[meas PERPUSTAKAAN MAKSL PPAK FEUL SELIGRAM, INC.: ELECTRONIC TESTING OPERATIONS We put ina piece of automated equipmenta year ago that only fits the requirements of one customer. This, sqiinment reduced the direct Iabor required to test his components and, because of aur labarsbased burden al: location eystem, substantially reduced his costs. But putting a $40,000 machine into the general burdan pool Falsed the costs to our other customers It just doesn't ineke sense shooting yourself in the foot at ce sane Lime you are lowering the company’s cost of operations, Past care, Manager Introduction The Electronic “Testing Center (ETO), a division of Seligram, Inc., provided centralized testing for elec- {omic components such as integrated circuits. ETO was cevented as a resul of a decision in 1979 tn consolidate electronic testing fom 11 different divisions of Seligram. ETO commenced services co these divisions in 1983, {t was estimated that centralization would save Soligram in excess of $20 million in testing equipment investment over the next Ave years ETO operated as a cust center and transferred prod iets 0 other divisions at full cast (diect costs plus allo- cated burden}. Although ETO was captive division, other divisions withit Seligram were allowed (0 use outside testing services if ETO could not meet their cost fo service requirements. ETO was permaitted to devote up to 10% of lis testing capacity 0 outside customers but chose to work mainly with otter Seligcam divisions clue to limited marketing resources, ETO employed approximately 60 hourly. personnel and 40 administrative and technical staff members. Bud- _geted expenses were $7.9 million in 1988 (see Fxhibit |). Testing Procedures ETO expected t0 test between 35 and 40 million com- ponents in 1988, These components included integrated ciscuits ICS), diodes, transistors, capacitors, resistors. “hace wos pepe by Profesor Fete BB, Tuny, Plan Sit Ua sn inp {Coprzn © 18% the Preeti aad Oo vard Busnes Scho cnt 9-08 Her uransformers, relays, and crystals. Component testing was eoquised for two reasoas. First, if defective compo- nts were nol caught carly in the manufacturing cycle, the cost of repair could exceed the manufacturing cost of the pradhict itself Studies indicated that a detective resisior eaughl before use in the menufacturing process cost two cents, If the resistor was no caught until te end produet was in the ld, however, the cost of repair could man into the thousands of dollars. Second, 2 large proportion of Seligram’s work was defense related. Mil- itary specifications frequently requited extensive testing of componcals utilized in neraspace and naval products By 1988, ETO had the ability to test 6,500 different components. Typically, however, the division would test aout 500 clfferent components. each month and be- toveen 3,009 and 3.500 per your. Componcars were 1 ceived from customers in jots; in 1988 ETO woald re- ceive approximately 12,00 lots of components, ETO performed bath electrical and mechanical est ing. Electrical using involved measuring the electrical charactertstcs of the components and comparing these rosssirements with the components specifications. For example, the specifications for an amplifier may have valled for & L-voll input to be amplified into a 10-volt outpat. FTO would deliver a L-velt input tothe compo nent. By mousing tho amphitier’s output, ETO gauged its conformance with specifications. Mechanical testing included scldsrability, compo- dent bom-in, taeemal shock, lead straightening, and leak detection. Soldcrabilty involved the inspection of components to see if they held solées. Bucn-in was the extended powering of compenents at high temperature Thermal shack involved the cycling of components be- tween high and low temperatures. Lead straightening ‘was the detection and correction of bent Ieadls on com ponents sich a axial components. Leak detection ex amined hetmetivally sealed ICs for Feaks Components varied significanily in the number and type of electrical and meckasical testing praceduzes they required, This variation resulted in about 200 dif- ferent standard process flows for the division, Process flows were determined by the different combinations of tests end specifications requested by the customer, Based on these combinations, ETO planners deter- amined the rooting of components between testing 6768 CHAPTER 2 TRADITIONAL STAGE W COST SYSTEMS equipsnent and the type of tess to be performed at exch station. TCs, for example, could follow six different flows through the facility. While some ICs only re- quired clectrical testing at room temperaure, solder- ability, and leak detection, others also required thermal shock and burr, Each type of component required separate software development, and custom tool: and Axtures were often required. Software, tools, and fixtures were developed by the engineering group. which was made up of spe- Gialisis in. sofware development, equipment mainte ‘mince, calibration and repair, woling and fisturing, and {esting equipment operation, Software engineers devel ‘oped programs for specific applications. The programs Were then retained in a software library for fiture use ETO had 6,500 different sofewate programs on file, of ‘which 1,300 were programs developed in the past year ETO also had an inventory of 1,500 tools and fxrares, cf which 200 had been developed in the past year. The large number of ols and Axtures allowed the testing of components with a wide variety of leads, pin combina- tions, and mating configurations, ‘The testing facility was vided into to rooms. The ‘main testing room contained the equipment used for electrical testing. The mechanical zoom contained the equipment used for mechanical testing, plus incoming receiving and the stockroom. A total of 20. people worked in the two rooms on each of two main shifs, snd 10 people worked oa the night shift Cost Accounting System The cost accounting system measured two comporents of cost: ditect labor and burden. Furden was gronped into a single cost pool which included barden associ ated with each of the testing rooms, as well asthe engi- neering burden costs relating t0 software and tooling development and the administrative coste of the divi sion. Total burden costs were divided by te sum of testing and engineering labor dollars to artve at a bur- den rate per direct labor dollar. The division costed each lot of components. Burdon was calculated for each lot by mulkiplying the actual cect labor dollars associated with the lot by the 1459 of burden rate. The resulting burden was then added to the actual direct labor costs to determine the Tots total eost. In 1988, the facility-wide burden rate was 145% of exch direct labor dolla of ‘hich more than 25% was atrihurable to equipment de- preciation (ce Exhibit 2. Signs of Obsolescence eral trends pointed to the cbsolessenee of the labor based burden allocation process. Since the founding of the division i 1983, direct labor hours per lot tested had been declining steadily. This trend was aggravated by an increased dependence on vendor certification. Vendor certification wes a Key component of ju delivery, With vendor certification, gram’s suppliers did the primary testing of components ETO then utilized statistical sampling w verify that the suppher's production process was still in control. Thus, while JIT fed to an increased number of smaller lots be. ing received by ETO, vendor certitication reduved the number of tests performed. Barly indications were that JIT deliveries would account for 30% of Seligram’s shipments within the nent fve year. Tn adcltion to declining direct labor content and Fever fest lots, the obsolescence of the lubor-based allo cation system was intensified by a shi from simple in spection services to broader based test technology. On Complex pus requiring screening, environmental con- ditioning, and testing, the division was consistently chesper than outside services. Where only elementary testing was required, however, low-tech outside labora tories were often cheaper, especially on large lots. The ‘advantage that the division brought customers over the outside Tabs was that the lawer provided essentially 00 engineering support, whereas ETO with its resident en- gineering resources was able to suppert such service on 4 rapid and cost-effective hasis. The shift co more tech- ically sophisticated services prompted a shift in the la- bor mix from direct to indirect personnel. The division expeoted to see a crossover between engineering head sount and hourly head count early in the 1990s, Finally, the introduction of higher technology cam- ponents created the need for more automatic testing, Jonger test sysles, and more data per part. Digital com- Ponents, for example, were currenily tested for up to 100 ceneitions (combinations of electrics! input and ourput states). The new generation of digital compo seats, on the other hand, would be much more complex and require verification of up to 10.000 condition ‘These would require very expensive highly automated equipment, This increase in automation Would, 1m ara lead to a smaller base of direct laber (o absorb the de- precistion casts ofthis new equipment. ‘These were fears tet the resulting increase in Duaden rates would drive some customers away, EYO had Seli- i i soescarrraton eese DEANE LTT already noticed an increase in the number and fre quency of complaints fiom customers regarding the fates they were charged for testing. ‘The division’s accounting manager proposed a new cout accounting system to alleviate the problem. Under this new system, burden would be directly traced to two cost pools. The first pool would contain burden re- Jted to the adminisirative ard technical functions (di vision management, engineering, planning, and admin- istrative personnel). This pool would be charged on a rate per direct labor dollar. The second pool would in- lade all other burden costs and would be charged based on machine hours. Exhibit 3 provides the pro- posed burilen rates, Shorly sfter the sesoonting manager subrnitted his proposal, a consultant hired by Seligram’s corporate management prepared an assessment of ETO's cost sys- tem. He recommended the implementation of a chroe- burden pool system utilizing separate burden centers for ech test room and a common technical and adminisire- live pocl. Burden would he ditselly traced t each of the three burden pools. Liks the sccounting manager's system, burden casis inthe test rooms would then be al- located on a machine hour basis. Technical and admin- istrative costs would continue to be charged on a rate perdivect labor dota ‘To examine the impact of the two alternative sys tems, FTO management asked that ¢ study be con ducted on a representative sample of parts. Exhibit 4 provides a breakdown of actual direct labor and ma chine hour requirements per lot for the five components selected for the study, Technological Future In 1988, the division faced major changes in the tech nology of testing that required important equipment ac- uisition decisions. The existing testing equipment was getting ofd and woul! aot be able 1 keep pace with developments in componeat technology. Existing cost Donenis, for example, had between 16 snd 40 input! ‘output terminations (¢-g,, pins or other mating configu- CHAPTER 2 TRADITIONAL STAGE M1 COST S¥SIEMS 69 rations), and ETO's equipment could lume up wo 120 terminations. Although the 120-terminetion limit hed only been reached & couple of times in the past few yeaTs, a neW generation af components with up to 256 icrininstions was already deing developed. Similerly, the upper limit of frequency on existing components was 20 MHz (nllion eyeles per sevond), wherens the frequency on the next generation of components was expected to be 5D MHz. The equipment requited io text the next generation of components would be expensive. Each machine cost ap- proximately $2 million. Testing on this equipment ‘would be mere automated than existing equipment, with longer test oycles and the generation of more text. cua per part It was aso likely that lot sizes would be large. The new equipment would not replace the exist ing equipment but would mescly add capabilities ETO did not currently possess. Additionally, the new equip- ‘ment would only be needed to service ihe requirements ‘of one or two costomers in the foreseeable future, Ex hibit S provides a sammary of te new equipinent’s eco- nomies und operating characteristics. ‘The impact of this new equipment would be an ac- ation inthe decline in dseet labor hours per lot of omponenis. At ihe same time, Durden would increase with the additional depreciation and engineering costs associated with the new equipment, This would resat in a large increase in the buiden sate per direc lator dol Jar. As Faul Carte, Manager of ETO, saw it the acquisi tion of the new equipment could have a disastrous effect oa the division’ pricing stevewire if the labor-based il location system remained in use: We plan on investing $2 milion ona large electronic test- jing machine tts the chips of ore or wo customers. Tis machine will be very fast and will require litle divest a bor. Its acquisition will nave «sigaifeant effect. on ou+ por sireat labor dar burden fate, which will esl ip ani crease in charges to our othor customers, (tis clea ht ¢ number of customers will alle away if we ty 10 pass bis ‘increas on. Tam afraid that we will lose 25% of ur cus tomer base if we don’t change our cast system70 charter 2 TRADITIONAL STAGE 8 COST SISTEMS EXEIBIT 1 SELIGRAM, INC. ELECTRONIC TESTING CENTER 1988 Budgeted Expenses Direct labor $3,280,018 Indirect labor 859,242 Salary expense 304.211 Supplies & expenses 538,029 Services! 245.226 Personnel allocations’ 229,140 Service allocations’ 2,498, 134 Total budgeted expenses $7,973,097 "nudes foc pai compute expenses, maltanencr Sos, and sevice ‘ot tres fom ote divs, Uncudes nde ndaatacied erployes fing bons, penome! depart inert, secant. stres/warchonsng aed hobs vacalins Spcudee bulking occupancy, tephones depestion,sloraation ss tens, and dla corte exHieir 2 SELIGRAM, INC. ELECTRONIC TESTING CPERATIONS Caleulation of Burden Rate Baced on L988 Plan BURDEN ELEMENT. Indirect labor 9859,242 Salary expense 394,211 Supphes & expenses 538,029 Services 245,225 Personnel allocations 229/149 Service allocations 2448 144 Total burden! Burden rate TOVAL BURDEN § DIRECT LABOR $ 7 $4,713,982 3,260,015 = 144.6% Effective rate 145% vente sistent Depresiion 1256.00 Other ved Livaisss “otal budeCHAPTER 2. TRADITIONAL STAGE 11 COST SiSTEMS 71 exHieir 2 SELIGRAM, INC, ELECTRONIC TESTING OPERATIONS Proposed Burden Rates Based on 1988 Flan MACHINE HOUR RATES (MACHINE HS BURDEN ¢ Main test 290m, 33,201 32,103,116 Mechanical test rooin 1,926,263, Total st0,579 Burden $__ Machine Hi Eifective machine hour rate = $30.00 Rate per Direct Labor Hour Total engineering & aciministrative buiden $ = $684,603" Total direct labor dollars = $3,260,015" surden rate = BAKES admin $ _ $684,603 _ Burden rat Direct Ibr $ $3,260,015 ane Bifective burden rate per direct labor $ = 20% ‘Machine Hoar Rate = $80.10 Bardens oad esate Dypectaion Cir Tuat Mantes oon 887379 $ 48779" $,126953. $2.103,16 Mectantal ten mom 493833808103" e7037 “Lgne 262 ‘Taal burden Ta Hea THI How css ration Nachble —§ 95,05 Fixed Depredaton 341.18 Other 196.390 Tora Sesta08 “includes all dec liber costs, including direct labor coxtsincuried in both tat soca se wll engineceng EKHIBIT a SELIGRAM, INC ELECTRONIC TESTING OPERATIONS Direct Lator and Machine Hour Reguiremeats ‘Actuals for One Lot MACHINE HOURS puect MAIN MECH, EABORS OGM ROOM Tora ICA S17 85 100s cB 205114026. 500) e Capacitor 1094 3.0 45 78 : Amplifier 525 40 10 50 Diede 519 70 50D