Defaults reproduce the published DIN figures exactly: 200 N/mm², no safety factor, sharp corners. Change any of them and the groove number moves; the ring number cannot, because it is catalogue data rather than a calculation.
A circlip retains an axially-loaded part by seating in a groove machined into a shaft or bore. The groove diameter sets how deep the groove is cut, and the depth is what carries the thrust: the ring presses on a flat annular wall, and that wall's area times the material's yield strength is the whole load rating. The width sets how far the ring can float, which matters more than it looks.
A 25 mm shaft takes a DIN 471 external circlip with a groove diameter of 23.9 mm — 0.55 mm deep. The bearing face is 42.2 mm², so at the standard's assumed 200 N/mm² the groove carries about 7.0 kN. Cut the same groove in C45 shaft steel and it carries 10.9 kN — a 55% gain for nothing but a different line on the drawing. The ring's own 16.2 kN is still well clear of that, and stays clear: this size would need a groove material of 462 N/mm² before the ring became the limit, so on a 25 mm shaft the groove is what stops you whatever you make it from.
Whichever is lower. Groove strength is the shaft or bore material yielding under the ring's edge, and this page computes it from your material. Ring strength is the ring deforming or rolling out of the groove, and it comes from the ring maker's testing. On the standard's own soft assumption the groove goes first on every one of the 41 sizes, which is where the folklore comes from. Put a real material in and that stops holding: at S355 the ring governs 15 of them — 4 of the 21 external sizes but 11 of the 20 internal.
Because it isn't a shear failure. Running the published ring figures back through a shear formula gives an implied strength scattering by more than 20% and averaging under 200 N/mm² — a fraction of what spring steel shears at. The column describes the ring dishing and lifting out, which depends on the ring's tapered section rather than on any dimension the standard publishes. Computing it here would be invention.
If the groove sits on the outside of a shaft, protruding outward to stop something sliding off the end, that's external (DIN 471). If the groove sits inside a bore or housing, protruding inward to stop something sliding out, that's internal (DIN 472). The nominal size always matches the shaft or bore diameter at the groove — not the diameter of whatever part is being retained.
More than most drawings admit. A 0.2 mm tool nose radius in the groove corner removes about 39% of a 20 mm shaft's rating, because it eats the bottom of the bearing face while the groove still measures full depth. Set the corner radius above and watch the number move. A chamfer on the retained part is a separate and larger loss that needs the ring maker's data.
No — groove diameter, width and depth are all matched to the specific nominal size. A circlip sized for a different shaft either won't seat correctly or won't develop its rated spring force in the wrong groove, undermining the whole retention mechanism.
A groove wall is a flat ring of metal, and the thrust it carries is just its area times the stress the material stands. Nothing about the published table is measured — every row is this sum, which is why you can redo it for your own material instead of accepting the one the standard assumed.
| Worked at 20 mm | Value |
|---|---|
| Shaft diameter | 20.00 mm |
| Groove diameter | 19.00 mm |
| Groove depth | 0.50 mm — half the difference, always |
| Bearing face | π/4 × (20² − 19²) = 30.63 mm² |
| At 200 N/mm² | 6.13 kN if the ring pressed on all of it |
| × 0.83 open ring | 5.08 kN — the published figure is 5.0 |
Why 0.83? A circlip is an open ring. It cannot press on the whole circle, so only part of the groove wall carries the load. Reading that fraction back out of the published table across all 41 standard sizes gives 165 N/mm² against the 200 N/mm² the standard names — consistent to better than 1%, and exactly where ring makers put a correction factor of their own.
The standard figures are worked at 200 N/mm², which is softer than almost anything a shaft actually gets made from. Capacity is directly proportional to yield strength, so the correction is a single multiplication — the same multiplier at every size.
| Groove material | Yield N/mm² | vs table | |
|---|---|---|---|
| DIN table reference | 200 | 1.00× | the softness the published figures assume |
| S235JR mild steel | 235 | 1.18× | EN 1993-1-1 |
| Aluminium 6082-T6 | 250 | 1.25× | a housing that dishes before it shears |
| S275 structural | 275 | 1.38× | EN 1993-1-1 |
| C45 / EN8 normalised | 310 | 1.55× | the usual shaft steel |
| S355 structural | 355 | 1.77× | EN 1993-1-1 |
Whether the extra strength buys you anything depends on the size, and mostly it does not. Taking the groove material all the way to S355 puts the ring in charge on 15 of the 41 standard sizes — but only 4 of the 21 external ones, against 11 of the 20 internal. A shaft almost always stays groove-limited; a bore often does not, because the internal ring is the thinner of the pair at most sizes. The earliest crossover anywhere is a 30 mm internal, where the ring takes over from 244 N/mm² — barely above mild steel.
Grey cast iron does not belong on this list. It has no yield point — it fractures instead of deforming, so a figure derived from yield strength does not describe how it fails. Housings in cast iron need the foundry's own data.
Every published figure assumes a groove with sharp corners. Real grooves are cut with a tool that has a nose radius, and that radius eats the bottom of the bearing face. The groove still measures the right depth — it just stops carrying load over part of it.
| Shaft | Depth mm | Sharp | 0.2 mm radius | Lost |
|---|---|---|---|---|
| 6 mm | 0.15 | 0.46 kN | nothing left | 100% |
| 10 mm | 0.20 | 1.02 kN | nothing left | 100% |
| 20 mm | 0.50 | 5.08 kN | 3.08 kN | 39% |
| 30 mm | 0.70 | 10.70 kN | 7.69 kN | 28% |
| 50 mm | 1.50 | 37.94 kN | 33.02 kN | 13% |
| 80 mm | 1.75 | 71.41 kN | 63.41 kN | 11% |
| 100 mm | 1.75 | 89.67 kN | 79.58 kN | 11% |
The same tool costs a small shaft everything and a large one almost nothing. One 0.2 mm nose radius throws away 39% of a 20 mm groove and 11% of a 100 mm one, because depth grows with size and the radius does not. On a 6 mm shaft the groove is only 0.15 mm deep, so that tool removes the bearing face altogether.
This is the groove side only. A chamfer on the retained part is a separate loss and a larger one — it tips the ring and lets it roll out of the groove. Quantifying that needs the ring maker's own maximum-chamfer figure, which is not derivable from the standard.
The standard groove is deliberately wider than the ring that goes in it. Across every size in both standards the difference is 0.10 to 0.15 mm — 12 of the 21 external sizes sit at the smaller figure. Made exactly to nominal, with nothing worn and nothing out of tolerance, the retained part can still move that far along the shaft.
Under a steady load this does not matter. Under a reversing one it is the whole problem. The gap closes at speed each time the load changes sign, so the ring arrives at the groove wall as an impact rather than a push. Thrust ratings are static figures and describe none of that — a joint that reverses wants a shouldered shaft, a wavy washer to take up the float, or a ring rated for the job, not a bigger circlip.
And the figure only goes up from here: groove width carries a plus tolerance and ring thickness a minus one, so a conforming pair can float meaningfully further than 0.10 mm.
Releases in which this page changed, newest last. Derived from the archived copy of every release, not from notes written afterwards — so it reflects what actually shipped. Site-wide passes are left out; they are in the full changelog.