Motor Lamination Quality Control: SPC for Lamination Stamping and Stack Dimensions

A stamped lamination can meet every reported dimension and still produce a poor stack.

The bore moves after interlocking. Stack height changes by batch. One strip lane develops burr faster than the other. Tooth width remains inside tolerance, but it has been drifting for hours.

Final inspection sees a part. Statistical process control should see the process that made it: coil, strip feed, die, individual lamination, joining operation, and finished stack.

A useful SPC plan connects those stages. A large inspection plan does not automatically do that. Sometimes it only creates more numbers.

Which Dimensions Should Be Controlled?

Control four groups of characteristics:

• Material inputs: sheet thickness, coating condition, and strip width
• Functional geometry: bore or outside diameter, tooth width, slot opening, slot location, pitch, concentricity, and runout
• Tool-health indicators: burr height, edge profile, feed progression, and selected punch-dependent dimensions
• Stack results: height under a defined load, face parallelism, bore or OD runout, skew, and interlock condition

A dimension belongs in routine SPC when it predicts product function, reveals tool wear, identifies a feed or cavity problem, accumulates during stacking, or carries a high escape cost.

Why Good Laminations Can Build a Bad Stack

A stack is not merely sheet thickness multiplied by lamination count.

H_stack = sum(t_i + c_i + b_i + e_i) - compression_set

Where:

• t_i = base sheet thickness
• c_i = effective coating contribution
• b_i = burr or edge-contact contribution
• e_i = joining or interlock deformation
• compression_set = height removed by the specified load

This is not a design equation. It is a reminder that stable sheet thickness does not guarantee stable stack height.

Punching changes the cut-edge region through deformation, residual stress, fracture, and burr formation. Burrs may damage insulation or create contact between adjacent sheets when the stack is compressed. So burr is a process variable, not only a visual defect.

Motor Lamination SPC Checklist

This is an architecture, not a universal control plan. Bonded, interlocked, and welded stacks do not accumulate variation in exactly the same way.

Do Not Copy Universal Burr Numbers

“Burr must always be below 10 micrometers” looks useful. It may be wrong for the product in front of you.

The acceptable limit depends on sheet thickness, material grade, coating, burr direction, edge location, stack compression, joining method, electrical validation, customer requirements, and the measurement method.

Use two limits where the risk supports it:

1. Product limit: the drawing or validated acceptance limit.
2. Maintenance trigger: a lower internal level that starts tool inspection before rejection.

A plant might place the maintenance trigger at 60% to 80% of the product limit after studying its own tool-life data. That is a policy example, not an industry rule.

Use Standards Without Letting Them Replace Engineering

The ISO 22514 family provides methods for process capability, performance, and measurement-process capability. The current ISO 22514-2 edition addresses continuous characteristics; ISO 22514-7 addresses whether a measurement process is suitable for a defined task.

The IEC 60404 series covers electrical steel materials, magnetic-property tests, grades, geometric characteristics, and inspection requirements. IEC 60404-8-4, for example, covers fully processed cold-rolled non-oriented electrical steel in specified nominal thicknesses. It does not replace the finished lamination drawing or validated stack specification.

Standards define methods and terminology. The drawing defines acceptance. The control plan defines how early to react.

Build the Sampling Plan Around Process Events

Always inspect after a die setup, die maintenance, coil change, strip rethreading, press jam, feed adjustment, stacker adjustment, joining change, or gauge-program revision. Each event can create a new process state.

For high-volume stamping, three to five consecutive laminations are often a sensible starting subgroup. Keep them from the same short period, lane, cavity, and process condition. Do not form a subgroup from the start, middle, and end of a shift; that mixes time variation into one range.

Add hit-based or time-based checks. Set the interval from tool-life history, press speed, tolerance, previous drift rate, measurement time, containment cost, and escape severity. “Every two hours” may be correct. The history should explain why.

Keep part and stack sampling separate. Part data reveals die and feed behavior. Stack data reveals accumulation, joining, compression, and indexing. Link both by coil, lane, cavity, hit range, stacker, and batch.

Choose the Chart That Matches the Data

Use an Xbar-R chart when several consecutive values can be measured as a subgroup. It often suits tooth width, bore diameter, slot opening, thickness, and repetitive stack height.

Use an I-MR chart when only one result is practical at each interval, such as a burr profile, full stack runout, skew measurement, destructive test, or low-volume stack.

Use an attribute chart for counts or proportions, such as missing interlocks, coating damage, or rejected stacks.

Do not combine strip lanes, cavities, dies, or stackers into one chart merely because the parts share a drawing number. One chart should represent one process stream. Usually.

Validate the Measurement Process First

Measurement noise can look like process movement.

Define the datum scheme, fixture, force, support points, temperature condition, optical edge rules, burr trace direction, stack-compression force, dwell time, operator method, and gauge resolution before collecting baseline data.

A stack-height result without a specified load is incomplete. A burr reading without a fixed location is hard to compare. An optical program with changing edge thresholds is not a stable measurement process.

Capability Formulas in Plain Text

Calculate capability only after the process is sufficiently stable.

Cp = (USL - LSL) / (6 * sigma_within)

Cpk = min(
  (USL - mean) / (3 * sigma_within),
  (mean - LSL) / (3 * sigma_within)
)

Pp = (USL - LSL) / (6 * sigma_overall)

Ppk = min(
  (USL - mean) / (3 * sigma_overall),
  (mean - LSL) / (3 * sigma_overall)
)

USL and LSL are specification limits. sigma_within estimates short-term variation. sigma_overall includes variation across the selected period.

Worked Example: Tooth Width

The figures below are only a teaching example.

Nominal = 5.000 mm
LSL = 4.960 mm
USL = 5.040 mm
Mean = 5.014 mm
sigma_within = 0.006 mm

Cp = (5.040 - 4.960) / (6 * 0.006)
Cp = 2.22

Upper Cpk = (5.040 - 5.014) / (3 * 0.006) = 1.44
Lower Cpk = (5.014 - 4.960) / (3 * 0.006) = 3.00

Cpk = min(1.44, 3.00)
Cpk = 1.44

Cp is strong because the spread is relatively small. Cpk is lower because the process is shifted toward the upper limit. Before adjusting the die, check the control chart, lane, cavity, gauge stability, and recent tool events.

Example: Combined Data Can Hide a Problem

Consider a two-lane strip with hypothetical tooth-width averages:

The combined mean looks ideal. Neither lane is behaving ideally.

Separate charts may point toward different wear, alignment, or clearance conditions. The same masking happens when data from multiple cavities or stackers is pooled.

Reaction Plan for an Out-of-Control Signal

1. Contain material from the last verified good check.
2. Confirm the signal with the defined gauge and method.
3. Identify the coil, lane, cavity, hit range, press, stacker, and batch.
4. Check linked characteristics; compare burr with slot size and edge profile, for example.
5. Inspect the likely source: feed, pilot, punch, die clearance, stripper, fixture, interlock station, or joining settings.
6. Correct the cause and verify consecutive conforming subgroups.
7. Release or sort contained material using a documented rule.
8. Record the cause, action, and verified restart point.

Do not recalculate control limits because one point crossed a limit. Change them after a real, understood process change.

What Buyers Should Ask a Lamination Stamping Supplier

Ask for more than one Cpk value:

• Lane- and cavity-separated capability data
• Measurement-system evidence for bore, tooth, burr, and stack checks
• Exact stack-height load, dwell time, and fixture method
• Burr maps and die-maintenance triggers
• Traceability from coil to lamination batch to stack
• Reaction, containment, and process-change rules
• Evidence that part-level and stack-level records are linked

A green dashboard is not enough. The process identity behind each number matters.

FAQ

What are the most important SPC dimensions for motor laminations?

Start with sheet thickness, bore or OD, tooth width, slot opening and location, runout, burr height, flatness, interlock geometry, stack height, face parallelism, and skew. Adjust the list to the product failure modes.

Should every drawing dimension have a control chart?

No. Some dimensions need setup approval or periodic audit only. Routine SPC should focus on functional features, wear indicators, high-risk dimensions, and characteristics that accumulate in the stack.

Is there a universal burr-height limit?

No. The approved limit should come from the drawing, coating system, stack compression, electrical validation, customer requirement, and agreed measurement method.

How often should burr height be checked?

Check after setup and maintenance, then at an interval based on hit count and observed wear rate. Measure fixed high-risk locations, not whichever edge is easiest to reach.

Why can stack height vary when sheet thickness is stable?

Coating, burr, waviness, debris, interlock deformation, joining conditions, part count, compression response, measuring load, and dwell time can all contribute.

Which Is More Important, Cp or Cpk?

Cp reflects potential capability based on spread. Cpk also reflects centering. Neither is reliable without a stable process and suitable measurement method.

Can Good Lamination SPC Guarantee a Good Motor Core?

No. The finished stack still needs controls for height, runout, parallelism, skew, joining, insulation condition, and product-specific performance tests.

A Practical Rule

Control dimensions that predict function. Track measurements that reveal wear. Verify the finished stack under defined conditions.

Keep the data separated by its real source: coil, lane, cavity, tool, press, and stacker.

That is how SPC catches a changing process before final inspection has to explain a failed stack.

Request a Lamination SPC Review

For a useful manufacturability or quotation review, provide the lamination and stack drawings, material grade and thickness, coating, annual volume, joining method, stack-height measurement load, burr and flatness requirements, skew requirement, required capability indices, and traceability expectations.

A quote becomes more accurate when the control method is defined with the part, not added after production begins.