Engineering note: ISO tolerance fits (H/s, H/r): typical δ/d ≈ 0.001–0.003. Dry steel-steel friction μ ≈ 0.12–0.15. Design for transmitted torque with safety factor 1.5–2.0. For large diameters use oil-injection assembly to protect bore surface.
What is a Press Fit?
🙋
What exactly is a press fit? I see parts like gears on shafts, but how do they stay on without screws or keys?
🎓
Basically, it's a mechanical joint where the shaft is made slightly larger than the hole in the hub. When you force them together, the elastic "squeeze" creates a huge contact pressure. This pressure, combined with friction, is what transmits torque and holds the assembly together. Try moving the "Diametral interference δ" slider in the simulator above to see how a tiny change, often just a few thousandths of an inch, creates that pressure.
🙋
Wait, really? So the parts are literally stretched and compressed? How do we calculate that pressure without breaking them?
🎓
Exactly! The hub stretches outward, and the shaft compresses inward. In practice, we use the Lamé equations for thick-walled cylinders. The key is balancing the interference with the material strength. For instance, if you set the "Hub outer diameter (do)" much closer to the contact diameter (d) in the simulator, you'll see the pressure skyrocket for the same interference, which could cause yielding.
🙋
That makes sense. So the main design goal is to get enough pressure for the required torque, but not so much it fails. How do friction and assembly force come into play?
🎓
Great question. The contact pressure (p) is the source of everything. The friction force is simply that pressure times the friction coefficient (μ) and the contact area. The assembly force you need to press them together is that friction force. In the simulator, try reducing the "Friction coefficient μ" to 0.05—you'll see the assembly force drop, but so will the torque capacity. It's a direct trade-off.
Physical Model & Key Equations
The core of the press-fit analysis is the Lamé equation, which calculates the interface pressure (p) caused by the diametral interference (δ). It accounts for the elastic deformation of both the hub (treated as a thick cylinder under internal pressure) and the shaft (under external pressure).
Where: p = Contact pressure at the interface [MPa, psi] δ = Diametral interference (shaft OD - hub ID) [mm, in] d = Nominal contact diameter [mm, in] E_h, E_s = Young's modulus of hub and shaft [GPa, psi] ν_h, ν_s = Poisson's ratio of hub and shaft d_o = Outer diameter of the hub [mm, in] d_i = Inner diameter of the shaft (0 for solid) [mm, in]
Once the pressure is known, the assembly force and torque capacity are derived from basic friction mechanics. The maximum stress, which dictates whether the part will yield, occurs at the inner surface of the hub.
$$F = \mu \, p \, \pi d L \quad \quad \quad \sigma_{\theta_{max}}= p \times \frac{d_o^2 + d^2}{d_o^2 - d^2}$$
Where: F = Axial force required for assembly (or separation) [N, lbf] μ = Coefficient of static friction L = Axial engagement length [mm, in] σθ_max = Maximum hoop (circumferential) stress in the hub [MPa, psi] — this must be compared to the material yield strength (Sy).
Frequently Asked Questions
If the contact pressure exceeds the material's yield stress, plastic deformation will occur in the hub or shaft. In particular, the circumferential stress on the inner diameter side of the hub becomes maximum, so during design, please consider the material's yield strength and safety factor, and ensure that the allowable interference is not exceeded.
For a solid shaft, input 0 for the shaft inner diameter di. By setting di=0 in Lamé's equations, the stress distribution for a solid shaft is correctly calculated. However, to avoid division by zero in the formulas, internal processing automatically compensates for this.
The main causes are that the mating surface diameter is too small relative to the interference, or the hub outer diameter is too close to the mating surface diameter. Additionally, unrealistic input values for Young's modulus or Poisson's ratio can also produce abnormal values. Please recheck the material constants and dimensions.
The current version does not account for thermal stress. It is intended only for mechanical interference at room temperature. For analyzing shrink fits using thermal expansion differences, please separately calculate the equivalent interference from the thermal expansion coefficient and temperature difference, and input that value.
Real-World Applications
Gear and Sprocket Mounting: Press fits are ubiquitous for mounting gears, pulleys, and sprockets onto shafts in powertrains and machinery. A common case is a steel gear pressed onto a steel motor shaft. The design ensures the torque from the motor is transmitted purely by friction, eliminating the need for keyways which are stress concentrators.
Bearing Seats: Rolling element bearings are often mounted with a light press fit (or "interference fit") on their inner race to prevent creeping and fretting corrosion under cyclic loads. The simulator's parameters for engagement length and modulus are critical here to ensure a secure fit without over-stressing the bearing race.
Automotive Wheel Hubs: The connection between a wheel hub and a constant velocity (CV) joint or axle often uses a precision press fit. The diametral interference is carefully controlled during manufacturing to withstand high alternating torques and bending moments from driving forces.
High-Pressure Cylinder Liners: In engines and hydraulic cylinders, a ductile iron or steel liner is press-fit into an aluminum engine block or outer housing. This creates a pre-stress that helps the assembly contain extremely high internal fluid pressures without the liner separating or leaking.
Common Misconceptions and Points to Note
First, let's talk about the misconception that "a larger interference fit is always stronger." While contact pressure does increase, the material's yield strength has a limit. For example, if you apply an excessive interference fit to an S45C hub (yield strength ~350 MPa) to create 400 MPa of contact pressure, plastic deformation will occur the moment it's assembled, actually reducing the holding force. If the stress distribution turns completely red in the simulator, consider that a "danger signal."
Next, using an "assumed value" for the coefficient of friction μ. Are you using a catalog's typical value (e.g., 0.15 for steel-on-steel) as-is? Actual parts may have cutting oil residue or different surface roughness. For instance, with a mirror-like finish on the same steel, μ can drop below 0.1. A difference of just 0.05 in this value can change the required press-fit force by over 30%. If a press-fit feels heavier or lighter than expected on an actual machine, reconsider μ first.
Finally, the pitfall of thinking "a longer fitting length L is always safer." While transmitted torque increases proportionally with length, so does the press-fit force. Forcing an excessively long fit risks shaft buckling (bending). This is especially critical for hollow shafts. As a rule of thumb, if you design a length exceeding about 1.5 times the shaft diameter d, a separate buckling calculation is often required.