Surface Finishing

32 Ra surface finish requirements for machined parts

What a 32 Ra surface finish means

A 32 Ra surface finish usually refers to a roughness average of 32 microinches Ra on U.S. inch-based machining drawings. In metric terms, 32 µin Ra equals 0.8128 µm Ra, commonly rounded to 0.8 µm Ra. It is a fine machined surface, but it is not a mirror finish, and the number alone is not a complete surface-texture requirement.

The callout should be read together with the roughness parameter, unit, drawing standard, measurement conditions, lay direction, and the surface’s functional role. In shop terms, 32 µin Ra is often achievable by controlled turning, milling, boring, or reaming. It becomes more difficult on unstable setups, long bores, soft gummy materials, interrupted cuts, or surfaces that also have tight geometry limits.

shadow, sc32, honda, cruising, chrome

Ra is the arithmetic average roughness of a measured profile. It represents the average absolute height deviation from a mean line over a defined evaluation length. That makes Ra useful for general machining communication, but it also has limits. Two surfaces can share the same Ra value and perform differently if one has sharp peaks, deep scratches, directional lay, chatter marks, or waviness. Standards such as ASME B46.1 and the ISO 21920 series treat surface texture as more than a single number for this reason.

For more background on finishing methods and surface-texture terminology, see MechMeld’s surface finishing section.

Why the units and parameter matter

The most common mistake is reading “32” without confirming what the number means. In many U.S. machine shops, a bare “32 finish” is understood as 32 microinches Ra. That convention may work inside one supply chain, but it is risky when drawings move between countries, customers, vendors, or inspection departments.

There is a major difference between 32 µin Ra and 32 µm Ra. A value of 32 µm Ra is roughly 1,260 µin Ra, which is dramatically rougher than a 32 microinch precision-machined finish. Ra is also not the same as RMS roughness, Rz, Rt, or Sa. RMS is a root-mean-square profile value, Rz describes peak-to-valley behavior in a different way, Rt is a total profile height parameter, and Sa is an areal roughness parameter measured over a surface area rather than a two-dimensional profile.

Callout wording Likely meaning Risk if unclear
32 µin Ra Roughness average of 32 microinches Usually clear on inch-unit drawings
0.8 µm Ra Metric equivalent, rounded from 0.8128 µm Should still state whether it is a maximum or target
32 Ra Often interpreted as 32 µin Ra in U.S. machining Ambiguous without units
32 RMS Root mean square roughness, not Ra Cannot be treated as identical to Ra without permission
N6 Common roughness grade comparison near 0.8 µm Ra Not a substitute for a complete drawing requirement

A clear drawing states the parameter and unit directly, for example “Ra 32 µin max” or “Ra 0.8 µm max.” If the surface has a functional role, the note should also define any required lay, measurement direction, cutoff, filtering, sampling length, or reference standard.

Where 32 Ra fits among common machined finishes

A 32 µin Ra finish is in the fine-machined range. It is smoother than many general-purpose machined surfaces, but rougher than finishes typically associated with precision grinding, lapping, polishing, or some sealing and bearing surfaces. Common applications include machined faces, shafts, bores, mounting surfaces, tooling details, and parts where visible tool marks must be controlled without requiring an unnecessarily expensive mirror-like finish.

Ra in microinches Approximate Ra in micrometers General manufacturing interpretation
125 µin 3.2 µm General machined or rougher functional surface
63 µin 1.6 µm Controlled machine finish for many noncritical surfaces
32 µin 0.8 µm Fine machined surface with controlled finishing pass
16 µin 0.4 µm Fine ground or carefully finished surface in many cases
8 µin 0.2 µm Very fine finish, often requiring special process control

These comparisons are useful for communication, not for process guarantees. The actual method depends on material, feature geometry, machine condition, tool selection, and inspection method. A short, rigid turned diameter in free-machining steel may meet 32 Ra with a normal finishing pass. A deep blind bore in stainless steel, a thin wall that vibrates, or a milled floor near a tight corner may require a slower strategy, sharper tooling, more stable workholding, or a secondary operation.

Can machining achieve a 32 Ra surface finish?

Machining can often achieve 32 µin Ra, but it should not be treated as automatic. The finish produced by a cutting process depends on both the theoretical feed marks and real process behavior. Feed per revolution or feed per tooth, tool nose radius, cutter runout, insert edge preparation, tool wear, chip evacuation, built-up edge, coolant delivery, spindle condition, and workholding rigidity can all affect the result.

In turning, a suitable feed rate and nose radius can create a smooth theoretical profile. The actual finish may still fail if the insert is worn, the material tears instead of shears, the part deflects, or chatter develops. In milling, step-over, tool runout, cutter geometry, toolholder stiffness, and finishing direction may matter as much as feed rate. A wide face mill can leave a clean surface on a rigid part, while a long-reach end mill may leave visible vibration marks even when programmed conservatively.

Boring and reaming add their own limits. A reamer may produce a good bore finish when alignment, stock allowance, lubrication, and chip control are correct. A boring bar, especially a long one, is sensitive to overhang and vibration. Internal features are also harder to inspect consistently because access, stylus orientation, and measurement location may be constrained.

  • More likely to meet 32 Ra by machining: rigid external diameters, accessible milled faces, stable bores, free-machining materials, and features with room for a finishing pass.
  • More likely to need extra control: slender parts, thin walls, deep bores, interrupted cuts, gummy stainless or aluminum conditions, heat-treated materials, and surfaces near sharp internal corners.
  • More likely to need another process: surfaces requiring much finer roughness, controlled peak height, special lay, tight waviness limits, or cosmetic consistency beyond ordinary machining.

Secondary finishing is not automatically better. Polishing, blasting, tumbling, grinding, honing, or lapping can improve or alter surface texture, but these processes can also round edges, change dimensions, smear material, embed media, or disturb flatness. The process should be selected for the surface function, not only for the roughness number.

What a complete 32 Ra callout should include

A surface finish note should tell manufacturing and inspection what must be controlled. “32 Ra” may be understood as an informal shop note, but a production drawing should remove avoidable interpretation.

  • Parameter: Specify Ra instead of leaving the number alone.
  • Unit: Use µin for inch drawings or µm for metric drawings, and avoid mixing units without explanation.
  • Limit type: State whether the value is a maximum, a range, or a nominal target.
  • Standard: Reference the applicable surface texture standard used by the drawing system, such as an ASME or ISO framework.
  • Measurement conditions: Define cutoff, filtering, sampling length, or evaluation length when they affect acceptance.
  • Lay direction: Specify lay if sliding, sealing, fluid flow, or visual appearance depends on direction.
  • Location: Apply the requirement only to the surfaces that need it.

A clearer callout might say “Ra 32 µin max, lay perpendicular to motion,” or “Ra 0.8 µm max on sealing land only,” depending on the design need. The exact notation depends on the drawing standard and company practice, but the principle is the same: the surface requirement should be measurable, limited to the relevant feature, and tied to the function of that feature.

Blanket notes such as “32 Ra all over” can add cost and confusion. They may imply unnecessary finishing on hidden clearance faces, saw-cut stock, cast surfaces, or nonfunctional areas. A better drawing separates general surfaces from functional interfaces. For example, a noncritical pocket floor might allow a rougher finish, while a bearing seat, seal land, or locating face receives the 32 Ra requirement. See also: CNC Machining.

How a 32 Ra finish should be inspected

A visual comparison block can help a machinist estimate surface finish, but it is not the same as a measured result. For acceptance, a calibrated roughness instrument is usually more appropriate. Contact stylus profilometers remain common for profile roughness measurement, while optical systems may be used for some surfaces and industries. The selected method should match the drawing requirement and the surface geometry.

Inspection should also consider direction. Roughness is often measured perpendicular to the lay because that direction crosses the tool marks and tends to capture the profile variation more clearly. If the drawing specifies another direction or a special functional direction, that instruction should control. The surface should be clean and free from burrs, loose particles, and handling damage before measurement.

One measurement may not represent the entire surface. A milled face can have different roughness near entry, exit, tool overlap, or chatter-prone regions. A turned shaft may vary near shoulders or interrupted sections. A bore may differ near the mouth and deeper inside. For critical parts, the inspection plan should define the number of measurements, locations, acceptance rule, and treatment of isolated defects such as scratches.

Measurement uncertainty matters more as limits become tighter. National metrology organizations such as NIST publish calibration and uncertainty information for roughness parameters, reinforcing a practical point: a roughness number is a measured quantity, not a visual impression. If a part measures close to the limit, the drawing, quality plan, and customer agreement should determine how uncertainty is handled.

When 32 Ra is not enough by itself

Ra is popular because it is simple, but that simplicity can hide functional problems. A seal may leak because of directional scratches even when the average roughness is acceptable. A sliding surface may need controlled valleys for lubricant retention. A fatigue-sensitive part may be affected by sharp peaks, machining marks, or residual stress. A cosmetic surface may be rejected because the pattern looks inconsistent, even though roughness readings are similar.

For these cases, designers may need additional controls. Rz or Rt can help describe peak-to-valley behavior. Waviness requirements can address longer-spaced surface variation that Ra filtering may separate from roughness. Flatness, cylindricity, or profile tolerance may be more important than roughness on sealing or locating features. Lay symbols can control directionality. Process restrictions may be justified if the function depends on grinding, honing, lapping, burnishing, or another specific finish.

The goal is not to make every surface smoother. The goal is to specify the surface texture that supports the part’s function. A 32 Ra requirement can be appropriate, economical, and straightforward to verify when it is applied to the right feature with clear units and measurement rules. It can also be wasteful or incomplete when used as a default note without understanding what the surface must do.

Frequently asked questions

Is 32 Ra the same as 0.8 Ra?

It depends on the unit. 32 µin Ra equals 0.8128 µm Ra, which is usually rounded to 0.8 µm Ra. The drawing should state the unit because 32 µm Ra would mean a much rougher surface.

Is 32 Ra a smooth finish?

Yes. It is generally considered a smooth or fine machined finish, but not a mirror finish. It is smoother than 63 or 125 µin Ra and rougher than 16 or 8 µin Ra.

Can a mill hold a 32 Ra surface finish?

A mill can often produce 32 µin Ra on a rigid, accessible feature with the right cutter, toolpath, step-over, finishing pass, and machine condition. Chatter, tool runout, recutting chips, worn inserts, and unstable workholding can prevent the surface from meeting the requirement.

Should 32 Ra be specified on every surface?

Usually no. It is better to apply 32 Ra only where the function requires it. Noncritical faces may work with a rougher finish, while sealing, locating, bearing, or visible surfaces may need tighter and more complete requirements.

Does a better Ra value always improve part performance?

No. A smoother surface can increase cost, reduce lubricant retention, affect coating adhesion, or change how two surfaces interact. The best surface finish is the one that matches the functional requirement, not automatically the lowest Ra value.