Compressed smaller than its free cross-section. Too little can’t track vibration or wear; too much shortens life through friction, heat and compression set.
The share of the groove the ring occupies, checked against a 60–90% target at service temperature. It needs headroom for thermal expansion and fluid swell.
Rubber is nearly incompressible, so stretching a ring thins the cord as 1/√(1+stretch). The groove, cut for the nominal cord, does not follow.
A static seal never moves, so it takes more squeeze; a dynamic one pays for every percent in friction and wear, so its groove is looser.
An elastomer expands about 2×10−4 per kelvin against steel’s 1.2×10−5, roughly fifteen times faster: a gland at 85% fill on the bench is near 89% a hundred degrees later, before fluid swell. Past the ceiling the ring hydraulically locks: contact pressure spikes and the only way out is the clearance gap. Below about 70% it can shift or twist. A Nitrile ring at 180 °C is the wrong compound, not a groove problem.
Volume is fixed, length goes as circumference, so area goes as 1/(1+s) and cord diameter as 1/√(1+s). That is why stretch is held to about 5%: past it the cord is thinned enough to matter and the rubber sits under permanent tension. Compressed more than about 3% into a smaller circle, a ring buckles or twists in its groove: a leak path, not a seal.
The same step in diameter is a far bigger fraction of a small ring: 10 mm to 11 mm is 10% stretch, the same 1 mm on a 50 mm ring is 2%. The tight, awkward small-bore seal is the one losing most squeeze to the fitting.
A ring left compressed for months does not fully spring back: expect roughly 15–20% of the original squeeze permanently lost over a typical service life at moderate temperature, more when hot. Design the initial squeeze with that margin already spent.
A harder compound resists being pushed into a clearance gap, which is why high-pressure static seals often specify 90 Shore A. The same hardness makes a poor dynamic seal: it will not conform to the mating surface and wears the bore or rod faster. Chemical compatibility can matter more than temperature: EPDM handles brake fluid and steam but degrades in petroleum oil, while Nitrile is the reverse.
The AS568 dash number pairs a standard cross-section with a standard ID for ordering. The groove depends only on cross-section and seal type, so you can type any ID and get a correct groove. A radial seal is squeezed between a shaft and a bore, the standard choice for anything that slides or rotates; a face seal between two flat surfaces closing together, like a lid.
Squeeze is what makes the seal — but it's also what makes friction. A static seal never moves, so it can afford more squeeze for a more robust seal. A dynamic seal (a rod, piston or rotating shaft) pays for every percent of squeeze in friction, heat and wear, so the design target is deliberately lower.
Above about 90% fill, the O-ring has nowhere to go as it heats up or absorbs fluid (swell) — it can hydraulic-lock in the groove, spike contact pressure, and extrude out through any clearance gap. Below about 70%, the ring can shift, twist, or fail to fully occupy the groove under pressure.
Not the groove — but it changes what the groove does. Depth and width follow the cross-section, so those are the same at any diameter. What the diameter decides is how far the ring has to be deformed to get into that groove, and rubber that is stretched gets thinner in the cord. A ring stretched 5% onto its groove floor arrives about 2.4% thinner than the one the groove was cut for, so it has less squeeze than the number on the slider. Small diameters need proportionally more stretch for the same step in diameter, so they lose the most.
About 5% is the customary ceiling for a ring installed on a shaft, and less for anything dynamic. Beyond it you are trading away cord thickness and living on rubber under permanent tension, which relaxes and cracks faster. Going the other way, more than roughly 3% of circumferential compression makes the ring buckle and twist in its groove instead of sitting flat.
A face seal is squeezed between two flat surfaces closing together, like a lid — simple to design and assemble. A radial seal is squeezed between a shaft and a bore, sealing sideways — the standard choice for anything that slides or rotates, like a piston or rod.
A dash number just pairs a standard cross-section with a standard inside diameter — the groove design itself depends only on the cross-section and seal type, not the specific ID. Entering your own ID here gives an exact, correct groove for any size, standard or custom, without needing to look up a dash-number table.
Chemical compatibility can matter more than temperature — EPDM handles brake fluid and steam beautifully but degrades in ordinary petroleum oil, while Nitrile is the reverse. Picking a compound by temperature range alone can still pick a seal that dissolves in service.
Rubber compounds are close to incompressible, so squeezing a circle in one direction has to push that same volume out somewhere else — sideways, into the groove's width, is the only place left to go. The diagram conserves cross-sectional area between the free and installed shapes for exactly this reason.
This page has changed in 189 archived releases, but each of those was a site-wide pass, so none is attributable to this tool on its own and none is listed here. That is not a claim that the tool never changed — a release that reworked many pages at once may well have altered this one too. The changelog has them.