I’ve pulled seals from equipment where the orientation was “correct” on paper yet the application still leaked within weeks. The engineer picked axial or radial based on a textbook rule, but nobody checked whether the groove geometry could actually deliver the required squeeze under real tolerances. Standard radial gland dimensions are rated for 1,500 psi — yet a 70 durometer O-ring in a properly engineered radial groove has sealed 200,000 psi. That gap between “standard” and “possible” is where most orientation selection failures live, and it has nothing to do with picking the wrong label.
Radial seals default to dynamic applications while axial seals handle higher static pressures — but the real engineering challenge is the overlap zone where either orientation works and groove design determines which performs better.
How Axial and Radial Seals Actually Work
An axial seal compresses an O-ring between two flat, parallel faces — think of pressing your palms together to block water. The sealing force acts along the shaft axis. A radial seal squeezes the O-ring between a bore and a shaft (or a groove wall and a cylindrical surface), with the sealing force acting perpendicular to the axis.
The fundamental design difference comes down to one thing: the extrusion gap.
Radial seals always have a clearance between the groove wall and the mating cylindrical surface. Under pressure, the elastomer tries to flow into that gap. Without backup rings, this limits pressure capability. Axial seals eliminate that gap entirely — the O-ring seats flat against the low-pressure face of the groove with no clearance to extrude into. Fewer failure modes, fewer design steps.
Inside-mounted versus outside-mounted configurations change which surface the seal compresses against, shifting whether extrusion gap management is a factor.
Yet here’s the counterintuitive part: despite axial being simpler from a sealing standpoint, industry practice recommends radial for most static applications. The reason is manufacturing, not sealing. Radial grooves are machined on a standard lathe. Axial grooves require face machining — a more complex setup. Engineers default to radial because their machine shop prefers it, not because the seal performs better.
Operating Parameter Boundaries
A face seal needs 15–25% squeeze while a static radial gland needs 22–32% — and those ranges come from AS568 and SAE standards. I keep this table pinned above my desk:
| Application Type | Squeeze Range | Orientation |
|---|---|---|
| Face / flange seal (small cross-section) | 15–25% | Axial |
| Face / flange seal (large cross-section) | 8–14% | Axial |
| Static gland seal (small cross-section) | 22–32% | Radial |
| Static gland seal (large cross-section) | 15–20% | Radial |
| Reciprocating | 10–20% | Radial |
| Rotary shaft | 0–10% | Radial |
All values must fall within the 5–30% absolute window. Outside that range, you either lose contact or overstress the elastomer.
Why Rotary Squeeze Looks Wrong
The 0–10% range for rotary applications looks like a typo until you understand the mechanism. At shaft speeds above a few hundred RPM, the O-ring doesn’t seal through compression alone — a hydrodynamic lubricant film forms at the interface. Low squeeze is intentional because higher compression generates heat and accelerates wear. This only works when the sealing surface is finished to 16 rms or smoother. Static surfaces, by contrast, tolerate 32–63 rms because there’s no relative motion to build a film.
The Hidden Stretch-Squeeze Tradeoff
Radial seals require the O-ring to stretch over a shaft or compress into a bore during installation. That stretch directly steals squeeze capacity: the cross-section reduction is approximately half of the ID stretch percentage. Specify a 5% stretch on a piston seal and you’ve quietly lost 2.5% of your designed squeeze before the system even pressurizes. Axial seals avoid this entirely — the O-ring drops into the groove with zero stretch.
For piston seals, the maximum allowable stretch is 5%. For rod seals, it’s 2%. Exceed these limits and the O-ring thins enough to compromise the squeeze range you designed for.
When Either Orientation Works: The Overlap Zone
The axial-versus-radial choice is not always binary. A large number of static and slow-speed applications fall into an overlap zone where both orientations seal reliably — and a wrong groove dimension causes more leaks than a wrong orientation label.
A medical device manufacturer learned this the hard way. Their handheld device needed IPX7 waterproofing, so they designed a pure axial face seal: a 1.2 mm diameter gasket in a 0.9 mm deep groove, creating 0.3 mm of axial compression. Nominally, gland fill sat at 90% — within the 75–85% target. But at worst-case tolerances of plus/minus 0.05 mm, gland fill hit 115%. The cases wouldn’t close. First-pass yield dropped, rework costs spiked.
The fix wasn’t switching to radial. It was redesigning the groove geometry to a hybrid 50-degree angled interface that combined axial and radial compression forces. The new gasket — 1.5 mm wide by 4.35 mm tall — achieved 95% worst-case gland fill and passed IPX7 submersion testing. Orientation wasn’t binary; the groove angle was the variable that mattered.
This case also destroys the oversimplification that radial seals are only for “low-pressure” applications. Standard radial groove dimensions per AS568 are rated at 1,500 psi — but with proper backup rings and groove engineering, radial seals have reliably sealed 200,000 psi using a 70 durometer compound. Groove design and material selection determine pressure capability, not the orientation label.
The real engineering tradeoff in the overlap zone is manufacturability versus sealing robustness. Radial grooves are easier to modify because groove dimensions control the seal — swap an O-ring size, re-cut the groove on a lathe, and you’ve adjusted performance. Axial grooves are harder to machine but inherently simpler to seal because there’s no extrusion gap to manage. When both work, pick based on your manufacturing capability and tolerance budget, not a textbook rule.
Selection Decision Framework
Stop asking “axial or radial?” and start asking these five questions:
1. Is There Relative Motion?
Reciprocating or rotating motion almost always demands radial orientation. The cylindrical contact geometry accommodates shaft movement without losing seal contact. Axial seals in dynamic service risk walking out of the groove.
2. What Is the System Pressure?
Below 1,500 psi static, both orientations work with standard groove dimensions. Above 1,500 psi, radial needs backup rings to prevent extrusion. Axial has no extrusion gap, so it handles higher static pressures without additional hardware — but only when faces remain parallel under load.
3. How Tight Are Your Tolerances?
Axial seals are sensitive to face parallelism and flatness. A few hundredths of a millimeter of tilt can create a leak path. Radial seals are more forgiving of bore-to-shaft misalignment because the circumferential squeeze compensates for minor eccentricity. If your manufacturing process can hold face flatness, axial works. If not, radial gives you margin.
4. What Machining Is Available?
Radial grooves are lathe-friendly — standard tooling, standard setups. Axial grooves require face machining, which adds cost and complexity. For a balanced or unbalanced seal assembly with multiple sealing points, the easier machining of radial grooves often tips the decision in production environments.
5. Is Contamination a Factor?
In applications exposed to dirt, slurry, or abrasive particles, axial face seals can be preferable. The flat mating surfaces are easier to protect with gland packing or mechanical seal arrangements, and particles are less likely to lodge in the sealing interface compared to the cylindrical gap of a radial configuration.
Work through these five parameters in order. By question three, the answer is usually clear. When it isn’t — when both orientations survive all five checks — the groove geometry and your shop’s machining capability make the final call.
The Bottom Line
The axial-versus-radial decision is not about memorizing which orientation goes where. It’s about matching groove geometry to operating parameters — pressure, motion, tolerances, machining, and contamination. In the overlap zone where both orientations work, I’ve seen groove design turn a leaking axial seal into a reliable hybrid, and I’ve seen radial grooves handle pressures a hundred times beyond their “standard” rating.
Default to radial when there’s relative motion and your tolerances are moderate. Default to axial for static, high-pressure, contamination-exposed faces. And when both seem viable, let your tolerance stack-up analysis and machining capability decide — not a textbook table.
