y 
 
 ic /rs<?<? 6Jm Ww\ S , 
 
 PRELIMINARY EVALUATION Of A VACUUM- 
 
 INDUCED CONCENTRATED-LOAD 
 
 SANDWICH TESTER 
 
 June 1932 
 
 
 
 This Report is One cf a Series 
 
 Issued in Cooperation with the 
 
 AIR FORCE-NAVy-CIVIL SUBCOMMITTEE 
 
 on 
 
 AIRCRAFT DESIGN CRITERIA 
 
 Under the Supervision of the 
 
 AIRCRAFT COMMITTEE 
 
 of the 
 
 MUNITIONS BOARD 
 
 No. 1832-A 
 
 UNITED STATES DEPARTMENT OF AGRICULTURE 
 
 FOREST SERVICE 
 
 FOREST PRODUCTS LABORATORY 
 
 Madison 5, Wisconsin 
 
 In Cooperation with the University of Wisconsin 
 
PRELIMINAR Y EVAL UATION OF A VACUUM- IN DUCED 
 CONCENTRATED-LOAD SANDWICH TESTER- 
 
 By 
 
 17. S. ERICNSEN, Mathematician 
 E. 17. KDEHZI, Engineer 
 
 and 
 B. G. EEEBINK, Engineer 
 
 Forest Products Laboratory,- Forest Service 
 U. S. Department of Agriculture 
 
 Summary 
 
 This progress report presents a description of a testing device designed 
 to aid in the inspection of aircraft sandwich constructions. Included is 
 a discussion of the performance of the tester on a limited number of sand- 
 wich constructions. Results of a theoretical analysis for determining 
 deflections and maximum stresses are presented. Suggestions are given for 
 improving the performance of the device. 
 
 Introduction 
 
 Increasing use of sandwich construction, particularly for primary structural 
 elements, has created a greater demand for inspection devices to determine 
 the quality of sandwich constructions. Various methods have been suggested 
 and some successfully demonstrated to locate and explore the extent of 
 unbonded areas. Obviously, a testing device would be most useful if a 
 nondestructive test could be made that would not only locate but would also 
 
 1 
 
 ""This progress report is one of a series prepared and distributed by the 
 
 Forest Products Laboratory under U. S. Navy, Bureau of Aeronautics Order 
 
 No. NAer 01519 and U. S. Air Force No. USAF-l8(S0O)-7O. Results here 
 
 reported are preliminary and may be revised as additional data become 
 
 available . 
 
 2 
 
 ""Maintained at Madison, Wis., in cooperation with the University of 
 
 Wisconsin. 
 
 Report No. 1832-A Agriculture-Madison 
 
evaluate the strength of poorly bonded areas. A vacuum tester was devised 
 and patented by an aircraft company. 1 This vacuum tester was submitted to 
 the Forest Products Laboratory for test at the request of the Air-Force- 
 Navy-Civil Panel 23. 
 
 Description and Operation of the Sandwich Tester 
 
 The tester consists of a dish-shaped casting, approximately 10 inches in 
 diameter, with a rubber gasket or washer around the outside rim. Figure 1 
 shows an external view of the device. The gasket is used to form a 
 pressure seal between the tester and a sandwich panel. An internal view 
 of the tester (fig. 2) shows a central rubber-covered steel foot that is 
 pressed against the panel as the dish-shaped cavity is evacuated. Foot 
 sizes of different diameters from 1 to 2 inches in steps of l/4 inch are 
 supplied so that various ratios of compression and shear can be applied. 
 The foot is fastened by means of a convenient snap-on fitting to a threaded 
 bolt that extends through the casting. A vacuum gage is attached to the 
 casting to measure the applied load. 
 
 In use, the tester is operated by placing it on a sandwich panel, adjusting 
 the position of the foot by turning the threaded bolt until the panel 
 contacts both the foot and the rubber gasket, and drawing a vacuum on the 
 casting until failure occurs or until some desired proof load, determined 
 by the setting on the poppet valve, is reached. 
 
 Exploratory trials of the tester showed that the rubber gasket may be 
 deformed so much, by the deflection of the panel, that the casting makes 
 contact with the sandwich. If this occurs, continued evacuation merely 
 places small additional uniform load on the sandwich with little further 
 deflection. This can be easily demonstrated by applying the tester to a 
 thin sheet of aluminum; the casting makes contact with the sheet at low 
 vacuum and then the sheet is slightly concave around the foot until maximum 
 vacuum is attained. 
 
 In order to indicate rim contact, a buzzer was introduced in a circuit 
 between the casting and sandwich facing. If a nonconductive facing material 
 is on the sandwich being tested, thin strips of metal foil can be placed on 
 the surface and used in the circuit. 
 
 If rim contact occurs at low loads, the vacuum can be partially released 
 and the foot extended toward the sandwich by turning the foot adjusting 
 screw. 
 
 It was conceivable that some sandwich constructions would deform consider- 
 ably under test with no visible or audible indication of failure. 
 
 2Devised by P. M. Matlock and patented by Lockheed Aircraft Corporation. 
 Licenses for manufacture and use granted to Aircraft Die Cutters, 
 Los Angeles, Calif., and Zenith Plastics, Gardena, Calif. 
 
 Report No. 1832-A -2- 
 
Indications of failure might appear on a load-deflection curve. 
 Accordingly, a deflection dial was arranged to read the central deflection 
 of the sandwich with reference to three points located opposite the rim of 
 the tester. The arrangement of this apparatus is shown in figure 3« It 
 was subsequently demonstrated that failure of a balsa core was apparent on 
 load-deflection curves (fig. h), but after failure the load increased until 
 rim contact occurred and eventually increased to maximum vacuum with no 
 audible signs of failure. 
 
 Stresses Induced by the Tester 
 
 Of primary interest are compression and shear stresses induced in the core 
 of the sandwich construction by the concentrated load at the foot of the 
 tester. Failures that may occur in the bond between facings and core can 
 be interpreted in terms of the shearing stress developed in the core. 
 
 The area covered by the casting and gasket of the tester decreases 
 slightly as the vacuum is increased as a result of deformation of the 
 gasket. The area of the tester used at the Forest Products Laboratory was 
 computed to be 90.76 square inches. From this area the compressive and 
 shear stresses in the core were calculated according to the values in 
 table 1. 
 
 The expressions given in table 1 were obtained from approximations of the 
 formulas given in the appendix of this report. For unusual constructions 
 having either thick facings or extremely soft cores the more exact 
 expressions in the appendix should be used to compute the stresses. 
 Expressions for deflections and facing stresses are also included in the 
 appendix. 
 
 Experi mental Work 
 
 In order to determine whether the tester would perform satisfactorily 
 in measuring the quality of sandwich constructions, a few panels were 
 tested to determine either core shear strengths, the location of unbonded 
 areas, or the strength of poorly bonded areas. 
 
 For more accurate load readings the dial vacuum gage was replaced with a 
 mercury column, shown in figure 3« A needle valve was placed in the 
 vacuum line to permit sensitive adjustment of load application. 
 
 A preliminary test on a sandwich having 0.012-inch 2*kST clad aluminum 
 facings on a core of end-grain balsa wood l/k inch thick developed core 
 failures in shear at approximately the shear strength of the balsa as 
 evaluated in shear tests. Load-deflection curves for this construction 
 are shown in figure h. 
 
 Preliminary testing on a sandwich panel known to have unbonded areas 
 demonstrated that the tester was capable of detecting the unbonded areas 
 
 Report No. 1832-A -3- 
 
regardless as to whether the unbonded facing was the one on which the 
 tester was applied or whether it was the opposite one. 
 
 As a further check on the performance of the tester, two flat panels were 
 tested of each of two sandwich constructions . One panel of each construction 
 was well bonded and one panel was poorly bonded on one side. Poor bonding 
 was obtained by using less adhesive than is required for good bonding. The 
 constructions were (l) 0.020-inch 24ST clad aluminum facings bonded with 
 adhesive 35 to a l/2-inch-thick core of aluminum honeycomb of 0.003-inch 
 foil formed to 3/8-inch cell size (core 52), and (2) facings of 8 plies of 
 glass cloth 112-114 impregnated with resin 2, wet-laminated to a l/2-inch- 
 thick core of glass-cloth honeycomb of 112-114 cloth formed to 1/4 -inch 
 cell size (core 36). 
 
 Each panel was large enough to permit four tests with the tester (two from 
 each side) without overlapping the test areas. Load-deflection curves for 
 the aluminum panel are given for each test in figure 5 and for the glass - 
 cloth panel in figure 6. Values of the shear stresses developed by the 
 tester at failure are given in table 2. The average strength values as 
 measured by the tester show that the poorly bonded aluminum panel had 
 approximately 80 percent of the strength of the well-bonded panel and that 
 the poorly bonded glass -cloth panel bad approximately 35 percent of the 
 strength of the well-bonded panel. Figure 7 shows a cross section through 
 an aluminum control panel (the two halves laid face to face) that illustrates 
 a typical failure under the foot of the tester, and figure 8 shows a typical 
 bond failure in a control glass-cloth panel. Poorly bonded panels of both 
 types failed in a similar manner, respectively, except that the aluminum 
 panels failed in the poor bond as well as in shear in the core, and the bond 
 failures in the poorly bonded glass-cloth panels were more extensive. 
 
 After the tests had been made with the tester, the panels were cut into 
 minor coupons to be tested in bending to determine shear strength, and in 
 tension normal to the facings to determine bond strength. The bending speci- 
 mens were 1 inch wide and were tested under loads applied at two-third points 
 on a total span of 6 inches for the aluminum sandwich and 4-1/2 inches for 
 the glass-cloth sandwich. The strong direction of the core was placed 
 parallel to the span length. Tensile specimens were 1 by 1 inch in cross 
 section. The results of these tests are also given in table 2. 
 
 The shear strengths as determined for the bending-test coupons were from 
 30 percent to 70 percent of the shear strengths as determined by the tester. 
 The bending-test coupons were cut from portions of the panel adjacent to the 
 area tested by the tester, and, therefore, this reduction in shear strength 
 may have been due to damage caused by the tester load or may also have been 
 due to stresses caused by the saw when the minor coupons were cut. 
 
 Shear strengths as determined from bending tests showed the poorly bonded 
 aluminum panel to have 88 percent of the strength of the well -bonded panel, 
 and the poorly bonded glass -cloth panel to have 19 percent of the strength 
 of the well-bonded panel. 
 
 Tensile strengths of the poorly bonded aluminum panel were 27 percent of the 
 strengths of the well-bonded panel, and strengths of the poorly bonded glass - 
 cloth panel were 11 percent of the strengths of the well -bonded panel. 
 
 Report No. 1332-A -4- 
 
General Observations 
 
 In the evaluation work on the tester to date, it appears that the device 
 has considerable promise of providing a practical means of proof -loading 
 flat, and possibly curved, sandwich panels to a precalculated stress. 
 The addition of the spring loaded relief valve provides a means for 
 controlling the load to any desired level within the range of 1 to 2k 
 inches of mercury. The accuracy of the load application appears to be 
 about ±l/2 inch of mercury. The trials showed that tests can be made at 
 the rate of 12 to 15 per minute if load-deflection data are not obtained. 
 
 Conclusions 
 
 The tester can be used to determine the location of unbonded or poorly 
 bonded areas in panels of sandwich construction. 
 
 Although the data of this report do not positively show direct correlation, 
 the tester may be used to determine shear strengths of sandwich constructions. 
 
 In the realm of nondestructive testing the tester should find use in careful 
 application of certain proof loads as an aid in inspection of the quality of 
 sandwich panels. 
 
 Eeport No. 1832-A 
 
APPENDIX 
 
 The deflection and the stresses induced in the facings and in the core by 
 the tester have been analyzed by the use of formulas derived in U. S. 
 Forest Products Laboratory Report No. 1828.- In that report the core 
 and facing materials are assumed to be isotropic and, consequently, the 
 present results are limited to this special case. 
 
 In order to carry out the analysis, the distribution of load over the foot 
 of the tester must be specified. An attempt has been made to keep the 
 distribution uniform by covering the foot with a rubber gasket, and it is 
 possible that at small vacuum loads it actually is quite uniform. At 
 large vacuum loads, however, the tester forces the panel to bend, and the 
 load is possibly concentrated more heavily at the rim of the foot than at 
 the center. On the basis of these considerations, the analysis has been 
 carried out for two assumed distributions, namely, (l) a load uniformly 
 distributed over the foot, and (2) a load concentrated at the rim of the 
 foot. For a given vacuum a uniform distribution yields higher predicted 
 shear stress in the core and lower bending stresses in the facings than a 
 load concentrated at the rim of the foot. It is expected that if the true 
 distribution of load over the foot of the tester could be determined, the 
 results would be intermediate between those of the two extreme cases 
 considered. 
 
 The formulas given below are derived on the assumptions that the test panel 
 is of infinite extent and that the facings are of equal thickness. More- 
 over, the formulas are given in forms applicable to a panel with thin 
 facings and with a core that, like end-grain balsa, has a relatively high 
 shear modulus. Specifically, it is assumed that oa and ab, defined below, 
 are so large that the Bessel functions that appear in formulas taken from 
 Report No. l828i can be expressed in the forms (6k) and (65) of that 
 
 -oIq —\) ) 
 report, and that the quantity e ~ J _ can be neglected in comparison 
 
 with unity. It is believed that these conditions will be fulfilled for 
 most constructions that are likely to be of practical interest. 
 
 The deflection and the components of stress in the facings and in the core 
 are given in terms of the following symbols: 
 
 a: radius of the tester gasket 
 
 b: radius of the foot 
 
 £: thickness of the core 
 
 E = - 
 
 - 1 « yd 
 
 h ' 
 
 -W. S. Ericksen. The Bending of a Circular Sandwich Panel under Normal Load. 
 Report No. 1852-A -6- 
 
E_: Young's modulus of the facing material 
 
 — f 
 
 f : thickness of the facings 
 
 G: shear modulus of the core material 
 
 I = x m + X f 
 
 I = f ( c + f )' 
 
 — m " jT 
 
 I - — 
 
 £: applied vacuum (p.s.i.) 
 
 a 2 _ 2GI 
 
 Ecflf 
 
 v: Poissons ratio of facing material 
 
 For a load distributed uniformly over the foot of the tester, the deflec- 
 tion of the sandwich, wjj, at the center of the tester relative to the 
 outer rim of the tester~~is given by the formula 
 
 W U ■ w bU + w sU 
 
 (1) 
 
 where, 
 
 w bU - ^ ! 5 (a 2 - b 2 ) - to 2 log £ 
 
 b 
 
 w 
 
 sU " 
 
 qa 
 
 ! log a - 
 L b 
 
 2EI. r a2 -2v2 
 
 (2) 
 (3) 
 
 For a load concentrated at the rim of the foot, the central deflection is 
 given by 
 
 w c " w bc + v sc 
 
 where, 
 
 w b 
 
 qa 
 
 w 
 
 6^EI 
 qa 
 
 5a 2 - 8b 2 (1 + log £) 
 
 b 
 
 sc 2EL 
 
 -k ! log a - i (1 + i-) 
 
 s2 b 2 ob 
 
 L 
 
 (5) 
 (6) 
 
 The shear stress in the core evaluated at the rim of the foot is given by 
 the formula 
 
 Eeport Wo. 1852-A 
 
 -7- 
 
%-r ~ 
 
 a 2 -b 2 
 
 U " 2(c + f) 
 
 ab 2 
 
 (7) 
 
 for a load distributed uniformly over the foot, and by the formula 
 
 r = a 
 
 c 
 
 r»o 
 
 2(c + f) ; 2b 
 
 (8) 
 
 for a load concentrated at the rim of the foot. 
 
 The stress at the outer surface of the facing upon which the tester is 
 placed is evaluated at the rim of the foot by the formula 
 
 U,c 
 
 q(c + 2f) 
 321 
 
 * I 6(c + f) (3c + 2f) 
 l W,c + f ( c + 2 f) 
 
 T.X 
 
 sU, c 
 
 (9) 
 
 where the subscripts U and o again designate quantities associated with a 
 uniform foot load and a load concentrated at the rim of the foot, 
 respectively. 
 
 For a load distributed uniformly over the foot, 
 
 r^u = k(l + y) a 2 log g - (3 +7 ) (a 2 - b 2 ) 
 
 m, 
 
 sU 
 
 m (1 + y) (a 2 - b 2 ) 
 
 a 
 
 2 2 
 
 Ix b 
 
 2 2 
 2a b 
 
 1 - 
 
 (i -r) 
 
 ab 
 
 L... 
 
 (10) 
 
 (11) 
 
 and for a load concentrated at the rim of the foot, 
 
 m bc = Ml + 7) a 2 log | - 2(1 + 7 ) a 2 + (3 + 7 ) b' 
 . (1 -y) J l_ + 7_ 
 
 - 1 * " «T-« " O 
 
 m sc ■ J£b 
 
 ab 
 
 (12) 
 
 (13) 
 
 At the present time a reliable means of estimating the transverse pressure 
 on the core is not available. However, if the load can be taken to be 
 uniformly distributed over the foot, it is estimated that the pressure 
 exerted by the loaded facing upon the core is about one-half the pressure 
 on the foot. 
 
 Application of formulas 
 
 Figure h shows the load-central deflection curve for a test panel to which 
 the preceding formulas are applicable. The results given below were 
 
 Report No. 1832 -A 
 
obtained from these formulas for a foot diameter of 1-3/4 inches. In the 
 computations, the shear modulus of the core material, which was not 
 measured, was taken as 20,000 pounds per square inch. The following is a 
 complete list of the dimensions and elastic properties used. 
 
 a = 5.38 inches 
 
 b = 0.875 inch 
 
 c = 0.25 inch 
 
 E.„ = 10? pounds per square inch 
 
 f = 0.012 inch 
 
 k 
 
 G = 2 x 10 pounds per square inch 
 
 y = 0.3 
 
 These yield 
 
 7 
 E = 1.1 x 10 pounds per square inch 
 
 I = 4.12 x 10 -lf inches 5 
 I f = 2.30 x 10"? inches 5 
 
 I = 4.12 x 10 "^ inches 5 
 a = 41.6 
 From (2) and (3) it is found that the deflections are given by 
 
 w bU = 0.0135 q 
 
 and (14) 
 
 W sU = 0.0049 q 
 
 Therefore, if the load is uniformly distributed over the foot, the central 
 deflection obtained from (l) is 
 
 v„ = 0.0181, 9(pis>1>) = 0.0090 q ( . n< EG) (15) 
 
 Similarly, from (5) and (6) 
 
 w bc = 0.0127 q 
 
 and (16) 
 
 W sc = 0.0034 q 
 
 Report No. 1832-A -9- 
 
so that for a load concentrated at the rim of the foot, {k) yields 
 
 » c - 0.0161 i (p>s>i>) = 0.0079 q (in> m) in) 
 
 Formulas (15) and (17) yield, respectively, 
 
 ^(in. T IG) = HI W[J and ^(in. HG) " 12 ? w c ( l8 ) 
 
 The slopes of the lines represented by these equations would change 
 slightly if the shear modulus of the core was changed and if the radius 
 of the tester, which presumably decreases with increasing load, was 
 varied. They are, however, of the same order of magnitude as the slopes 
 of the linear portions of the two curves in figure h for the foot 
 diameter of 1-5 /h inches, which are 125 and 152, It is of interest to 
 observe that the deflections due to shear represented by w s y and w sc 
 
 are about one-third of the deflections due to bending represented by 
 w^u and w^ c . These relatively large shear deformations for an aluminum- 
 balsa panel are attributed to the small dimensions of the area of the panel 
 over which deflection takes place. 
 
 The shear stress in the core obtained from (7) and (8) are, respectively, 
 
 T U = 60 V.i.) =29q (in. IIG) (19) 
 
 and 
 
 ip.s.o.,; (m. IIG) 
 
 The stress predicted on the basis of a uniform load on the foot is thus 
 about twice as great as that obtained by taking the load to be concentrated 
 at the rim of the foot. 
 
 For the present core and facing thicknesses, formula (9) for determining 
 facing stresses takes the form 
 
 °U,c = - 20 ' 8 %,c " 20 > 500 * m sU,c (21) 
 From formulas (10) to (13) 
 
 20.8 m^uq = 3760 q (22) 
 
 20,300 mgyq = 267 q (23) 
 
 20.8 rn^q = 4170 q (2'l) 
 
 20,300 m sc q = 3950 q (25) 
 
 Report No. 1832-A -10- 
 
Therefore, for the load distributed uniformly over the foot, the stress 
 at the surface of the facing at the rim of the foot is 
 
 °U - "»* 1(p.s.i.) ■ - 1930 »(la. B» (26) 
 
 and for the load concentrated at the rim of tlie foot the corresponding 
 stress is 
 
 °b ■ - 8120 1 " -3990 3 (ln> HG) (27) 
 
 The minus sign in these equations indicates a compressive stress. 
 Formula (27) indicates that if the load on the foot was concentrated at 
 the rim, the proportional-limit stress of 27,000 pounds per square inch 
 would be reached at 6,6 inches of mercury. 
 
 Report No. 1C32-A -11- 
 
Table 1. --Stresses induced by the tester 
 
 Diameter 
 of foot 
 
 : Core stresses 
 
 . n p 
 ; Compressions — 
 
 : cr 
 
 : Shear- 
 
 : t 
 
 In. 
 1.00 
 
 1.25 : 
 
 1.50 
 
 1.75 
 
 2.00 
 
 : P.s.i. 
 57. 2q 
 
 56. 5q 
 
 25. 2<i 
 
 iQ.hq 
 
 15. 9q 
 
 ; P.s.i. 
 
 : 57.2 -JL- 
 h + c 
 
 h + c 
 
 ! 57.8 -4- 
 h + c 
 
 : 52.2 , <1 
 
 h + c 
 
 2 ?- 8 E I o 
 
 "£ = applied vacuum (p.s.i.), h = total sandwich 
 thickness (in.), c = core thickness (in.) 
 
 -See Appendix, page 12. 
 
 Report No. 1332-A 
 
Table 2. — Strengths of panels as determined by the tester 
 
 
 Shear i 
 
 strengths 
 
 
 Tensile strength 
 
 Hell-bonded 
 panel 
 
 Poorly bonded 
 : panel 
 
 : Uell- 
 , bonded 
 : panel 
 
 Poorly 
 ■ bonded 
 , panel 
 
 Tester 
 
 : Bending 
 : test 
 
 . Tester 
 
 Bending 
 test 
 
 (1) 
 
 • (2) 
 
 (3) 
 
 {h) 
 
 <5) 
 
 (6) 
 
 P.s.i. 
 
 P.s.i. : 
 
 P.s.i.: 
 
 P.s.i. 
 
 P.s.i. : 
 
 P.s.i. 
 
 Sandwich Construction: Facings — 0.020-inch 24ST Clad 
 Alu m i num, Core — 1/2 -inch- thick Aluminum Honeycomb 
 " of 0.003-inch Foil formed to 3/8-inch Cell Sizel 
 
 191 
 2^2 
 211 
 204 
 
 Av. 212 
 
 78 
 
 95 
 179 
 196 
 
 86 
 175 
 
 135 
 
 2l63 
 
 2170 
 
 163 
 
 170 
 
 76 
 
 156 
 
 56 
 163 
 
 166 
 
 119 
 
 370 
 330 
 350 
 320 
 310 
 230 
 230 
 250 
 270 
 280 
 280 
 270 
 300 
 270 
 250 
 340 
 34o 
 300 
 310 
 310 
 
 296 
 
 110 
 
 110 
 
 100 
 
 70 
 
 70 
 
 60 
 
 70 
 
 60 
 
 50 
 
 60 
 
 50 
 
 80 
 
 110 
 
 100 
 
 100 
 
 80 
 
 (Sheet 1 of 2) 
 
 Report No. 1832-A 
 
Table 2. -- Strengths of panels as determined by the tester (Cont,) 
 
 
 Shear i 
 
 strengths 
 
 
 : Tensile 
 
 strength 
 
 Well-bonded 
 panel 
 
 : Poorly bonded 
 
 : panel 
 
 : Uell- 
 : bonded 
 : panel 
 
 Poorly 
 bonded 
 panel 
 
 Tester 
 
 Bending 
 test 
 
 Tester 
 
 . Bending 
 test 
 
 (1) 
 
 : (2) 
 
 (3) 
 
 : W 
 
 • (5) 
 
 (6) 
 
 P.s.i. ! 
 
 P.s.i. : 
 
 P.s.i, . 
 
 P.s.i. : 
 
 P.s.i. : 
 
 P.s.i. 
 
 Sandwich Construction: Facings -- 8 plies of Glass 
 Cloth 112, Core -- 1/2-inch-thick Glass -cloth Honey- 
 comb of 112 Cloth Formed to 1/4-inch Cell Size3 
 
 451 
 544 
 521 
 515 
 
 Av. 506 
 
 180 
 252 
 34 
 
 55^ 
 
 226 
 282 
 
 272 
 
 2.146 
 2215 
 
 151 
 
 I85 
 
 17^ 
 
 38 
 66 
 50 
 52 
 
 52 
 
 310 
 270 
 250 
 240 
 420 
 220 
 340 
 290 
 500 
 230 
 290 
 240 
 34o 
 
 280 
 290 
 300 
 280 
 320 
 290 
 300 
 
 290 
 
 
 40 
 
 60 
 
 20 
 30 
 
 20 
 50 
 50 
 50 
 
 32 
 
 Tested with a 2-inch-diameter foot. 
 
 2 
 
 -Tester on poor -bond side of sandwich. 
 
 •7. 
 
 -Tested with a l-l/2-inch-diameter foot. 
 
 (Sheet 2 of 2) 
 
 Report No. 1832-A 
 
Figure 1. --External view of the sandwich tester, showing 
 rubber gasket, attached vacuum line, and vacuum gage. 
 The center bolt extends to the interior foot. An auxiliary 
 poppet valve has been installed (immediately below the gage on 
 the photograph) for control of vacuum during use as a tester. 
 
 z M 88(Al F 
 
Figure 2. --Internal view of the tester, showing the central foot 
 and alternate foot sizes available for use. The cantilever spring 
 actuating the poppet valve for control of vacuum is also shown. 
 
 Z M 880k0 F 
 
Figure 3. --The tester in use during evaluation tests, showing the addition of 
 a deflection-measuring device on the opposite facing. 
 
 Z M 88038 F 
 
0.02 
 
 0.04 0.06 0.08 
 
 FOOT DEFLECT/ON (INCH) 
 
 0.12 
 
 7. M 90070 F 
 
 Figure k.- -Load-deflection curves produced by applying tester 
 to a sandwich having 0.012-inch 2^ST clad aluminum facings on 
 a l/4-inch-thick core of end-grain balsa. 
 
16 
 
 14 
 
 12 
 
 Cfc; 
 
 8 
 
 
 Q 6 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ____ — -o 
 
 
 
 
 // r ' 
 
 v/ — ^ > ""~ ' 
 
 a 
 
 
 
 
 ' / 
 
 
 ____ — — - 
 
 "~ 
 
 
 
 y// / 
 
 ' _ 
 
 — 
 
 
 
 
 
 fr 
 
 
 
 
 
 
 1 //// 
 
 fjjhf 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 W 
 
 
 
 
 
 
 
 i 
 
 
 
 
 
 
 1/ ' 
 
 
 
 LEGEND: 
 D TEST 1 
 
 
 
 ill 1 
 
 
 
 o TEST 2 
 
 r • T C C* T ~2 
 
 
 
 
 
 
 v / tS 1 o 
 
 
 
 h i 
 
 
 
 A TEST 4 
 
 
 
 
 
 CONTROL PANEL 
 
 Jfl i 
 
 
 
 POORLY BONDED PANEL 
 
 & i 
 
 
 
 
 
 
 
 ¥ 
 
 
 
 
 
 
 
 i 
 
 
 
 
 
 
 
 J/ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 o 
 
 0.01 
 
 0.02 0.03 0.04 
 
 FOOT DEFLECTION (INCH) 
 
 0.05 
 
 0.06 
 
 Z M 90071 F 
 
 Figure 5» — Load-deflection curves produced by applying tester 
 with a 2-inch-diameter foot to a sandwich having 0.020-inch 
 24ST clad aluminum facings on a l/2-inch-thick core of 0.003- 
 inch aluminum foil formed to 3/8-inch hexagonal cells. 
 
24 
 
 20 
 
 o 
 
 16 
 
 12 
 
 8 
 
 
 
 
 
 . ... i 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 RIM CONTACT^ 
 
 
 
 
 c 
 
 > 
 
 
 
 yy^ 7 ^ 
 
 
 LEGEND: 
 
 
 /5^__^ 
 
 7 : 
 
 P D TEST 1 
 O TEST 2 
 
 
 
 
 V ' 
 
 TEST 3 
 
 
 
 
 A ; 
 
 TEST 4 
 
 
 
 
 CONTROL PANEL 
 
 
 
 
 POORLY BONDED 
 
 
 
 
 PANEL 
 
 i i 
 
 o 
 
 0.02 
 
 0.04 0.06 0.08 
 
 FOOT DEFLECTION (INCH) 
 
 O.IO 
 
 0.12 
 
 Z M 9007? F 
 
 Figure 6. --Load-deflection curves produced by applying tester with 
 a l-l/2-inch-diameter foot to a sandwich having facings of eight 
 pli^s of glass cloth 112-114 on a l/2-inch-thick core of glass 
 cloth 112-114 formed to a honeycomb of l/4-inch cell size. 
 
& 2Li 
 
 
 CO 
 
 H ID 
 
 +2 n 
 
 O -H 
 O eS 
 O tH 
 
 IS 
 O ID 
 
 £? ■** 
 
 1>> CO 
 
 (D 
 
 § ° 
 
 3 O 
 
 CI -H 
 
 3 O 
 H -P 
 
 a) ca 
 
 ■H 
 
 d -d 
 
 3 
 ■p 
 
 O 
 
 -P 
 
 ft CD 
 O 
 
 | 
 
 3 
 
 § 
 
 ■H 
 -P 
 
 o 
 o 
 
 m 
 
 ca O 
 
 CO 3 
 
 O -P 
 ft 
 
 U <D 
 
 H 
 CD CD 
 
 CD • 
 
 h in 
 
 CD H 
 
 p H 
 5 
 
 •H O 
 
 fa 
 
 cy 
 
 CO 
 
 s 
 
A 
 
 #. 
 
 r* 
 
 n 
 
 ^ 
 
 ■^ #• 
 
 ^4 
 
 
 Figure 8. — Typical failure in a control glass-cloth panel. Failure is confined to 
 the bond between the facing and the core on the side opposite the tester. 
 
 z M 88863 F 
 
UNIVERSITY OF FLORIDA 
 
 3 1262 08928 5810