A poor turning surface finish is not always a feed-rate problem. Regular helical marks, torn patches, isolated scratches and broad waves point toward different mechanisms. Before changing cutting conditions, establish whether the failed requirement concerns roughness, waviness, surface damage or appearance. Then compare the defect with the tool, material and operation that produced it.
That distinction matters when a shaft measures correctly but damages a seal, or a turned face looks bright yet fails a roughness check. A useful investigation connects three things: where the mark occurs, how the surface must function, and what changed during manufacture. It does not begin with a blanket instruction to slow the machine down.
Four surfaces that need four different investigations
Turning leaves a directional texture as the cutting edge advances across a rotating workpiece. Some visible texture is therefore expected; visibility alone does not establish rejection. A photograph under grazing light can exaggerate a harmless pattern, while a small scratch on a sealing track may matter more than the rest of the surface.
| Observed pattern | First hypothesis to test | Useful evidence | Misleading shortcut |
|---|---|---|---|
| Regular, evenly spaced feed marks | Feed and effective cutting-edge profile | Feed per revolution, insert identification and measured profile | Rejecting every visible line |
| Irregular torn or smeared patches | Adhesion, built-up edge or unsuitable cutting action | Edge condition, actual material and changes along the cut | Assuming the surface only needs polishing |
| Isolated grooves crossing the normal lay | Chip contact, handling damage or a contaminated later operation | Inspection immediately after cutting and after handling | Changing the entire finishing program |
| Repeated waves or bands | Vibration, changing support or machine-related movement | Location, sound, engagement and workholding records | Treating all periodic marks as normal feed texture |
These are starting hypotheses, not diagnoses from appearance alone. Inspect a clean surface and preserve an unmodified sample. If parts have already been tumbled or hand-polished, the original evidence may be gone.
The feed-radius calculation has a boundary
For an ideal conventional round-nose tool in stable turning, a common geometric estimate is Ra ≈ f²/(32r), where feed per revolution f and nose radius r use the same length unit. The result uses that unit as well. It estimates the texture produced by geometry; it does not predict adhesion, tool damage, vibration or material tearing.
For an illustrative calculation, f = 0.10 mm/rev and r = 0.8 mm give Ra ≈ 0.000391 mm, or about 0.39 µm. Doubling feed to 0.20 mm/rev increases that estimate to about 1.56 µm. These are calculated examples, not measured results or Jucheng capability commitments. Wiper inserts and other edge geometries require their own application guidance.
A larger radius is not an unlimited route to a better finish. On a flexible workpiece, the associated force distribution can increase deflection or vibration. Reducing feed too far can also make the edge rub rather than cut effectively. Tool geometry, edge preparation, depth of cut and material condition have to work together. The Processus de tournage CNC provides the wider context for stock support and operation sequencing.
Use the location of the damage to narrow the cause
A good start followed by a poor finish
If the first portion of a cut is acceptable and the remainder deteriorates, check what evolves: edge wear, chip accumulation, temperature, unsupported length or proximity to a shoulder. Record whether the effect repeats on every part or appears after several parts. A repeating spatial pattern and a gradual batch trend imply different tests.
One damaged band on an otherwise sound diameter
Examine the corresponding operation transition. A dwell, tool entry, clamp contact or chip trapped at a shoulder can leave a localized mark. Compare the surface before and after secondary handling. A defect created in a washing basket will not be solved by another turning insert.
A face and diameter that behave differently
The cutting conditions on a faced surface change with radius. Constant-surface-speed programming is also subject to a spindle-speed limit. The center region therefore does not experience exactly the same conditions as the outer edge. Inspect the region that actually performs the function, not whichever area is easiest to reach with a probe.
Accept the functional surface, not the shine
Ra is an average profile-height parameter. Two surfaces with similar Ra can differ in isolated valleys, peak shape and lay. A seal contact, sliding interface and visible decorative face may therefore need different acceptance information even when the drawing shows the same numerical roughness limit.
Consider an illustrative turned sleeve: the outside diameter locates a bearing, one shoulder carries an axial load, and another surface is merely visible. Applying the tightest finish everywhere can add work without protecting the critical interface. Instead, identify each surface, its mating part and the failure that the finish requirement is intended to prevent.
Do not convert this into a generic table promising one Ra value for every seal or bearing. Requirements come from the actual component design and the relevant mating-component specification. Specify additional texture or defect limits when function needs them. If the question is about anodizing, plating or another added treatment, use the separate surface-finishing options guide; treatment selection and as-turned texture are not the same decision.
A finish dispute can start at the measurement setup
Before comparing two inspection reports, compare the methods. Identify the parameter, units, evaluation settings, trace direction, sampling location and instrument suitability. On a cylindrical surface, a trace parallel to the lay can produce a different result from one across it. Curvature, short accessible lengths and probe access can also affect whether a measurement represents the specified area.
Clean the part without altering its texture. Separate form from texture using the agreed evaluation method rather than accepting whatever default filter an instrument happens to use. A handheld roughness comparator is useful for screening, but it should not automatically replace the specified quantitative measurement.
Keep a map of the measured positions, especially when damage is localized. A report containing only a single average number may conceal the fact that the functional track was never measured. Jucheng’s quality-control overview is a starting point for discussing inspection scope; the measurement method and acceptance plan still need agreement for the particular part.
Run a small trial that can distinguish cause from coincidence
- Preserve the baseline: sample, surface photos, tool identity, material batch and measurement settings.
- Choose one mechanism suggested by the marks. Check the cutting edge before changing numerical settings.
- Make one controlled change within the tool and machine’s safe operating limits.
- Measure the same locations using the same method, and check dimensions as well as texture.
- Repeat enough parts to determine whether improvement survives tool use and normal production variation.
A surface that improves while its diameter drifts is not a successful fix. Nor is a polished first sample evidence of a stable production process. Record which conditions must remain controlled and what triggers an insert change or additional inspection.
For a supplier discussion, send the drawing revision, identified failed surface, measurement report and an unaltered sample or clear images. Request the cause-and-verification record rather than simply asking for a “smoother” replacement. That gives both parties an observable criterion for closing the issue.
Questions that arise after the first inspection
Can the part pass Ra and still be rejected for a scratch?
Yes, if a separate defect or functional-surface requirement applies. A localized scratch may not be represented by the particular roughness trace. The drawing and inspection plan should make the acceptance basis clear.
Should every turned surface be polished?
No. Polishing is an additional process that can change dimensions, edges and texture. Use it when the surface function requires it and allowance, method and final inspection are defined.
Does a quiet cut prove that chatter is absent?
No. Sound is supporting evidence, not a measurement. Check the surface profile and geometry, and compare where the marks occur with support and engagement conditions.