An O-ring is one of the cheapest parts in a product and one of the most common reasons a product comes back. The seal is not really made by the ring — it is made by the groove the ring sits in. Get the squeeze, the gland fill, the extrusion gap, and the surface finish right and a two-cent ring will hold for years. Get the groove wrong and no substitution of rubber will save you, because fixing it means recutting steel or reprogramming a machining operation. That is why gland design belongs in the first CAD pass, not in the week before a leak test.

Start With the Duty, Not the Catalog

Every sealing decision starts with four questions: what temperature range does the joint see, what pressure and in which direction, what fluid or gas touches the ring, and for how long. Elastomer choice follows from chemical compatibility, not from price:

  • Nitrile (NBR) — the default for oils, fuels, and water at moderate temperature. Inexpensive and widely stocked, but poor in ozone and sunlight.
  • EPDM — excellent for water, steam, and outdoor exposure. Attacked by petroleum oils, so never use it on a lubricated shaft.
  • Fluorocarbon (FKM) — high temperature and broad chemical resistance. Costs several times NBR and stiffens badly in the cold unless you specify a low-temperature grade.
  • Silicone — very wide temperature range and physiologically benign grades, but low tear and abrasion resistance. A static-only choice in most designs.
  • Perfluoroelastomer and PTFE-encapsulated rings — for aggressive chemistry and clean processes, at a large cost step.

Hardness matters too. A softer compound conforms better and seals at low pressure; a harder one resists being extruded through the clearance gap at high pressure. Durometer is a design variable, not a detail — the practical scale is explained in the Shore hardness guide.

Axial vs Radial: Two Different Design Problems

In an axial or face seal, the ring is squeezed between two parallel surfaces — the simplest and most forgiving arrangement, and the right default for a bolted enclosure lid. One rule gets forgotten constantly: when pressure acts from the inside out, the ring must be able to seat against the outer wall of the groove, so the groove is placed accordingly. Reverse the pressure and the geometry reverses too.

In a radial seal, the ring is compressed between an inner and outer diameter — appropriate for round covers, plugs, and shafts. Radial glands are less forgiving because the squeeze depends on the diametral fit of two parts, so it inherits everything wrong with your dimensional chain.

Motion changes the numbers again. A static seal, where the parts do not move relative to each other, can carry higher squeeze. A dynamic seal on a rotating shaft or a sliding piston must run lower squeeze, a finely finished counterface, and appropriate lubrication, or friction generates heat and the ring wears out fast.

Gland Design Rules of Thumb

  • Squeeze. Aim for roughly 15 to 30 percent compression of the ring cross-section. Static seals tolerate the upper end; dynamic seals live near the bottom.
  • Gland fill. The ring should occupy about 75 to 90 percent of the groove volume, leaving room for thermal expansion and fluid swell. A completely filled groove hydraulically locks and can split the housing.
  • Stretch. Installed stretch on an inner diameter should stay within a few percent. Stretch it more and the cross-section flattens, squeeze disappears, and the seal leaks.
  • Extrusion gap. As pressure rises, the diametral clearance between the mating parts must shrink, or the elastomer is forced into the gap and nibbled apart. At high pressure, add anti-extrusion backup rings.
  • Finish and lead-in. Groove surfaces should be smooth and free of radial scratches, with a chamfer or generous lead-in on any edge the ring passes during assembly so it is not sliced going in.
  • Corner radii. Sharp groove corners concentrate stress in the ring and are hard to machine cleanly. Specify small radii deliberately.

Tolerances Decide Whether the Math Survives

Perfect nominal squeeze is worthless if the tolerances allow it to vanish. The chain from lid to base to boss to fastener determines the squeeze that actually occurs, so compute the worst case, not just the nominal — the method is set out in tolerance stack-up analysis, and the drawing has to communicate it, which is what GD&T is for. Two more practical constraints: molded parts shrink and warp, so a long straight sealing face on a plastic lid will not stay flat without ribs or a raised sealing land, and the groove itself needs draft if it is molded rather than machined. Sealing an enclosure is one of the reasons electronics enclosure design is harder than it looks.

Work to standard AS568 size series wherever possible. Off-the-shelf rings in standard cross-sections are cheap, stocked everywhere, and available in every compound. A custom size locks you to one supplier at a long lead time for no functional gain.

How O-Ring Seals Fail, and How to Catch It Early

Field failures cluster into a short list: compression set, where the ring takes a permanent flat and stops pushing back after prolonged heat; extrusion and nibbling from an oversized clearance gap; spiral failure on a slow-moving piston where the ring twists instead of sliding; chemical swell or shrinkage from the wrong compound; and installation damage from a sharp edge or a missing chamfer.

Test accordingly. Do not certify on one fresh unit. Test a unit that has been opened and closed several times, a unit that has been through thermal cycling, and a unit assembled by a real operator rather than the designer. Pressure decay or immersion testing against a declared ingress rating is the acceptance gate, and it belongs in the prototype phase — the practical approach is described in building a waterproof prototype. Material compatibility should be confirmed against the actual fluid, not a similar one, which is part of the wider question of choosing materials for a product.

If your enclosure or fluid assembly is failing leak test, or you want the glands checked before the tool is cut, Projects House designs sealed assemblies and validates them on real parts. Send us the design through our contact form and we will review the groove geometry, the stack-up, and the compound choice together.