Product Development Systems & Solutions Inc.-click to homepage
News from PDSS Inc.
"Leading the Future in Product Development" 
November 2014- Vol. 7, Issue 10
In This Issue
QG18: Adjustability (Metric 3 of 7)
QG 19: Independence, Interactivity & Stat.Sig.(Metric 4 of 7)
Links to Prior CPD&M Quick Guide Newsletter Issues
Metrics 3 and 4 of the "Big 7" are introduced in this issue. They are:
3) Adjustability and
4) Independence, Interactivity & Statistical Significance.
We will be back in January with the next Metrics in the "Big 7" series!
  
Have a happy Thanksgiving!  
-Carol
QG18: Adjustability (Metric 3 of 7)

With Stability under control, we can move on to explore X's that are effective at adjusting the mean of the NUD functional responses over ranges of application. This can define how useful a new technology will be across a range of potential applications, as well as the knowledge to make trade-offs between X's that can affect both the mean and the standard deviation.

 

Main Effects Plot 
Main Effects Plot

 

Surface Plot
Surface Plot

Minitab or JMP can produce a number of graphs and regression equations to characterize your adjustment parameters across all the sub-level designs, as well as the production process control parameters.

 

Adjustability is essential for shifting a mean response on top of the required target. This is how Cpk is adjusted to equal Cp. In this context, we adjust the difference between the mean and target to zero! High integrity (known "gain") adjustment parameters are crucial to the design of control systems within a product or manufacturing process. Sometimes we will have to measure undesired noises and compensate for their unwanted effects on the function by using adjustment parameters as a form of robust design. If a design has poor adjustability of desired ranges of performance, we may have to re-design the concept to improve its adjustability.

 

The table below illustrates a tiered structure of % Knowledge-In-Hand to rate the risk a NUD Function or Characteristic is carrying due to the current level of adjustability.

 

 

How to Assign % Knowledge-in-Hand for 3) ADJUSTIBILITY

100%

I can produce clear regression analysis results, quantify the Correlation Coefficient (R2), assess Residuals & Fits data, and identify statistically significant terms (using ANOVA) for the Y = f(X) model that describes the adjustable range of the mean of the function, as controlled by one or more X's, under both nominal and stressful conditions.

80%

I can routinely characterize the mean and standard deviation of adjusted performance of the measured function. I can characterize the Coefficient of Variation (COV = std. deviation / mean) across the adjustable range of the function under nominal conditions.

60%

I have extended the range of adjusted performance and have enough sample data (>30) at each point to characterize the stability of Y's mean and standard deviation performance across this adjustable range.

40%

I have simulated the performance over appropriate ranges of adjustment. I have physically adjusted the function over a very limited range of performance for limited sample runs for limited periods of time. I need to conduct further development to expand the range of Y's performance.

20%

I have one or more ideas on how to adjust the function.

0%

I do not have the ability to adjust the mean performance of the function over a range of desired outputs of my measured Y yet.

 

QG19: Independence, Interactivity & Stat.Sig. (Metric 4 of 7)

Knowing which X's are independent of one another or interactive with one another is crucial for CPD&M. Interactivity between X variables is just like a third parameter in an empirical equation derived from a designed experiment (Y= C0 + C1X1 + C2X2 + C3 X1*X2 + error). Statistical significance of independent and interactive X's is how we know what parameters are real, have strong effects, and should be included in our empirical model. These empirical models help us improve and refine our analytical models. All First Principles models are in error to some extent, and this kind of data is essential to increasing their validity and utility for repeated applications, trade-off studies, and optimization.

 

What if your functions with measured response Y's and their controlling X parameters possess interactivity and you don't know it? You will likely discover sensitivities and intermittent behaviors when enough variation affects the design causes the "internal sensitivity" imbedded in the interactive component of your control factors to awaken and thus change your functional performance. This parametric behavior is not usually developed or understood in technology and product development. Without it, corrective action is forced down-stream, where the sustaining engineering team must "investigate" the unforeseen trouble. Better that this statistical investigation and analysis is done on a preventive basis during development projects.

 

The table below illustrates a tiered structure of % Knowledge-In-Hand to rate the risk a NUD Function or Characteristic is carrying due to the current level of independence, interactivity and statistical significance of control parameter effects on the measured Y across the function.

 

How to Assign % Knowledge-in-Hand for 4) Independence & Interactivity

100%

I have clear ANOVA results. I can quantify F Ratios and p values that describe the statistical significance, independence, and/or interactivity between the X's that were intentionally varied within a DOE under controlled, nominal conditions. I know which X's to include for Model Building DOE's in the future.

80%

I have conducted Screening DOE's by changing the X's and measuring Y's changes. I can separate signal from noise in my experimental results using ANOVA.

60%

I have designed an orthogonal array-based factorial experiment to exercise intentional changes in my controllable X's to study their effect on the measured Y's that characterize my functions.

40%

I have a credible plan on how to measure and intentionally change my X's to assess statistical significance, independence and interactivity as they affect the measured Y's of my functions.

20%

I have identified the candidate controllable engineering factors (X's) to intentionally change to assess how they affect the measured Y of the function.

0%

I do not know if my X's are statistically significant, independent or interactive with one another as I make purposeful changes in their set points.

Links to Prior CPD&M Quick Guide Newsletter Issues

There are 24 Critical Parameter Development & Management (CPD&M) Quick Guides being published in installments in this newsletter. Below are links to each of the prior newsletters with CPD&M Quick Guides: 

 

The CPD&M Quick Guide TOC (Nov 2013)

CPD&M QG1&2: Intro & Process (Jan 2014) 

QG3&4: Prioritize Req'ts & Design Guide (Feb 2014)

QG5&6: Functional Diagramming & Functions, Complexity & Risk (Mar 2014)

QG7&8: Fn's, Design Controls, DG O'view & I-O-C Diagrams (Apr 2014)
*Note: there was no May 2014 issue

QG9: Design Failure Modes & Effects Analysis (DFMEA) (June 2014)

QG10&11: Fishbone & Noise Diagrams (July 2014)

QG12&13: Base-line & Robust P Diagrams (Aug 2014)

QG14&15: Robust Design P Diagrams & Big 7 Metrics O'view (Sep 2014)

QG16&17: Measurability & Stability (Metrics 1&2 of 7) (Oct 2014)

 
Is there a topic you'd like us to write about? Have a question? We appreciate your feedback and suggestions! Simply "reply-to" this email. Thank you!
  
Sincerely,
Carol Biesemeyer
Business Manager and Newsletter Editor
Product Development Systems & Solutions Inc.
About PDSS Inc.
Product Development Systems & Solutions (PDSS) Inc.  is a professional services firm dedicated to assisting companies that design and manufacture complex products.  We help our clients accelerate their organic growth and achieve sustainable competitive advantage through functional excellence in product development and product line management.
  
Copyright 2014, PDSS Inc.
Join Our Mailing List!
 
See PDSS Inc.'s Archived E-Newsletters