Script

How to manage common cause and special cause to produce consistent good quality product?

There are a lot manufacturers who do NOT know about using process control to minimize process input variations  and much less about common cause and special variations.  Quality issues could happen due to poorly controlled common cause process variation  or  there is a sudden change in 5Ms 1E which is not well managed which become special cause variation.  Sometimes quality issue already exist,  however it is not detected in the manufacturing process due poor inspection or test coverage  or incorrect data collection and analysis.

In order to successfully reduce common cause and eliminate special cause,  we must able to know if the quality issues are contributed by common and special cause type of variation.  The approaches to manage special cause and common cause are very different. 

From article posted on 11 Sep 2017,  we have learned on how to detect common and special variations thru preset rule for reject count or using more advance statistical analysis,  statistical process control (SPC) chart.

Once we have detected and identified the type of variation then we should learned how to manage those variations to ensure we reduce or if possible eliminate those variations  to produce consistently good quality product.

Common cause variations could derived from several root causes, in 5Ms and 1 Es process input and cannot be eliminated.  However it can be reduced through systematic approach of measuring correctly, collecting and analyzing data statistically  to understand which 5Ms or 1E contributed to the source of variations.  Then we can create action plan to address source of variations such as optimizing certain machine settings which have an impact on quality parameter.  This methodology is known as six sigma pioneered by Motorola engineers in 1980s.  Six sigma methodology consists of 5-6 phases, abbreviation as DMAIC or DMAICR in some organization. Below are a snapshot of the six sigma approach.

Phase 1 – Define – Detect, understand, scope problem such as high reject rate for certain reject symptom

Phase 2 -  Measure – Identify  what is the measurement metrics which represent quality feature to understand current status.  Conduct measurement system analysis (MSA) to check if measuring process is  reproducible and repeatable to ensure data collect from an inspection or test or measuring process can be trusted and does not have too much variations.  Then only collect and analyze  quality metrics data to understand current reject rate or process performance.

Phase 3 -  Analyze -  Root cause analysis  through using fish bone diagram, 5 whys or fault tree analysis or other organized root analysis  tools.

Phase 4-  Improve -  Once root causes is isolated,  which could be from multiple source,  then we generated improvement plan to counter the root causes  and check if improvement plan effective by collecting and analyze data and compare against data collected from measure phase.  The data will show improvement if the improvement plans are effective

Phase 5 -  Control, once we have improvement plans which effective to control variations,  then we will need document the actions taken for  standardization.  Then we will continue monitor the process to ensure improvement plan is sustainable.

Phase 6 -  Report all work done, some organizations  consider reporting is done in control phase, thus their 6 sigma phase is DMAIC

Special cause variations which are associated to changes in process sometimes could process improvement which is desirable or changes in unknown source which cause process to behave worse than before.  Our focus would be to eliminate this unwanted change (special cause) to ensure our process is behaving normally again.  We would need to use systematic problem solving approach such as corrective action approach 8- discipline approach, 5C methodology and 7 steps which I will write in details in my next article.

Attached is the table which compare both cause of variations in manufacturing process.
  

Common cause
Specials causes
Strategy
Reduce
Eliminate
Source of variation
Already part of process input
Sudden change in process input
Common Tools used
6 sigma approach

8-D, corrective action, error proof, 5C, 7 steps problem solving, Kaizen
Time taken
Could be months
Days or  1-2 weeks
Application
Continuous improvement plan
Customer complaint, sudden change with reject rate spike

As the strategy to address both common cause and special cause variations is different , it would be fatal and costly to treat special cause variations as common cause variations vice versa. 

Unfortunately there are many manufacturers who does NOT know how to identify common cause and special cause variations when there is quality issue due to :-

  • Manufacturer handle common cause as special cause issue when they did not realize common cause variation already exist during the start of new product manufacturing.  They are unable to detect the problem until their customer complaint.
  • Manufacturer treat special cause as common cause variation when both causes exhibit same failure symptom.  For example cosmetic defect such as scratch always happen in mechanical parts which is a common cause variation.  If there  is a sudden high spike of the scratch reject symptom,  then we could have a special cause variation mix with common cause variations.  Common cause issue root cause could be natural variation contributed by process input as there are zero defect process which are non existence.  The high spike could be due to a single source problem where a new jig had caused scratch to the part. 

The mishandling of common and special cause are the reasons why some issue are NOT properly resolve.  Therefore the key to a consistent good quality product is to understand the type of variation that  is contributing to poor quality followed by improvement to address type variation from root cause.

No comments:

Post a Comment

Sharing is caring

Continuous Improvement Program CIP - 6sigma Methodology