Tell Me Why You Disagree

I asked to perform a comparison of the product thickness that molded with 3 different resins.

I ask my technician to mold 10pcs sample from each resin type with same process parameters, and randomly select 10pcs sample from the productive lot as reference.

All samples thickness was measured once with dial gauge, below table shows the samples data collected.

Resin 1	Resin 2	Resin 3	Ref
7.497	7.498	7.496	7.497
7.492	7.495	7.490	7.490
7.487	7.505	7.490	7.500
7.503	7.492	7.497	7.488
7.505	7.498	7.498	7.492
7.506	7.498	7.493	7.501
7.508	7.493	7.494	7.506
7.499	7.501	7.498	7.498
7.502	7.492	7.501	7.499
7.499	7.497	7.485	7.500

For graphical presentation;

I use box plot to assess the sample data distribution between groups.

I use ANOVA for equal mean test

I use I-MR chart for process condition assessment when sample being mold.

I was questioned by a group of people from quality control department when I present my result. So many question throw to me at the same time, really make me blur and don’t know what to answer on the spot.

I review the questions again when I m driving home. I would like to answer some of the question on this post for further discussion.

Question: Why I use I-MR chart, why not X-Bar R chart?

Answer: I-MR chart was choosing in this case because each sample was measure once. X-Bar R chart cannot be use since subgroup size is equal to 1

Question: Why I stack up the data points and plot it on a single chart whereas the data were came from 4 different groups?

Answer: Box plot shows data distribution for the 4 group are quite even, ANOVA tested no significant different on the mean between the group. Therefore I don’t see any problem to do this. We shouldn’t only focus on mean, UCL and LCL, comparing the mean against USL and LSL and the data distribution also equally important.

Honestly, I’m quite unhappy in the meeting because of too many why throwing to me. I welcome question with constructive comment, either positive or negative comment.

If you spot anything wrong I did, or you disagreed with what I did, just tell me. But don’t keep asking why without telling me the reasons why you disagree.

Six Sigma Is Not For Fire Fighting Problem Solving

A senior engineering manager told me Six Sigma DMAIC is good for solving problem that already exist, it is not so good to us it for identification of potential problem and prevent it to happen in new product development.

 

Do you agree with her statement?

 

I don’t agree.

 

Let me tell you why I m not aligned with her statement.

 

The original intention Bill Smith (1985) develop DMAIC process is to improve product quality by eliminating variation exist in the manufacturing process.

 

Source of variation come from

Man / people

Machine

Method

Material

Design and

Environment

 

Among the 6, Man/People is the most difficult factor to control and it exists every where and anytime in any manufacturing process. In reality, there is no effective way to control of variation caused by human consistently. The only way we can do is to minimize human interaction between manufacturing processes so that they don’t cause so many problems to us.

 

Minimize human interaction can be done before production start by careful consideration of variations that potentially caused by Machine, Method, Material, Design and Environment. By nature, process variation from the sources could be minor, controllable and consistent; however, it could be amplified on the hand of human being and become out of control if no proper preventive measures were taken care before mass production start. Integration of FMEA in Six Sigma DMAIC process could be very powerful in identifying and assessing potential process problem induced by the source of variation during product development stage.

 

Today we are busy solving process related issue induced by all the six source of variation in the production floor. This is because the five major non human related sources of variations were not taken care during development stage. We end up busy in correcting process methodology, fine tune product design to improve manufacturability, standardize material selection, capital investment in environment control and finally we point the finger to the people for not doing a good job!

 

This is the reason why we often see DMAIC were use in fire fighting, but seldom have chance we seen DMAIC were uses start from product development stage. It is not the weakness of DMAIC in product development, but the willingness of the people in development job put in a little more effort to come out a better solution before hand over to the production folks.