Technical Deep-Dive 08

"A Previous Supplier Failed My Part — Can You Review It, Salvage What's Possible, and Tell Me Why It Went Wrong?"
Independent Process Review & Non-Conformance Recovery

When a 5-axis job goes wrong at another shop, the buyer is left with scrap risk, a slipped schedule and — most dangerously — no explanation. We offer an independent process review: we examine the failed parts and the supplier's data, separate real defects from measurement disagreement, recover what is salvageable, and give you a root-cause answer you can act on. This page explains how that review works and what you receive.

Reviewed by: Goldcattle Engineering & Quality Team Last updated: August 2026
Home / Services / CNC Machining / 5-Axis CNC Machining / Process Review & Failure Recovery

The Short Answer

Datum & Fixture Strategy

Yes. A failed 5-axis job is rarely unsalvageable, and almost never unexplained — if you review it methodically. Our process review follows three steps:

  1. Triage: separate real non-conformance from a measurement-system disagreement between two CMM reports. On 5-axis parts this is the single most common false alarm.
  2. Recover: disposition each part — use-as-is (with engineering concession), rework, repair, or scrap — based on safety, contract and cost, not default scrap.
  3. Explain: classify the root cause (drawing / process / capability) and deliver a structured report with corrective-action recommendations.

Roughly 60% of rejected PPAP submissions fail on PFMEA, control plan, MSA or capability — not on the part itself. The same pattern holds for incoming rejects: the report often reveals a documentation or measurement-method gap before it reveals a bad part. We check the method before we blame the metal.

Why 5-Axis Failures Need a Different Review

A 5-axis part is machined in a single setup, so its entire feature relationship is defined by the machine's rotary-axis kinematic chain — not by sequential re-fixturing. That is the manufacturing advantage, and it is also the inspection trap.

The datum-reference-frame trap: the most common incoming-inspection error on 5-axis parts is measuring features against a datum established by surface contact, when the drawing's DRF (Datum Reference Frame) calls for a different precedence — e.g. resting the part on its largest flat face when the drawing defines datum A as a bore axis. If the supplier's outgoing CMM and your incoming CMM used different datum-establishment strategies, the reports are not comparable. The first corrective action is alignment on measurement methodology — not a process change. We request the supplier's CMM datum sequence before raising any NCR.
Deviation pattern on the report What it usually means First action
All affected parts shifted the same direction Datum offset in machine setup or CMM program zero-point Verify datum sequence; compare zero-point definition
Non-conformances scattered across the batch Thermal drift or tool wear mid-batch Check batch sequence vs measurement time
One specific feature wrong on every part Rotary-axis calibration error at a specific attitude Inspect rotary-axis calibration record
Scatter around nominal, some out Fixturing repeatability / inconsistent datum establishment Audit fixture and clamping method
First N parts OK, later parts out Tool wear or thermal growth partway through run Review tool-life and warm-up discipline

The Non-Conformance Workflow We Follow

Whether the failing parts come from another supplier or from our own floor, we handle them through a standard NCR (Non-Conformance Report) workflow. An NCR records what failed; CAPA fixes why it failed. Both must exist together.

1

Detect & Segregate

Quarantine suspect items physically or in the system so they cannot ship or be consumed. No part leaves the red bin until dispositioned.

2

Document

Raise the NCR with part/lot/serial, the exact requirement violated, and measured values — not just "out of spec". Evidence over adjectives.

3

Contain

Check whether the same problem affects other lots, machines or already-shipped product. Scope the blast radius before deciding disposition.

4

Disposition

Decide use-as-is, rework, repair, scrap or return-to-supplier — with the required approvals. Balance safety, contract and cost; default scrap is rarely the right first move.

5

Analyze & Close

For significant or repeat issues, run root-cause analysis and open corrective action. Verify the disposition executed, update records, feed data into trend review.

What Can Be Recovered — and What Cannot

"Failed" does not mean "scrap". Here is how we classify recovery options for parts from a previous supplier.

Disposition Meaning When it applies
Use-as-is Deviation formally accepted, with customer / engineering approval Non-critical feature; function unaffected; documented concession
Rework Brings the item back into full conformance e.g. undersized bore re-machined to size; cost below scrap + re-make
Repair Functional but not fully conforming; needs explicit authorization e.g. weld build-up + remachine on non-critical area
Scrap Removed from the value stream Structural/safety feature breached; recovery not feasible
Return to supplier Cost and correction pushed upstream Supplier responsible; under their warranty / NCR process
Honesty on salvage limits: we will not promise to recover a part we cannot recover to conformance or safe function. If a safety-critical feature is breached, the only correct disposition is scrap — and we will say so even if it costs you the part. A recovery promise that hides a structural risk is worse than the original failure.

Root Cause — Classified, Not Guessed

"We'll fix it" is not a root-cause analysis. We classify the failure into one of three buckets, because the corrective action is completely different for each.

Drawing / Specification

The requirement itself was ambiguous, inconsistent, or impossible as toleranced. Corrective action: clarify the drawing, adjust GD&T, re-baseline the DRF — not change the process.

Process

The process was capable in principle but executed wrong: tool wear, wrong offset, thermal drift, missed step. Corrective action: tool-life control, setup verification, SPC.

Capability

The process simply cannot hold the requirement as specified on the available equipment. Corrective action: process change (more axes, better machine) or tolerance/design revision.

For significant failures we use structured methods — 5-Why, fishbone (Ishikawa), or full 8D — rather than opinion:

D1 Form a cross-functional team (design, process, quality)
D2 Describe the problem: quantity, batch, defect type
D3 Interim containment: isolate, stop flow to next stage
D4 Root cause via 5-Why / fishbone
D5 Permanent corrective action on the root cause
D6 Verify effectiveness (e.g. pass rate ≥ 99.5% over 3 batches)
D7 Prevent recurrence: fold fix into standard work
D8 Close, archive, update work instructions

What You Receive From a Process Review

A review is only useful if it ends in something you can act on. Our standard deliverable package:

Findings

  • Triage verdict: real defect vs measurement disagreement
  • Per-part disposition with reasoning
  • Deviation-pattern map and likely root-cause class
  • Salvage feasibility per part (rework/repair/scrap)

Recommendations

  • Root-cause classification (drawing/process/capability)
  • Corrective-action proposal with verification plan
  • Re-qualification scope if you re-source to us
  • DFM feedback to avoid the same failure next time

Frequently Asked Questions

My supplier's report says all parts are in spec, but our incoming CMM disagrees. Who is right?
Possibly neither is "wrong" — you may be measuring different things. On 5-axis parts the most common cause is a datum-reference-frame mismatch: the two CMM programs establish the datum in a different order or from different features than the drawing's DRF precedence. Before raising an NCR, request the supplier's CMM datum sequence and compare it to your drawing's DRF. If they differ, the first fix is methodological alignment, not a process change. We do this comparison as step one of every review.
Can you recover parts another shop scrapped?
Sometimes — that is exactly what disposition analysis determines. Undersized bores can often be re-machined; minor out-of-position features may be recovered by rework; non-critical deviations may be accepted via engineering concession. But if a safety-critical or structural feature is breached, scrap is the only correct answer, and we will tell you that plainly. We evaluate each part individually rather than condemning a whole batch by default.
What if the root cause is the drawing, not the supplier?
That happens, and it is the most useful outcome — because fixing the part without fixing the drawing just moves the failure to the next batch. We classify drawing/specification root causes explicitly and give you DFM feedback: tighten or relax a tolerance, fix a GD&T datum scheme, add a local datum pad. You then issue an ECN and we re-qualify only the changed characteristics. Honest root-cause classification protects your program more than defensive supplier-blaming.
Will you take over the part after the review?
If you choose to re-source to us, we treat the reviewed part as a new program: we re-run our own FAI on production-representative parts, build the process flow / PFMEA / control plan, and confirm capability before volume. The review findings feed directly into that ramp so the same failure mode is designed out from the start. Re-qualification scope is defined in the review report so there are no surprises.

Have a Failed 5-Axis Job? Let Us Review It.

Send the failed parts (or their CMM reports) and the drawing. We will triage real defects from measurement disagreement, recover what is salvageable, and give you a root-cause answer you can act on — usually within a few business days.

Drawing + CMM reports + failed parts accepted. Confidential. No obligation to re-source.

Related Topics

Inspection & Verification → Capacity & Volume Production → Datum & Fixture Strategy → Back to 5-Axis Overview →

Recommended Reading