How much load can a circlip take?

Two limits, and you get the lower one

A retaining ring assembly has two independent thrust capacities. The first is the load at which the ring itself shears. The second is the load at which the groove wall yields and permanently deforms. The usable capacity of the joint is whichever of those is lower — not the ring's own rating, which is the number people tend to quote.

Ring manufacturers publish both. Smalley notes that the thrust load based on ring shear must be compared to the thrust load based on groove deformation to determine which is the limiting factor, and that groove deformation is by far the most common design limitation. Ring shear only becomes the governing case when the groove is cut in hardened steel.

The practical consequence: a ring rated well above your load will still fail if the groove is cut in soft material, cut shallow, or cut with rounded walls. Choosing a heavier ring does not raise the capacity of a joint that is groove-limited.

How a ring actually fails

The failure is progressive rather than sudden, which is why it often looks like the ring simply fell off. As permanent groove deformation occurs the ring begins to twist. As the angle of twist increases the ring enlarges in diameter. Eventually it becomes dished and extrudes — rolls — out of the groove.

By the time the ring is on the floor the groove has usually been out of tolerance for a while. This is the reason a ring that keeps coming loose is almost never fixed by fitting another one of the same size: the new ring inherits the same deformed groove and takes the same path.

Safety factors are already in the published numbers

Manufacturer thrust tables are not ultimate values. Smalley's published ring-shear capacities already include a safety factor of 3, and a factor of 2 is recommended for groove deformation analysis. Applying your own factor on top of a catalogue figure therefore double-counts, and treating a catalogue figure as an ultimate limit under-counts.

The tabulated groove figures also assume a groove material. Values are quoted against a stated yield strength — cold drawn 1018 steel at around 70,000 psi is a common baseline. A groove cut in aluminium, brass or a free-machining grade will yield well below the tabulated load, and needs recalculating rather than derating by feel.

The retained part matters as much as the ring

This is the detail most often missed. Thrust ratings assume the retained component presents a sharp corner to the ring, because a sharp corner contacts the protruding portion of the ring and gives the best load support.

Where the retained part has a chamfer or a corner radius instead, contact moves outward toward the edge of the ring. Rotor Clip describes this as creating a lever action against the loaded groove wall, which under high load deflects the ring and leads to failure. A bearing with a large chamfer bearing against a circlip is therefore a materially weaker joint than the same ring against a square-shouldered part, at the same nominal rating.

If a joint is close to its limit, the cheapest fix is often not a bigger ring. It is a spacer or washer that presents a square face to the ring and moves the contact back where the rating assumed it was.

Groove geometry, briefly

Both groove walls should be parallel to each other and perpendicular to the axis of the shaft or housing. Depth must be held to specification, because the depth and the groove bottom radius together determine how much support the load-bearing wall provides.

Edge margin — the distance from the groove to the end of the shaft or housing — is calculated to leave enough material to resist shearing at maximum load. Moving a groove closer to the end than the standard allows removes that margin, and the failure mode becomes the shaft end shearing away rather than anything to do with the ring.

When a circlip is the wrong answer

Retaining rings are a shoulder, not a clamp. They resist axial movement and do nothing about rotation, so anything that must not turn needs a key, spline, pin or interference fit regardless of the ring.

They also assume the load arrives along the axis and square to the ring. Where a joint sees significant reversing load, vibration with a clearance gap, or load applied off-square, the ring is being asked to do something its rating does not cover. In those cases a shoulder machined into the shaft, a locknut, or a bolted retainer is the honest answer.