How much interference does a press fit need, and what pressure, stress and holding torque does it give? Enter the shaft and hub sizes, the smallest and largest interference your tolerances allow, and the two materials. The calculator gives the contact pressure, the stress in the hub and the shaft, the force and torque the joint holds, the force to press it together, how all of that changes from the coldest to the hottest service temperature, and how far to heat the hub or cool the shaft for a shrink fit.
Engineering estimate, not a quote. The equations assume elastic, round, smooth cylinders of equal length; confirm critical joints with push-out and torque tests on real parts.
A press fit holds because the shaft is slightly bigger than the hole. Forcing them together stretches the hub outward and squeezes the shaft inward until both meet at a common diameter, and the pressure between them, the contact pressure p, is what produces friction against sliding and turning. The interference is not shared equally: the softer and thinner part does most of the moving. Lamé's solution for thick-walled cylinders gives exactly how much each part moves for a given pressure, and setting the two movements to add up to the interference gives p.
Step 1: compliance of each part. For the hub, the bore grows by p × d × [(do² + d²)/(do² − d²) + νo] ÷ Eo. For the shaft, the outside shrinks by p × d × [(d² + di²)/(d² − di²) − νi] ÷ Ei; for a solid shaft the bracket is simply 1 − νi. The geometric term in the hub bracket is always greater than 1 and grows quickly as the hub gets thin, which is why a thin hub gives a low pressure.
Step 2: contact pressure. The two movements add up to the diametral interference δ, so p = δ ÷ (d × [hub term + shaft term]). The calculator does this for the minimum interference, which sets how much the joint holds, and the maximum, which sets the stress and the press force.
Step 3: stresses. At the hub bore the hoop (tangential) stress is tensile, σt = p × (do² + d²)/(do² − d²), and the radial stress is −p. These are combined with the von Mises criterion and compared with the yield strength. A solid shaft is in uniform compression, −p in both directions, so its von Mises stress is just p. A hollow shaft is worst at its bore, where the hoop stress is −2p d²/(d² − di²).
Step 4: holding and pressing. The axial force to slide the joint is the friction on the contact area, F = π d L μ p, and the torque to turn it is F × d/2. The same expression with the maximum pressure gives the press-in force at full engagement.
Step 5: temperature. If the hub and shaft expand differently, the interference changes by (αo − αi) × d × ΔT from the assembly temperature. An aluminum or plastic hub on a steel shaft loses interference when hot and gains it when cold. The calculator rechecks every result at the coldest and hottest temperatures you give, and for a shrink fit it gives the temperature change that opens the bore, or closes the shaft, by the maximum interference plus a clearance.
A gear hub in aluminum 6061-T6, 40 mm outside diameter, is pressed onto a solid 1045 cold-drawn steel shaft of 20 mm diameter with 25 mm of engagement. The tolerances give a minimum interference of 0.020 mm and a maximum of 0.035 mm. The product is assembled at 23 °C and must work from −30 °C to 70 °C. Representative values: aluminum E = 68,900 MPa, ν = 0.33, yield 276 MPa, α = 23.6 µm/m·°C; steel E = 207,000 MPa, ν = 0.29, yield 530 MPa, α = 12 µm/m·°C. Friction is taken as μ = 0.15.
At room temperature the joint is comfortable: it holds 72.7 N·m before slipping and the hub bore is at less than half its yield strength even at maximum interference. The temperature range changes the picture in both directions. On a 70 °C day the aluminum has grown away from the steel and more than half the minimum interference is gone, so the torque capacity drops to 33.1 N·m; if the gear has to carry 50 N·m, this joint slips hot even though it passes on the bench. At −30 °C the aluminum has shrunk onto the steel and the worst-case hub stress rises to 170 MPa, still below yield but with much less margin. For assembly, a hot plate or oven at about 120 °C opens the bore enough to drop the hub on with no press at all. Press Load the worked example above to reproduce every number.
Work backwards from what the joint must do. The minimum interference has to hold the required torque or axial load at the worst temperature, with a margin for the uncertainty in friction. The maximum has to keep both parts below yield, or below a long-term design stress for plastics, at the other temperature extreme. Everything in between is the tolerance band, and the band is the sum of the shaft tolerance and the bore tolerance. If the gap between the two limits is too narrow for the tolerances the process can hold, the joint needs a different design rather than a tighter drawing: a longer engagement, a thicker hub, a knurl or spline, or adhesive as well as interference.
For metal shafts in metal hubs, the standard hole-basis interference fits (H7/p6 for a locational interference fit, H7/s6 for a medium drive fit and H7/u6 for a force fit) give tolerance limits a machinist can hold and make a sensible starting point; take the resulting smallest and largest interference from a fit table and enter them here. Surface finish matters on small fits: pressing flattens the peaks of both surfaces, so on a small diameter a few micrometres of the measured interference can be lost, and a rough bore holds less than its measured size suggests.
Pressing a metal pin, bushing or bearing into a molded boss is one of the most common uses of this calculation, and the same equations apply with two changes in how you read them. First, the plastic's modulus is so much lower than steel's that the pin barely moves: the hub takes nearly all the interference, and its hoop strain at the bore is close to δ/d. A 6 mm steel pin in a 12 mm acetal boss with 0.06 mm interference gives p = 13.8 MPa, a hoop stress of 23.0 MPa and a hoop strain of 0.995% (using the calculator's POM preset, E = 2,800 MPa and ν = 0.35), almost exactly 0.06/6. That is why resin suppliers express press-fit limits as an allowable strain, and why you should compare the strain this calculator reports with the value in the supplier's design guide rather than with the short-term yield.
Second, the pressure does not stay put. Plastics creep, so a press fit relaxes over time, faster when warm, and the holding force after months can be a fraction of the value on the day of assembly. Use the long-term creep modulus from the supplier for the holding calculation, and treat the short-term result as the ceiling. Moisture adds a third effect for nylon, which softens and swells as it absorbs water. A molded boss also usually carries a knit line where the melt fronts met around the core pin, and a knit line is weaker in exactly the hoop tension a press fit creates. Gate so the knit line sits where the boss is thickest, or support the boss with ribs, and test assemblies for cracking after a temperature cycle, not just at the press.
| Method | Use it when | Limit |
|---|---|---|
| This calculator (Lamé equations) | Round shafts in round hubs of uniform wall; choosing the interference band; checking hold, stress, press force and temperature effects | Elastic only; equal lengths; no keyways, slots or uneven walls; friction is an input |
| Fit tables (hole-basis interference fits) | Turning a fit class into machinable size limits for metal parts | Gives sizes, not pressure or torque; use the limits as inputs here |
| Resin supplier allowable strain | Plastic hubs or plastic shafts; the governing limit is strain, and long-term relaxation matters | Grade-specific; needs creep data for long-term holding |
| Finite element analysis | Bosses with ribs or slots, hubs shorter than the shaft (stress rises at the hub ends), plastic flow beyond yield | Needs a model, contact settings and material curves |
| Push-out and torque tests | Confirming the friction coefficient and holding of the real process, finish and plating | Needs parts; destructive |
The contact pressure, hub and shaft stress and diameter-change equations are the Lamé thick-walled cylinder solutions for interference fits as presented in Shigley's Mechanical Engineering Design and in Roark's Formulas for Stress and Strain, assuming plane stress, elastic behaviour and parts of equal length. Stresses are combined with the von Mises (distortion energy) criterion. The holding force and torque use Coulomb friction on the cylindrical contact area. The temperature terms use the linear expansion relation ΔL = α·L·ΔT. The press-fit strain approach for plastic hubs follows the method used in resin suppliers' design guides. All material values in the menus are representative published room-temperature values and are fully editable; the friction coefficient is an assumption the user should confirm by test. The thresholds the calculator uses to flag results (a hub safety factor under 1.5, a hub ratio under 1.5) are page-set engineering practice, not standards.
Use the Lamé thick-wall cylinder result for two cylinders forced together: p = δ ÷ (d × [(1/Eo)((do² + d²)/(do² − d²) + νo) + (1/Ei)((d² + di²)/(d² − di²) − νi)]), where δ is the diametral interference, d the interface diameter, do the hub outside diameter, di the shaft bore (zero for a solid shaft), E the modulus and ν Poisson's ratio of the outer (hub) and inner (shaft) part. For a solid 1045 steel shaft of 20 mm in a 6061-T6 aluminum hub of 40 mm outside diameter with 0.020 mm interference, the hub term is (2000/1200 + 0.33)/68,900 = 2.898 × 10⁻⁵ per MPa, the shaft term is (1 − 0.29)/207,000 = 0.343 × 10⁻⁵ per MPa, and p = 0.020 ÷ (20 × 3.241 × 10⁻⁵) = 30.9 MPa.
The axial force is the friction on the contact area: F = π × d × L × μ × p, where L is the engaged length and μ the coefficient of friction. Size the press for the maximum interference, because that gives the highest pressure. For the 20 mm steel shaft in a 40 mm aluminum hub with 25 mm of engagement, the maximum interference of 0.035 mm gives p = 54.0 MPa, and with μ = 0.15 the press-in force at the end of the stroke is π × 20 × 25 × 0.15 × 54.0 = 12,700 N, about 12.7 kN or 2,860 lbf. The force rises roughly in proportion to the engaged length as the shaft goes in. The friction coefficient is the least certain input, so measure it on trial parts for a production process.
The slip torque is the friction force times the radius: T = π × μ × p × d² × L ÷ 2. Use the minimum interference, because that gives the lowest pressure. For the 20 mm steel shaft in the aluminum hub with 0.020 mm minimum interference, p = 30.9 MPa, the holding force is 7,270 N and the slip torque is 7,270 × 0.010 m = 72.7 N·m. If the product runs hot, recompute at the hottest temperature: the same joint at 70 °C holds only 33.1 N·m, because the aluminum hub has grown away from the steel shaft.
When the hub and shaft have different expansion coefficients, the interference changes by (αhub − αshaft) × d × ΔT. An aluminum hub at 23.6 × 10⁻⁶/°C on a steel shaft at 12 × 10⁻⁶/°C loses 11.6 × 10⁻⁶ × 20 × 47 = 0.0109 mm of interference on a 20 mm fit heated from 23 to 70 °C, more than half of a 0.020 mm minimum interference. The same fit gains 0.0123 mm at −30 °C, which raises the hub stress. With 0.020 mm minimum interference the fit goes loose at 23 + 0.020 ÷ (11.6 × 10⁻⁶ × 20) = 109 °C. Check the holding torque at the hot end and the hub stress at the cold end when the hub expands more than the shaft, and the reverse when it expands less.
Heat the hub until its bore has grown by the maximum interference plus an assembly clearance: ΔT = (δmax + c) ÷ (αhub × d). For a 20 mm aluminum 6061 hub with 0.035 mm maximum interference and 0.010 mm clearance, ΔT = 0.045 ÷ (23.6 × 10⁻⁶ × 20) = 95 °C, so about 118 °C from a 23 °C start. Cooling the steel shaft instead would need a drop of 0.045 ÷ (12 × 10⁻⁶ × 20) = 188 °C, to about −165 °C, which only liquid nitrogen reaches. Heating the hub is usually the practical route, but keep heat-treated aluminum well below its aging temperature and plastic hubs below the temperature at which they distort.
Yes, and the same Lamé equations apply, but the plastic hub does almost all of the deforming, so its hoop strain at the bore is close to the interference divided by the diameter. A 6 mm steel pin pressed into a 12 mm acetal boss with 0.06 mm interference gives a contact pressure of 13.8 MPa, a hoop stress of 23.0 MPa and a hoop strain of about 1.0% at the bore, which is 0.06 ÷ 6. Plastics creep, so the pressure falls over time, and a molded boss usually has a knit line, which is weak in hoop tension. Check the hoop strain against the resin supplier's allowable value for press fits, use the long-term creep modulus for holding force, and consider knurled inserts or heat staking for joints that must hold for years.
A thicker hub raises the contact pressure for a given interference and lowers the hoop stress, but the gains flatten out. For an aluminum hub on a solid steel shaft with an interference of 0.1% of the diameter, the contact pressure is 13.5 MPa when the hub outside diameter is 1.25 times the bore, 30.9 MPa at 2 times, 37.9 MPa at 3 times and can never exceed 44.0 MPa however thick the hub is. The hoop stress at the bore falls from 61.3 MPa at 1.25 times to 51.4 MPa at 2 times and 47.4 MPa at 3 times. Most of the benefit is reached by an outside diameter of about two to three times the bore.
Pressing a pin into a molded boss, or a bearing into a die-cast housing? Ask our AI assistant. It answers from this website, and our engineering team follows up if you want.
AI assistant. Answers can be wrong; nothing it says is a quote or a promise.
This calculator gives an engineering estimate. Our engineering team reviews your CAD for interference bands, boss walls, knit-line position and insert options together, while each change is still a design edit and not a tool modification.
Book a 30-min call Send a drawing or sample