Case: Turning Wear Diagnosis

A lathe insert fails prematurely. Following the method from page 27, we read the edge, name the wear type, isolate the cause, change one variable, and verify — instead of just swapping inserts.

Concept

This case walks the page-27 loop on the locked CTT-T1 part: OD turning of a Ø50 raw bar down to Ø40 over 60 mm in 42CrMo4 QT at a classroom 240 HBW. We read the worn insert and the measured size together, separate gradual wear from sudden failure, and decide when to adjust — without releasing a part that is over the limit.

Why It Matters

Premature insert failure in production costs money per stop. A structured diagnosis turns a recurring nuisance into a fixed parameter or grade, and the record feeds tool management (page 30).

How It Works

Input: CTT-T1, Ø50→Ø40×60, classroom fn = 0.20 mm/rev, n = 1000 rpm, so vf = n·fn = 200 mm/min. Roughing arithmetic: radial ap = 2 mm means the diameter drops 4 mm per pass (50→46→42), then a final radial 1 mm to Ø40 — material-removal arithmetic, not a full tool path. Measure: read VB on the insert at fixed cutting time (page 25). Compare size: measured diameters 40.006 / 40.014 / 40.028 mm against a drawing upper limit of 40.020. Decide: the third part is 0.008 mm over the upper limit; the mean (40.016) being inside does not release it (page 18). Before compensating, rule out temperature, seating and measurement (page 27).

Worked Example

The evidence points to gradual flank wear growing over the batch, not a sudden fracture: the diameter creeps up (the worn edge pushes the tool away) and the finish degrades together, while the VB land grows with cutting time. A crater on the rake or a chipped corner would branch to a different lever (grade/heat vs toughness/rigidity), as page 27 requires. Here the corrective direction is a planned wear limit and a small compensation — but only after the cheap setup checks.

Reading the record

The three measured diameters — 40.006, 40.014, 40.028 mm — tell a story the average hides. The drawing upper limit is 40.020. Parts one and two are inside; part three is 40.028, 0.008 mm over. A shop that reports only the mean (40.016) as “within spec” is shipping an oversized part, because it is the maximum, not the average, that the tolerance controls. The trend (0.006→0.014→0.028) accelerates, consistent with a wear land developing rather than a one-off setup error. Before compensating, rule out the cheap causes (page 27): insert seating, swarf in the pocket, offset and nose-radius compensation, bar rigidity, and thermal equilibrium of part and gauge.

What the follow-up tells you

The verify step is what turns a guess into knowledge. In CTT-T1, the three values are measurements of three consecutive parts, not three sections of one part. The out-of-tolerance third part must be held for disposition. Diameter drift alone does not establish wear as its cause: check seating, offsets, measurement conditions and thermal effects, and compare the diameter trend with recorded flank-land growth over cutting time. Writing down which outcome occurred is the point of the trial log, so the next person does not have to re-derive it.

Putting numbers on the drift

To make the wear problem concrete, put the measured trend on the limit. The upper limit is 40.020 mm. Part three at 40.028 is over by 0.008 mm, and the gap from part two (40.014) to part three (40.028) is 0.014 mm of growth between parts — faster than the 0.008 mm step from part one to two. An accelerating trend is the signature of a wear land reaching its useful limit rather than a stable offset. The classroom feed and speed (fn = 0.20, n = 1000, vf = 200 mm/min) are held fixed while we observe, so the size growth is not caused by changing parameters; that makes the edge a credible candidate, but fixed parameters do not by themselves rule out thermal drift, measurement or seating. This is why the case measures VB at fixed cutting time (page 25): the curve of VB against time tells you when to stop and index, before part three repeats on the next batch.

Why you rule out setup before changing the grade

The reason the diagnosis loop has a “rule out” step is that many premature-failure complaints are not about the edge at all. A loosely seated insert, a built-up chip in the pocket, a wrong tool offset, or a holder that is not drawing the insert fully home all look from the outside like a tool that does not last: the finish goes off, the size wanders, the edge seems to wear fast. If you skip the setup check and jump straight to ordering a more expensive grade, you have spent money on a hypothesis that was never tested — and the new insert fails the same way. So the first action on a recurring failure is the cheapest one: confirm the insert seats cleanly, the pocket is free of swarf, the offset and nose-radius compensation are correct, and the bar is rigid. Only after those are known-good do you read the wear morphology and change a process variable. This is the same zero-cost-first principle as page 27: every setup check costs minutes, every new grade and every trial costs money and stops the machine. Diagnose the cheap causes first, then spend.

Why the corrected speed is recorded, not just remembered

The point of the wear-trial log is that a solved problem this week should not reappear as a mystery next month. When the measured limit — index before the diameter can drift past 40.020 — is written against the grade, geometry, material and these settings (fn = 0.20, n = 1000), the next batch starts indexed at the right point and never repeats part three being oversized. Without the record, knowledge lives in one operator's head and evaporates when the shift changes; with it, the shop accumulates "what actually worked here" instead of re-deriving it each time. This links to tool management on page 30: individual diagnoses become shop-wide starting points. The log is what turns one lucky fix into a repeatable process.

What each failure mode would have looked like instead

To make the diagnosis sharper, it helps to hold the alternatives side by side. If the edge instead showed a smooth, even flank land that grew steadily until finish failed, that would be normal wear whose rate may point toward a more wear-resistant grade or lower vc, but the rate alone is not proof of a speed error — it must be read against the recorded conditions. If it showed a small crescent crater on the rake face at high temperature, that is crater wear from the long steel chip sliding at speed, again pointing at speed and grade rather than setup. If it showed a chipped corner with a rough, interrupted or scale-covered entry, that is mechanical shock and wants a tougher substrate and a more rigid setup, not a speed change. A smooth, steadily growing flank land — as in this CTT-T1 case — points at planned wear and a settable limit, not at speed. Practising this matching — holding the morphology and asking which of these stories it tells before reaching for a tool — is what turns page 27's loop from a procedure into a reflex, and it is why the worn insert is photographed or kept rather than thrown away: it is the evidence that decides which lever to turn.

Why changing only one variable keeps the diagnosis honest

The loop says change one variable at a time, keep the others fixed, because a trial that moves speed and feed together cannot tell you which lever helped. Note the physical direction: with the rest of the geometry fixed, raising feed generally increases the undeformed chip thickness rather than thinning it, so feed is not a lever that "thins the chip". If two levers move at once and the finish improves a little, you never learn which one mattered; if it gets worse you do not know which to back off. The one-variable rule is what makes the verify step meaningful: the before-and-after difference is attributable to the single thing you changed. Keep diameter, feed and depth constant during the trial, change the one lever the morphology points to, and let the measured VB and finish decide the next move.

Why a stable edge geometry matters once the wear limit is set

Once the VB limit and the indexing point are established, the geometry choice quietly matters too. A built-up, blunt edge pushes and rubs, making the finish worse exactly as the diameter drifts, and a chip that does not break can rewrap around the part and scratch it. This is why the case pairs the wear-limit decision with a check of the edge and chipformer rather than only moving the offset: the corrected indexing point is paired with geometry that shears cleanly for this material (page 12). If after setting the limit the size holds but the finish still drags, the next lever is geometry and chipformer window, not a further offset. Substrate, coating and geometry are chosen together, and a wear limit is rarely the whole fix.

Why this case is deliberately not about a broken tool

The scenario — three consecutive parts whose diameter drifts gradually out of tolerance, with a smooth growing flank land on the edge — is chosen on purpose, because the most expensive recurring failures on a lathe are quiet rather than dramatic. A snapped insert stops the machine immediately and gets attention; a slowly wearing edge that produces drifting size and short, predictable life instead creeps along, and the cost shows up in rejected pieces and edge changes nobody planned. The structured loop exists to catch exactly this quiet case before it becomes a scrap issue. It also illustrates the diagnostic habit the course is building: reach for the morphology first, not for the alarm. A growing flank land together with a slow size trend is not a crisis, it is a readable sign, and treating it as information rather than annoyance is what lets you set the indexing point once instead of replacing edges all week. The dramatic fractures are covered on page 26; this case is the everyday, low-noise problem most shops actually live with.

Read the CTT-T1 evidence as a trend, not as a single number. The three diameters rise in order, and the measured flank land grows with cutting time; those two trends agreeing is what makes wear a credible explanation rather than a coincidence. A single out-of-tolerance part could come from seating or temperature, but three consecutive parts moving the same way points to a gradual edge process. The discipline the case builds is to compare the diameter trend with the recorded wear trend before touching the offset, so the correction follows the evidence instead of a guess on the first bad part.

A suspected out-of-tolerance result requires immediate verification and control of the affected part according to the inspection plan. Do not release it or continue making suspect parts merely to collect a trend. Use existing records and controlled measurements to investigate seating, offsets, temperature and wear. A confirmed wear trend may support a planned indexing limit or an approved dimensional correction; an offset does not restore edge condition or finish capability.

Common Mistakes

• Releasing part 3 because the mean is inside the limit. Each of the three consecutive parts is judged separately; the mean does not release an over-limit part. Fix: judge the worst of the three parts (page 18).
• Applying a wear offset before checking setup and thermal state. Fix: rule out the cheap causes first (page 27).
• Changing speed and feed at once. Fix: one variable, then verify.

The case holds fn and n fixed while observing, so any size growth is not from changing parameters. Measure VB at fixed cutting time and watch the trend: an accelerating growth from part one to three, after seating, offset and thermal checks are ruled out, supports scheduling an index before the next batch repeats an oversized part.

After the corrective offset, keep recording the next few parts: an offset brings size back toward the middle of the tolerance but does not reset the wear process. The VB-vs-time trend tells you when the next index is due, and that interval — not the clock — becomes the tool-life rule for this job.

Practice