THE UNIVERSITY OF MICHIGAN COLLEGE OF ENGINEERING Department of Mechanical Engineering Heat Transfer Laboratory Final Report AUTOMOTIVE RADIATORS MANUFACTURED BY THE ELECTROFORMING PROCESS ~J A:f Clar'k C. clare n,e Siebert,. Robert' B'. RellerrMi"'ichael Bor.den. * *W.?:'.... 4.. ORA Project 05335 under contract with: INTERNATIONAL COPPER RESEARCH ASSOCIATION, INC. NEW YORK, NEW YORK administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR December 1964

I?

TABLE OF CONTENTS Page NOMENCLATURE v ABSTRACT ix 1. INTRODUCT ION 1 2. RESULTS, CONCLUSIONS, AND RECOMMENDATIONS 3 3. HEAT EXCHANGER DESIGN SELECTION AND TESTING 5 4. RESULTS 11 APPENDIX A. HEAT EXCHANGER ANALYSIS 15 1. Analysis 15 2. Reference Heat Exchanger and Reference Conditions 20 31. Heat Exchanger Matrices Studies 21 4. Heat Transfer and Friction Data 21 5. Results and Conclusions 21 APPENDIX B. RADIATOR HEAT TRANSFER PERFORMANCE TEST APPARATUS 27 1. Introduction 27 2. Heat Transfer Apparatus 27 3. Instrumentation for Heat Transfer Apparatus 28 4. Wind Tunnel 29 5. Instrumentation for Wind Tunnel 30 6,. Wind Tunnel Development Tests 30 601 Bypass Performance 30 6.2 Test Section Velocity Distribution 31 7. Operating Experience and Procedures 31 7ol Radiator Installation 31 7.2 Stabilization of Radiator Water Inlet Temperature 31 7.3 Typical Operating Procedure 33 APPENDIX C. METALLURGICAL STUDIES 35 1. Literature Survey of Solders 35 1.1 Tensile and Shear Strengths 35 1.2 Creep Properties 38 lo3 Fatigue Properties 39 1.4 Summary 40

TABLE OF CONTENTS (Concluded) Page 2. Literature Survey of the Mechanical and Physical Properties of Electroformed Copper 40 2.1 Summary 50 3. Tensile and Fatigue Test on Electroformed Copper 50 3.1 Conclusions 54 APPENDIX D. ELECTROFORMING PROCEDURES FOR HEAT EXCHANGER FABRICATION 55 REFERENCES 59 iv

NOMENC LATURE 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 F heat exchanger correction factor, dimensionless G mass velocity, (w/Ac), lbm/hr-ft2 conversion factor; go = 32.2 (lbm/lbf) (ft/sec ) h heat transfer coefficient, Btu/hr-ft -~F k thermal conductivity, Btu/hr-ft ~F Se length of fin, ft L total flow length of heat exchanger, ft -I. m fin parameter, ft, see Eq. (7) NTU number of transfer units, dimensionless p wetted perimeter of fin, ft Pr Prandtl number, dimensionless, see Eq. (9) Ap pressure drop, psf q heat transfer rate, Btu/hr

NOMENCLATURE (Continued) English Re Reynolds number, dimensionless, see Eqg (10) rh hydraulic radius (AcL/A), ft, 4rh = hydraulic diam St Stanton number, dimensionless, see Eq. (9) t fin thickness, ft T temperature, OF ATOL log-mean temperature difference for pure counterflow, ~F U overall heat transfer coefficient, Btu/hr-ft2-~F V volume, ft3 VB air volume between plates, ft w mass flow rate, lbm/hr Greek 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/fts water channel thickness, ft heat transfer effectiveness (Ref.b ), dimensionless total surface temperature effectiveness, dimensionless, see Eq. (5) Tlf fin efficiency, dimensionless, see Eq. (7) C Ac/A.FR, dimensionless p', density, lbm/ft3 vi

NOMENCLATURE (Concluded) Greek viscosity, lbm/ft-hr see Eq. (32) 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. Radiator Identification.EFHX-I first model of a partially electroformed radiator (Figs. 6 and 9) EFHX-II second model of a partially electroformed radiator (Figs. 7 and 10) SPHX soldered radiator having physical characteristics similar to EFHX-I and -II. (Figs. 8 and 11) LMHX literature matrix heat exchanger (radiator) (Fig. 11).

ABSTRACT This is the final report of a study to investigate the feasibility of manufacturing an automobile radiator (air-water heat exchanger) by the process of electroforming. It was concluded that owing to the present state of the technology of electroforming it is not economically practical at this time to fabricate radiators either partially or completely by the electroforming processes. The desirability of producing an electroformed radiator remains an important goal as such construction would eliminate the solder and could be expected to have superior strength, improved performance and higher reliability in comparison with conventional radiators. It is recommended that further studies be undertaken to translate available information on transport phenomena in electrolytic systems to practical manufacturing processes. Significant improvement in industrial electroforming methods and techniques and the competitive position of this industry are considered possible by a careful and well planned application of transport theory. Two partially electroformed radiators and one soldered radiator were designed, constructed and tested for their thermal and frictional pressuredrop performance. The first electroformed radiator displayed superior thermal performance and inferior frictional pressure-drop performance. This was attributed to an artificial surface roughness induced on the air-side surfaces by the processes of manufacture of the radiator and was not a direct consequence of the electroformed construction. Accordingly, further research was started on an investigation of the heat transfer characteristics of thin electroformed sheet which can be made with rough surfaces. This report outlines the method of selection of the radiators for testing, the design and operation of a wind tunnel for radiator performance testing, test procedures and the results of the thermal and pressure drop testing of all radiators. These results are given in Appendices A and B. A limited literature survey was made to obtain selected-data on the mechanical properties of solders, soldered joints, and electroformed copper. Tensile and fatigue strengths were determined on nominal 0.010 and 0.020 inch thick strip specimens. The tensile strength averaged 27,818 and 27,324 psi respectively for these two average thicknesses. The endurance limit of these two nominal thicknesses based on ten million cycles was 11,500 and 10,500 psi respectively. These results are presented in Appendix C. The electroforming procedures are presented in Appendix D and have been prepared by Mr. Frank K. Savage, President, Graham-Savage and Associates, Inc., ix

Kalamazoo, Michigan. Mr. Savage was the consultant on electroforming manufacture processes and also supervised construction of all electroformed radiator subassemblies. x~~~~~~~~~~~~~~~~~~~

1. INTRODUCTION This is the final report on "Automotive Radiators Manufactured by the Electroforming Process,," a research program sponsored by the International Copper Research Association, Inc., New York, New York. The objective of this study was to investigate the feasibility of manufacturing an automotive type radiator (air-water heat exchanger) wholly or in part by the process of electroforming. It was considered desirable from both an economic and performance reliability standpoint to eliminate the solder joints in presently manufactured radiators. To this end two partially electroformed radiators were assembled and performance tested in a wind tunnel designed for that purpose. Comparative performance of these partially electroformed radiators was obtained by testing a similar but soldered radiator and by using performance data from the standard literature on extended surface heat exchangers. A partially electroformed heat exchanger was decided as a practical initial step in view of the state of the electroforming technology. This construction consisted of electroforming approximately 0.012 in. thick copper faces to the crests of pre-existent corrugated copper automotive radiator strip stock. Prior to electroforming the strip stock was encased in a meltable alloy and the faces machined to expose to edges of the copper strip stock. The electroforming then completed a sub-assembly with two parallel faces connected laterally by the corrugated copper strip stock. Approximately twenty such subassemblies were soldered together to form the heat exchanger matrix with a 0.094 in. spacing between adjacent subassemblies serving as water coolant channels. Headers were attached at each end of the heat exchanger to provide for water coolant connection and distribution. Photographs of a typical subassembly and an assembled radiator (heat exchanger) are shown in Figs. 1 and 2. All electroforming operation were done by the Graham, Savage and Associates, Inc., Kalamazoo, Michigan, under the supervision of Mr. Frank K. Savage. The heat exchanger design and testing were completed by The University of Michigan. This included the selection and evaluation of radiator core designs from the standpoint of their mechanical, thermal and frictional pressure drop performance, the selection of a specific design and a metallurgical study of the physical properties of electroformed copper sheets and electroformed copper joints. This latter involved a study of the microstructure characteristics and a determination of the tensive and fatigue strengths of;the electroformed parts.

Mr. RichardD. Chapman, Director, Automotive Development, Copper and Brass Research Association, contributed to the general coordination of this work with interested parties in industry. This report will be divided into two major parts. The first will be a brief summary of principal results, conclusions and recommendations, heat exchanger (radiator) design selection and testing, wind tunnel facility design, and experimental results. The second part consists of appendices in which much detailed information is presented on the summaries in the first part. The results of the metallurgical studies and the electroforming procedures are included as appendices. An initial six-month progress report on this work was issued in May 1963 under the title "Automotive Radiators Manufactured by the Electroforming Process," some of the material from that report is reproduced here as appendices.

2. RESULTS CONCLUSIONS AND RECOMMENDATIONS Probably the most significant result of this research was the finding that with the current state of the technology of electroforming it is not economically feasible to manufacture an automotive radiator from partially PI oc+.r rn-Pn-rmPri BlaiTh maq -ct'imh 1 P Ti'r+i'm":. im'r; n r- ow," A~,,mn r I'-,,.... -rm4 —,

by this laboratory under the present sponsor and results of a preliminary study are expected to be ready at the end of this year. Apart from this investigation of the characteristics of rough electroformed surfaces, the following recommendation is made as a result of this study. There is at present an apparent information separation between the theoretical workers in electrochemistry and those in the transport mechanics of heat and mass transfer and the practitioners in the electroforming and electro-plating industry. As a consequence the electroforming technology has not been developed to the stage where it is practically possible nor economically feasible to electroform shapes as complex as those of an airwater heat exchanger. It has been found, however, that a great deal of useful information dealing with transport theory applied to electroforming processes is presently available, mostly in the form of scientific papers and monographs. This information ought to be carefully surveyed by competent and experienced personnel and a position paper formulated in a way which would be useful to application in the electrochemical industry. Such a study also would reveal areas in the technology of electro-deposition where additional investigation of the transport processes of heat and mass can produce potentially useful results. It is felt that significant improvement in industrial electroforming methods and techniques, and possibly the competitive position of this industry can be achieved by a careful and well-planned application of transport theory. In specific terms it is visualized that such an application would enable the prediction of rates of deposit (or removal) of metal on various geometric shapes in terms of electrolyte properties, flow rate, turbulence level, temperature, electric potential fields and diffusion coefficients, among others*. Such a study as is recommended here should include experimental investigations in circumstances having commercial importance but not be limited to these aloneo New conditions of geometry, flow, electrolyte, potential field, etc., should be studied in order to investigate improved methods for the electroforming of new shapes. These results would apply as well to metal removing processes such as electro-chemical machining. NIt is recognized that surface phenomena not influenced by the transport processes will have an important effect onthe deposition process and must be included in a comprehensive studyE

3. HEAT EXCHANGER DESIGN SELECTION AND TESTING At the start of this research it was desired to select a heat exchanger matrix for possible electroforming manufacture which would have thermal and frictional characteristics acceptable for automotive application and representative of that class of air-water heat exchangers. To accomplish this an optimization procedure was adopted which would select that heat exchanger matrix having the maximum rate of heat transfer per unit volume per unit frictional pressure drop. In other words, the matrix corresponding to this optimum would represent that design requiring the minimum volume and producing the minimum frictional pressure drop for a specified rate of heat transfer and air flow rate. The calculation of this optimum design required the use of experimental data on extended surface heat exchangers. The data employed were those published by Kays and London in Compact Heat Exchanger (1), an authoritative publication on this subject. The details of this optimization procedure and selection are outlined in Appendix A. As a result of this study it was determined that a matrix which was of a suitable optimum type on the basis of its thermal and frictional characteristics should be of a plate-fin design having 10 fins per in. with waviness, or turbulent promoters, on the fins. The air flow channel was to be between 0.40 and 0.50 in. wide, 2 in. deep and 12 in. in height, the latter two dimensions being selected arbitrarily. Such a design, it was felt, also would have a good chance of being fabricated, in part at least, by electroforming. Three test heat exchangers were fabricated. All were essentially of the design shown in Fig. 2. Two of the exchangers were made of electroformed sub-assemblies and are referred to in this report as EFHX-I and EFHX-II (electroformed heat exchanger I and II.). The third exchanger was basically the same, except that it was fabricated in a conventional manner by solder and is referred to here as SPHX (soldered prototype heat exchanger). For comparative purposes the experimental results from EFHX-I, EFHX-II and SPHX are compared with published results from the heat transfer literature on a similiar heat exchanger matrix. This is identified as LMHX (literature matrix heat exchanger). Both electroformed heat exchangers were manufactured from sub-assemblies such as that shown in Fig. 1. Pre-existent automotive radiator spacer stock was placed in a mold into which a meltable alloy was poured. After the alloy hardened it was removed from the mold and its lateral faces were machined exposing the thin edges of this copper spacer stock. This was then placed in an electrolytic bath and approximately 0.012 in. of copper was deposited on the machine faces forming a copper-copper joint with the spacer stock. The alloy was then melted out leaving a sub-assembly as shown in Fig. 1. Approximately 20 such sub-assemblies were placed adjacent to each other, separated

about 0.094 in. and soldered to form the basic heat exchanger matrix. Headers were attached at each end of the matrix to complete the heat exchanger. Detailed drawings of EFHX-I and -II are given in Figs. 6 and 7. The soldered heat exchanger was assembled from the identical copper spacer stock as EFHX-I and -II. The air channels are wider in this design because the crests of the strip stock were not machined off. A detailed drawing of the soldered heat exchanger is shown in Fig. 8. Detailed air-channel and water-channel dimensions as seen from the air flow direction are given in Figs. 9, 10 and 11 for EFHX-I, EFHX-II and the SPHX, respectively. The EFHX-I was fabricated by the Graham, Savage and Associates, Kalamazoo, Michigan. The EFHX-II was fabricated under the supervision of this laboratory using air channel sub-assemblies electroformed by the Graham, Savage and Associates and the SPHX was fabricated by the Young Radiator Company, Racine, Wisconsin. A summary of the principal geometric characteristics of each of these heat exchangers is given in Table 1. The heat exchanger matrix referred to as LMHX is matrix number 11, Table 2 in Appendix A. This matrix was chosen for comparative purposes with the experimental data obtained in this study in order to provide an approximate independent check on the measurements obtained from EFHX-I, EFHiX-II and the SPHX. The experimental data for LMHX were obtained by Kays and London and reported in Ref. 1. These data are reproduced here in Fig. 31. The mechanical strength of an electroformed sub-assembly was tested by subjecting one of the electroformed faces to air pressure and measuring the corresponding deflection of the matrix. Air pressure was applied by the use of an air channel soldered to one side of the sub-assembly as shown in Fig. 12. This arrangement simulates the state of stress imposed on the matrix by the pressure of the water in the water channels. As may be observed from Fig. 12 the pressure-deformation characteristics are very complex as would be expected. Both elastic and plastic type behavior is seen with a certain amount of permanent deformation which increases at about 40 psig. However, even at a pressure of 60 psig, or about 4 atmospheres gage, the total deformation is only 0.038 in. The matrix was tested to 75 psig without rupture. On the basis of these results it was felt that the subassembly was of sufficient strength and rigidity for an automobile radiator. Designs having even greater stiffness can readily be fabricated. The radiators were tested in a windtunnel specifically designed for this purpose. The wind tunnel, its auxiliaries, instrumentation and operating pro-cedures are described in detail in Appendix Bo The rate of heat transfer q was determined from the enthalpy change of the heated water flowing through the core of the radiator. This enthalpy

TABLE 1 SUMMARY OF HEA.T EXCHANGER CHARACTERISTICS* SYMBOL EFHX-I EFHX-II SPHX LMHX' ft2 j 38.8 353 38.4 Ac, ft2 0.727 0. 660 0.729 AFR 0ft i 0.93 0.852 0. 924. a in n. 0.012 0.012 0.010 -- b, in. 0o.434 O.434 0.480 0.413 Ic, in 0.002 0.002 0.002 0.o006!L, in. 2.00 2.00 2.00 4r (airside), ft 0.01246 0.01246. 01266 0.olo6 4rh(waterside), ft 0.01396 0.01509 0.01588 Iv,ft3 o.156 ft3 o0.142 0.154 Vg, ft3 O. 1212 O. 1212 10. 0 1215 - ft2/ft3 249 248 249 294 ft2/ft3 321 321 316 351, in o.09og4 094. 099ogg 0, 0|776 0. 775 0o789 o.836 Ae~, ft 2 0.0251 00246 0.0271 h, ft2 6.53 5.90 5.90 fins per in. 10 10 j 10 11.44 *Key EFHX-I is the first electroformed heat exchanger fabricated (see Figs. 6 and 9) EFHX-II is the second electroformed heat exchanger fabricated (see Figs. 7 and 10) SPHX is the soldered heat exchanger (see Figs. 8 and 11) LMMHX literature heat exchanger (Fig. 31).

change was computed from measurements of the flow rate of the water and its temperature change. The objective of these tests was the determination of the average air-side heat transfer coefficient and the corresponding value of the Stanton number, the principal heat transfer parameter. Because of the test arrangement, the air-side heat transfer coefficient could not be obtained directly but had to be determined from indirect measurements and computations. The following is a brief description of the method of calculation of both the heat transfer and the frictional performance data. The overall coefficient of heat transfer is determined from the rate equation for the radiator as, q = UAF ATOL = (WCph (To-Ti)h where, ATOL is the logarithmic mean temperature difference for pure counterflow. F is the heat exchanger correction factor for the actual flow configuration (cross-flow, both fluids unmixed). In all cases F was approximately 0.99. From Eq. (1) the product UA is written 1 FATOL F ATOL (2)'VA. - q = (wcp)h (To-Ti)h also, by the definition of U, we have 1 = 1 + a + 1 (3) UA 1ochcA A Akk ohhhAh however, for the radiator configurations employed the thermal resistance of the water channel wall is negligible and )oh is unity. Hence, = 1 + 1 (4) UA nochcAK hhA the fin efficiency 1oc is given as

O = 1 _ (1 - Anf) and Tanh(m) ( f m, where, m = ___ (p) MtQ (7) kta As may be noted from Eqs. (2), (4), (5), (6), and (7) the determination of he, the average air-side heat transfer coefficient, requires values of hh, the water side heat transfer coefficient. Furthermore, the air side fin efficiency T itself involves hc. As a result once numerical values are Oc obtained for hh and UA, the final calculation of he is a trial and error procedure. This computation is not difficult, however, since hh is considerably greater than hc, in most cases, and the value of Toc falls within fairly fixed limits. Numerical values for hh were taken from the standard heat transfer literature for flow in channels. The air side heat transfer data are presented in terms of the dimensionless parameters Stanton number, St, Prandtl number, Pr and Reynolds number Re. Data formulated in this manner provide a simple, meaningful and generalized method of presentation for comparative purposes. These results are summarized in Fig. 3 and are discussed below. In addition to heat transfer data, measurements were taken of the frictional pressure drop characteristics of each radiator configuration. These results are summarized in Fig. 4 in terms of a friction factor f as a function of Reynolds number and are discussed below. The friction factor is given as f rgoPo (Ac APc (8) 2 L - c where, APc is the drop in total pressure across the radiator.

4. RESULTS The experimental heat transfer and frictional pressure-drop results for the three radiators tested and the one taken from the literature are summarized in Figs. 3, 4 and 5. Individual results on each of the radiators are presented in Figs. 13 through 20. Three experimental radiators were tested and an additional one of similar design was taken from the literature for comparative purposes. Physical property data on each of these radiators is summarized in Table 1. The radiators are identified as follows: 1. EFHX-I. First model of a partially electroformed radiator (Figs. 6 and 9) 2. EFHX-II. Second model of a partially electroformed radiator (Figs. 7 and 10) 3. SPHX. Soldered radiator having physical characteristics similar to EFHX-I and -II. (Figs. 8 and 11) 4. LMHX. Literature matrix heat exchanger (radiator) (Fig. 31). Two partially electroformed radiators were fabricated as the first radiator was felt to be not of a sufficiently high standard of construction to be used for conclusive studies. In particular there were irregularities and excessive roughness in the air channel which would not be found in a commercial automobile radiator. Using the experience gained on the first, a second partially electroformed radiator was constructed and identified as EFHX-II. This radiator was assembled using electroformed sub-assemblies produced by Graham-Savage and Associates, Kalamazoo, Michigan. Its construction was done by soldering the sub-assemblies and was supervised by The University of Michigan. A completely soldered radiator (SPHX) was fabricated using the same spacer stock and tested for comparative purposes. The heat transfer data are presented for the air-side as the product St Pr2/3 as a function of the air-side Reynolds number Rec. These dimensionl2ess groups are defined as 2/3 he ( 2/3 St Pr/ - (9) Gcp 11

Re - 4rcG (10) c Friction data are given in terms of a friction factor f as a function of Reynold's number. These methods of representation generalize the data and allow for comparison between radiators as well as between various flow conditions for a given radiatoro It also permits comparisons to be made without the requirement that air or water flow rates be identical for the comparison. In all data reported the Prandtl number was that for air and hence varied only slightly. Superior heat transfer performance among these essentially similar radiators is noted by a large St Pr2/3 product for a given Reynold's number whereas superior friction pressure-drop performance would be indicated by small friction factor for a given Reynold's number. The summary heat transfer data in Fig. 3 indicates that EFHX-I has superior performance to the others. Close inspection of the matrix disclosed a rough air-side surface condition which is believed to have caused an increase in the turbulence in the air-flow and thus the increased heat transfer. This is further borne out by the friction data summarized in Fig. 4 where EFHX-I also exhibits the greatest friction factor. This is consistent with the conclusion that air-side surface roughness is responsible for the increased heat transfer. The surface roughness is thought to be caused by a residue of the meltable alloy used to encase the copper spacer stock during the electroforming operation. Because of this it is concluded that the increase in heat transfer is not a result of improved performance owing to the electroformed copper spacer stock (fin) attachment to the water channel. The heat transfer and friction characteristics of the other radiators are summarized in Figs. 3 and 4 also. Differences in the performance characteristics are less and the electroformed radiator (EFHX-II) exhibits somewhat better heat transfer and slightly poorer friction characteristics than the others, SPHX and LMIIXo The soldered radiator (SPHX) has essentially identical heat transfer performance as the radiator matrix taken from the literature above a Reynold's number of about 1000. Its friction characteristics are generally superior to all other radiators9 a probable result of its excellent manufacture. Combined heat transfer and friction data for all radiators are presented in Fig. 5 as St Pr2/3/f as a function of the Reynold's number. Representation of this kind emphasizes both the heat trans er and friction characteristics of a radiator. Thus large values of St Pr2/ 3/f indicates both improved heat transfer and friction performance. As may be noted on this figure, the soldered radiator SPHX has superior performance while the first electroformed model EFHX-I exhibits the poorest performance. This latter is in line with the characteristics of BFHX-I as shown in Figs. 3 and 4~ The heat transfer/ 12

friction characteristics of EFHX-II and SPHX are about the same and each is superior to that of EFHX-I. The heat transfer and friction data for each of the individual radiators are given in Figs. 13 through 20. Each data point corresponds to a separate run having a different water velocity. All tabulated data are on file with the principal author. 13

APPENDIX A HEAT EXCHANGER ANALYSIS 1. 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 basic 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.l 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 = E(wcp)min (Thi-Tci), (11) in which e 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, ioe., counter flow, cross flow, etc., and the ratio (wp)min/(wcp)maxo1 Hence, with the specifications a to e above, it is evident from Eqo (ll)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 15

NTU = UA (12) (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)l = (UA)2 ~ (13) Now, if A is based on the air-side or finned-side of the exchanger, we have, 1 1 a 1 l + — a + (14) UA flochcA Awk 1ohhhAh For the usual automobile radiator the last two terms in Eq. (14)are negligible, i.e., the air-side controls, so, UA = TochcA' (15) Dropping the subscript c as all symbols now refer to the air-side, we have from Eqs. (13) and (15) (foh A)1 = (0oh A)2 (16) where ho = total surface temperature effectiveness. Hence, for purposes of comparison between various heat exchanger matrices at constant NTU, their relative volumes may be given as V2 (UA/V)2 [(Lh)(A/V)]1 (17) V2 (UA/V)1 [(qoh)(A/V) ]l 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, 16

V = LTx -h)/ vA/'o 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 (1~9) V8 (V/V)8 ( Now, the heat transfer parameter, the Stanton number, the flow parameter, the Reynolds number are defined as St = h (20) Gcp Re =- 4rh (21) so, ho StoGo (Sto)(w/Ac)o St G (22) h St G (St)(w/Ac) which for constant air-side flow w, becomes, ho Sto Ac -- = -—, --- (23) 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 StoPrl/30 Thus, Sto (St pr2/3)o St St. Pr2/3 Hence, from Eq. (18)we have for a fixed AFR, the frontal area of the heat exchanger, 17

(V o) St Aco (A/V) (rAo) C/A (A/V) (ko) \St (AC/AFR)o (A/V) (rio)o (St0 (24) (no) St CI where Ac Ca -, (25) AFR a A = a (26 V and A VB Kays and Londoni 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~Ly where L is the depth of the matrix in the air flow direction, Eq. (24) may also be written for fixed frontal area and total surface temperature effectiveness, as L _ -Sto o V Lo a / o- (S( VO The Reynolds number ratio is written for constant w and AFR from Eq. (21) as Re (rh/rho) Reo (Ca/a 0) 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 a= 1 =(29) l+6/b 18

In these calculations 5 is taken to be 0.080 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 Ap = L (w/Ac)2 L 1 3 rh 2gop rh p?2gop) Hence, for the various matrices having constant frontal area, the relative frictional pressure loss is Ap (f/fo)(L/L) APo (rh/rho) ( /a o) 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, = (qZV), (32) Ap we have the formulation of the relative value of this parameter as L ( q/V)/ZAP o0 (q/V) O/Ap (L/o) (AP/Apo) Equation (33) 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 volume ratio (and depth ratio), Eq. (27), frictional pressure drop ratio, Eq. (35J, and heat transfer per unit volume per unit frictional pressure drop ratio, Eq. (33). All are based on a reference heat exchanger for reference conditions, defined below. 19

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. 21-37,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 hydraulic diameter as the selected Honeycomb core. 2. 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. 34. 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 H = 1.285 x 10-5 lbm/ft/sec p = 0 071 lbm/ft3 Hence, 4rho = 0O1042 in. = 0.oo00868 ft ao = 1/(l+o o80/0o414) = o.838 0o = 417 ft2/ft3 ao = Boao = (417)(0o838) = 350 Re0 = 4rhoGo/' = 4rho(W/Ac) o/t = (o.oos68)(o o7?)(o 1) 1o0/(o858)(. 285) = 572. From Fig. 34 corresponding to Reo = 572 we find, 20

(St-Pr2/3)0 0. 0155 and fo = 0 093. 3. EEAT EXCHANGER MATRICES STUDIES A summary of the 18 heat exchanger matrices selected for study is given in Table 2. 4. HEMAT TRANSFER AND FRICTION DATA The computed data for heat transfer and friction including the matrix parameters of a, aO, and D and their relative values are summarized in Tables 3 and 4. 5. RESUITS 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, Ir/rIo are plotted in Figs. 38-40. 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 2 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. 38), and the highest heat transfer per unit volume per unit pressure drop (Fig. 40), although its frictional pressure drop is only moderately favorable (Fig. 39). 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. 21

TABLE 2 Summary of Heat Exchanger Matrices* Matrix Hydraulic A/V No. Classification Matrix Type Kays and London Fins/in. Diameter, ft2/ft3 in. Flat tubes, Flat tube, plain 9.68-0.87 9.68 0.1416 229 continuous fins continuous fin 2 Flat tubes, Flat tube, ruffled 9.68-o.87R 9.68 0.1416 229 continuous fins continuous fin Flat tubes, Flat tube, plain 9.1-0.737-S 9.1 0.1656 224 continuous fins continuous fin Flat tubes, Flat tube, ruffled 9.29-0.757-SR 9.29 0.1622 228 continuous fins continuous fin Flat tubes, Flat tube, ruffled 11.52-0.757-SR 11.52 0.1582 270 continuous fins continuous fin 6 Round tube, Round tube, 8.0-3/8T 8.0 o.143 179 continuous fins continuous fin Plate fin, Plain fin, 5.5 5.5 0.2420 161 plain fins b = 0.470 in. 8 Plate fin, Plain fin, 9.05 9.05 0.1828 222 plain fins b = 0.825 in. Plate fin, Plain fin, 15.o8 15. 8 0.1052 348 plain fins b = 0.418 in. Plate fin, Plain fin, 19.86 19.86 0.0738 424 plain fins b = 0.250 in. 11 Plate fin, Wavy fin, 11.44-3/8w 11.44 0.1272 294 wavy fins b = 0.415 in. 12 Plate fin, Wavy fin, 12.9t i, ayfn 17.8-3/8w 17.8 o.o836 432. wavy fins b = 0.415 in. Plate fin, Strip fin, 3/32-12.22 0.1343 292 strip fin b = 0.414 in. 14 Plate fin, Strip fin, 1/8-15.2 15.2 o.lo42 350 strip fin b = 0.414 in. Plate fin, Louvered fin 15 Louvered fin 14/-11.1 11.1 0.1214 278 Louvered fin b =0.250 in. 16 Round tube bank Staggered S-1.50-1.25(S) -- 0.1980 80.5 17 Round tube bank In-line I-1.50-1.25(S) -- 0.1980 80.4 18 Honeycomb 1/4-in.-2-1/4 in. --- -- 0.1410 192 *Round tubes with circular fins not considered since for available basic heat transfer data, tubes are larger than 3/8-in. dia and area/volume ratio is less than 170. These would not complete with the above configurations.

TABLE 3 Heat Transfer Data No.ix 4rh rh/rho a a/ao a a/ao Re/Reo Re StPr2/3 St/Sto L/Lo 4'h~~~~~~~~~~~~~~~R St-Prh/ tSo L/L I 0.1416 1.36 0.697 0.833 229 229 0.655 1.635 935 0.0063 0.405 3.14 2 0.1416 1.36 0.697 0.833 229 229 0.655 1.635 935 0.0079 0.510 2.50 3 0.1656 1.59 0.788 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 0.654 1.65 945 0.0115 0.740 1.98 5 0.1382 1.33 0.780 0.930 270 270 0.771 1.43 820 0.0110 0.710 1.70 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 0.855 1.02 188 161 0.460 2.28 1300 0.0059 0.380 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.O1 0.840 1.00oo 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 0.0110 0.710 1.05 11 0.1272 1.22 0.836 1.00oo 351 294 0.840 1.22 699 0.0170 1.10 1.07 12 0.0836 0.800 0.835 1.00 514 432 1.235 0.800 468 0.0172 1.11 0.730 13 0.1343 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 0.1214 1.162 o.758 o.go905 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 A:,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~33

TABLE 4 Friction Data Matr ixL Mati Re f f/fo L/L0 rhjrho a/a, Ap/Ap0 &L.A,) */* No. * p\ApQ/g/ 1 935 0.023 0.247 5.14 1.6.833 0.685 2.1 0.465 2 935 0.035 0.377 2.50 1.36 o.833 0.832 2.08 0.480 3 967 0.036 0.387 2.07 1.59 0.941 0.535 1.1o8 0.904 4 945 0.042 0.452 1.98 1.53 0.941 o.614 1.218 0.822 5 820 0.042 0.452 1.70 1.33 0.930 0.621 1.0o6 0.946 6 1250 0.028 0.501 2.04 1.57 0.637 0.704 1.435 o.696 7 1300 o.oi6 0.172 5.83 2.52 1.02 0.425 2.465 o.4o5 8 920 0.025 0.247 4.48 1.75 1.09 0.633 2.835 0.355 9 572 0.036 0.587 1.83 1.01 1.00 0.705 1.290 0.775 10 447 o.o4 o0.441 1.05 0.707 0.905 0.725 0.741 1.550 11 699 0.092 0.990 1.07 1.22 1.00 0.869 0.950 1.075 12 468 0.083 o.893 0.750 0.8oo 1.00 0.815 0.595 i.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 1 15 756 0.070 0.753 1.26 1.162 0.905 0.900 1.155 0.880 16 2750 0.071 0.764 2.08 1.90 0.398 2.10 4.19 0.239 17 2550 0.055 0.570 2.74 1.90 0.465 1.78 4.88 0.205 18 1150 0.0125 0.1545 3.38 1.55 0.671 0.501 1.70 0.587

'rt 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-l178. 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.

APPENDIX B RADIATOR HEAT TRANSFER PERFORMANCE TEST APPARATUS 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. 41 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. 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.

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. 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 rotametero A turbine flow meter is provided for an independent check. A typical set of automotive radiator performance conditions is given in Table 5 and may be compared with the design capabilities of the test loop shown in Table 6o TABLE 5 Typical Maximum Automotive Radiator Performance Water flow rate 300 lbm/min Air flow rate 1760 ft3/min Air inlet temperature 940F Air outlet temperature 151~F Water inlet temperature 184~F Water outlet temperature 190~F Heat transfer rate 1800 Btu/min Radiator inlet airflow. Using the constant area inlet section described below, a Pitot tube was used to measure the inlet airflow velocity. The dynamic pressure was measured with a micromanometer. Radiator air side pressure drop. Total pressure surveys downstream of the radiator indicated that a single location was sufficient to provide a representative or average value of downstream total pressure. Therefore, the air side pressure drop was measured as the difference between the inlet total pressure (from the Pitot tube) and the downstream total pressure. 28

TABLE 6 Heat Transfer Loop Design Conditions Heat transfer rate 100 hp or 4200 Btu/min Water temperature in 190~F Water temperature out 1800F Water flow rate 50 gpm Inlet water pressure 30 psia Inlet steam temperature 350~F Inlet steam pressure 50 psig 4. WIND TUTNNEL The wind tunnel consists of a contraction section, test section, diffuser, fan, and discharge ducting as shown in Fig. 42. 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. 43, and the combination of all three methods of flow control are shown in Fig. 44. Exhaust air from the tunnel which is some 500F 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. 29

Photographs of the wind tunnel assembly, the test section region with removable, viewing, and access doors, and the bypass are shown in Figs. 46-48 5. INSTRUWENTATION 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 7. TABLE 7 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 6. WIND TUNNEL DEVELOPMENT TESTS 6.1 Bypass Performance First testing of the basic tunnel performance involved the determination of the effect of bypass opening upon the test section air velocity. With test section air velocity determined by a Pitot tube located at the center of the tunnel test section, the fan operating at maximum rpm, the bypass initially closed, readings of airflow velocity were obtained for various bypass openings. Results are shown in Figure 43 with the test section velocity in terms of the zero bypass air velocity and the bypass opening in terms of the hydraulic diameter of the tunnel test section. Comparison of the predicted 30

with the actual values shows approximately a 2% difference between the two sets of values, in terms of the reference velocity. 6.2 Test Section Velocity Distribution The velocity distribution in a plane normal to the tunnel longitudinal axis and located near the middle of the test section provides information on the uniformity of the airflow. In order to obtain this information, horizontal and vertical Pitot tube traverses were made at 3-inch intervals except near the walls where the spacing was reduced to 1-1/2 and then to 1/4 inches from the walls. The center of the tunnel was taken as the baseline station (0, O) with all distances measured from this point. Also the velocity at this location was taken as the reference velocity and all velocities are expressed in terms of this velocity. Results are shown in Table 8. Since the tunnel test section dimensions are 24 x 24 inches, station 11-3/4 is 1/4 inch from the wall. At a distance of 1-1/2 inches (station 10-1/2) velocities are generally within 3% of the reference velocity. At all points 3 inches or further from the walls (stations 9 to 0 to 9) velocities are within 1.5% of the reference value. 7. OPERATING EXPERIENCE AND PROCEDURES 7.1 Radiator Installation Since the frontal area of all of the test radiators was considerably less than the wind tunnel test section area, the radiators were mounted on baffles with a constant area section located immediately upstream of the radiator. A bellmouth in the inlet end of the constant area section was used to insure proper inlet flow conditions. Sealing was used around the radiatorbaffle joint and also the baffle-test section wall joint to insure that all of the airflow which entered the constant area section flowed through the radiator. This installation is shown schematically in Figure 45. 7.2 Stabilization of Radiator Water Inlet Temperature The radiator inlet water was heated with low pressure laboratory steam with the heat exchanger. With only a manual throttle valve in the steam line an oscillation of some 100F double amplitude with a period of approximately 10 minutes was observed. This temperature variation reflected the response of the steam supply boiler to the boiler controls. A steam regulator was installed upstream of the heat exchanger and a combination of the regulator setting, heat exchanger throttlevalve position, and bypass cooler operating condition resulted in a maximum variation of radiator inlet water temperature of 0.5~F. 31

TABLE 8 Wind Tunnel Test Section Velocity as a Function of Position* in Terms of Reference Velocity,** V/V Reference Vertical: Horizontal Distance from Vertical i, in. Distance from Horizontal, 11-3/4 10-1/2 9 6 3 0 3 6 9 10-1/2 11-3/4 in. 11-3/4.714.748.791.781.773.797.812.767.771.777.718 10-1/2.869.977.984.983.992.990.990.990.980 ~977.917 9.936.988.981.985.981.981.981.982.979 ~979.863 6.945.995.998 1.0.994.996.987.985.981.975.822 3.946.992.995.990.988.992.985.985.985.981.848 0.925 1.004 1.o 1.0 1.0 1.0 1.002.999.996.996 ~951 3.987 1.004.999 1.003 1.003.997.988.987.987.987.925 6.901.999 1.0.998.995.995 ~993.992.989.992 ~957 9.872.992.995 ~993.995 ~9973 995.9976.991.991.987.944. l10O-1/2.789.990.992.764.997.990.987.985.979.977.845 11-3/4.597.743.790.801.867.867.867.867.828.814.759 *Tunnel test section 24 x 24 in. **Reference velocity, measured at mid-point of test section (station 0,0), = 77.1 ft/sec.

7.3 Typical Operating Procedure a. Fill system with water. b. Operate pump while continuing to add and bleed water and air from the system. c. When bleeding of air is complete, turn on steam supply to bring water temperature to about 1850F, while keeping a small addition and bleed of water. d. Start fan and when steady state temperatures have been established shut off water addition and bleed, thus isolating the system. e. Set desired air velocity through test section by means of bypass opening, set desired water flow rate using throttle valve, and set radiator water inlet temperature by adjusting setting of steam pressure regulator. f. When steady state values have been reached, record radiator water and air temperatures and pressures and flow rates. g. Repeat steps (e) and (f) for successive data points. 33

APPENDIX C METALLURGICAL STUDIES 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. 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 100~C. The tensile strengths were determined using a constant head speed of 0.5 in./min. The results are given in Table 9. TABLE 9 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 40o 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, 0.010% 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 10.

TABLE 10 Composition Tensile Strength, psi 15% Sn, 85% Pb 5,270 20% Sn, 80% Pb 5,730 30% Sn, 70% Pb 6,810 40% Sn, 60% Pb 6,890 50% Sn, 50% Pb 6,400 Table llgives the results reported by Turkus and Smith4 They did not give the details of their tests. TABLE 11 Composition Tensile Strength, psi 20% Sn, 80o% Pb 4,940 30% Sn, 70% Pb 5,390 40% Sn, 60o Pb 5,660 20%o 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 12 TABLE 12 Composition Tensile Strength, psi 15%o Sn, 85% Pb 13,300 33% Sn, 67% Pb 17,100 40% Sn, 60O Pb 14,100 50%0 Sn, 50o Pb 23,900 63% Sn, 37% Pb 29,000

Gonser and Heath2 prepared lap joints using 70 copper, 50 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 13. TABLE 13 Tensile Strength, psi Composition Open System Closed System 15% Sn, 0.25% Ag, 0.015% Bi, balance Pb 4,265 4,050 20% Sn, 0.25q 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 5,240 635 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 14. The details of their test procedure was not given. TABLE 14 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 15. 37

TABLE 15 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 40o% 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,o60 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 16. TABLE 16 Composition Shear Strength, psi 15% Sn, 85% Pb 5,640 33% Sn, 67% Pb 6,450 40o% Sn, 60% Pb 8,280 50% Sn, 50% Pb 7,580 63% Sn, 37% Pb 8,ooo 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 17 give the stress to produce a strain of 1 x 10-4 per day.

TABLE 17 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 40% 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%9 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 18 are the 500-day stress-rupture strengths. TABLE 18 At Room 40o Sn, 2% Sb, Balance Pb Temperature, psi At 80~C, psi For steel Joints 325 120 For copper Joints 390 120 For brass Joints 470 120 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 19 and 20 give the maximum stress level which did not produce a decrease ~in the tensile strength. The mean stress used for the tests of Table 19 was 600 psi, and for the tests of Table 20 it was 900 psi. 39

TABLE 19 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 20 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 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. 2. LITlERATURE 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. Res

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 21-23. TABLE 21 Current density 500 amp/sq ft Initial temperature 35~C Tensile Strength, RPM lb/sq in. 1,750 37,000 2,500 49,000 3,500 51,000 5,500 58,000 TABLE 22 Initial temperature 20~C 2,500 rpm Tensile Strength, Amp/sq ft lb/sq in. 300 60,000 400 68,ooo 510 40,000 1,100 35,000 1,700 14,000 41

TABLE 23 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 24. TABLE 24 5,500 rpm Current density 500 amp/sq ft Tensile Strength, * CuS04 5H20 H2S04 lb/sq in. 12 15 60,000 20 15 58,000 25 15 55,000 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 25.

TABLE 25 5,500 rpm Current density amp/sq ft Tensile Strength, Temp., ~C lb/sq in. 25 63,000 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 22 and 23 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 26. TABLE 26 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 0.366 26.5 33 0.305 24.9 35 Shakespearll discussed the Anaconda commercial process for producing electrosheet copper and stated that the tensile strength varies from 50,000 to 40,000 psi, and elongation from 15 to 25%.

Altmeyerl2 reported a tensile strength of 39,000 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 27. TABLE 27 Mechanical Properties of Acid Copper Deposits Tensile Elongation Hardness Addition Agent Adi Agn Strength, psi % in 2 in. V.H.N. None 36 9 150 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 Read14 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.ll1 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/1 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: 44

Tensile Strength, psi Thickness 25, 50'' 754 Minimum 28,400 29,800 30,700 Maximum 33,000 33,200 32,300 Average 31,200 31,600 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/1 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/1 bath. Such17 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 S04.5H20, 50 g/l H2SO4 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 9,500 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 S04 5H20, 60 g/l H2S04, rhen plated at 350C and a current density of 30 amp/sq ft.

TABLE 28 Physical Properties of Deposits Fluoboric Yield Tensile Elongation Rockwell, g Acid Temp., Point, Strength, % in Hardness g/lg/1 psi psi iT 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 29,4oo00 15.5 Avg 20,230 29,500 14.5 60 4 120 12,800 17,000 8.o 44-45 13,200 17,200 7.5 12700 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) 20,200 300 12.0 Avg 21,480 30,300 10.5 *With 1.2 ml/1 molasses. **With 2 g/l Dacolyte.

TABLE 29 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 30. TABLE 30 Test Temp._ C.D., Test Type of Bath Composition pH Temp., C.D., 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 Thickness, Change of Calculated Stress,* Test No. in. Deflection, psi (wt area/den) in. Steel Base Bronze Base 1 0.0010 0.0002 1400 0 2 0.0005 0.000ooo8 9900 7200 3 0.0005 o o 4 0.00154 0.0013 5200 *The stress values are all positive or tension stresses. 47

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 31. It can be seen that the residual stress values vary from 15,400 psi in tension to 5,000 psi in compression. TABLE 31 Cathode Stress in Test No. Coefficient, Deposit, Variable Studi/ed 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.o 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.o - 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. 48

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 32. TABLE 32 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, Modulus, from Ref, Modulus, from Ref., as psi x 106 Lpsi 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: 49

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. 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. 3. TENSILE AND FATIGUE TEST ON ELECTROFORMED COPPER Tesile tests were performed on copper which had been electroformed at the Savage Rowe Company in Kalamazoo. The plating conditions are given in the table in Appendix D. The test samples were standard 0.5 in. wide strips and were run on an Instron self-aligning machine. The fracturing of the specimens appearedto 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 percentages 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 were 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 thesheet 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. F1 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 G5 0.0131 28,090 7.3 Avg 27,818 9.0 The microstructure of the copper is shown in Figs. 49-52. 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. The fatigue results were obtained using a Krouse machine shown in Fig. 53. 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 thousands to almost two thousands 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 re51

sults. The fatigue data are shown in Figs. 54 and 55, and in Tables 33 and 34. TABLE 33 Fatigue Data —0.010-in. Nominal Thickness Sheet Specimen Stress, Average Load, Number of Cycles No. psi Thickness, lb to Failure in. 107 11,500.0103.0739 10,000,000+* 108 11,500.0103 0739 10,000, 000+ 109 11,500 0131.1196 10, 000, 000+ 110 11,500.0127.1124 10,000,000+ 113 11,500.0100.0697 10,000,000+ 101 12,000.0132.1267 10,000,000+ 102 12,000.0128.1192 414,100 103 12,000.0130.1229 10,000,000+ 104 12,000.0135.1362 10,000,000+ 105 12,000.0102.0757 304,400 106 12,000.0105.0802 415,300 128 12,000.0102.0757 10,000,000+ 129 12,000.0110 0880 538,900 130 12,000.0110.0880 10,000,000+ 114 12,500.0105. 835 2,663,700 115 12,500.0101.0773 162,600 116 12,500.0103.0804 14,700,000+ 117 12,500.0115.1002 2,216,100 118 12,500.0113.0967 9,124, 600 119 12,500.0113.0967 8,234,600 125 12 500.0101.0773 717,400 126 12,500.0115.1002 3,982,800 127 12,500.0111.0933 10,000,000+ 120 13,000.0117.1079 424,500 121 13,000.0109.0936 371,000 122 13,000.0110.0953 320,400 123 13,000.0117.1079 2,305,500 124 13,000.0118.1097 710,100 *+after the figure indicates the specimen did not fail.

TABLE 54 Fatigue Data-0.020-In. Nominal Thickness Sheet Specimen Stress, Average Load, Number of Cycles No. psi Thickness, lb to Failure in. 1 10, 000.0263.4097 12,000, 000+* 12 10,500.0259.4269 767,400 15 10,500.0220.03080 10,000,000+ 22 10,500.0260.4302 10,000,000+ 14 10,500.0271.4674 10,000,000+ 19 10,500.0273.4743 10,000,000+ 8 11,000.0180.2160 10,000,000+ 7 11,000.0225.2275 9,791,500 9 11,000.0223 ~3315 4,943,500 18 11,000.0210.2940 10,000,000+ 13 11,000.0263.4611 5,427,400 16 11,000.0223 ~3315 10,000,000+ 2 11,500.0242.4017 1,594,200 24 11,500.0183.2334 10,000,000+ 10 11,500.0208.3015 10,000,000+ 39 11,500.0258.4639 4,904,200 37 11,500.0251.4391 10,000,000+ 17 11,500.0299.6231 10,000,000+ 4 12,000.0200.2909 10,000,000+ 33 12,000.0243.4291 1,276,200 6 12,000.0244.4330 1,060,400 5 12,000.0213.3300 535,200 20 12,000.0195.2765 1,479,700 27 12,000.0220.3520 10,000,000+ *+After the figure indicates the specimen did not fail. The endurance limit, based on ten million cycles, is 11,500 psi for the nominal 0.010-in. thick copper, and 10,500 psi for the nominal 0.020-in. thick copper. A number of broken fatigue specimens were examined metallographically. Fatigue cracks were observed on some of the specimens and these were always on the electrolyte side of the specimen. Figures 56 and 57 show one of the cracks on the unetched and etched condition, respectively. It can be seen 55

that the crack originated at a rather sharp irregularity in the surface and progressed diagonally across the columnar grains,. The surface of.the electroyte side of the copper was much more irregular than the mandrel side. Figure 58 shows this difference and also shows a pronounced change in deposition characteristics, i..e.,. grain structure and orientation, resulting. from a foreign particle attaching itself to' the surface shortly after electrodeposition had started.3..1 Conclusions a. The tensile strength of the 0.010 inch thick copper was approximately 500 psi higher than that for the 0.020 inch thick copper.. b. The endurance limit based on ten million cycles is approximately 1000 psi higher for the 0.010 inch thick copper than that for the 0..020 inch copper. c. A microscopic examination of failed fatigue specimens run.in reverse bending indicated that.the fatigue failures originated. on the electrolyte side of the electroformed copper. 54

APPENDIX D ELECTROFORMING PROCEDURES FOR HEAT EXCHANGER FABRICATION Prepared by Mr. Frank K. Savage, President Graham, Savage and Associates, Inc. Kalamazoo, Michigan Savage Plating & Anodizing Co., Inc., 1724 Clinton Street, Kalamazoo, Michigan, electroformed components for two (2) tube and fin heat exchangers which could be used for automobile radiators under direction of Frank K. Savage and Charles H. Bommerscheim per consulting agreement with Graham Savage & Associates, Inc. The first heat exchanger was made by Savage Plating & Anodizing Co., Inc., complete. The electroformed components of the second heat exchanger were made by Savage Plating & Anodizing Co., Inc., and the assembly was made by The University of Michigan. The electroforming was accomplished in the following solution and under the conditions delineated. Fluoborate Copper Bath Cu(BF4)2 60 oz/gal Cu(Metal) 16 oz/gal HBF4 6 oz/gal Density Be 37-1/2 to 39 at 800F-control to correct low Be~ was to add Cu(BF4)2 Temperature 100~F Agitation Vertical oscillation 60 cycles per minute with an 8-in. stroke Current Density 75 Asf Note: The range of current density in this solution, with reasonable agitation and proper cooling equipment, is up to 400 Asf. No attempt was made to go this high in the prototype production. At 400 Asf the solution heats far above the normal limits of operating temperature. This requires cooling equipment not thought to be justified for prototype manufacture. Speed did not seem to be a factor at this time.

Both prototypes were produced by electroforming a very thin film of copper over a zig-zag copper fin material (spacer stock) which is available and presently used in the manufacture of the present tube and fin construction of automobile radiators. The zig-zag of copper foil approximately one-half inch wide by 2 inches deep by one foot long was inserted in a steel mould which would just contain it. The mould was poured full of a low eltingalloy of bismuth and tin. The first used was Cerrotrue. A primary problem was effective coating procedure as follows: 1..Lack of fluxing and tinning of fin stock caused boundary voids along the line of contact between the copper fin stock and the cerro alloy. This produced a reverse meniscus at the junction which penetrated deeply into the composite coating, producing excessive porosity in the resultant overlay of electroformed copper. 2. There was pin point porosity throughout the cerro alloy which prevented the propagation of a continuous film of copper cover plate within the thickness of the deposit allowed. 3. One and two above, point out two unsolved problems; first, work should be done on fluxing techniques so the cerro alloys will wet the copper and fill all the voids not leaving deep meniscus impossible to bridge without porosity. Second; the pin point porosity which is inherent in the cerro alloys when casting under atmospheric pressure must be eliminated by pressure coating. This technique is known to Cerro Sales Corporation and was made available to Savage Plating & Anodizing Co., Inc., during this project. Neither time or money was available in this project to evaluate either of the above. However, without any budget, several tinning methods were tried with various fluxes. Not enough time was available to achieve comprehensive results. There were some definite indications that further work could produce very good results. In the case of porosity, in order to get some semblance of a pressure coating the mould was deepened by 2-1/2 inches to create head pressure. Further a cold water quench technique was developed to expand the cerro alloy at the surfaces and to push the porosity to the center of the casting. Definite improvements in soundness of castings was achieved by each of quenching and the additional head pressure. This indicated progress but the castings were still unsound. Upon the recommendation of Cerro Sales Corporation the alloy was changed from cerrotrue to cerro cast. Cerrocast expands to a greater extent upon freezing, thereby filling the mould better. This made little extra improvement in the soundness of the castings.

In order to attempt to eliminate surface defects in the casting, several thousandths of stock was milled from each side of the casting. This eliminated most of the cold shut areas,eliminated or shallowed the V shaped voids of the reverse meniscus. At the same time equally bad defects of hair line voids were opened up along the edges of the milled copper fin. The feed and speed of the milling cut was explored. The speed made little difference. Slightly better results were obtained when the feed was made very slowly by hand. Dry peen blast after milling helped close some of the voids and flattened the top of the copper fin stock which produced an area two to three times as wide on which to anchor the electrodeposit. This technique was, however, only a compromise. In the entire program a large number of castings were made to get the few needed for the two (2) prototype radiators. It was necessary to repair all the castings used in the program with a very small needle point soldering iron. The conclusion drawn by the Graham, Savage & Associates, Inc., and the Savage Plating & Anodizing Co., Inc., group are as follows: 1. This is not a practical design from an electroforming standpoint. The product contains a low percentage of electroformed copper. Electroforming approximately.005 inch of copper cover plate, normal to vertical fins of rolled copper, which results in a junction of virtually right angles is unsound practice. 2. In any similar future work, where thin copper fin stock is used, pressure casting technique should be employed. 3. A fluxing and tinning technique should be developed for use on any copper base insert in a cerro casting.

REFERENCES 1. W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Califo, 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, Jro, "Low Tin Solders Containing Silver and Bismuth," Metals and Alloys, 15, p. 412 (1942). 5. F. N. Rhines and W, A. Anderson, "Substitute Solders," Metals and Alloys, 14, po 704 (1941). 6. J. B. Russell and J. 0. 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, Noo 2, po 277 (1939) o 8. M. McKeown, "Properties of Soft Solders and Soldered Joints British Non-Ferrous Research Association, Research Mograph, Noo 5, Po 57 (1947) 9. 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), pO 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),pO 367-3700 Chemical Abstracts, 28, 6640. 135 Co Eo Huessner, Ao Ro Balden, Lo Mo Morse, "Some Metallurgical Aspects of Electrodeposits," PLATING, 35, Po 554-557, 719-723, 768 (1948)

REFERENCES (Concluded) 14. To A. Prater and H. J. Read, "The Strength and Ductility of Electrodeposited Metals," PLATING, 36, p. 1221-1226 (1949).o Ibidem, 37, p. 830-834, 850 (1950)o 15. N. P. Fedotev and Iu. M. Pozin, "Effect of Surface-Active Substances on the Mechanical Properties of Electrolytic Deposits," J. of Appl, Chem, of the USSR, 1, p 406 (1958). 16. C. Sruyk and A..E. Carlston, "Copper Plating from Fluoborate Solutions," Monthly Review of Electroplaters' Soc., 3.3, p. 932-934 (1946). 17. T. E. Such, "The Physical Properties of Electrodeposited Metals," METALLURGIA, 56, p. 61-66 (1957). 18. A. Pocalyko, M.S. thesis, Pennsylvania State University (1951). 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, po 43-49 (1956). 23. H. J. Read and A. H. Graham, "The Elastic Modulus and Internal Friction of Electrodeposited Copper," Electrochemical Soco Journal, 108, P. 73 (1961). 24. P. Hinnert and H. S. Kr.ider, "Thermal Expansion. of Some Copper Alloys," Journal of Research, National Bureau of Standards, 39, p. 419 (1947). 25. Kreith, F., Principles of Heat Transfer, International Textbook Co., 1958.

..~~~. ~ ~. H, ~~~~~~.. ~ ~ ~ ~ ~ 4~~~~ U)~~e U) 7. ~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~ 2~~~~~~~~~~~~~,. H.~~~~~~~~~~~~~~~~~~r C) a) ~ ~ \ H H bD

:::::;:-:,.:i;:i:(~~::;:::;;_s;:,i;i~, iiif:~iiii;ii:~ilii-:i-:.~SiZ~~ii;:i~p::-:::;:::~::::::;:: iii:~ia-sii:iiiiiiii:: - B esas8aas' - gi:: li%iiii:i:::::: i ::':ssllaareasPpastii iii.:ii')21 1L-L131~pl~LSIP~~~~~~~~BpidBB?lr:: I: j:i I r I%:iiTa I IllbllllL-~r ~slp8ssaan,r~*lmr::(:::i::::::j;:::r.~:'a:Qi: -:;I~~::::::i::i~~~::::i::::/:::::::::.-~::-I::':~:: j:jl~-':::'':i::::.::::~:::ji::::::,:.:i:'::::::.:::::~:..:.iiiiil:l:8i:~.;:.ii; " Iiiifi:.i:iiii::~:::::r:::a "'::iiiiii:i~ii~li:::.:::,I:: Blii iiisiiai idIlli iiiiiiiir'i'iii?iifi'8i3iiiiii'~;:'';~;~ bii~i~ii:::::._::: aBe881'II,:I:::::::"' r a9ss&:l:p:ii:'::~::~ iiii Iiiiiii CsR~'L:I""":'::i:j:ii8lsis:::::r ~::i~:~: iii'':il3iliSiillE i::iii:iiiii:::~i:1'111::-:1'::~::.:~::.i::::::i:I:;:i:l::::::~rr:::::li'i;,l:i:::li:~:-s:~.:i:i: iii ~ire:r:~r~::~a iiiiiIbiiE:iii:"'"''~'::::;~:::::::'':'::":::::':';~:~.:; ~i::i::i':ili"':::':'rii:il': ~ ~:.~::l::::l.:i.'ii~i:::::,.':.::'. ~:::~;:::::r:::~:;:::~I~; iai$~::::~:~::::::::~~;~:~.:::::1: I"'~":i ilS::~:::;:::;-::::I::::: 1-:rll::;:::...::i.-.:::::;:: ~::::.li~r~: ~::':::::::'::':::':;:::iR1:iii:::::-:-:::..~. ~:i-: — -':::::: —::::i:i'i:::i:i nl~iQI1"1` i,-~~~.a~~~,,,,,,; ~il:~:~::.~:: ~: ~:~ ira:i:ciii: i:r::::::::::::~::.. a~~l,_,: ~' 4~iPI'"i",_*i-:il."'Lr" ig a I,:a?:~~!~i::~::~~, _,?:- —_i::~::li:::'::::'iiiiiiiii~ji:: r:i::i:r:I::~:::::::;-i~rii:i-r-i::r:::~i:;i'::~:i: r:i::.r:i::jg~~ii':ji:i: iti9ib:n:ii:lij':::::::::::::::':~ ~~'::':' ~:~:: ~:~;'":':': Fig. . Partially electroformed automobile radiator (heat exchanger)

0.1 S PHX EHI LMHX- FHX- I I o 01 c, 001102 103 104 Rec Fig. 3. Summary heat transfer results for all radiators. 63

1.0 f EFHX- I 0.1 LMHX EFHX-II S PH.01 102 103 10' Rec Fig. 4. Summary friction results for all radiators. 64

1.0 S PHX EFHX- II I _ ~~~~~~~~~EFHX-I LMHX 0. 1 L..01 102 103 104 Rec Fig. 5. Summary heat transfer/friction results for all radiators.

CORRUGATED C PPER A ///2" 1E I B —---— 7bnFL~nn~LU~l___ I ~ 25] 3/32"SPAER ACERSl SOF1 T SOFT SOSLDER T E I." 5/32HOLE 6" ~I.WATER CHANNEL ILVER I "COPPER PPE OLDE20 AIR PASSAGES 2" 19 WATER PASSAGES 1O B I I 112- I 14 I 4 MOUNTINGE D LUG SER CLVEVER PLTE SECTION AA I SOlDER COvER PLATE A-24 END CAPS 2 SOFT SOLDER NOTE: ELECTROFORMED SURFACES E TO BE.007" TO.008" THICK. ASSEMBLY TO BE TESTED AT 25 psig CORRUGATED COPPER SPACER TO SOFT7 SOLDER HAVE APPROXIMATELY 10 FINS per INCH. TO BE SUPPLIED BY MR. RICHARD D. CHAPMAN 4"CCOPPER PIPE EFHX-I GRAHAM-SAVAGE S ASSOC. KALAMAZOO, MICHIGAN ~~~~~~SECTION BB PROTOTYPE HEAT EXCHANGER DESIGN FOR INTERNATIONAL COPPER RESEARCH -DRAWN- FKS SCALE 1/2 DATE 2-16-63 NO. 6SA-2 Fig. 6.

. 50 A'" 1 112 NPT Bronze Fitting 2 Places-Centered on Unit.25 dia. 6 holes spacing to suit 1.00 Mounting 1t - 13.70_ —— (. 70 Flange 2 Places Section AA Fig. 7. EFHX-II heat exchanger.

Make Top and Bottom Tanks and,/ I, Fittings of Copper 12" Long, 10 Fins/inch as furnished by RDC Material has been sheared to fulfill these requirements Tube Tube to be Copper. 008 wa ll Inside Dimensions to be 2 2"x 3/32".008 Cu.-i -H K-Channel 20 Air Passages * Spacer 19 Water Passages Each End *Air Passage at each Tank per 43052 end place copper strip over end air passages rad. flpacers must be This side of Tank make Extend 1/4" This side of Tank'" 2 make Extend 1/4" Fig. 8. SPHX heat exchanger. 68

Electroformed Sides 0. 434, Fin Stock (simplified) 10 fins/inch Air Channe CD 0012 A0I458 Direction of Water Flow (typ.) 0. 094 Fig. 9. Section of EFHX-I showing air channel. 69

Electroformed Sides O. 434 Fin Stock (simplified) 10 fins/inch 0 0.0121 p V Direction of Water Flow (typ.) 0. 094 Fig. 10. Section of E II showing air Channe. To 0.~35~~~~~~~~l O. 012 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ —-,. C] Dieto fWtrFo (typ.)~~~~~~~~~- ~v 00 Fig 1. Scton f FH-IIshwin ar canel

Electroformed Sides \ 0.480 Fin Stock (Simplified) 10 fins/inch Solder. _ ~t 1;1 Ityp. 0 Air C annel 6 06-0.260 a= 0.010 0.500 (typ.) 099Direction of Water Flow,. 099 Fig. 11. Section of SPHX showing air channel. 71

.040 Electroformed Subassembly 2" 0. 400 _ o =.030 - Air Chamber.020 Permanent Deformation.010 0 - 0 10 20 30 40 50 60 p, psig Fig. 12. Mechanical testing of an electroformed sub-assembly. 72

0. 1 0 0 r') N; 8.01 102 103 Rec Fig. 13. Heat transfer data for EFHX-I.

1.0 0. 1 2L1 104 102 103 10 Rec Fig. 14. Friction data for EFHX-I. 74

0.1.01 01 102 103 104 Rec Fig. 15. Heat transfer data for EFHX-II.

1.0 0.1 I- I I 1 1 III I iii!.01 2 104 10 10 Rec Fig. 16. Friction data for EFHX-II. 76

0. 1 1.................... a.01 4-, 1102 JIIIL_ 001 102 104R Rec Fig. 17. Heat transfer data for SPHX. 77?

1. O 0. 1 I I I I I I I I I I I I I I I 1 I. 10 Rec Fig. 18. Friction data for SPHX. 78

ro N- 901.001 I I I I I I I I 1 1020 103 10 Rec Fig. 19. Heat transfer data for IMHX. 79

1.0 0. 1-.01 II I I 1 I{ I I I II. I I I I I I I, 102 103 104 ReC Fig. 20. Friction data for LMX. 80

0. 1 2 O" 0.050 D- 0. — — 012 120" 0. W1 cznzzzD czzzzD 1zI ~ o.I —2o 1.o' 6 0.04,0 0, I. 050 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~501 1 ___ __)_. 00.870 -ll - 0. 87o,, 00 - -- Nio3(0.1003 - -/uR-'/, -D ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0I 03 0 0.6 0.8".. o 3. 4 & a 00 oo 8~~~~~25" 0.0209 0.015~~~~~~~~~~~~~~~~~~~~~~~~~~.2 a 0.0 1 FINEF-A-BEST INTERPRETATON 0.0IS \, —E S T 1NT=RRRE'Tj%1 1 10N ~ P 1 0.0' 030'0' 0 ~L ~~~~~~~~~~~~~~~~.008 "~~~Fnpth- 9.68,,princh r~~~~~~~~~Oo 0.006.. F m.006 ea 0.005 I -_ 0.004 -re-lwarafona re 0.9 inmtl hcnes-0.0 n r 0.004 r~~~~~~~~~~~~~~~~0.0 0.003 ~ota.ettase Fe-owrafrnaoae 0.81012 aetoa 1 Total heat transfer are/tta volume -49 f(4rhG/t./ft 0.4 0.5 0.6 0.8 1.0 1.5 2.0 3.0 4.0 1 o 0 0. 4 0.5 0.6 0 1.0 1.5 2.0 3.0 4~0 50 6.0 8.0 IQ 9.68 - 0.87 9.6a - 0.79 F a WoM asadA Fig. o21. Fi. 22 Pl Ao ClBE. 15 (bNpEr.Iso FI)NED FLAT TUBES SURFACE 9.68 - 0.81 SURFACE 9.68 - 0.87 - R Fin pitch - 9.68 per. Inch nun pitch - 9.68 per inch nowr passage hydraurlic diameter - 4rhrO.01180 ft. Flow passage hy~draulic diam~eter - 4r.70.01180 ft. Fin metal thickness - 0,O0in. Fin metal~ thickness - 0.004 in. Froe-fl aw area/irontal area - or.0. 691 t! Fee-flow area/froutal area - a-.0.6972 Toalhet rasfr re/tt~ vlue (X22 f./ Total heat transfer area/total volume 4.229e ft./ft! Fin area/total area - o.795 Fin area/total area - 0.705 W. M. Kays and A. L. London, Compact Hleat Exchangers, The National Press, Palo A-Ito, Calif., 1955 (reproduced by permission).

0.100" O.OG-i,,~a.__.~ —0.060 -9" - - N — n0.55" o.o" 0.04 -- _________.050", -I C-0 0.737-N ["1 ) o 0 I02.08".4- I IIIII~ aI I I I II1I 1 1030.020 0."15 BEST INTERPRETATION.015 - EST INTERPRETATION Fig. 25 Fig. 2)i-~~. o 0.0.0, o INTE EATION.01BEST INNERPFEATION 0.00 ~ a. - z'.01o~,_' 0~~~~~~~~~~~~~~~~008' F.00 0.0 PaloAlto, C NRa. 19 (r4eorhGoc b 3riso) I I I I I O~~~~~~~~~~~.004 ~~~~~NR x lo- (4rh G/,u) 04 5 0.650.8 1.0 1.5 2.0 3.0 4.0 6.0 8.0 II10) 0.4 0.5 0.6.8 1.0 1.5 2.0 3.0 4.0 60 8.0 9.1 - 0.737- S 9.29 - 0.737- SR Fig. 23 Fig. 24 FPNNED FLAT TUBES FINNED FLAT TU)BE SUIRFAC] 9.1 - 0.37 - S SURFALCE 9.29 - 0.317 -SR Fln pitch - 9o1 per inch Fii pitch - 9.29 per inch Flow passage hydraulic diameter - 4rh-0.01380 rt. Flow passage hydraulic diameter - 4rh-0.01352 rt. Fin metal thickness - 0.004 in. Fin aetal thickmess - 0.004 in. Free-flor area/frontal area - r.0.T88 Free-flow area/frontal area - d -0.788 Total heat transfer area/total volume -0=224 ftf/ft3 Tota1 heat tr..nsfer area/total volume -oc=228 fOt/ft3 Fin aroa/total area - 0.813 Fin are&/total area - 0.814 W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission)

0.100" zzcz 0.07 G- - - - - - ____ - T~cz~D III __ 0 0.060 --:"1437.06 0.050-I 0 I - I I c IIk-0.79J. 1 "0.08" - s0 0.040.01 0 C f 0 E. 5 - I4.k-O )2 0.040 0-.40. 0.030 - -B125"-ES NTERPRETAON I.66' 0.030 0.020 0.020 \ Q01I BE ST INTE RPRE TA TI ON OD010 i O.Olq-C, 171 Isr~~~ I i 1 I I I I 111 1~~0.008 Z co 0 0.00 0.004 -NRX O N~~x~o~ (4r GI) 0 - - 0_ 0.4 0.'6 0.8 1.0 2.0 3.0 4.0 6.0 8. IO 0.004 NR x 10-3 Gp 0.4 0.5Q0.6 0.8 1.0 1.5 2.0 3.0 4.0 6.0 8.0 10.0 Fig. 26 11 - 32 - 0.7 37- SR FINNED CIRCULAR TUBES Fig. 25 SFACE 8.0 - 3/8 T (Data of Trane Co.) Tube outside diameter - 0.40.2 in. Fin pitch - 8.0 per inch FINNED FLAT TUBES Flow passage hydraulic diameter - 4rh70.01192 ft. SURFACE 11.32 - 0.737-SR Fin thickness - b.013 in. Free-flow area/frontal area - or 0.534 Fin pitch - 11.32 per inch Heat transfer area/total volume -( 0.179 ft!/ft? Flow passage hydraulic diameter - 4rh0.01152 ft. Fin area/total area - 0.839 Fin metal thickness - 0.004 in. Notet Minimas free-flow area in spaces transverse Free-flow area/frontal area - 0 -0.780 to flow. 2 These data are included in this compilation Total haeatotal/total volume -- 0.845/ft becense they apply to a compact surface conFin area./total care& - 0.845 figuration of considerable technical interest for which no data have been obtained en this project. W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission).

-l/4rhn10.3.030 -- 0.040 I ~.-0.47~1 ____ o0.2216_ - - - _ - -.020 - - a_ _ _,015~~~~~~.015 ----- (P4rh)= 65 k- 823-* 0.020-'I, -.010 0 ~~~~.008 --- _ _ _ 0.019.006 0.008- - EST INTERPRETATION-SIRE O _____ ____________ - - EST INTERPRETATION IcQ _ 0.006ell _ __ _ _.00 3> —- 0o 0.005 Z_- _ _ N /. - ____ CL IIII "~L.003 (3~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1 0003 NG)3_.002 — )N___ (4_rh___ __NR 11q3 (4 rh G /,0.6 0.8.0 L5 2.0 3.0 4.0 6.0 8.0 10.0 15.0 0.4 0.5 0.6 0.8 1.0 1.5 2.0 - 3.0 4.0 5.0 6.0 80 1tO 9.03 5.3 Fig. 27 Fig. 28 PLAIN PLATE-FIN AIN PLATE-F SURFACE 5.3 SFACE 9.03 Fin pitch - 9.03 per inch Fin Pitch - 5.3 per inch Fin pitch - 9 03 per ih Plate spacing - h=0.470 in. Plate spacing - b.0.823 in. Flow passage hydraulic diameter - 4rhro!02016 ft. Flow passage hydraulic diameter - 4ri..O.01522 ft. Fin metal thickness - 0.006 in. h ~~~~~~~~Fin metal thickness - 0.008 In. f23 Tin metal nficknesa - 0l006 in. p 8 / Total transfer area/volume between plates -,9a ft4/ft. Total trsnsfer area/volume between plates fr188fd /r Fin area/total area - 0.886 Fin area/total area - 0.719?'/4rh)-10.3 W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission).

~~~~~.05~~~0 —----- 1 7~~~~ -rLLI~ -~ -I~ 1~ —~ 1~i/~54rh04, 350.040. GN - - - - 0 _ _ _ _ 0.1326 4.030 0,020 020 (1Xkrh)r 65 0.418t ~.015 1 1 I I- -0 - N. - 003.010.. I I.015 o008 ~'~~ 91. I I I I I \ I I I I I i.008 PLN PL BEEST INTERPRETAT ION F ps hri it- r-BEST INTERPRETATION.006. Fi rattlaea-080-i re/oa ra-005. CL~ z z 0. K00a o Palo AO C3 - on.003 0 NRx 16-3 (4rhG/pJu N! 4 1&3 (4rG/ 0.4 0.6 0.8 [0 15 2.0 3.0 4.0 6.0 8.0 0 0.50.6 408 1.0 1.5 2.0 3.0.0)u 5. 6. 15.08 Fig. 29 Fig. 30 PLAUN PLATE FIN X PLATE F]3 SURF&CE 15.08 SURFACE 19.8N Fin pitch - 15.08 per inch Fin pitch - 19.86 per inch Plate spacing - b=0.418 in. Plate spacing - b10.2p0 in. Fln pa ssage hydraulic diameter - 4r. h0.00876 ft. Flor passage hydraulic diameter - 4rh=0.00615 ft. Fin metal thickness - 0.006 in. Fin metal thickness - 0.006 in. Total aransfer area/volume belroon plates - 0 -414.f2ff3 Total transfer area/volume between plates -P.561 ft2/ ft! Fin area/total area - 0-8F0 Fin area/total area - 0.849( ~y4rh)-865 y )3. W. M. Kays and A. L. London, Compact Heat Exchangers, The-National Press, Palo Alto, Calif., 1955 (reproduced by permission).

*.9 4 A 0054 1-0 0 _ _'-f —XX r.0874.-413".0562" 3.0 _ -.6..... 04 ——.08.08 Plate -06-X -0.07" A R. ~.02- - -- - ~ - - _,_ BEST INApRETAION.[ -:.ft. F_ _ _- i 0 5 z- -.05_ C-C 0.008- -.075 AP RO X. 00r.04 ~.006-1- iI N.006. 0.3 0.4 0.6 0.8 1.0 1.5 2.0 3.0 4.0 0 dO 10.0ee 0.4 0.6 0.8 1.0 1.5 2.0 30 4.0 6.0 8.0 10.0 11.44 - 3/8W 178 Fig. 1 Fig. 32 TWA-F IN PLTE-FIN WAV-FIN PLATE-FIN SURFACE 17.8 - 3/81 SURFACE 11.44 - 3/81 Fin pitch - 17.8 per inch. Fin pitch - 11.44 per inch Plate spacing - baO.413 in. Plate spacing - b.0.413 in. Flow passage hydraulic diameter - 4rh.0.00696 ft. Flow passage hydraulic diaseter - 4rh0.01060 ft. Fin metal thickness - 0.006 in. 23 Fin metal thickness - 0.008 in. Total heat transfer area/volume between plates -,.514 ft/ft. Total heat transfer area/volume hetween plates -..351 ft!/ft! Fin area/total area - 0.892 Fin area /total area - 0.847 Note: Hydraulic diameter based on free-flow area normal to Note: Hydraulic diameter based on free-flow area normal to mean 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).

-_ _3z_"-i, K85'.150 - - - — z — -.100-.031"- - - - 0 +4 5___ 11.os1'4-"' = A'.06- - _____.080 ----.010 044 i i i i I\ I I 1.'050 O030- - - - - BEST INTERPRETATION ~~'-.02. -_BEST INTERPRETATIO 0 1~ ___1~ ____I~.020 a. n.015 coa1.013- 12.2 cc _'.050.... u.006 NR x 10 (4rG/J) - - 0.3 0.4 0.5 0.6 0.81 1.0 1.5 2.0 30 4.0 5.0 6.0 8.0.005 0.5 0.6 Q8 1.0 1.5 20 3.0 40 5.0 6.0 80 10.0 1/8-15.2 3/32 - 2.22 Fig. 33 Fig. 34 STRnP-FIN PLATE-FIN STRIP-FIN PLATE-FIN SURFAICE /32 - 12.22 SURFACE 1/8 - 15.2 Fi pitch - 12.22 per inch F'in pitch - 15.2 per inch Plate spacing - bo0.485 in. Plate spacing - b.0.414 in. Fin legth - 0.094 in. Fin length - 0.125 in. Finslu staggered symtrically Fins staggered symnetrically Flow passage hydraulic diameter - 4rhO.0O1120 ft. Flow passage hydraulic diameter - 4rh..00868 ft. Fn metal thickness - 0.04 In i Fin metal thickness - 0.006 in. to h t rur r/l 1 plte -0 Total heat transfer ara/ea/vlmolume between plates - ft417 ft.4ft. Fiu area/total area - 0.862 Fin area/total area - 0.873 Notes PFin leading and trailing edges slightly scarfed from Note: Fin leading and trailing edges slightly scarfed from fi cutting operation. Frietion factors may be lower fin cutting operation. Friction factors may be lover with lean fis. ith cean fins W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission).

.035".055 ~ -25H 0090.100.070.080 0 - - - - - ~~~~~ ~~~ ~~~.060 -_ _ STEADY-STATE TEST DATA.0050.340 - _040 EST INTERPRETATION SIR.040 FLO — 010 _ -~~~~~~~ —.030 15 TUBE ROWS BE-ST INTERPRETATION -2.020 1/1 1 0.015 MaF. ig. 5 010 - Plat, spacing - b.O.280 in..010 -008.008 1 1 1 I 10.006 NR -A 10 N 3 K rh G/bu) 03,NRX IC) (4 rhG/)J) 0.3 0.4 0.5r0.6 0.8 1.0 1- 5 2.0 3.0 4.0 outsid 8.0 e0.6 Q8 I 0 1.5 2.0 W 4.0 6i0 1 8.0n 1. 1/4Hr11.1 yf Figc - 30 FigH 3v LOUVMM M PWM-Pii SURFACE 1/4 - I 1.1 FLOW NOMAL TO A STAGGEM TMJB BANK~ (Steady-state tests) Flar pitch - 11.1 per inch SUIBFACE S-1*50-1.25(s) Plato spacing - boO.,50 In. Lower spacing - 0.250 In Tabo outside diamter - 0.250 In. FLU gap - 0.035 In. Ry~~~~~~~~~~~~~~~Bdrealic diameter - 4M*~0.0166 ft. Louver gap - 0.055 In. Fr~oflowarea./frout&I nrn - W-0.333S Flow parrerge bhydrcolic diameterP - 4rhm.01012 ft. Heat transfer &re&/total volum -O 80.3 rt!/ft! Fla notal thickness - 0.006 La. Total heat transfer area/volum, between plates -,a 361 ft!/ft! Note: Mini free-flow area is in spaces Fin area/total area - 0.156 transverse to flow..W. M. Kays and A. L. London, Compact Heat Exchangers, The NationaL Press, Palo Alto, Calif., 1955 (reproduced by permission).

.060 ToA'.040'o I -15-1.5.030 o t'~ AIR.020 0.3 5' BEST INTERPRETATION.015 035' — a *0.1 1x=.25.010 * co006 NR X 10-3 (4rhG/p) 0.6 0.8 1.0 1.5 2.0 3.0 4.0 5.0 6.0 8.0 100 15.0 I - 1.50 - 1.25 (s) Fig. 37 IlOW NORML TO AN IN-LINE TUBE BANK (Steady-state teats) SURFACE I-1.30-1.25(s) Tube outside diameter - 0.250 in. Hydraulic diameter - 4rh=0.0166 ft. Free-flow area/frontal area -'-0.338 Heat transfer area/total volume -.80.4 ft./ft3 W. M. Kays and A. L. London, Compact Heat Exchangers, The National Press, Palo Alto, Calif., 1955 (reproduced by permission).

o6 rlr o <I< C 00 P.c-') s\ ~ —~ I P rl v4 Round Tubes Continuous Fins CD Cr r CIIw~~~~~~~~~~~~~ rPlate- Fin,,,., Wavy-Fin — [~~~ c J ~~~~~~Strip-Fin \,I~~~~~~_ I Plate- Fin Louvered - Fin ]\ Round Tube Bank Without Fins,. _ _ __ Honeycomb McCord Type GN

114L CCa -o~"~~~~~~~CC co - C3 C -#-.-.,,, ~~~I —-::3 I L.J.'Q I-. ":'~~~ ~ ~ ~~ =JIu Flat Tubes Plate Fins -- >' o a a_ u3ro:3o Continuous Fins Plain Fins. - ~ ~ ~ ~ c - 2.0 Ap, APO 1.0 0{... 1 1,. 1 1 1,, 1.. 1 1 1 I. I, I. I I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Matrix Number Fig. 39

IC 3r C> a I o c Round Tubes Continuous Fins ~ O-0 __. — | -" Pavy-Fini, Wavy- Fin / ~ -~~Plate- Fin P late-Fin Louvered- Fin Round Tube Bank 00 __ __ Honeycomb McCord Type GN

Steam IReservoir Steam In Control Valve in Steam- Cooling Bypass Test Water Water Cooler Radiator Heat out Exchanger Flow Meter Accum u lator Return to Boiler LPump 0 ( Fig. 41

CONTrRACTrION CONE WESTINGHOUSE8030 9:/ RAT O 7 25 HP MOTOR /0000 RPM at 7"SP 2 ET2 VANE FLOW ON' IFFUER At <_ TESTSECT/ON 2'X2~-4"LONG DIFFUSER~. 6' x.O l~- 11 l i p,~ O ig. 422' Fig. 42

1.0 o Measured Performance (Maximum Fan RPM, Full Open Vane Setting) 0 0.6 0a. 6 0 Test Section Velocity, Zero Bypass 70. 2 ft/sec 0C >Lo VN Q, 0, O) -,1. O 1' Predicted Performance 0 0.2 0.4 0.6 0.8 1.0 Axial Travel x Bypass Opening, Test Section Equiv. Dia. 4A Fig. 43

Match Fan I n let Point Speed Bypass Vanes A 100 0 Full Open B 100 0 Part Closed C 100 Part Open Full Open D 100 Part Open Part Closed E 50 0 Full Open F 50 Part Open Full Open Fan Characteristic 100% Speed System Full Open Vanes Characteristic 0o' i |(zero bypass) Pressure C Fan: 100% Spee Rise Rise ] // Part Closed Vanes D Fan Characteristic System Bypass |E/ 50% Speed Characteristic Full Open Vanes Volume Flow Rate Fig. 44

Water Out By-Pass Flow Total Pressure Pitot Tube Constant Area I nlet Section Test Test ~~~~~~~~~~~~~Air Fluw \D ~~~~~~~~~~Radiator Be Ilmouth Baff le Water In Fig. 45. Schematic of test radiator installation (elevation view).

Fig. 46. Wind tunnel. Fig. 47. Wind tunnel test section showing access panel. 98

Fig. 48. Wind tunnel test section with access panel removed and air bleed shown~ I~~~~I H202~~~~~~~~~~~~~~~~~~~~~~~~~?~t etch ma.ni':io OO ii~'iiiii'!i9 ~~~~~~~~~~~~.~i~~i~:ii_:~i —i:,,.:i-_ii-ii ~icopper~ NH40 H20 a ~'-~i''!Fiii9i et ch ~i~i m agificaio 0X.- i8_-:::_:_ i ji::::hi::: —.:::::-::::::I:-:: _I: -: 99..:.:::

Fig. 50. Microstructure of.0.010 in." copper, NH4OH, H202 etch, magnification 500X. $-;iiii~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Prli ioo!.:!::?~i —-. Fig 51. Microstructure of "O0100 in." oper NH40H) ~ ~ ~ ~ ~ ~ ~ ~ ~: ~ —-::. cope H202 etch., magnification 100X. 100 NA, Aw~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i

Fig. 52. Microstructure of "O.020 in. " copper, NH40H, H202 etch, magnif ication 500X............,_ -Ci~~~~~~~~~~~~~~~~~:i i:i: -E? EyEi-EiE ER.id: —:iEEiE-EjEE _E:i EB-ii;EE i TE- -g EE E S> Eg i V,il i;;;;i; 0;:gg; gg;; t; g ^B t:tg gl;s g; s B;22ago2BaX Sg.a.....a.aa —:: —-_: E E::E. v:s N>+>B.a. > >> a i l S >.B:as.-Vs,-..::.-.... 70: 4 i 0:::::i:;;:Jl;;:: t~l t:::: aB::asiaa gR i-:: 8a-:-:::::._-:::::_::- i _ i - gsBigigig88 aaB88aa-~sssg. St~ ~ ~ ~ ~ ~~-:::: —:-';': f:c j:::::: iii:-:-:, 0 t < w - ^ h 2 2~~~~~~~~~~~~~~~~~~~~~ii:- Bi -::i::-i;i~':-i. —.. -::i::-Z ii::,,:B-i i —~iii:-::iii:::i- iiiiii: —-iiiii,iii-i- i-'!^s-y -—:9..,,::;,:i::_:: Q: -:-iis'i i~~~~~~~~~~~_iZig."i-:::: —: —:i-ii 53. Krus fa igue:-::i tsingm chns i~i —-i~i-:-:E:iiii-:-::-10 1a:

13, 000 \ooo o 12,500 oO O\O Oi 12,000 o 0 00o,0 - o. 0 N. \'- 11,500 1 10 100 1,000 10,000 Cycles (thousand) Fig. 54. Fatigue data for electroformed copper of 0.010 in. nominal thickness.

12, 500 12,000 0 %O000 \0 0 _11,500 0~~~~. HI S ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 0 10,5000 1 ~~~~10 100 1,000 10,00 Cycles (thousand) Fig. 35. Fatigue data for electroforrnec copper of 0.020-in. nominal thickness.

Fig. 56. Crack initiated from the electrolyte side of the specimen in a fatigue test. Unetched. Magnification 500X. Fig. ii. Crack initiated from the electrolyte side of the pecimen Fig. 57. Crack initiated from the electrolyte side of the specimen in a fatigue test. Etched with NH40H, H202. Magnification 500X. 104

Fig. 58. Structure of specimen showed a stronger fatigue resistance than normal. Note the blister on the electrolyte side and the inclusion below it. Magnification 10OX. 105

UNIVERSITY OF MICHIGAN III3 9015 02827 4671ii 3II 9015 02827 4671