Control Phase Blog

Control Phase Blog

The Control stage is the final stage of the DMAIC process. For the purpose of this this project the Control Phase will include the following; an outline of how the process will be controlled, how ongoing compliance will be achieved, an early review of the effectiveness of the process improvements via control charts and by comparing metrics before and after improvements. a project of savings achieved will also be included in the final report.

A P-Chart will be used to control the process. The data will be gathered by the team members on the line, these data sheets are collected by the production administrator and inputted into an Excel spreadsheet, this data is automatically generated into a P-chart as it is inputted. This P-Chart can be reviewed at any time by the Production Manager. After implementation of the changes a review will be carried out weekly by the team leader (me). This review will continue for a significant period of time into the future.

The Table below was taken from the Measure phase and updated to include the metrics after the improvements were implemented.

	%Rejects before improve	%Rejects after improve	Rejects before improve	Rejects after improve *	DPMO before improve	DPMO after improve*	Yield before improve	Yield after improve*	s Level before improve	s Level before improve*
P 1	8%	6.6%	5,010	2,464	531	198	99.58%	99.84%	4.7s	5.04s
P 2	47%	40%	27,823	14,916	2947	1195	97.64%	99.04%	4.2s	4.54s
P 3	45%	53.4%	26,502	19,910	2807	1595	97.75%	98.72%	4.2s	4.45s
T	100%	100%	59,335	37,290	6285	2988	94.97%	97.61%	3.9s	4.25s

 * Signifies a projection of results based on the data gathered since the implementation of the improve phase. The implementation of the improve phase was the 1st of April 2016.

"Objective Statement - Reduce the DPMO to <3140 within 6 months, to do this scrap must be reduced by 50%. This is around 2,472 rejects p/m. If a DPMO of 3140 is achieved; this will allow a sigma level of 4.2 to be achieved."

When we compare our projected metrics from the table to our Objective Statement, we appear to be well on track to achieve our target. However we must not get ahead of ourselves, the metrics in the table are projections from the 1st - 14th of April 2016. The imperative part of this project is now to control and maintain or improve upon these early results.

The addition of the visual inspection system will save between 0.8 - 2 operators per annum after introduction. A projection of savings for the year will be included in the control phase of the final report. This projection will allow for savings with and without the vision system.

Taking part in this project has been a great experience for the team, as mentioned in the improve phase a DOE was set up for process two. This DOE  will form the basis of a new project, with the aim of again halving the DPMO for process two. This project stems directly from the ground work during this project.

Overall this project can be deemed a success so far, the report shall outline the achieved results in more detail and over a greater length of time i.e. 1st - 28th of April, this will allow for more accurate projections.

Improve Phase Blog

Improve Phase Blog

The goal of the Improve Phase of Six Sigma is to identify and trial/test the potential solutions to the problems that has arisen thus far in the project. Depending on the trial/test results the solutions can then be implemented. Contained below are the solutions that have been implemented or will be introduced in the near future. Changes that were tested and ruled out are also contained. In the Six Sigma report, there is a list of changes that are recommended which where however beyond the scope of this project due to time constraints.

Process 1

During the Brainstorming session of the Define Phase, it was discovered that 3 parts where rejected at the beginning of each shift and 3 after each break. On investigation it became clear that this practice began when an older model of the machine was used for manufacturing, when the newest model was purchased and introduced this practice was needlessly continued. This resulted in 4,500 parts being rejected last year. A rational was formed and approved, an internal validation was then performed. As part of this validation a number of samples where gathered and subjected to rigorous testing. The validation was approved and the number of rejects due to this practice was reduced to 156 per year.

Process 2

A number of changes where made to this process, these are outlined in the below bullet points;

• Horn cleaning product changed from IPA to Klericide
• Inspection lamps installed at workstation, this will enable the operator to insert the Septum the correct orientation.
• Piercing Pin manufacturers will supply two pins per month, this will eliminate downtime in the event that two pins break in quick succession. The average pin lasts 6 weeks, however may vary from 1 day to 10 weeks.
• A sensor which detects a falling drop of silicon has been changed to a more reliable sensor which gives a signal for the the falling drop at a high level of repeatability, without being affected by external factors such as light or vibrations. The older sensor required adjustment / re-tuning at least once per shift. This resulted in 10 - 20 minutes of downtime per shift., since the new sensor was installed it has required no adjusting or re-tuning.
• The opening product jaws, which on occasion catches the product as the nest closes where modified. If a part was caught in the jaws it would be rejected and the machine would error, needing reset. The nest was modified to leave only a lip in place to hold the part - below this lip the diameter of the closed nest was increased to a diameter larger than the part itself.
• The gripper moves slightly overtime, resulting in the robot placing the part in the wrong place.These grippers keyed into place to make movement impossible.
• A new welding nest was manufactured with the aim of reducing or bringing to a minimum the movement of the part as it is being welded. This movement is one of the causes of residual stresses in the finished product. The ideal position was found for the nest and was marked out.
• The response of the septum to temperature / environmental changes was investigated;by drying the septum out at 65 degrees for 3 hours, by cooling the septum at 4 degrees for 3 hours and by checking the moisture content of the septum the processing these septum's and testing the final product - No significant changes where noticed.

Process 3

The Inspection bench was modified to a bench with a lip around the edges in order to prevent parts falling off the inspection bench.

Changes for the combined processes

• A vision system which  will inspect the product after welding but before being the silicon is applied; this camera system has been approved and is due be installed within 6-8 weeks. After installation the time spent inspecting will be eliminated / reduced.
• A visual aid was assembled, printed and laminated for all processes. This defines clearly what is / is not acceptable.
• Operators have been made aware of Tappi cards - one has been supplied to the line for reference purposes.

These changes have been implemented or will be implemented as a result of the three previous Six Sigma phases, with the final decisions being made at the beginning of the Improve Phase. These changes will be monitored continuously and any improvements will be evident quickly as a result of the data collection method defined in the Measure Phase and the construction and continuation of the control charts.

Analyse Phase Blog - March 2016

Analyse Phase Blog

The Analyse Phase is the statistical review of the process, for the Analyse Phase of this project each of the three assembly processes were analysed individually and as a combined process. The following tools were used to analyse each process; Process 1 - A detailed process map, A P-Control Chart (both Excel and Minitab), An Ishikawa diagram and a brainstorming session. Process 2 - A detailed process map, A P-Control Chart, An Ishikawa diagram and a brainstorming session. Process 3 - Brainstorming session. Combined Processes - A P-Control Chart. 

Process 1 - The P-Chart shown on the left was of Process 1 from Jan - Dec 2015. There was special cause variation in both Jan & Dec 2015 as the defective parts exceeded the UCL. In Oct 2015 a point falls on the LCL, which indicates that the process was running at its lowest proportion of defects. Overall process 1 is statistically out of control. On the 2 occasions where the UCL is exceeded, this special cause variation could be attributed to the fact that 2 new personnel where being trained on that process. After reviewing the data, it was found that the night shift did not run in Oct, this could explain why the number of rejects declined in Oct 2015. I found the P-Charts useful throughout the Analyse phase as they clearly illustrate how a process is performing, and it is easy to see when a process is veering out of control. When there is special cause variation or unusual trends it is important to investigate thoroughly and pinpoint anything that was done differently. From the P-Chart for process 1 - we checked our training records and they coincided with the hike in rejects in Jan & Dec 15. As for Oct 15 when the defective parts where at the lowest it became apparent that the night shift did not run that month, it is understandable to have a higher reject rate at night due to no Quality / SPC / Engineer present.

Process 2 - An Ishikawa diagram was constructed in Minitab in order to determine the root causes for the high rate of defective parts produced. When the diagram was complete, all the causes where listed as well as a comment to explain the cause and a potential solution. A detailed process map and a P-Chart where also constructed for process 2. The Ishikawa diagram was found to extremely useful throughout the Analyse phase. The result was a clear and categorical root cause analysis of each process, which was extremely beneficial for the team.

Process 3 - It was decided for process 3 to concentrate on the brainstorming session that took place in the Define Phase. All the points suggested in the brainstorm warrant being investigated further in the Improve phase, however for the Analyse phase these point were elaborated on.

Measure Phase Blog - Feb 2016

Measure Phase Blog - Feb 2016

Data Collection Process Map

The Measure Phase of DMAIC involves how the process data is gathered, determining what is important, studying the collected data and defining the baseline process capability. For the process X assembly line, there is a data collection method in place which allows the Production Administrator to obtain accurate Output and Reject figures for each stage in the process line, this can be used to obtain historical data to find baseline figures and to accurately keep track of future data. This data is gathered using the method shown on the process map to the right;

The gathered data was studied and the Output, Rejects, Yield and Sigma Level were calculated using the formula below; (Please Note the following, Number of Opportunities for defects = 8, Number of units 1,180,000 and the Sigma Level was obtained via conversion tables from the DPMO)

The information obtained from the data and the resulting calculations is shown in the table  to the right;

This table is extremely useful as it gives us clear information on each process and an overall indication of how the combined processes are performing. This will be of benefit when the Analyse phase begins.

A Pareto chart was generated for Process 3; this chart is useful as it gives us a clear  indication of why the parts have been rejected and will be a useful metric in the Analyse phase.

The tools that were used in the Measure Phase have allowed an accurate baseline for the project to be calculated, the data collection method shown in the above process map will be used throughout the duration of this project. As a result of the Measure Phase the 'Project Statement' and the 'Objective Statement' have been updated to the following;

'Project Statement; The product X process has an average reject rate from Jan - Dec 2015 of 4,945 parts p/m. This corresponds to DPMO = 6285. Which indicates a Sigma Level of  3.9 sigma'

'Objective Statement; Reduce the DPMO to <3140 within 6 months, to do this  scrap must be reduced by 50%. This is around 2,472 rejects p/m. If a DPMO of 3140 is achieved, this will allow a Sigma Level of 4.2 sigma to be achieved'

Define Phase - Feb 2016

Project Statement - The production process has an average reject of 4,000 parts p/m from Jan - Dec 2015. This corresponds to a DPMO = 6275 and a Sigma level of 3.9 sigma. 

Objective Statement - Reduce the DPMO to <3125 within 6 months, to do this scrap must be reduced by 50% to around 2,000 rejects p/m. If a DPMO of 3125 is achieved, this will allow a sigma level of 4.2 to be achieved.

Team members were chosen the week beginning 18th of January 2016. I held a team meeting on the 22nd of January, this meeting followed the following procedure;

•  A brief introduction on the six sigma process was given for the benefit of the team members who had no previous experience of the process.
• Critical to Quality (CTQ) and the Voice of the Customer (VOC) where defined and the 'Project Statement' and the 'Objective Statement' were defined.
• The whole process was mapped, with a focus on the three manual assembly processes that will be looked at.
• A brainstorming session took place for these three manual assembly processes.
• The processes were written in the centre of a flip-chart. Each team member anonymously wrote down process improvement suggestions / reasons for high rejects on post-it pads. The post-it's were stuck around the central process. The suggestions were reviewed and combined if similar.
• An action plan was formed.

Brainstorm of process 2

VOC - The customer wants a good quality product that is fit for its intended purpose. For product X this is a clean and solid weld in both the 'filter and plug' and the 'septum swage' weld process, with no leaks from either weld.

CTQ - Septum Weld - A good weld will not leak, it will have a thin rim of flash around the septum, it will sit level in product x, there will be no holes and there should not be excessive flash. Filter and Plug weld - The plug will be placed the correct way, filter will be positioned correctly, no holes in the filter and no cracks in the plug.

Brainstorm of process 3

The CTQ, VOC and the process mapping allowed the operators to see the full scope of the product they are making, and the importance of it when in use by the customer. The brainstorming allowed them to express their opinions anonymously and so, truthfully. Overall the tools used in the define phase where extremely useful in making the problem clear to all involved, getting primary feedback from the process line and outlining possible process improvement ideas.

We aim to achieve the lower reject rate, higher productivity and lower downtime by a combination of the following;

• Better welding horn cleaning products
• Modifications to; welding horn, product opening nest, welding nest, sensory equipment and the inspection table.
• Train operators to change the pierce pin, tune sensor and to adjust the welding parameters as required
• Have a stock of spare parts.
• Define a reject
• Reduce paperwork
• Introduce a camera inspection system.

An extract from the process flow, showing the 3 assembly processes that will be looked at