‘Them’ are my ‘babes’, my algorithms, my little friends inspired from another world.
No, not from the three kinds of antibiotics I originated as an idea to cure ulcers.
‘US’ is a culture that prefers minimal rule sets, being drawn into a global economy
with countries that have a basis of culture approximated by maximal rule sets.
To ‘us’, freedom and independence and state law versus big government make sense.
To ‘them’ our system sells products designed within the context of a small window of
technology, functionality, rationality and Quality.
What we need is more thought about how materials can be put into arrangements and
sold as a component in spite of the fact that they evolve from different sources and
controls. As energy prices rise and competition for energy increases, we will see more
changes away from ‘Americanization’ and toward replacement with much more
pervasive and invasive standards for the sale and distribution of products of good quality.
Already Wall Mart stocks 80 percent of it’s stock from China, and the average employee
makes $15,000 per year, working harder than ever before in terms of hours spent and
opportunity to advance lost. Here where I live in the midwest, firms founded by local
businessmen pledged to achieve maximum hiring of local state residents are managed
exclusively by MBA’s from coastal schools. The founders of these institutions would
roll over in their graves! Our minimal rule sets often fail us, and we have been slow to
catch on. I believe that with enough patience and focus, we can state rules that better
describe our long term plan and goals, and hopefully the rules do not continue to degrade our
own people and workforce as much as has been done in the past. We have been acting
like the drunk that looks for his keys to his car under the lamp post rather than in the
dark, because there is just an easier way to search under the light.
We must learn to look in the dark, to focus on function and it’s goals. It may be too
late for some of us who had expected better, but it is never too late to ask for change.
In this case, we need change from more exhaustive studies of subject matter and less
governance from pledgelings of the common social good. Competition has long since
failed to provide what safeguards against poor design it once provided, if it ever did at all.
There are other examples, equally complex, of tragic miscalculation in science.
Two of them that I have had experience with (and spent countless hours examining are:
1) recurrence plots and so called recurrence ‘quantification’
2) standard practices for geometric tolerances in shop math
The interesting question is why these bad ideas hang on so persistently, shamelessly.
Anyway, I will make this page as brief and turs as possible as this web site is devoted
to thermal coatings. Yet, these other two math problems show similar lack of attention.
Cross Plotting A normal distribution (Kqrd.exe available on the internet free)
Here is an example of the analysis taken light years ahead of today technically.
It is showing chaotic noise in the upper high part of the data, probably a result
of unbalanced rotor of the machine that may have made this. A lobe or an out
of round condition. My statistics wares shows this distribution has an entropy
a little tighter than normal and that adding ten percent pure noise brings it back
into normality exactly, with a stationarity of that pseudo-process of 18 percent of
coefficient of variation. The vertical structures you see are stationarity in the plot.
It helps do lot identification and traceability using coatings and heat signatures.
Also you may note that the sine wave extracted from the data has a little more
of a point on the nipple on top than on the bottom. This is the case as there is
skew to the high side and too many outliers on the high side, determined with
histogram software I have written. This screen is analysis that is based on fast
Fourier used in math co-processors for computers. (Pronounced FOO-Y AAY)
The FFT is on the bottom left. The big noise like distribution plot is the raw data
and the sine wave is a procedure I invented or designed that yields this advanced
visual representation of data that enhances the output of my histogram programs.
This data comes from the burning of coal in exact measures, done by the British
Statistical Society. The question was asked whether nature tends to give normally
distributed distributions. The answer is that it does not. The model of a normal
distribution is useful, but misleading because the value of using it to parametize
data is in it’s ability to give fine metrics like differential entropy and stationarity.
These advanced techniques are able to show a picture of all patterns in data, and
used in astronomy to look at signals. Let’s hope to find one some day from E.T’s
or EBU’s (extraterrestrial biological units). This technique is limited to larger data
sets like 200 readings, it’s not seen in shop math. It has been used in nuclear power
and research papers have said that it can show the unset of runaway nuclear reactor
activity sooner than other method (older methods). Pattern recognition is another
wave of computer power that needs to be technology transfer empowered so that
the public can move away from what broad activity exists that borders on hoaxes
based on accepted faulty math and underlying assumptions about our physical world.
2) Quality Standards and hoax ‘geometric’ standards (Qs-9000 and others)
A number of years ago, when I was in the coordinate measurement contracting
business, I came across this article about how tolerances for cylindrical shapes
could be done in a new way. It was written by a professor, but he did not have
time or computer algorithms developed to write a program to prove his point.
I on the other had, had already developed a new form of spline algorithm based
on Bezier (pronounced baysiah) spline techniques, and wrote a program to do it.
I also was re-writing programs for coordinate measurement machines at the time,
and offering them to the authors of the software free on condition that they would
put them into their softwares and replace the faulty sub-program (center of circles
calculations). For the reader who has not used a coordinate measurement machine,
all most all of them back in the 1980’s had very bad center of circle sub-programs.
One algorithm was so bad it gave up 21 errors on 10 circles taken. I was then able to
persuade Better Business Bureau to intervene in that instance, after authors refused
to change their method. This resistance to better technology is common. It is better
described in the Berkeley California article, the link is on the hoaxes button page.
I have found the only reason to use these exotic design callouts is to confuse the enemy.
I have gone on and developed spreadsheets for commercial building bay design and steel
specification callouts, but though they are spreadsheets they don’t fit well in this website.
You should make inquiry by email if you need to design your own barn or steel frame
building and aren’t comfortable with $30,000 minimum for a garage and similar pricing.
I will scan that article and include it here because it is an important look into my past
history of making efforts to persuade the science and technology sector that change is
needed. You may want to see the software output and spline algorithms, but I will save
space and your time and leave you with the article that I read that started all my work.
Education
Let’s talk about ‘size’..
by Herb Voelcker, Cornell University
Size is a
nice intuitive notion that everyone uses. It’s a safe notion when used
relatively, hut can he ambiguous when small distinctions on absolute values are
needed.
The Y14.5 dimensioning and
tolerancing standard defines size narrowly, through features of size (FOS).
These are, “One cylindrical or spherical surface, or a set of two
<opposing> plane parallel surfaces, each of which is associated with a
size dimension.” A FOS can be external or internal (for example, a shaft or a
hole), but let’s talk here only about solid cylinders.
Figure 1
distinguishes between a nominal FOS, which is an ideal specified by a designer,
and an actual or physical FOS. A nominal FOS carries
two limits of size
termed MMC (Maximum
Material Condition)
and LMC (Least
Material Condition),
which in Figure 1 are
1.1 and 0.8 respectively.
Clearly the
limits of size should
control the variation in actual size, but the exact nature of the control
depends on how size is defined.
Traditional Notions Of Size
Actual (measured) size is determined traditionally by two-point caliper
measurements. The circular cross-section in Figure 2 would yield a (small)
range of size values if rotated under the jaws of a caliper; these values, and
similar values from other cross-sections, should all lie within the MMC, LMC
limits of size. But note that all two-point measurements on a Reuleaux section
yield very similar values, as do all two-point measurements on lines through
the center of a Fourier section... yet neither is circular.
Figure 2: (Can two point
measurements distinguish these sections?)
The
issue is even murkier when size is mixed with form, as in American (but not
ISO) tolerancing practice. Rule I of Yl4.5 requires that our cylinder lie on or
within a perfect cylinder of MMC diameter, as in Figure 3, and that the LMC
limit be met on all cross-sections. But how can one check all cross-sections,
and how is ‘cross-section’ defined for imperfect cylinders with ambiguous axes?
(
Figuree 3: The Y14.5
interpretation of size under Rule 1
In summary, definitions of
size based on two-point measurements may be intuitively appealing, but they
carry ambiguities that are very difficult to resolve. So — let’s
look at a different approach.
The Offset-Zone Approach To Size
Requicha provided a rigorous definition for size through containment zones
defined by offsets. Our cylinder is in-tolerance if its
boundary lies in
the zone shown in
Figure 4. Note that
the MMC boundary
honors Y14.5’s
Rule 1 (perfect
form at MMC),
but the perfect-
form LMC boundary
is more stringent
than Y14.5 or
ISO requires; it
could not be met, for example, by the Y14.5-acceptable ‘bent cylinder’ shown in
Figure
3. Thus Requicha’s definition was
rejected, but its primary mechanism a containment
zone
— is used in the latest definition of size.
The Y14.5.1 Definition
Y 14.5.1 M- 1994, “Mathematical Definition of Dimensioning
and Tolerancing Principles”, is the new ‘mathematical companion’ for the latest
edition
of Y l4.5(YI 4.5M- 1994). Figure
5 applies Y14.5.l’s new definition of size to our example.
(END OF PAGE 1)
• Roll a
solid ball of MMC diameter along the ‘spine’ (line or curve segment) labeled
SMuC to generate the solid RMMC.
• Roll a ball of LMC diameter
along the spine SLMC to generate RLMC.
• Construct zone Z as the set difference of the two R-solids, as shown in
Figure 5.
• Our cylinder ‘G’ is in-spec if it lies within Z.
MC LMCMccG
Figure 5: Conformance to the Y145 I definition of
size
The spines are the key to this
definition. SMMC is a line segment under Rule 1 ... but
SLMC (and SMM( if Rule 1 does not
apply) can be any curve segment that satisfies
some mathematical conditions. Thus, if
you can produce spines that show G to be in
-spec, you’re done; if not, you must keep searching for better spines unless
you can prove
(somehow) that no such spines exist.
Remarks
The new size definition may be a precursor of things to come, as tolerancing is
‘mathematized.’ The definition is mathematically elegant, and it seems to
capture the
physical aspect of size that Y14.5 gropes for in prose. But it presents major
challenges to
dimensional metrologists, and may not be viable in practice.
We need to think more deeply about size. For starters: do we really need the
notion of
size? (Whenever I ask that question,
someone reminds me that tool cribs the world over
are organized on the basis of size, so
there must be something in it . .
Reinventing
Education To Grow Future Talent
By Professor John W. Sinn,
Bowling Green State University
We simply must figure out how
to provide technical leadership to rebuild our
manufacturing base — including
the quality and metro logy fields —
for the future. In
short, we must reinvent
education. But the challenges abound!
One of the key questions is the issue of how to grow talent and leadership
within our
organizations. Recognize that
this is not a separate issue (or at least it should not be) from
the question of what our
public schools should be doing. How do we design teaching and
learning functions within our
organizations —and in our schools —
and communities to
best facilitate individual
growth for people? How can this be done within the context of
maintaining our current
responsibilities and commitments, and in a cost-effective
manner? In other words, if
there could be an articulated educational system at all levels
and targeting all persons,
what would it look like, and how would it help us grow talent,
and leaders, for the future?
Perhaps part of the answer will come by better understanding
the overall current
situation.
First, it is vitally important to remember that students generally (hopefully)
do turn into
workers and employees.
Employees, as responsible citizens, must continue learning,
assuming they really are
serious about continuous improvement, as individuals,
organizationally and
culturally. The point is, we are all students, at every level, and we all
must participate in the
teaching and learning process. Thus, we all must be part of the
solution. The opportunity for
improvement is tremendous!
Second, few would argue that one of the key drivers in the total mix of change
and
competitiveness is
technology. Generally, if we wish to “reinvent education” for the
future, we all must study and
learn more about our technological culture, history, and
evolution, impacts and
implications, the change process, technology assessment,
technology transfer and so
on. Focused more technically, we all must better understand
how to solve technical
problems, the design process, quality functions, and our
manufacturing processes as
they relate to our customers upstream and downstream.
Finally, technology means
much more than simply computers, yet it includes the
computer with its broad
implications for communication, documentation, presentation,
computation, data
acquisition, and process control. The point is, we must not make the
mistake of thinking that
technology equals the computer — it is much, much, bigger than
any one device.
Third, we must pause and carefully reflect on how we learned what we value most
— and
what we really know — and
where and how we learned it. For most of us, sadly perhaps,
we probably did not come to
know what we know and value in play, and elsewhere.
Significantly, and as related
to our broader discussion, we have often learned through
technical projects and
problems to be solved or undertaken through our work. Moreover,
the old adage, “you never know
it as well as when you have to teach it,” along with
applications of theory and
information studied through real world experiences, would
seem to equal knowledge.
Clearly, the way we teach and learn to communicate, work
together, solve problems, document
the findings and solutions for future reference, are all
part of the missing link— and
solution — for the future of metrology education.+
I
Acknowledgments
This article is ha,ced on a
paper by K. Suresh
and
HB, Voelcker published in the December, 1994, issue of ASME Manufacturing
Review. The paper
summarizes what is known about assessing conformance to the new
definition.
HB.
Voelcker holds the Charles W. Lake Chair in the Sibley School of Mechanical and
Aerospace Engineering at
Cornell
University, Ithaca, NY. He can be reached
by E-mail at voelcker@cornell.edu, and by
telephone at 607-255-9654.
mfg. The Brown 8 Sharpe Publication of Precision Manufacturing 41
Back To the Thermcoat.com Home Page