Gearboxes fail at the bearings

Why the supporting components outlive neither the gears nor the housing.

A gearbox usually fails at its bearings first because every force the gears generate, including the sideways separating push that tooth contact creates in addition to the useful turning force, ultimately has to be carried by whatever is holding the shafts in place, and the bearings doing that job are routinely chosen from a standard catalogue sized to just meet the load rather than engineered with the same generous margin the gear teeth themselves usually enjoy.

Why the humble part carries the whole load

Gear teeth get the design attention a visible, load-carrying component naturally attracts, since a broken tooth is dramatic and immediately obvious, which tends to push designers toward comfortable safety margins on tooth strength almost by instinct. The bearings supporting the shafts those gears turn on receive comparatively less individual attention, often selected as an off-the-shelf part sized against a published load and life rating rather than custom-engineered with the same margin, even though those bearings are the ones actually absorbing every reaction force the gear mesh produces, the useful turning force and the sideways push neither the tooth profile nor the shaft itself was ever asked to resist directly.

That sideways push is easy to forget precisely because it produces no useful work of its own, unlike the turning force everyone is actually trying to transmit, so it rarely gets the same deliberate attention during design even though the bearings have no choice but to resist it in full on every single revolution the gearbox ever makes. A helical gear pair, chosen earlier in this set specifically to run quieter than an equivalent spur pair, makes this particular problem worse rather than better, since its angled teeth generate an additional axial thrust load along the shaft that a straight-cut spur tooth never produces at all, a genuine trade a designer chasing quieter gears has to account for in how the bearings are then selected. That thrust load has to be carried by a bearing actually capable of resisting force along the shaft's own axis rather than only around it, which is why a gearbox built around helical gears commonly pairs them with angular contact or tapered roller bearings chosen specifically for that axial capacity, rather than the simpler deep-groove bearings that would have been perfectly adequate for an equivalent spur gear pair carrying no thrust load at all.

The drawer-runner comparison

A well-built wooden drawer can look, and largely be, as solid on its hundredth opening as it was on its first, the box itself barely showing any wear at all, while the small metal rollers or plastic runners it slides on quietly absorb every bit of the sliding motion's actual wear, gradually roughening, loosening or gathering grit until the whole drawer starts sticking or rattling long before the drawer box itself shows any comparable sign of age. Nobody admires a drawer's runners the way they might admire its joinery, yet the runners are doing all of the actual mechanical work of supporting a moving load through repeated cycles, exactly the relationship a gearbox's bearings have to its gears, an unglamorous supporting part quietly absorbing the wear a more visible, more admired component is spared. A drawer's worn runners can often be replaced in a few minutes with a cheap spare part, the drawer box itself needing no attention at all, and a gearbox's bearings are very often designed with exactly that same repairability in mind, sized and mounted so a single bearing can be pressed out and swapped during a service interval without disturbing the gears, the shafts or the housing surrounding it.

Why fatigue makes this worse over a long service life

A rolling-element bearing's inner race is subjected to the same contact point passing under load thousands of times per minute during ordinary operation, and that repeated cyclic loading is a textbook setup for fatigue failure, a slow accumulation of microscopic damage that eventually surfaces as pitting or spalling on the race, well after the bearing has already delivered a long, apparently trouble-free service life. Gear teeth experience a broadly similar cyclic loading themselves, but a well-cut tooth in a properly lubricated mesh generally tolerates it for a considerably longer working life than an equivalently loaded bearing does, simply because the contact geometry, contact area and typical safety margins applied to each component differ enough in practice that the bearing usually reaches its own fatigue limit first.

The number that matters here

A gearbox's bearings are frequently selected against a catalogue-rated life measured in a defined number of hours of continuous running at a specified load, a number chosen to be adequate rather than generous, while the gears meshing inside the same housing are commonly designed with a considerably larger margin above their calculated minimum tooth strength. That asymmetry in how each component's margin was actually chosen, not any inherent weakness in bearings as a class of part, is a large part of why a well-designed gearbox so often gives its first sign of real trouble at a bearing rather than at a tooth, a pattern that catches out anyone assuming the more complicated, more precisely machined part must automatically be the one that fails first.

What follows from this

Servicing a gearbox with this asymmetry in mind means paying closer attention to bearing condition, vibration, noise, play in the shaft, than to gear tooth condition specifically, since the bearings are statistically the more likely first point of failure precisely because of how each component's margin was originally set rather than because bearings are inherently less durable as a class of part. It also argues for choosing a bearing with a real margin above the calculated minimum rating during design, rather than the smallest catalogue part that technically satisfies the load calculation, closing the very gap in margin that makes bearings the predictable weak point in the first place.

Where this stops being true

A gearbox subjected to genuinely severe shock loading, sudden reversals or occasional overload spikes well beyond its steady rated capacity, can just as easily break a tooth first, since fatigue-driven bearing wear assumes a reasonably steady, predictable load pattern, and an occasional sharp overload plays instead to whichever component has the least margin against a single large event rather than the least margin against many repeated small ones. The general pattern favours the bearings as the more likely failure point under ordinary, steady use, and an unusual loading history can just as easily overturn that pattern entirely, which is exactly why the specific application a gearbox is destined for, steady running or occasional shock, deserves to shape which component actually receives the larger design margin rather than defaulting to the same assumption every time.