88 Designing capabl€ componentl and assembli€s

Example - determining the failure cos|6 for produd designWe will now consider calculating the potenrial cosls of failure in more detail for thesover support leg shown earlier. The process for calculating the failure costs for acomponent is as follows:

. Determine the value of 4- or ./,

. Obtain an FMEA Severity Rating ("9)

. Estimate the number ofcomponenls to b€ produced (1v)

. Estimate the component cost tPc),

For €xample, the characteristic dimension 'A' on the cov€r support l€g was critical tothe success of the aDtomated assembly process, the potential failure mode beinga major disruption to the production line. An FMEA Severity Rating (S)- 8 isallocated. See a Process FMEA Severity Ratings table as provid€d in ChryslerCorporation er al. (1995) for guidance on process orientated failures. The componentcost, P(= t5.93 and the number planned to be prcdrced Fr annum, N = 50000.

Tbe characteistic was analysed using CA and gm was found to be 9, The values of4. = 9 and S = 8 are found to intersect on the Conformability Map above the l0%isocost line. (lftheyhad intersected b€twe€n twoisocost lines, thefinal isocost valuersfound by interpolation.) Ifthere is more than one critical characteristic on the com-ponent, then the isocosts are added to give a total isocost to be used in equation 2.15.The total failure cost is determined frcml

isocost(%) xf fxPc

(2. rs)

Figur(mouldeaPotentia

Risk Prgiven (lcould ca, :8 bof Occu256. Thr

further.of the c,

Some

The posneeded tpositionion the t

mould pThe p,

angular l4 + 0.08th€ deptltrica) colikely on

The r€variabilidescriber

mould p

Follot

out of tlrRisk Antbund in

li is e,tolr and

Severityaddition

Total failure cost =

Total failure cost =

100

10x50000x5.93= r29 650100This frgure is ofcourse an estimate oflost profit and may even be conservative, but itclearly shows that the designer has a signincant rol€ in reducing the high costs offailure reported by many manufacturing companies. The results are repeated in theConforrnability Matdx in Figure 2.33.

2,7,1 Electronic power assisted steering hub design

Under this heading, a flexible hub design for an automotive steering unit is analysed.The application of CA resulted from the requir€ment to explain to a customff howdimensional charact€ristics on the product, idenlified as safety critical, could beproduced capably. A key component in this respect is the hub. The component ismade by injection moulding, the material being unfilled polyburylene terephthalare(PBT) plastic. The moulding process was selected for its ability to integrate anumber of functional elements into a single pi€ce and reduce assembly costs. Thedesign outline of the hub is showr in Figure 2.34(a) together with an oprical platethat is mounted on il.

Cas€ studier 89

Figure 2.34(b) shows a line from the design FMEA relared to rhe plasticmoulded hub. lt gives the component function! the potential failure mode, thepotential effects and polential causes of failure. In addition. the columns of currentcontrols, Occurrenc€ (O), Severity (,t, Detectability (D) ntings and associatedRisk Priority Number (RPN) have been completed. A high severity rating wasgiven (S = 8) since faulty posilional rsdings due to out of tolerance variationcould cause loss of car control and driver injury. The Detectability was rated atD = 8 because complex inspection proc€sses would be required. The possibilityof Occurrence was provisionally estimated at O = 4, giving an RPN equal to256. The design FMEA specifies compatible dimensioning as the current controlto avoid failure, and it was this aspect of the design that needed to be exploredfurther. A quotation ffom a supplier had been received for volume productionof the component,

Some detail on the analysis ofthe hub design is given below. The hub performedseveral functionsin the controllerand therefore carried several critical characteristics,The positional tolennce of the rec€sses to accommodate a system of location pegsne€ded to bc clooe, Faces on the hub for mounting the optical plate required pre{isepositioning to provide th€ necessary spacing between two optical grids (one mountedon the hub and the other carried on a torsion shaft). The depth of the mouldedrecesses needed to be controlled as they were part ofa tolerance chain. lt is importantto note that the recesses and faces were in ditrerent planes and the depth was across amould parting line.

The positional tolerance on a l0mm dimension was +0.1mm (to\), providing anangular position of0.6'. Additionally, the widths of the rec€sses necded to be held to4 + 0.08mm (tolr), the dimensional tolerance on the faces to I + 0.07mm (toll) andthe depth to l0 + 0.12mm (tola). Also, the thin sections of the hub gave two geome-tdcal concerns as these vane6 wer€ on th€ limits of plastic flow and distortion waslikely on cooling.

The results of the analyses carried out by the business on the hub are given in thevariability risks table shown in Figlre 2.34(c). The four critical characteristicsdescribed abovc were examin€d. The positional tolerance (tolr) set across the thinvanes resulted in high geometry to process risk Gp) and gave qm = 9. This equatesto an out of control C , as does the {m value for the le€ess depth across themould parting line (tola) which came to 8. The qD scores for the chamcteristics tolzand tolr suggest initially that the process will be in control giving estimated Cbkvalues of 1.33 and l.?5 respecrively.

Following th€ completion of th€ variability risks table, a CoDformability Matrixwas produc€d. Tbis was used to relate tbe failure modes and their severity comingout ofthe d$ign FMEA to the results of tle Component Manufacturing VariabilityRisk Analysis. The portion of th€ matrix concerned with the moulded hub car befound in Figure 2.34(d) and was completed using the Confomability Map.

It is evident that the two characteristics described earlier as being out of controt,to\ and tol4, give costs of failure greater than l0%. Also, the characteristics tol2and to13 which may have been rcgaded as havirg acc€ptable Cp! values are shownto have costs of failure of greater than l0% and 0.2% respectively, due to the highSeverity Rating (S) = 8 for th€ potential failure mod€s in question. Note ttlat twoadditional failure modes are also illushaied.

(a) Hub Design Uould€d hub Optical plat€

:l

Jt

o_, , -T\"a's'-\J- /

(b) Design FMEA

tailu1€ tailure

Exi8ting conditione

o

I

o

o

z

a b 256

(c) Variability dsks

(c) Variability risks

Figure 2.14 Hlb anayss resr ts

j

€i r

E€

a!

I

5

3

r3

i ! " '

!'T3

3 !a

ItPI

{! *-k :

Ig

(d) Conformability matrix

5B3;

6s

FJ|nbo!..!so|!dE4dydE{9

; ;!c

aa

ea

92 Designing capable (omponents and assemblies

The analysis indicnted that the confomrance problcms associated with thc hubdcsign had a cost of failure of more than 3{l%. This would represent at the annualproduction quartity required and largel selling price. loss to the business ofscverulmillion pounds. As ir result ofthe siudy the business had lhther detailed discussbnswilh their suppljers ilnd nol suryrisingly it turned oul thal the supplier would only bcprep:rrcd b sland by its origiral quotarion provided thc bleranses on the hub.discussed abov€. wcrc opcncd up considorably (more than 50yo). Subsequently, thisresull supporled thc adoption ol anolher nore coprble d€sign schcme.

2.7.2 Solenoid security cover

This clsc study concerns the inirirl design and redesign ol ! security cover assemblylbr:r solcnoid. The analysis only foouses on those critical specls ofthe assembly ol'thc prodnct that mu!t1 be addressed to neet the require cnr that the electronics insidcthe unil arc scrlcd liom the oulside environlnenl. An FMEA Scvcrity R,lting (S) forthc asscmbly w{s dctcrmined 0s S = 5. a wilrranty return il fuilure is experienced.

Cover assembly initial designThc inilial dcsign ol thc covcr $scmbly as $hown in figure 2.15 uscs an O-ring to seolthe electronics agninst any contaminlltion. Concerns were raiscd irbout thrcc mainsaspects of thc sscmbly. thcsc bcing:

. Thc comprcssion ol thc O-ring m.ry work agajnst thc nccds ol lln rdhesive cure onl ln, i l asscmbly with an end uni l .

. Thcrc is n risk thal thc O-ring will not maintain its piopcr oflcnhtion in lhe coverrecess during subscqucnl asscmbly proc€sses, .rnd thercforc mny not be correctlyposilioned on linal asscmbly. Rcstriclcd vision of the insidc ol lhc cover is thekey problem here.

. The wire cable may present problens using cithcr manurl or rutomalcd ,tsscmbly.

The analysis in Figurc 2.15 shows thal lhcre is a high risk of non,conlblmance for theinsertion oflhe liamc into thc cover, the process relying on thc position of rhe O-ringbeing maintiincd bpcration ,i8). The situation is complicatcd by the restriction ofvision during O-dng placcmcnl. and this is refl€cted in rhe analysis. Using theConformability Map, it is possiblc b calculate the porenrial f^ilurc co(s fiu thisdesign scheme in meeting the sealing inlegrity rcquirement, as docDnenLcd in rhcConlbrmability Matrix. The final failure cosr is calculaled ro be 1805000. Thispolential failure cost for rhis single failure mode is far too high. representing ovcrl0'% ol lhe totd product cost. A more reasonable targcl vnlue would be less thanl%. An alternalive d€sign scheme should be devclopcd, lbcusing on reducing thcrisks of the final assembly operation to reduce the potcntial for non-conformrnccas highlighlcd by lhe analysis.

Cover asse m bly redes ig nUnforrunately. the design ofthe wirc could not be chlng€d to a more sirnplc arrange-ment. forexanrple usinga spade con nechf integra ted with a reccss for lhe O-ring. Thewire is part of the customer's requiremcnls and will inevitably present problcms using Figure 2.

Case studies 93

hubruj llral

bl:

idcfor

!alins

Cover .3aembly

Ar3.mblyaoquencedlagnm

tl]he

lr

rg

b€

is

I

4e'd@re4 n$1tu h v.tea xM tr) . ad !a5,D_ b.1

- alr]!14149r190...tuh M6

Conformabttity matrix

figure 2.35 Cover assemby des gn afa ys s

' !EIgr g' 9€

o

rt

o

94 Designing (apable components and assemblies

eilher mannal or automatic asscnrbly operalions. Looking to rhe O-riDg. a beucrdesign would bc to eliminare it rllogether and inregraie thc seal wirh the wire asshown in Figure 2.36. Thc wirc is then posilivety joc,trcd lvhh thc seal in rt,ecover hol€. Clcrrly. the risks associ,rred wirh thc cover asscmbty have bccn reduced

l inlowin!

Mrn. i ( ) i

2.7.3 Teass€mblyredosign

cllagram

(,s) - 8. .

wilh thc.

! . l r i l lb i l i l '

n l lys is I

(rssembh

The coturther lu

including

failur.

involvc tl

Conformabitity matrix

Figure 2.36 Cov€r assemby red€srgn analysu

F8

o

cas€ studies

following the elimination ofofle positionally unstable component and its int€grationwith adother. Again, with an FMEA (S) = 5, and referring to the ConformabilityMap, isocosts for each assembly vadability risk can be€valuated and the total failurccost is calculated to be €7000.

Comparing this value with the initial design's high potential cosl of failure, it isevident that amajordesign fault in the cover has been eliminated although the assem-bly proc€ss must remain within special control. Subsequently, the rcdesign solutionwas chosen for further design development,

2.7.3 Telescopic lever assembly

Consider the telescopic lever assembly, Design A in Figure 2.37, which b parr of astretcher, and hence safety critical. The assembly has an FMEA Severity Rating(S) = 8, ana is used in a product having a cost of €150. lt is estimated thar 5000units are produced Z?/ dnr?4m. The assembly is subjected to bending in operation,with themaximum bendingstresses occurring ata pointonlhemain tubecorrespond-ing to the stop ring r€€ess.ln order to provide additional support, areinforcing tube ispositioned as shown. It is crucial that the tube is placed where it is. since fracture ofthe telescopic lever may result in injury to users and third parties. Figl]re 2.37 showspart of the results from the Conformability Matrix for Design A. At each node in thematdx, consideration has b€en given to the effect of the component or assemblyvariability risk, represented by 4h or {r. on the failure mode in question dercrminedfrom an FMEA.

While the desigD is satisfactory from a design for strength point of view, theanalysis highlights a number of areas where potential variability and failure severitycombine to make the risks unacceptably high. For example, there are no designfeatures which ensure the positioning of the reirforc€ment tube in the assembly(assembly process a3 in Figure 2.3?). There are also several critical componenttolerances which need to be controlled if its position is to be maintsined i[ service,such as the irtne! diameter of the main tube, cl, the outside diameter of the reinforce-ment lube, c2, and the application of the adhesive. a2.

The conclusion frcm the analysis is that tle assembly should be redesigned. This isfurther justified by calculating the potential costs of failure for the assembly. lfthisdesign of telescopic lever assembly fractured i'l service, user injury, high losses,includitlg legal costs, could be incurred. A cost of failure of€25?OO0 was calculatedfrom the analysis, which is far too high reprcsenting more than 36% of annualrevelue from the product. This figure was calculated by summirg the isocosts loreach characteristic/assembly prccess whose variability risks potertially contributeto €ach failure mode, and then multiplying the total failure mode isocost (%) bythe product cost and number of items produced. The calculation of the costs forthe second failure mode type (reinforcing tube moves our of position) on Design Ais shown in detail in Figure 2.3?.

Failure ofthis design jn service did in fact resulr in user injury. High losses of theorder ofthose calculated above for the particular failure mode, includjng legal costs,were incured. A number ofaltemalive designs are possible, and one which does notinvolve the above problems isincluded with its Confonnability Matrixin Figure 2.38.

95

EEan

96 Designing capable components and ass€mbliej

A88E[IBLY SEOUET{CE DlqcRAII

r1)-

{FI-

tt=t !

l f iEI t rt ! r| 4 i

l5 lt;Ir-r+1"2 rtrl--rE

Flgure 2.31

r-

'9anpL aalau aian d It urc ao6n l

Nonb.t at o^tta = @AA

?rcaudao6r(rc)= 50

lotal tailorc naAa i.aaa.r (\) = 2 + ta + a 05 = 12,a5

n.doa.wttrturc .obr - Zg:@:2 - rso37"1AA

CONFOR AAI|TY TIATRX fOR OESIGN A

tigure 2.37 Te es.opic ever assemb y ana ]/sis for Des 9n A

amoull!.The ab

manufactimporlanharm elirIosses thaof this tythe prodr

2.7.4 S

for a sol(

OUTER TIJBE

E

E3,t

ComponenV

de3cdplloni

f.nunI|.d.D.*rlr{s.||dFrE &rllym|il{!J

comm6nb5+u

Eft :*FB

6. t

Fi*, "5

a 0 aMA N TIJ'E

0 1.6 o UNI'CCETAOLE

MA N TI]DE 1.6 ata2 2.4 O I]NI€CErlA'LE

c2l 1,5 !

| ,e 4,5 oc5l r5PLtfl Le a) ACCE4AAte

o UNACCErIADLE

OIJTER TLJ'E

t toToblF.lluE |tod. l!o@l (%) 12.6 2

,:b.bK t257K

Case studi€s 97

AASE BLY SEQUENCE OIAGRAMDESIGN B

iFs*E

E'J,58

6g

Fdtr[N.odd|,{cnr'dflu..wiyMd(51

:E

ora

lu2 r9 o

Toid F. u6 Mod€ !o@l (%)

COIIFORMABIUTY MATRIX FOR DESION B

tigur€ 2.38 Telescopic lever assemby anays s lor Design B

The costs ol failure for rhis Design B were subsequenlly reduced to a n€gligible

The abovc examplc demonstrates thc use ofCA in supporting thc identincation of

manufactu ng and assembly problcms before production comnences' but mor€

inrporrantly s;f€ty problems can bc syslematicallv identified and rhe potential for

harm eliminated before lhe product is in use by the cuslomer' Given th€ huge

losses lhat are associat€d with safcly criticnl products when ihcy fail, considerations

ofthis type musl be on the ag€nd,t of all manufrcturing compani€s, €spccially when

tbe product has a high degree oI interaction wirh the user'

2.7.4 Solenoid end assembly

The followingcase studydetermines the manufacturing and assembly variability risks

lbr a solenoid end assembly design, shown in Figure 2.39, and projects the potential

98 Designing capable components and ass€mblies

flgur€ 2,39 Soeiod end assembly

costs of failute associated with th€ capability of an ass€mbly tolerance stack Thesolenoid is to operate as a fuel cut"off device in a vehicle, operated when a signal isrepeived from the ignition. The signal allows currenl to ffow to the inductor coilwhich then withdraws the plunger seal and allows the fuel io ffow. The solenoidassembly is to b€ sorewed into an engine block at the fuel pott in a counterboredhole. An importanl requirement is that the plunger displacement from the engineblock face through the solenoid tolerance siack to the plunger seal face must bewilhin a tolerance of +0.2mm, If this requirement is not met, fuel flow restrictioncould occur, this being the main failure mode. The product will be in the warrantyr€turn category as it has little effect on user safety if it fails in service, which relatesto an FMEA Severity Rating (S) = 5. The product cost is !7 66 and it is estimatedthat on€ million units will be manufactured in total.

solenoid end assembly initial designTh€ inilial design is analysed using CA st a component level for their combinedabilily to achieve the important cusiomer requirement. this being the iolerance of+0.2mm for the plunger displacement. Only those charactetistics involved in thetolerance stack are an*lysed, The 'worsl case' tolerance stack model is used asdirecl€d by the customer, This model assumes thai each component tolerance isat its maximum or minimum linut and that the sum ofthese equals the ass€mbly tol'eranc€. given by equation 2.16 (s€e Chapter 3 for a detailed discussion on tolerancestack modeh):

\ - r ,<, .

/i : bilateral tol€rance for ith componcnt characteristic

r* : bilateral tolerance for assembly stack

Figure 2.40 shoqrs the initial dctailed design including the tolerances required or enchcomponent in the stack to achieve the +0.2Ir]fi assembly tolerance, ,a (not included isthe dimensional toterance on the fuel port block of 12 + 0 05 mm which is set bv the

(2.16)

Figure 2.10

supplier).

back to I,ilates throl

Case studies 99

210.05'

22 +0,A35

@28 r0,05 0.210.025

I}rci is

inebe

i)n

ted

NUT PROFIIE INTERFERENCE FITH7p6

t6)

oIlbe

r isDI

Chl isb€

i.qulr.d plung.| dlrpt.c.m.nt r 0.6 r 0,2 mm

' Tor.trno tu.c by ruDl.rtb.l{.d|ns..i.|v.|.'fu.,tnn

Figufe 2.40 Sol€noid end assemby in tatdes gn

supplier). Also showr is a table describing the process used ro manufacture eachcomponent and^an assemhty sequence d,dgfam rr gi \en in trgufe 2.4t. Referr ingbacl ro Figure 2.40. lhe tolerance qlack starr. at face A on rbe fuel porr and accumuilates through the individual components to face B on the Dlunser seal.

Wo6t caae lolerance stack

EI{DASSEMBLY

ggE?;Y=f9 r5; iEgl?Eig

KEY fO ASSEMBLY SEQUENCE

a ASSA'SLY SEdJENCE IIUiI|AER

p PO6T MiI]UFACIURING PROCESS NlJIls€R

o V\ORK l€rc€R, c.t hcndrirs 9ad lo agdy

il ASS€MBLY PRoCESS. s rtitg, itsid

iDDtro{ uPosr AssEiGlY PRoCESS,as

't€n{,\85',lg. b d.,

ASS€II|EIY IN MANt'frcruRNG PROCESS.& is.*b h nnJfi!r'€djig

FOSr Hi TUFACTURTNC PROC€SS,..0 b!& d $fte n€, t .f ialsrc€urss

Figur€ 2.41 tusembly ,equence diagram for the solenoid end asembly desiqn

! IJT l - r -

j t ei | $itg

Case studies

The body is impact extruded from a cold forming sreel. The characreristic dimen,sion to be analysed in rhe tolerance srack js the base thickness of3mm (on a Z20rnmbore) and this dimension has been assigned a tolcrance of I0.02mm.

Following the tolerance stack through the end assembly. the bobbin dimension of22mm from the ourside face to the back face of the magnetic polc is analysed next.This charact€ristic dimension does not include the tolerancc on the imoaci cxtrudedpole. l he pole i , lo be moulded inlo lhe bohh,n and rh€ pole tace i , considered lo bepart of a mould related dimension. The bobbin is inj€ction moulcled usins 30% filledpob hurylene rerephthalare tPBTr. The rolerrnce ssigncd ro rhe bohhin dimension i ,I0.035mm.

r01

l rT '{ l

3r

t3I,

t3

I

I

tI

I!

'

lili

Ii

l rIrIt

iII

I

I

!Ilc!

I

I

' lt . ,

Figur€ 2-42 Varab lity r6ks analJ6is for the soteioid end assemblv nitialdeson

102 Designing capable corhponents and assemblies

The pol€ has a characteristic dimension of8Inm from tle rear of thc bobbin to therecess face and has a tolerance assigned to it of +0.02mm. From the pole recess face,the tube base tolerance is the lasl component to make up the tolerance stack, Thebrass tube has been given a dimensional tolerance on its base of 0.2 + 0.025mm.Note, the dimensional tolerance on the plunger is 28 + 0.05mm, but the analysiswill concentrate on the silicone rubber seal lcngth of 6mm because this is mouldedonto the plunger and again is a mould related dimension.

Figure 2.42 shows the variability risks analysis based on the tolerances assigned tomeet the l0.2mm tolerance for the assembly. Given that an FMEA Severity Rating(S) = 5 has been determined, which relates to a 'definite return to manufaclurer',both impact extruded components are in the unacceptable design region, as w€llas thc bobbin and plunger end seal as shown on the Con,brmability Matrix inFigurc 2.43. The lolerance for the brass tube base thickness has no risk and is anacceptnble design.

The assf3 milliorThis Iigur

FUt{bSO

48FI

a6

f.nuD rbd. D.lrrrd.n .id FNE .mdt iiri. {51

>=

I o DE5I6N

.1 IHCKNEgg9 o DE96N

gEAL LENGTN oFOLELEN6TH

9 or)

Tor.rF6iruE Modo boco6l (%)

E3.A1V

' Nrnb.r d rnttb = 1 OAA Oaatuoauar co.r (Pc) = t7.66lara la urc noa. i6oao6r ('A) = 10 + 14 + a,O1 + 14 + 1a = 4a.o1

f..-lor..w 'a1.-naa. -a - 4pg::-@-@::.@.- 13064 7bb

Figure 2.43 Conformabiiiy malr x for the so eno d end assembly nitialdesign Figufe 2.4

D theface.The

mm.tly"itided

dtotring

[ t rnban

Case studies 103

The associated cost offailure for the solenoid end ass€mbly is calculaied to be over!3 miUion for a product cost of !7.66 and production volume of one miuion units.This fignre is for th€ tolerance stack failure mode alone as this is most importantto lhe customer. Although the assembly variability risks are analysed, they are nottaken into account in the final costs of failure. ln conclusion. the orocess caDabilities

0.2 r0.015

t

INTERFERENCE FITflTpo

R.qulod plung.r.ll.pl.c.mnt. 0.8 r 0,2 hh

.]o|.fuc.nr.dby!UFl|.rb. u..d In rh. rn.t.l.. nd 2t mm

tigur€ 2.44 Solenoid end assemby redesign

-)-r---J-

104 Designing capable componen$ and ass€mblies

ofseveral characteristics in this tolerance stack are inadequate and will not meet thecustomer's requirements consistentiy.

Solenoid end assembly redesignThbissimilarto the inilial design, butinvolving turningas a secondary process on thebody to improve a key tolerance capability as indicated in Figure 2.44. The body isstill impact extruded, but the face which mates with the fuel port block is machined,together with a shoulder on the inside diameter. The front face of the pole, now fullymachinedand assuming no component cost increase, is assembled up to the machinedshoulder on the body. Only the tolerances on the pole length, tube and plunger endseal r€main in the stack, reducing the number ofcomponents to five,

Conforma

the *0.2nThe ass

€3000. H(

involvingthese criti(the cost ol

As high

Figure 2.4

aI

t !t ll ;

3_

ttIt

t

I

1

l ! ! lllliilf

.;

tI

FItI

tII

it

I

I!

!itI

t "

t . ,

Fsr-dE5E

f83A

Figure 2.45 Varabilly risks analyesfor the solenoid end assemblyrcdesign

casa studi€s

The variability risks tabl€ for the redesign is shown in Figure 2.45 and theConformability Matrix in Figure 2.46. Clearly, machining the critical faces on theimpact extruded components has rcduced the risks associated with conforming rothe +0.2mm tolerance for the plunger displacement.

The associated potential cost of failure has reduced significandy to a little overf3000. However, there is an additional cost associated with the extra machinidgprocess which adds to the overall product cost. Since it is likely a secondary processinvolving machining *ill take place on the body thread an)'!vay, the case for turningthese critical faces may be furtherjustified. Although machining these faces will raisethe cost of the component slightly, this must be secondary to satisfying the overridingcustomer requirement ofmeeting the plunger displac€ment tolerance,

As highlighted by the diference in the potential failure costs, the redesign schememust be chosen for fu her design developfient, Of course. other design schemes

105

6

FEa65E

l :5a

assemuyproc6ss

F.n!'r r!.. D.*in|d d FI.,EA h'ny tulne 6)

ftmmonls

(lncludlng

tor Bupplloru)

i6

i ;V

5

lftcKNE96 1,OS r)t

MAoIETICPOTE I.EN6TH n

DESIGN

1U* 1.06 a)oE6t9tl

Toiar Fdlur. Mod. bo.osl

'Number ol unils = 1000000Product cost (Pc) = e7.66Total lai lur€ modo l8oco3l (%) = 0.01+0.01+0.01 + 0.01 = 0.04

0.04 r 1000000 x 7.60= c30t4Thorolor€, toral lalluro mod6 cost = 100

Figure 2.46 Confornrabi ity mair x tor ihe solenoid end assembly redesign

106 Designing capable (omponents and assemblies

could also be explored, bul the initial design shown here is ofinhercntly poor qualityand lherelbre must be rejected.

Decisions made during the design stage of thc prodlctdevelopment process account fora large proportion ofthe problems that incur failure costs in production and service. Itis possible to relale thcse failwe costs back to the original design intent whcrevariabilily, and the lsck of understanding of variability, is a key failure costs driver.The correct choice of tolerance on a dimensional characteristic can be ca cial for thecorrect funclioning of the product in service and tolerance selection can have a largecontribution to the overall costs of the producl, both production and quality loss.

Process capability indices arc not generally specificd by designers and subsequentlylhe impact of design decisions made on the production department cannot b€ fullyunderstood because tolerances alone do not contain €nough informfilion. Variabilityin cornponent manufacture has proved difrcult to predict in the early stages of thedesign process and there are many influencing factors that the designer may notneccssrrily be abl€ to anticipnte. The material and geometrical configuration of thedesign, rnd the compatibility with rhe manufacturing process are the main variabilitydrivcrs. Although design rules and general manufacturing capability information areavailable, they are rarely presenled in a useful or practical form, especinlly wheninnovative design is required. There is a need to set realislic tolerances and anticipatethe variability associated with the design to help reduc€ tailure costs latcr in theproduct s life-cyclc.

Th€ CA melhodology is useful in this respect. It is comprised ofthree sections: theComponent Manufacturing Variability Risks Analysis. thc Component AssemblyVariability Risks Analysis rnd the determination ofthe EFects ofNon-conformnncethrough the Conformability Map.

The Component Manulncturing Variability Risks Analysis prescnted. models theimportsnt design/manufacture inrerface issu€s which reflect the likely processcapability that can be achieved. Included is th€ assessment of toleranoe, geometry,malcrial and surface roughness variability in componert manufacturc. euantitativernd qualitnlive manufacturing knowledge is used to support various asp€cts of theannlysis and is taken from a widc range of sources. Th€ concept of an ideal designallows the annlysis to generate risk indices, wherc values gr€rrer rhan unily have apotential for increased variation in production. A simple cost tolerance relationshipis used in the Proccss Capability Maps. developed for over 60 manufacruring procestmat€rial combinations. The maps are subsequently employed to determinc theprocess capability estimates for th€ component characreristics analysed. Throughempirical studies. a clos€ correlation between the process capability €stimates usingthe Component Manufncturing Variability Risks Analysis and shop-floor processcapability has been observed.

Most literature tends only to focus or tolerance stack analysis when assessing thecapability of assemblies- The variabiliry of the actual assembly operations is rarelyconsideted and does not rely solcly on th€ tolerances accumulating throughout theasscmbly, but on the feasibility and inherent technical capability of the assembly

DFA techquality Pr

tion. Theaddress viof design!facilitatinfor th€ c,and. thenthe produ

The Cc

quantifyiranalysls IAgain, thassemDlyprocess ritne assefistag€ if 1l

Currenimproven

typicauyproducll(FMEA dfailures Iquanry-(

critical aof th€ pl

cost), anlikely prmodels Iithey bec(quality l,for chariof the pr

Throuhisl ighr

which tcthe desilsupplier!

Summary

operations performed, manually or automatically. Develolrers and practitioners ofDFA techniques reason that an ass€mbly with a high asscmbly efficiency is a bederquality product. The natural outcome of having a high assembly efficiency leads tofewer parts in the assembly and, therefore, fewer quality problems to tacklein produc-tion. The outcome is not due to any specific analysis process in the DFA technique toaddress variability, and there still exists a need for analysing the assembly capabilityof designs, rather than a production cost driven approach. A useful technique forfacilitating an assembly risks analysis is the declaration of a sequence of assemblvfor the components. Through such a diagram. each component in the assemblyandj therefore, the potential areas for assembly risk can b€ logically mapped throughthe product design.

The Component Assembly Variability Risks Analysis has the purpose of betterunde^tanding the effech of a componenCs assembly situation on variability, byquantifying the risks that various assembly operations inherently exhibit. Theanalysis processes are supported by expe knowledge and presented in charts.Again, thc theory is that an ideal component assembly situation exish wher€ theassembly risk is unity. Using the charts to refl€ct the handling process risk, fittingprocess risk and th€ risks associated with additional assembly and joining processes,the assembly situation of the component is questioned, accruing penalties at eachstage if the design has increased potential for variability.

Current quality-cost models are useful for identifying general trends in a long-t€rmimprovement programme, but are of limited use ill the identification of the failurecosts a$ociated with actual design decisions. A link between the costs that can betypically €xpected in practice due to failure or non-conformance of the product inproduction or service, and the probability of fault occurrence, is made usincFMEA rhrough rhc Conformabiliry Map. The underlying concept assumes rhat aifailures b€come more s€vele, th€y are going to cost more wh€n they fail. Thequality cost model embcdded in the Conformability Map allows the designer toassess the level of acceptability, special control or unacceptability for non-safetycritical and safety critical component charactedstics in the design by determinationof the process capability measures from the previous two stages of the analysis.The Conformability Map also allows failure isocosts (perc€ntages of total productcost), and, therefore, the total lailure cost to be estimated wit! knowledse of thelikely producr cost and production volume. The narure of rhe underliinc cosrmodels limits the accuracy of the failure con estimares al an absolute level ind sothey become useful in evaluating and comparing design schemes for their poteDtialquality loss. The model can alternatively be employed in setiing capability targetsfor characteristics to incur allowable failure costs depend€lt on the failure severitvof the product.

Through performing an analysis using CA, many modes of applicatiol have be€nhighlight€d. This has resulted f.om the way that the CA design performance measuresallow a non-judgemental 'language' to develop between the design team. tt has alsobeen found not to inhibit the design process, but provide a structured analysis withwhich to trace design decisions. The knowledge €mbedded within CA also allowsthe designer to generate process capable solutions and ope up discussion withsuppliers. The analysis is currently facilitated thjough the use of a paper-basedass€ssment. This has many benefits, including improved team working, and provides

101

!rqua1i1y

I .:::,I -:.::.

Eounlforrnicc. ItIt whereB drivcr.rl tor the. a largel loss.EquenllyI bc fullyFabi l i ryts of thela\ not! of theLiabi l i lyl|on areb \r hen|rrcrpatef In the

bns: rhe

Fem blyFmance

il€ls thc

P'o:y.lartahveI of theldcsignihave alon,rrip

he rhefrouehr using

hg thel rarelylut theFnbly

0 610 15 20 2! 10 !5 / r0

108 Designing capable components and assemblies

fA ol btel Producllon cGl)

Flgure 2,47 nfoence ol the team.based application of CA on severa product n$oduct on ploiecls

a more unconstrained approach than if the analysis were computer based, lt alsoallows the knowledge to be readily visible and aviilable at anytime for th€ designerto scrutinize and manipulate iflhey chose to do so.

The potential benefits ofusing CA in the early slages of the product developmentprocess have been found to b€:

. Early awareness ofpotential design problems through a systematic analysis

. Produces more process capable desig:ns with regard to their manufacture andassembly

. Reduces internal/external failure costs

. Reduces l€ad times

. Focused disaussions with suppliers.

Finally, the main beneft as far as competitive business pe ormance is conc€rnedis the potential for reduction in failure cosh. Studies using CA very early in thedevelopment process ofa numb€r ofprojects have indicat€d that the potential failurecosts were all reduced through an analysis. This is shown in Figure 2.47. where thispotential failure cost reduction is showr as the difference between pfe-al and pocr-C,{ application by the teams analysing the product designs.

The analysthe designto this wolcomponentolerance,stack is t\\design funIt involvescritical clcments, Nethe frnal a

l99l) . Io,lolerance

Toleranphy bccarrcqulred ?

assembly

knowled8philosopl

Today'

facturabi