Documente Academic
Documente Profesional
Documente Cultură
• Heights • Pressure
• Weights • Roughness
• Lengths • Strength
• Widths • Conductivity
• Diameters • Loudness
• Wattage • Speed
• Horsepower • Torque
• Miles per Gallon • Etc. etc. etc.
Eli Whitney in 1798
• Won a U.S. Military contract to supply 10,000 guns
• Reduced variation and created interchangeable parts for
assembly and service by:
– Installing powered factory machinery
– Using specialized fixtures, tools, jigs, templates, and end-stops
– Creating drawings, routings, operations & training
Manufacturing in the 21st Century
• International competition to provide defect-free products
at competitive cost
• Reducing variation and providing interchangeable parts
for assembly and service by:
– Using machine tools
– Using specialized fixtures, tools, jigs, templates, and end-stops
– Using drawings, routings, operations & training
Graphing the tolerance and a measurement
.512 .513 .514 .515 . 516 .517 .518 .519 .520 .521 .522 .523 .524 .525 .526 .527 .528
It’s useful to see the tolerance and the part measurement on a graph.
Suppose that:
Graphing the tolerance and a measurement
Specification
Limit MIN
.512 .513 .514 .515 . 516 .517 .518 .519 .520 .521 .522 .523 .524 .525 .526 .527 .528
It’s useful to see the tolerance and the part measurement on a graph.
Suppose that:
--the tolerance is .515”
Graphing the tolerance and a measurement
Specification Specification
Limit MIN Limit MAX
.512 .513 .514 .515 . 516 .517 .518 .519 .520 .521 .522 .523 .524 .525 .526 .527 .528
It’s useful to see the tolerance and the part measurement on a graph.
Suppose that:
--the tolerance is .515” to .525”
Graphing the tolerance and a measurement
Specification X Specification
Limit MIN Limit MAX
.512 .513 .514 .515 . 516 .517 .518 .519 .520 .521 .522 .523 .524 .525 .526 .527 .528
It’s useful to see the tolerance and the part measurement on a graph.
Suppose that:
--the tolerance is .515” to .525”
--and an individual part is measured at .520”.
Graphing the tolerance and measurements
X
X
X
Specification Specification
Limit MIN
X Limit MAX
.512 .513 .514 .515 . 516 .517 .518 .519 .520 .521 .522 .523 .524 .525 .526 .527 .528
XX
XXX
XXXXX
Specification XXXXXXX Specification
Limit MIN Limit MAX
.512 .513 .514 .515 . 516 .517 .518 .519 .520 .521 .522 .523 .524 .525 .526 .527 .528
(*A number of common industrial measurements, such as flatness and straightness, do NOT tend
to distribute in a bell shape; their proper statistical analysis is performed using models other than
the bell curve.)
What is a “standard deviation”?
XX XX
Typical distance XXX XXX Typical distance
from the center: -1 XXXX XXXX from the center: +1
standard deviation XXXX XXXX standard deviation
XXXXX XXXXX
XXXXX XXXXX
XXXXXX XXXXXX
XXXXXX XXXXXX
XXXXXXX XXXXXXX
XXXXXXXX XXXXXXXX
XXXXXXXXX XXXXXXXXX
XXXXXXXXXXX XXXXXXXXXXX
-6 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5 +6
(NOTE: We usually show the bell from –3 to +3 to make it easier to draw, but in
concept, the “tails” of the bell get very thin and go on forever.)
What is Cpk? It is a measure of how well
a process is within a specification.
A
Cpk =
A divided by
B B
Specification Specification
Limit Limit
Cpk = A divided by B
A = Distance from process mean to closest spec limit
B = 3 Standard Deviations (also called “3 Sigma”)
Specification Specification
Limit Limit
…a LARGER “A”
…and a SMALLER “B”
…means BETTER “Process Capability”
An Analogy
A
Cpk =
A divided by
B B
Specification Specification
Limit Limit
Analogy:
The bell curve is your automobile.
The spec limits are the edges of your garage door.
If A = B, you are hitting the frame of your garage door with your car.
How can we make Cpk (A divided by B) better?
A
Cpk =
A divided by
B B
Specification Specification
Limit Limit
Specification Specification
Limit Mean Limit
Note that when Cpk = 2, our process mean is 6 standard deviations from
the nearest spec, so we say it has “6 Sigma Capability.”
What does a problem Cpk look like?
A
This Cpk is just
B slightly greater
than 1. Not good!
Specification Specification
Limit Limit
Specification Specification
Limit Limit
A significant part of the “tail” is hanging out beyond the spec limits.
This process is producing scrap, rework, and customer rejects.
Notice that if distance “A” approaches zero…
…the Cpk would approach zero, and…
…the process would become 50% defective!
Free software is available to draw a histogram
and calculate average, standard deviation, and Cpk.
Specification Specification
Limit Limit
The following few slides depict some original messages from that time.
Robust Design – part of the original Six Sigma
Known
Existing
Process
The new design above has tolerances set “tight” to a known existing process, while
the one below has tolerances that allow “six sigma capability”.
Products have thousands of tolerances. They result from choices about shapes,
thicknesses, grades of materials, and grades of components. “Robust design” is
NOT about permitting “sloppiness.” It requires very smart engineering to allow
ample tolerances AND achieve satisfactory function.
Known
Existing
Process
CAUTION:
Known
Existing
Process
The process above varies so much that it “fills” the design tolerance. The different
process below has good repeatability for “six sigma capability”.
It’s a false-economy to choose an allegedly lower-cost process that “uses up” all
tolerance. The resulting scrap, rework, rejections, recalls, damage to reputation,
crisis communications, and fire-fighting cancel out the alleged economy. “Robust
Process” requires skillful insight to choose ways to make defect-free product at the
lowest real cost.
New Process
choice “Y”
The 3
Requisites
Of Self-Control
Ability to Process
Measure Results Capability and
Ability to Control
Summary:
• To call a process “capable” typically requires at least a
Cpk of 1.33 (+ and - 4 standard deviations within
tolerance)
• Many customers desire a Cpk of 2.0 (+ and - 6 standard
deviations within tolerance)
• Organizations need:
1. Feasible designs
2. Capable processes
3. Process self-control
Conclusion:
Process Capability:
Yes: No:
too wide
Yes: No:
potentially capable
if re-centered
Yes: No:
potentially capable
if re-centered
Review Question 1
B
The defect PPM is
the area outside
spec limits
Specification Specification
Limit Limit
Specification Specification
Limit Limit
The “6 Sigma Philosophy” includes the premise that real-world processes move around
to some extent, and produce more defects than a static process. As an arbitrary
convention, this is represented as an “unfavorable shift” of 1.5 sigma in Parts Per
Million tables for Six Sigma programs. The intention is to plan conservatively.
(This means that the “PPM vs. Sigma” charts published for “6-Sigma Programs” show higher defect rates than the
similar but traditional “Z-tables” in statistical textbooks.)
The following page is a table showing the relationships
among the following:
• Cpk,
• “How Many Sigma Capability,”
• Parts Per Million according to traditional statistical tables
• Parts Per Million taking into account the “6-Sigma
Philosophy” of an unfavorable shift in the mean of 1.5
Sigma
Cpk, PPM, and "Six Sigma"
PPM of the Bell
Cpk "How Many PPM of the Bell PPM of the Bell
Curve
(Defined as Within Spec Out of Spec
Sigma Curve Curve
distance from (Process Perfectly The column AT with Six-Sigma
Capability?" Out of Spec Out of Spec
process mean to Centered, LEFT equates to 1 Philosophy
Distance of (Process Perfectly (Process Not
the nearest spec, Both Tails defective out of how of 1.5 Standard
Process Mean to Centered, Centered,
divided by 3 Considered) many total? Deviation Penalty
Spec Limit in Both Tails Only One Tail
Standard Good Units Per for Anticipated
Standard Deviations Considered) Considered)
Deviations) Million Unfavorable
Process Mean Drift
0 0 0 1,000,000 500,000 2