Návrh a výpočty

A spiral retaining ring relies on square-corner contact with the groove wall to maximize its thrust capacity. If the groove has a large radius at the bottom, or if the mating part has a large chamfer, the point of contact shifts, creating a moment arm that tries to 'dish' the ring (turn it inside out). The reduction in thrust capacity can be modeled by a factor $C_f = \frac{d - (r + c)}{d}$, where $d$ is the groove depth, $r$ is the groove radius, and $c$ is the chamfer of the retained part. If $C_f$ is significantly less than 1, the ring will fail prematurely by being pushed out of the groove. In aerospace gearboxes, 'sharp-cornered' grooves are often specified, and shim rings are used between the chamfered bearing race and the retaining ring to ensure the load is applied as close to the groove wall as possible.

The thrust capacity of a spiral retaining ring is limited by two factors: the shear strength of the ring and the deformation of the groove. The shear-limited thrust $P_r$ is calculated as $P_r = \frac{D T \pi au_s}{K_s}$, where $D$ is the shaft/bore diameter, $T$ is the ring thickness, $ au_s$ is the shear strength of the material (typically taken as $0.6 \times \sigma_{tensile}$), and $K_s$ is a safety factor (usually 3). However, in many cases, the groove material (often aluminum or soft steel) yields before the ring shears. The groove-limited thrust $P_g$ is $P_g = \frac{D d \pi au_y}{K_s}$, where $d$ is the groove depth and $ au_y$ is the yield strength of the groove material. Designers must take the lower of these two values as the assembly's safe thrust capacity.

The rotational capacity of an external spiral retaining ring is the speed at which centrifugal force causes the ring to expand and lose contact with the groove bottom. The maximum allowable RPM $N$ is calculated as $N = \frac{1}{\pi} \sqrt{\frac{4.48 E I g}{w r^3 (1+K)}}$, where $E$ is the modulus, $I$ is the moment of inertia of the cross-section, $g$ is gravity, $w$ is the weight of the ring per inch, $r$ is the radius to the center of gravity, and $K$ is a factor related to the 'clinch' or interference fit of the ring in the groove. If the operating RPM exceeds this value, the ring may 'float' out of the groove, leading to a catastrophic failure of the assembly. For high-speed applications like turbochargers, 'self-locking' spiral rings are used, where a tab on the inner turn locks into a slot on the outer turn to mechanically prevent expansion.

The natural frequency $f_n$ of a wave spring is critical in high-speed automotive transmissions to avoid resonance-induced fatigue. The fundamental frequency is given by $f_n = \frac{1}{2 \pi} \sqrt{\frac{K g}{W}}$, where $K$ is the spring rate, $g$ is the gravitational constant, and $W$ is the weight of the spring. Because wave springs have a much lower mass than traditional coil springs, their natural frequency is significantly higher, often moving the resonance point outside the operating range of the machinery. However, for multi-turn springs, the individual turns can also exhibit local vibration modes. Engineers use Finite Element Analysis (FEA) to verify that the excitation frequency of the system does not align with the first or second harmonic of the spring, which would otherwise cause 'wave jumping' and premature failure.

As a wave spring is compressed from its free height towards its solid height, the waves flatten, which naturally causes the mean diameter to expand. This expansion can be approximated by the formula $D_{expanded} = \sqrt{D_{free}^2 + (h^2 / \pi^2)}$, where $h$ is the height of the wave. If the spring is installed in a tight bore, this radial expansion can lead to 'binding' or friction against the bore wall, which artificially increases the spring rate and causes unpredictable hysteresis. Designers must ensure that the clearance between the Outer Diameter (OD) and the bore, or the Inner Diameter (ID) and the shaft, is sufficient to accommodate this expansion at maximum deflection. In precision medical devices, the use of a 'shim-end' wave spring can mitigate this by providing a flat contact surface that stabilizes the expansion.

Nested wave springs are produced by coiling multiple layers of the same wire in parallel, whereas Crest-to-Crest springs are coiled in series. The primary advantage of a nested spring is that the spring rate $K$ increases linearly with the number of turns $n$ (layers), expressed as $K_{total} = n \times K_{single}$. This allows for extremely high loads in very tight axial spaces where a single-turn spring would fail due to over-stressing. Mathematically, while a Crest-to-Crest spring increases deflection for a given load, a nested spring increases load for a given deflection. In subsea valve actuators, nested springs are used to provide thousands of Newtons of force with a stack height that is a fraction of what a traditional coil spring would require, though they are more susceptible to internal friction and galling between layers.

The number of waves $N$ is inversely proportional to the operating stress for a given deflection. The bending stress $\sigma$ is calculated by $\sigma = \frac{3 \pi P D}{4 N^2 b t^2}$, where $P$ is the load and $D$ is the mean diameter. Increasing $N$ reduces the stress at a specific load point because the total deflection is distributed across more contact points, effectively shortening the beam length between peaks. However, increasing $N$ also increases the spring rate $K$ exponentially ($N^4$), which can lead to a very stiff spring. For fatigue-critical applications, such as automotive clutch packs, $N$ is optimized to keep the alternating stress amplitude below the endurance limit of the material, typically $SAE 1070-1090$ carbon steel, while maintaining the required clamp force.

The spring rate $K$ for a multi-turn Crest-to-Crest wave spring is derived from the beam deflection formula adapted for a curved geometry. It is expressed as $K = \frac{E b t^3 N^4}{R^3 5.88 n}$, where $E$ is the modulus of elasticity, $b$ is the radial wall thickness, $t$ is the material thickness, $N$ is the number of waves per turn, $R$ is the mean radius, and $n$ is the number of turns. Non-linearity occurs primarily at the extremes of the deflection curve. At the start, 'bedding in' of the waves against the mating surfaces causes a lower initial rate. As the spring approaches 'solid height,' the waves begin to flatten and touch, causing a sharp increase in the rate. In high-precision applications, engineers must account for the $K_{factor}$ which adjusts for the change in the moment arm as the wave crests shift radially during compression.

In cryogenic environments (e.g., liquid nitrogen at -196 C), the yield strength $\sigma_y$ and tensile strength of stainless steels like 302 or 316 increase, but ductility decreases. The thrust load capacity formula $P_g = (D \cdot d \cdot \pi \cdot \sigma_y) / S_f$ suggests an increase in capacity. However, the risk of brittle fracture becomes the dominant failure mode. The safety factor $S_f$ must be increased from 2 to 4 to account for the reduced fracture toughness $K_{IC}$. Furthermore, the thermal contraction of the ring ($L = L_0 \cdot \alpha \cdot \Delta T$) must be calculated to ensure that the ring does not contract so much that it binds on the shaft or expands out of the groove due to differential cooling rates between the ring and the housing.

The expansion limit for an external ring is the maximum diameter it can be stretched to without permanent set. The maximum stress during expansion is $\sigma_{exp} = (E \cdot t \cdot (D_s - D_i)) / (D_i^2)$, where $D_s$ is the shaft diameter and $D_i$ is the ring's free inner diameter. If $\sigma_{exp}$ exceeds the yield strength $\sigma_y$ of the material (e.g., SAE 1070), the ring will not return to its original shape and will lose its grip on the groove. This is why spiral rings are often manufactured with multiple turns; two or three thinner turns can provide the same total thickness $T$ as a single heavy turn but with significantly lower individual stress $t$ during installation, allowing for much greater expansion ratios.

Dishing occurs when the axial load $P$ creates a moment $M = P \cdot (b/2)$ that twists the ring's cross-section. This is exacerbated if the groove walls are not perpendicular or if the groove has a large radius at the bottom. The resistance to dishing is proportional to the material's modulus $E$ and the thickness $t$ cubed. If the ring dishes, it reduces the contact area with the groove, leading to premature failure of the groove edge. To mitigate this, engineers specify a maximum groove radius of 0.1 times the thickness and ensure the groove depth $d$ is sufficient to support at least 70 percent of the ring's radial wall $b$.

The thrust capacity of a spiral retaining ring assembly is the lesser of the ring shear and the groove material shear. The ring shear capacity $P_r$ is $P_r = (A_r \cdot \tau_{ring}) / S_f$, where $A_r$ is the shear area of the ring and $\tau_{ring}$ is the shear strength (approx. 0.6 times the tensile strength). The groove shear capacity $P_g$ is calculated as $P_g = (D \cdot d \cdot \pi \cdot \sigma_y) / S_f$, where $D$ is the shaft diameter, $d$ is the groove depth, and $\sigma_y$ is the yield strength of the groove material. It is a common mistake to ignore the edge margin; the distance from the groove to the end of the shaft must be at least $3 \cdot d$ to prevent the groove wall from shearing out. A safety factor $S_f$ of 2 or 3 is typically applied.

The centrifugal force acting on an external spiral retaining ring can cause it to expand and lift out of its groove. The maximum allowable RPM $V$ is calculated using $V = \sqrt{(4.48 \cdot 10^{12} \cdot E \cdot I \cdot g) / (w \cdot \rho \cdot R^3 \cdot (1 + \nu))}$, where $E$ is the modulus, $I$ is the moment of inertia, $g$ is gravity, $w$ is the material weight per unit length, $\rho$ is the density, $R$ is the groove radius, and $\nu$ is Poisson's ratio. This limit is reached when the ring's expansion due to centripetal acceleration exceeds the interference fit within the groove. In high-speed aerospace turbines, self-locking features (tabs and slots) are added to the spiral ring to mechanically prevent this expansion, effectively negating the RPM limit imposed by the centrifugal calculation.

The radial wall width $b$ is a linear multiplier in the load formula $P = (E \cdot b \cdot t^3 \cdot f \cdot N) / (D_m^3 \cdot K)$. While it directly increases the load capacity, a wide radial wall can introduce 'dishing' effects where the cross-section of the spring tilts under load. This non-parallel compression alters the effective mean diameter $D_m$ during the stroke, leading to a progressive (non-linear) spring rate. For high-precision medical devices, the ratio of $b/t$ is kept within a range of 8:1 to 12:1 to maintain linearity. If the wall is too wide, the spring behaves more like a Belleville washer, losing the characteristic soft-rate benefit of the wave design.

Fatigue life is dictated by the stress range between the initial preload height $H_1$ and the final operating height $H_2$. The alternating stress $\sigma_a$ is defined as $(\sigma_{max} - \sigma_{min}) / 2$, and the mean stress $\sigma_m$ as $(\sigma_{max} + \sigma_{min}) / 2$. Utilizing a Goodman diagram, engineers must ensure that the point $(\sigma_m, \sigma_a)$ lies within the safe region for the specific material, such as SAE 1070 Carbon Steel. In dynamic hydraulic valves, if the deflection $f$ exceeds 50 percent of the available travel, the probability of fatigue crack initiation at the wave crests or troughs increases due to micro-plasticity. Precise work height control ensures the stress remains below the endurance limit $\sigma_e$.

The solid height $H_s$ of a nested wave spring is not simply the sum of material thicknesses. It is calculated as $H_s = (n \cdot t) + (n-1) \cdot \delta$, where $n$ is the number of turns, $t$ is the thickness, and $\delta$ is the nesting gap factor, though in a perfectly nested spring, $\delta$ approaches zero. The actual height must also account for the radial expansion of the material during compression. As the spring is compressed toward solid, the mean diameter $D_m$ increases according to the formula $\Delta D = 0.02 \cdot (f^2 / D_m)$, where $f$ is the deflection. If the bore clearance is insufficient to accommodate $\Delta D$, the spring will bind, leading to an unpredictable non-linear spring rate and potential catastrophic failure of the assembly.

In wave spring design, the number of waves $N$ per turn is a primary driver of both load and stress. The stress $\sigma$ is calculated as $\sigma = (3 \cdot \pi \cdot P \cdot D_m) / (4 \cdot b \cdot t^2 \cdot N^2)$. This equation demonstrates an inverse square relationship between the number of waves and the stress. By increasing $N$ for a fixed load $P$, the stress level decreases exponentially. However, an increase in $N$ also increases the spring rate $k$. Designers in automotive transmission systems often optimize $N$ to balance the required axial force against the fatigue limit of the material, typically aiming for stress levels below 80 percent of the minimum tensile strength of 17-7PH CH900 to ensure longevity during high-cycle operation.

The spring rate $k$ for a Crest-to-Crest wave spring is defined by the relationship between the applied load $P$ and the deflection $f$. Using the specialized Munter's formula, the load is expressed as $P = (E \cdot b \cdot t^3 \cdot f \cdot N) / (D_m^3 \cdot n^4 \cdot K)$, where $E$ is the Modulus of Elasticity, $b$ is the radial wall, $t$ is the material thickness, $N$ is the number of waves per turn, $n$ is the number of turns, and $D_m$ is the mean diameter. To find the spring rate $k = P/f$, we rearrange to $k = (E \cdot b \cdot t^3 \cdot N) / (D_m^3 \cdot n^4 \cdot K)$. The factor $K$ is a correction constant for the wave profile, typically around 1.0 for theoretical sinusoidal waves but adjusted for actual crest contact geometry. For high-precision aerospace applications, it is critical to realize that the number of turns $n$ is in the denominator with a power of 4, meaning increasing the number of turns significantly reduces the spring rate, allowing for high deflection in restricted spaces.

Dishing occurs when the axial load $P$ causes the ring to deflect into a cone shape. The coning angle $\phi$ can be estimated by considering the ring as a circular plate with a hole, subjected to a moment $M = P \times (R_c - R_g)$, where $R_c$ is the load contact radius and $R_g$ is the groove reaction radius. The angle $\phi \approx \frac{M R_m}{E I}$. When $\phi$ exceeds approximately 5 to 7 degrees, the risk of the ring 'rolling' out of the groove increases significantly. To mitigate this in heavy-duty applications, engineers specify materials with higher Modulus $E$ or increase the material width $b$ to increase the moment of inertia $I$, thereby stiffening the ring against torsional deformation.

Multi-turn spiral retaining rings (often 2 or 3 turns) provide a $360^{\circ}$ retaining surface without the 'ears' or lugs found on stamped circlips. This leads to a more uniform distribution of the axial load around the circumference of the groove. In a multi-turn ring, the load is shared across the turns, although the turn closest to the load source bears the highest stress due to the axial gap between turns. The total thickness $T$ in the shear formula $P_s = \frac{D T \pi \tau}{K}$ is the sum of the individual turn thicknesses. This design eliminates the gap found in single-turn rings, which is vital in applications requiring uniform clamping or preventing the passage of small particles in a sealed assembly.