June 12th, Thursday, 10.30 am

The team met to discuss the outcome of the concentration chart. Singh, the operator had recorded the rejection data cavity wise. They were using a 4 cavity mould and Singh had recorded the rejection % cavity-wise for one week. The following was the pattern of rejection. Cavity-1 - 3400 ppm, Cavity-2 - 4500 ppm, Cavity-3 - 4000 ppm and Cavity-4 - 4200 ppm. Singh said "It is clear that there is no pattern cavity wise and hence mould cannot be the reason for the problem". Everybody agreed and decided the next course of action. Sinha and Harish were trained on Shainin techniques and they decide to identify the appropriate technique for the suspected sources of variations which had been identified in the Problem definition. Sinha said "Guys, the work we are doing now is called the Measure and Analyze phase of Six sigma. This is the phase, where we are going to use the Shainin techniques to pinpoint the real reasons creating the problem". He and Harish guided the team and came up with the following table illustrating the technique to be used for each of the suspected sources of variation.

Harish proposed to do Product/Process search for the four SSV's which everybody agreed. The responsibility for doing PPS was given to Poonam. Since Poonam was not trained in Shainin technique, Harish instructed her on how to collect the data. "Poonam, you collect 8 pieces with sand blasting and 8 pieces without sand blasting in one hole. Then you take them to the moulding machine after application of the chemical and do the moulding. Check whether the 8 pieces which have been shot blasted were having the defect of no bonding. If they had, then scale as either 1 or 2. 1 meant no bonding at all and 2 meant that there was rubber bonded to metal, but only partially. If the bonding was perfectly ok, scale it as 0. Do the same for the 8 pieces which have not been shot-blasted in one hole". Poonam said "OK, but how to collect the data for Time delay". Harish said "You take 8 pieces after chemical application and mould them after 24 hours. Take another 8 pieces after chemical application and mould them after 2.5 hours. Again, check the response after molding and scale as I had told you earlier". Poonam said "Yes, I get it now, I will do the similar procedure for handling with bare hands and also for rework done due to underfill". Now, Singh interrupted saying "Madam, rework ke liye, aisa karna zarrori nahi hai. Maine already dekhliya hai, agar hum rework karenge tho, definetly un components mey bonding sahi nahi atha hai". Sinha agreed, but he told the group that now they would have to work on how to reduce the rejections due to underfill. For this, the team agreed that they would do Product/Process search for the input blank weight to establish the relation between blank weight and underfill. Everybody agreed and decided on the following actions:

• Product/Process search for No sand blasting, Time delay and Handling
• Product/Process search for Blank weight to check the relation between Blank weight and Underfill.

Sinha said "Lets us now meet on June 14^th, Saturday with the above data to discuss the findings and decide the further course of actions". The team dispersed.

The team now analysed the data. It is very clear that Sand blasting miss was one of the confirmed causes for the problem, since all the 8 pieces which were not sand blasted were either rejected with scale 1 or 2 and all the 8 pieces which were sand blasted were ok. Harish said that in Shainin terminology, the Total count for this was 16 and it was a confirmed cause. Time delay was not the reason since the pieces with 2.5 hours as well as 24 hours time delay were ok. For weight, Harish said that they had to do the counting and the Total count was 8. Since the total count was > 6, this was a confirmed reason for Underfill. Harish said "Now we can also fix the specification on how much should be the blank weight to avoid underfill from the same data. The specification should be minimum 9 gms, since after 9 gms all the pieces are good".

The team summarized the findings of the M&A phase as follows:

Sinha said "Now, since we are pretty clear about the causes, we will come up with an action plan for improvement and the responsibilities for implementation. Production head Durai, who was also present in the meeting insisted that these causes should be validated before implementing the actions. Sinha completely agreed with him and planned to use B vs C (Better Vs Current) as the tool for validation. For doing B vs C, Sinha suggested that all the causes be validated at a time by creating the Better condition, where the causes would not be present and Current condition, where all the causes would be present. He told Sinnarkar to do this and monitor the rejections in the Better and Current condition. The team worked out the improvement plan with responsibilities as follows:

The team decided to implement these actions within 15 days and meet on July 5^th to decide the controls required to sustain these actions.

July 5^th, Saturday

The team decided the controls that needed to be in place to sustain the actions. Durai also joined the meeting and he was happy that the rejections had already started decreasing. He was confident that it would become zero once the controls were implemented. The team identified the following controls for each action to ensure that the actions sustained over time.

After implementing these controls, the rejections were tracked for over 2 months and it was continuously zero. Annual tangible savings to the Organization was Rs. 2 lakhs. More importantly, the Organization was able to get a new Customer order worth Rs. 2 crore /month. Previously, the Customer was importing this part and was facing the same defect.


Mr Ramnarayan is a Six Sigma Consultant with CII-IQ and can be reached at ram1270@sify.com