Machining‍‌‍‍‌ marks, roughness direction and local defects that are visible when a seal leaks although it is correctly made of material and dimensions.

A pressure-decay test at a certain point is stuck at the same low leak rate for three samples. The O-rings are of the proper size, the chemical certificate corresponds to the order, and the groove dimensions conform to specifications. The disassembly did not show that the seal got cut. It was just under turned flange with angled light the narrow machining line across the contact band showed itself. The problem is not only the rubber. It is the interface the rubber was intended to seal.

Such a distinction is necessary because, as drawn, one can look at the metal face as a passive boundary whereas the seal meets it as a working terrain.

Peaks in the texture may cause stress concentrations, valleys may open a path connecting the high- and low-pressure sides, and a machining direction in a way may produce an endless opening right through the sealing band. So a rubber seal surface finish specification should not just copy one well-known roughness height from an old ‍‌‍‍‌drawing.

The Trace on the Flange

The first useful question after a leak is not, “Is the surface smooth?” It is, “What path could the fluid actually follow?” A witness band, transfer film, developer, or carefully controlled dye check may show where the seal contacted the hardware and where that contact weakened. The position of the leak should then be preserved relative to the groove, fasteners, ports, and machining direction before anyone polishes the evidence away.

A single numerical result can miss the reason for failure. Two surfaces may report similar Ra values while one has short, random texture and the other has a helical feed mark that crosses the seal line. A stylus trace taken in only one direction may also overlook a scratch running parallel to that trace. This is why machining marks and seal leakage should be investigated as a geometric relationship, not as unrelated inspection items.

Three Surfaces Can Hide Inside One Number

Texture amplitude

Ra is an arithmetic average of profile deviations. It is useful for process control, but an average does not fully describe the highest peak, deepest valley, peak spacing, or one isolated defect. Parker's O-Ring Handbook discusses Ra, Rz and other profile concepts and notes that the mating face and groove bottom influence seal life. It also separates static from dynamic surface recommendations. [1]

Waviness, form and isolated damage

Longer-spaced waviness, dish, taper, parting-line mismatch, porosity, coating nodules and handling dents may sit outside what one short roughness trace describes. The seal must bridge the entire local gap after compression, pressure, temperature and fastener distortion are applied. Surface roughness for seals is therefore only one layer of interface control; flatness, concentricity, lead, edge condition and cleanliness may be equally decisive.

In short, surface roughness for seals must be interpreted together with machining marks and seal leakage evidence from the actual contact band.

A‍‌‍‍‌ Public Case That Began With Reinspection

Parker published a case concerning a container that housed inertial measurement unit (IMU) electronics, commonly used in airplane and spacecraft navigation. One requirement for shipment of the enclosure was a helium leak test that the container had to successfully pass. Once the O-ring arrangement of the previous vendor did not achieve the required performance level, the customer re-examined the mating components for finishing correctness and checked the O-ring installation. The operation resulted in downtime and delayed shipments. [2]

Pressure washer cleaning at the field level caused the seal to break down as well as damage to the electronics.

The vendor replaced the configuration with a one-piece bonded enclosure seal. Parker says that with the new system, a machined groove for a mating surface is no longer needed, leak points are fewer, manufacturing is somewhat less strict, and reclamation efforts are less.

It is a valuable case to use publicly since it links interface design with a real finished-unit leak test. It is not a case of customer satisfaction with product Yida and the published page does not offer sample size details, leak-rate figures, working life, or a dataset comparing before and after in a controlled experiment.

Engineering takeaways are narrower than "bonded seal is always superior." The example shows that an interface is quite vulnerable to the assembly and machining variation if inspection of O-ring groove surface finish is repeated. Using another seal design could make it less susceptible, however, it still requires verification of pressure-media, temperature, opening cycles, and production tolerances,

Static and Dynamic Seals See the Same Surface in Different Ways

An impermanent seal (usually static in nature) needs to maintain the pressure continuously throughout the entire interface between the parts that are being joined. It shall handle hardware tolerance variation and rubber relaxation happening gradually over time. A continuous scratch across the contact area could be more harmful than a generally smooth surface finish. Also, after the drawing of parts is inspected and the parts are assembled, bolt spacing and flange stiffness might still affect local compression differently. A great surface finis can never be the reason to ignore the presence of a warp, an overfilled grove, or too shallow compression in a flange for example.

A seal for a reciprocating or rotating shaft constantly passes over its mating surface. Sharp peaks can wear down the rubber and scratch a coating, the lubricant film can be disrupted. The surface finish with a directional lead could be pumping some of the fluid. On the contrary, ultra-highly polished surface may not be the best for every moving mechanical system because what is good at lubrication retention and the break-in behavior is determined by seal type, speed, load, and medium. Thus, setting limits on dynamic seals should not follow one general surface roughness requirement.

The Standards Set Limits and Also Leave Some Things Unfixed

ISO 21920-1:2021 specifies a set of indications for the profile surface texture in technical product documentation whereas ISO 21920-2:2021 introduces a vocabulary of parameters and definition to determine the profile texture. [3][4] Such a set of documents allows engineers to describe requirements precisely and to read them without a problem. Still, the standards are not meant to recommend a safe working value that will ensure seal performance against a particular liquid, a set of pressure cycles, and a manufacturing procedure.

Current ISO 3601-2:2025 specifies general industrial use requirements for O-ring housing dimensions and the housing is defined as the groove or cavity and the mating surface in combination. It adds that a user and a manufacturer are expected to reach a mutual agreement on the design for special applications. [5] Having mentioned that, we see the point: it is one thing to comply with the housing standard as a baseline design, but another one to be sure the real working conditions will not result in any ‍‌‍‍‌leaks.

Write‍‌‍‍‌ the Drawing So Two Inspectors Reach the Same Result

It is a mistake to copy the surface finish of the O-ring groove just because the same finish was used in another component. Note down if the interface is static, reciprocating, oscillating or rotary; pressure direction; whether it is fluid or gas; vacuum requirement; temperature range; speed; if lubrication is present; and hardware and coating details; groove tolerances; anticipated squeeze; as well as the assembly route of the seal and hardware suppliers will be capable of identifying the same service problem from this data.

• Highlight the area under control on the schematic, including the full sealing path and groove transitions.

• State explicitly the surface feature parameter, the limit, measurement direction and reference standard rather than just writing "smooth".

• Take control of lay, lead, waviness, burrs, scratches, porosity and coating defects at locations where such factors may result in a leak or wear path.

• Give clear directions for cleaning, protection, lubrication and installation controls to ensure an undamaged machined face reaches the assembly point.

• Connect the surface demand to a completed-assembly leak or functional test using a designated media, pressure, temperature and tolerance limit.

Recreate the Leak Before Changing the Finish

If possible, make use of the returned or faulty assembly as a base for testing.

Observe, for example, the contact band; locate the leak direction by making markings; look out for rubber deformations in one area such as flattened area, cuts, abrasion and extrusion. Furthermore compare the rubber seal with the relevant hardware. Carry out measurements along and transverse to the path under suspicion at several positions. Should the process leave a directional texture, check if the direction varies from part to part or with different tool conditions.

Testing methods should simulate the conditions under which a failure could occur. Examples of tests could be detecting air pressure decrease, vacuum seal testing at ultra high vacuum, thermal cycling test by media soak, media pressure pulsation test, endurance testing at reciprocal movement or rotary motion, etc. As an additional requirement, one can include disassembly inspection. Both leakage and structural changes should be measured for the final part.

Finished-parts testing is the one point where a drawing hypothesis turns out to be an industrial rubber seals performance evidence.

The roughness measurement can support traceability and a sealing functionality verification has to be a representative assembly test.

A Practical Release Decision

An excessive surface finish requirement can lead to higher machining time and cost without eliminating the actual sources of leakage. On the other hand, overly generous limits can simply move the problem to inspection and warranty. The most cost-efficient requirement would be the biggest process window that does not compromise sealing performance and backed by the measurement procedures and validation on a finished-assembly. This is where, instead of assumed performance, seal industrial rubber seal performance starts to be controllable.

How Yida Is Helping in Interface Validation

You need to provide details, besides the 2D drawing, as the example of a 3D interface, surface requirements, type of machining, vacuum or pressure target, media, temps, type of movement, expected life, and lubrication if possible. You need to provide also failed samples. This way, the discussion around rubber seals surface finishing will be precise to the level where tooling, inspection, and verification will be influenced instead of generating another generic roughness ‍‌‍‍‌suggestion.