THE UNIVERS ITY OF MI CHIGAN COLLEGE OF ENGINEERING Department of Mechanical Engineering Heat Transfer and Thermodynamics Laboratory Six Month Progress Report AUTOMOTIVE RADIATORS MANUFACTURED BY THE ELECTROFORMING PROCESS (Covering the Period Oct. 1, 1962 to Apr. 1, 1963) John A. Clark Clarence'A.: Siiebert Robert B', -Kelle t' David M. Mellen Jo C.'Hoo''-... ORA Project 055335 under contract with: INTERNATIONAL COPPER RESEARCH ASSOCIATION, INC. NEW YORK, NEW YORK administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR May 1963

/ "', / 9c(7, 1 C(/d

TABLE OF CONTENTS Page NOMENCLATURE v ABSTRACT vii 1. INTRODUCTION 1 2. HEAT EXCHANGER ANALYSIS 3 2.1. Reference Heat Exchanger and Reference Conditions 8 2.2. Heat Exchanger Matrices Studies 9 2.3. Heat Transfer and Friction Data 9 2.4. Results and Conclusions 9 2.5. First Electroformed Model Heat Exchanger 13 3. RADIATOR HEAT TRANSFER PERFORMANCE TEST APPARATUS 15 3.1. Introduction 15 3.2. Heat Transfer Apparatus 15 3.3. Instrumentation for Heat Transfer Apparatus 16 3.4. Wind Tunnel 17 3.5. Instrumentation for Wind Tunnel 18 4. METALLURGICAL STUDIES 19 4.1. Literature Survey of Solders 19 4.1.1. Tensile and Shear Strengths 19 4.1.2. Creep Properties 22 4.1.3. Fatigue Properties 23 4.1.4. Summary 24 4.2. Literature Survey of the Mechanical and Physical Properties of Electroformed Copper 24 4.2.1. Summary 34 4.3. Tensile and Fatigue Test on Electroformed Copper 34 REFERENCES 37

NOMENCLATURE English A total heat transfer area on one side, ft2 Ac minimum air side free flow area, ft2 AFR total frontal area air side, ft2 a plate thickness, ft b plate spacing, ft Cp specific heat at constant pressure, Btu/lbm-~F f friction factor, dimensionless G mass velocity, (w/Ac), lbm/hr-ft2 go conversion factor; go = 32 2 (lbm/lbf)(ft/sec2) h heat transfer coefficient, Btu/hr-ft2- ~F k thermal conductivity, Btu/hr-ft-~F L total flow length of heat exchanger, ft NTU number of transfer units, dimensionless Pr Prandtl number, dimensionless q heat transfer rate, Btu/hr Re Reynolds number, dimensionless rh hydraulic radius (ACL/A), ft, 4rh = hydraulic diam St Stanton number, dimensionless T temperature, ~F ~U overall heat transfer coefficient, Btu/hr-ft2-~F

NOMENCLATURE (Concluded) V volume, ft3 VB air volume between plates, ft3 w mass flow rate, lbm/hr Greek (x ratio of total heat transfer area on one side to total volume of heat exchanger A/V, ft2/ft3 ratio of total heat transfer area on one side of a plate-fin heat exchanger to the volume of air between plates on that side, A/VB, ft2/ft3 water channel thickness, ft heat transfer effectiveness (Ref. 24), dimensionless total surface temperature effectiveness, dimensionless a Ac/AFR, dimensionless p density, lbm/ft3 viscosity, lbm/ft-hr see Eq. (22) Subscripts c cold (air) side of heat exchanger h hot (water) side of heat exchanger w wall max maximum min minimum i inlet o reference heat exchanger. vi

ABSTRACT This is a progress report covering the first six-month period of effort, Oct. 1, 1962, to Apr. 1, 1963, on research directed at the problems of manufacturing an air-water heat exchanger suitable for automotive application by the electroforming process. Such an exchanger would be constructed without soldered joints. The work has been divided into three parallel phases: (1) selection and evaluation of core designs; (2) a metallurgical study of the physical properties of electroformed copper sheets and electroformed copper joints; and (3) the production of an electroformed heat exchanger. Phase (1) and (2) are underway at The University of Michigan, and Phase (3) is being done at the Graham-Savage Associates, Electroformers of Kalamazoo, Michigan. Eighteen different heat exchanger core configurations have been evaluated on the basis of their thermal performance and the most compact type identified. This design consists of the plate-fin core with the fins ruffled in order to promote the turbulent flow of air, thus providing improved heat transfer capability. Wind tunnel apparatus suitable for determining the experimental performance of an electroformed core has been designed and constructed. The design of the first electroformed core has been established, and its constuction, as well as that of a similar core having soldered joints for comparative study, was started. Tensile test data on electroformed copper from a fluoborate bath produced at Savage Rowe Company averaged approximately 27,500 psi with a uniform elongation of 8.0*. The limited fatigue data available indicate a fatigue limit of 10,000 to 11,000 psi based on 10 x 10b cycles. Published data on the tensile strength of soldered joints vary from 13,000 to 29,000 psi with fatigue strengths of 200 to 400 psi for stresses to produce damage in 3,000,000 cycles. The published literature on the properties of electroformed copper give tensile strengths varying from 17,000 to 90,000 psi depending upon plating conditions. Grain size has a pronounced affect on the strength, the finer grained deposits being the stronger. vii

1. INTRODUCTION This is a six month progress report covering research activities from Oct. 1, 1962 to April 1, 1963. Owing to the problems of the recruitment of personnel and the start of the fall semester at the University, essentially full-time studies began Oct. 1, 1962. The research reported here involves a study of the feasibility of design and the performance characteristics of an air-water heat exchanger suitable for automotive application which is partially or wholly manufactured by the electroforming process. At present, common automobile radiator design and manufacture process consists of preformed tubes or water channels furnace soldered to dimpled or otherwise corrugated strip spacers forming a fairly compact and inexpensive matrix. Owing to the presence of the solder joints, however, there appears to be a definite possibility for design improvements if such joints can be eliminated. From the standpoint of thermal performance, the presence of solder is disadvantageous as it introduces undesirable thermal resistance at a critical point in the heat-flow path. This tends to require larger and heavier radiators to accomplish a given cooling task. Furthermore, soldered joints are subject to selective chemical attack from salt deposited on the highways to melt ice in the winter and to suppress dust in the summer. Such chemical attack causes further loss in thermal capacity by separating the spacer from the water channels as well as by providing a source for engine coolant leakage. Soldered joints are also susceptible to fatigue failures in service and the amount of solder necessary for manufacture adds undesirable weight to the radiator. In view of these problems it has been proposed to study an entirely new method of forming the joints between the spacers (fins) and the water channels. This method is to produce an integral and metallurgically uniform joint either by electroforming both the spacer and water channel as a unit or by joining a preformed spacer and water channel by an electroforming process using copper as the material of construction. Either method, if successful, would produce a thermally superior joint and result in a radiator core of improved structural integrity and lower coolant leakage probability than the present soldered cores. The resulting core would probably be smaller, lighter, and more reliable than present automobile radiators. It is recognized that such a core must also be economically competitive. This latter consideration will probably be decided by the success of the electroforming process phase of this study in which it is expected that current techniques will require extension or that new ones will have to be developed based on an improved understanding of the mechanics of electroforming.

The research is divided between three related but parallel phases: (1) selection and evaluation of radiator core designs from the standpoint of their thermal and frictional pressure drop performance; (2) a metallurgical study of the physical properties of electroformed copper sheets and electroformed copper joints, including microstructure characteristics, and tensile and fatigue strengths; and (3) the production of an electroformed radiator satisfactory for automotive application. This last phase is being conducted by the Graham-Savage Associates, Kalamazoo, Michigan, under the direction of Mr. Frank K. Savage, Vice-President. Close technical liasson is maintained between the Graham-Savage Associates and The University of Michigan but contractual arrangements with the International Copper and Brass Research Association, Inc. are conducted independently. The first two phases of the research are being carried on at The University of Michigan with Prof. C. A. Siebert of the Department of Chemical and Metallurgical Engineering responsible for the metallurgical phase and Prof. R. B. Keller of the Department of Mechanical Engineering responsible for the radiator test program. Professor J. A. Clark of the Department of Mechanical Engineering is overall Project Director as well as having responsibility for the heat exchanger analysis and selection phase of the research. Mr. Richard D. Chapman, Director, Automotive Development, Copper and Brass Research Association, contributes to the general planning and assists in the coordination of the research between The University of Michigan, Graham-Savage and Associates, and the International Copper and Brass Research Association, Inc.

2. HEAT EXCHANGER ANALYSIS The selection of the type of heat exchanger matrix suitable for an automotive radiator begins with an examination of the performance characteristics of typical existing air-water heat exchanger cores, The determination of the performance characteristics depends on the availability of generalized baseic friction and heat transfer data of an experimental nature, Probably the most recent and comprehensive data of this kind presently available are those of Kays and London published in their book Compact Heat Exchangers.1 This study reports the heat transfer and friction characteristics of 88 different kinds of extended surface heat exchangers suitable for gas turbine (gas-to-gas) or automotive (air-to-water) application. From this group 17 different matrices were selected for study and their relative thermal and friction performance determined. An eighteenth core included for comparison is a McCord Corporation type "GN" Honeycomb core for which the heat transfer and friction data has to be estimated from the best available source, A brief presentation of the analytical procedures follows. The heat-transfer rate q in a heat exchanger may be expressed as follows q = (wcp) min (Thi-Tci)' (1) in which c is the heat transfer effectiveness.1 The remaining symbols are defined in the nomenclature of this report. In this analysis the relative performance of air-water heat exchangers will be determined for the following conditions: a. q is fixed b. Thi is fixed c. Tci is fixed d. (wcp)min is fixed e. (wcp)max is fixed. For the usual automobile radiator (wcp)min corresponds to the air-side and (wcp)max corresponds to the water-side. The effectiveness e is determined by the flow arrangement, i.e., counter flow, cross flow, etc., and the ratio (wcp)min/(wcp)maxol Hence, with the specifications a to e above, it is evident from Eq. (1) that e will be fixed for all possible matrix shapes for a heat exchanger of a given flow arrangement. As shown by Kays and London1 under these circumstances e will then be a function only of the NTU, known as the number of transfer units in a heat exchanger. Furthermore, the NTU is defined as

NTU = (2) (WCp) min Thus, we may now conclude that the relative performance of air-water heat exchangers may be determined on the basis of a fixed NTU which for the imposed restraints becomes (UA)1 = (UA)2 ~ (3) Now, if A is based on the air-side or finned-side of the exchanger, we have, 1 1 a 1 + — + (4) UA nochcA Awk TohhhAh For the usual automobile radiator the last two terms in Eq. (4) are negligible, i.e., the air-side controls, so, UA = lochcA. (5) Dropping the subscript c as all symbols now refer to the air-side, we have from Eqs. (3) and (5) (rloh A) 1 = (roh A)2 (6) where no = total surface temperature effectiveness. Hence, for purposes of comparison between various heat exchanger matrices at constant NTU, their relative volumes may be given as V, (UA/V)2 [(T]h)(A/V) ] V2 (UA/V) [(oh)(A/V) ]1 A similar formulation may be written in which a "reference heat exchanger" matrix is used for comparison between all the others. Designating this reference exchanger by the subscript o we have,

v _[(roh)(A/v)0 (8]o -- =. (8) Vo [('oh)(A/V) ] With this formulation, then, any other two matrices, say 4 and 8 may be compared on a volume basis by V4 (V/Vo)4 (9) V8 (V/Vo)8 Now, the heat transfer parameter, the Stanton number, the flow parameter, the Reynolds number are defined as St = - (10) GCp Re = 4rhG (11) so, ho StoGo (Sto) (w/Ac)o (12) h St G (St)(w/Ac) which for constant air-side flow w, becomes, ho Sto Ac — = ~- ~ ~ --— ~ (Z13) h St Aco Because the air-side flow is the same in all comparisons, the Prandtl number is constant. Basic heat transfer data are given in terms of St-Prl/3 Thus, Sto (St.Pr2/3)o St St. Pr2 /3 Hence, from Eq. (8) we have for a fixed AFR, the frontal area of the heat exchanger,

V ho(r) (St /Ac) (A/V)t ( o) (Sto (Ac/A) (A/V) o (ro) \St (Ac/AFR)o (A/V) (r0)0 (St ) (a\(~ (14) (o) St G/aC where Ac ( ) AFR a= = Ca (16) V and A VB Kays and Londonl report corresponding values of St, Re, a, and a for the first 17 of the 18 heat exchanger surfaces considered in this report. Since the total volume of the matrix is written as AFR. L, where L is the depth of the matrix in the air flow direction, Eqo (14) may also be written for fixed frontal area and total surface temperature effectiveness, as L _ Stt o/ V _ =Lo (S4O) ( aO) (a ) (17) The Reynolds number ratio is written for constant w and AFR from Eq. (11) as Re (rh/rho) Re0 (a/ao) For tube-fin matrices, a is uniquely specified by the particular matrix under consideration. In the case of plate-fin designs the ratio a depends on the thickness, 6 of the water channel and is computed from = (19) l+6/b

In these calculations 5 is taken to be 0o080 in. The above relationships permit the relative comparison between the heat transfer matrices from the standpoint of their thermal characteristics. A second important basis for comparison is the relative frictional pressure loss. Entrance and exit pressure losses are not included in this. The frictional pressure drop is expressed as p = L (w/Ac)2 f L 1 (W2 (20) AP = I. ( rh 2gop rh Ac =2gopJ Hence, for the various matrices having constant frontal area, the relative frictional pressure loss is p (f/fo) (L/L) (21) APo (rh/rho)(c /ao) The last type of comparison to be made combines both the relative heat transfer and the relative friction. This is on the basis of heat transfer per unit volume per unit pressure drop. On this basis a favorable exchanger is one which has a large value of this parameter. Thus defining, = qLV), (22) Ap we have the formulation of the relative value of this parameter as.L - (q/v)/ZAp'o (q/V) o/Apo 1.... J- --. e. (23) (L/Lo) ( Ap/Apo) Equation (23) also may be regarded as the relative value of the heat transfer rate per unit volume to frictional pumping power corresponding to the condictions specified. Comparison between the various matrices is made for the volwume ratio (and depth ratio), Eq. (17), frictional pressure drop ratio, Eq. (21), and heat transfer per unit volume per unit frictional pressure drop ratio, Eq. (23). All are based on a reference heat exchanger for reference conditions, defined below.

The 18 matrices considered here consist of six tube-fins, nine platefins, two tube-banks and one Honeycomb type. The geometry and basic friction and heat transfer data of the first 17 are given in Figs. 1-17, taken from Kays and London.1 The eighteenth is a McCord Corporation type "GN" Honeycomb replacement core having 1/4-in. square air passages, 2-1/4 in. long. Since basic heat transfer and friction data are not available for this matrix, its performance was estimated from data in Kays and London for a dimpled tube having approximately the same hydraiulic diameter as the selected Honeycomb core. 2.1. REFERENCE HEAT EXCHANGER AND REFERENCE CONDITIONS The reference heat exchanger is a plate-fin matrix having strip-fins and is the Kays and London surface designated as 1/8-15.2, indicating that it is made of 1/8 in. wide fins and has 15.2 fins/in. The width b of the air-flow channel is 0.414 in. In this report the reference heat exchanger is designated as surface 14 and its geometry and basic heat ransfer and friction characteristics are given in Fig. 14. In order to have a specific flow condition for which all comparisons may be made, the following are taken Air velocity = 10 ft/sec Air temperature = 100~F = 1.285 x 10-5 lbm/ft/sec p = 0.071 lbm/ft3 Hence, 4rho = 0.1042 in. = o.00oo868 ft Co = 1/(1+0o080/0.414) = 0,838 Po = 417 ft2/ft3 Co = oao = (417)(0o838) = 350 Reo = 4rhoGo/L = 4rho(w/Ac)o/A 4rhoPVAFR/g( aoAFR) = 4rho/ao (pV/4) = (o.oo868)(0.071o)(10) 105/(o838)(1.285) = 572. From Fig. 14 corresponding to Reo = 572 we find,

(St-Pr2/3) = 0.0155 and fo = 0.093 2.2. HEAT EXCHANGER MATRICES STUDIES A summary of the 18 heat exchanger matrices selected for study is given in Table 1. 2.3. HEAT TRANSFER AND FRICTION DATA The computed data for heat transfer and friction including the matrix parameters of a, a, and P and their relative values are summarized in Tables 2 and 3. 2.4. RESULTS AND CONCLUSIONS The relative volume or relative depth, V/Vo or L/Lo, relative friction pressure drop, Ap/Apo, and the relative heat transfer per unit volume per unit pressure drop, r/ro are plotted in Figs. 18-20. From these results it is possible to compare all of the 18 matrices according to volume, pressure drop, and heat transfer per unit volume per unit pressure drop. As is evident by inspection, matrix 7 has lowest pressure drop but largest volume. From Table 1 it will be noted that this matrix is of the plate-fin' type and very open with only 5.3 fins/in. and the second lowest area to volume ratio. The heat transfer per unit volume per unit pressure drop for this matrix is, however, rather poor, which is a result of its open design. From the standpoint of compactness matrix 12 is especially outstanding. This matrix has the smallest volume of them all (Fig. 18), and the highest heat transfer per unit volume per unit pressure drop (Fig. 20), although its frictional pressure drop is only moderately favorable (Fig. 19). The round tube banks (16 and 17) suffer from high pressure drop and high volume and consequently have very poor heat transfer per unit volume per unit pressure drop. The Honeycomb matrix, 18, also shows up poorly from the standpoint of compactness, Its pressure drop is second lowest but its heat transfer performance is not significant compared with the plate-fin designs. These results indicate the superiority of the plate-fin type of matrix from the standpoint of maximum heat transfer per unit volume, minimum weight, and probably minimum cost. The benefit of turbulence promoters obtained by deforming the fins (wavy) is evident. The penalty is, of course, increased pressure drop but compactness is gained. 9\

TABLE i Summary of Heat Exchanger Matrices* Hydi'aulic / Matrix y Classification Matrix Type Kavs and Ljonnon Fins/i Dianeter ft2/ft3 No.i Flat tubes, Flat. tube, plain 1 9.68-c.87 c. i16 2, contirnuous fi n s continuouas fin i U~~b s -ube, vu-~f fed.~~( a 2 1-'ub eseS Flat cue irfe 9.68-0.,(- 7R o. i2ll continiucu fins cont4nuous fir! Flat'C1.nDs Flat tube, 10/50a62 cortiriuouos ds t-in. )oti -iOu s`7'n Fiat t, fs, lat tue rJi Qle2 L be s - 1_?'i; -SK21 con4~ o. o ins contin u in Fliat tab!Fin4 t Flat; ube, r11- c'J f2ld -' C~~~~~~~~~~~~~~~~~~~~~~~~~- t7 SnR o.i0 7 Rowjan tue, Ruiotn tube) 6 3. c, - 5/8rr, kJ~~~~~3-. Ic ~ 17 C ont.-I._r_."IouS f 1 I s conitin-iaoi.s'L i te 1PrIrtn 7.b0~ 6 mu o.rii o 03~ 5~ 2 aIte fJn Plain tir d ii r~~~~~~~~~~~~~~~~~~~~~~~ plii Iins b 0. 4Ino pla in Tins 1 30u Plate fin, Wlain fin, plain fins b = 0.41 in./1 12 Plate fin, Wlain fin, 10 Plate'_` n.Y Y plain f- n s b =020A11 - 1-1 Plate f in. avy 1"in, 4 4- 8, -ii.4 0 I -17 wavy fins b = o.4i3 in. D Plate fin, Wavy fin, 17.8- 3/8W 7. o o836 strip fins b = 0o414 in./ Plate fin, Strip fin, 5152- 12. 2_ 0 1`45 strip fin 1m 0 414 in. /i 3 Plate fin: Strip fin, i4 ~~~~~~~~~~~~o.io42 ~ -f" fn -c.14 in. 1Plate fin. Louverei fin i Louvered fin b = 0 2b i. in.0. 214 16 Round tLe bank i- Staggered -1.30 -1.(S) 0. 9yO 80.5 17 Round tube bank Tn-lin -. -- 804 18 Honeycomb -. —- --.1O 192 *Rouand tubes with circcular fins not cons —ide`-red sn c fr a:Vaia be bs hatrn, dat, tubes are large than 3/S-in, din and area/'volicse ratio i L I ess on 170. Thes o c s i above urations.

TABLE 2 Heat Transfer Data Matrix 4rh rh/rho a a0/o a a a/%o Re/Reo Re StPr2/3 St/Sto L/Lo No... i 0.1416 1.36 0.697 o.833 229 229 0.655 1.635 935 0.0063 0.405 3.14 2 0.t416 1.36 O.697 0.833 229 229 O.655 1.635 935 0.0079 0.510 2.50 3 0.1656 1.59 O.788 0.941 224 224 o.640 1.690 967 0.0110 0.710 2.07 4 0.1622 1.55 0.788 0.941 228 228 o.654 1.65 945 0.0115 0.740 1.98 5 0.1382 1.33 O.780 O.930 270 270 0.771 1.43 820 O. 0110.710!.70 6 0.1430 1.37 0.534 0.637 179 179 0.511 2.15 1230 0.0095 0.611 2.04 7 0.242 2.32 O.855 1.02 188 161 O.460 2.28 1300 0.0059 O.380 5.83 8 0.1828 1.75 0.911 1.09 244 222 0.635 1.61 920 0.0061 0.384 4.48 9 0.1052 1.01 o.840 1.00 414 348 0.995 1.00 572 0.0085 0.547 1.83 10 0.0738 0.707 0.757 0.905 561 424 1.215 0.782 447 O.Oll0 0.710 1.05 11 0.1272 1.22 0.836 1.00 351 294 o.840 1.22 699 0.0170 1.10 1.07 12 o.0836 0.800 0.835 1.00 514 432 1.235 0.800 468 0.0172 1.11 0.730 13 0.1345 1.29 0.858 1.02 340 292 0.833 1.265 725 0.0175 1.13 1.090 14 0.1042 1 0.838 1 417 350 1 1 572 0.0155 1 1 15 0.1214 1.162 0.758 0.905 367 278 0.795 1.285 736 0.0140 0.905 1.26 16 0.1980 1.90 0.333 0.398 80.3 80.3 0.228 4.77 2730 0.0130 0.840 2.08 17 0.198 1.90 0.388 0.463 80.4 80.4 0.228 4.10 2350 0.0115 0.742 2.74 18 0.141 1.35 0.563 0.671 192 192 0.549 2.01 1150 0.00562 0.362 3.38

TABLE 3 Friction Data Matrix L o/A Matrix Re f f/fo L/Lo rh/rho a/aoPo /Ap No. & 1 935 0.023 0.247 3.14 1.36 0.833 0.685 2.15 0.465 2 935 0.035 0.377 2.50 1.36 0.833 0.832 2.08 0.480 3 967 0.036 0.387 2.07 1.59 0.941 0.535 1.108 0.g04 4 945 0.042 0.452 1.98 1.55 0.941 0.614 1.218 0.822 5 820 0.042 0.452 1.70 1.33 0.930 0.621 1.056 0.946 6 1230 0.028 0.301 2.04 1.37 0.637 0.704 1.435 0.696 7 1300 0.016 0.172 5.83 2.32 1.02 0.423 2.465 0.405 8 920 0.023 0.247 4.48 1.75 1.09 0.633 2.835 0.353 9 572 o.o36 0.387 1.83 1.01 1.00 0.705 1.290 0.775 10 447 0.041 0.441 1.05 0.707 0.905 0.725 0.741 1.350 11 699 0.092 0.990 1.07 1.22 1.00 0.869 0.950 1.075 12 468 o.o83 0.893 0.730 o.800 1.00 0.815 0.595 1.680 13 725 0.100 1.075 1.090 1.29 1.02 0.891 0.971 1.030 14 572 0.093 1 1 1 1 1 1 15 736 0.070 0.753 1.26 1.162 0.905 0.900 1.135 0.880 16 2730 0.071 0.764 2.08 1.90 0.398 2.10 4.19 0.239 17 2350 0.053 0.570 2.74 1.90 0,.463 1.78 4.88 0.205 18 1150 0.0125 0.1345 3.38 1.35 0.671 0.501 1.70 0.587

It was on the basis of these calculations that the matrix for the first electroformed model heat exchanger was selected. While matrix 12 is the most compact of the group it has a fairly high number of fins/inch-1708. It was felt that for the first electroformed model the number of fins/inch ought not to exceed about 10 to reduce the electroforming problems and to have a design which is similar to current automotive practice. The model should however be of a plate-fin design with waviness in the fins, if possible. 2.5. FIRST ELECTROFORMED MODEL HEAT EXCHANGER In view of the results of the study outlined above it was decided to construct the first electroformed radiator of a plate-fin design, To accomplish this in the most direct manner as quickly as possible, existing spacer stock deformed to provide waviness in the fins was obtained and a design established in which water channels would be electroformed at the fin roots. Mr. Richard D. Chapman secured the spacer stock. While it is recognized that such a design is not a completely electroformed heat exchanger, the critical points in the matrix, namely the joints, are electroformed. Undoubtably some development work will be necessary even to successfully accomplish this construction. When this is done, attention may then be given to more complete electroforming manufacture, should such appear desirable. When the spacerwater channel subassemblies are formed they will be soldered together and then joined to headers at each end. This design is shown in Fig. 21. As a check on the thermal effectiveness of the electroformed joints in the first model, a second heat exchanger was planned which would be similar in all respects to the first except that the fins would be soldered to the water channels. The heat transfer performance of both exchangers would be determined experimentally in the test loop and wind tunnel. 13

I I I

3. RADIATOR HEAT TRANSFER PERFORMANCE TEST APPARATUS 3.1 INTRODUCTION In order to determine experimentally the actual heat transfer performance of various radiator configurations, a heat transfer test loop was designed. Major emphasis was placed upon flexibility of operation: the capability for testing both scale models and full size automotive radiators over a range of vehicle speeds, air-side pressure drops, heat- transfer rates, and heating water temperatures comparable to automotive application. The general configuration required to satisfy these requirements is shown schematically in Fig. 22 and consists of a heat transfer apparatus to supply a constant flow of water and heat transfer rate to the radiator, a wind tunnel to remove this heat from the radiator by a flow of cooling air through the radiator core at appropriate velocities and suitable instrumentation and controls. 3.2. HEAT TRANSFER APPARATUS The heat transfer apparatus consists of a low-pressure steam supply used to heat the water, a bypass cooler, an accumulator, reservoir, pump, test radiator core controls, flow meter, and other instrumentation. The selection of water instead of steam as the hot-side heat transfer medium to the radiator was based on two considerations: first, water represents the fluid used in an automotive radiator; and second, the measurement of the heat transfer rate from water-wide temperature drop appeared to offer a convenient and reliable method for this determination. Radiator performance tests using steam are difficult to perform since the measurement of the heat transfer rate must be obtained by separating the liquid condensate flow from the vapor flow at the radiator exit. The principal problems associated with the use of water as the heat transfer medium is the accurate measurement of low temperature difference between the inlet and outlet water temperatures and the determination of precise adiabatic mixed mean temperatures at these points. These problems have been solved by using a thermopile at the inlet and outlet and a baffled flow mixer such that a true mixed mean temperature can be obtained. Low-pressure steam is, used in a shell and tube heat exchanger to heat the water in which the steam flow can be controlled either manually or automatically. The automatic controller takes the radiator water inlet thermocouple signal as an input and produces a pneumatic output signal to a pneumatically actuated control valve, A centrifugal pump is used to supply the pressure losses in the system. Control of the flow through the system is by means of a pump bypass, a pump discharge throttle valve, and a vernier bypass valve around the main control valve~ 15

An accumulator is installed near the radiator inlet to maintain a given pressure level at the radiator inlet. An air regulator with constant bleed is used to pressurize the accumulator bag. The use of an accumulator to keep the water and the pressurizing gas separate permits the entire system to be free of entrapped air and also reduces the air dissolved into the water. A bypass cooler in parallel with the test radiator permits the system to be operated without the test radiator or wind tunnel in the circuit. The bypass cooler may also be useful in stabilizing the system when tests are run at low mass flow rates and low heat transfer rates to the test radiator. Copper lines and fittings are used throughout for ease of assembly and freedom from corrosion. Flexible hoses and swing joints are used to make radiator connections. 3.3. INSTRUMENTATION FOR HEAT TRANSFER APPARATUS Water temperatures are measured by thermopiles (thermocouples connected in series) positioned in a flow-mixing baffle at both the radiator inlet and outlet stations to provide true mixed mean water temperatures. All temperatures will be continuously monitored and recorded on a self-balancing recording potentiometer. Radiator inlet water pressure is measured with a bourdon tube gage, and radiator pressure drop with a manometer. Water volume flow rate is measured with a rotameter. A turbine flow meter is provided for an independent check. A typical set of automotive radiator performance conditions is given in Table 4 and may be compared with the design capabilities of the test loop shown in Table 5. TABLE 4 Typical Maximum Automotive Radiator Performance Water flow rate 300 lbm/min Air flow rate 1760 ft3/min Air inlet temperature 94~F Air outlet temperature 151~F Water inlet temperature 1840~F Water outlet temperature 190~F Heat transfer rate 1800 Btu/min 16

TABLE 5 Heat Transfer Loop Design Conditions Heat transfer rate 100 hp or 4200 Btu/min Water temperature in 190~F Water temperature out 180'F Water flow rate 50 gpm Inlet water pressure 30 psia Inlet steam temperature 350~F Inlet steam pressure 50 psig 3.4. WIND TUNNEL The wind tunnel consists of a contraction section, test section, diffuser, fan, and discharge ducting as shown in Fig. 23. The wind tunnel geometry and fan size were determined by the combination of the maximum simulated vehicle velocity, maximum radiator pressure loss, and maximum size of test radiator. The minimum operating conditions were determined by the minimum vehicle or inlet air velocity, and minimum test radiator size. These two limits, the maximum and minimum conditions, then provided the operating range and the flexibility requirements of the tunnel. Three flow control methods were used to meet the air flow requirements of the tunnel: (1) a 2:1 fan speed reduction by means of switching the voltage of the electric motor which reduces fan air flow by this same ratio; (2) variable inlet guide vane geometry capable of producing a continuous variation in fan air flow from 15% to 100% of maximum air flow; and (3) a bypass around the test section which can produce a continuous variation in test section air flow from 10 to 100% of maximum test section air flow. The change in fan speed and the inlet guide vane geometry produce changes in fan characteristics while the tunnel bypass alters the total system flow characteristic to vary the test section air flow. The bypass also allows the fan to operate in a surge-free region at all times. The bypass is accomplished by constructing the contraction section and the test section as a single unit and then translating this unit forward relative to the fixed diffuser and fan. The bypass flow characteristic is shown in Fig. 24, and the combination of all three methods of flow control are shown in Fig. 25. Exhaust air from the tunnel which is some 50'F above inlet or ambient air temperature is ducted downward through an opening in the second floor to the first floor area of the laboratory. This will eliminate the problems of re-ingestion of heated air and insure a nearly constant radiator inlet air temperature. 17

Photographs of the wind tunnel assembly, the test section region with removable, viewing, and access doors, and the bypass are shown in Figs. 26-28. 3.5. INSTRUMENTATION FOR WIND TUNNEL A fairly elaborate pressure and temperature survey grid will be used for a detailed initial calibration and check-out. This unit will also be available for determination of local radiator performance. Final instrumentation after check-out will consist of eight total-static pressure tubes, four upstream, and four downstream and located so that the average flow conditions on both sides of the radiator. The design conditions for which the wind tunnel is capable of meeting are given in Table 6. TABLE 6 Wind Tunnel Design Conditions Maximum radiator inlet velocity-equivalent to a vehicle speed of 100 mph 45 mph Radiator pressure drop-at max tunnel velocity 5.1 in. H20 Total tunnel losses (incl radiator loss) 6.o in. H20 Maximum radiator frontal dimensions 18 x 18 in. Minimum radiator frontal dimensions 6 x 6 in. Test section flow dimensions 24 x 24 in. Inlet station of contraction section dimensions 72 x 72 in. Fan volume flow rate (at 1580 rpm, 7 in. H20) 11,400 cfm Fan drive motor, 440 V AC, 3 phase 25 hp Construction and assembly of the basic wind tunnel was completed on April 1, 1963. The initial testing of its flow characteristics and the installation of the heat transfer loop and associated instrumentation is expected to be completed by the first part of May. 18

4. METALLURGICAL STUDIES 4.1. LITERATURE SURVEY OF SOLDERS A limited literature survey was made to obtain selected data on the mechanical properties of solders and soldered joints. The properties included in this report are: tensile strength, shear strength, creep, and fatigue. 4.1.1. Tensile and Shear Strengths Gonser and Heath2 prepared tensile bars by casting solder into a split steel mold. The specimens were 7-5/16 in. long with a 2-in. parallel central section 3/8-in. in diam. The bars were annealed for 16 hr at 1000C. The tensile strengths were determined using a constant head speed of 0.5 in./min. The results are given in Table 7. TABLE 7 Composition Tensile Strength, psi 15% Sn, 0.25% Ag, 0.015% Bi, balance Pb 5,680 20% Sn, 0.25% Ag, 0.015% Bi, balance Pb 5,800 30% Sn, 0.25% Ag, 0.012% Bi, balance Pb 5,990 40% Sn, 0.25% Ag, 0.010% Bi, balance Pb 6,250 50% Sn, 0.25% Ag, 0.010% Bi, balance Pb 6,090 63% Sn, 0.25% Ag, 0.010l Bi, balance Pb 7,490 Thompson3 prepared standard 0.505 test bars by casting solder into a split steel mold. The bars were tested in the as-cast condition using a head speed of 0.5 ipm. Their results are given in Table 8. 19

TABLE 8 Composition Tensile Strength, psi 15% Sn, 85% Pb 5,270 20% Sn, 80% Pb 5,730 30% Sn, 70% Pb 6,810 40o Sn, 60%o Pb 6,890 50o Sn, 50% Pb 6,400 Table 9 gives the results reported by Turkus and Smith4 They did not give the details of their tests. TABLE 9 Composition Tensile Strength, psi 20% Sn, 8o% Pb 4,940 30% Sn, 70% Pb 5,390 40o Sn, 60o Pb 5,660 20% Sn, 2% Ag, 78%o Pb 5,620 20% Sn, 1.5% Ag, 3% Bi, 74.85% Pb 8,120 30% Sn, 1% Ag, 69% Pb 8,810 Rhines and Anderson5 prepared specimens by joining 0.75 in. copper rods which had been faced at the joining ends. A gap of 0.005 in. was used and the bars were heated to 60~C above the liquids of the solder, fluxed, soldered, and cooled. Their results are given in Table 10. TABLE 10 Composition Tensile Strength, psi 15% Sn, 85% Pb 13,300 33- Sn, 67%o Pb 17,100 40% Sn, 60o Pb 14,100 p0o Sn, 50% Pb 23,900 63% Sn, 37% Pb 29,000 20

Gonser and Heath2 prepared lap joints using 70 copper, 30 zinc brass. Two methods of joining were used: (1) the closed method, i.e., maintaining a constant gap of 0.007 in.; and (2) the open method, i.e., the gap open to 0.030 in., solder applied, and then the gap closed to 0.007 in. Compensation was made for the offset character of the specimen and the head speed of the machine was 0.5 ipm. Their results are given in Table 11. TABLE 11 Tensile Strength, psi Composition Open System Closed System 15% Sn, 0.25% Ag, 0.015% Bi, balance Pb 4,265 4,050 20% Sn, 0.25% Ag, 0.015% Bi, balance Pb 4,490 4,880 30% Sn, 0.25% Ag, 0.012% Bi, balance Pb 5,450 4,790 40% Sn, 0.25% Ag, 0.010% Bi, balance Pb 5,550 5,250 50% Sn, 0.25% Ag, 0.010% Bi, balance Pb 5,750 5,240 63C Sn, 0.25% Ag, 0.010% Bi, Balance Pb 6,170 5,750 (Maximum value of each alloy was taken) Turkus and Smith4 results on lapped joints are given in Table 12. The details of their test procedure was not given. TABLE 12 Composition Tensile Strength, psi 20% Sn, 80% Pb 5,680 30% Sn, 70% Pb 5,770 40% Sn, 60% Pb 6,270 20% Sn, 2% Ag, 78% Pb 5,550 30% Sn, 1% Ag, 69% Pb 5,620 Gonser and Heath2 performed shear tests on bulk solder samples. Their results are given in Table 13. 21

TABLE 13 Composition Shear Strength, psi 15% Sn, 0.25% Ag, 0.015% Bi, balance Pb 4,470 20% Sn, 0.25% Ag, 0.015% Bi, balance Pb 4,740 30% Sn, 0.25% Ag, 0.012% Bi, balance Pb 5,500 40% Sn, 0.25% Ag, 0.010% Bi, balance Pb 5,680 50% Sn, 0.25% Ag, 0.010% Bi, balance Pb 5,870 63% Sn, 0.25% Ag, 0.010% Bi, balance Pb 6,060 Russell and Mack6 reported the following shear strengths on bulk solder samples: 15% Sn, 85% Pb 4,280 psi 40% Sn, 60% Pb 4,900 psi Rhines and Anderson5 used butt soldered joints using 0.75 in. round copper bars which had been faced at the joining ends. The shear strengths were determined using a standard torsion test. Their results are given in Table 14. TABLE 14 Composition Shear Strength, psi 15% Sn, 85% Pb 5,640 33% Sn, 67% Pb 6,450 40% Sn, 60o Pb 8,280 50% Sn, 50% Pb 7,580 63% Sn, 37% Pb 8,ooo 4.1.2. Creep Properties Baker7 studied the creep properties of cast solders using specimen 0.564 in. in diam and 4 in. long in the parallel section. The results shown in Table 15 give the stress to produce a strain of 1 x 10-4 per day.

TABLE 15 At Room Composition Temperature, psi At 80~C, psi 30.4% Sn, balance Pb 115 39 49.5% Sn, balance Pb 125 28 62.2% Sn, balance Pb 335 68 40o Sn, 2. o0 Sb, 0.10% Bi, 0.011% Ag, balance Pb 420 60.2% Sn, 0.24% Sb, 0.015% Bi, 0.018% Ag, balance Pb 800 62.2% Sn, 0.002% Bi, balance Pb 450 45 54.5% Sn, 3.6% Sb, 0.003% Bi, balance Pb 1,030 110 Baker7 prepared lap joints using the "open method" described by Gonser and Heath2 with a final gap of 0.006 in. The results reported in Table 16 are the 500-day stress-rupture strengths. TABLE 16 At Room 40% Sn, 2n Sb, Balance Pb Temperature psi At 800C, psi For steel Joints 325 120 For copper Joints 390 120 For brass Joints 470 120 4.1.3. Fatigue Properties McKeown8 conducted fatigue tests on lap Joints. The fatigue machine was designed to apply alternating shear stresses to the specimen. Fatigue damage was determined by obtaining the decrease in tensile strength of a specimen after sustaining 3,000,000 cycles in the fatigue test. The results shown in Tables 17 and 18 give the maximum stress level which did not produce a decrease in the tensile strength. The mean stress used for the tests of Table 17 was 600 psi, and for the tests of Table 18 it was 900 psi.

TABLE 17 Composition Maximum Alternating Stress, psi 63% Sn, balance Pb 400 50.4% Sn, balance Pb 400 31.5% Sn, balance Pb 360 18.9% Sn, balance Pb 350 TABLE 18 Composition Maximum Alternating Stress, psi 63% Sn, 37% Pb 200 56% Sn, 3.2% Sb, balance Pb 310 30% Sn, 1.0% Sb, balance Pb 260 4.1. 4. Summary A limited literature survey on the mechanical properties of solders and soldered joints indicated tensile strengths of 5,000 to 8,000 psi for bulk solders and 13,000 to 29,000 psi on butt soldered joints depending on the composition of the solder. The fatigue strength of lap Joint soldered interfaces was reported as 1000-1100 psi to produce damage in 3,000,000 cycles. 4.2. LITERATURE SURVEY OF THE MECHANICAL AND PHYSICAL PROPERTIES OF ELECTROFORMED COPPER The majority of properties of electroformed copper have been determined on deposits from copper sulfate baths. The mechanical properties are either taken from standard tensile tests or from a hydraulic bulge test which is a modified Olsen cup ductility test. The tensile strength of electrodeposited copper varies from 17,000 to 90,000 psi depending upon the plating conditions. In general the strength of the copper is related to the grain size and structure, being high for the finer grained copper deposits. There does not appear to be any correlation between the tensile strength and the hardness of the deposit, or the tensile strength and ductility as measured by percent elongation in the tensile test. Decreasing the bath temperature or increasing the current density appear to increase the tensile strength of the deposit, but specific additives to the bath in most cases have a greater effect on increasing the strength. Res24

idual stresses are usually low in electrodeposited copper. However, plating conditions are reported where the tension stresses have been found to be as high as 21,000 psi. The following summary of published papers gives specific details on the mechanical and physical properties of electrodeposited copper. Bennett9 deposited copper on a rotating cathode from a solution composed of 20% Cu S04-5H20 and 12% H2S04. The cathode was an aluminum pipe 1 in. O.D. and 5.5 in. long. After the copper was deposited, the pipe was placed in a lathe and a section 1 in. wide and 2 in. from one end was turned down to a uniform thickness ranging from 0.040 to 0.060 in. The pipe was then cut into longitudinal sections in a milling machine, separated from the aluminum and pulled in an Olsen tensile testing machine. The strength reported was the average of five or more tests. The effect of speed of rotation, tempera ture of the bath, and current density are shown in Tables 19-21. TABLE 19 Current density 500 amp/sq ft Initial temperature 350C Tensile Strength, RPM lb/sq in. 1,750 37,000 2,500 49,000 3,500 51,000 5,500 58,000 TABLE 20 Initial temperature 20~C 2,500 rpm Tensile Strength, Amp/sq ft lb/sq in. 300 6o,oo000 400 68,000 510 40,000 1,100 35,000 1, 700 14,000 25

TABLE 21 Initial temperature 50~C 5,500 rpm Tensile Strength, Amp/sq ft lb/sq in. 34o 34,000 500 50,000 1,000 41,000 1,600 32,000 2,400 28,000 4,000 13,000 At a constant current density of 500 amps/sq ft, increasing the speed of rotation resulted in an increase of tensile strength. At constant speed of rotation, increasing the current density first results in an increase in strength, then reaches a maximum, and finally decreases. Variations in the bath composition has little, if any, affect on the strength of the deposite as shown in Table 22. TABLE 22 5,500 rpm Current density 500 amp/sq ft Tensile Strength, % CuSO' 5H20 H2SO4 lb/sq in. 12 15 60,000 20 15 58,ooo 25 15 55,o000 15 12 60,000 15 25 57,000 Increasing the temperature from 25~C to 75~C has a pronounced affect on the strength as shown in Table 23.

TABLE 23 5,500 rpm Current density amp/sq ft Tensile Strength, Temp., ~C lb/sq in. 25 63,ooo 50 49,000 75 30,000 The author states that at current densities higher than 500 amps/sq ft no attempt was made to control the temperature of the bath, and it is his opinion that the decrease in strength at the high current densities noted in Tables 20 and 21 is due to an increase in temperature brought about by the plating conditions. Sonoda10 determined the strength of copper deposited as a sheet 120 cm x 30 cm and in varying thicknesses. The details of the electrodeposition process are not given. The author states that the properties varied from sheet to sheet, and even on sections of the same sheet. This is undoubtedly due to variations in thickness as he obtained the average thickness from weight measurements. The results are given in Table 24. TABLE 24 Nominal Thickness, Tensile Strength, Elongation, g/sq cm kg/sq. mm 0.671 24.9 26 0.610 26.8 -- 0.549 23.8 36 o.488 24.7 39 0.427 25.6 35 o.366 26.5 33 0.305 24.9 35 Shakespear11 discussed the Anaconda commercial process for producing electrosheet copper and stated that the tensile strength varies from 30,000 to 40,000 psi, and elongation from 15 to 25*. 27

Altmeyerl2 reported a tensile strength of 39,aoo000 psi and an elongation of 34%. These results are for copper deposited on a cathode, 68 cm in diam and 4 m long, rotating at 30-40 rpm. Huessner, Balden, and Morse13 discussed the effect of grain size and structures on the mechanical properties of electrodeposits. They include eight photomicrographs of copper deposited from baths of various compositions. Their mechanical property data are given in Table 25. TABLE 25 Mechanical Properties of Acid Copper Deposits Tensile Elongation Hardness Addition Agent Strength, psi * in 2 in. V.H.N. None 36, 150 22 81 Molasses 33,000 21 81 Molasses and Thiourea 80,280 3 170 It can be seen that the addition of Thiourea to the bath produced a marked increase in the tensile strength and a marked decrease in ductility due to the fine grain structure resulting from the addition agent. Prater and Readl4 used a hydraulic bulge test to determine the mechanical properties of copper. The bath consisted of 45 gm/l of copper and 200 gm/l of H2S04. Glue was used as an addition agent. The sheet material was prepared by the Anaconda Process as described by Shakespear.ll The average thickness was determined by weighing the test piece and calculating it, using 8.9 g/cc for the density of copper. Seven or more determinations were made for each thickness. The tensile strength of the 0.66 mil copper varied from 46,100 to 50,300 with an average of 48,000 psi. The dutility varied from 0.5 to 0.9%. The tensile strength of the 3.6 mil copper varied from 38,100 to 46,000 with an average of 41,000 psi. The ductility varied from 1.9 to 2.9%. It should be pointed out that the ductility measured in the bulge test are not related to ductility as measured in the tensile test. Fedotev and Pozinl5 studied copper deposited from a bath containing 250 g/l Cu S04*5H20, 50 g/l H2S04, at a current density of 1 amp/sq dm and a temperature of 18~C. The cathode was a stainless steel plate 115 x 50 mm. Eight determinations were made for each thickness and the minimum, maximum, and average strengths are as follows: 28

Tensile Strength, psi Thickne s s 25j 504 75l Minimum 28,400 29,800 30,700 Maximum 33,000 33,200 32,300 Average 31,200 1, 200 31,400 Rochelle salt as a bath additive has the following affect on the tensile strength of the deposit. Rochelle Salt, 0 0.025 0.1 0.2 1.0 3.0 5.0 g/l Tensile Strength, 33,500 12,100 12,100 11,900 8,050 1,690 1,240 psi Struyk and Carlson6 determined the tensile strength of copper deposited from fluoborate baths of various compositions. The current density used was 300 amp/sq ft, and the thickness of the deposits was 0.020 in. Their results are given in Table 26. The deposit from a bath where the copper concentration is 120 g/l has a much higher tensile strength than one from a bath of 60 g/l of copper. However, the addition of 1.2 ml/l of molasses, or 2 g/l of Dacolyte to the 60 g/l copper bath resulted in producing deposits of comparable tensile strengths to those from the 120 g/l bath. Suchl7 indicates maximum tensile strengths of 43,000 psi and 60,000 psi may be obtained from Cyanide, and pyrophosphate baths respectively. However, the data are very meager and should be considered as only a rough approximation. Read and Whalenl8 studied the behavior of electroformed copper under alternating stresses. The copper was deposited from a bath containing 230 g/l Cu S0451I20, 50 g/l H2S04 using a current density of 10 amp/sq ft. The specimens were not machined on the solution side of the specimen prior to testing in a Krouse bending fatigue machine. The thickness was 0.025 in. The tensile strength of the copper was 26,000 psi. Their results are given in Table 27. At a stress level of 9,5Q0 psi there is a range of cycles to failure from 5.1 x 104 to 1.4 x 107. This is undoubtedly due to surface irregularites and possibly non-uniform thickness of the specimen. Barklie and Daviesl9 report a residual stress of essentially zero in copper deposited from a bath containing 200 g/l Cu S04o5HF0, 6o g/l H2S04, when plated at 35~C and a current density of 30 amp/sq ft. 29

TABLE 26 Physical Properties of Deposits Fluoboric Yield Tensile Rockwell, Copper, Yield Tensile Elongation Rockwell, Acid, C Point, Strength 2 Hardness g/l g/l psi psin in. 120 30 95 -- 30,600 3.5 68-74 -- 22,500 2.5 28,000 34,500 3.5 Avg 28,000 32,500 3.2 120 30 120 19,850 30,100 14.0 59-63 19,200 29,100 14.0 21,750 29,400 15 Avg 20,230 29,500 14.5 6o 4 120 12,800 17,000 8.o 44-45 13,200 17,200 7.5 12,700 15, 700 6.5 Avg 12,900 17,100 7.3 60* 4 120 22,600 30,800 11.0 64-70 23,600 28,700 -- Avg 23,100 29,800 11.0 60** 4 120 22,750 30,300 9.0 57-57.5 (16,400) (23,100) (5.0) 20,200 30,300 12.0 Avg 21,480 30,300 10.5 *With 1.2 ml/1 molasses. **With 2 g/l Dacolyte. 3o

TABLE 27 Stress, psi Cycles to Failure 10,800 3.9 x 10 10,800 3.3 x 10 10,800 6.6 x 10 9,500 5.4 x 10 9,500 4.2 x 10 9,500 5.1 x 10 9,500 1.1 x 10 9,500 4.0 x 10 9,500 3.0.x 10 Phillips and Clifton20 reported residual stress measurements of copper deposited from various types of baths as shown in Table 28. TABLE 28 Test Temp., C.D. NType of Bath Composition pH OF asf No. OF asf 1 Acid copper CuSO, 32 oz/gal 0.9 Room 30 H SO, 4 oz/gal 2 Copper cyanide NaCN, 9 oz/gal 12.8 125 15 CuCN, 6 oz/gal Ha CO, 2 oz/gal 3 Copper "K" Recommended make-up 13.5 180 15 4 Copper "L" Recommended make-up 140 15 Test Thickness, Change of Calculated Stress,* in. Deflection, psi (wt area/den) in. Steel Base Bronze Base 1 0.0010 0.0002 1400 0 2 0.0005 o.ooo8 9900 7200 3 0.0005 0 0 4 0.00154 0.0015 5200 *The stress values are all positive or tension stresses. 31

Graham and Lloyd21 determined stress values for copper deposited from alkaline cyanide baths. The standard bath consisted of 4 oz/gal copper, 0.8 oz/gal rochelle salt, and 4 oz/gal sodium carbonate to which a number of variables were introduced as shown in Table 29. It can be seen that the residual stress values vary from 15,400 psi in tension to 5,000 psi in compression. TABLE 29 Test Cathode Stress in No. Coefficient, Deposit, Variable Studied *% 1000 psi 1 87.3 8.7 Standard 2 53.7 11.6 Current reversal 3 73.0 14.7 cd, 40 asf 4 82.5 11.5 Temperature 130~F 5 91.0 5.8 Temperature 180~F 6 95.0 6.4 No rochelle salt 7 78.1 9.0 Na CO, 9 oz/gal 8 86.o 9.0 Free NaCN, 1.6 oz/gal 9 87.0 8.6 NaOH to pH 13.0 10 88.4 9.7 Copper, 2.5 oz/gal 11 88.0 8.o Copper, 5.0 oz/gal 12 88.0 8.4 CaCO, 100 ppm 13 86.1 8.6 K Fe(CN), 0.5 g/1 14 85.2 8.8 K Fe(CN), 1.0 g/l 15 87.2 11.3 Lead, 2 ppm 16 90.3 15.4 Lead, 75 ppm 17 90.0 - 4.0 KCNS, 2 oz/gal 18 95.3 - 4.7 KCNS, 2 oz/gal, no rochelle salt 19 98.5 3.1 A 20 98.9 3.6 A, lower pH 21 92.8 7.8 A, lower temperature 22 93.0 11.1 B 23 95.2 - 5.0 C 24 88.0 - 3.8 D Fisher, Huhse, and Pawlek22 studied the effect of gelatin additions on the residual stress in copper deposits. The bath was 1N CuS04 and 1N H2S04. Increasing the gelatin content from zero to 0.1 g/l resulted in a gradual increase in the residual stress from approximately zero to 21,000 psi in tension. A further increase in gelatin content resulted in a gradual decrease in the residual stress, reaching zero at 0.2 g/l and a compressive stress of 1,400 psi at 0.25 g/l of gelatin. 32

Read and Graham23 determined the elastic modulus of electro-deposited copper using the sonic technique. Thin wall tubes approximately 0.4 in. in diam and 4 to 6 in. long were prepared by plating on a steel mandrel coated with a thin layer of a low melting alloy. Their results are given in Table 30. TABLE 350 Elastic Moduli of Cu Deposits Compared to Modulus for Drawn Cu Tubing 18.1 + 0.1 x 106 psi As-Plated Surface Machined Surface Current -......... Elastic Deviation Elastic Deviation Plating Bath Density, asf Modulus, from Ref, Modulus, from Ref, asf psi x 106 psi x 106 Purified acid 10 13.9 16.2 Cu sulfate 13.9 16.1 14.0 16.0 14.2 16.1 (avg) (14.0 + 0.2) 22.7 (16.1 + 0.1) 11.0 Impure acid 10 15.7 17.1 Cu sulfate 15.8 16.9 16.2 16.9 (avg) (15.9 + 0.3) 12.2 (16.9 + 0.1) 6.6 DuPont's P. R. 40 16.1 16.8 Cu cyanide 16.0 16.9 15.8 16.8 (avg) (15.9 + 0.2) 12.2 (16.8 + 0.1) 7.2 A considerable difference exists in the modulus value reported for the three different baths investigated in the as-plated condition, but not a very significant one on the machined surfaces. We may raise the question as to whether or not the differences noted in the as-plated condition are not due to nonuniform thickness, since thickness enters into the calculation of modulus when using the sonic technique for its measurement. Hinnert and Krider24 published the following expansion data for electrolytic copper: 33

Temperature Average Coefficient Range, of Expansion, ~C ~C x 10 20- 60 16.6 20-100 16.8 20-200 17.3 20- 300 17.7 20-400 17* *From Esser and Eusterbrock, Archiv Eisenhuttenwesen, 14 (1941), 341. 4.2.1. Summary Tensile strength data on electroformed copper reported in the literature vary from 17,000 to 90,000 psi. The strength is a function of the bath composition and plating conditions. The hardness of the plate does not have any correlation with the strength or ductility of the plate. In general, reducing the temperature or increasing the current density increases the strength of the copper. This is probably related to the fact that the grain size of the copper is decreased, which increases the strength. 4.3. TENSILE AND FATIGUE TEST ON ELECTROFORMED COPPER Tensile tests were performed on copper which had been electroformed at the Savage Rowe Company in Kalamazoo. It is our understanding that the copper was deposited under the following conditions: Temperature: 115 ~F Current Density: 100 amp/sq ft Agitation: Air Bath Composition: Copper fluoborate 443 g/1 Copper metal 120 g/l Free fluoboric acid 45 g/l Additives None The test samples were standard 0.5 in. wide strips and were run on an Instron self-aligning machine. The fracturing of the specimens occurred in an irregular manner in that part of the fracture appeared to be brittle and part of it appeared to be ductile. The ductility as measured on a fractured specimen was erratic due to the nature of the fracture. It was therefore decided to use "uniform elongation" as a measure of ductility, i.e., the percentage of elongation in two inches at the point where the test specimen reaches the maximum load.

The tensile tests on the first shipment of copper in December, 1962, indicated a wide scatter in the strength ranging from 30,000 to 39,000 psi. The results of the tensile test on the second shipment of copper in January, 1963, are somewhat lower but more uniform than the results on the first shipment of copper. Since fatigue tests are being run on the second shipment of copper, only the tensile test results on this shipment of copper are included in this report. They are somewhat lower than the values obtained on the first shipment of copper and no explanation has been found for this difference since the plating conditions were supposedly the same. The thickness of the sheet material supplied varied from one end of the sheet to the other by as much as 0.007 in. The variation in thickness of a 2-in. gage length was as much as 0.003 in. However the point of fracture was not always at the point of minimum thickness. Therefore a thickness gradient was determined for each test piece and the cross-sectional area was computed at the point of fracture. The results of the tensile test are: Tensile Uniform Specimen Thickness, Strength, Elongation No. in. psi in 2 in. Fl 0.0249 27,550 7.8 F2 0.0260 27,690 7.8 F3 0.0265 27,470 7.8 F4 0.0269 26,915 7.0 F5 0.0263 26,995 8.2 Avg 27,324 7.7 G2 0.0106 27,735 9.0 G3 0.0109 26,970 8.2 G4 0.0118 28,475 11.0 G3 0.0131 28,090 7.3 Avg 27,818 9.0 The microstructure of the copper is shown in Figs. 29-32. The specimens for microscopic examination were taken from the tensile test bars. The microstructure is a typical columnar structure of electro-deposited metals from baths that do not contain addition agents to produce a fine grain structure. It was originally intended to use a Sonntag constant-load fatigue-testing machine to evaluate the properties of the electroformed copper under alternating stresses. However the design of the machine did not permit the use of loads below two pounds and the stress levels needed required loads of 0.5 pounds or less. It was therefore decided to use a constant deflection testing 35

machine, and this caused a few weeks delay in starting the fatigue program as it was necessary to obtain a Krouse testing machine, which is shown in Fig. 33. This machine was purchased from funds supplied by the University and made available to this research at no cost. The thickness of the electroformed copper varied over the length and width of the test specimen. The range in variation of thickness was from a few ten thousandths to almost two thousandths of an inch. The average thickness was used in computing the load necessary to produce a given stress in the test specimen. The Krouse machine is designed to permit a desired weight to be suspended from the crank end of the specimen and electrical contact established at the deflection produced by the weight. The specimen is then attached to the crank and the eccentric on the crank adjusted until electrical contact is again established, which indicates that the specimen is again deflected to the same extent as it was when the weight was employed. This procedure eliminates the necessity of having to know the modulus of elasticity of the material. Variations in thickness, surface condition, etc., enhance the scatter that is normally encountered in fatigue test results. The data obtained to date cycling around zero mean stress are as follows: Average Specimen Thickness, Stress, Number of Cycles No. in. psi to Failure 4 0.020 12,000 10x106* 33 0.024 12,000 1.28x106 6 0.024 12,000 1.06x106 5 0.021 12,000 0.54x106 2 0.024 11,500 1.6x106 3 0.025 11,500 0.9 —2.2x106** 7 0.023 11,000 9.8x106 9 0.022 11,000 4.9x106 18 0.021 11,000 10x106 13 0.026 11,000 5.4x106 1 0.026 10,000 12x106 *Did not fail. **Broke during night. Machine did not shut off. From the limited data available it would appear that the endurance limit of this electroformed copper is somewhere between 10,000 and 11,000 psi.

REFERENCES 1. W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif,, 1955. 2. B. W. Gonser and C. M. Heath, "Physical Properties of Soft Solders and the Strength of Soldered Joints," AIME Transactions, 122, p. 349 (1936). 3. J. G. Thompson, "Properties of Lead-Bismuth, Lead-Tin, Type Metal and Fusible Alloys," U.S. Bureau of Standards, Journal of Research, 5, p. 1085 (1930). 4. S. Turkus and A. A. Smith, Jr., "Low Tin Solders Containing Silver and Bismuth," Metals and Alloys, 15, p. 412 (1942) 5. F. N. Rhines and W, Ao Anderson, "Substitute Solders," Metals and Alloys, 14, p. 704 (1941). 6. J. B. Russell and J. 0O Mack, "Substitute Solders of the 15-85 TinLead Type," AIME Transactions, 161, p. 382 (1945)o 7. W. A. Baker, "The Creep Properties of Soft Solders and Soft Soldered Joints," Journal of Inst. of Metals, No. 2, po 277 (1939). 8. M. McKeown, "Properties of Soft Solders and Soldered Joints British Non-Ferrous Research Association, Research Mograph, No. 5, PO 57 (1947). 90 CO W. Bennett, "Tensile Strength of Electrolytic Copper On a Rotating Cathode," Transactions, Am. Electrochemical Soc., 21, p. 245-274 (1912). 10. S. Sonoda, "The Properties of Sheets Deposited on Rotating Cathode," Transactions, Am. Electrochemical Soc., 52 (1927), p. 233-247. 11. W. H. Shakespear, "Development and Use of Anaconda Electro-Sheet Copper," AIME Transactions, 106, po 441-448 (1933). 12. M. Altmeyer, "Manufacture of Sheets of Electrolytic Copper," Original in CUIVRE ET LAITON, 7 (1934),p. 367-370~ Chemical Abstracts, 28, 6640. 13. C. E. Huessner, A. R. Balden, L. M. Morse, "Some Metallurgical Aspects of Electrodeposits," PLATING, 35, po 554-557, 719-723, 768 (1948). 37

REFERENCES (Concluded) 14. To A. Prater and H. J. Read, "The Strength and IDuctility of Electrodeposited Metals," PLATING, 36, p. 1221-1226 (1949). Ibidem, 37, p. 830-834, 850 (1950). 15. N. P. Fedotev and Iu. M, Pozin, "Effect of Surface-Active Substances on the Mechanical Properties of Electrolytic Deposits," J. of Appl, Chemo of the USSR, 31, po 406 (1958). 16. C. Sruyk and A. E. Carlston, "Copper Plating from Fluoborate Solutions," Monthly Review of Electroplaters' Soc., 33, P. 932-934 (1946). 17. T. E. Such, "The Physical Properties of Electrodeposited Metals," METALLURGIA, 56, p. 61-66 (1957). 18. H, J. Read and T. J. Whalen, "The Ductility of Plated Coatings," Proceedings, Am. Electroplaters' Soc., 46, p. 318 (1959). 19. R. H. Barklie and H. J. Davies, "The Effect of Surface Conditions and Electrodeposited Metals on the Resistance of Materials to Repeated Stresses," Proceedings, Inst. of Machanical Engineers (London), 1, p. 731-750 (1930) 20. W. M. Phillips and F. L. Clifton, "Stress in Electrodeposited Nickel," Proceedings, Am. Electroplaters' Soc., 34, p. 97-110 (1947). 21. A. K. Graham and R. Lloyd, "Stress Data on Copper Deposits From Alkaline Baths," PLATING, 35, p. 449-450, 506 (1958). 22. H. Fishcer, P. Huhse, and F. Pawlek, "Internal Stress in Electrodeposited Copper," Zeitschrift fur Metallkunde, 47, p. 43-49 (1956). 23. H. J. Read and A. H. Graham, "The Elastic Modulus and Internal Friction of Electrodeposited Copper," Electrochemical Soc. Journal, 108, P. 73 (1961). 24. P. Hinnert and H. S. Krider, "Thermal Expansion. of Some Copper Alloys," Journal of Research, National Bureau of Standards, 39, p. 419 (1947). 38

0050- - __ ____o_ _ __. 12 o" 0040 - - - - - ___~T 9 000 Zn. 20 0.40 0 -0 —.o6"-4 -_ _ 0.02 0 ~0.00-0-_.10 -__._ C.. N0,j~s (4rhG/u) - - - NR xI0~-(4r~G/,jJ) - - -- - _0__ 0 40 5 0 0.3 4Q.'0. 4050 05 3 4 580 Q 1 0~~~~~~ii.6 0. 015~~~~~~~~~~~~~~~~~~~~.1 WRFA~~ ~ ~ 9.6 - 0.8 FDI ILL I tJUB Q01 BX I I~~t-b~~~I III i)EEST INTERPRETATO 0.008, EST'I INT RFF)RR E TAST 10 01 loSC6. 7.00_ 0.006 0~ ~~~-. 0"040' ~"-~ ooz ~II I I II ibs'~ -~-8 I im "Itch - 9.6 Ir op 9 e 10 ~ ~ ~ ~.. ~ ~ -0.01.vpsa 0.02e d t CL0_ %l0.004 10 t ar a -3 0 (49 tn Gotu) / 0.4 Q5 Q6 YT 1.5 ZO 0 4.0 Ob 05M R a 10 o),s t Q8 1.0 6-C -~~~~~~~~~F0.01 a/o- a - FigW M. Fig. 2L Palo AloWClf."93568 - (Epo by pERmATION via Pitch - /.68 per inC 0 F1WW Pui'W h~draulle diametr 4%.Olr olleo fti. Fin patch - 9.68 per inch FIA 1OW thddcaeas - 0.0041a flow passage 1draulic diameter - 4r.0.018 t IrrY-ij~l ~I~l~rosw area - 4r0o~wi Tla otl thlelmoss - 0.004 1a TOW beat transfe~r erea/tow veime -(Da f At Pr*-low uva/froute.1 area - a,.0,891. Fla area/total area - 0.195 Total 61at transfe~r &re&/total Talmo-1r9 f~i~ Fin ares./totl &ft& - O.'M5 W. M. Kays and A. L. L~ondon,, Compact Heat Exchang ers., The National Press, Palo Alto, Calif.. 1955 (reproduced by permission)

0. 100* 0.060 -.1 0 _ _ _ _ _ _ _ _ _ _ _ _ _ _ 0.55" -Ft - __ __ __.0 4 — - 0- 00sm_ _ _ O jQIW410 Q737e4 -I -i 0 1 1( 40.050 f III 0.040 02, 60 j 040.030 ~c~o _I II I I 1.020.15.015 - __- -t —- -~~~~~~~~~~~~~~~01 L-ES IT BEST INTERPRETATION.0 I 0~& —____ Z.010 " 0.008~- I- W- a_.008 - 0.006 ~ ~ ~ NR~0 (rG,~) I0.006,~~~~~~~~~~~~~~~~~~~.0 0 0.005~~~~~~~~~~~~~~~~~~~.0.004~~~~~~~~~~~~~~~~~~~.0 d4 506~ NRIO~ (4hGI) oo1 R xi RR~o- (4 rh Gj) 0.4 Q5 0.6 I I1.0 1.5 2.0 3.0 4.0 6.0 8.0 1QO 0.4 0.5 Q6 0.8 1.0 1.5 2.0 3.0 4.0 O Q Q 9.1 -0.737-S 9.29-0.737-SR Fig. 3. Fig. 4. FINNED FLIT TUBES FINED FLAT 1TJU SURFAE 9.1I - 0.131 - S SURFACE 9.29 - 0.T31 -SR Fin pitch - 9.1 per lach Fln pitch - 9.29 per Inch Flow passage hydraulic diameter - 4r.b-0.0130 ft. Flew passage hydraulic disater - 4rhrO.OlS6l ft. Fin metal thickness - 0.004 in. Pl& metal thickness - 0.004 In. Free-floew aea/froetal area - V 0.T88 23 Fre-flow are/fronl area - v 0o.71 Total heat,transfer area/total volume -o-824 ft.fft Total heat transfer area/totl volmme - Fy228 fte/ftB Fl area/te"taI area - 0.813 Fin area/tota ar - 0.814 W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission)

0.070 - 11 0.1 0600. 0.060 0.060 0,, I %,05 O i" 3?-i7 0050 -2 0.040 0.040 - -4 O.I - - BEST INTERPRETATIONGOIO1 0.004' 0 - Thb. outsid di emete1Ir - ~0.401 Ia. SF 11 - 0.006737 /r Flu pitch - 11.38 pr luckf arsa/ttlaa —0.839 I''lew pasa I c ditr - O 0.06 ft04 Flu ta th e - 0.004 j - MRotes Minim freefle a LspeR tra10r freflow area/froa a o0.780 to fle.:4 0.50 Total eat 1.0 1.5 2.0 3t0 4.0 6.0t/ft Te d idd i thi itis6. Fu area/total ara -.S b s yapplt o a t 1.32 - 0.ur SR Tube iuratsion ofcnieal - in402rest pifrch - 8. T ER i e ta o SURFACC 11*39 - 0.731-SR Fre"I ow wma/8roatal -,&,,,,,0, W. N. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reprod0.00uced by permission). f're*-rlow re&/iroz area - d.0.0 to fz *W. for whchs:d. b ~ boon obtA Pal Alto, Calif., 1955 (reproduced by permission).

! 4rh - 10.3 -F —-9o~ F —--.030 0.040.020 - - Q008- - _____ __ _____ __......_._._ z ~~~~~~~~~~~~~~~~~J0.47~., I I~ 0.030 ~ ~ __ _.0203 _ ~~~~~~~~~0020 0.001510 Rx - 4rG/u.0.6081 5 2....0S 00 1. _______ __ 1.5 2.0 3. 4.0 5.00,.~O1 090 ~'~SITPET~IN o' (~4rh) 6500 [ —~-+-~IW S 0. 8~,$~.-~I 5.3019-( 0o.os o 0.008 —-~~~.0 NxO (rG/J Fig. 7. Fig. 8. RJSI PLIT~~~~~~~~IP~~~~IIO 396191~~~EID ELITE-FI.0906 SURFACE 6.3SIPI 00 0.008 "'"4 p Fin Pitch - 6.3 p r inch flt spaslg h.0.SM 13. Plate spacing - h-O.470 in. Mwc pseeng. - diamete - r.00l6 t Flo passage hydraulic disaster - 4rh0.Ss Oft F etaly thisse - wf 0.006 I. in ml hi - 0.006 in. // Total trfer areavoum beteen plates l188ft./rft Total transfer ma/volume betwee plates -B9..a ft.1t.? (2004 rhr) /A'- / o~~~~~~~~~~~~~~~~~~~~.3 Fin Pitch - 5.3 per inc0.005 pitch - 90""a " mealticn"s-. 0.003 i: h"" Fl'Wtikes. n -~ ~ ~ ~~~~~~l~&*/oa r&- 001 Fla're "" O0.00.c: -~.14h~.0. W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1953 (reproduced by permission)

.050 4rh o 251.040 1. - T6'.... _- I 3FD 0 30.0210 — (V4h5 k-O.5 -8 I 20 -t-5 pi e~t - b^0.418 ia. n-te *pecit - W0 0 i.)0 05 ~Fl p rldr.01r-.07 t op08 ol ~~~~F 0081 _____ - -.008 -BEST INTERPRETATION - _____! BEST INTERPRETATION (. 00....003 - -.4 06 0.8 W. 15 2 3.0 4.0 6.0 8.0 004 0.5 0.6 1 ID 1.5 2.0 3D6 Fig. 9. Fig. 10 15.08 A C9.86 PI PLATE FIN PLAUI PLATE FIN SrMFAC 18.08 SURFACE 19.868 pFlu pitch - 15.08 per inch Fin pitch - 19.86 per inch Plate spacing - b0o.418 In. Plate spacing - b0.250 in. Flow passage hydraulic diameter - 4r hO.00876 ft. Flow passage hydraulic diameter - 4r -0.00616 ft. Fin metal thickness - 0.006 in. Fin metal thicknes - 0.006 in. Total transfer &rea/volume between plates -414 ft*/ft. Total transfer are&/volume between platesf -p-51 ft. / f. Fin area/total are& - 0.870 -Fuin ara/total area - 0.0849 ('/4rh)-685 (?/4rh)-35.0 W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission).

.0~~4" —) 1.4)341.0562"(.10 --.10 _08-.08 -'.0775_APPROX. -- _.06 - _ __.0775"APPROX. 005-.04 - 004-.0304-.030 __ EST INTERPRETATION.. BEST INTERPRETATION.015 IN. z_ ___.010.010 I / I I I II I 01 O~~~~~~~~~~~~~~~~~~~~~~.008 - QI I r1 00 0.008 —. —N-. -.006C-C jll/ I~~ —NR X 10-3 (4rh G /A D o i NR x 10- (4 rh G/ V 0.3 0.4 0.6 0.8 1.0 1.5 2.0 3.0 4.0 6 80 a01100 0.4 0.6 0.8 1.0 1.5 2.0 3.0 40 60 80 100 11.944- 3/8 W8 - 17.8 -3/8W Fig. 11. Fig. 12 WAVY-FBI PLATE-FIN WAVY-FN PLATE-FIN SURFACE 11.8 - 3/81 SURFACE 11.44 - 3/8W Fin pitch - 11.8 par inch. Fin pitch - 11.44 per inch Pinte spacing - b.O.413 in. Plato spacing - b.0.413 in. Flow passage hydraulic dianeter - 4rrh..000698 f. Flow pnasage hydraulic diameter - 4rh-.01060 ft. Fin netal thickness - 0.006 in. Fin notal thickness - 0.008 in. Totni bent transfer nren/volnn between plntoa -j ~ 5i4 ft./ft. Total bent transfer nren/volune between pint.s -4r351 ft 3 Fin nren/totnl aren - 0.892 Fin area /total nren - 0.847 Notes Hydrnuiic dianeter bnsed on free-flow nren nornni to Notes Hydraulic diameter based on free-flow aren nornal to nenn flow direction. mean flow direction. W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission).

' L85'-cl Ia NTER I.4,4"-..150 - __.100.031- _"__ ___- -......_. APPRO.05 —-- -.060- -- i 0 _ __ _ _ i b.00 EST 1I-N.0.0 E T - NR 0 - -. EST INTERPRE —'_..006 NR x'0 (4r - 0 00..05 080'"'5 2.0 3L0 4.0 5.0 6. 8.0 I I I I02 I III 1 1.060 Fig..015 14a. Via ite - 18.33- BES INo FTi ptch I0 16.2 p n V01 ia me t. Fi etlt 0.0.sat., 0.63'.ieot - 0./ Palo Alto, Calif., 131955 (reproduced by permission).Fig. 14. Fla length - o0oer i Fih lng-.008,10-125 in, F06 iN.I xhiclkO- - 0.0. 056 0. 8in metal thicke - 4.0 506 In. 005 cung op5Q6e 0.8 L O 1.5.D 3n. factors my be lowerfi, cutting ope6t 8o0 Fr i/t-n5.2 W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto., Calif., 1955 (reproduced by permission).

.100 - -- - -.035.0550 h0250 - II 00oo.080- o.. 070- 0 co _.060 - STEADY-STATE TEST DATA.060 "-'5O.050 *1 - l T T I T t r 1 1 1 1 1 1 1~040- I' - _040 ___ BEST INTERPRETATION AIR Q3Q -! - -- _ FLO 15 TUBE ROWS BEST INTERPRETATION 020 IX4 _ _.030-0 0 0 ___ 1/4 - 11.1_ NOS TO A STAD BUA FL *p - 0n - T d ditr-.. l. - 0E S Lower Pr - 006 I. Wralie d0r - 4r__.01" ft. o IV a h t i NR - 40.101 (4h t,/ -flw NRoat3 &rhG/& ) 0.3 0.4 0.506 0.8 1.0 1.5 2.0 3.0 4.0 600800 La.-Bea 0.6 08 1. 1.5 2.0 &0 4.0 Ot vm 10 15 nu area,/toud sr" - t - a.n S- 50re to flo(. Fig.W. M. Kays and A. L. ondon, Co act H15eat Echangers, The National Press, -RFACE /o4 -o T ~'"a>'o — S mE UAN -'1. 0 -'00a -~ 0 er gp - 0.005I. R8ralio diamtr - ~'0.015' "ft Palo Alto, Calif., 1955 (reproduced by pemission)

.060 -,.Q4Q4~- - X o ~ - - _ -.040 X,. 0o~~ ~15 TUBE ROWS.03 0 L 0 ~xt=1.5o.015 / o —. I-. —-x03125 1.25 0& NOMU TO M =-Lnj" TM B.._ Ta outid di -r - 0.250 la. d08iter -.,- - Fr *low -/-ro - - 0:338 -- N NRxIlO0 (4rhG/p) 0.6 0.8 1.0 1.5 2.0 3.0 4.0 5.06.0 8,0 IQO 15.0 Bet tra r area/totl volI - 801.50.4- 1.25 t/)t Fig. 17. FLOW NOWAL TO AN IN-LIKB TUBE BANK Tube outside diameftr - O.20O i.. Hydraulic diameter - 4rj-OoOI66 ft. Pree-floa area/froutl azea - V.0.338 Beat transfer arsa/total volumo - ~.80.4 ft./ft. W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission).

rIr ~ o<l< C- ) C C) C) C) C)J WJ 0 0A 0 f "NJ Round Tubes Continuous Fins Cord Type GN Louvered - Fin Round Tube Bank --...JWithout Fins JI J I ~ I i IHoneycomb McCord Type GN

6t 0 -0 ~- " I *rX "3 CC C — n Round Tubes oPe - Fin..........9 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~a, Continuous Fins -, - _ ~ I~ ~ — "- j Plate-Fin ~1 Wavy-Fin )I'~- Plate- Fin,l\ _.~_Strip-Fin Plate - Fin Round Tube Bank oo McCord Type GN

C) C ID,vin Round Tubes Continuous Fins WoF I )+ I" > Pa,, C~~~I' I IT1 ~~-" IPlate- Fin _, -- IJr Wavy- Fin Strip- Fin _____________________ Honeycomb McCord Type GN

CORRUGATED C PPER A 11/2" I/2 -- I 21- 3132 1SPACERSA'I I 3 1;rl 1'5/32 " HOLE SURFACE -' K —- 6" —--— ( 1jI iWATEMR CHA NNEL 6" SLVER 2" COPPER PIPE I OLDER 20 AIR PASSAGES /V 19 WATER PASSAGES'V I~~~~~~~~~~~~~~~~~~~~~~~~~~14I B I I B I 4 MOUNTING LUGS SILVER SECTI SOLDER COVER PLATE A 2 H 4 END CA PS SOFT SOLDER NOTE: ELECTROFORMED SURFACES E TO BE.007" TO.008" THICK. ASSEMBLY TO BE TESTED AT 25 psig CORRUGATED COPPER SPACER TO SOFT SOL DER HAVE APPROXIMATELY 10 FINS per INCH. TO BE SUPPLIED BY MR. RICHARD D. CHAPMAN 4"CCOPPER PIPE T 2" GRAHAM-SAVAGE, ASSOC. KALAMAZOO, MICHIGAN SECTION BB PROTOTYPE HEAT EXCHANGER DESIGN FOR INTERNATIONAL COPPER RESEARCH DRAWN - FKS SCALE 1/2 -DATE 2-16-63 NO. 65A-2 Fig. 21.

Steam I n Reservoir Control Valve i n Steam- Cooling Bypass Test Water Water Cooler Radiator ro Heat out Exchanger Flow Meter Accum u lator Return to Boi!ler Pump Fig. 22.

CONTRACT/ON CONE WESTINGHOUSE 8O3O 9.:/ RATr/O 725 HP MOTOR I0000 RPM at 7"SP INLET VANE FLOW CON TROL ~'[ ~~~~~22' Fig. 23.

1. 0 0. 8 m. 0. 6 CL. U) I~ U) V) 0.2 U)I 0 0. 2 0. 4 0. 6 0.8 1.0 Axial Travel Bypass Opening, Test Section Dia. Fig. 24.

Match Fan n let Point Speed Bypass Vanes Point Speed A 100 0 Full Open B 100 0 Part Closed C 100 Part Open Full n D 100 Part Open Part Closed E 50 0 Full Open F 50 Part Open Full Open Fan Characteristic 100% Speed System / \ 4 Full Open Vanes Characteristic (zero bypass) Pressure C Fan: 100o Speed Rise Part Closed Vanes D Fan Characteristic System Bypass E/ / 50%l Speed Characteristic FullOpen Vanes Volume Flow Rate Fig. 25.

Fig. 26. Wind tunnel. Al Fig. 27. Wind tunnel test section showing access panel. *-~~~~~~5

Fig. 28. Wind tunnel test section with access panel removed and air bleed shown. Fig. 29. Microstructure of "0.010 in." copper, NH40H, H202 etch, magnification lOOx. 57

jy:::i~ ii.~iiiii::iiiilii-ii~i::i~ii:iii:i-:::il..:.::._::::::-:::-~-::::::-:::i::::::~:ii:~:i-i-::-ii_.:: -::::::-:-:-:-::::~i::-::.-_:::-::i I:..-:: — i:: -i i-iil:-li:ilii -:i-i::i:i;i ~-::~:::-::::::_::::::-:::::_.i:i:::::-:::_-:-:::::::::-;.:.-..i i:::-.. j;.:::: I~-:i~i' i;-;i:_-.::_-. i:::l"':;::::::: —''- -:i-i:iiiiitii-.:::::-:_::::-;::.I..:::_:::_~ii: -i::ii.-_::::::::-:-: ~:: i:- i:iii:i_:ijiiiiii:-_i:-:__:i::::::'::~:::::::.:_:::-:: I: -::::::::~:::::::::-:::-::::::::::::-::~::::::::::'::::::::-:-::~siio~i i-i-::j::::::::~:: ~:::::::-:..::-:_-:::i:l:.-:::i:: 3::: -:: --i::::_:i: iii i.i-i —_ii: —' —'::.: —.-':.:-:':C:i&:i2iii:':::j:i:i::j:::::::- - -::-ii-:-i::- -:::~:::-:::::::::::~ —::I:~:~: r:~:~-~:i-i.i::::: -_:-::::::_:-p.:~i:..:.-:::. ^::;.::::::::: zi::,ilgst_: ij,::: —_~:iiil:iC: t%-Etl -_l-:_-:::::-iai:::': j:::::::::-::-:-:::-:::::::::.:-:::-.::;::::::~:::::::.:;:: i: -::::::-: —: ii$i~:iiii;:::-.:::.:-::::':":':- ~:-~,di::;Wi'::"j::::::::::-::i:: i::I::::~:' *::ii:~i:~-:::::::::::::i:'is.BW -: —:::::i::.:.: i:i:~Ii'~ii'-~i'j-'::;:'~ iu~i ~,-.L.:::::::i::::i::-:: i~ii:ii:i i-_:i.i::ii::::::iiziii -i.:-i:ia iij ii:ii8iii i-:C -~,s~: jdi ~::::-:-:::::::::: i-:i ii-T.~i —~i a_ "" iiii-ii-li _;E-i:-:-::.. i-:::~: --— ii; i::::-::: il-:::-::::::-:::-::::-::::::~:-::::::':::-: ~::~::::::-i:::: i:i-iiii. Ilt.sii~i:.:i:::::i:i:_ i-::::':::i;.::::;i~:nii";-::.,-_j~i::::::I:::::::: ii:~ —- i:::i-i:?-::::::::-:-:~::::-.::,:. i-i: I~i-i''"i::i:: ~:-:~*~:~::::::::;::::::::r::r:-.::::::-::::i:i::.::. ~:~~_-:-:-:i:::::j:::i~: i:i~:i::_::::: Wiali. i ~-i- _i —.:::.-::::::''":::~::::: ii::iii:i.:i:i:;::iiii;iii i::iii-:i-:i9 i-i':i:'i:i:-:i::~:::::::/'-~:i:i:i:l.':i.iir:~:::::-:,-'- l.i-jr i.::..; _':::-:"::-::::i~-::j -:.i —:-:~::.ii:~:::::i. l:::-:~::I:::::~'::'i::-i~:~'i:l:::.::::;g_:,::~::~-~~~~::::~::::: Fig. O. Microstructure of "0.010 in. rr NH40H, H202 etch, magnification 500x. Fig. 1. Microstructure of "0.020 in. It NH40H, copper, H20a etch, magnification 100x. 58

ii iia-s:i:::::::: iig. 32. Microstructure of "0.020 in." copper, NH4OH, H202 etch, magnification 500x. i~ Fig. 33. Krouse fatigue testing machines. 59

UNIVERSITY OF MICHIGAN l3 90502827 4911 THE UNIVERSITY OF MICHIGAN DATE DUE