This Month in SMT007 Magazine—Continuous Improvement: As Simple as X = Xc – 1

Estimated reading time: 2 minutes

X=X_c– 1 is a conceptual equation for continuous improvement. You define X and work to reduce it by a factor of 1. This could be one work hour, one process step, one day less in a cycle, and so on. We recently met with Ronald Lasky, Ph.D., PE, Sr. Technologist, Indium Corp. and Professor of Engineering at Dartmouth College, to discuss the concept of X=X_c – 1 and get his advice on generating enthusiasm among readers and the next generation about continuous improvement. As this discussion illustrates, many process improvements are small in scale, not yearlong, major efforts.

Nolan Johnson: Continuous improvement is not a new idea, but we would like to shine a light on the idea of “X=X_c – 1.” To do this, first, you define what X is, and then you make an action plan to reduce it by one. It could be reducing your design spins, the number of steps in your process, or the number of gates in manufacturing. It could be any approach where you can incrementally improve, take something out, and get it done with fewer steps and iterations.

Happy Holden: It’s the concept of solid base hits rather than going for the home run all the time.

Ron Lasky: It’s sort of a formulaic approach, but maybe that’s better. Continuous improvement is the essence of Lean Six Sigma. There is a term in Lean Six Sigma called DMAIC (define, measure, analyze, improve, and control), assuming you want to improve something. You define what it is that you want to improve (D). You measure where you are (M). You collect some data and analyze it (A). Then, you improve it (I), usually with a designed experiment. Once that’s all set up, you must develop a plan to control it (C). That’s statistical process control.

For example, let’s say a small mom-and-pop shop collected the data for the year 2020, and they found that at the end of the line before repair, they had 2% fallout: 2% of the boards had to be repaired. They collected the data in a Pareto chart. If they made 100,000 boards, they had about 2,000 boards that were defective, and they found that the primary defect was shorts. That was 1,200, and then the second defect was a missing component and on down the list. Usually, you would want to attack the most significant defect mode. That was shorts. They started to look at what typically causes shorts. Two-thirds of end-of-line defects can be attributed to stencil printing, so that is good place to start.

This mom-and-pop shop may hire a local college student as an intern, who has a Lean Six Sigma Green Belt, and they teach the intern about electronic assembly. The company may also have pictures of all 2,000 defects. They analyze those images and decide that the main reason they had too many shorts is that there’s too much solder paste on the pads. When the component is placed, excess solder paste spills over the pad that occasionally, when it melts, creates a solder bridge to an adjacent pad. They do some more work and decide that maybe they should make all of their stencil apertures a little tighter, or maybe they should get a different solder paste. They call their current solder paste vendor and discuss some of these issues.

To read this entire interview, which appeared in the January 2021 issue of SMT007 Magazine, click here.