Lock Washers: Do They Actually Work? (And What to Use Instead)

Reach into any toolbox drawer and you'll find a handful of split-ring lock washers — the little spring-steel rings with a gap and a twist. Most of us drop one under a nut without thinking, trusting it to keep the bolt from rattling loose. But does it actually do anything? The short answer from decades of fastener engineering may surprise you: on a properly torqued joint, a split lock washer does very little — and there are far better ways to keep a bolt tight.

What a Lock Washer Is Supposed to Do

The theory behind that little spring ring.

A threaded fastener stays tight because of preload — the clamping force stored in the stretched bolt. A split lock washer is meant to act like a spring that keeps tension on the joint and digs its sharp ends into the nut and the surface to resist backing off. That logic works on a loose thumbscrew. The problem is that on any joint torqued to a real spec, the bolt's own preload flattens the washer completely — and a flattened spring can't spring.

Do Split Lock Washers Actually Work?

Independent vibration testing — most famously the Junker test, and the guidance NASA publishes for its own fasteners — has shown repeatedly that helical spring (split) lock washers are largely ineffective at preventing self-loosening once a bolt is properly torqued. NASA's fastener design standard explicitly advises against relying on them. Some tests have found they can even aid loosening, because the ramped gap acts like a tiny inclined plane the nut can walk down. In short: by the time the washer is doing its supposed job, it has already been crushed flat.

Why a Tightened Nut Wants to Unwind

The thread is a ramp, and a ramp always has a downhill.

Unroll one turn of a thread and it stops looking like a screw and starts looking like a simple machine: an inclined plane wrapped around a cylinder. The base of that ramp is the circumference the nut travels in one turn (π·d); the rise is the thread pitch, p — how far the nut advances axially in that same turn. Preload, F, is the clamping force locked into the joint once the bolt is torqued, and it acts straight down the bolt axis, onto that ramp.

Because the surfaces are inclined, F never points purely "into" the joint. A component of it always resolves along the ramp itself, trying to walk the nut back down — the same reason a parked car rolls downhill without the brake on. That component is what generates an unwinding torque:

T_unwind = F · p / (2π). For an M8×1.25 stud carrying 10,000 N of preload, that works out to about 1.99 N·m (≈2.0 N·m) of constant, ever-present torque trying to spin the nut loose — every second the joint exists, whether it's moving or sitting still.

Friction along the same surfaces resists that unwinding torque. It acts over the thread flanks and the under-nut/washer bearing face, and because a standard thread flank is angled 30° from vertical, the thread-side contribution gets divided by cos 30° before it's added to the bearing-face contribution:

T_friction = F · μ · ( r_t / cos 30° + r_n ), where r_t is the effective thread radius and r_n the effective under-nut bearing radius.

Set T_unwind equal to T_friction and solve for μ, and you get the friction coefficient below which a joint unwinds on its own with no vibration at all — the critical friction coefficient, μ_crit = p / (2π · (r_t/cos 30° + r_n)). For this M8 stud that's about 0.02 — roughly a sixth of the μ≈0.12 that dry, unlubricated steel-on-steel threads actually deliver.

The 6-to-1 Paradox

If friction outnumbers the unwinding torque six to one, why does anything ever loosen?

Plug μ = 0.12 into the friction formula and the same M8 joint returns 11.73 N·m of resisting friction torque against just 1.99 N·m trying to spin it loose — a 5.9 : 1 margin. By that math the nut should never move. And on a joint that only ever sees axial load, it doesn't.

The margin only collapses to 1:1 right at μ_crit ≈ 0.02 — a coefficient closer to a lubricated or worn thread than a dry one. So what actually causes loosening at μ = 0.12? Transverse vibration — side-to-side motion perpendicular to the bolt axis, not rotation around it. While the mating surfaces are momentarily sliding sideways against each other, they cannot simultaneously grip and resist rotation the way they do when sitting still; friction along the ramp drops toward zero for that instant. With nothing to hold it back, the ever-present 1.99 N·m unwinding torque wins, and the nut rotates a fraction of a degree. Repeat that a few hundred or a few thousand times — exactly what a shaker table (or a running engine, or a wheel over washboard) delivers — and the preload is gone in increments too small to feel one at a time. This is precisely what the Junker test (Gerhard Junker's 1969 SAE paper, since codified into DIN 65151) was built to reproduce and measure.

Lock Washer & Locking Methods Compared

Wedge-locking washers are worth a special mention because they're the one washer that genuinely works. They install as a pair: radial teeth on the outer faces bite into the bolt head and the joint, while the two inner cam faces lock against each other. Because the cam rise angle is greater than the thread's pitch, any attempt to rotate loose would have to lift — and stretch — the bolt, so the joint stays put.

What Actually Keeps a Bolt Tight

If a split washer isn't the answer, here's what is, in order of how often it matters:

When a Split Lock Washer Is Still Fine

They aren't useless everywhere. On low-stress, low-vibration hardware — a bracket, a cover, a bin of shop fixtures — a split washer adds a little take-up and a hardened bearing surface, and it's cheap. Just don't rely on one to hold a wheel, a control arm, or a caliper. For those, use the right torque and the maker's specified locking method.

Frequently Asked Questions

Bottom line: the best "lock" is the right clamping force. Find the exact torque and bolt spec for your vehicle with our free Vehicle Fastener Specs tool, brush up on strength ratings in the Bolt Grade Guide, confirm thread sizing in the Thread Pitch Guide, or grab the printable torque chart.