A thermometer calibration record (or the certificate a lab gives you) has to show seven things:
- Which thermometer: asset ID, the serial numbers of the display and the probe, type, range and resolution.
- The tolerance and its source: for food thermometers, ±1 °C, or ±2 °F if they read only in Fahrenheit, across the range of use (FDA Food Code, 4-203.11).
- The reference and the method: the calibrated reference thermometer and its certificate, the bath or dry block, and your procedure.
- The conditions: immersion depth, stabilization time and the immersion check.
- The readings: at temperatures covering the range you use, lowest and highest included, with 0 °C if it is in the range, as found and, after any adjustment, as left.
- The uncertainty at each point and the decision rule for pass or fail.
- The decision: in or out of tolerance, what was done, who did it and reviewed it, and the next due date.
The free template below has a field for each of these, and under it there is a one-page log for daily ice-point checks. Print either one from this page: no sign-up needed. Then comes the same record filled in with a worked example, and what to look for on a lab certificate.
Thermometer calibration record template
Thermometer calibration record
One sheet per thermometer, probe and display together. Comparison with a calibrated reference thermometer in a stirred bath or dry block; temperatures in the units ticked below.
Use one sheet per thermometer, with the probe and display calibrated together. For a thermometer calibrated by an outside lab, copy the key results from the certificate, record your own decision on the last block and attach the certificate. For in-house calibration, the sheet is the record. The test points, readings and immersion check follow our thermometer calibration procedure. An infrared thermometer needs a radiation source instead of a bath; see the FAQ below.
Thermometer calibration log for ice-point checks
Most kitchens and food plants don’t calibrate thermometers every week: they check them in ice water and write it down. This log is for that, with many thermometers on one sheet.
Thermometer calibration log (ice-point checks)
Crushed ice packed to the top, water filling the gaps, stirred; probe in the middle at least 2 in deep, touching nothing; read after at least 30 s. Expected: 32 °F (0 °C). Food thermometers: within ±2 °F (°F only) or ±1 °C (°C or both scales).
Make the ice bath the USDA way: finely crushed ice packed to the top, clean water filling the gaps, stirred, with the probe at least 2 inches deep in the middle, touching neither the sides nor the bottom, and read after at least 30 seconds (USDA FSIS). A refrigerator or freezer thermometer that measures air is held to ±1.5 °C or ±3 °F instead (Food Code, 4-203.12). Keep in mind what the log can’t show: a thermometer that passes at 32 °F can still be off at 165 °F. Our procedure page covers the boiling-water check and its altitude correction.
Example: a filled-in thermometer calibration record
The record below is for the same fictitious food-plant thermometer as the worked example on our procedure page: same readings, and every mean and error is calculated from them. The company, people, numbers and uncertainties are invented. Click a number to jump to its note.
Thermometer calibration recordExample
One sheet per thermometer, probe and display together. Comparison with a calibrated reference thermometer in a stirred bath or dry block; temperatures in the units ticked below.
Probe and display together
The asset ID ties the record, the label and any certificate together. Write down the serial numbers of the display and the probe: a calibration is of the pair, and a calibrator that simulates the sensor signal tests only the display (EURAMET cg-11, 1.2). TH-117 is a digital thermometer with a Pt100 probe, −50 to 150 °C, used from −20 to 100 °C.
The tolerance and where it comes from
±1 °C because it is a food thermometer that reads in Celsius (FDA Food Code 2026, 4-203.11(A)); one scaled only in Fahrenheit would be held to ±2 °F. The limit applies across the range of use, so it is the same at −20 °C and at 100 °C (Food Code, Annex 3). Write the source next to the number.
Reference, source and conditions
A calibrated reference PRT in the same stirred bath, compared with the thermometer once both are stable (DKD-R 5-1, 4 and 7). The ice-point reading before the calibration is the first sign of a shift since last time, and DKD asks for it to be stated separately on the certificate (DKD-R 5-1, 5–6). The expanded uncertainty of 0.10 to 0.15 °C is well under a quarter of the ±1 °C tolerance.
Immersion depth, and the check
At least 10 times the probe diameter from the top of the sensing element as a first estimate: 120 mm for a 4 mm probe. Withdrawing the probe slightly changed the reading by 0.01 °C, under 10 % of the uncertainty aimed for, so heat conduction along the stem is not a problem (DKD-R 5-1, 8.3). DKD asks for the depth on the certificate so the calibration can be repeated.
As found: passes at 0 °C, fails at 100 °C
Four temperatures from the lowest to the highest of use, three readings each. The error grows with temperature: +0.13 °C at the ice point, +0.63 °C at 50 °C and +1.23 °C at 100 °C, where all three readings are above the limit. Even after subtracting the 0.15 °C uncertainty, the 100 °C result is outside. An ice-point check alone would have passed this thermometer.
As left: a separate set of readings
After a two-point adjustment in the thermometer's calibration menu, every point is measured again and recorded as a new set; the largest error is now +0.3 °C. Results before and after adjustment both go on the record (DKD-R 5-1, 10; VIM 3.11). If nothing is adjusted, tick "No" and write "as found = as left".
Decision, next due and sign-off
Simple acceptance, with the uncertainty reported (name the rule: Global ACI-TECH-1-002, formerly ILAC G8). The as-found failure opens an out-of-tolerance report: what was measured above about 80 °C with this thermometer since its last good calibration? See out-of-tolerance results. Weekly ice-point checks go on the log below until the next calibration.
The error on this form is the reading minus the reference temperature, so a positive error means the thermometer reads high, as on our procedure page. To use a thermometer with a known error, you subtract it (USDA FSIS), but only between the temperatures you calibrated, because results are valid only within that range (DKD-R 5-1, §1).
What a thermometer calibration certificate should include
ISO/IEC 17025 sets the minimum content of any calibration certificate from an accredited lab. Clause 7.8.2.1 covers the general items (a unique identification, the customer, the method, the item calibrated, the dates, the results and who authorized the report), and clause 7.8.4.1 adds the calibration items: measurement uncertainty, the conditions that affect the results, a traceability statement, results before and after any adjustment or repair when available, and a statement of conformity when one is given (NIST’s crosswalk of clause 7.8). DKD’s guideline for resistance thermometers adds what is specific to temperature (DKD-R 5-1, §8.3 and §10):
| On a lab’s thermometer certificate | Where it goes on the template |
|---|---|
| Thermometer and probe identification, range, resolution | Thermometer block |
| The method or procedure | Method |
| Reference thermometer, readout and their traceability | Reference thermometer and readout ID, reference cert. no. |
| Bath or dry block, and the insert | Temperature source |
| Immersion depth | Immersion depth |
| Reference temperature, reading and error at each point, before and after adjustment | As found and as left tables |
| Expanded uncertainty at each point; for resistance thermometers, measuring current, self-heating and hysteresis | U column (add the rest under Comments or attach the certificate) |
| Pass or fail, the specification and the decision rule | Tolerance, decision rule, decision |
Red flags on a thermometer certificate
- One temperature. DKD asks for at least two, and the results are valid only between the lowest and highest calibrated temperatures (DKD-R 5-1, §1 and §5). A certificate at 0 °C says nothing about 75 °C.
- Temperatures outside your range of use. A calibration at 0, 50 and 100 °C does not cover a freezer thermometer used at −20 °C.
- The display calibrated without its probe. Electrical simulation tests the indicator only (EURAMET cg-11, §1.2). If your thermometer has a probe, the certificate should say the probe was in the bath or block.
- No immersion depth. Without it, the calibration can’t be repeated, and a short probe may have been reading the stem’s heat (DKD-R 5-1, §8.3).
- Only as-left results. If the lab adjusted the thermometer and reports only the results after adjustment, you have lost the as-found data you need to judge past readings (ISO/IEC 17025, clause 7.8.4.1; DKD-R 5-1, §10). Ask for both when you order.
- No uncertainty. An accredited calibration certificate states it at each point (clause 7.8.4.1).
- “Pass” with no tolerance or decision rule. A statement of conformity must say which specification was used and how uncertainty was taken into account (clause 7.8.6, as described in Global ACI-TECH-1-002).
- “NIST traceable” and nothing else. NIST does not certify other organizations’ traceability claims; the supplier has to support the claim and the user has to judge it (NIST traceability policy). Look for results, uncertainties and the reference used.
Our calibration certificate guide covers accreditation symbols, scopes and the full review checklist.
How to fill in the template
| Field | What to write |
|---|---|
| Serial no. | Both the display and the probe. If you swap the probe, the calibration no longer applies to the pair |
| Range of use | The lowest and highest temperatures the thermometer actually measures in your process. They set the test points |
| Tolerance | Your specification, the manufacturer’s, or for food thermometers the Food Code limit (4-203.11). Write the source |
| Reference | ID and certificate of the reference thermometer and its readout. Its uncertainty should be a small fraction of the tolerance: about a quarter or better is a common target. Check it with the TUR calculator |
| Temperature source | Bath or dry block. In a dry block, use an insert no more than 0.5 mm wider than the probe (up to 660 °C), and read an external reference thermometer in the block: the block’s calibration relates to the block, and your probe can be at a different temperature (EURAMET cg-13, §2.2 and Annex C; DKD-R 5-1, A.3) |
| Ice-point reading | The thermometer in a proper ice point before you start. A change since the last calibration is an early sign of damage, and DKD asks for this initial value on the certificate (DKD-R 5-1, §5–6) |
| Immersion depth and check | At least 10 × the probe diameter as a first estimate; then withdraw the probe slightly and write the change. It should be under 10 % of the uncertainty you aim for (DKD-R 5-1, §8.3) |
| Test points | The lowest and highest temperatures of use, at least one in between, and 0 °C if it is in the range (DKD-R 5-1, §5; EURAMET cg-8, §12.6). Calibrate thermocouples from the highest temperature down (EURAMET cg-8, §10.1) |
| R1 to R3 | Readings after both thermometers are stable, taken at the same time as the reference (DKD-R 5-1, §7). No free standard fixes the number: the example uses three; DKD’s worked example averages 20 |
| Errors | Reading − reference. The largest one decides pass or fail, because the tolerance applies to every reading |
| U | From the lab certificate or your own budget (try the uncertainty calculator) |
Where to keep it and for how long
File each record under the thermometer’s asset ID, with the lab certificate attached, and keep the ice-point logs for the same thermometers alongside: together they show whether the thermometer drifts between calibrations, which is what lets you lengthen or shorten the interval (see calibration intervals). How long to keep them depends on the standards and regulations you work under; our calibration records guide sets out the rules by standard. Print a matching status sticker with the calibration label generator.
Get the PDF pack
The templates on this page are free to print, as they are. If you’d rather have the calibration record as a PDF, subscribe to Traceable, our newsletter, with the form on this page: subscribers get the printable pack with our calibration program starter kit and the instrument record templates, this one included.
Related:
- How to calibrate a thermometer: the method this record follows, with the ice-point and boiling-water checks.
- Thermometer calibration hub: standards, accuracy limits and what NIST traceable means.
- How to read a calibration certificate.
- Out-of-tolerance results: what to do when the as-found result fails.
- TUR calculator and uncertainty calculator.
FAQ
What should a thermometer calibration certificate include?
The thermometer and probe identification, the method, the reference thermometer and its traceability, the bath or block used, the immersion depth, the results at each calibrated temperature (reference temperature, reading and error), results before and after any adjustment, the measurement uncertainty at each point, and, if the certificate says pass or fail, the tolerance and the decision rule. It also needs a unique number, the dates and the name of the person who authorized it. For resistance thermometers, DKD's guideline adds the measuring current, self-heating and hysteresis.
What temperature should a thermometer be calibrated at?
For a quick check, at the ice point: a properly made ice-water slurry, where the thermometer should read 32 °F (0 °C). For a calibration, at several temperatures covering the range you actually use, including the lowest and highest, with at least one point in between and 0 °C when it is in the range. A calibration is valid only between the lowest and highest temperatures tested.
How often should a thermometer be calibrated?
There is no single interval. The FDA Food Code says food thermometers must be calibrated per the manufacturer's specifications as often as needed to keep them accurate, which in practice means regular ice-point checks and a check after any drop. NIST recommends checking liquid-in-glass thermometers at the ice point at least once a year. For industrial platinum thermometers used as reference standards, DKD recommends recalibration at most every one to two years, depending on their history.
What is the difference between a thermometer calibration log and a calibration record?
A log is a running list of quick checks, usually at the ice point, with many thermometers on one sheet: date, thermometer, reading, pass or fail, action and initials. A calibration record documents one full calibration of one thermometer: several temperatures against a calibrated reference, every reading, as-found and as-left results, the uncertainty and the decision. Most food businesses need the log; quality systems and labs need both.
My thermometer came with a NIST-traceable certificate. Is that enough?
Check what is on it. NIST does not certify other organizations' traceability claims: the supplier has to support the claim and you have to judge it. A useful certificate shows the results at stated temperatures, the uncertainty at each one and the reference used, ideally from a lab accredited to ISO/IEC 17025 for that calibration. A paper that only says 'NIST traceable' tells you little.
Can I use this template for an infrared thermometer?
Use the same structure, with a different setup: an infrared thermometer is calibrated against a radiation source (a blackbody or flat-plate calibrator), not in a bath. Record the source and its temperature, the emissivity setting, the distance and alignment, and the test points (one for a narrow range, three for a wide one). ASTM E2847 is the US test method for handheld wideband infrared thermometers.
What tolerance applies to a refrigerator or freezer thermometer?
Under the FDA Food Code, thermometers that measure ambient air or water, such as the one in a refrigerator, must be accurate to ±1.5 °C, or ±3 °F if they read only in Fahrenheit (4-203.12). Food thermometers are held to ±1 °C or ±2 °F (4-203.11). Your state or local code decides which edition of the Food Code applies.
Sources
- U.S. Food and Drug Administration. FDA Food Code 2026 (4-203.11, 4-203.12, 4-502.11 and Annex 3). 2026 — Published September 17, 2026; pp. 91, 106
- USDA Food Safety and Inspection Service. Kitchen Thermometers (section: Calibrating a Thermometer) — Read via the Internet Archive copy of 2025-03-11 (the live page blocks automated access)
- PTB / DKD. DKD-R 5-1, Calibration of resistance thermometers. 2023 — §1, §4–§7, §8.3, §9 (Table 1), §10, Appendix A.3
- EURAMET. EURAMET cg-8, Guidelines on the Calibration of Thermocouples (v3.1). 2020 — §10.1, §12.6
- EURAMET. EURAMET cg-11, Guidelines on the Calibration of Temperature Indicators and Simulators by Electrical Simulation and Measurement (v2.0). 2011 — §1.2
- EURAMET. EURAMET cg-13, Guidelines on the Calibration of Temperature Block Calibrators (v4.0). 2017 — §2.2, Annex C
- NIST (Cross, Miller, Ripple, Strouse). NIST SP 1088, Maintenance, Validation, and Recalibration of Liquid-in-Glass Thermometers. 2009 — §4, §7.6
- ASTM International. ASTM E2847-21(2026), Standard Test Method for Calibration and Accuracy Verification of Wideband Infrared Thermometers. 2026 — Paid standard; scope only
- Fluke Calibration. Infrared Thermometer Calibration – A Complete Guide
- NIST Office of Weights and Measures. ISO/IEC 17025 Crosswalk – Reporting the Results (clause 7.8). 2018 — Free summary of ISO/IEC 17025:2017 clause 7.8 (7.8.2.1, 7.8.4.1, 7.8.6)
- NIST. NIST Policy on Metrological Traceability. 2024
- Global Accreditation Cooperation (Global ACI). Global ACI-TECH-1-002 (G), Guidelines on Decision Rules and Statements of Conformity (formerly ILAC G8:09/2019). 2026
- JCGM / BIPM. International Vocabulary of Metrology (VIM), JCGM 200:2012. 2012 — 2.39 calibration, 3.11 adjustment