Home ] Up ] Solve Problem ] Quality 1 ] Quality 2 ] SPC Trends ] Capability ] Quality Cost ] [ Lean Quality ]

Lean Quality

When starting Lean Manufacturing it is assumed that quality systems are fully in place and that all employees understand SPC.  Lean Manufacturing concentrates its efforts on determining ways to improve quality.  Lean methods help quality but it should not be forgotten that high quality is essential if lean processes are to function effectively.

BATCH QUANTITIES

The reduction of batch quantities implemented to accelerate product through the factory considerably assist quality levels.  The reason for this is if bad parts are made somewhere within the process and not detected at the time of manufacture, the defect will soon come to light when the batch is passed to subsequent down stream machining or assembly processes.  You then have the opportunity to go back and examine and rectify the problem process before the problem gets out of hand.  In Mass Production, large quantities of parts are usually run on one set-up and if defectives occur the resultant large batch would have many defectives present.

ERROR DETECTION

In Lean Manufacturing great emphasis is placed on shop floor people being given the authority to stop the process if a fault is noticed.  Once the process is stopped it is vital to rectify the problem immediately because the lean system does not have comfortable in process buffers that allow you to take your time to resolve such issues.  One way to detect errors other than through SPC is to build in error detection on subsequent tooling of down stream processes.  This error detection checks the in-coming assemblies as the new process is about to take place.  For example if a part could be assembled incorrectly on a process the following location fixtures would be designed to check that the part has been assembled in the correct fashion.  If an error had been made in the original positioning then the subsequent fixture would not accept the in-coming assembly thereby alerting the person working the process that an error had occurred up-stream.

WORK PLACE INSTRUCTION

Lean Manufacturing recognizes that it is vital to ensure that people working the particular processes know what is required of them to produce quality product at an economically viable rate.  People’s knowledge of the various processes is noted in a Knowledge Matrix.  This attempts to ensure that a person does not start manufacturing without knowing what to do.  The lean system relies on people flexibility so it is vital that each person understands the processes and quality issues in their area.

USE OF SPC  

There is one major problem with SPC in Lean Manufacturing is the size of the batch.  We have seen that SPC X-bar & R chart control limits are constructed using previously collected data from 20 samples taken through the course of a production run.  So, what if the batch size is too small to gather sufficient data?  First, you should conduct a capability study using 25 readings; this would be the minimum requirement.  It is always wise to conduct these studies because you have to know how the machine performs before setting off blindly manufacturing product.  If the machine proved to be process capable i.e. +-3 sigma distribution was within 75% of the total tolerance and centralized you could establish Pre-control Charts.  In this case you would divide the total tolerance that you considered in the study into +-4 sigma. 50% of the tolerance would be covered by +-2 sigma and would encompass 95.4% of the expected results.  This would mean that 4.6% would fall outside the +-2 sigma range, 2.3% each side.  One reading in 43 would therefore fall outside of the 50% central band.  So to construct the chart we would draw two pre-control limits at +-2 sigma and start plotting the results of samples of one part only.  If a result fell outside a pre-control limit, take a second reading.  If this too fell outside then you would have an indication that the process had moved out of control.  

If the process is found to be considerably inside the drawing limits when the capability study is undertaken the following may well occur.

In this case you can see that the process varies from batch to batch, set-up to set-up.  The process is under control relative to the drawing specification, there is no danger particularly if you get Mr Stat to establish the “modified Limits” and run within them.  However if you see this you know that your set-up procedure is not under control and you need to revisit to see why it varies so much.   Modified control limits are established by setting at a distance of “mean range x 0.95” if using a sample size of 5.  The range chart modified control limit would be set at “mean range x 2.34 if using a sample size of 5.  A purist may well criticise using the above chart in its current condition but it does give control and it does show up when the set-up needs attention.