Thermometer calibration: standards, accuracy limits and what NIST-traceable means

Thermometer calibration is the comparison of a thermometer’s readings with a reference of known temperature, usually a calibrated reference thermometer in a stirred bath or dry block, or a fixed point such as melting ice, at stated temperatures. The result is the thermometer’s error at each point and the uncertainty of that error (VIM 2.39). Calibration does not change the thermometer; correcting it is a separate step, adjustment (VIM 3.11). A one-point ice-bath check is useful, but a calibration covers at least two temperatures across the range you use, and its results hold only within that range (DKD-R 5-1, §1 and §5).

This page is the overview: thermometer types, the documents that apply, typical limits and intervals, and what “NIST traceable” on a thermometer really means. For the method itself, from the ice bath to a lab comparison, see how to calibrate a thermometer.

Types of thermometers and how each is calibrated

Type	Typical use	How it is checked or calibrated
Bimetal dial	Kitchens, food service	Ice-point check; many have an adjustment nut under the dial (USDA FSIS)
Digital, thermistor probe	Food, refrigerators, labs	Comparison in a bath or block; not all models can be adjusted (USDA FSIS)
Digital, thermocouple probe	Ovens, process, high temperatures	Comparison, from the highest temperature down; immersion depth matters (EURAMET cg-8, §10.1)
Platinum resistance (RTD, PRT)	Reference thermometers, process	Comparison with a reference PRT; ice-point checks between calibrations (DKD-R 5-1, §9)
Liquid-in-glass	Labs, older references	Ice-point checks plus comparison (NIST SP 1088, §4 and §7.6)
Infrared	Surfaces, non-contact	Against a blackbody or flat-plate radiation source, not in a bath (ASTM E2847, scope)

For a digital thermometer, the probe and the display are calibrated together: an electrical simulator only tests the display (EURAMET cg-11, §1.2).

Standards and guides that apply

Document	Scope	Access
FDA Food Code 2026, 4-203.11, 4-203.12, 4-502.11	Accuracy and calibration of thermometers in retail food and food service (a model code; states adopt it)	Free
USDA FSIS, Kitchen Thermometers	Ice-point and boiling-water checks for food thermometers	Free
NIST SP 1088 (2009)	Maintenance, validation and recalibration of liquid-in-glass thermometers	Free
DKD-R 5-1 (2023)	Calibration of resistance thermometers: comparison method, uncertainty, certificate	Free
EURAMET cg-8 (v3.1)	Calibration of thermocouples	Free
EURAMET cg-13 (v4.0)	Calibration of temperature block calibrators (dry blocks)	Free
EURAMET cg-11 (v2.0)	Temperature indicators and simulators, by electrical simulation	Free
ASTM E2847-21(2026)	Calibration and accuracy verification of wideband infrared thermometers	Paid

Typical accuracy limits and intervals

• Food thermometers: ±2 °F if scaled only in Fahrenheit, ±1 °C if scaled in Celsius or both, over the range of use; ambient air and water thermometers ±3 °F or ±1.5 °C (Food Code, 4-203.11 and 4-203.12). Calibrate them per the manufacturer as often as needed to keep them accurate (4-502.11).
• Liquid-in-glass thermometers: an ice-point check at least once a year, more often when new or used hot (NIST SP 1088, §4 and §7.6).
• Reference PRTs: DKD recommends recalibration at most every 2 to 3 years for SPRTs and every 1 to 2 years for industrial PRTs, with ice-point checks in between (DKD-R 5-1, §9, Table 1).
• Thermocouples: no fixed frequency; EURAMET recommends a recalibration program built on your own experience (EURAMET cg-8, §13).

Then let your results move the interval: see calibration intervals.

What a NIST-traceable thermometer really means

“NIST traceable” on a box or a catalog page is a claim that the thermometer’s calibration results can be related to a reference, here the SI units as realized by NIST, through a documented, unbroken chain of calibrations, each adding to the uncertainty (NIST traceability policy, item 1; VIM 2.41). Four things follow:

• Traceability belongs to a result, not to the thermometer. What you actually buy is a calibration certificate with results (NIST FAQ, 5.1.3).
• NIST does not vouch for it. NIST does not certify the traceability of results it did not provide; supporting the claim is the supplier’s job, and judging it is yours (NIST policy, items 3 and 4). A NIST test report number on a certificate is not proof of traceability (NIST FAQ, 5.2.4).
• Traceable is not the same as accurate enough. The uncertainty also has to be small enough for your use (NIST policy, item 5). Check it against your tolerance with the TUR calculator.
• NIST rarely calibrates your thermometer itself. Its industrial thermometer lab calibrates platinum resistance, thermistor and thermocouple probes and digital readouts with probes from −196 °C to 550 °C, by comparison with SPRTs calibrated on ITS-90, and no longer calibrates liquid-in-glass thermometers (NIST, Industrial Thermometer Calibrations). Otherwise, the chain runs through other labs:

LevelWhat you can check

1. Reference

   The SI unit of temperature (kelvin)

   Traceability runs to a reference, normally the SI units.

   What you can check

   The traceability statement points to the SI, for example "traceable to the SI through NIST".

2. National metrology institute

   NIST, or another national institute

   Calibrates reference thermometers by comparison with its own SPRTs, calibrated on ITS-90.

   What you can check

   Only results NIST issues itself come from NIST. It does not certify anyone else's claim.

3. Calibration lab

   A lab accredited to ISO/IEC 17025

   Calibrates your thermometer, or your reference thermometer, against its own calibrated references.

   What you can check

   The accreditation symbol and number on the certificate, and temperature in the lab's scope.

4. Your thermometer

   Its calibration certificate

   Results at stated temperatures, with their uncertainty.

   What you can check

   Readings and errors at the temperatures you use, the uncertainty, and results before and after any adjustment.

An accredited certificate, with the accreditation symbol, is the form accreditation policy treats as demonstrating traceability (ILAC P10, Note 3; ANAB AR 2251, §2.1.2). Our guide to reading a calibration certificate shows how to check each of these points, and what calibration is covers the basics.

FAQ

What is the difference between NIST traceable and NIST certified?

On a thermometer from a commercial supplier, both phrases usually stand for the same claim: that its calibration results link to NIST standards through a documented chain of calibrations. Neither means NIST tested the thermometer or approved the claim. NIST says it does not certify the traceability of results it did not provide itself, and that supporting the claim is the supplier's job and judging it is yours. Only a calibration report issued by NIST comes from NIST.

How long is a NIST-traceable thermometer good for?

The certificate does not expire: it reports what the thermometer did on the calibration date. How long you can rely on it is your decision, based on how the thermometer drifts, how it is used and what a wrong reading would cost. Labs accredited to ISO/IEC 17025 do not put a recommended recalibration date on the certificate unless you asked for one or a regulation requires it. Ice-point checks between calibrations tell you early if a thermometer has shifted.

Do I need an accredited calibration, or is a traceable certificate enough?

For a kitchen or a food-service thermometer, a regular ice-point check against the Food Code limit is the usual practice. For a quality system, a lab or an audit, a certificate from a lab accredited to ISO/IEC 17025, with the accreditation symbol and the calibration inside the lab's scope, is the form that accreditation policy treats as demonstrating traceability. A non-accredited certificate can be acceptable, but you then need other evidence of the lab's competence and traceability.