fOR

America's newest "secret weapon" boosts quality, lowers

costs and beats the Japanese at their own game.
by GERRY KOBE
F
or years Japan has been accused
of doing cheap imi tat ions of US
designs. But nothing coul d be
fur t her from the t r uth.
In fact , the Japanese h ave
been care ful NOT to copy American
desi gn s , opti ng in stead to look at
how the product can be made better
a nd easier . As a res ult , J apanese
products r eflect a high degr ee of sim-
plicity and qual ity, while offering an
inexpensive alternat ive.
The name of t his approach is Design
For Ma nu fact ur e and Assembly or
DFMA. And it' s becoming a way of life
in many Amer ican indust ries. "One of
my cu st ome r s - a truck manufac-
turer - went from 800 parts to 120 on
one assembly, an d qualit y improved
dr amat icall y, " says DFMA exper t
Sandy Munro, pr esident of MTS for
Productivity Inc. "Also, a machine t ool
com pa ny we ' ve worke d wit h went
from 120 parts down t o four!"
Although exist ing automotive exam-
CURRENT
HEADLAMP
ASSEMBLY
pies may be less dramat ic, a good illus-
tr ation might be one-piece door stamp-
ings (Fig. 1), which J apanese aut o-
maker s pioneer ed in the early 1980s.
Then cur rent American designs wer e
using a complicated t hr ee-piece config-
uratio n t hat added cost and qual ity-
robbing variability to the door - a mis-
take t he J apanese resolved. In simple
terms, t he J apanese were using DFl'i1A
principles t o ask t he quest ion: How
can we design complexity (cost) out ,
and quality in?
As simple as t hat quest ion sounds, it
DFA
HEADLAMP
ASSEMBLY
Ford Ranger head/amp benefits from DFMA analysis. Number of parts reduced 22%. assembly cost reduced 52%, mater ials reduced 2%
Reprinted from Automotive Industries, March 1990 issue.
Fig. 1
Credit: Harbour Report
MULTIPIECE
DDORPANELS
has taken a long time for domesti c au-
tumaker s to catch on to t he scen-t of
J apa n's success. And now th at t hey
haUl' caught on, automukers must pn-
aitio n themsel ves to implement wha t
they' ve learned.
The first order of business, then. is
to formulate "clean sheet" product lift,
cycle s th at are compet it ive with t he
Japanese -curr ently four year s.
Chrysler ' s Denn is Pawley, ge ne r al
manager, advanced manufacturing en-
gineering, notes, " .. You're limited as
to what you can do hy maintaining the
same platform. New technologies that
allow for a more efficient manufactur-
ing process cannot be implemented un -
less designed int o the product. "
Pawley' s ins ight illustrates how the
J apanese ca pitalize on thei r quick
t urnaround time to ta ke advantage of
new mater ials, designs and man ufac-
t uring methods . When cost -cutt ing
tech nology is avai lable, t he J apanese
incorporate it quickly. In contrast, cur-
rent domestic pr oduct life cycles run
from eigh t to 10 years for a complete
redesign, often at a cost of being non-
responsive to customer needs.
That' s t he bad news.
The good news is that successful
mel
0001PINEL
QII/JpmCII aoor dnsign pio/1l'f'mrl by
Jilp dllfls e, Offt!l S lowe, cosl , IHSS v<I",iIJ/iJly
IhiJlI IJWII/plix e dos/gn
1H' l\l A is growing rapidly in tlw US
auto industry. In fnct , wlu-n compan-d
to thu comh int-d t' ffor t s of till' Big"
'l'hn-r-, .Japan may -for ont"t' - find it-
self playing" catch-up.
On one end of the scale, Chrysler is
focu sing" its IH'MA efforts on major
operutions : s t a mpi ng and t rim!
chass is /final assembly. Rol a nd
Mueller , manager of manu facturin g
fea s ibility at Chrysl er' s a dvance d
mfg.opcraticns says, "We had a philos-
ophy at one point, and t hat was during
the K-car program. Th e philosop hy
was to maximize the size of t he stamp-
i ngs and minimize t he n umbe r of
part s. We have r eversed on th at ," he
confides. "Now we want to mi ni mize
the number of st amping operat ions."
Chrysler has learned one of the most
important lessons of DF:VlA, and t hat
is to analyze the entire process to de-
termine where the greatest gains can
be made. In this case , their analysis r e-
veal ed t hat t he costs of makin g large
panels - requiring up to nine stamping
operations -negatively offset the sav-
ings of a reduced part count. In con-
trast , body panels for the up coming
LH platform are running at an average
3.9 hits , and will be done in higher effi-
ciency transfer pr esses.
The same pla tform also promi se s
improved t echniques for final assem-
bly. J ohn Hinckley, program manager
for Chrysler ' s LH-LX platfor ms, says
th e new vehicle was designed to elimi-
nate the need for workers to per form
operations over their hea ds . " \Ve de-
cided we wanted to ta ke t he chassis
and buil d it [engine, fr ont suspensi on,
exhaus t system, brakes , etc. l at
ground level. where we can get good
lighting, load parts simply, and get our
tools on it. And we want ed to build it
on a pallet. " Hinckley t hen offers that
t he body and chassis will be assembled
aut omat icall y, usingal l vertical at-
DFMA DESIGN TIPS
SandyMunro, president of MTS for
Productivity, Inc. offers these design
tips :
TEST /'ONMINIMUM
NUMHEH OPPARTS
Do purt s move relative to each
ether?
Must part s be made of dif-
ferent materials?
Would combination uf tbcse parts
prevent assembl y or disassembl y of
other parts'!
lias ser vici ng of the assembly
bf>t'll adversely affected'!
If the answer to all que stions is
no, consider combining t he parts.
GOOD DESIGN PRINCIPLES
Teamwork -the difference be-
tween good and bad designs.
Minimize the number of parts.
Des ign so the assembly process
can be completed in a layered fash-
ion - preferably from above.
Consider easy part handling.
Design mating parts that are easy
to insert and align.
Avoid expensive fastening opera-
tions; i.e. screwing.
Avoid part designs that will cause
tangling with identical parts.
Make part symmetrical to aid in
automatic orientation.
If symmetry can't be achieved, ex-
aggerate asymmetrical features.
Avoid adjustments.
-.i..g. 2 ArtCredit: Harbour Report
Japanpioneeredlap-jOints toreplaceflangedbuttjoints, Designmaintainsjomt quality while al-
lowing forpanel variation
DFMA
tachment s - ef fectively reducing as-
sembly time and improving quality.
No Silve r Bullets?
One of the more promising applica-
tions of DFMA comes in the form of
soft war e progr ams-run on PCs -
that allow engi neers t o do a part-by-
part design for manufacture and as-
sembly analysis. And of the syste ms
available, one- Boothroyd Dew-
hurst - distinguished itself among en-
gineers as being the most compre-
hensive and pov.... erful. Moreover, we
discover ed that for the past sever al
vears , thi s svstem has been named the
;' Number One t echnol ogy" at Ford,
and is current ly spreading through
GM like wildfire.
We contact ed Gary Cowger, man u-
facturing manager at Cadillac and cor-
porate champion of the Boot hroyd
Dewhurst approach. " BD! [Boot hroyd
Dewhur st Inc.] soft ware is ano t her
tool - t hat ' s all , " Cowger says.
" There's no one great big t hing t hat
you go out and learn , and t hat 's par t of
t he problem - people t hi nk t her e' s a
silver bullet. "
Cowger ' s poi nt is well t aken, but
since he pu t on his best poker face
when t he to pic turn ed to BD! - we
t hink he might be sandbagging. You be
the judge.
" It 's magic," boast ed a B-O-C engi-
neer who spoke on condition of ano-
nymity. " It's spre ading like cr azy
around her e!"
" Getting the voice of the assembler
up front," says Cowger , " It 's the little
things - the bunts and the singles."
"You guys hit a home run "...hen you
keved in on thi s, " sai d a namel ess
D"-leo employee. "This is hot st uff! "
SECTIONAFLANGEO JOINT
" DFM is an ena bling tec hnology,"
Cowge r opine s, " It 's evo lutionary
rat her t han revolutionary."
" It's like we're re-learning the de-
sign process," says a secretly sat isfied
CpC user . "Only t his time we'll do it
right !"
"It's the removal of two screv...'s here,
or the addition of a lap j oint (Fig. 21
there that makes a difference," says
Cowger.
" I can't ta lk about it , but BD! has
made a big difference...real big!" says
a Saturninsider.
"You're looking for minutes...even
seconds," Cowger offered.
"We' ve pulled 35<;7, out of the assern-
bly t ime on that [ 1993 GM]
car."-satisfied BDI user.
Silver Bullets!
If we had any doubts in our minds
that we wer e ont o so met hi ng, t hey
were gone by the t ime we left t he office
of Don Smit h, senior computer appli-
cat ions engi neer at Ford's roboti cs
cente r . Ford ha s been invo lved wit h
Booth royd Dewhurst DFMAprinciples
since Sandy Munr o-now with MTS
for Pr oduct ivit y - br ought it to the
company in 1983.
" Look at the compari son (Fig. 3) be-
tween the 1980 Fair mont/Zephyr AC
unit and the DFA' d redesign in 1986,"
Smith boast s. "The numbers ar e dra-
matie, but we usually tell people to ex-
pect a 20<;7, 30<;7, improvement. "
We spoke with Munro about Smith's
claims and he was even more optimis-
t ic. " Whe n I was at Ford we used to
think that if we got a 15<;7, reduction we
SECTION ALAP JOINT
were rea llydoing somet hing. But every
time we looked at an assembly we got
anot her 15<;7, . Now [ won't accept 15<;7,
anymore, I look for 40<;7" 60<;7, and in
some cases 100<;7, - that' s where we
eliminate a sub-assembl y complet ely
and do somet hing else. "
Smith point s out that the soft ware
helps balance out the relati onship be-
tween ass embl y and manufactu ring.
" It may cost a little more to manufac-
ture a more compl ex part ," Smith
says, " But the point is thi s: If the new
part cost s more t o make, you have t o
consider howmuchit would have cost
t o assemble th e old one . Invar iably,
when you compare unifi ed parts t o
separate, the unified part is cheaper."
Smith says that as of the 1990 model
year, virtually every vehicl e in produc-
tion at Ford has had some involvement
with DFl\l A. Some cars like Taurus!
Sable had extensive analysis (40<;7, of
vehicle), an d yet it probably won't be
until MY 1994 that For d produ ces a
veh icle t hat has bee n analyzed from
bumper to bumper . Even so, side-by-
side comparisons reveal t hat a typical
Taurus has hund reds fewer part s than
a GM 10 car - most of t hem fastener s'
If you have doubts t hat there is a
payback in such an extensive proc-
ess - and ther e is st ill resi stance to
DFMA- Munro ca n ease your fears.
" One of the years I was at Ford I was
nomi nated for an award and we had to
find out how much DFl\IAhad saved or
deferred in cost s,"l\lunro says . " When
we came up with a figur e of $1.2 bil-
lion, that let us know it was the right
thing to do."
Fueled by Mu nr o' s numbers, we
as ked Smith t o talk us through the
BD! analysis process-a request t hat
he answered with a full-blown demon-
st rat ion of the system.
Unlike ot her analysis systems on t he
market, one of t he virt ues that recom-
mends BD! is that it uses te rms t hat
can be ea sily under st ood. It deal s in
dollars and cents, minut es and sec-
onds -words that are meaningful.
The first thing the program asks is
cost per hour, so it can calculate your
labor costs. Then it want s t o know if
you wish to analyze the part during as-
DFMA
1980 " FOX" AlC EVAPORATOR
Ford Ale evaporator was analyzed with BDI software. Assembly shows 20% quality im-
provement, 59% part reduction . 22% cost savings.
int o the system, the soft ware prod uces
a chart t hat - aside from doubling as a
pr ocess sheet - evaluat es design effi-
ciency and highlight s part s that could
be eliminat ed. Th e designer can t hen
go back a nd evaluat e t he software's
recommendati ons, int er act ing wit h
t he program so th at he can assess the
improvement s.
If it sounds simple, t hat' s because it
is. Admittedly, we' ve only been given a
t humbnail sket ch of how one of t he
soft ware programs opera tes. What we
haven ' t ment ioned is t hat t his pro -
gram is jus t part of a seri es of modules
t hat work t oget her to consider : DFMA;
Assembly System Economics and Ma-
chine Simulat ion; Design for Robot As
sembly; Design for Automati c Assem-
bly and Handling.
In addit ion, BDI has software pack-
ages for : small metal st ampings; plas-
t ic inj ect ion mo lding; machini ng;
pr int ed circuit s. Future programs will
include powdered met als, die cast ing
and possibly large panel ma nipulat ion.
Mun ro says DFMA has impa ct ed
every t hing he' s tried it on, a nd t hat
t he biggest gains are yet to be made.
" I' ll t ell you what , if you really want t o
find a way to where the big money is,"
Munro says, " look at t hose sub-assem-
blies th at th e au to compan ies farmed
out because t hey couldn' t find a way to
manufact ure th em economically. Look
a t t hose pr od uct s wit h DFMA a nd
you' ll be looking at cost reducti ons of
50'70 or 60%."
Becau se Munro is unable to t alk
about his current involvement s, he of-
fers a chall enge: " I' m looking for a
company t hat has a product t hat at one
time had a large market share, but has
fallen on hard t imes. I' m willing to do-
nat e my t ime t o t urn t he company
a r ound in exc hange for ha vin g the
freedom t o talk about how it was done.
If th at happens, t hen you' ll hear a re-
ally amazing story!"
To accept t he ch all enge or lea rn
more about DFMA contact:
Sandy Munro
MTS for Product ivity
911 W. Big Beaver
Troy, MI 48084
313-362-5110
(Tell him AI sent you). [i!]
1986 " FOX" A/ C EVAPORATOR
(DFA PROPOSAL)
Accns 129 PARTS!

} HU ff"
evA_ATOII
CASf
Ul' PU us.

HUf_
Of.Il0n

you tojusti fy t he reason for the design .
It does t his by asking t h ree basic
quest ions: 1. Does t his part need t o be
separate becau se it moves relative to
par t s alr eady assembled? 2. Does it
need to be a different material or iso-
lat ed fr om parts al ready asse mb led?
3. Would combini ng t his part wit h
previous part s ma ke asse mbly or
disassembly impossible? If you answer
"yes" t o any of th e quest ions, th e soft -
ware accept s th at t he part must be sep-
arate. But if you were able t o answer
" no, " when the soft ware is finished
wit h it s analysis it will flag the part as
a candidate for elimi nat ion.
As each part is added, t he soft ware
looks closely at any " irregular" an-
swers . For inst ance, if you wer e to an-
swer t hat a part was difficult t o grasp,
t he pr ogram would want t o know why.
In t his case a window appears and asks
what sort of pr oblem it is. Is t he part
hea vy? Oily? Does it t a ngle ea sily?
Nest s wit h ot her pa rts? Too flexible?
Sharp? Too small? etc. The program
th en considers howmuch ti me t he spe-
cial handling adds to assembly, and fig-
ures it int o th e cost of t hat st ep.
When all t he parts have been logged
CAS(
cOWU U S,
Fig. 3 Credit: Ford
sembly or disassembly. The reason for
t his feature is to accommodate analyz-
ing exist ing designs - which you can do
dur ing di sassembly -or new designs
as t hey are being bui lt.
For demonst rat ion purposes, Smit h
chose to analyze during assembly.
The first prompt is t o name t he ini -
ti al part in t he ass embly, in t his case a
" block." It t hen as ks in format ion
a bout t he block such as: I s it rota-
tional? What are it s dimensions? Is it
easy t o gra sp? Inserti on sy mmet ry?
(when t he par t is in correct insert ion
posit ion, how many degrees must it be
rotated before it is in correct positi on
again)? Is t here a clear view of mating
locati ons? Good access? Easy to align?
and so on.
As each qu est ion is answer ed, t he
program builds a " profile" of t he part
t o determine the cost of assembly, and
the necessary relat ionship of t hat part
t o t hose aro und it .
As t he next part is bro ught i n to
place, t he soft ware "examines" it s re-
lat ionship t o the first part . Aod since
the " key" t o the soft wa re is that it
keeps track of parts t hat are potent ial
ca ndidat es for eliminat ion , it forces