Scribd
Upload
119 views

Common Mistakes Commirtted in Statistical Process Control

This document discusses common mistakes made in Statistical Process Control (SPC). It outlines several issues such as control limits being adjusted too often or incorrectly, not using range charts, ignoring points beyond control limits, poor knowledge of non-random patterns, and not analyzing charts. It also discusses problems like others maintaining charts instead of operators, inadequate training follow-up, using the wrong control charts, mixing multiple processes in one chart, and not recording corrective actions. The document emphasizes properly understanding and implementing SPC to control and improve processes.

Uploaded by

Arvin Babu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
119 views

Common Mistakes Commirtted in Statistical Process Control

This document discusses common mistakes made in Statistical Process Control (SPC). It outlines several issues such as control limits being adjusted too often or incorrectly, not using range charts, ignoring points beyond control limits, poor knowledge of non-random patterns, and not analyzing charts. It also discusses problems like others maintaining charts instead of operators, inadequate training follow-up, using the wrong control charts, mixing multiple processes in one chart, and not recording corrective actions. The document emphasizes properly understanding and implementing SPC to control and improve processes.

Uploaded by

Arvin Babu
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Statistical Process Control ( S P C )

Common mistakes committed in Statistical


Process Control ( S P C )
When SPC is properly understood and implemented it will be a very powerful tool helps both
operators and management to control and improve the process. Whereas many SPC
applications are having lot many problems in them. Some of the most common problems are
explained below:

• Control limits are adjusted too often :


If the process is under statistical control and is a stable one, there may not be any need to
adjust the limits. The obvious question would be when to adjust the limits.
Limits are adjusted onlyif..
a) To verify process improvements
b) When the process inputs are significantly changed ( which may result in change in CPk)

We should ensure whether the computer protocol automatically adjusts the control limits.

• Control limits are incorrectly set :


Some of the common errors would be
a) Forcing the limits to be symmetrical about the product nominal or some other arbitrary target
value.
b) Failure to use process data to establish the limits.
c) Fixing control limits even without homogenization
d) Arbitrarily setting the limits to some predetermined values( Eg 50% of the tolerance range. We
should understand the formula for establishing control limits do not make use of product
specifications.

An Intellectual Property of Sri Padhmam www.spel.net.in +91 99621 17222


Common mistakes committed in Statistical Process Control ( S P C )

• The range chart is not plotted :


There is a wrong notion, that controlling the average alone will ensure process control.
Whereas, without a range chart process variability (piece-to-piece variation) will not be
understood completely.

• Points plotted beyond a control limits are ignored :


The general excuses are :
a) “I ensured the sizes of few parts afterwards…”
b) “It is only one point which has exceeded…”.
c) It is a very clear case of assignable cause is ignored.
d) This cannot be just like that ignored as a “Statistical freak”, may result in reduced Cpk.

• Poor or no knowledge on Non-random patterns :

a) The common misconception is, that process is out of control when points go beyond the
limits.
b) Within the limits also there could be non-random patterns such as unexpected trends,
cyclic patterns, runs, hugging of control limits
c) Seven or more consecutive points in one side of the chart (in case of X-bar average)
may be a clear case of run.

• Charts are completed and duly filed (!), without any analysis :

a) Management to take the initiative to introduce a system in place to ensure periodical


analysis and suitable actions.
b) SPC charts are for process controlling and improving not just to satisfy the auditors.
c) If analysed properly, one can gain lot of clues from a control chart.

An Intellectual Property of Sri Padhmam www.spel.net.in +91 99621 17222


Common mistakes committed in Statistical Process Control ( S P C )

• Charts are plotted and maintained by someone other than the operators concerned :

a) The best way to implement SPC is to provide adequate knowledge to the operators
concerned and making them responsible to plot control charts.
b) I have seen the good practice of Operators plotting the Control charts - in very few
organizations. One such organization is Lucas TVS, Pondicherry plant.
c) As long as someone else is checking the process/product characteristics and charts are
plotted, the ownership is under question mark.

• SPC training doesn’t end with a practical hand holding in respective work area :

a) Most of the organizations pay very little attention towards this.


b) A classroom education on any subject is a beginning but not an end
c) Equal importance to be given by the management to ensure an atmosphere for
implementation.
d) Without this, the money spent for the training is only an expenditure but not an investment.
e) It is ideal to ensure implementation within two weeks’ time for better effectiveness.

• Applying SPC to the process, whereas wrong control chart is selected :

a) It is of paramount important to introduce right control chart for the right purpose.
b) One example would be Introduction of X-MR chart for controlling a process parameter.
c) I have seen in one place, wherein X-bar R control chart was deployed to control a Blanking
process.
d) The process being very consistent, the range chart was not sensitive at all.
e) Here the ideal option would be X-bar S chart.
f) A wrong control chart will yield absolutely no results.

An Intellectual Property of Sri Padhmam www.spel.net.in +91 99621 17222


Common mistakes committed in Statistical Process Control ( S P C )

• Mixing more than one process in a single chart :

a) This is generally happening where the machine is having multiple stations such as
multiple head drilling
b) Though this may appear to one machine, each spindle must be considered as a separate
process
c) The same logic is applicable for multiple cavity mold also.

• Control limits are not reviewed and adjusted based on the data :

a) This is almost as bad as adjusting the limits too often.


b) Based on the behavior of the process, control limits to be recalculated and adjusted by
an SPC expert.
c) All apparent changes to the process should be validated before any changes to the limits.

• Using SPC charts without having control limits in them :

a) This may be acceptable in 2 situations viz. Stability check through control chart and in the
initial stage of SPC application, where our interest is to gather data.
b) But for the above situations, control charts without control limits are not acceptable.

• Fixing the control limits based on the Product specification :

a) This is a very common error.


b) This will never help to statistically control the process and to achieve the required CPk.

An Intellectual Property of Sri Padhmam www.spel.net.in +91 99621 17222


Common mistakes committed in Statistical Process Control ( S P C )

• Corrective actions are not recorded:

a) This is a great mistake as will lead to loss of knowledge base.


b) Over a period of time, the organization will lose the grip over the process.
c) It would be ideal to introduce a coding system for the causes and the corrective actions.
d) Recurring problems would be analyzed based on this record.

Best Wishes for bringing a Better

Process Control through right

control charts and SPC practices!

An Intellectual Property of Sri Padhmam www.spel.net.in +91 99621 17222


Sri Padhmam Consultancy & Training
Program Topic
Measurement Systems Analysis
( MSA 4th Edition )

Resource Person : A V Manivannan, Principal Consultant & Trainer

Practical exposure to Understand the Concepts & the application of Get your e-Certificate after
statistical techniques to estimate the Capability & performance of successful completion
measuring / inspection systems of the course

Program Date & Time : 16-03-2018 (Friday) & 17-03-2018 (Saturday) [ 2 days ] 09.00 am to 05.00 pm
Target Participants : Engineers & above - from QA, SQA, Manufacturing ,NPD & other engineering functions

Date Course Deliverables

✓ MSA –what & why


✓ Role of statistics in MSA. Terms and MSA over view
✓ Key Changes in the 4th Edition
✓ Application of Statistics on MSA
16th
- Sigma factor
March 2018
- Accuracy ( Bias ) & Precision
(Friday)
- Confidence level with a Case study
- t distribution with a Case study
✓ MSA on Variable systems
- Stability check with a Case study
- Bias with a Case study

- Linearity (Numerical method ) with a Case study


- Linearity (Graphical method) - an introduction
- Gauge R & R studies
17th
- Range method with a Case study
March 2018
- Average & Range method with a case study
(Saturday)
- Significance of ndc
✓ MSA on Attribute Systems
- Kappa method, Effectiveness, False alarm & Miss rate with case studies
- Signal detection method

Program Fee: Rs. 4500 Rs.3,900/- (incl. GST) per participant Venue:
(Complementary Work book, Stationary, Refreshments & Lunch) THE VIJAY PARK Hotel
Corporate offers : Next to Arumbakkam Metro Station, Near
CMBT Bus Stop, Arumbakkam, Chennai.
No. of participants Program fee ( incl. GST )

3 employees from the same company Rs.3,600/- per head Participants are encouraged to bring
• Scientific calculators
4 to 6 employees from the same company Rs.3,300/- Per head • their cases for clarifications

Last date of Registration: 14-03-2018 ( Wednesday )

For Registration: +91 9962 117222 [email protected] 23

You might also like