name or number thermometers so your records are traceable to the correct unit. Thermometers can be purchased for just a few dollars these days, so if you have many of them, finding then checking calibration can be quite an exercise.

Have a Backup Plan

Lets say you calibrate your thermometer on 1st January and it shows 0C / 100C, but when you check calibration on 1st March it doesn’t show these values and it is clear that the thermometer has lost calibration. When you check something that is 75C, your thermometer should 76C. This means that when the faulty thermometer is reading 75C, the actual temperature is 74C and non compliant. How can you be confident that all the chicken you produced between 1st January and 1st March is safe?

I worked in a factory early in my career that checked the calibration of their metal detectors every 30 mins, and if one failed, they stopped producing and fixed the problem. They collected all the stock that went through that machine and ran it through again. They couldn’t be sure the product was free of metal (it was a process where there was a risk of metal), so they recalled and reworked all the product.

In a situation where you are only checking calibration every 3 months, how can you be sure the product you have produced is safe for the past 3 months. Unless you had a plan for this in advance, the answer is you probably can’t be sure.

  • Do double checks. For instance, use two thermometers for each batch.

  • Use an oven probe and manual probe. This is a variation on the above method. Sometimes it is very difficult to check the calibration of an oven probe, but these are typically very reliable. If you have this as a backup method, then you might need to check the oven calibration after the incident with the faulty thermometer.

  • Validate your cooking method. For instance, if you cook chicken at 180C for 60 mins every single time, and get very consistent results, then you might be able to be sure that the times the thermometer was faulty still resulted in safe food.

  • Give yourself a safety margin. For instance, some rotisserie chicken cookers aim for temperatures in the early 80’s, rather than the minimum of 75C.

Calibration Check of Oven Probes

In the first instance, check the equipment manual and speak to the supplier of the equipment. In my experience, they are often unhelpful here. Common responses are:

  • “You don’t need to check calibration on our equipment because it is so good”

  • “It can’t be checked”

  • “We’ve never been asked this before”

Verifiers and MPI will not accept any of this and neither should you.

In most cases it will just be a case of finding a way of doing an ice point check and boiling water check on this probe. Ice is probably easier than boiling water, but boiling water is more important. You may need to invest in an extension cord so you can get a kettle to the probe.

Calibration Check of Temperature Loggers – Probe Type

Lets say you have a device that has a probe which is attached to a recorder.

Some of these have a temperature read out and in this instance, it is easy, just check the calibration in the same way as any probe thermometer.

Others I have used do not have a read out. These get really tricky because you can’t see the temperature reading so don’t know if you have hit 0 / 100.

Firstly, change the recording interval to 1 second.

What you want to do is set the recorder going, then put it into an ice bath, then put it in boiling water. You then need to download the data to your computer to check the graph reached 0 and 100.

When calibrating a probe thermometer, it can take a long time to reach 0 / 100c – usually several minutes. If it takes 3 mins to reach 0.0C and you took it out of the ice bath after 2 mins, you will have missed your mark and will have to do the exercise again. This can get painful. This is why it is important to follow the instructions above for an ice point and use an insulated container.

Lastly, change the recording interval back to whatever it was to begin with e.g. 15s. If you keep it at 1s then you might run out of memory when it is needed.

Calibration of Air Temperature Recorders

Often the instructions will say they cannot be immersed in water (so ice point and boiling point are not suitable methods).

The first task is to check the accuracy of a probe thermometer. You then have to check the accuracy of the air temperature recorder against the probe thermometer. Often this means checking just one temperature point, not two (like ice/boiling for instance)

This sounds easy, but it often difficult in practice. One piece of advice is to do this in a very static temperature situation if you can.

If you want to do 2 temperature points, then you can choose room temperature / fridge temperature / freezer temperature.

With this, you might need a little bit of trial and error because, it really can be more difficult than it seems in practice.

Calibration of Infrared (IR) Thermometers

Either follow the calibration instructions that come with the thermometer or ask the business you bought it from for advice on when it should be calibrated, how this should be done, and who should do it.

  • The reading in iced water should be between –2°C to +2°C; if outside this range, the unit should be replaced or returned to the supplier to be recalibrated but this often can’t be done.

  • Infrared thermometers don’t work on shiny surfaces so it is worth laying a tissue over the ice slurry.

  • A boiling point calibration is not practical.

Follow the link below to download your own thermometer calibration form.

Thermometer calibration