L UN 1 I J ”WW I E WNWINWWNHNIWMI)I 889 will Ho’s m mllllllllllll mm in I Ill ._ 31293005712181 'LIBRARY Michigan Stat. L University: This is to certify that the thesis entitled A COMPARISON OF PEAK ACCELERATION IN CUSHIONED DROPS: ACCELEROMETERS VS. HIGH-SPEED VIDEO METHOD presented by Jeffery S. Waldeck has been accepted towards fulfillment of the requirements for M. 5. degree in _P_as:ka£in2 aj r professor Gary urgess Date_ApLil_ZB.._J.9_8.9__ 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution )V1£3I_J RETURNING MATERIALS: \ Place in book drop to LIBRARIES remove this checkout from Jun-(silln. your record. FINES will be charged if book is returned after the date stamped below. A COMPARISON OF PEAK ACCELERATION IN CUSHIONED DROPS: ACCELEROMETERS VS. HIGH-SPEED VIDEO METHOD BY Jeffery S. Waldeck A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Packaging 1989 5914QJ7 ABSTRACT A COMPARISON OF PEAK ACCELERATION IN CUSHIONED DROPS: ACCELEROHETERS VS. HIGH-SPEED VIDEO METHOD BY Jeffery S. Waldeck A quick summary is presented of some conventional methods for measuring the deceleration of a cushioned product in an impact and their limitations. A new method involving a high-speed video camera is developed. The procedure involves filmimg the compression-expansion process of the cushion during impact and afterwards measuring displacement values over time from the video screen. By fitting a high order polynomial function to this data and then obtaining its second derivitive, the peak deceleration produced during the impact is obtained. The peak deceleration value obtained with an accelerometer is compared to that obtained using the high-speed video camera. The peak compression values of the cushions during the impact are also measured. Copyright by JEFFERY SCOTT WALDECK 1989 DEDICATION I would like to dedicate this thesis to my mother, Carol Hart Melvin. in recognition of the constant support she gave me and the sacrifices she made throughout my education. ACKNOWLEDGEMENTS I would like to thank Dr. Gary Burgess for acting as my advisor on this study. The constant assistance and direction he provided was greatly appreciated, especially at those times when all did not proceed as planned. Above all I cherish his friendship and humor. without which I would have been hard pressed to finish this study. I am also greatful to the other members of my committee for their assistance. To Dr. Diana Twede for her moral support and constant encouragement. To Dr. George Hase for his time and effort. And to Dr. Paul S. Singh for suggesting this topic of study. I would also like to thank Susan and Gene HcDonald for generously allowing me the use of their personal computer in typing this thesis. Their assistance and support was invaluable to me throughout the editing process. TABLE OF CONTENTS List of Tables .......... ...... . ................... iv List Of Figures OOOOOOOOOOOOOOOOOOOOOOOO0000000000. v List of Symbols ............... ............... ..... vii Chapter 1 -' INTRODUCTION 0 o o e o o e o e o o o e o o o o e o o e o o o e e 1 Chapter 2 - A NEW METHOD . ................... . ..... 13 Chapter 3 - DATA AND RESULTS .....o................ 36 Chapter 4 - LIMITATIONS AND SUGGESTIONS FOR FUTURE RESEARCH .OOOOOCOOOOIIOOOOOOOOOI 48 Chapter 5 - CONCLUSION OF RESULTS 0.000000000000000 52 Appendix A - EXPERIMENTAL DATA ........... ........ . 53 Appendix B - MEASUREMENT ERRORS ........ .......... . 57 Appendix C - FINITE DIFFERENCE ANALYSIS ........... 61 Appendix D - EQUIPMENT SPECIFICATIONS . ....... ..... 66 References .......... ........ ...................... 69 Table Table Table Table Table Table Table Table LIST OF TABLES PAGE Values of inches/pixel calculated in StUdy OOOOOOOOOOOO 00000 00000000.. 0000000 29 Results of both methods using 2" thick cushions ............................. ..... 37 Results of both methods using 3" thick cushions ........... . ............ .... ...... 38 Percent total compression at peak cushion deformation for 2" and 3" thick cushions .. 46 Position of platen vs. time from high—speed video method for 2" thick cushions ........ 53 Position of platen vs. time from high-speed video method for 3" thick cushions ........ 54 Data from accelerometer method for 2" thick cushions ............. ............... 55 Data from accelerometer method for 3” thick cushions ............................ 56 iv Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 5a 5b 6a 6b 6c 10 11 LIST OF FIGURES PAGE Trade-off between package cost and damage cost . ....... ... .................. 2 Methodology for determining cushioning needed .................................. 2 Typical cushion curve . ..... . ............ 4 Typical shock pulse recorded by an accelerometer ....... .... .......... . ..... 8 View screen of Kodak’s Ectapro 1000 system ...... ..... . ...... . ............... l4 Cushion as it appears before compression .. ......... . . . ............. 14 Cushion as it appears during compression ..... ........... . ............ l4 Illustration of angular offset error .... l9 High-speed video setup ...... ............ 19 Test platen interference ................ l9 Trigonometric error analysis ....... ..... 19 Test setup for the accelerometer method ......... . ........... ... .......... 22 Test setup for the high-speed video method ......... . ........................ 23 Typical displacement plot of cushion compression ............ . ................ 27 Determination of P (inches/pixel) from sample . ............... .. ........ . ....... 28 Weights and static loadings used in tests 000... 00000 O O O 0000000000 O 0000000 33 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 12 13 14 15 16 17 18 19 20 21 22 23 24 Summary of test conditions Results of both methods at a 30” drop height Results of both methods at a 36” drop height Results of both methods at a 42" drop height Dow Chemical Co. data for a 30" drop height 00.0.0.0.000.........OOOOOOOOOOOO Dow Chemical Co. data for a 36" drop height Dow Chemical Co. data for a 42" drop height Percent compression at peak cushion displacement Video error analysis. maximum case .. Video error analysis. minimum case Typical 3-point data spread Typical 5-point data spread Typical 7—point data spread ... 0.0000000000. ......00......0........... vi 40 41 42 43 44 47 59 S9 61 63 65 g000OOO. p.s.i. v ...... b.000... LIST OF SYMBOLS Peak Deceleration (G’s) Pounds per Square Inch static loading of cushion Impact Velocity of test platen Acceleration due to gravity (386.4 in/seczl Equivalent Drop Height The size of 1 display dot on the CRT screen Cushion Displacement at time n Sampling Time Seperation (3 msec in this study Conversion of Inches/Pixel Cushion Thickness Static Loading of cushion vii CHAPTER 1 INTRODUCTION Cushioning is often the only way to protect a product in many situations. If a product is not sturdy enough to withstand the rigors of the distribution environment by itself, some form of protection, hereafter refered to as package cushioning, is generally required. However. since package cushioning can be expensive and depends on product fragility, it is important to use only enough cushioning to adequately protect the product. For a given packaging material and product fragility it can be shown that a trade off generally exists between the cost of excessive packaging and the cost of excessive damage. This is shown in figure 1. Ideally. the amount of protection built into a package should just make up for the difference between the hazards of the environment and the ability of the product to withstand these hazards. This is graphically illustrated in Figure 2. If the product can withstand the rigors of the distribution environment. no cushioning is needed. The product's fragility may be determined using the procedure outlined in ASTM procedure 0-3332 [1]. The hazards of the distribution environment may be determined using measurement techniques on actual test shipments of the product until-II 6-0th x / \ I \\ / \ I cl \ l l / I I A ‘ / CI \ /. o x I U 'l \x / \ .0 z \ 0 i (-0 .l i r T I l I T Amount of Packaging Used Figure l. Trade-off between package costs and damage costs. Cushioning Rigors of Product Needed . Distribution - Ruggedness Environment Figure 2. Methodology for determining cushioning needed [2.3.4]. This allows for the amount of package protection required to then be determined. The function of a cushion is to protect a fragile product by dissipating the free fall energy accumulated in a drop more slowly than would be dissipated in an unprotected drop. This energy is dissipated by a number of mechanisms such as heat transfer from the air in the cushion to the surrounding cell structure IS], damping I6]. and plastic deformation of the cushioning material (7]. The peak deceleration in the impact is commonly used as a measure of cushion effectiveness. A larger peak deceleration indicates that the free fall energy is dissipated too rapidly and therefore. less protection is offered. The standard method of portraying the protection potential of a cushion in an impact is the "cushion curve" as shown in figure 3. A cushion curve relates cushion thickness. load bearing area, and product weight to the peak deceleration (g’s) that can be expected in a free—fall drop from a certain height. The method for generating cushion curves can be found in ASTM procedure D-1596 (8]. As the material properties of a cushion often change somewhat with repeated drops, 30‘ Deep Height. 2-5 PM?! “I 9N ON 71 6t \\ /// \ ill ‘\ 13 ZN 1N Poak acceleratlon (9's) 1 Static Loading (psi) Figure 3. Typical Cushion Curve the cushion curve will depend on the number of drops. The curves most often used in actual package design are the_ ‘2 - 5 drop’ curves which represent the average deceleration experienced for the 2nd.3rd,4th, and 5th impacts under the same conditions. For this reason, the results of this study will be presented in the standard format of cushion curves, with static loading (P.S.I.) on the X—axis, and shock (g's) on the Y-axis. The curves generated from the video results will then be compared to published cushion curves for ‘2 - 5 drops' at the indicated drop height. The results should not be interpreted as new cushion curve data to be used for design purposes. This format is used merely for comparison purposes. The standard method for developing a cushion curve requires some technique for measuring the peak deceleration during the cushioned impact. There are many used to measure this. Mechanical shock measuring devices include spongy balls, cantilevered beams, and in general any spring-mass system whose response to shock can be measured and translated into peak deceleration. The problem with all mechanical devices is that their response is slow and is determined to a great extent by the duration of the shock [9). The response time of a spring-mass type of mechanical accelerometer is governed by its natural frequency. If the response time is comparable to the shock duration. then the mechanical accelerometer will provide useless results. Only mechanical accelerometers with natural frequencies that are high enough to avoid this problem will produce accurate results. The piezo-electric accelerometer is such a device which relies on the deformation-dependent electrical properties of a crystal to measure deceleration. In spite of the widespread use of these types of accelerometers, they are still plagued with problems, some of which are outlined below: 1. CALIBRATION ERRORS: Accelerometers are transducers that produce minute voltages in proportion to the forces they are subjected to. Since force is proportional to acceleration through Newton's Law, the voltage produced becomes a measure of the deceleration experienced. While relatively reliable, they nevertheless require calibration to determine the ratio of voltage output to the deceleration experienced. The accuracy of the results will depend on the technique used and the skill of the operator. WhiCh makes exact calibration difficult. CONNECTION IMPEDANCE: If an accelerometer is not properly connected to the test item. inaccurate measurements will result. The high strains generated when the impact occurs may alter the electrical resistance of the connection. This may alter the output voltage and introduce an error in the measured deceleration. COUPLER ERRORS: A piezo-electric accelerometer requires a capacitance coupler in order for an oscilloscope to measure the output. This coupler has an associated error. Also, as in #2 above. the connections to and from the coupler may induce an error. The power source for the coupler may magnify this error. OSCILLOSCOPE ERRORS: Since the trace width of the accelerometer signal displayed on an oscilloscope is typically on the order of 1 mm at the very least. the peak height of the trace will be in error by this amount. See Figure 4. In addition, The divisions in the photo are each 1 cm x 1 cm. The width of the shock pulse trace is appx. 1 mm (determined visually). ‘ The settings for the oscilloscope in this photograph were 5 ms/division horizontally. 20 mV/division vertically. The drop height was 42". the static loading 1.0 p.s.i.. and the cushion sample tested was 3" thick Ethafoam 220. The sensitivity of the accelerometer used was 2.0 mvlg (see Appendix D). Figure 4. Typical shock pulse recorded by an accelerometer the oscilloscope itself has inherent errors such as drift (the gradual shifting of the ocsilliscope's beam over time), connection errors. and calibration errors. The user may also induce an additional error by failing to differentiate between the base line (start of impact) and the peak (maximum deceleration) on the trace. AMBIENT ERRORS: Electromagnetic interference from overhead lights. power sources. and nearby equipment can induce a current in the cable through induction. the result of which is to alter the signal from the accelerometer. This may be minimized but not completely eliminated by shielding the cables or reducing nearby electrical activity in the environment. An improper ground loop between equipment may also result in an error. Temperature variations may cause errors in virtually all of the equipment involved due to temperature dependent electrical properties of the types of electrical components found in the equipment. TRIBO-ELECTRIC NOISE: Since the signal cable cannot be entirely immobilized throughout the 10 impact. cable flexing will alter the signal. This error can be reduced by using a signal cable with no more slack than necessary for the operation of the test equipment. TRANSVERSE ACCELERATION ERRORS: Acceleration: not confined to the axis of the accelerometer will cause an output error. These can be reduced by maintaining alignment of the accelerometer in the vertical direction during impact and by using well- braced mounting and proper test equipment. OFFSET DUE To DAMAGE: When overloaded with accelerations beyond their measurement capabilities, accelerometers may be permanently affected in the form of an offset from the calibration value. While this error can be prevented through careful handling of accelerometers, it is often impossible to know the complete handling history of an accelerometer beforehand. RINGING OF THE TEST FIXTURE: During impact. the test fixture and associated support equipment will 11 ”ring”, or vibrate at their natural frequencies. This causes transverse vibrations which affect the signal from the accelerometer and lessen the impact by dissipating some of the impact energy which the cushion would otherwise have to absorb. This effect may be reduced by eliminating free play within the equipment as much as possible. If this is not possible, the output of the coupler may be electronically filtered to take out the (undesirable high frequency components in the signal associated with externally induced noise; transverse accelerations, and fixture ringing. The positioning of the accelerometer can also effect the transmitted shock pulse. Taking into account all of the above possible sources of error. the accelerometers used to generate deceleration data are likely to give results which are accurate to :14 z. This figure was calculated as outlined in Appendix B. As mentioned earlier, some of the sources of error may be removed electronically through filtering. The remaining sources of error still exist however. For 12 this reason, a new method for determining the peak deceleration in cushioned drops is desired. The purpose of this study is to compare the peak deceleration obtained in an impact in a cushioned free— fall drop using a piezo-electric type accelerometer to the peak deceleration obtained using the high-speed video method outlined in Chapter 2. Although certain test conditions were used. it is not the intent of this study to generate any conclusions about the cushioning materials based on the results obtained under these conditions. Rather. the intent is to compare the two methods under the same set of arbitrary test conditions. CHAPTER 2 A NEW METHOD This chapter evaluates a new and simpler visual method for determining the deceleration in a cushioned impact. The technique uses a real time plot of product displacement versus time obtained from a high-speed video camera. The camera used in this research was a Kodak Ectapro 1000 which is capable of capturing images seperated by l millisecond intervals (Fig. 5]. When a suitable mathematical function is fitted to this discrete displacement vs. time data, the deceleration function may be obtained as the second derivitive of the displacement function. This method offers several key advantages over the conventional accelerometer method, such as: 1. NO CONNECTION ERRORS: The very nature of the video recording system eliminates connection errors by eliminating the connections themselves. 2. N0 COUPLER ERRORS: There is no coupler. 3. NO TRIBO-ELECTRIC NOISE: Again, there is no physical connection to the test sample. This of 13 i 1 1) 3 s Figure 5a. Cushion as it appears before compression. ... Figure Sb. Cushion as it appears during compression. Figure 5. View screen of Kodak’s Ectapro 1000 system. is course assumes that the cable between the camera and recorder does not move appreciably during the shock. FEWER AMBIENT ERRORS: Ambient conditions can only effect the high-speed video recording system and this possibility has been reduced by the manufacturer by enclosing the electronics of the recording system in a metal enclosure (a Faraday Cage) which essentially eliminates electromagnetic interference. NO TRANSVERSE VIBRATION ERRORS: The high speed video system may be used to analyze motion in one direction only by orienting the camera so that the on-screen grid system coincides with this direction. N0 RINGING ERRORS: Unless the amplitude of vibration of the test fixture is greater than the resolution of the camera (approximately .03") . which is rarely the case, the displacement measured by the camera will be essentially that of the product on the cushion. 16 7. NO ADDED MASS: Even though present day accelerometers are relatively light-weight, certain applications are precluded because of their weight. An example would be measuring the dynamic characteristics of thin films. The high- speed video method is ideal for such applications. In spite of the obvious advantages, there are disadvantages, few of which however affect the accuracy of the technique. These include: 1. HIGH xCOST: The current price of the necessary test equipment is around $80,000, far greater than the cost required for all of the equipment used in the accelerometer method. This limitation can be expected to decrease as high—speed camera technology and use improves. 2. LIMITED USAGE: Because of the strict set-up requirements of the high-speed video system, this technique is limited to laboratory use only except in special cases. 17 FREQUENCY CUT-OFF: Since the video system has a maximum “capture speed , displacement resolution decreases with increasing speed of the event to be captured. This limiting capture speed effectively acts as a low-pass filter which eliminates high- frequency motion superimposed on the dominant compression/expansion motion during the impact. This limitation may actually be regarded as an advantage since it automatically eliminates ringing problems with the test fixture. CAMERA RESOLUTION: The amount of detail, or displacement resolution,‘ of the camera will limit its accuracy. Movements less than the resolution of approximately .03” for the camera used here cannot be accurately displayed as this is the approximate size of a “pixel" on the video screen. This becomes critical around the peak displacement, when the change in displacement is less than the camera resolution. This most likely accounts for the greatest source of error in the method. A possible solution would be to focus in on the lower 1/3 of the cushion in order to record only the critical moments during peak displacement. 18 SCREEN ERRORS: Both image blurring and reading errors also limit the accuracy. Blurring may be minimized by proper lighting and adequate camera speed. In the tests performed here, the speed of the Kodak video system was adequate but less than optimal. This resulted in occasional blurring at critical moments during measurement. It may be assumed that some degree of reading error was also present. ANGULAR OFFSET: Since the camera must capture continuous motion through a range of elevations, it must film at an angle most of the time. The actual displacement will therefore be distorted by the changing camera angle. Angular offset tends to make the perceived compression about 21 greater than the actual compression. This is illustrated in Figure 6. This may be minimized during the critical moment of peak displacement by adjusting the height of the camera lens to that of the bottom of the test cushion. In any event. if necessary this error can be removed by using trigonometric principles. See Figure 6. This error will not be present at all if the dimension of the cushion sample being tested is the same 19 f '1‘ ---- —— ------------ _--r ’ Foam ’ , Sample , ”' ’ d’ Figure 6a. High-speed video set-up l lo— 1 Test -.| -_ ....... --.... :- 1.- ID- Figure 6b. Test platen interference Appx. 48" C D A E l B F Line DA 2 Distance of camera to sample a 48" Line EF a CD - Distance from platen edge to sample a 1" Line DF a Perceived compression Line BC . Actual compression CD DA CDsEF (1")(perceived comp.) -- a -- Error - EB - ---- - EB DF DA (48“) Therefore. Actual Compression a 1.02 x Perceived Comp. Figure 6c. Trigonometric error analysis Figure 6. Illustration of angular offset error 20 from front-to-back as the test platen. 7. GROUND-BORN VIBRATIONS: Vibrations of sufficient magnitude transmitted through the ground to the camera by the force of the impact can cause the camera’s aperature to move. This could result in a displacement error which can be minimized through the use of a proper seismic mass to anchor the test equipment and by isolating the camera from the ground. To utilize the high-speed video method in determining the peak deceleration experienced during a cushioned drop, the following equipment and materials are needed: TEST MATERIALS Although the test samples may be any material of a resilient nature, as described in ASTM 0-1596 [8), the material used in these tests was Ethafoam 220. a product of Dow Chemical Company. Test samples are usually 8” x 8" x Thickness, but for these tests, it was necessary to reduce the surface area to 6“ x 6“ to reach the upper static loadings used. 21 TEST APPARATUS The test apparatus used was a model 23 free-fall drop tester manufactured by Lansmont. A piezoelectric accelerometer mounted to the platen was used in conjunction with a digital storage oscilliscope to record the shock pulse. This set-up is illustrated in Figure 7. A Kodak Ectapro 1000 high-speed video recording system was used to gather the displacement versus time data. The camera was set up so that the entire compression range of the cushion from first contact 'to maximum deformation was within view of the camera lens. This is shown in Figure 8. TEST PROCEDURES The procedure to. record a single drop using the high- -speed video method is as follows: 1. RESET VELOCITY DETECTOR 0N DROP TESTER: This is standard on most drop testers in order to ensure that the proper impact velocity is reached for the desired free fall drop height. 22 l. Lifting mechanism 2. Test weight 3. Test Platen 4. Accelerometer 5. Guide rods 6. Cushion sample 7. Seismic mass 8. Coupler or amplifier 9. Oscilloscope 0 9 '7 Figure 7. Test set-up for the accelerometer method. 23 Guide Rods Light B Test Platen With Weights /L Seismic . Mass Light A Figure 8. Test set-up for the high-speed video method 24 RESET STORAGE OSCILLISCOPE: (optional) The use of a storage oscilliscope is required only if it is desired to compare the result with that obtained from an accelerometer. TURN ON CAMERA LIGHTS: The high-speed camera requires very bright lighting for proper filming. As such lights are very hot, it is desirable to leave them on only when required for filming. ADJUST CAMERA POSITION: In the ”live“ or camera-on mode, the view can be lined up precisely. This should be checked before each drop to avoid possible loss of data due to an accidental bump to the camera.‘ At this time the optimum focus can also be obtained. START HIGH-SPEED CAMERA RECORDING: The high-speed camera uses up video tape very quickly. When running at the speed used in these experiments (23 ft/sec). a single tape can hold approximately 30 seconds of real-time data. For this reason it is desirable to film as little as possible for each drOp. 10. 11. 25 INITIATE DROP TEST: Self explanatory. TURN OFF HIGH-SPEED CAMERA: See #5 above. TURN OFF CAMERA LIGHTS: See s3 above. READ VELOCITY DETECTOR: This is done to verify that the impact velocity was correct for the desired free fall drop height. where; READ OSCILLISCOPE: Optional. see 42 above. TRANSCRIBE CUSHION DISPLACEMENT DATA: The high- speed video system used was equipped with a cursor location readout system which allows for measurements of events to 'within 1 pixel ( .03" for these tests). By selecting a reference point, such as the test platen edge. the displacement of this point can then be measured to within .03” at 1 millisecond intervals. Continue measurements throughout the displacement cycle, noting where the peak displacement occurs. Since the read-out system measures displacement in pixels, these 12. 26 values can be converted to inches only after the entire sequence is viewed. PICKING DATA POINTS: The mathematical analysis used here requires that displacement data be sampled at regular time intervals. There should be the same number of points on either side of the peak displacement, determined upon viewing the recording. In these tests. the sampling time interval used was 3 milliseconds. A typical transcription of the data appears in Figure 9; This method requires a conversion from pixels to inches. To determine this. one must measure the thickness of the sample cushions before testing. After the video record has been made, the reticle location system can be used to determine the locations for the top and bottom of the cushion on screen. The value for inches per pixel can then be determined as shown in Figure 10 and determined as follows: Thickness of cushion (inches) P = --------------------------------- eq.(2) Pixels from btm to top of cushion 27 13. All 110. um. / l 90. .04 "1 1' / / . \\ Y 5.. y 7 Posttlon on Sore-n (pmxols) All” ”-< \ \ as. 3%: 3h 3%: 3% :4: ah 3‘: fine (nilliseeeeds) Figure 9. Typical displacement plot of cushion compression 28 TOP of Cushion Foam Sample Thickness (pixels) _L Bottom of Cushion Figure 10. Determination of P (inches/pixel) from sample Table 1. Values of inches/pixel calculated in study 29 30“ drop height, ' cushion .024 in/pixel 36” drop height, ‘ cushion .024 in/pixel 42” drop height. ' cushion .026 in/pixel 30" drop height, ' cushion .032 in/pixel 36" drop height, cushion .031 in/pixel 42" drop height, ' cushion .030 in/pixel 13. 30 FINITE DIFFERENCE ANALYSIS: The method of splines involves fitting a polynomial function to any number of displacement versus time data points. For greater accuracy, such a spline was fitted to the 7 equally spaced data points obtained in #12 and shown in Figure 9. Such a function represents the displacement of the cushion over time during the impact, and its second derivative is’ the instantaneous acceleration. The details of the derivation are carried out in Appendix C and the result is that the peak deceleration experienced during the impact in terms of the displacement values YI ,Yg ,...etc in pixels taken at evenly spaced intervals t seconds apart is ; tEq.lS] 2(y,)-27(y2)+270(y3)-490(y‘)+270(y5)-27(y3)+2(y7) 180(32)(386.4 in/sec2)(P) See Appendix C for the derivation of this equation. 31 For this study, s was .003 seconds. The values for inches/pixel are shown in table 1. This value is also the effective resolution of the system, as discussed earlier in disadvantage #4. The value for inches/pixel changed from test-to-test due to camera relocation between most tests. Therefore. it was calculated for every drop test recorded. This procedure is illustrated in Figure 10. The procedure for recording a single drop using the accelerometer method is detailed in ASTM 0-1596 [81. Tests were conducted using the above procedures to determine how closely the results of the high-speed video method compared to the results obtained from the accelerometer method. The following test conditions were used. STATIC LOADING : Static loads of 1.0, 1.5, 2.0. and 2.5 P.S.I. were used. This represents the upper range of static loadings that are normally used in cushion design and was singled out for testing not only for this reason but also since accelerometers are likely to have the greatest difficulty in gathering data under these high G conditions. Since the static loadings were 32 chosen to be in this upper range, a smaller than normal cushion sample size was required due to the weight limitations of the test equipment used. The cushion size used was 6”x6” and the weights and static loadings produced are shown in Figure 11. DROP HEIGHTS : Actual drop heights of 30 .36 ,and 42“ were chosen since they are typical of industry cushion curves. The distance the test platen actually falls during the tests must be made somewhat higher than these values since friction between the guide rods of the drop tester and the test platen slow the test platen down as it falls. For this reason, an impact velocity gate was utilized to set the platen drop heights so that the desired free fall drop heights listed above were achieved. The platen drop height is adjusted until the target impact velocity for each of the drop heights is achieved Impact Velocity (Vi) = (J2Gh eq.(l) 33 WEHGHT l CUSHION ./ 36s ---------- = 1.0 9.8.1. 36 sq. in 54s —————————— =3 1.5 9.301.». 36 sq. in 72* ---------- = 2.0 p.s.i. 36 sq. in 90s —————————— = 2.5 9.3.1. 36 sq. in Figure 11. Weights and static loadings used in tests 34 For the desired test drop heights of 30”, 36”, and 42” the required impact velocities are 152 inches per second, 168 inches per second, and 180 inches per second respectively. SAMPLE CHARACTERISTICS : The samples used were 6" x 6" blocks of Ethafoam 220, a low density expanded polyethylene foam weighing 2.2 pounds per cubic foot made by Dow Chemical. Two and three inch thick cushions were chosen as typical thicknesses that would be used in industrial applications. Twice as many samples were constructed as needed for the testing and checked for accurate measurements. Any samples that failed to meet specifications were discarded. From the remaining samples, the actual test samples were chosen at random for each drop. TEST REPETITIONS : In order to maintain a minimum degree of statistical significance, two repetitions of each test condition were performed. While this is perhaps insufficient to gather statistically valid data for plotting cushion curves, it was judged to be adequate for the comparison between the accelerometer and the video methods since both acceleration measurements were taken simultaneously during the same 35 drop. This allowed for a direct comparison of the results as conditions were identical for each method. These test conditions are summarized in Figure 12. Test Drop Heights 3 30" , 36” , 42" ' Test Static Loading: 4 1.0, 1.5, 2.0, 2.5 p.s.i Test Sample Dimensions 2 2" x 6” x 6" 3" x 6" x 6" Test Samples 1 Ethafoam 220 Test Repetitions 2 Total Drop Tests 48 Figure 12. Summary of test conditions CHAPTER 3 DATA 8 RESULTS The data for the displacement versus time observations obtained as described in step 412 of the high-speed video method can be found in Tables 5 and 6. This data can i be used in Equation 15 to calculate the instantaneous acceleration values. The data obtained by using an accelerometer to measure the peak I deceleration can be found in Tables 7 and 8. The results of both methods are compared directly in Tables 2 and 3, and the data is displayed again in the format of cushion curves in Figures 13 through 15. Cushion curves generated from the data provided by the Dow Chemical Corporation are shown in Figures 16 through 18. 36 37 Table 2. Results of both methods using 2“ cushions. Video Accelerometer h / T / S-L Method Method 30 / 2 / 1.0 psi 98 g's 9S g's 89 g's 76 g’s 30 / 2 / 1.5 psi 140 g's 143 g's 139 g’s 140 g's 30 / 2 / 2 0 psi 185 g's 175 g's 185 g's 178 g’s 30 / 2 / 2.5 psi 203 g's 255 g's 203 g’s 260 g’s 6 / 2 / 1.0 psi 129 g’s 105 g’s 146 g’s 113 g's 6 / 2 / 1 5 psi 226 g's 280 g's 201 g’s 27S g’s 36 / 2 / 2 0 psi 236 g's 310 g’s 246 g’s 325 g’s 36 / 2 / 2 5 psi 302 g's 460 g’s 255 g’s 460 g’s 42 / 2 / 1.0 psi 159 g’s 140 g’s 166 g's 158 g's 42 / 2 / 1.5 psi 286 g's 350 g's 226 g’s 345 g's 42 / 2 / 2.0 psi 256 g's 430 g's 235 g's 430 g’s 42 / 2 / 2.5 psi 304 g's 590 g's 284 g’s S70 g’s h 8 equivalent drop height in inches T = cushion thickness in inches S-L = cushion static loading in p.s.i. 38 Table 3. Results of both methods using 3” cushions. Video Accelerometer h / T / S-L Method Method 30 / 3 / 1.0 psi 48 g's 36 g's 46 g's 35 g's 30 / 3 / 1.5 psi 43 g’s 47 g's 16 g's 44 g’s 30 / 3 / 2.0 psr 65 g's 63 g's 52 g's 65 g's 30 / 3 / 2.5 psi 85 g's 103 g's 75 g’s 93 g’s 36 / 3 / 1.0 psi 82 g’s 50 g’s 106 g's 54 g's 36 / 3 / 1.5 psi 146 g's 93 g's 147 g's 90 g's 36 / 3 / 2 0 psi 169 g's 118 g's 133 g's 110 g's 36 / 3 / 2 5 psi 190 g’s 158 g’s 189 g’s 163 g’s 42 / 3 / 1.0 psi 101 g’s 57 g’s 99 g's 63 g's 42 / 3 / 1.5 psi 171 g's 125 g's 183 g’s 135 g’s 42 / 3 / 2 0 psi 179 g's 160 g's 202 g’s 168 g’s 42 / 3 / 2.5 psi 236 g's 245 g's 200 g's 235 g's s equivalent drop height in inches cushion thickness in inches -L = cushion static loading in p.s.i. (fit-53‘ ll 39 36' I!!! Height. 2'5 DFOFS A nigh-3pm m» mm D Accelerometer Method Ii ¥ 3 .— .8 § .1 Punk Goo-loratdon (9's) .— M 1E1 (It 2:- I If! i 2!: 3 Static leading (Fsi) Figure 13. Results of both methods at a 30" drop height 40 36' Deep leisht. 2-5 Drops A list-8F!“ Video hthed 55" [J Meier-enter Method 3'. fig: j Peak Goo-lorntdon (g‘s) ... c 2r 1 if: i if: 2 Static Leading (psi) Figure 14. Results of both methods at a 36” drop height 41 42" Deep Height. 2-5 been 6. A list-Speed Video mm / 55" [J lecelereneter hthed 5'4 450. 2 n w u f fii .N _Ll i u 2 Poak neocloratdon (g's) p I .s l 135 i 235 a sum mum (psi) Figure 15. Results of both methods at a 42“ drop height 42 30' be! Height. 2-5 PM” 6“ 0 Dee Genital Data 550. . 5.. S 55 j 35 Paula hooolonatlon (g's) ... o in— 1 1 if: i 2.5 Static mum (m) Figure 16. Dow Chemical Co. data for a 30" drop height 1.; 43 36' Drop Height. 2-5 MP! 6' 0 Dee musical Data SEA 5“. 450.] N u a a g u 1.; P 8 _.L j W3 Peak nooolorntdon (g's) u 0- -. E i l .5 1 if: i 215 Static Leading (psi) Figure 17. Dow Chemical Co. data for a 36" drop height 44 42" he! Height. 24 he?! 6“ O Dee Chenical Data 550. 5'. 450 .1 i E i l Punk Acceleration (g' s) .‘i l r I if: i :3: Static mum (psi) Figure 18. Dow Chemical Co. data for a 42“ drop height 45 An additional benefit derived from using the high-speed video method is that the maximum cushion strain can be readily determined upon playback of the recording. In comparison, peak compression values are difficult to obtain, if not impossible, using the accelerometer method. The percent compressions during peak cushion deformation are calculated in Table 4 and shown in Figure 19. Note the extreme degree of cushion compression that occurs. Since the sides of the cushions were not observed to bulge appreciably during this compression, it follows that the volume of the cushion decreases dramatically. When this occurs, it is natural to assume that the density, and therefore the shock-pulse frequency experienced during the impact, increases dramatically. 46 Table 4. Percent total compression at peak cushion deformation for 2“ and 3“ thick cushions h=30” h=36” h=42” S-L, sample T=2" T=3" T=2" T=3" ' T22" ta3" 1.0 A 82% 69% 89% 77% 91% 80% 1.0 B 83% 68: 90% 79% 92% 81% 1.5 A 922 792 99% 92% 99% 95% 1.5 B 95% 791 971 922 99% 942 2.0 A 96% 90% 99% 94; 99% 96% 2.0 B 982 90% 99% 93x 99x 95% 2.5 A 99% 95% 992 972 99x 99% 2.5 B 99% 941 99% 97% 992 99% .1“ 47 954 9t 85. 654 Percent Coo-apron- a on “J 5L Figure 19. / /“ g/ A 30' Deep height. 2' Cushion A 30' Drop Height. 3' Cushion U 36" he! leisht. 2' Cushion | 36' «IMF Height. 3' Cushion .s 1 If: i :3: Static Leadin! (F31) Percent compression at peak cushion displacement “am—“...— CHAPTER 4 LIMITATIONS AND SUGGESTIONS FOR FUTURE RESEARCH When comparing the results obtained using the high- speed video method with the results obtained using the accelerometer method, it is clear that they may differ by as much as 100%. Only part of this discrepancy may be attributed to the high-speed video method. The remainder must be associated with the accelerometer method. Some obvious problems encountered during the drop tests using the high-speed video method that were not covered previously are covered below. The contrast between the white cushion material and the moving test platen tended to decrease the most at peak displacement and this may have increased reading errors somewhat. Blurring or wash-out of certain ”key" frames also resulted in greater reading errors. Ambiguous portions were double-checked in an attempt to reduce these errors. Possible improvements include improving the contrast between the test platen and the cushion material, and using a faster camera speed (not currently possible on the camera used) with a greater resolution. 48 49 The contrast could also be improved by using the proper combination of surface characteristics. Both black-and- white paper, and black paper and aluminum foil were. tried as platen coverings. The best results were obtained with the black-and-white paper. It is ‘possible that other surface coverings may work better. This would have helped to positively locate the pixel corresponding to maximum cushion compression. A faster camera speed would result in a smoother image on the video screen. Often the platen edge would "jump" 5 pixels or more in 1 "frame . While this occured to a lesser extent at the critical moment of peak cushion .deformation, it must still be assumed that reading errors occurred whenever the platen edge was blurred over a span of many pixels. A faster camera speed would also allow for greater resolution through ”zooming-in” at the critical moment of peak cushion displacement. This assumes some prior knowledge of the approximate value of this peak compression however. Since a greater number of pixels would cover the smaller critical area, finer measurements could be made. This would in turn require a greater camera speed in order to maintain a reasonably smooth record. 50 The number of replications was limited by the availability of film for the Kodak video system. A test project using a greater number of replications may give better results but would require more film. While the above arguments point to the high-speed video method as being responsible for the difference between the two methods, it is entirely likely that the accelerometer method is equally responsible for the discrepency. The accelerometer method has an accuracy of £141 (See Appendix B) which accounts for at least part of the difference. The data shows a tendency towards greater disparity at higher static loadings with greater drop heights. See Figures 13 through 15. These impacts produce very high cushion strains (from 90% - 99%). In these instances, the density of the cushion increases dramatically (from 10 to 100 times in a very short time) which in turn raises the cushion stiffness and leads to high g values. The accelerometer must therefore be capable of responding to large changes in acceleration over small time periods and may not be able to do so depending on its natural frequency. The accelerometer may also be producing exaggerated data as 51 the result of shock amplification [101. This may also help explain the variation in g’s measured between the accelerometer results and the cushion curves for Ethafoam 220 provided by Dow Chemical. It does not however completely explain the variation between the high-speed video method and the other results.Further testing needs in order to more fully understand the errors inherent in each method. CHAPTER 5 CONCLUSION OF RESULTS The high-speed video method is a useful alternative for for determining peak acceleration during cushioned impacts. While there are problems associated with this method, few affect the accuracy of the results. Furthermore these problems, outlined in Chapter 2. may be expected to decline as the technology of high-speed video becomes more prevalent. The high-speed video method also is capable of several functions not currently possible using the accelerometer method. Since a time record is made of the entire compression- expansion cycle, peak displacement is very easy to obtain. This study shows that cushions deform much more upon impact than previously thought. This is shown in Figure 19. In addition, since there is no contact with the test sample, it is possible to record the dynamics of cushioned drops on extremely lightweight samples. One possible example would be to measure the dynamic characteristics of thin films without having to deal with the weight of an accelerometer. Any analysis situation where the weight of the accelerometer is a concern is a possible candidate for the high-speed video method. 52 APPENDIX A: Experimental Data all LOO-3:3" measurements in pixels. 53 Time for 2” Table 5. Position of Platen vs. ICshnI h T / S-L I TopI -9 -6 ................ +-_--+---_--_----- 30" 2"/ 1 0 psiI 93 I 66 46 I 89 I 66 45 I l 30” 2”/ 1 5 psiI 92 I 68 47 I 92 I 62 40 I I 30“ 2"/ 2.0 psiI 92 I 69 47 I 92 I 69 47 I I 30" 2 / 2.5 psiI 91 I 68 45 I 91 I 66 43 ................ +—-_-+--——---——--- 36" 2”/ 1 0 psiI 94 I 71 46 I 90 I 74 50 I I 36” 2 / 1.5 psiI 89 I 76 51 I 89 I 71 45 I I 36“ 2 / 2.0 psiI 93 I 74 49 I 91 I 75 49 I I 36“ 2 I 2.5 psiI 90 I 76 53 I 91 I 74 49 ———————————————— +—-—-+--—_---——--- 42” 2”/ 1.0 psiI 86 I 70 47 I 82 I 69 44 I I 42” 2“/ 1 5 psiI 81 I 73 47 I 79 I 69 42 I I 42” 2 / 2.0 psiI 82 I 75 48 I 81 I 70 44 I I 42”/ 2”/ 2.5 psiI107 I102 76 I109 I 98 73 ................ +----+-——------_-_ equivalent drop height in inches cushion thickness in inches cushion static loading in p.s.i. cushions +1 +6 +9 29 41 56 26 38 52 21 34 50 21 36 52 18 33 49 18 32 48 19 34 50 20 36 52 26 42 60 24 40 57 23 41 60 26 43 62 21 39 S8 22 40 58 22 38 56 24 41 60 20 35 53 22 39 S7 15 33 49 17 35 54 20 38 57 23 41 61 45 64 81 47 64 81 54 Table 6. Position of platen vs. time for 3” cushions ICshnI h / T / S-L I TopI -9 -6 -3 42"/ 2"/ 2.0 psiI 112I 63 40 18 42"/ 2“/ 2.5 psiI 114I 71 45 23 I 116I 69 44 22 ................ +_-_-+-----------_ all measurements in pixels h a equivalent drop height in inches T a cushion thickness in inches S-L a cushion static loading in p.s.i. +3 +6 +9 44 50 61 47 53 63 35 42 52 34 42 52 28 37 48 29 39 51 25 35 48 28 39 52 38 48 62 37 48 63 24 37 52 23 35 50 21 34 50 23 37 53 19 33 50 19 34 50 34 45 60 33 46 61 18 33 51 17 32 50 17 33 52 16 32 49 17 34 S3 17 35 54 55 Table 7. Data for accelerometer method, 2“ cushions IOscilloscopeI Pulse IAcceleromtrIAccl. 42"/ 2"/ 2.5 psiI 200 mv/div I 200 mV/div h / T / S-L I Setting I Height ISensitivityIIg's) I I in Div.I I ---------------- +---—--------+--------+-----------+----- 30 / 2 / 1 0 psiI 50 mV/div I 3.8 diVI 2.0 mv/g I 95 g I 20 mv/div I 7.6 diVI 2.0 mv/g I 76 g I I I I 30"/ 2”/ 1.5 psiI 50 mv/div I 5.7 diVI 2.0 mv/g I143 g I 50 mV/div I 5.6 divI 2.0 mVIg I140 q I I I I 30"/ 2“/ 2.0 psiI 50 mvldiv I 7.0 diVI 2.0 mv/g I175 g I 50 mv/div I 7.1 diVI 2.0 mv/g I178 g I I I I 30"/ 2“/ 2.5 psiI 100 mv/div I 5.1 diVI 0 mv/g I255 g I 100 mvldiv I 5.2 diVI 0 mvlg I260 g ---------------- +------------+--------+-----------+----- 36“/ 2"/ 1.0 psiI 50 mv/div I 4.2 diVI 2 0 mv/g I105 g I 50 mv/div I 4.5 diVI 2 0 mV/g I113 g I I I I 36"/ 2”/ 1.5 psiI 100 mv/div I 5.6 diVI 2.0 mvlg I280 g I 100 mv/div I 5.5 divI 2.0 mv/g I275 g I I I I 36"/ 2"/ 2.0 psiI 100 mV/div I 6.2 diVI .0 mv/g I310 g I 100 mV/div I 6.5 diVI .0 mv/g I325 g I I I I 36"/ 2”/ 2.5 psiI 200 mv/div I 4.6 diVI 2.0 mv/g I460 g I 200 mV/div I 4.6 diVI 2.0 mV/g I460 g ---------------- +------------+--------+-----------+----- 42“/ 2“/ 1.0 psiI 50 mv/div I 5.6 diVI 2 0 mv/g I140 g I 50 mv/div I 6.3 diVI 2 0 mv/g I158 g I I I I 42"/ 2”/ 1.5 psiI 200 mvldiv I 3.5 divI 2 0 mv/g I350 g I 100 mv/div I 6.9 diVI 2 0 mV/g I345 g I I I I 42"/ 2”/ 2.0 psiI 200 mv/div I 4.3 diVI 2.0 vag I430 g I 200 mv/div I 4 3 diVI .0 mv/g I430 g I I I ... h equivalent drop height in inches T a cushion thickness in inches S cushion static loading in p.s.i. I I‘." II Table 8. Data for accelerometer method, 3” 56 IOscilliscopeI Setting mV/div mv/div mv/div mv/div mv/div mv/div mv/div mv/div mV/div mV/div mv/div mv/div mv/div mv/div mv/div mVIdiv mv/div mv/div mv/div mv/div mvldiv mvldiv mv/div mv/div 3 equivalent drop height h T = cushion thickness in inches S i cushion Pulse IAcceleromtrIAccl. Height ISensitivityI(g’s) in Div.I I ........ +-_--—---—--+----- 3.6 divl 2.0 mv/g I 36 g 3.5 diVI 2.0 mv/g I 35 g I I 4.7 divl 2.0 mv/g I 47 g 4.4 diVI 2.0 mV/g I 44 g I I 6 3 diVI 2.0 mv/g I 63 g 6 5 diVI 2.0 mv/g I 65 g I I 1 diVI 2.0 mv/g I103 g 7 diVI 2.0 mv/g I 93 g ........ +--..----....--+-_--- 2 0 diVI 2 0 mVIg I 50 g 5 4 diVI 2 0 mv/g I 54 g I I 3.7 diVI 2.0 mv/g I 93 g 3.6 diVI 2.0 mV/g I 90 g I I 4.7 diVI 2.0 mv/g I118 g 4.4 diVI 2.0 mv/g I110 g I I 6.3 divl 2.0 mv/g I158 g 6.5 divI 2.0 mv/g I163 g ........ +------_---_+_---_ 5.7 diVI 2.0 mv/g I 57 g 6.3 diVI 2.0 mv/g I 63 g I I 5.0 divl 2.0 mv/g I125 g 5.4 divl 2.0 mv/g I135 g I I 3.2 diVI 2.0 mV/g I160 g 6.7 divl 2.0 mv/g I168 g I I 4.9 diVI 2.0 mv/g I245 g 4.7 diVI 2.0 mVIg I235 g ———————— +-——--—---—-+-—--— n inches -L a cushion static loading in p.s.i. Appendix B Measurement Error 3 S7 The errors associated with this study may be divided into accelerometer method errors, drop tester errors, and video method errors. Accelerometer method errors: Accelerometer: :21 Coupler: 152 Oscilloscope: 13% Reading error: 14% Sensitivity error: unknown (See Appendix C) Total Error : 114% 58 Drop tester height errors: Drop height error: .121 Total Error : :21 Video method errors: A An obvious technique error resulting in the blurring of some of the recorded video images is known to have occurred. When this occurred, the exact position of the platen edge was not easily determined. The range of uncertainty was at most 1 pixel. To maximize the effect of this uncertainty on the instantaneous acceleration [eq.lSJ, the displacements y],y3,ys, and y? should be increased by 1 pixel, and the displacements y2, y‘ , and ya should be decreased by 1 pixel. This is illustrated in Figure 20. Likewise, to minimize the instantaneous acceleration, the displacements 59 1— Figure 20. Video error analysis, maximum case Figure 21. Video error analysis. minimum.case 6O y1 .y3 .y5 . and y7 should be decreased by 1 pixel and displacements y2 ,y4 , and V6 should be increased by 1 pixel. This is shown in Figure 21. Since the maximum and minimum_ accelerations obtained in this way represent the two extremes (assuming no more than a :11 pixel error) the actual acceleration is expected to be somewhere in between them. The average error for the acceleration associated with this range of values using the test data is 3:252. A more accurate assessment of the position error is .5 pixels. which would reduce the error to :t12.52. Another source of error the changing camera angle. See Figure .6. This error is greatest at peak displacement if the camera is initially focused on the top of the cushion (as in these tests). By focusing on the cushion bottom, this error would have been less. This error was estimated to be :22 for a typical drop. Total Error : i271 Appendix C Finite Difference Analysis 61 The method of splines involves fitting a polynomial function to a collection of equally spaced data points. The degree of the polynomial is determined by the number of data points to be fitted. In general, if there are N points to be fitted, the degree of the polynomial is N-l. The greater the degree of the polynomial, the better it is expected to represent the data. For example. to fit a quadratic spline to the 3 data points represented in Figure 22.; Displacement (pixels) / N -S 0 +5 Time (t) Figure 22. Typical 3-point data spread A quadratic spline can be fitted to these points by forcing the displacement vs. Time equation, Y s a + bIt) + cIt)2 [Eq.41 62 to fit the data, which requires that: When: t = -3, Y = a - b(s) + c(s)2 [Eq.S] When: t = 0 Y = a + b(0) + c(O) = a [Eq.61 When: t = +3 Y = a + bIs) + c(s)2 [Eq.7l The solution of this system of equations for c gives: c = ............. [Eq.BJ The values of a and b are not needed as the acceleration is the second derivitive of Eq.4 with respect to time and this involves only c; Y = a + bIt) + c(t ) [Eq.4] Y’ = 0 + bIl) + 2c(t) Y - 2y + Y Y" = 0 + 0 + 2c = ------------- [Eq.9] 63 The value of y” in Eq.9 is the peak deceleration associated with the 3-point finite difference approximation to the true acceleration. An improved finite difference expression for the acceleration uses a S-point data spread as illustrated in Figure 23. Displacement (pixels) \ I -2s -s 0 +5 +25 Figure 23. Typical 5-point data spread The spline to be fitted to this data is a quadratic polynomial, r . a + b(t) + c(tz) + d(t3) + c(t‘) [Eq.lOJ r" a o + o + 2c + 6dIt) + 12eIt2) [Eq.lll | Time (t) 64 Eq. 11 is the instantaneous acceleration at any given moment in time, It). At t=0 where the deceleration is desired. the result is seen to be dependent again only on c: Y" = 2c [Eq.lZJ The solution to the system of equations obtained by forcing the spline in Eq.10 to fit the displacement y at the five sampling times in Figure 23 gives: c g ............................. [Eq.l3 g = Y" = ----------------------------- [qul4] Fitting a 7-point data spread as in Figure 24 to a 6th order polynomial gives an even better finite difference approximation to the true acceleration, [Eq.ISJ 2(y])-27(y2)+270(y3)-490(y4)+270(y5)-27(y6)+2(y7) Displacement 65 (pixels) / N -35 -28 ‘5 0 0s ‘25 +35 Figure 24. Typical 7-point data spread Time It) APPENDIX D EQUIPMENT SPECIFICATIONS 66 Cushion Test Machine: Lansmont model 23 cushion tester Machine Type Platen Size Platen Weight Ballast Kits Controls Table Lifting/Positioning Maximum Equivalent Free- Fall Drop Height Power Requirements Pneumatic Requirements Free-fall 9" x 9” 12 8 lbs. 1 x 6.4 lbs. 1 x 12.8 lbs 2 x 25.6 lbs. 2 x 32.0 lbs 24 vdc control system (standard) By electric hoist 45 inches 115 vac, 60 hz, 8 amps. Air or nitrogen at 80 -120 peSeie Accelerometer: PCB model 305 A05 piezoelectric Resonant Frequency Reference Voltage Sensitivity 40 khz 2.00 mv/g @100 hz Note: Although this value was used for this study, it must be assumed that an error due to the frequency response characteristics of the accelerometer was present. For best results, the accelerometer used should be calibrated over the entire range of _shock frequencies expected. Range 1000 g's Coupler: Kistler model 5116 Frequency Response Input/Output Coupling Full Scal Impedance Noise e Signal Power Sour C e Oscilloscope: Screen Type Accelerat ion Voltage Writing Speed Vertical Sensitivity Frequency Bandwidth Rise Time Input Imp edance Maximum Allowable Input Voltage 67 Kikusui model 055 6520 5% from .5 hz to 250 khz Input is AC coupled to output buffer amplifier. 20 v p-p 20 ohms .18 mv rms 110 vac, 60 hz, 10 mA Direct viewing, bi-stable storage tube. Appx. 3.15 Kv 25 div/msec 5 mv at SV/div DC - 20 mhz, -3 dB 17 nsec 1 Mega-Ohm 400 v High-Speed Video Camera: Imager Resolution Imager Lens Mount Tripod Mount Recording Technique Recording Medium Tape Handling Recording Rates Recording Time Playback Video Output Power Requirements Kodak Ectapro 1000 192 x 240 pixels C-mount 1/4-20 and 3/8-16 standard ANSI Linear PM 1/2” high-density magnetic tape Cassette (700 ft) 30,60,125.250.500, 1000 frames/sec From 30 seconds to 16 minutes based on speed used Continuous. jog, or single step NTSC and PAL 110 vac, 60 hz, 8 amps LI ST 0? REFERENCES 10. 69 ASTM 0-3332. Mechanical:§hgc3_firagility_o£-£rqduct§l using Shock Machines, 1983. . James w. Goff and Diana Twede, §h§£§_§QQ_§I§§k;- Laborato:2-§dzenturg§-in_§h9g3_and_222amics. (Michigan State University, School of Packaging) 1983. PP. 11-17. James W. Goff. Control 0: Damage in Shipment: The *-- “——----- _—------‘- Dexeloemsnt 9f 5 Telemet;2-§h9§3_8gasurins-§xstgm. Technical Report No. 14. (Michigan State University, School of Packaging). 1968 'Richard K. Brandenburg and Julian Lee, gundgmgntglg 9£_2ackagins_Drnamics. 2nd Ed. Minneapolis. MTS Systems Corporation, 1985. Gary Burgess. "Some Thermodynamic Observations on the Mechanical Properties of Cushions.” Journal of Q:llnlar-£la§tics. 1987. ----------- flgndbggg, 2nd Ed. McGraw-Hill Book Company, 1976. E§§§§Q§_Qfléhi9nigq fléterials» 1983. Harris and Creed. §h925_and_!ibratign-8andbgox. pp. 12.1-12.5 Dynamics. PP. 86-93. ST TE UN MICHIGAN IV. LIBRQRIES IIHIHMIWIH IWNWWINIHIWIHHI 312 12181 n 9J0057