New Control Chart for Electronics Assembly
Paper offers extension of the defects per million opportunities control chart. Examples of in-control and out-of-control processes are offered.
DOWNLOAD
Authored By:
Daryl L. Santos, Ph.D.
SSIE Department
Small Scale Systems Integration and Packaging S3IP
A NYS Center of Excellence
Binghamton University, Binghamton, NY, USA
Summary
As six sigma and better processes are demanded for higher yields and as organizations move from measuring defects in terms of parts-per-million (ppm) towards parts-per-billion (ppb), the resolution of extant control charts is becoming insufficient to monitor process quality. This work describes the development of a new statistical process control (SPC) chart that is used to monitor processes in terms of defects-per-billion-opportunities (dpbo). A logical extension of the defects-per-million-opportunities (dpmo) control chart, calculations used to derive the dpbo control limits will be presented and examples of in-control and out-of-control processes will be offered
Conclusions
From an examination of the example presented, it is apparent that the chart limits and plot points for the dpbo chart are 1000x their respective values for the dpmo chart as they should be. The dpbo has been presented as a scaled version of not only the dpmo chart, but also of the u chart. The interested reader might then pose the question (which is similar to one posed in a recent graduate engineering class covering SPC concepts (Santos, 2009b): Why don't we just get the dpmo control limits and multiply those by 1000 instead of performing all of the calculations of the dpbo chart? The answer is simple the author does not propose the use of both charts in realistic applications, only the use of one chart. Two charts were developed for demonstration and comparison purposes. Thus, if the dpbo chart is to be used, it is not suggested to be used in addition to the dpmo chart, but to be used instead of the dpmo chart.
The benefit from the use of the dpbo chart will ultimately become evident as processes with large opportunities for defect counts reach high quality (i.e., low defect) levels that are measured in terms of ppb, as opposed to ppm
Initially Published in the SMTA Proceedings
Comments
Aug 29, 2012
I really appreciate the articles you share with us. However, I'm baffled by this one.
I read through this article, and it looks like the author took the opportunity to refresh us on the math for control charts (thank you) just to get credit for saying that we should use "DPBO" instead of "DPMO".
Use of significant figures (eight decimal places for numbers that have 2 significant digits) and grammatical issues with this paper notwithstanding, the example data set provided did not even support use of DPMO (yes, it came from a previous paper; not actually helpful).
From my understanding of this paper, it does not make any meaningful contribution to our industry. Am I'm missing something here?
Drew Kidder, Eurotech Inc., USA
|
