Revision
00-QE20-00014
B
Statistical Process Control Procedure
Approver(S) Designation
Rocio Ivanov Quality Director
Germana Lameira Operations Development Director
Mario Baeza Sr Operations Development Manager
Fabricio Ferreira Sr Engineering Manager
Distribution List
Global Document Control Distribution
Rev Author Change Details
<A to Last
NA To see history of changes please click HERE.
Revision>
New procedure to define the requirements for SPC (Statistical Process
A Fernando Rodriguez
Control) and eSPC (Electronic Statistical Process Control).
1. Purpose
1.1. This procedure explains to Jabil sites how to implement Statistical Process Control in their facility and
the actions which need to be followed to maintain a superior level of quality in Jabil’s products.
2. Scope
2.1. This procedure should be used when SPC tools are set up to monitor the critical or special
characteristics (CTQ) of the manufacturing process or the product, it is no limited to specific function or
department. This procedure also addresses the changes on the process with containment and corrective
actions performed for process optimization.
2.2. This procedure applies to all Jabil Manufacturing sites with the exception of healthcare dedicated sites.
3. Definitions/ Terminology
3.1 SPC - Statistical Process Control
eSPC - Electronic Statistical Process Control
Xbar-R Chart - Control chart for continuous variables based on mean and subgroup range.
Xbar-S Chart - Control chart for continuous variables based on mean and subgroup standard
deviation.
I-MR Chart - Control Chart for continuous variables when samples are single unit.
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 1 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
P Chart - Control chart for attribute data used for defective count with variable sample size. Commonly
used for percentages.
NP Chart - Control chart for attribute data used for defective count with constant sample size.
C Chart - Control chart for attribute data used for defect count with constant sample size.
U Chart - Control chart for attribute data used for defect count with variable sample size.
Mean - Mean is the mathematical average of a set of two or more numbers.
Range - Distance between minimum value and maximum value
Standard Deviation - Measure of variation
USL - Upper Spec Limit
LSL - Lower Spec Limit
UCL - Upper Control Limit
LCL - Lower Control Limit
Cpk – Estimated Process capability
Ppk – Real Process capability
Cmk - Machine Capability
Normality Test - Validates that data has a symmetrical and bell shape probability distribution.
Subgroup - A group of units that are created under the same set of conditions.
CTQ- Critical to Quality
CTP – Critical to Process
ME – Manufacturing Engineer
TE – Test Engineer
CP- Control Plan
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 2 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
4. Responsibilities
4.1. This procedure is developed, updated and maintained by Global Quality Engineering, changes to this
procedure must be required to this group and incorporated on future revisions based on evaluation.
4.2. The Global Quality Engineering Director, Global Operations Development Director, Sr. Operations
Development Manager and Sr. Engineering Manager are responsible for the approval of this document.
4.3. Quality Engineer is responsible:
- To ensure that CTQs are monitored with SPC as required by Jabil and/or customer requirements.
- To verify that SPC is correctly implemented based on the following:
o MSA before SPC implementation is done and acceptable.
o Specification and Control SPC limits are correct.
o Stability and capability of the parameters
o Reaction plan derived from process control plan.
o Escalation plan is set-up properly and maintained.
o Process parameters are in control.
- To verify that corrective actions are adequate to address the uncommon causes of variation.
4.3.0. Manufacturing Engineer or Industrial Engineer is responsible:
To determine which critical parameters, require statistical process control.
For the implementation of SPC process
For the selection of proper Control Chart type, frequency, and sampling plan (if applicable)
To ensure accurate calculation and when applicable, the recalculation of Control Limits for SPC
To provide Operators and Technicians with SPC training, work instructions and applicable tools
To perform timely root cause analysis performed by multifunctional team as applicable, when
causes of variation are observed.
To review escalations from technicians and trigger an in-depth analysis as it may apply (DMAIC,
DOE, etc.).
4.3.1. The technician (ME, IE, and TE) is responsible to collect sample(s), perform the measures
defined by Engineering and plot the results as required. When critical SPC rule is broken, the
technician is responsible for implementing and documenting containment action.
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 3 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
4.3.2. Quality Technician and/or Quality Engineer is responsible for verifying and approving the
corrective action.
4.3.3. Production Supervisor is responsible to ensure that Production Team members
communicate when SPC rules are broken and stop the line when required. The matrix below
shows which SPC rules require the line to be stopped, and which are only warnings.
4.3.4. Quality Manager is responsible to facilitate the resources and ensure the SPC procedure is
followed and effectively implemented.
4.3.5. Engineering Manager is responsible to support the effective implementation of SPC and
provide support for corrective actions as required.
5. Associated Documents
5.1. 00-QE20-00007 FMEA and Control Plan Procedure
6. Process
6.1. Select the product’s critical characteristics (CTQ) and critical process parameters (CTP) that should be
monitored to ensure the required quality level. The parameters to be monitored could also be required
by the customer or other regulatory agency. The critical parameters to be monitored are found in the
process control plan under the section “CHARACTERISTICS (To be controlled – process).”
6.2. Process and Machine capability.
6.2.1. For critical characteristics, the Ppk (long term process capability) or Cpk (short term process
capability) the value required to validate a product parameter is 1.33 as a minimum for non-
automotive processes. For critical parameters and overall automotive processes the value required
is 1.67 as a minimum if the process is between 1.33 and 1.67 process may be acceptable. Contact
the authorized customer representative for a review and approve.
6.2.2. The Process Capability PPk and Machine Capability Cmk should be validated when a significant
change is made to the process or machine, when machine is relocated or at minimum once per
year since last Process Capability validation, the annual validation is required to be performed
after the annual gage RR validation for testers. When process is in statistical control monitored by
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 4 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
SPC and Capability was proved to meet the criteria (before current SPC limits were calculated),
there is no annual capability validation required.
6.2.3. For PpK, CmK, and CpK calculations, 125 datapoints are recommended, when customer
recommend a different sample size it needs to be follow or also can follow VDA table for capability
studies sample size:
6.2.4. For closed loop processes or systems, SPC is not required if the system assures on critical
characteristics a Cpk > 1.33 for non-automotive and Cpk > 1.67 on automotive applications, if the
process is between 1.33 and 1.67 process may be acceptable. Contact the authorized customer
representative for a review and approve.
6.2.5. For Cmk calculation the VDA table 6 could be used to validate the machine or reach a Cmk of
1.33 with a sample size of 85 samples.
6.3. When SPC charts are deployed they should be visible and near the machine or process being
monitored, or a periodically monitoring method should be documented.
6.4. Select the appropriate control chart based on the type of variable to monitor and sampling size.
6.4.1.1. For continuous data:
6.4.1.1.1. For continuous data single sample must use the Individual Moving range chart
(I-MR). For example: Flux Weight, contamination test, destructive test of 1 piece.
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 5 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
6.4.1.1.2. For continuous data with subgroups at size 8 or less, use the average range chart
(Xbar-R), for example: Torque.
6.4.1.1.3. For continuous data with subgroups larger than 8, the recommended chart is the
average standard deviation chart (Xbar-S).
6.4.1.2. For attribute data:
6.4.1.2.1. Any inspection or test that provide attribute output as a defective unit, should use
the P Chart for attribute data (pass or fail) that are common for testers or inspection
stations.
6.4.1.2.2. For counting defects when sample size is constant use a c chart.
6.4.1.2.3. For counting defects when sample size is variable use a u chart.
6.5. Sampling frequency should be defined by the functional team of the work cell based on the Control
Plan.
6.6. Process steps for Limits Calculation - Once proper chart has been selected, the following steps must be
done before SPC limits are defined:
6.6.1. Measuring device should be calibrated and validated through gage R&R. Having a reliable
measurement is a basic rule that must be accomplished before moving further in Statistical
Process Control (SPC) implementation.
6.6.2. Before defining control limits, data should be validated for stability. Process should be stable.
It is recommended to collect 25 subgroups as a minimum.
6.6.3. Normality verification is recommended for continuous data.
6.6.4. To release an SPC chart for continuous variables, the Ppk value required must be 1.33 at
a minimum for non-automotive. For critical or automotive parameters, it may be recommended
to use 1.67 as an acceptance reference.
6.6.5. Control limits should be calculated using Minitab or the formulas provided in the SPC AIAG
Manual.
6.6.6. Once limits are calculated, limits will be plotted into the charts or set up in the
software designated to follow SPC.
6.6.7. Some software may have the automatic limits recalculation. In this case, the Quality
Engineer must verify that the rules follow this procedure and specify them in the local SPC
procedure.
6.6.8. Discrimination level on the measuring device should be at least 1/20 of the distance between
USL and LSL.
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 6 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
6.7. The following SPC rules are required to detect special causes of variation:
6.7.1. Rule 1 - One point out of the control limits – Critical Rule
6.7.2. Rule 2 - Six consecutive points ascending or descending – Critical Rule
6.7.3. Rule 3 - Fourteen points alternating Up and down.
6.7.4. Rule 4 - Nine points on one side of the mean
6.7.5. Rule 5 - Three points out of four points in the 3-sigma band, upper or lower – Critical Rule
6.7.6. Rule 6 - Four points out of five consecutive points more than 1 sigma from center line in the
same direction
6.7.7. Rule 7 - Fifteen points between plus minus 1 sigma.
6.7.8. Rule 8 - Eight continuous points in the 2-sigma zone.
NOTE: These are. Lloyd S. Nelson in his April 1984 in the Journal of Quality Technology and are the
rules used in Minitab Software.
6.8. For cases when using eSPC and conveyor stop feature is activated, if rules 1, 2 or 5 are broken or the
parameter being measured goes out of specification, this will cause the conveyor to stop. A corrective
action will be required to re-start the conveyor.
6.9. Rules 1, 2 and 5 require a reaction plan to address the special cause of variation. Line stops criteria
based on SPC should be defined in the site “stop matrix.”
6.10. These rules should apply or any other more restrictive criteria such as AIAG, VDA or ISO TS 16949.
6.11. When the process is mature and stable and the CpK, Ppk and/or Cmk is over 2.0, the site can decide to
use only Critical Rule #1 or continue with the Rules #1, #2 and #5 to react to uncommon causes of
variation.
6.12. Sampling, plotting, and recording data.
6.12.1. Operator or Technician will take the readings and plot them into the charts.
6.12.2. It is important for charts with subgroups that samples are taken randomly and according to the
frequency specified on the Control Plan (CP).
6.12.3. For P charts it is recommended to collect at least eighty events to calculate each chart point.
6.13. Limits recalculation
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 7 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
6.13.1. Limits should be reviewed based on the volume and recalculation must be done if the variation.
has been reduced or shifted significantly. When limits are recalculated, the requirements of
stability needs to be validated and Cpk should not be lower than the minimum required. Limits
should be reduced not expanded because quality improvement is reflected in variation.
reduction. When a Cpk of 2.0 is achieved, the limits are not required to be reduced.
6.13.2. Limits should be recalculated using at least the last 25 datapoints or a week of data at a
minimum. Points out of control with assignable cause should be removed for the recalculation.
6.13.3. Control Limits are recommended to be reviewed every 3 months, after significant changes of the
process or when (but not limited to) when the rules 4,7 and 8 are broken frequently.
6.13.4. Limits recalculation will require a PCN before implementing the recalculated control limits, in case
of automatic limit recalculation software, PCN is not required and limit changes need to be
reviewed periodically.
6.13.5. If PpK drops below 1.33 for non-automotive and 1.67 for automotive, the team needs to put in
place an action plan to recover the process capability.
6.14. Containment Actions
6.14.1. When a point or group of points break a critical rule, it is required to provide an immediate
correction before continuing to run.
6.14.2. Process Owner must review the containment and corrective actions to verify that actions are
not recurrent. When a containment or corrective action is recurrent, a JOS action must be opened.
When the causes of variation have not been identified, a deeper root cause analysis is required
(A3 or DMAIC).
6.15. Corrective actions (A3, DMAIC, etc.) must be promptly completed and implemented. In the absence of
timely closure, escalation should be followed.
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 8 of 9
Revision
00-QE20-00014
B
Statistical Process Control Procedure
7. Risk & Controls
7.1 Risk & Mitigation
7.1.1 The risks of not having SPC would be lack of process control, poor quality, slow reaction or
delayed improvement to non-conformities. Implementing SPC would mitigate these risks because
special causes of variation will be visible, and sites can address them before quality is negatively
impacted.
7.2 Effectiveness verification
7.2.1 The effectiveness of using SPC is verified by Internal Audits, External Audits and Customer Audits.
8. Revisions History & Change Details: Go back to cover page – click HERE
Rev Release Author (s) Change Details
Date
NA NA NA NA
Documentation Category Mandatory
Business Confidential – All rights reserved Printed copies are reference only.
Page 9 of 9
