CNC thermal expansion calculator

Linear expansion: ΔL = L₀ · α · ΔT, where α is the linear expansion coefficient, L₀ the dimension at the reference temperature and ΔT = T_final − T_initial. The final length is L₀ + ΔL. For a diameter or radius the same formula applies to that dimension.

In the GetGCode catalog we use µm/m·K (micrometers per meter and kelvin), equivalent to 10⁻⁶ /°C. A steel with α = 12 µm/m·K expands 12 µm for each meter and each degree of rise.

A 300 mm aluminum bar that warms up 10 °C grows ~70 µm. That's enough to lose a fit, push an IT7 dimension off or fight the gauge at 20 °C. In long parts, mixed fixtures (steel/aluminum) and rooms without climate control it's one of the classic causes of "the part doesn't measure".

Yes. Search for C45, 42CrMo4, AISI 304, AW-6082, POM… and we load the typical α from the shop data sheet (µm/m·K) and, if available, the E modulus to estimate the stress if expansion is constrained. Values are catalog guidance, not certificate data.

It's the growth difference between two materials with the same dimension and the same ΔT: ΔL_diff = (α_A − α_B) · L · ΔT. Useful for bushings, inserts, shafts in housings of another metal or fixtures where the pin and the plate don't expand alike.

If the part can't expand freely (clamped, welded, embedded), the approximate thermal stress is σ = E · α · ΔT. It's indicative: it doesn't replace a finite element analysis nor account for creep or complex geometry.

The usual metrology reference is 20 °C (ISO 1). If you measure a part still hot from machining or washing, the dimension is "swollen". For fine tolerances, let it stabilize or correct with this calculator knowing part T and reference T.

That is the Reading at 20 °C mode, the reference standard of ISO 1. Correcting the part alone is not enough: the gauge expands too, so you enter both temperatures and both materials. You get the corrected dimension, how much each one contributes and, if you type the nominal and the two drawing deviations, whether it passes and how much of the band the correction eats.

What you measure with, not what the part is made of. The Measurement at 20 °C mode offers the instrument materials with their coefficient: steel for calipers, micrometers and gauge blocks (11.5), long 500 mm blocks (10.6), zirconia ceramic (9.2), chrome carbide (8.4), granite surface plates (6.3), tungsten carbide (4.5) and Invar (1.2), all from table 3.2 of the NIST Gauge Block Handbook. There is no option by brand because the brand does not decide it: Mitutoyo, Mahr and Tesa all sell steel, carbide and ceramic blocks. And watch the easy case: measuring steel with steel at the same temperature gives zero correction, which is exactly why gauge blocks are made of steel.

The Assembly mode works it out from the shaft Ø, the bore Ø, the clearance you want to slide it in and both materials. You get the three ways out: heating the bore alone, chilling the shaft alone, or splitting between the two, with presets for a freezer, dry ice and liquid nitrogen. It warns you if your means do not reach and if the temperature goes past the tempering point of a hardened part.

It depends on thickness, material and whether air is moving. The tool estimates it by lumped capacitance: a 25 mm steel plate in still air takes hours, and with a fan it drops to less than a third. It is an estimate, not a measurement: the real shape, how it rests and the draughts change the result.

Yes. POM, PA or HDPE can have α 5–10× that of steel. A cold-fitting metal-plastic fitting can seize or loosen within a few degrees. Always use the material from the catalog or an α from the supplier.

No, and that is an error source almost no calculator corrects. The α on a material datasheet is the mean value measured between 20 and 100 °C. Above that the material expands more: a C45 goes from 11.1 to 13.9 µm/m·K at 500 °C, 25 % more. This tool uses per-range tables from standards and manufacturers (EN 10088-1, thyssenkrupp, Lucefin, Deutsches Kupferinstitut, Iron Castings Handbook) and tells you which α it is applying for the temperature step you entered. At shop temperature nothing changes: the correction only shows up when you really leave the range.

There it works the other way: the material contracts LESS than its datasheet says, because α drops on cooling. A 316 stainless shaft at −196 °C uses 13.0 µm/m·K instead of the catalogue's 16.5, 19 % less. On a Ø60 shaft that is 41 µm of contraction you are not going to get, and that is exactly the difference between the assembly going in or getting stuck halfway. The sub-zero data comes from the NIST cryogenic fit for 304/316 and from the ASME B31.3 table for carbon steel.

As far as each family's table goes, and the tool warns you when you leave it. There are also physical limits we flag separately: carbon steel stops working above 650 °C because it starts transforming to austenite and the part shrinks while you keep heating it; duplex is not designed above 250-300 °C because of 475 °C embrittlement; grey iron grows permanently past 400-500 °C; and a T6 aluminium or a tool steel can give a correct expansion yet lose hardness without recovering it on cooling.