FAQ

When a wave spring is compressed from its free height ($H_f$) toward its work height ($H_w$) or solid height ($H_s$), the waves flatten, causing the mean diameter $D_m$ to increase due to the geometry of the arc segments. The approximate expansion $Δ D$ can be estimated using the relationship $Δ D = 0.02 \cdot \frac{(H_f - H_w)^2}{D_m}$. In tight-tolerance bore installations, this radial growth is critical to prevent binding. Engineers must calculate the 'Maximum Expanded OD' using $OD_{max} = OD_{free} + \sqrt{L^2 + R^2} - R$, where $L$ is the wave length and $R$ is the radius of curvature. If the spring is constrained by a bore, this expansion induces secondary hoop stresses that can lead to premature yielding or 'scuffing' against the bore wall, necessitating a reduction in the initial OD specification.

Nested wave springs consist of multiple turns coiled in parallel rather than in series. This configuration drastically increases the spring rate $k$ by a factor equal to the number of turns $n_{parallel}$. The total load $P$ at a given deflection $f$ is expressed as $P = n_{parallel} \cdot \frac{E \cdot b \cdot t^3 · n^4 · f}{K · D_m^3}$, where $K$ is a geometry-dependent constant. Unlike Crest-to-Crest springs, the nested design provides high forces in very limited radial and axial envelopes. However, frictional hysteresis occurs between the nested layers during compression, which can lead to energy dissipation and slightly higher loading forces during the downstroke compared to the upstroke. This damping effect must be modeled in high-frequency automotive damping systems using a modified coefficient of friction $μ_{eff}$ in the load equation.

The spring rate for a multi-turn Crest-to-Crest wave spring is calculated by treating each turn as a series of individual waves. For a spring with $N$ turns and $n$ waves per turn, the total number of active waves is $Z = n \cdot N$. Using the standard deflection formula for a curved beam, the rate is $k = \frac{E \cdot b \cdot t^3 \cdot n^4}{I_{coeff} \cdot D_m^3 \cdot N}$, where $E$ is the Young's Modulus, $b$ is the radial wall, $t$ is the material thickness, and $D_m$ is the mean diameter. The factor $1.15$ or similar empirical constants ($I_{coeff}$) are used to account for the actual boundary conditions at the contact points. In a multi-turn configuration, the total deflection $f_{total}$ is the sum of the deflections of each turn. Because the load $P$ is constant throughout the serial stack, the total rate is inversely proportional to the number of turns $N$. For precision aerospace applications, the operating stress $\sigma$ must be checked using $\sigma = \frac{3 \cdot π \cdot P \cdot D_m}{4 \cdot b \cdot t^2 \cdot n^2}$, ensuring it remains below the yield strength at the operating temperature.

End-play is the axial clearance between the retained part and the retaining ring. If excessive end-play exists, the retained part can act like a slide-hammer during machine start/stop cycles, delivering high-energy impacts to the ring. This leads to 'Impact Fatigue'. The ring will show signs of mushrooming or localized deformation on the contact face. To resolve this, designers should use 'Cusp' or 'V-shaped' wave springs behind the spiral ring to take up all axial clearance, ensuring the ring is always under a slight preload and preventing the kinetic energy buildup of moving parts.

In reciprocating pumps, the rapid reversal of axial loads can cause the spiral ring to 'chatter' in the groove. This creates a hammered effect, where the ring repeatedly impacts the groove walls. This dynamic loading can lead to fatigue of the groove land even if the static load is well within limits. Furthermore, if the frequency of the pump matches the natural frequency of the ring, resonance can occur, causing the ring to vibrate out of its seat. Increasing the ring's thickness $t$ to increase its natural frequency and using a 'Tight Fit' groove are the standard mitigations.

Centrifugal lift-off occurs when the outward radial force $F_c = m \omega^2 r$ generated by rotation exceeds the inward spring tension of the ring. The ring expands, loses contact with the groove bottom, and then is easily ejected by any axial thrust. This is a common failure in high-RPM electric motors. The failure is 'clean'—the ring is often found intact but outside the groove. The solution is either a heavier cross-section ring to increase the spring force or a 'Self-Locking' ring design which uses a tab to physically limit expansion.

Cracking during installation is almost always due to over-expansion (external) or over-compression (internal) beyond the material's rupture point. For high-carbon steel, this is often linked to 'Quench Cracks' from manufacturing or hydrogen embrittlement. If the crack surface is crystalline and bright, it is a brittle fracture. The installation stress $S_{inst}$ must be recalculated. For example, if $S_{inst} = \frac{E t (D_g - D_i)}{(D_g - t) D_i}$ exceeds the UTS, the ring will fail. Using a more ductile material like 302 Stainless or reducing the radial wall width can prevent this.

Groove dishing is characterized by the groove wall becoming angled rather than perpendicular to the shaft axis. This is caused by a thrust load exceeding the groove material's yield strength. As the wall deforms, the spiral ring begins to twist ('dish'). The radial force component, which normally keeps the ring seated, now develops a vector pointing out of the groove. If the ring's angle exceeds approximately $7$ to $10$ degrees, it will lose all grip and pop out. Engineers should look for a 'shiny' wear pattern on the outer edge of the ring as a precursor to this failure.

Fretting corrosion appears as reddish-brown (for steel) or black (for stainless) debris around the contact points of the wave crests. It is caused by small-amplitude oscillatory motion between the spring and the mating surface. This motion removes the protective oxide layer, leading to continuous oxidation and wear. In aerospace actuators, this can lead to 'frozen' assemblies. The remedy is to increase the preload to stop the movement or to plate the spring with a sacrificial or lubricating layer like Silver or Cadmium (where permitted).

'Shingling' occurs in multi-turn wave springs when the waves of adjacent turns slide over one another instead of remaining crest-to-crest. This usually happens because the spring is not properly guided by a shaft or bore, or because the radial wall $b$ is too thin relative to the wave height. Once shingling occurs, the spring rate doubles or triples instantly, and the spring can no longer reach its design work height. Diagnosis is easy: the spring will appear 'telescoped' or tilted upon disassembly. The fix involves increasing the radial wall or improving the guidance tolerances.

This phenomenon is known as 'Set' or 'Relaxation'. If the spring is compressed to a height where the internal stresses exceed the material's elastic limit, plastic deformation occurs. For a new design, engineers should perform a 'Preset' operation where the spring is compressed to its solid height (or maximum work height) during manufacturing. This 'cold works' the crests and induces beneficial compressive residual stresses. If a spring loses height in the field, it indicates the operating stress $S$ was underestimated or the operating temperature exceeded the material's limits.

Fatigue failure in wave springs starts at the point of maximum tensile stress, typically the inner diameter of a crest. Under Scanning Electron Microscopy (SEM), the presence of 'striations'—parallel lines marking each load cycle—confirms fatigue. If the striations are widely spaced, it indicates high-stress, low-cycle fatigue. If very fine, it is low-stress, high-cycle fatigue. In 17-7PH, failures are often accelerated by 'Alpha-prime' martensite embrittlement if the heat treatment was improper, which would be visible as a more granular fracture surface near the initiation site.

Noise in wave spring assemblies is usually caused by 'Stick-Slip' friction between the spring crests and the mating surfaces or between the spring $O.D.$ and the housing. This occurs when the spring is under-lubricated or when the surface finish of the housing is too rough ($R_a > 1.6 \mu m$). As the spring compresses, the waves must move slightly; if they stick and then suddenly slip, high-frequency vibrations are excited. The solution is applying a dry-film lubricant like $MoS_2$ or increasing the axial preload to prevent micro-movements.

In rotating assemblies like automatic transmissions, centrifugal oil pressure can build up behind the retaining ring. If the ring does not have sufficient 'drainage' (i.e., the gap between the ring ends), the pressure can force the ring out of the groove. Design engineers must ensure the ring's radial wall does not completely block oil flow-through paths. Additionally, the centrifugal expansion of the ring itself (see Q62) must be modeled alongside the hydraulic pressure to ensure the net radial force remains inward.

Spiral retaining rings are inherently more tamper-resistant than stamped rings because they lack the eyelets used by standard circlip pliers. Removal requires a small screwdriver or dental pick to 'unwind' the first turn from the groove. For high-security applications, the ring can be designed with a 'removal notch' that is only accessible through a specific port in the housing. This makes them ideal for consumer electronics and military hardware where field-disassembly is discouraged.

In automated systems, seating can be verified through either vision systems or mechanical probing. A vision system checks for the presence of the 'gap' or the overlap ends at the correct radial position. A mechanical probe can measure the axial position of the ring; if the ring is not seated, it will sit higher than the nominal groove location. For safety-critical automotive drivelines, a 'torque-to-turn' test of the retained component can also indicate if a ring is improperly seated, as the resulting misalignment would cause abnormal friction.

A radius at the bottom of the groove is often necessary for stress concentration reduction in the shaft, but it reduces the effective contact area for the retaining ring. The allowable thrust load $P$ must be derated if the radius $R$ exceeds $10\%$ of the groove depth $d$. The modified capacity is $P' = P (1 - \frac{R}{d})$. If a large radius is required for shaft fatigue life, a 'heavy-duty' ring with a larger radial wall or a square-edge shim must be used to ensure the ring does not climb out of the groove.

Internal spiral retaining rings are installed by compressing the ring to a diameter smaller than the bore. This is done by 'winding' the ring, essentially reducing its circumference. Unlike stamped circlips, spiral rings do not have lugs for pliers. They are often installed using a tapered mandrel and a plunger. The mandrel compresses the ring as it is pushed through, and it snaps into the groove. Care must be taken not to over-compress the ring beyond its elastic limit, calculated by $S_{comp} = \frac{E t (D_i - D_g)}{(D_i - t) D_g}$, to avoid loss of 'cling' in the groove.

Wave springs require flat, parallel mating surfaces to distribute the load evenly across all wave crests. If the mating surface is concave, the inner diameter of the spring will be loaded more than the outer diameter, creating a moment that can cause the spring to twist. The parallelism should be within $0.02$ mm per $25$ mm. In many automotive clutches, hardened and ground shims are placed between the wave spring and cast-iron housings to provide a suitable mating surface and prevent the spring from 'digging' into the softer material.

Nested wave springs must be installed such that each turn is perfectly nested within the other. If the waves are out of phase, the spring will function as a series stack or a chaotic hybrid, resulting in a much lower spring rate. Most manufacturers provide nested springs pre-aligned. During assembly into a valve or seal, the technician must ensure no debris enters the nested layers, as even a small particle can prevent the waves from nesting fully, creating a 'thick spot' that alters the load at the specified work height.