PFMEA Failure Analysis

🧭 Why Step 4? #

You defined what each station must achieve in Step 3. Now ask:

“How can each function fail, why would it fail, and what happens if it does?”

This step creates the backbone of your risk evaluation: unambiguous chains that will be rated in Step 5 and improved in Step 6.

---

🎯 Objectives #

• Write precise, testable Failure Modes (deviations from Step-3 specs).
• Trace Effects at the next operation/line and the end user/vehicle.
• Identify specific Causes using 6M (Man, Machine, Method, Material, Measurement, Mother Nature).
• List current Prevention & Detection controls (as-is), and set initial Severity (S).

---

🧩 What “good” looks like #

• Function: Press impeller to 12.00 ± 0.05 mm with OK force-distance signature
• Failure Mode (FM): Press depth undersize (<11.95 mm) or oversize (>12.05 mm); signature out-of-window
• Effect (Line): Low retention → slip at run-up / early leak
• Effect (End User): Low/zero flow → engine overheat; leak complaint
• Cause: Fixture hard-stop drift; excessive lube; shaft OD low; press profile mis-set
• Prevention: Hard-stop setting SOP + PM; incoming OD control; lube spec
• Detection: LVDT 100% + signature window; SPC depth chart; daily master checks
• S (initial): 10 (engine overheat if undetected & shipped)

Rule of thumb: Write FMs as measurable deviations; write Effects at two levels (line + end user); write Causes at component/tool/method granularity.

---

🔗 Severity Anchors (examples) #

• S=10: Safety or regulation-critical user effect (e.g., engine overheat due to insufficient flow).
• S=8–9: Major vehicle function loss (e.g., coolant leak leading to breakdown).
• S=6–7: Performance degradation/latent field failure (e.g., intermittent electrical).
• S=4–5: Nuisance/customer dissatisfaction (noise, rework only).
  (Confirm exact scale per your AIAG-VDA table/customer CSRs.)

---

📋 PFMEA Failure Chains — Worked Examples (copy into your sheet) #

#	Station (Op)	Function (from Step 3)	Failure Mode (deviation)	Effect – Next Process/Line	Effect – End User/Vehicle	S	Likely Causes (6M)	Current Prevention Controls	Current Detection Controls	Evidence/Notes
1	OP05 Press-fit	Press impeller to 12.00 ± 0.05 mm; signature OK	Depth undersize/oversize; signature out-of-window	Low retention, slip at OP11 spin	Low/zero flow → overheat	10	Machine: hard-stop drift; Method: wrong press profile; Material: shaft OD low; Man: lube over-apply	Hard-stop setup SOP; profile locked; incoming OD spec; lube spec	LVDT 100%, force signature match; SPC; daily master	Keep last 30 subgroup chart; master results
2	OP05 Press-fit	same	Spline/impeller crack (overforce)	Crack propagates; early failure	Leak/low flow field failure	8	Machine: press over-force; Material: brittle impeller lot; Method: no soft-start	Force window limits; supplier COA	Signature window; visual audit	COA + incident log
3	OP06 Seal install	Install seal orientation; cleanliness	Seal mis-orientation / damaged lip	Fail at OP10/12; rework	Coolant leak at customer	8	Man: handling; Machine: nest mis-align; Environment: cleanliness	Guided fixture; chamfers; handling WI	Vision 100%; particle gate; OP10 pre-leak	Vision false accept rate trending
4	OP06 Seal install	Maintain shaft Ra ≤0.30 µm	Surface roughness high	Accelerated wear	Leak in service	8	Material: incoming shaft finish; Method: skipped audit	COA; Ra audit plan	OP10/12 tests (late); random Ra measurement	Consider earlier Ra check
5	OP07 Potting/ESD	Dispense mass/time; cure profile; ESD <100 V	ESD event uncontrolled (no strap/mat fail)	Latent PCB damage; passes OP11	Intermittent failure in field	9	Measurement: ESD tester faulty; Method: bypass checks; Environment: dry air	ESD policy; start-of-shift check	ESD monitor log (manual); OP11 may not detect	D likely high; plan continuous monitor
6	OP07 Potting	Mass/time; cure profile	Mass low / voids; under-cure	Moisture ingress; OP11 pass; OP12 pass	Early electronics failure	7	Machine: dispenser drift; Method: wrong recipe; Material: resin aged	Recipe lock; PM on dispenser; FEFO for resin	Scale log 100%; cure profile log	Add in-process mass check
7	OP09 Torque	Tighten to 4.0 ± 0.3 N·m + angle	Under-torque	Joint loosening; leak at OP12	Leak / vibration in field	9	Machine: tool calibration drift; Method: wrong strategy; Man: wrong socket	Torque strategy; tool cal schedule	100% torque/angle trace; socket ID	Daily audit curves
8	OP09 Torque	same	Over-torque / cross-thread	Cracked housing; scrap/rework	Leak in service	8	Man: mis-start; Machine: tool strategy; Material: thread burrs	Thread-start aid; SOP	Curve pattern check; visual thread check	Add thread forming guidance
9	OP10 Pre-leak	Air decay ≤ X @ 1.0 bar	Bench recipe mis-set / leak at fixture	False pass/false fail; late discovery at OP12	Leak reaches customer if OP12 bypassed	8	Method: unlocked recipe; Machine: valve seal wear	Recipe lock; PM seals	Daily master part; result to MES	If both weak → high D
10	OP12 Final leak/flow	Leak ≤0.5 @1.5 bar; Flow ≥35 L/min @ ΔP 20 kPa	Recipe wrong / test bypass	Unit ships untested	Leak / overheat in field	10	Measurement: role mis-config; Tool: bench comms lost; Method: manual override	MES role-based lock; No scan → no progress gate	Auto-upload curves; bench heartbeat	Audit logs reviewed?
11	OP12 Final flow	Flow ≥35 L/min @ ΔP 20 kPa	Meter mis-cal / clogged filter → false pass	Low flow escapes	Overheat in field	10	Measurement: missed cal; Environment: temp/viscosity	Cal matrix; temp comp; filter PM	Daily master at 2 points; curve plausibility	Add dual-check (ΔP vs RPM sanity)
12	OP01 Kitting	Assign serial; verify kit	Wrong impeller variant in kit	Fail at OP12 (low flow)	Overheat in field if undetected	10	Material: barcode duplication; Man: kit mix; Method: no double-scan	Two-scan check; kit layout; color coding	Scanner + MES kit check 100%	Add vision shape check
13	Interface – MES	No scan → no progress	Gate disabled / role override	Downstream tests not tied to unit	Untested unit ships	10	Method: IT role; Machine: network outage	Role policy; IT change control	Gate audit log	Add automated alerts
14	OP08 Connector	Fit/Join per spec	Mis-seat / incomplete crimp	Intermittent current; OP11 fail or pass	Intermittent field failure	7	Man: technique; Machine: crimp wear; Measurement: pull test gap	Crimp PM; WI with photos	Pull test (sample); vision 100% for features	Increase sample on pull

How to use: Keep Functions identical to Step 3 wording. Phrase FMs as quantified deviations. List Effects at both levels. Causes should be specific and verifiable. Controls reflect current state only (don’t pre-write actions).

---

🧠 Writing Tips (with examples) #

Bad FM: “Press not OK.”
Good FM: “Press depth <11.95 mm or >12.05 mm OR force signature outside window.”

Bad Cause: “Operator error.”
Good Cause: “Operator applied excess lube (2× spec) causing hydraulic lock,” or “Hard-stop set +0.15 mm post-PM.”

Effects: Always include line effect (scrap/rework/late detection) and end-user effect (safety, regulatory, performance). Severity is set from the end-user effect if the defect escapes.

---

🧯 Prevention vs Detection (don’t mix them) #

• Prevention controls influence Occurrence (O) (design of process: fixtures, recipes, tooling strategy, PM, training, incoming controls).
• Detection controls influence Detection (D) (what checks/tests can catch the defect before ship: gages, benches, vision, SPC, interlocks, curve storage).

In Step 5, justify O with evidence (capability, PM compliance, MSA). Justify D with coverage & timing (100% vs sampled, early vs late, automated vs manual).

---

🧱 6M Cause Library (starter list to seed brainstorming) #

• Man: certification lapsed; fatigue; ESD strap not worn; wrong socket; improper lube technique.
• Machine: servo press drift; DC tool calibration; vision lighting; leak bench valve leak; flow meter drift.
• Method: recipe unlocked; wrong press profile; missing thread-start step; potting start/stop timing; no “no scan → no progress.”
• Material: impeller ID spread; brittle batch; seal hardness out of spec; resin past shelf life; wrong screw finish.
• Measurement: LVDT out of cal; master part unstable; pull tester mis-zero; temp probe offset.
• Mother Nature: humidity <30% (ESD risk); temp shift affecting viscosity and ΔP; airborne particles at seal station.

---

📑 Template (paste into your PFMEA workbook) #

Process Step	Function (from Step 3)	Failure Mode (deviation)	Effect – Line	Effect – End User	S	Cause (6M)	Current Prevention	Current Detection	Notes/Evidence

(Fill one row per failure chain. Add your PFMEA Row ID to link later to Control Plan and Actions.)

---

✅ Outputs of Step 4 #

• ✅ Clear Cause → Failure Mode → Effects chains for each critical station & interface.
• ✅ Initial Severity (S) assigned from the end-user effect.
• ✅ Current controls captured, ready to justify O and D in Step 5.
• ✅ A prioritized list of chains to rate first (press-fit, seal, torque, leak/flow, ESD, MES gates).

---

⚠️ Common Pitfalls (and fixes) #

Pitfall	Fix
Vague FMs (“not ok”)	Use measurable deviations & thresholds.
Effects only at line level	Add end-user effect; S comes from there.
Generic causes (“operator error”)	Make causes specific and auditable.
Controls written as future actions	Document current controls only; actions are Step 6.
Skipping interfaces (MES, recipe lock)	Add interface rows—they prevent major escapes.

---

🔗 What’s Next #

Proceed to Lesson 5.5 — Step 5: Risk Analysis (S-O-D & AP). You’ll assign Occurrence and Detection ratings using the controls listed here and determine Action Priority (AP) to drive Step 6 improvements.

---

🧠 Pro Tip #

If you can’t point to evidence (MSA, calibration, SPC, PM logs, master checks) for a control, assume D will be weaker (higher) and O higher—and plan to fix it in Step 6.