"Interference: 18 to 59 µm" is the input to the question, not the answer. Squeeze a shaft into a slightly smaller hole and the hub stretches, the shaft is compressed, and the two press against each other at a contact pressure that depends on how stiff both parts are and how much metal the hub has around the bore. Everything you actually need follows from that one number: the force to press it together, the torque the joint will transmit before it slips, the hoop stress trying to split the hub, and how hot the hub must be to drop over the shaft instead.
Materials come from the thermal panel above — those two selectors now change these numbers at every temperature, including 20 °C.
A fit does not have a capacity. It has a range of capacity, because every part you make lands somewhere different inside its tolerance band — and contact pressure is proportional to interference, so the range carries straight through to torque. Specify H7/s6 at Ø50 mm and the interference runs from 18 to 59 µm: the worst assembly you are entitled to receive holds less than a third of what the best one holds. Both are in specification. If the design needs the torque, it must be sized on the minimum, and if the hub must survive, it must be checked at the maximum. Using the middle checks nothing.
That is also why a bigger hub is not simply a stronger one. Thickening the hub raises the contact pressure for the same interference, so torque goes up — but with steeply diminishing returns. Going from an outer diameter of twice the bore to three times buys about 18% more pressure; everything beyond three times, all the way to an infinitely thick hub, is worth only another 13%. Drag the outer diameter and watch the curve flatten.
A 50 mm steel shaft in a steel hub of 100 mm outer diameter, 50 mm engaged, specified H7/s6.
At maximum interference — 59 µm — the contact pressure is about 93 MPa. Pressing it together takes roughly 88 kN, which is a serious press, and the joint will then transmit around 2,190 N·m before it slips.
The hub bore carries a hoop stress of 155 MPa — but it is also being squeezed radially at the full 93 MPa at the same time, and the combination is what matters. The von Mises stress there is 217 MPa against a 250 MPa yield, so the hub is at 87% of yield rather than the 62% the hoop figure alone suggests. That is why the readout reports the combined stress and not the pretty number.
At minimum interference — 18 µm, the same drawing, a different pair of parts — pressure falls to 28 MPa and the torque to 668 N·m. That is the number the design has to survive on.
And pressing is not the only option. Heat the hub by about 180 °C and its bore grows past the shaft, so it drops on with no force at all and grips as it cools. That is why the thermal panel above is not a curiosity: it is the assembly method.
Both parts deform. The interference is shared between the hub growing and the shaft being squeezed, in proportion to how stiff each one is. Put a steel shaft in an aluminium hub and the aluminium does most of the moving, so the contact pressure — and the torque — comes out around 40% of the all-steel figure for the same interference. A plastic hub, an order of magnitude less stiff again, barely grips at all.
Up to a point, and the limit is the hub, not the shaft. Hoop stress at the bore rises with pressure, and once it reaches yield the bore takes a permanent set — the joint loses grip the moment it is assembled, and may crack outright in a brittle hub such as cast iron. The readouts here flag it when the hub bore passes yield.
No. It is the theoretical slip point at the stated friction coefficient, with no safety factor, assuming clean dry surfaces and full contact along the whole engagement. Friction is the least certain number in the whole calculation — anything from about 0.05 oiled to 0.25 dry on rough surfaces — which is why it is a slider rather than a constant. Apply a real factor of safety to the minimum-interference case.
Heating is usually kinder. A press drags the surfaces across each other under full pressure, which scores them and can shave material off the shaft — a shrink fit lands with no sliding at all and holds better as a result. Press fits win on convenience and where heating would ruin a temper, a bearing or a seal.
The fundamental deviation fixes where the band sits against the nominal size. Lowercase is a shaft, uppercase a hole. H/h sit at the line; letters before give clearance, after give interference.
The IT grade sets the band width: IT6–7 precise, IT9–11 general, IT12+ rough. The same grade widens with size.
Holes are cut with fixed tools (drills, reamers), so fix the hole at H and vary the shaft, which can be turned to any size. Why H7/g6 dominates drawings.
A gap everywhere is clearance (slides); overlap everywhere is interference (pressed); a band straddling the line is transition (accurate location).
Start from the function: free-running (H8/d9, H9/d9), close-running (H7/g6), location or transition (H7/k6, H7/n6), press or interference (H7/p6, H7/s6), drive or shrink (H7/u6). Then confirm against your load, material and assembly method: this tool gives the numbers, not the engineering judgement.
Both the IT grade width and the fundamental deviation grow with nominal size: a 10 µm band on a 5 mm part would be impossible on a 400 mm one. ISO 286 scales them using the standard tolerance unit, which rises with the cube root of size.
Every material grows with temperature at its own rate: an aluminium housing expands roughly twice as fast as a steel shaft in it (steel grows about 1 mm per metre per 85 °C, α ≈ 12 µm/m·°C). Push either part far from the 20 °C reference and a clearance fit can tighten into interference, or the reverse. The thermal section makes it concrete for your materials.
Hole-basis keeps the hole at H and varies the shaft letter, the usual choice because holes are harder to machine to size. Shaft-basis keeps the shaft at h and varies the hole. Both reach the same fits; hole-basis needs fewer reamers and gauges.
Clearance: the hole is always bigger — parts slide. Interference: the shaft is always bigger — parts must be pressed or shrink-fit. Transition: depending on where each part lands in its band, you might get a slight gap or a slight squeeze — used for accurate location.
Hole-basis keeps the hole at H (lower deviation zero) and varies the shaft letter — the usual choice, since holes are harder to machine to size. Shaft-basis keeps the shaft at h and varies the hole. Both reach the same fits; hole-basis needs fewer reamers and gauges.
Both the IT grade width and the fundamental deviation grow with nominal size — a 10 µm band on a 5 mm part would be impossible on a 400 mm one. ISO 286 scales them using the standard tolerance unit, which rises with the cube root of size.
Start from the function: free-running (H8/d9, H9/d9), close-running (H7/g6), location/transition (H7/k6, H7/n6), press/interference (H7/p6, H7/s6), drive/shrink (H7/u6). Then confirm against your load, material and assembly method — this tool gives the numbers, not the engineering judgement.
Every material grows or shrinks with temperature, but at different rates — an aluminium housing expands roughly twice as fast as a steel shaft inside it. Push either part far enough from the 20°C reference temperature the tolerance was specified at, and a clearance fit can tighten into interference, or vice versa. The thermal section below makes that concrete for your actual hole and shaft materials.
This page has changed in 193 archived releases, but each of those was a site-wide pass, so none is attributable to this tool on its own and none is listed here. That is not a claim that the tool never changed — a release that reworked many pages at once may well have altered this one too. The changelog has them.