Why is 20 °C the reference temperature for dimensional measurement?

Because length changes with temperature, a specified dimension is only meaningful at a stated temperature, and international dimensional standards fix that temperature at 20 °C. A part measured at 25 °C is genuinely a slightly different size from the same part at 20 °C. For ordinary tolerances this does not matter; for anything toleranced in micrometres it is the dominant consideration, and it is why precision dimensional laboratories are temperature-controlled rather than merely comfortable.

How much does temperature affect a steel measurement?

Steel expands by roughly 11.5 micrometres per metre per degree Celsius. On a 100 mm dimension that is about 1.15 micrometres for every 1 °C away from 20 °C, so a workshop at 25 °C puts a steel part about 5.75 micrometres over its 20 °C size. If the instrument is also steel and at the same temperature the two expansions largely cancel; measuring aluminium with a steel instrument, they do not.

Why does a micrometer have a ratchet?

To make the measuring force repeatable. Tightening a micrometer harder gives a smaller reading, because the part and the frame both deform under the load — so a reading without a controlled force is a reading at whatever force that operator applied that day. The ratchet or friction thimble is the mechanism that removes the operator from the result, and bypassing it by turning the thimble directly throws away most of what makes a micrometer better than a caliper.

How accurate is a digital caliper really?

Far less accurate than its display suggests. A caliper resolving to 0.01 mm typically has a real measurement uncertainty in normal use of several hundredths of a millimetre, because of jaw parallelism, play in the slide, the force applied and the angle it is held at. It is an excellent instrument for what it is — quick, versatile, robust — and the mistake is asking it to police a tolerance it was never capable of policing.

Should I use the deviation on a gauge block certificate?

Yes. A block certified at 25.0001 mm is a better reference when used as 25.0001 mm than when treated as exactly 25 mm — you have paid for the information, and ignoring it discards the accuracy you bought. Using nominal values is acceptable only when the certified deviations are genuinely negligible against the uncertainty you are claiming, and that is a calculation rather than an assumption.

Dimensional calibration: calipers, micrometers and gauge blocks

Why 20 °C is written into dimensional standards, what thermal expansion does to a steel measurement, and how calipers and micrometers are actually checked.

Calibration by Parameter ValiTrac AI editorialUpdated 2026-09-133 min read

ShareLinkedIn X Email

Dimensional calibration has one feature no other parameter shares quite so sharply: the reference temperature is written into the definition. Lengths are specified at 20 °C, and a measurement made anywhere else is a measurement of something that is not quite the size it is specified to be.

Why 20 °C, and what happens away from it

Materials expand. Steel moves by roughly 11.5 micrometres per metre for each degree; aluminium by about 23. Over a 100 mm dimension, a workshop at 25 °C rather than 20 °C makes a steel part about 0.6 micrometres longer than its specified size. For most work that is irrelevant. For gauge blocks, precision bores and anything toleranced in single micrometres, it is the measurement.

Thermal expansion over a 100 mm steel dimension
Temperature	Departure from 20 °C	Change in length
20 °C	0 °C	0.00 µm
22 °C	2 °C	2.3 µm
25 °C	5 °C	5.75 µm
30 °C	10 °C	11.5 µm

Thermal expansion over a 100 mm steel dimension

The saving grace is that if the instrument and the part are the same material at the same temperature, the two expansions largely cancel — a steel micrometer measuring a steel part at 25 °C is closer to right than the table suggests. The danger is measuring aluminium with a steel instrument, where they do not cancel at all, or measuring a part that has just come off a machine and is still warm.

Handling is part of the measurement

Holding a gauge block in bare fingers warms it. A 100 mm block held for a minute can grow by a micrometre, which for a grade-K block is the whole tolerance. This is why gauge blocks are handled with gloves or tongs, why they are allowed to stabilise on the surface plate before use, and why a hurried measurement is a worse measurement.

Calipers

Check zero with the jaws closed, and check it again after the rest of the calibration — a caliper that no longer zeroes has been mishandled.
Measure gauge blocks at several points across the range, on the outside jaws.
Check the inside jaws separately; they are ground differently and often disagree with the outside jaws.
Check the depth rod, which is usually the worst part of the instrument and the least often tested.
Check parallelism of the jaws — measure the same block near the tips and near the base of the jaws.
Look at the slide for play; a caliper that rocks measures whatever angle you happen to hold it at.

A caliper is a convenience instrument. Expecting micrometre performance from one is a category error — it resolves to 0.01 mm and its real uncertainty in a normal hand is several times that.

Micrometers

A micrometer is a far better instrument and has a complication calipers do not: measuring force. Screw it tighter and you get a smaller reading, because you are compressing both the part and the frame. The ratchet or friction thimble exists to make that force the same every time, and using the thimble directly because it is quicker discards the instrument's main advantage.

Zero with the anvils closed, using the ratchet, on a clean and wrung surface.
Measure gauge blocks at several points through the range, including points that put the thimble at different angular positions — screw error varies around a turn.
Check anvil flatness and parallelism with optical flats where the grade warrants it.
Use the ratchet for every reading, including the zero check.

Gauge blocks as the reference

Gauge blocks are the working length standard behind most shop-floor dimensional work, graded by how closely they hold their nominal length and how flat and parallel their faces are, under ISO 3650. They are calibrated by mechanical comparison against a better set, or interferometrically at the top of the chain, and their own certificate carries a deviation from nominal that ought to be used rather than ignored — a block certified 25.0001 mm is a better reference used as 25.0001 than as 25.

Frequently asked questions

Why is 20 °C the reference temperature for dimensional measurement?: Because length changes with temperature, a specified dimension is only meaningful at a stated temperature, and international dimensional standards fix that temperature at 20 °C. A part measured at 25 °C is genuinely a slightly different size from the same part at 20 °C. For ordinary tolerances this does not matter; for anything toleranced in micrometres it is the dominant consideration, and it is why precision dimensional laboratories are temperature-controlled rather than merely comfortable.
How much does temperature affect a steel measurement?: Steel expands by roughly 11.5 micrometres per metre per degree Celsius. On a 100 mm dimension that is about 1.15 micrometres for every 1 °C away from 20 °C, so a workshop at 25 °C puts a steel part about 5.75 micrometres over its 20 °C size. If the instrument is also steel and at the same temperature the two expansions largely cancel; measuring aluminium with a steel instrument, they do not.
Why does a micrometer have a ratchet?: To make the measuring force repeatable. Tightening a micrometer harder gives a smaller reading, because the part and the frame both deform under the load — so a reading without a controlled force is a reading at whatever force that operator applied that day. The ratchet or friction thimble is the mechanism that removes the operator from the result, and bypassing it by turning the thimble directly throws away most of what makes a micrometer better than a caliper.
How accurate is a digital caliper really?: Far less accurate than its display suggests. A caliper resolving to 0.01 mm typically has a real measurement uncertainty in normal use of several hundredths of a millimetre, because of jaw parallelism, play in the slide, the force applied and the angle it is held at. It is an excellent instrument for what it is — quick, versatile, robust — and the mistake is asking it to police a tolerance it was never capable of policing.
Should I use the deviation on a gauge block certificate?: Yes. A block certified at 25.0001 mm is a better reference when used as 25.0001 mm than when treated as exactly 25 mm — you have paid for the information, and ignoring it discards the accuracy you bought. Using nominal values is acceptable only when the certified deviations are genuinely negligible against the uncertainty you are claiming, and that is a calculation rather than an assumption.

References

[1]ISO 3650:1998 — Geometrical Product Specifications (GPS): Length standards, gauge blocks
[2]EURAMET cg-2 — Calibration of Gauge Block Comparators
[3]ISO/IEC 17025:2017 — General requirements for the competence of testing and calibration laboratories
[4]JCGM 100:2008 — Evaluation of measurement data: Guide to the expression of uncertainty in measurement (GUM)

General technical guidance written against the cited sources. It is not regulatory or legal advice and does not replace the applicable standard, guideline or a qualified reviewer's judgement.

Have a question on this topic?

Ask ValiTrac AI and see the evidence and calculation behind the answer.

Ask ValiTrac AI

Free calculator: Calibration interval

Paste an instrument's as-found history and see how fast it is actually drifting, how much of its tolerance that consumes, and how long the evidence supports trusting it. The strongest input to an interval is the instrument's own record — this reads it.

Open the calculator

Metrology

Measurement uncertainty explained

What an uncertainty statement means, how it is built up, and why a coverage factor matters.

2026-09-13 · 4 min read

Calibration

How to set calibration intervals

There is no universal interval. Set it from risk, manufacturer data and — above all — the instrument's own as-found history.

2026-09-13 · 3 min read

Calibration by Parameter

Multimeter calibration: what the specification actually promises

A DMM specification is a formula, not a number, and it changes with range, function and how long since the last adjustment. How to read one and what a calibration has to cover.