My first ever Six Sigma project was thrilling.
At the time, I was working towards my Six Sigma black belt, and I remember fielding all sorts of questions at the plant – mostly from people who wanted to know if I thought Six Sigma was going to work.
There I was, I’d finished the training, learned all about Six Sigma and complicated statistical tools, and taken the tests – but I’d never applied it! I was supposedly the champion of six sigma in my workplace and – I honestly wasn’t sure! I figured, “let’s find out, let’s give it a shot!”
The Problem
The problem I was tasked to solve involved bearings being installed into a housing with a press. Every time the press installed a bearing, it measured the amount of force required to install the part. Over the course of a couple of months, the operators noticed the press force reading was often outside of the acceptable spec.
Now, the bearings we were installing were shipped pre-oiled, from China. The press operators had noticed that sometimes the bearings were dry, so they took it upon themselves to oil them.
All previous attempts to systematize a solution (“okay, always oil’, ‘okay, never oil’, ‘okay, measure the new box of bearings, and if they measure too big, then oil them’) had fallen short.
This was happening multiple times per shift per day, and we were losing a lot of money on this problem. There was a very complex interaction of parameters on the press: speed, position sequence of the programming, the size of the two pieces, whether there was oil or not, and the surface finish on the inside of the housing (which turned out to be the key variable!).
Identifying the key variable.
See, everything in Six Sigma is x’s and y’s. The x’s are inputs, and the y’s are outcomes. The key Y (outcome) we were after was the press force reading on each bearing. We had all these X’s (inputs) and we didn’t know which one was the important one. It was about sorting through ‘the trivial many’ inputs, to narrow down to ‘the vital few’.
It was only after we used six sigma to break apart the problem, and analyze every angle of that we discovered the issue. Our housing supplier, thinking they were doing the right thing, was making the inside of the housing too smooth. The finish was still technically ‘in spec’ but it was always on the high side of the spec as far as surface finish.
The Solution
The surface finish of the housing turned out to be the cause of the high press force. We brought the surface finish down, and there was less variation. Instead of surface finish between 5 and 20, we requested a range of 5 and 12. And voila!
The problem virtually disappeared! Instead of getting 20 rejects per shift, we were only getting one. This was my first taste of applying the tools, and it took us a while to narrow down on that key factor – but we got there.
I’ll never forget this – a thrilling, exciting, formative adventure in Six Sigma!
