Allowing Defective Chips to Go to Market
Author(s):
Jeremy Hanzlik
by Jeremy Hanzlik
Texas A & M University College
Station, TX 77843
Introduction
A production line engineer, Shane, checks every chip for
quality control (QC). His workers find errors approximately
every 150 chips. Either the defective chips must be sent back
for repair or they must be axed (thrown away). The manager,
Rob, has mandated that workers must axe all defective chips.
Rob walks over to Shane's line and declares, "Why some lines
sink more dollars into a chip that's failed, I don't
understand. We only make 25 cents off of each chip anyway!
Spending an additional $2.00 per chip will only be
more money down the drain. Shane, in our line of work we can't
afford to flush money down the toilet."
The following afternoon, Rob calls a meeting in his office.
Rob informs Shane, that Shane's line is axing too many chips.
"One chip every hundred and fifty is unacceptable! This is
becoming a substantial cost to the company. I believe that it
would be more beneficial to allow defective chips to go out the
door." Shane asks, "What about the defective chips? Won't
customers complain?" Rob replies, "Yeah, yeah, but that's not
your problem, the company has a return department that will
replace them as customers complain." Rob further estimates that
allowing defective chips on the market will yield a $416,000
profit for the company.
Facts:
- The line produces 100,000 chips per year.
- Every chip is purchased.
- Chips cost about $9.00 to produce.
- Chip testing costs about $4.00 per chip.
- Chip repair (manpower and material) is about $2.00 per
chip.
- This repair cost includes re-testing.
- Profit per chip is $0.25 after testing.
- There are fifteen full-time employees working under
Shane.
- Two part-time employees work under Shane's
supervision.
- Shane's manager, Rob, has been with the company for about
7 years.
- Shane has been working under the same manager for several
years and has had relatively good relations with Rob.
Additional information regarding the Shane's line:
The engineer's line consists of the final inspection between
the bond wires, which attach the chips to the prongs and spot
plates (the prongs that protrude from the final product), just
before the chips are encased in molding compound for final
packaging. You may assume that all defects are caused by faulty
bond wire attachment and not by any problem with the chip
itself, because the chips were tested in the preceding phase
before the bond wires were attached.
Back to Top
Numerical and/or Design Problem(s)
- What percent of the chips may fail if Xanthum Inc. Orders
15,000 chips from Shane's production line?
- Do you believe this is an acceptable failure rate? From
the perspective of Xanthum? From the perspective of the
manufacturer? Why or why not?
- If Shane's line produces 100,000 chips per year, how much
will it cost to:
- Test and repair each defective chip?
- Test all chips and throw away the defective
chips?
- Not to test any chips and to replace customers chips
as needed?
- Is Rob's estimate reasonable? What about his assertion
that it is cheaper to axe the chips?
- An additional problem could be added by incorporating
combinatorial mathematics to compute time required to test
each chip.
Back to Top
Questions on Ethics and Professionalism
- What issues are involved in following Rob's
recommendation?
- Is it acceptable to follow Rob's suggested course of
action (based on your calculations above)?
- If Shane has a differing opinion, how could he present
his case to Rob?
Back to Top
Additional Scenario
(To be introduced in lecture after the students have
completed the work above on their own.)
The chips ordered by Xanthum, Inc. are to be placed in
aircraft navigation units. Shane's boss still believes that the
failures are inconsequential. Rob claims, "They always have
back-up navigation systems anyway. Besides they fail less than
one percent of the time! You should know that -- calculate the
percent of chips that will fail."
- What flaws can be found (based on your previous
calculations and present observations) with the superior's
argument? Hint: look at the logic used in the quote.
- Perform a cost/benefit analysis based on the above data.
How much will it cost the company in litigation, etc? Make
any necessary assumptions such as dollar values for your
calculations (as long as your assumptions are not in direct
conflict with the stated facts above).
- How does this scenario influence your response to
question 2 from part III?
- How can Shane constructively present an argument against
his superior's opinion?
Back to Top
Solutions for Numerical Problems
Part I
- 1/150 chips will not work so: 1/150 = 0.67% will not
work.
- This question in neither right nor wrong. It is meant to
provoke thought for later ethical issues.
-
- Test all: 100,000 x $4.00 = $400,000 (testing cost)
Repair cost: 667 bad chips x $2.00 repair cost = $
1,334
Profit on repaired chips: 667 bad chips x -$1.75
profit loss = -$1,167.25
Profit on good chips: (100,000 - 667) x $0.25 =
$24,833.25
Net profit: $24,833.25 - $1,167.25 = $23,666
- Test all: 100,000 chips x $4.00 profit = $400,000
(testing cost)
Ax cost: $0.00
profit on axed chips: 667 bad chips x ($9.00 + $4.00
- $0.25) = -$8,504.25
Profit on good chips: (100,000 - 667) x $0.25 =
$24,833.25
Net profit: $24,833.25 - $8,504.25 = $16,329
- There are two polar scenarios based on different
student assumptions (and equally correct answers based
upon differing assumptions as to the number of chips
returned):
- "Best Case" - no returns of defective chips not
test any chips: 100,000 x $4.25 = $425,000
- "Worst Case" - all defective chips are returned
to find the number of chips which actually generate
profit:
- 100,000 - 667 returns = 99,333 - 667
replacements = 98,666 chips to generate
dollars
- Profit from "satisfied" customers: 98,666 x
$4.25 = $419,330.5
- Original profit on the 667 returned chips:
667 x $4.25 = $2,834.75
- Profit from replacement chips: 667 x -$9.00 =
-$6,003
- Net profit: $419,330.5 + $2,834.75 -$6,003 =
416,162.25
- Here the student should see that the answer is clearly
NO. The manager, Rob, has based his calculation on an
improbable event that no chips will be returned.
Consequently, his estimate for profit is too high. The
manager is also wrong to advocate axing chips; it is
profitable if the chips can be repaired, but the line loses
money if the chips are axed.
Back to Top
Solutions to Ethics and Professionalism
Questions
- One major issue in the narrative above is deceiving the
public, because the public is sold chips with no precautions
taken to ensure their correct assembly. A second issue
involved here could be public safety: what if the part is
used to build a critical device such as a navigation computer
for missiles or airplanes?
- This answer depends is on the assumption made by the
student. If the student perceives this issue as a safety
issue, then it would clearly not be acceptable to follow the
superior's suggested course of action. Moreover, this answer
hinges on whether or not the student perceives this as a
management decision or an engineering decision. Again, this
is decided by the assessment made by the student: Is this a
safety issue? If the welfare of the public will not be
compromised then this is decisively an economic decision
which falls in the domain of a management. If however, the
assumption is made that the chips involve public safety, this
falls in the domain of engineering.
- To present a differing opinion to a manager, it may be
necessary to compromise. In this situation, economics can be
powerful aid to the engineer. As demonstrated by the above
calculations, the manager's estimates were for a best case
scenario. Therefore, it can be asserted that actual profits
should be less, especially if there is a return cost involved
with the defective chips in terms of personnel, paper trails,
and loss of reputation. It can be further asserted that by
repairing the chips, the loss is significantly minimized in
comparison to axing chips: the line is still profitable and
the company's reputation for quality could help increase the
overall corporate value. So, one creative middle way would be
for the engineer to suggest the repair of all chips, and
present his calculations to his boss in doing so. This solves
the problem of the line losing money, turns a reasonable
profit (although nowhere near what could be made via the
worst case scenario for the first year), and establishes a
tradition of quality within the company.
Back to Top
Reference
- Jaegar, Richard C. Introduction to Microelectronic
Fabrication. Addison-Wesley Publishing Co., pp.
153-172 (Fete, 1991).
Back to Top
These problems were originally developed as part of an
NSF-funded project to create numerical problems that raise
ethical issues for use in engineering and other course
assignments. The problems presented here have been edited
slightly for clarity.