CUSTOM METAL PARTS MANUFACTURER IN CHINA · OEM DRAWING-BASED PROJECTS
Processes Technical reference guide 8 min read

Executive Summary

Surface roughness is the fine-scale texture left on a metal surface after machining, forming, grinding, blasting, or finishing. For an OEM part, it can influence sealing, sliding friction, lubricant retention, coating adhesion, fatigue behavior, appearance, and the reliability of dimensional inspection. It is therefore a functional manufacturing characteristic, not simply a visual preference.

The most familiar roughness value is **Ra**, the arithmetic average of the measured profile’s absolute deviations from a mean line. Ra is useful for setting a common language between buyer, designer, and supplier, but it does not describe every feature of a surface. Two parts can have the same Ra while differing substantially in peak height, valley depth, lay direction, waviness, or isolated defects. A robust specification identifies the parameter, unit, measurement location, evaluation conditions, and relationship to the manufacturing process.

For international sourcing, the practical objective is not to demand the smallest possible Ra. It is to select a surface condition that supports the part’s function and can be produced consistently at an appropriate cost. The best RFQ communicates where the finish matters, which areas are excluded, and how acceptance will be verified.

What Surface Roughness Actually Describes

A manufactured surface contains texture at several scales. **Roughness** refers to relatively fine, closely spaced irregularities. **Waviness** describes broader undulations caused by machine vibration, work deflection, tool runout, thermal effects, or fixturing. Form error is an even larger-scale departure from the intended geometry. These characteristics can overlap in a real measurement, so the instrument and filtering settings matter.

Ra is normally expressed in micrometres, or µm, in metric documentation. It averages the absolute profile deviations over a defined evaluation length. Because positive and negative deviations are treated as magnitudes, the result is stable and easy to compare. However, averaging can hide a sharp scratch, a torn edge, a deep isolated pit, or a directional pattern. A sealing face, bearing journal, or fatigue-sensitive fillet may require additional controls beyond an Ra limit.

Other parameters may be relevant when the application demands more detail. Rz describes a peak-to-valley characteristic over sampling lengths, although the precise calculation convention should be stated because standards and software settings can differ. Rt or Rmax may describe the total height of the measured profile within an evaluation length. Rq, the root-mean-square roughness, gives greater weight to larger deviations. Waviness parameters and lay direction can also be important for sliding or sealing interfaces.

| Surface consideration | Why it may matter to an OEM part | Typical specification question | |---|---|---| | Ra average roughness | Provides a general texture limit for machined or finished areas | What Ra maximum is required, and at which locations? | | Peaks and valleys | May affect leakage, wear, contact stress, or coating coverage | Is an Rz, Rt, or defect limit also necessary? | | Lay direction | Influences lubricant flow, sliding behavior, and sealing paths | Must tool marks run parallel, perpendicular, or in a defined direction? | | Waviness | Can prevent full contact even when Ra is acceptable | Is waviness controlled separately from roughness? | | Local defects | Scratches, burrs, pits, and chatter may cause functional failure | Are visual or tactile defect criteria defined? |

Choosing a Surface Requirement by Function

The design team should begin with the interface, not a preferred number. A surface that supports a static gasket, a rotating seal, a press fit, a painted enclosure, and a decorative panel may need very different treatment. The drawing should identify functional zones and avoid applying a demanding finish to every face when only one area needs it.

For a sliding or rotating interface, roughness interacts with material hardness, contact pressure, speed, lubrication, temperature, and alignment. A very rough surface can increase abrasive wear or disrupt a lubricant film. A very smooth surface is not automatically better: it can reduce lubricant retention in some arrangements, and achieving it may require grinding, honing, lapping, or additional polishing. The mating material and the complete tribological system must be considered together.

Sealing performance depends on more than Ra. A static metal-to-metal seal may require controlled flatness, low waviness, an appropriate lay, and freedom from scratches crossing the sealing path. A polymer seal may tolerate a different texture, but sharp peaks can damage the seal and deep valleys can become leakage channels. Ask the seal supplier and manufacturing engineer to define the surface requirement for the actual seal geometry rather than copying a generic “fine finish” note.

For coating, plating, painting, or adhesive bonding, texture affects mechanical keying, coverage, and the exposed area available for adhesion. Yet the pretreatment may change the final surface, and coating thickness can partially obscure the original profile. The specification should say whether roughness is measured before or after treatment, especially when the functional requirement concerns the finished assembly.

Fatigue-sensitive parts deserve special caution. Machining marks, burrs, tool tears, and grinding burns can act as stress concentrators. Average roughness alone cannot guarantee a suitable fatigue condition. Fillet geometry, residual stress, material condition, heat treatment, edge quality, and direction of tool marks may be equally important. If fatigue life is critical, define the relevant process and inspection controls with the design authority instead of relying on a single Ra value.

How Manufacturing Processes Create Texture

Every process leaves a characteristic surface signature. Turning commonly produces a helical lay whose spacing depends on feed, tool geometry, nose radius, cutting speed, material behavior, and tool condition. Milling can leave scallops, cusps, or periodic marks influenced by stepover, cutter diameter, tool runout, and machine dynamics. Drilling may create feed marks, burrs, smeared material, or a less uniform finish than the adjacent milled face.

Grinding can produce a lower average roughness than conventional cutting, but wheel selection, dressing, coolant, contact pressure, and thermal control are decisive. An apparently fine ground surface can still contain directional marks, burn, or microcracking if the process is poorly controlled. Honing and lapping can improve geometry and texture for selected bores or sealing surfaces, but they add operations and require appropriate access and fixturing.

Forming, casting, and additive processes generally create a different starting condition. Cast surfaces can include mould texture, embedded particles, and local variations. Forged surfaces may retain scale or die marks. Sheet-metal surfaces can carry mill finish, scratches, or forming marks. Additive parts often have layer-related texture and orientation effects. Machining only selected interfaces may be more economical than trying to make every external face uniformly smooth.

A supplier should choose process parameters capable of meeting the requirement with margin, but the buyer should not assume that a nominal machine capability applies to every geometry. Deep pockets, thin walls, interrupted cuts, long shafts, small internal radii, and hard materials can increase vibration or deflection. The accessible measurement area may also differ from the area that is functionally critical.

Measuring Ra Without Creating Ambiguity

The most common production instruments are contact stylus instruments and non-contact optical systems. A stylus traverses the surface with a small tip and converts vertical movement into a profile. Optical methods can capture a broader area and avoid contact, which is useful for delicate or difficult surfaces, but reflectivity, slope, transparency, and surface processing can affect the result. The method must suit the material and geometry.

Measurement settings are part of the specification. Cutoff length separates roughness from longer-wavelength waviness, while evaluation length typically includes one or more sampling lengths. A short trace can miss periodic variation; a long trace may combine different surface conditions. Stylus tip radius and force, filter type, traverse speed, instrument calibration, and measurement direction can all influence the reported value.

Measurement direction is particularly important on turned, milled, ground, or brushed surfaces. Reading across tool marks often produces a different value from reading along them. A drawing or inspection plan should state the direction relative to the lay when direction affects acceptance. If no direction is defined, buyer and supplier may obtain different results while both believe they followed the requirement.

The probe must sit on the actual functional surface, not on an adjacent fillet, chamfer, burr, or accessible substitute area. Curvature and limited contact length can distort the trace. A measurement plan should identify locations, number of readings, orientation, and how results are evaluated. For critical parts, agreement on a reference method before production prevents disputes after delivery.

Common Failure Modes and Trade-Offs

Specifying a number without a surface location

A general note such as “Ra 1.6 µm” may be interpreted as applying to every face, including hidden or nonfunctional surfaces. This can force unnecessary finishing and obscure the zones that truly matter. Mark functional areas directly on the drawing or model, and distinguish default surface condition from special requirements.

Treating Ra as a defect-free guarantee

An average does not prohibit every individual scratch or pit. A surface may pass Ra while containing a defect that crosses a seal or bearing track. Add visual defect criteria, edge requirements, local limits, or a different profile parameter where the application needs them.

Ignoring lay direction

A directional texture can create a leakage path or alter friction even when the average value is acceptable. Define lay direction when it affects fluid containment, sliding, optical appearance, or bonding.

Over-specifying smoothness

A lower Ra often means slower feed, sharper tooling, extra passes, finer abrasives, or a separate finishing operation. It may also increase handling and inspection sensitivity. If the function does not require the tighter finish, the added process can increase cost and lead time without improving performance.

Measuring after the wrong process stage

Oil, oxide, blasting media, plating, paint, deburring, and polishing can all change the measured profile. If the drawing limit applies to the bare machined substrate, inspect before coating. If it applies to the assembled or treated condition, state that explicitly and control the treatment process as well.

Comparing unlike instruments or settings

Two valid instruments can report different values when tip geometry, filters, cutoff, trace length, or direction differ. Supplier reports should include the method and settings, not only a number copied into a certificate.

RFQ and Pre-Production Checklist

Before requesting quotations, the buyer should provide a drawing or model that identifies critical surface zones, material, heat treatment, coating condition, datum structure, and the intended function of each special finish. A photograph or annotated view can help clarify areas that are difficult to label, but it should supplement—not replace—the controlled drawing.

Ask the supplier to confirm the proposed process route and the stage at which roughness will be measured. For example, a turned shaft may need a finishing pass, while a sealing face may require milling followed by grinding. The supplier should identify features that may be difficult to measure and propose a controlled method before tools or fixtures are finalized.

A practical pre-production checklist includes:

  • Specify **Ra, Rz, or another parameter**, including units and maximum or target limits.
  • Identify the exact faces, bores, journals, sealing paths, and edge zones covered by each requirement.
  • State the required lay direction or whether direction is unrestricted.
  • Define measurement stage: as-machined, after heat treatment, after coating, or at final inspection.
  • Agree on cutoff, evaluation length, instrument type or equivalent method, trace direction, and calibration practice.
  • Define the number and distribution of readings for critical surfaces.
  • Separate roughness from flatness, roundness, profile, burr, edge-break, and visual-defect requirements.
  • Review manufacturability for thin walls, deep features, interrupted cuts, and limited probe access.
  • Request a first-article or pre-production measurement plan rather than accepting an unexplained single reading.
  • Establish how nonconforming texture will be contained, reworked, or dispositioned.

The RFQ should also ask whether the quoted process includes deburring, polishing, blasting, cleaning, and protective packaging. A surface can be compliant at the machine and then damaged by bulk handling or by abrasive contact during shipment. Packaging is part of surface preservation when the finish is functional or cosmetic.

A Practical Decision Sequence for OEM Teams

First, classify each surface by function: sealing, sliding, fatigue-sensitive, coated, cosmetic, dimensional fit, or noncritical. Second, determine the minimum surface condition that supports that function using material, mating component, environment, and service loads. Third, choose a plausible process route and check whether the geometry permits stable production and inspection.

Fourth, write the requirement so that two independent inspectors can measure it in substantially the same way. Fifth, validate the requirement on representative first-article parts, paying attention to local variation rather than only the best-looking area. Finally, monitor the process characteristic that drives texture, such as tool wear, wheel condition, feed, vibration, or treatment media condition.

This sequence keeps roughness connected to design intent. It also gives the sourcing team a better basis for comparing quotations. A supplier proposing turning, grinding, or polishing may be responding to different interpretations of the drawing; process clarification makes the commercial comparison fairer.

Conclusion

For OEM metal parts, surface roughness should be specified as a controlled functional characteristic, not as an isolated appearance label. Ra is a valuable common measure, but it should be supported by surface location, lay direction, measurement conditions, defect criteria, and the correct process stage. Buyers who connect the requirement to sealing, wear, coating, fatigue, or fit can avoid both under-specification and unnecessary finishing cost.

The strongest specification is one that a designer can justify, a manufacturer can produce repeatedly, and an inspector can verify without guessing. That shared clarity is more valuable than selecting an impressively low Ra value that the application never needed.

References

[1]: https://www.iso.org/standard/680 ಅ "ISO 21920-1:2021, Geometrical product specifications (GPS)—Surface texture: Profile method—Part 1: Indication of surface texture"

[2]: https://www.asme.org/codes-standards/find-codes-standards/asme-y14-36-surface-texture-symbols "ASME Y14.36, Surface Texture Symbols"

[3]: https://www.nist.gov/publications/surface-texture-measurement "National Institute of Standards and Technology, Surface Texture Measurement"

Use this guide in a drawing-led RFQ.

Share the latest drawing or model, material, quantity, required finish, delivery target, and open technical questions. This lets the manufacturing discussion start from your actual component rather than a generic article.

Request a technical review