A loose bolt breaks and a tight one does not

Why insufficient preload is the common cause of fastener failure.

A loose bolt breaks and a tight one does not because a properly preloaded bolt is already carrying most of a joint's clamping load before any external force ever arrives, leaving very little extra stress for that external load to add, while a loose bolt is carrying almost none of it in advance, forcing the bolt itself to absorb the full swing of every external load directly and repeatedly until it eventually fatigues and snaps.

Preload as a buffer against every passing load

A correctly tightened bolt holds two parts together with a steady preload that is usually considerably larger than any external load the joint will see in normal service. As long as the external load stays smaller than that preload, the clamped parts never separate or shift against each other, and most of the extra load is taken up by a small, easily tolerated drop in how hard the two parts press together. The bolt itself stretches only a little further than it already was.

The reason lies in stiffness. The clamped plates are usually two to four times stiffer than the slender bolt passing through them, so when a load tries to pull the joint apart, it is mostly the plates relaxing their squeeze that carries it. In a typical steel joint the bolt sees perhaps a fifth to a third of each passing load, and the plates absorb the rest.

A loose bolt, one never properly tightened or one that has lost its preload over time, has no such buffer. With no large clamping force already present, every external force that acts on the joint has to be resisted by the bolt stretching and flexing directly, cycle after cycle, in a way the preloaded bolt was never asked to do. Fatigue accumulates from exactly that kind of repeated stress swing, and the loose bolt meets the whole of each swing where the tight one met only a fraction of it. That is why a loose fastener so often fails not all at once under one dramatic load, but by quietly cracking through after thousands of ordinary cycles that a tight one would have shrugged off.

A luggage strap on cobblestones

A luggage strap left loose around a suitcase, slapping against the case with every bump as it is wheeled over cobblestones or down a kerb, shows the same difference in a setting anyone can feel. The loose strap takes up none of each bump smoothly. It snaps taut suddenly every time the slack runs out, delivering a sharp shock to the buckle and stitching holding it on.

The same strap buckled down tight moves as one piece with the case and barely registers the bumps, because there is no slack left for a bump to take up, and the strap simply flexes gently. Over a long walk the tight buckle sees a small, gentle load at every cobble, the loose buckle a sharp jolt at every cobble, and it is the loose one that wears through and fails first, the pattern that plays out inside a loose bolted joint every time a load passes through it.

Reading a broken bolt correctly

When a fastener fails in service, the instinctive assumption is that it must have been overloaded. More often the real cause is that it never carried enough preload in the first place, whether from under-torquing at assembly, from a joint that was never retightened after its surfaces settled, or from vibration gradually working the preload down over time.

A fatigue-cracked bolt usually shows a distinctive fracture surface, a smooth region marked with faint curved lines where the crack crept forward cycle by cycle, and a small, rougher patch where the remaining metal finally tore through. An experienced eye reads that as evidence of many small cycles rather than one big overload, and finding it on a failed bolt is a stronger clue that preload was the problem than any guess about the external load.

It is also why the neighbours of a failed bolt are worth torque-checking before the broken one is simply replaced. A single loose bolt is rarely an isolated event. A batch of joints tightened incorrectly, or a design that never allowed for how much preload an application needed to stay ahead of its cyclic loads, tends to produce a string of failures over time, and the first one noticed is simply the first.

How much preload is enough

A joint's preload is normally set well above the largest external load it expects, by a comfortable margin, so that the external load never comes close to overcoming the clamping force already holding the parts together. That margin, more than the bolt's raw strength rating, decides how many cycles the joint survives before fatigue becomes a concern. A joint tightened to only part of its intended preload has thrown the margin away, and can begin accumulating fatigue damage almost at once under loads that the same joint, correctly tightened, would have felt only as a gentle flex.

More preload is not better without limit. Tightening a bolt beyond its elastic range stretches it permanently and reduces the clamping force it can reliably hold, and an over-tightened bolt can fail as readily as an under-tightened one, snapping outright from having been pushed past its yield point instead of slowly fatiguing. The target is a specific, calculated preload, and the joints left too loose and the ones wound down too hard fail for closely related reasons: an absent or a destroyed margin between what the bolt is holding and what it can hold.