01 Overview
Fatigue failures almost never begin at the point of maximum nominal stress. They begin at the point of maximum local stress, which is a function of geometry, surface condition, and microstructure rather than load alone. When a designer selects a steel for a fatigue-loaded part, the instinct is to reach for the strongest grade available. That instinct is understandable, but it is frequently wrong. Strength and fatigue resistance are related, yet the relationship is not linear, and it can invert once surface condition and notch sensitivity enter the picture.
The core issue is that fatigue is a local phenomenon. A crack initiates at a stress concentration, a machining mark, an inclusion, or a decarburized layer, and then propagates under cyclic loading. A very high-strength steel resists plastic deformation well, but it also tolerates far less stress concentration before a crack initiates. The same hardness that raises tensile strength can make a surface scratch or a sharp fillet far more damaging than it would be in a softer, tougher grade.
This article treats steel selection as a system decision rather than a materials-only decision. The goal is to help designers and machinists reason about how alloy choice, heat treatment, machining practice, and inspection interact. Shop safety matters throughout: heat treatment, grinding, and crack inspection all carry real hazards, and no fatigue improvement justifies bypassing guarding, ventilation, or personal protective equipment.
02 Designing for the process
A fatigue-critical part should be designed around the processes that will produce it, not adjusted afterward. If a shaft will be turned, heat treated, and ground, the drawing should anticipate the stock removal needed to clean up distortion after quenching. If a fillet will be machined, the drawing should specify a radius that the chosen tool can actually generate without chatter or rubbing. Fatigue life is often lost in the gap between what the drawing implies and what the process can deliver.
Stress concentrations deserve explicit attention. Sharp internal corners, undercuts, keyways, and abrupt section changes all raise local stress. A generous fillet radius, a gradual taper, and a relieved transition cost little in function but can dominate fatigue life. Designers should also consider how the part will be held and machined, because a feature that forces an interrupted cut or a thin unsupported wall invites chatter, which leaves a rough surface that behaves like a row of small notches.
Surface condition is a design variable, not a finishing detail. A turned surface with visible feed marks has a different fatigue behavior than a ground or burnished surface of the same nominal dimensions. Specifying surface finish and inspection on fatigue-critical zones is as important as specifying the alloy. Safety enters here as well: grinding and polishing operations produce dust and sparks, and any finishing step must be performed with proper guarding, extraction, and eye and respiratory protection.
03 Material and tolerance decisions
The common mistake is to equate higher hardness with better fatigue performance. Up to a point, increasing strength raises the fatigue limit, but the gain flattens and then becomes fragile. High-strength steels are more sensitive to inclusions, decarburization, and surface defects. A cleaner, moderately strong steel with a well-controlled surface often outperforms a harder, dirtier one. Through-hardening a large section can also leave a tough core and a hard case, which is useful, but only if the transition is gradual and the surface is not over-hardened.
Tolerance decisions interact with fatigue in ways that are easy to miss. A press fit that is too tight raises local stress and can initiate cracking at the bore edge. A fit that is too loose allows fretting, which is a potent fatigue initiator. Tolerances on fillet radii, chamfers, and surface finish are therefore functional, not cosmetic. The designer should identify which tolerances protect fatigue life and which are merely convenient, and then hold the critical ones tightly while allowing reasonable latitude elsewhere.
Heat treatment must be specified with the final properties in mind. Quenching and tempering, case hardening, nitriding, and shot peening all change surface residual stress, which strongly affects crack initiation. A compressive residual stress at the surface is beneficial, while a tensile residual stress from aggressive machining or poor heat treatment is harmful. The material callout, the heat treatment, and the finishing sequence should be written as one coherent specification, not three independent notes.
04 The machining operation
Machining sets the surface that fatigue will see. Rough turning with a heavy feed leaves a periodic profile that acts as a stress raiser, and the depth of that profile can exceed the depth of any beneficial residual stress. A finishing pass with a sharp, correctly ground tool and a modest depth of cut produces a smoother surface and a more favorable residual stress state. Tool wear matters: a dull edge rubs and burnishes unevenly, producing a surface that looks acceptable but varies in hardness and stress.
Heat generated during machining can also alter the near-surface microstructure. Aggressive speeds and feeds, or a lack of coolant where it is needed, can temper or even re-harden a thin layer, creating a soft or brittle zone that becomes a crack initiation site. For fatigue-critical parts, it is often worth leaving a small finishing allowance and taking a light, controlled final cut rather than trying to achieve the final dimension in one heavy pass. Grinding, if used, must be dressed and cooled properly to avoid burning, which leaves temper colors and tensile residual stress.
Safety is inseparable from these operations. Rotating workpieces, sharp edges, hot chips, and airborne dust all present hazards. Guards must be in place, work must be securely clamped, and personal protective equipment must be worn. No fatigue specification is worth an injury, and a machinist who is rushed or unprotected is more likely to produce a poor surface as well as to get hurt.
05 Measurement and quality control
Quality control for fatigue-critical parts goes beyond dimensional inspection. Dimensions confirm fit and function, but they do not reveal a crack, a decarburized layer, or a burned surface. Surface finish should be measured on the fatigue-critical zones, not just sampled somewhere convenient. A profilometer reading taken on a non-critical face tells you little about the fillet where the crack will actually start.
Crack detection is the key check. Magnetic particle inspection works well on ferromagnetic steels and can reveal surface and near-surface indications, particularly in fillets and at section changes. Dye penetrant inspection is useful where magnetic methods are impractical, though it detects only surface-breaking defects. For very high-risk parts, additional methods such as ultrasonic inspection may be warranted. The inspection plan should name the specific zones to be examined and the acceptance criteria, because a general note to inspect the part is not a plan.
Process control supports inspection. Hardness checks, microstructure examination on coupons, and verification of heat treatment records help confirm that the material actually received the specified treatment. Dimensional inspection should include the features that control stress concentration, such as fillet radii and chamfer angles. Records should be retained so that a field failure can be traced back to a specific lot and process route. Inspection is not a substitute for good design and machining, but it is the last line of defense against a fatigue failure.
06 Troubleshooting bad parts
When a fatigue-loaded part fails or fails inspection, the first question is where the crack started. A crack that begins at a fillet points to a stress concentration or a machining mark. A crack that begins at a bore or a keyway points to a fit, a sharp corner, or fretting. A crack that begins in the middle of a smooth surface suggests an inclusion, a decarburized layer, or an unexpected residual stress. The location of initiation is the most useful clue available.
If the failure is at a fillet, check the radius against the drawing and inspect the surface finish. If the radius is undersized or the surface is rough, the fix is geometric and procedural. If the failure is at a fit, review the interference and the surface condition under the fit, since fretting fatigue often leaves a characteristic dark, oxidized debris. If the failure is subsurface, examine the material for inclusions or an improper heat treatment, and verify that the specified hardness and microstructure were actually achieved.
Corrective action should address the root cause, not just the symptom. Increasing the fillet radius, improving the surface finish, changing the heat treatment, or adding a compressive surface treatment such as shot peening may each help, but only if the diagnosis is correct. Replacing the steel with a stronger grade without understanding the initiation site often makes the problem worse. Each failure should feed back into the design, the process specification, and the inspection plan, and safety procedures should be reviewed whenever a failure involves a part that could injure someone.
07 Frequently asked questions
Does a stronger steel always give better fatigue life? No. Strength helps up to a point, but high-strength steels are more sensitive to notches, inclusions, and surface defects. A cleaner, moderately strong steel with a good surface and favorable residual stress can outperform a harder grade. The controlling factors are local stress, surface condition, and microstructure, not tensile strength alone.
How much does surface finish matter? It can matter more than the alloy choice. A rough turned surface acts as a series of stress raisers and can reduce fatigue life substantially compared with a ground or polished surface of the same dimensions. For fatigue-critical zones, specify and measure the finish rather than assuming it.
Should I change the material if a part keeps cracking? First find where the crack starts. If it starts at a fillet, a fit, or a surface defect, changing the alloy may not help and could hurt. Fix the geometry, the fit, the surface, or the heat treatment first, then reconsider the material only if the evidence points there.
Is shot peening a substitute for good design? No. Shot peening introduces beneficial compressive stress and can improve fatigue life, but it cannot compensate for a sharp corner or a poor fit. It is an addition to good design, not a replacement for it, and it must be controlled and inspected like any other process.