Learning Statistical Analysis and Statistical Tools

I don’t have to tell anyone that times are tough!  When they’re tough like this, you don’t like to waste what money you have earned—it came dearly! That’s when design choices come to mind. Any bad choices go to the bottom line; so what are some of the tools to help improve decision making? Statistical choices and probability come to mind. Design trade-offs have always been difficult to predict in printed circuits. Those uncontrolled factors can always creep into our decision making process, but there is help.

All decision making is a feedback loop of sorts. But the element that I want to focus on is the “ANALYSIS” block, or more precisely, the ‘human decisions’ which make up analysis.

Human Decision Making

The first link in Analysis is the Human Link. The high-level objectives (goals) are to:

• Complete the Design on Time
• Reduce Variations
• Increase Manufacturing First Pass Yields
• Reduce Repair and Rework
• Improve Quality and Reliability
• Improve Workmanship

The Decision-Making Tools and Methods that a person may have to work with include:

• Pareto Charts
• Cause & Effect Diagrams
• Multi-Vary
• Design of Experiment
• Process Optimization
• Control Charts
• Process Capability Indices (Cp, Cpk)
• PPM
• Six Sigma

Of importance to the engineer are the statistical tools, as seen in Figure 1:

FIGURE 1.  Selecting the Right Statistical Tool-Four main techniques are basic to selecting the right statistical tool for problem analysis

Traditionally, statistical tools have been rather cumbersome, expensive and difficult to learn. I have good news for you. You can now get good statistical analysis training on the web—and at a bargain price: FREE.

The Engineering Statistics Handbook from the National Institute of Standards & Technology

I have just discovered this great statistics handbook on the Web. I was looking for Weibel Reliability Plot information and the SEMETECH/NIST Statistics Handbook popped up. After playing with it awhile, I discovered it was designed for just us working engineers. The organization of the handbook follows the statistical tools that a working engineer would need to ferret out a problem, measure the extent of the problem, look for root causes, uncover how many factors are involved in the problem, postulate a solution, verify the solution and then monitor the process to be sure the problem is gone and does not reappear. These eight (8) stages are the main chapters of the Statistics Handbook (outlined below) and can be seen in Figure 2, the first page of the Handbook:

1. Exploratory Data Analysis

   • Introduction
   • Assumptions
   • Techniques
   • Case Studies

2. Measurement Process Characterization

   • Characterization
   • Control
   • Calibration
   • Gauge R & R Studies
   • Uncertainty Analysis
   • Case Studies

3. Production Process Characterization

   • Introduction
   • Assumptions
   • Data Collection
   • Data Analysis
   • Case Studies

4. Process Modeling

   • Introduction
   • Assumptions
   • Data Collection
   • Interpretation & Use
   • Case Studies

5. Process Improvement

   • Introduction
   • Assumption
   • Choosing an Experimental Design
   • Analysis of DOE Data
   • Advanced Topics
   • Case Studies

6. Process or Product Monitoring and Control

   • Introduction
   • Test Product for Acceptability
   • Univariate and Multivariate Control Charts
   • Time Series Models
   • Tutorials
   • Case Studies

7. Product or Process Comparisons

   • Introduction
   • Comparisons: One Process
   • Comparisons: Two Processes
   • Comparisons: Three+ Processes

8. Assessing Product Reliability

   • Introduction
   • Assumptions/Prerequisites
   • Reliability Data Collection
   • Reliability Data Analysis

The usefulness of this statistics handbook comes from the experimental data sets supplied in it (case studies).  As you read the handbook, it encourages you to make statistical comparisons with its statistical tools. It does this by supplying a complete statistical software program for you to use! When used with the data sets supplied, it can then coach you through the interpretation of the results. You can then substitute your own data and look at the results. This “First we run the demo – then we run your problem” is a very effective way to coach a person through the statistical tools.

FIGURE 2.  The NIST / Sematech Engineering Statistics Handbook available over the Internet.

To get to the NIST/SEMATECH e-Handbook of Statistical Methods, go to: https://www.itl.nist.gov/div898/handbook/. The statistics software, Dataplot^®, is accessed by selecting the ‘TOOLS & AIDS’ button in the left border of the Web page, then selecting the ’10 STATISTICAL SOFTWARE’ option on the next Web page and finally, the ‘DOWNLOAD DATAPLOT’ on the software page.

NIST has prepared versions of the software for nearly all computers; Windows, NT, UNIX, Linus, MAC OS, etc.  The manual is a bit large (over 3,800 pages), and is hyper-text oriented, so it isn’t intended to be printed, but the download time is worth it.

This is a good opportunity to learn statistical analysis or how to use statistical tools, and you can’t beat the price!  After all, you paid for it through your taxes. It is oriented to be a ‘self-learning’ manual. I advise you to at least download it right-away, as I don’t know how long the government will be providing it! You never know when you might need it.

Look no further than Altium Designer^® for the best PCB design software. Have more questions about statistical tools, statistical method, or PCB design? Call an expert at Altium.