' ”Mi ; WWW,» I 553' 3‘ W” M w l'“" Jhfi: «Mg ha: . ‘3'" '~'~~>3 a .3, h I .'I I 3 ” a ' flu 1' M?” It“). 5.. hy' l 3‘ 39 ”film | Mill 3 1293 01093 1834 KEY unmfwgan S Univenlty l I I ”ll llllllllllllllflll .. This is to eertify that the thesis entitled AEVEwmeNT ,wb Fame/name TEST ”F A THZEE POINT Hn'cH bYNAmomereg F91 USE IN “WAGE QJEQGY QGSWCH presented by PASCAL. GITARI [(A UMBUTHO has been accepted towards fulfillment of the requirements for Adegree in fig“; ENG/VG . fi/wé/W 44%; Major professor Date //—/fg‘5/ 0-7639 MSU is an Afiirmative Action/Equal Opportunity Institution lV1SSI_J BEIURNING_MATERIALS: Place in book drop to LIBRARIES remove this checkout from .u—cu—L your record. FINES will be charged if book is returned after the date stamped below. ~— __ «-.—< {DM- DEVELOPMENT AND PERFORMANCE TEST OF A THREE POINT HITCH DYNAMOMETER FOR USE IN TILLAGE ENERGX RESEARCH BY Pascal Gitari Kaumbutho A THESIS Submitted to Michigan State University in partial fulfillment of the requirements ' tor the degree of MASTER OF SCIENCE Department of Agricultural Engineering 1985 ABSHMST DEVELOPMENT AND PERFORMANCE TEST OF A THREE POINT HITCH DYNAMOMETER FOR USE IN TILLAGE ENERGY RESEARCH BY Pascal Gitari Kaumbutho A. three point hitch dynamometer was developed and tested for performance capabilities and limitations. The design adopted is that of Chung et a1 (ASAE paper No. 83-1065). The dynamometer, composed of strain gaged sensing pins supported on a quick attaching coupler was adopted because of its relatively simple and inexpensive characteristics. It measures vertical and horizontal components of tractor tillage draft forces. A micro-computer based data acquisition system was utilized in calibration, verification and field-testing of the dynamometer. Tests made were aimed at gaining knowledge of the dynamometer's capabilities, error sources, accuracy and dependability as a draft measuring device. Approved:/fl%5:iié7€:;4<ézé£ézét;;£&r~\¢2 Major Professor Department Chairman ACKNOWLEDGMENTS I thank God for my good health and capabilities in accomplishing this work. My gratitude is due to the following: ** Dr. Robert H. Wilkinson, my major professor for his professional guidance and never-ending support provided throughout the duration of this study. ** Members of my guidance committee: Dr. Ajit K. Srivastava, Dr. Thomas H. Burkhardt and Dr. Gary L. Cloud for their time, understanding, constructive counsel and generosity in letting me share in use of space, equipment and materials. ** U.S.D.A. staff, especially Richard Wolthius, fellow graduate students like Steve Richey, David Jackson, Solomon Tembo, Pedro Herrera and Sergio Perez who were always willing to lend a helping hand. ** Ford Tractor Company for their generous donation of the Ford 7700 tractor for our use. ** Dr. Earl A. Erickson and Dallas A. Hyde for their continued support during field tillage tests. ** The Kenya government and the University of Nairobi for financial support and letting me have this tremendous opportunity to learn. ** My mother and father, relatives, brothers and sisters for their unfailing love; the many teachers, older or more experienced people, great friends like Mwongera and Guegbeh, who (all) have, in their own special ways, helped make me who I am. 11 LIST OF LIST OF CHAPTER 1 . 2. 3. TABLE OF CONTENTS TABLES FIGURES INTRODUCTION ooooooooooo00.00.000.000...00000000000000... PROBLEM DEFINITION AND OBJECTIVES :.................. REVIEW OF RELATED LITERATURE ......................... 3.1 Three Point Hitch Dynamometers 3.1.fiioneer Work . 3.1.2 Later Dynamometer Designs 3.2 Linkage Force Analysis 3.2.1 Modern ApproacHIto Linkage Force AnalYSi-s OO...OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO. 3.3 Data Acquisition and Analysis Systems MATERIALS AND METHODS.................................. 4.1 Three Point Hitch Dynamometer 4TI.1 Dynamometer Frame 4.1.2 Strain Beam Support 4.1.3 The Force Transducer 4.2 The Data Acquisition System 4.2.1 The Signal Conditioners 4.2.2 The Data Acquisition Control Unit 4.2.3 The Central Command Unit 4.2.4 The Power Supply 4.3 Theoret_i__cal Considerations 4.3.1 Resolution and—Accuracy of Measurement.. 4.3.2 Force Component Measured and Circuit Analysis 4.4 System Configuration 4.5 The Computer Program and Data Acquisition Brocedure OOOOOOOOOOOOOOOO0.0.000...OOOOOOOOOOOOOO00...... 4.6 System Calibration 4.7 System Verification 4.8 Field-Test§ 0.0.0.0...0......OOOOOOOOOOOOOOOOOOOOO. 4.8.1 Test Procedure RESULTS AND DISCUSSION ................................ 5.1 Calibration Results 5.2 Verification Results . 5.3 Field Test Results 5.3.1 NFise-Error Sources Results............... 5.3.2 Hardware Performance SUMMARY AND OBSERVATIONS 22............................ 6.1 Summggy 6.2 Observations 6.3 mure Work. BIBLIOGRAPHY ' APPENDIX iii 41 41 42 42 44 50 50 52 53 54 55 58 60 65 66 7O 75 78 81 85 85 108 115 137 138 141 141 142 144 LIST OF TABLES 1 Calibration Curve Data ..........................93 2 Averaged Verification Data 106 3a Field Test Data Summary (Mould Board Plow).......109 3b Field Test Data Summary (Cultivator) ............110 4 Permanent Strain Test Results ...................139 iv LIST OF FIGURES 1 Arrangement of Strain Gages on the cross-shaft of the plow (Lal, 1959) IOOOOIOOOOOOOOOOOOOOOOOOOOOOOOOOOO00.00.00.000011 2 A draft transducer of the three-point hitch dynamometer (Jensen,1954) 00....0......O...OOOOOOOOOOOOOOOOOOOOOOOOOOOOIS 3 Three-point hitch dynamometer, (a) three measuring elements, (b) measuring element, MA' (thiel, 1958)....................15 4 Dynamometer beams and strain gage arrangement (Sholtz, 1964)17 5 Three point hitch dynamometer a) the subframe assembly b) the force transducers (Scholtz, 1966) ...................19 6 Three-point hitch dynamometer using octagonal ring traDSducerS (DeVine’ 1973)....OOOOIO0.0000000IOOOO0.0.0.0.0021 7a)Three point hitch dynamometer right side schematic b)Design load conditions on transducer subassembly (JOhnson and VoorheeS’ 1979)....0000000.0.00000000000000000024 8 Three point hitch Dynamometer (Barker et. al.,1981).........26 9 Forces between tractor and mounted implement (Lal, 1959)....28 10 Transformation of forces and moments between the global XYZ and local xyz coordinate system ........................28 (Upadhyaya et. al., 1983) 11 Schematic of the lower link of the three point linkage system (Upadhyaya et. al., 1983) ...................30 12 Dynamometer assembly drawing ...............................43 13 Sensing-pin support beam with transducer hook-up detail ....45 14 Sensing pins with strain gage arrangement and force component direCtions Shown O...OOOOOOOOOOOOOOOOOOO0.0.0.0.0048 15 Schematic of microprocessor based data acquisition system for tractor tillage forces ..........................51 16 Strain gage Circuitry (Wheatstone Bridge) ..................64 17 Multichannel sampling flow-chart diagram ........... ....... 67 18 One channel sampling flow-chart diagram .......... 68 19 Calibration loading set-up ....................... 72 20 Verification dynamometer loading hook-up ......... 76a LIST OF FIGURES (cont'd) 21 Vector diagram for verification force ............ 74 22 Recording-penetrometer probe ..................... 77 23 Calibration curve of acceptable linearity ..... ...86 24 Calibration curve of acceptable linearity (Chan2).87 25 Calibration curve of acceptable linearity (Chan4).88 26a Unacceptable straight line plot of Channel 3 calibration data. First assembly loading ........ 89 26b Logarithmic plot of calibration data for Channel 3. Data from first loading following assembly....90 27 Acceptable repeatability calibration plot for Channel 3. Obtained after a series of load/ unload sequences . .............................. ..91 28 A comparison plot of outputs of orthogonally placed channels as one is loaded for verification ... ..... 96 29 A plot showing load/unload hysteresis as encountered in unacceptAble amounts.. ........... . ............ .97 30 A plot showing load/unload hysteresis as encountered in acceptable amounts ...................... .....98 313 Typical verification output at medium size loads.99 b Typical statistical analysis of output from verification at medium size loads ........ . ...... 100 32a Typical verification output at large size loads .101 32b Typical statistical output of large size load verification output ............................ 102 33a A comparison plot of measured and sampled draft force using actual verification data ........... 103 33b A curve fitting comparison plot of measured and sampled draft force using verification data ..... 105 LIST OF FIGURES (cont'd) 33c Verification draft force measurement error distribution graph ........ .................... 107 34a(i Typical output for mouldboard plow tillage....112 (ii Statistical analysis results for moldboard plow tillage forces ................. ...... 113 34b Typical output for a cultivator tillage run ....114 35 Penetrometer probe output chart showing probe area factors and typical output curves .......... 116 36a Comparison of horizontal with vertical force for a moldboard plow .... .......................... 118 36b Comparison of horizontal and vertical force for a cultivator ................ . ..... . .......... 119 37a(i Variation of horizontal force with moisture content (moldboard plow) ..................... 121 37a(ii Variation of vertical force with moisture content when working with a moldboard plow ..122 37b(i Variation of horizontal force with bulk density (moldboard plow) .................... 124 37b(ii Variation of vertical force with bulk density (moldboard plow) ........................... 125 37c(i Variation of horizontal force with penetration resistance (moldboard plow) ................. 126 37c(ii Variation of vertical force with penetration resistance (moldboard plow) ................ 127 37d(i Variation of horizontal force with tillage depth (moldboard plow) ...................... 129 37d(ii Variation of vertical force with tillage depth (moldboard plow) ...................... 130 vii 38a 38b 38c 39 40 LIST OF FIGURES (cont'd) Typical tillage force output showing evidence of hardware failure ................ ........... 131 Statistical analysis results for output with evidence of hardware failure ..... . ..... . ....... 132 Typical output showing evidence of short-circuiting problem in wiring of Chan. 1 ... ........ ........133 Noise-error test without strain gage excitation.l35 Noise-error test with strain gage excitation....136 viii 1. INTRODUCTION In 1976, agriculture used approximately 12% of total energy consumed in the United States, if the total chain (including transportation and storage of manufactured food products) was considered. Of this 12%, only 3% was actually used on the farm (Joint Task Force, 1976). Although 3% is a small amount, a shortage of fuel would force undesirable decisions redirecting the use of energy supplies. Researchers must therefore be prepared to help the farmer make the best use of available energy supplies. Based on these considerations many a researcher have sought to know more about the forces involved in tillage operations, between the tractor and implement. Drawbar pull (tractive effort less rolling resistance) of a tractor consumes a signifiCant amount of total energy input into the tractor implement system. Many forms of strain gaged dynamometers have been developed in recent years for measuring drawbarpull and other hitching forces. They have varied in design detail, range and types of measurement, manufacturing cost and versatility. Not many three point hitch dynamometers are available commercially. The British Hovercraft Corporation (Bast Cowes Isle of Wight, England) has one commercially available in a price range between $24000 1 and $35000, without instrumentation. The development of mounted and semi-mounted implements has considerably diminished the usefulness of the common drawbar dynamometer, both for measurement of tractive ,effort and implement draft (Kepner et al.; 1978). The performance of modern tractors is often dependent ,upon their use with correctly coupled mounted or semi-mounted implements and the draft resistance of a modern mounted implement is likely to be significantly less than that of a comparable trailed version. This situation has led to the demand for an instrument which will measure drawbar pull of a tractor connected to its implement by a three point linkage. Initial attempts to measure forces between tractor and mounted implement were made by measuring forces in the links themselves (Rogers et al.; 1952). The analysis of results was a formidable problem in three-dimensional vector addition. Today, with computer capabilities, it has become possible to measure these forces using completely separate or detachable components with fast, yet relatively accurate means. The micro-computer has made the difference. Until about 10 years ago most tractor testing was usually confined to the laboratory. Field test procedures were often complicated, cumbersome and instrumentation systems had limited capabilities. A tractor is a field machine and should be tested in the field if actual operating data is to be obtained. Micro-computer benefits which include availability, portability,' reliability, flexibilty and ease of operation have made it possible to collect and process much larger quantities of data. A three point hitch dynamometer based on the design by Chung et a1. 1983 (ASAE Paper No. 83-1035) was developed, calibrated, verified and field-tested. Reported here are details of the various steps and considerations involved in the development of hardware, software and methodology used in making it possible to obtain draft force data. The data obtained in the whole dynamometer development process and field tests are analyzed to check the capabilities and dependability of the system as a future tillage energy research tool. 2. PROBLEM DEFINITION AND OBJECTIVES Tillage energy requirements by agricultural tractors depends on tillage depth, implement type speed and to a major extent on the soil physical properties and prevailing environmental conditions. Tillage energy data required for various tillage implements and systems on different types of soils is available in very generalized forms. Soil physical conditions may vary widely over a small area. Therefore there is need for a means by which "on the spot” or localized tillage forces can be measured. An alternative is to make a similitude-analysis-based study of the influence of various soil parameters involved in the generation of forces by the tillage operation. This kind of study would then standardize the tillage operation, compensating for variation in the prevailing soil conditions. Regardless of the approach taken, there is obvious need to develop instrumentation to measure the various parameters involved in the energy input of the tractor, implement and soil interaction. While some of this instrumentation is available off the shelf, the uniqueness of the specific systems usually requires development of the specific parts to meet the intended use and research detail. Measurement of tillage forces encountered in work with the wide range of implements that exist called for the manufacture locally of a three A point hitch dynamometer. Prototypes of dynamometers that exist have been made for specific research needs and these not available in the market place at reasonable prices. With limited resources available, the development of the dynamometer described in this study was undertaken with the objective of developing a system which would accurately measure components of tillage forces and also also have limitations and dependability that were known. Many methods of draft force measurement have been used over the years. The particular methods selected in various research situations have been chosen according to performance required and expected results. The use of strain gages in draft measurement has been popular for many years. They have given good performance in requirements for fast response, accuracy, precision and sensitivity. These are important factors especially in tillage force measurements where changes are rapid. Strain gage performance has improved in recent years due to the specialized electronic circuitry and readout equipment that has been developed. Strain gage type load cells are now comparable if not superior to other kinds of precision force measuring equipment. The major requirement for good performance however, is proper choice and treatment of materials along with correct techniques for attachment of high-stability strain (gages. The recognition of the need for traction and energy data is stimulating this development of more efficient, convenient and less expensive tractor instrumentation systems. 3. REVIEW OF RELATED LITERATURE Since 1937 there has been an ever-increasing number of implements mounted integrally with the farm tractor. Mounted implements have many advantages such as maneuverability in the field and ease of handling and transportation. However, in recent years the size of implements and tractors has increased and this has stimulated the need for better information about forces in the tractor links and the effect of these forces on the implement and tractor. Because draft is the major component of forces between tractor and implement many efforts have been made to develop draft measuring devices. In this endeavor, various projects have been undertaken to study the magnitude, distribution and effects of draft force on the energy requirements for the tillage operation. With this knowledge about the tractor-implement interaction, the energy use on the farm can be optimized. Considerable research effort has been directed to this end and more recent means of doing this have necessitated simultaneous development of data acquisition systems. 3.1 Three Point Hitch Dynamometer Due to single pin hitching of “pull type" equipment it is considerably easier to measure the draft of this type of implement than it is on the more complex mounted equipment. Draft of trailed implements may be determined by a spring type dynamometer. Hoag and Yoerger (1974) developed a ring-type force transducer capable of accurately measuring draft. R. D. Singh et a1. (1981) also fabricated a similar transducer. While these components gave accurate draft measurements with adequate sensitivity for a wide range of magnitudes of forces, they were limited to a single point hitch for trailed implements. 3.1.1 Pioneer Work on Measuring Three Point Hitch Forces. Roger and Johnston (1953) were among the first people to develop a hydraulic dynamometer for measuring the draft of mounted. implements. They attached a hydraulic cylinder directly to each link of the three point hitch. Their technique required photographing the Bourdon pressure gages, lined up on a board and attached to the cylinders by flexible hydraulic hoses. All the gages were mounted on the gage-board in a cluster form so they could be simultaneously photographed by a 16mm movie camera. To test the accuracy of the system, a strain gage was installed in a chain connecting the back of the plow to an anchor post. Comparison of readings to the corrected strain gage readings showed the error in the instrument was 4%, attributed to friction in the cylinders. I Data collected was enough to show magnitudes, rates and amounts of fluctuation of forces with time. From these data, weight transfer to the rear wheels of the tractor was calculated. Roger and Johnston (1953) concluded that resistance strain gages had a high potential for use in force evaluation and with them, more accurate readings could be expected. They computed their drawbar pull as an algebraic sum of linkage forces since top and bottom links were nearly parallel to the ground surface. The normal variation in the link forces was about i 25%, an indication of how much the soil resistance to a plow can change in an average field. 3.1.2 Later Dynamometer Designs In the early days, the three point hitch dynamometers used springs or hydraulic cylinders of some form. Volkov and Klochev (1958) developed a three point hitch measurement system which consisted of dynamometer elements built into each link, and an integrating device. The movement of the. front half-bar of the dynamometer link was transmitted by an unequal-arm lever mounted on the body of the link on ball-bearings. The unequal-arm lever magnified the movement and forces were measured by spiral springs and a multicontact gage. One 10 main part of the integrating unit was a collector consisting of 15 contacts which passed impulses to the recorder in accordance with the deformations of measuring springs. All of the systems designed with mechanical components, i.e. springs, bearings and contact points were subject to problems of inaccuracy, non-durability and were complex in connstruction and operation. Hydraulic dynamometers were an improvement over those of spring type. The device built by Skalweit (1958) used a hydraulic cylinder which was located ahead of a large rectangular frame suspended beneath the tractor by four vertical links. This frame was connected to the rear vertical rectangular frame on which three links were attached. The' implement forces were transmitted through these frames to the hydraulic cylinder. Lal (1959) reported that to provide a complete definition of the force between a tractor and a mounted implement, it would in general be necessary to measure the axial forces in the three links, the two lift rods and the angular position of one link. A graphical analysis of the resulting information would probably be a more formidable task than to obtain the information by instrumentation in the first place. Lal came up with an interesting design. He replaced the normal cross-shaft of the plow by a straight one so that forces acting at the lower link hitch points caused only bending and 11 eliminated torsion (see Figure 1). The section of the shaft at equal distances from the center of the ball joints at either end was reduced. The bending produced at these sections in the horizontal and vertical planes was measured separately by using. strain gages. To measure the force on the top link, a top link dynamometer was built,which was supported on a frame so that it would measure only the horizontal force.The major shortcoming of this device was the difficulty of changing implements. Jensen (1954) demonstrated the superiority of experimental stress analysis as a means of obtaining exact quantitative information about distribution of loads in structures and behavior of materials under loads. He applied this technique to develop various dynamometers for tractor field tests such as a torque meter, a drawbar dynamometer and three point hitch dynamometer. ..see* Mum an cmcfldl or onu- uuu on coon-nan. “I! IP90“! an em no sun "Wm‘m'aam manna mutt-cw * Other part of this Figure is on next page..... 12 3‘ L E.» ~:.- . ‘Syf :3 ' an» ammonium) MOON-M0 FIGURE 1. Arrangement of strain gages on the cross-shaft of the plow(Lal. 1959) ' Since Jensen (1954) demonstrated specific application of the strain gage to tractor dynamometers, many trials have been made to develop more accurate and convenient three point hitch dynamometers. One of the important considerations in developing dynamometers has been the location of strain gages or the force transducer. Either one of the following four units seems to have been chosen for attachment of strain gages or transducers, a) hitch links , b) some part of the implement, c) tractor body or d) hitch frame, located between the hitch links and an implement. . Neuholt (1959) measured draft and vertical forces of the three-point links by attaching a set of strain gages connected as a wheatstone bridge on each link and a strain-gage-equipped proving ring on the upper link. The resultant pull of the implement was obtained by determining the angle of each link with respect to the direction of travel. Orlowski and Wolf (1963) used almost the same device as Neuholt but put one more set of strain gages on one of the lower links and tried to measure side draft force. Luth et a1. (1978) mounted transducers directly 13 on the lower links to obtain reaction forces in all three cartesian coordinate directions. The geometry of the three-point hitch was not altered by the transducer system. The third place available for attaching the strain gages or transducers is the tractor body. Jensen (1954) used three identical beams,. one at each hitch point on the tractor body (see Figure 2). The beam was supported at the point indicated by a subframe to which the tractor links were attached. A set of four strain (one wheatstone bridge) gages was placed on each beam. Thiel (1958) measured the forces by simply mounting two strain gages on each of the lower and upper link pins (see Figure 3). The system was sensitive only to the horizontal components of the forces on the links. Reece (1961) also mounted strain gaged cantilever pins on the_ tractor and supported the ball joints at the inner end of the three links. The ball centers were 76.2mm out from the tractor sides, as compared with the 70mm of the manufacturer's original mountings. The sensitivity of Reece's device can be calculated from the expression: 19 = VSe/(R+Rg) where I9 = ammeter reading V s strain gage exitation voltage S a gage factor 14 e = tensile or compressive strain R strain gage resistance Rg = ammeter resistance The strain in the surface of the cantilever was not accurately predicted by the theory of simple bending because the cantilever was very short relative to its depth. The actual strains were found to be only 83% of those calculated from the simple bending expression: e = 3PL / 3Ec2t load to be measured where P L effective length of cantilever E = modulus of elasticity 2c = depth of cantilever at center of gages t = width of cantilever at center of gages The Stokes-Wilson equation for short beams was used and results with closer agreement were obtained: e = 3P(L/2-c/r)/(2Ec2t)+(P/2Ect) Scholtz (1964) further developed Reece's method. He examined various factors affecting the link pin transducers. these factors included, friction in the ball joints, imperfect position of positioning the 15 \‘ F 0 SH :5 is 4:, \' O «avenue: e! 1.1 out seer-ea ~.\ ...._.. . ) weeoaruoa "an ..." / an IDOLIIINT M’Y‘GIIIO _ tom: not eaarr react ' L 00-00.le 0' 0.”? ll OII‘GNOI 0' VIAVCL FIGURE 2. A draft transducer of the three-point hitch dynamometer (Jensen. 1954) FIGURE 3. Three-point hitch dynamometer, (a) three measuring elements, (b) measuring element. M . Dimensions are in millimeters (Thiel, 1958) l6 gages, faulty cantilever position and hydraulic system characteristics. He estimated the magnitude of errors they' would cause in dynamometer measurements. Results from this study were later utilized in design of link pin transducers. To reduce hysteresis effects, the nominal length of the lower beams was fixed at 165mm (see Figure 4). This gave results which, compared in a field experiment with those of a hydraulic drawbar dynamometer, the difference was insignificant at the 5% level. As the above results show, the link pin size limits its use as a transducer. Making it longer, requires modification of tractor, otherwise errors due to link hitch position changes become inevitable. To eliminate the requirement for modification of either tractor or implement and to make the transducer easily interchangeable among tractors and implements, a sub-frame or a hitch frame three point hitch dynamometer came into being. Scholtz (1966) designed and built a hitch-frame type three point hitch dynamometer. It consisted of force measuring transducers firmly wedged and bolted inside vertical rectangular hollow sections (see Figure 5). The transducer on the top link was arranged to measure draft force only. While the dynamometer was capable of measuring vertical and horizontal components of draft force with good fulfillment of requirements, it was also bulky (118 Kg), and imlements were mounted 17 r—- FIGURE 4. Dynamometer beams and strain gage arrangement (Scholtz 1964) 18 229mm further to the rear of the tractor. The dynamometer 'introduced significant extra resilience in the hitch. It is this Scholtz's (1966) original design that has been re-designed by the National Institute of Agricultural Engineering Silsoe, Bedfordshire, England. The British Hovercraft Corporation has adopted the re-designed product for manufacture. under licence for sale. Carter (1981) developed a dynamometer shaped similar to that of Scholtz (1966). He attached three firectangular strain beams to each link hitch-point of the hitch frame. Strain gages were applied to the front and back of three aluminium strain beams and were connected so that the vertical forces cancelled and horizontal forces on each beam were summed algebraically. Devine (1963) developed a little different type of hitch frame dynamometer. Basing the structural design on load estimates, the frames were constructed of hot rolled rectangular tubing (76x51x5mm thick). The dynamometer consisted of two frames which were connected together by two octagonal ring transducers (see Figure 6). As Hoag and Yoerger (1975), and Godwin (1975) explained, this kind of extended ring transducer fitted with strain gages can measure strain proportional to two orthogonal forces and a torque about the center of the ring independently. Vertical and horizotal adjustment of the dynamometer was supplied so the hitch frame would fit most standard three point hitch equipment. The 19 (a) L .1'. 3 3.1M 8453 It! “lets: Satin A-A I'm-r A m! I 0)) FIGURE 5. Three point hitch dynamometer a) the subframe assembly b) the force transducers (Scholtz, 1966). 20 dynamometer was also designed to be removed easily from the tractor and -implement to protect the strain gaged transducers when measurements were not being recorded. The calibration results showed that the percentage error for draft, verticall and moment measurements were i 1.29, i 2.10, and q: 1.60% respectively. ,The dynamometer has a load limitation of about 13,000N and it could only be used with small size tractors. Johnson and Voorhees (1979) designed one of the most versatile three—point hitch dynamometers ever made. It was capable of measuring draft, vertical force and torque (in a vertical longitudinal plane). simultaneously and independently, with a dual-loading capability (low and high range). It also had a fast hitching capability. It was designed for use on both category II and III tractor hitches and also performed satisfactorily with two point semi-mounted implements. Maximum draft capacities were 66700N in the dynamometer's high range (Category III) and 36000N in low range (Category II). The dynamometer consisted of three subassemblies (see Figure 7a). A transducer subassemly made of an aluminium tube was located between the tractor and implement subassemblies (see Figure 7b). The tractor subassembly was attached at points b and c and the implement subassembly at points a and d. Thus a load applied to the transducer caused both bending and torsion. The complete dynamometer measured 1220mm wide and 790mm high. It extended the implement mounting point A: TRANSDUCERS FIGURE 53 Three-point hitch dynamometer using octagonal ring transducers (Devine, I973) 22 about 310mm to the rear of the tractor hitch points. Smith and Barker (1982) developed a thin three-dimensional measuring hitch frame dynamometer (see Figure 8). It was constructed in the shape of a triangle using a 152mm steel channel. The frame was composed of two triangular halves, symmetric except for tractor, implement and load cell connectors. The dynamometer used six BLH-UZMl load cells with capacities of 22kN. Three of them, placed at the corners and perpendicular to the dynamometer surface measured draft, two load cells located on the upper triangular arms provided vertical and side force information and prevented rotation of the halves. One load cell located in the lower triangular arms measured only the side force. The dynamometer shifted the implement 190mm to the rear, compared to 127mm for a category 111 type quick hitch. It was found that considerable time and effort were required, when the dynamometer was mounted, to adjust the linkage and clearances within the dynamometer so that all forces were well transferred to the load cells. The dynamometer halves had to be separated to adjust the linkage, a time consuming and tedious process. The above review clearly indicates that there are many strong points in favor of the design and use of the hitch frame dynamometer. Problems, however, have arisen that need to be considered in this type of design. These include : a) Construction of the hitch frame. Weight of 23 frame, structural strength and location of transducers, type of strain beam etc. are important considerations. b) Extension of the hitch points of the implement further away from the tractor, a cause of extra resilience (Sholtz, 1966). c) Mounting of the dynamometer (Scholtz, 1966, Devine, 1973, Smith and Barker, 1982). d) Use with the PTO shaft (Johnson and Voorhees, 1979, Smith and Barker, 1982). e) Versatility and interchangeability between tractors (Johnson and Voorhees, 1979). 3.2 Linkage Force Analysis Lal (1959) showed the desirability to simplify the problem of measuring total force between tractor and implement. This force had to be analyzed to see if it would be possible to ommit information concerning some of its components. Figure 9 shows Lal's (1959) concept of forces acting between a tractor and a mounted implement. The general force P has components in all major planes of the tractor and associated with it is a couple C perpendicular to the direction of the force. The force can be resolved into three components L, V and S. Assuming the tractor is proceeding along a horizontal 24 .o-o“ Hd I High range . 272 mm 1 Low range I 505 nun SCSI-T6 Aluminum m s 154mm Tubet IOl.6 mm 0.0. 76.2 mm 1.0. b FIGURE 7 a. Three point Hitch Dynamometer-right side schematic b. Design load conditions on transducer subassembly T=torque, V=Vertical force. H=horizontal force (Johnson and Voorhees, 1979) 25 plane surface, these three force components can be refered to as horizontal, vertical and side forces. The component L, which is parallel to the direction of travel is also refered to as the draft of the implement. Being the most important component of the total force, draft is the part which determines the traction required from the tractor and directly consumes drawbar horsepower.It is the component devoted to useful work by the implement. The vertical component V has the effect of adding load to the tractor rear wheels and removing load from the front wheels. It therefore has a profound effect on tractive ability of the tractor as well as its stability and steerability. No less important is this component's effect on the implement's ability to penetrate and maintain depth as well as on the draft on the implement because of friction forces associated with it. The side force S exists only in the case of implements which are not symmetrically disposed about the tractor centerline. In such cases it is difficult to maintain directional stability of either or both tractor and implement. The draft of the implement is increased because of the friction produced. This component may be considered of far less importance than horizontal and vertical components because with many implements it is non-existent. Lal (1959) found it could be disregarded in the case of multi-bottom plows since even with two 26 .. a m cm can mmeu uaop seem . 52 _ a 0.3 ~13; 2.2a 3...: m :8: access: vautwucm .tmumeoEe: . +wa. llr is. NH. 1 A 1.: 83 t...- xlfiahT/ C . i a . . 1 NR . SQ “.— lrtfl . h a , \\ A llllllll l lllll lllll llllllllllll ..v .4 a. as WN‘N 3.: I11 I» N.“ >>m=> a (m m 27 furrow plows it was quite small. The couple associated with the total force lies approximately in the vertical plane containing the rear axle of the tractor. Its effect therefore is to transfer the load. from one drive wheel to the other and it could therefore have an appreciable effect on tractive ability. However Lal (1959) reports that this couple is either small or non-existent. The force in the vertical plane perpendicular to the driving axle exerted by a mounted implement on a tractor is equal and opposite to the force exerted on the implement by the reactions at the lower link ends and at the top link end. Reactions at the lower link ends will have both horizontal and vertical components while the top link will have tensile or compressive force depending upon. conditions. Lal (1959) went ahead to practically support his analysis. He used a two bottom plow with a straight cross shaft (see sect.3.l.2), instead of the normal offset one. Lal established an overall accuracy of i 5%. He attributed the greater part of the error to friction in the hitch joints. Use of self-alligning bearings instead of ball joints at lower links seemed to reduce the hysteresis effects. 3.2.1 Modern Approach to Link Force Analysis Compared to Lal (1959), the more recent work of 28 FIGURE 9. Forces between tractor and mounted implement. (Lal 1959) LINK \ Pomr ‘1 \. 9y F / M Y. Mx 4’ 4' Y I:x X.X' FIGURE 10. Transformation of forces and moments between the label XYZ and local xyz coordinate system. IUpadhyaya et al.. 1983) 29 Collins et a1. (1981) and Upadhyaya et a1. (1983), provide an improved way of analyzing forces in the three point hitch linkage system. Following Lal's work, Upadhyaya et a1. investigated the cases of: a) A symmetric mounted implement in a free link mode. b) A symmetric mounted implement in a restrained link mode. c) An assymetric mounted implement in a free link mode. d) An assymetric mounted implement in a restrained link mode. Most tillage implements with the exception of moldboard plows and offset disk harrows are symmetric about their longitudinal centerlines, Kepner et a1. (1978 pp 181). The side components of the soil forces are balanced, the horizontal center of resistance is at .the center of the tilled width and the center line of pull is in the direction of travel. Plows and offset disk harrows can withstand substantial amounts of side draft and proper hitching is necessary to minimize adverse effects on the tractor and implement. Utilizing landsides and furrow wheels, mold board plows and disk plows, respectively, absorb side forces. Offset disk harrows do it by automatically changing disk angles to create a difference between the soil-force side components for the front and rear gangs. Pull type disk plows have essentially free-link pull members whereas mold board plows and disk harrows have 30 bz_oa xz_4 A82 3: um Fag—€83 layman manger. «are; worse on» Co x=__ sweep on» Co u_uuangum .Hq m¢=w~u O. . goosm >fl use acuauuoouuo omen Sosa-.233 as; magnum 2: 9:5: am._mm.z .cwa seduces unuuuuuosoa may . 331 53.3! tie a: 77 . ul .1... . ta .cwm nuances uuuaiuosoa Aav 3 a - -.." ...... 1m 1% weed use. baa “any .33 first too. “SQ a. 33 48:}; 23$ «use A L .2 WON-k 3m. bk! HQMJ Hanan. :M;uat .Pgtfluuuzq. .WOGDM .xfithus.bkuflnuK5u .eaa wouocoo Luna: Aav -4- III \—__/ 3* 49 (Measurements Group Inc.). Fine wire was used for making eletric circuit connections between gages. These fine wires ended up at terminals where they met larger conductors that continued the circuitry. The larger conductors were those from the shielded-four-conductor, insulated cable which supplied excitation voltage and carried output signals to signal conditioners. The cable was firmly held against the sensing pin with a hose clamp safeguarding the delicate wiring from damage by any possible external pull. Each of these cables constituted a channel. On its other end the cable's conductors were each soldered to the inner end of the pins of a S-pin Amphenol terminal connector. The cable's shielding made the conductor for the fifth pin. It is this connector that plugged into the female connectors at the back of the signal conditioner composing a channel. Strain gages and the delicate wiring were protected and secured from the surroundings by a thick coat of silicon rubber. This water, shock and heat resiStant substance also kept dust and chemicals away from the gages. It also pealed off easily for rechecking circuitry after hardware failures and for other needs. ',50 , 4.2 The Data Acquisition System The data acquisition system consisted of the strain gage circuitry, signal conditioners, interface box, a data acquisition control unit (HP3497A) and the central command and control unit, the HP85 microcomputer (see Figure 15). Power supply to run the system was obtained from the tractor battery through a 12V DC - 120VAC power inverter (Vanner Inc., Columbus, Ohio). 4.2.1 The Signal Conditioners Two units were used each capable of handling four channels. Each channel was composed of a card (module) plugged into a mainframe. The card had an amplifier, Wheatstone bridge circuit completion (1/4, 1/2 or full bridge) resistors and balancing circuitry. A power supply of a well regulated 0, 5 or 10V was available for strain gage excitation. The mainframe front panel had a voltage output readout display, calibration references and knobs for adjusting sensitivity, voltage supply and bridge-type selection. Each channel could be programmed individually with a choice of fine or coarse balancing. The amplified output analog signal leaving the conditioner was directed into an interface box which was labeled to assist channel numbering. The signal conditioner units measured about 25x25x30cm 51 r** COMMAND AND CONTROL UNIT MICRO-COMPUTER(HP 35) l SIGNAL CONDITIONER DATA (Strain gage exitation ACQUISITION r “game .———- power supply, bridge CONTROL UNIT 1 'balance. signal filter (HP 3497A) and amplification) l TRACTOR BATTER' POWER INVERTER (DC-AC) .— ------ 4 — - —-' - ' ‘ f | ' I ' I ' I I I lZ-VOLT : . ' l .J 3-PT HITCH DYNAMOMETER FIGURE 15: Schematic of microprocessor-based data acquisition system for tractor tillage forces. 52 each. The signal conditioners had a maximum full bridge input of 0.1V, and a maximum DC input of 10V. They had a maximum allowable input Overload voltage of i 15V, maximum usable output load of lOma (10V 9 1k load) an input impedance of 200 MEG. OHMS and a maximum usable excitation. current of 100ma. Their frequency response was -3 db @ 10 kHz. 4.2.2 The Data Acquisition Control Unit The analog signals from all channels interfaced travelled in one coaxial cable into the Data acquisition Conrol Unit (HP3497A, Hewlett Packard Inc.) The unit consisted of a mainframe which provided interfacing power supplies and support structure for optional plug-in cards and assemblies. It had an Outguard Controller capable of receiving instructions via front panel and/or the remote control interface. Instructions decoded into binary format became useable by the Inguard Controller and plug-in cards. Information returning from these areas was encoded into data suitable for output to the front panel display or remote control interface. Other interfaced parts included the timer which had date options, timer interval (for slow paced 24 hour range scan sequences) timer output (fast scan 100 micro second and 1 second ( 1 0.13) and an internal voltmeter. 53 An Interface Bus Address Switch inside the unit selected "talk" and "listen“ address, which made a two way communication possible with remote control interface. The HP3497A measured about 429 widex521 deepxl9lmm high with a maximum weight of 20.4kg (451b). It was capable of accepting upto 20 different transducer outputs. 4.2.3 The Central Command Unit The central command unit consisted of a microcomputer, the HP85 (Hewlett Packard Inc.). A data acquisition program written in BASIC language controlled the whole sampling, analyzing storage and retrieval of data. The HP85 had a CRT display, printer, tape drive, keyboard and room at the back for plug-in modules. The keyboard consisted in part of a typewriter keyboard, display control keys, 8 special function keys (user defined), system command, numeric and program control keys. Plug-in modules were the HP-Interface, 16 kilobyte extra memory module, ROM (Read Only Memory) drawer and plotter/printer ROM. The unit measured about 400mm wide 160mm high and 500mm deep. 54 4.2.4 The Power Supply In the use of micro computers and other components of a data acquisition system to obtain farm-tractor performace data, electric power can be a critical consideration. Generators have been used in many such applications but they have not been the most reliable, portable or convenient power source. Vanner Corporation (Columbus, Ohio) has developed a sinusoidal voltage inverter that was used in this study. The Vanner (Model 20-500) 12 VDC-120 VAC, 60 Hz, 500 Watt sinusoidal voltage inverter measured 22x24x40cm, weighed about 7kg and was therefore very convenient for use on the tractor. Input power for the inverter was obtained from the battery posts on the tractor to which the inverter was connected. An earlier test of the inverter showed its no-load output voltage to be 120.8 VAC with a voltage spike at points of maxima and minima in the output waveform. Frequency was content at 60 Hz with and without load. At intervals between field-tests tractor engine idle was kept high to maintain the necessary tractor battery charging capacity. With a tractor battery that was in a relatively good condition, this inverter proved a very reliable power source even for the delicate computer hardware. The inverter had two outlet power sockets and a circuit breaker switch which always had to be turned off before the tractor engie was stopped. The unit also had a plug-in remote switch. 55 4.3 Theoretical Considerations In recent years there has been a great increase in the use of digital computers and special-purpose digital circuitry for performing varied signal processing functions that were originally achieved with analog equipment. Stanley et a1. (1979) define an analog signal as a function defined over a continuous range of time and in which the amplitude may assume a continuous range of values. The same reference defines a digital signal as a function in which both amplitude and time are quantized (may assume only distinct values). The process by which an analog sample is quantized and converted to a binary number is called analog to digital (A/D) conversion. In general, the dynamic range of signal must be compatible with that of the A/D converter employed and the number of bits employed must be sufficient for the required accuracy. In the A/D conversion the signal is sampled only at discrete intervals of time and the fundamental question is whether something might be missed in the intervening time intervals. In sampling, if the sampling rate is greater than or equal to twice the highest frequency, the signal can theoretically be recovered from its discrete samples. 56 isP'th where f, a sampling rate fh = highest frequency in spectrum (Nyquist frequency) 2f“ 8 minimum sampling theoretical rate (Nyquist rate) This corresponds to a minimum of two samples per cycle at the highest frequency (Stanley et al., 1979). If a signal is not sampled at a sufficiently high rate a phenomenon known as aliasing results. This concept results in a frequency's being mistaken for an entirely different frequency upon recoverly. Given a set of sampled values we cannot relate them specifically to one unique signal. If the incoming signal content is restricted to frequency components less than gag/2 then no errors due to aliasing are possible. Ideally therefore the input signal spectrum should not extend beyond fa: 57 or fN g f; / 2 or fN 4 l / 2T where fN = frequency of highest spectral component T = sampling interval The signal from the tillage operation is expected to have a relatively low frequency. It is also of a very random nature, (have a wide spectrum) and might therefore be difficult to filter in attempting to hold f8 at a specific value without losing part of the wanted signal. Also if a signal-plus-noise spectrum is sampled, even though the spectrum of signal contains no frequency component above fh , the noise contribution may have them, resulting in a broadening of the sample-wave spectrum and a creation of significant errors. With no knowledge of the highest desired frequency, appropriate sampling interval for the tillage operation cannot be computed. In this study, f3 was selected dependent on memory space available in the micro-computer. A sampling rate of 5H2 was used. This would require fN to have a value of at most 2.5Hz. As discussed above, at this relatively low sampling frequency there is a chance of aliasing. With the lack of knowledge of nature of f“, in this study, a test was made to search for a more severe 58 error, that due to presence of a noise signal. 4.3.1 Resolution and Accuracy of Measurement The force measurement error resulting from the analog to digital signal modification depends on the number of bytes used in the converter. In this study an 8-bit converter was used. The output range was 0 - 10 Volts. Resolution of an Analog to Digital Converter (A/D) is an expression of the smallest change in . input which will increment (or decrement) the output from one code (000, 001, 010, 011 etc. on Binary scale) to the next adjacent code. It is defined as the number of bits or 1 part in 2 . As mentioned above, the A/D used in this study had a 1 part in 256 resolution (0.3663%). Quantization Uncertainty is a direct consequence of the resolution of the converter. Since all analog voltages within a given range are represented by a single digital output code there's an inherent conversion error. If the midpoint of the range is assumed to be the nominal value, there's an Uncertainty of :1/2 LSB (Least Significant Bit). It is common practice to offset the converter 1/2LSB in order to reduce the Uncertainty to 11/2 LSB. Quantization Uncertainty is expressed as :1/2 59 LSB or as an error percentage of full scale (+0.1832% F8 in the case of this study). The error in this study was therefore 10.001832 * 10 - 10.01832V. In this study the sensitivity was adjusted to give a maximum output of 7V. A maximum output reading of 7V therefore would actually have a range (7 1 0.018)V. During calibration the load measuring dial was analog and had divisions of 445N (1001b) and could therefore be read within a 1223N (SOlb) error. The largest calibration factor encountered was 3626 N/V. With reading error incorporated this was (3626 1 223)N/V. Therefore the force output of: (7 : 0.018)V * (3626 i 223)N/V has maximum error: [(7.018 * 3849) - (7 * 3626)] / (7 * 3626) = 0.0623 = 6.23% The smallest change in load that the dynamometer would sense in the 10V range was 0.018V equivalent. For a calibration factor of 3626N/V this is equivalent to 65.27N (2901b). At the maximum load producing 7V (25382N), 65.27N is 0.26% (of maximum 60 load). 4.3.2 Force Components Measured and Circuit Analysis The resultant draft force between a tractor and implement has three components when a side force exists. Part of the horizontal and the whole of the vertical force components in a longitudinal plane were measured by the lower transducers. The upper transducer measured the remaining fraction of the horizontal force. The strain gage arrangement (to make the wheatstone bridge and circuitry determines which force component is sensed by which bridge (see Figures 14 and 16) The total horizontal and total vertical forces were obtained by summing the fractions measured by the various transducers. It is therefore desirable that the bridge measuring horizontal force does not sense the vertical force and vice versa. The basic circuit used with the metal variable-resistance strain gages is the four arm bridge (see Figure 16) with constant voltage power supply. Potential drop across the diagonal b-d is originally equal to zero for the balanced bridge. 61 It can be shown easily that for a balanced bridge: R1/ R2 = R4/ R3 ..... (1) For a change in potential across the diagonal b-d dE due to small chan es in R dR R nd R ( bcl) 9 ll 1) 4 a 3 remain constant and the potential at point d is unchanged. Current flow through R1 is: I1 = a / (R1 + R2) .....(2) and potential drop across R1 is: Bab = I1R1= ER1/ (R1+ R2) oooeooo(3) Change in Eab due to small changes in R1 may be determined by differentiation: dEab a deRlE / (32 + R‘)2 ......(4) Since Eb = Ed at balance and Ed remains constant, the difference between b and d due to dR1 is equal to the change in potential of b and the change in potential at b is equal to the change across ab since potential at a remains constant: 62 Ebd = Ed- (Eb + 61%) = -dEb = -dEab a = - dee1 3 / (R2 + R1)2 ............(5) Using the same approach the equation can be written for changes in R2, Raand R‘ If R‘, R2, R3 and R4 undergo small changes simultaneously due to temperature change or load, the total effect on Ebd is the sum of these effects: 1_ "'1 'deRI Rldkz R dR R an 4 3 3 4 Ebd - ------ 2+ -------§- - ------“E- + 2 E e o o (6) (RI-+112) (R1+R2) (Rf-R4) (R3-l-R2) J By substituting the definition of gage factor: a a (dR /R) /. ...............(7) and considering the bridge circuit is made up using equal resistance strain gages: R8 = R1 3 R2 ’ R3 3 R4 Ebd = (FE / 4) (-‘l +¢2 - ‘3 + ‘4) ....(8) a (FE/4)‘net eooooooeoeoeo(9) 63 .From the above analysis it can be concluded that the imbalance of the bridge E is proportional bd to the sum of strain (or resistance) changes in opposite arms and to the difference of strain (or resistance) changes in adjacent arms. A temperature change will produce a strain a, on all four gages. A look at equation (8) will show that the overall effect on Ebd will be zero. Therefore the 4-arm bridge is fully temperature compensated. A look at Figure 14 will show that a force LS (on the lower left sensing pin for example) will cause equal strain on all four gages. The effect of this load will therefore not unbalance the bridge. The force component LV will cause the same strain on all four gages l, 2, 3 and 4 and will therefore not be measured. The. same force (LV - vertically downwards) will cause a tensile strain on gages 2' and '4' while producing a compressible strain on gages l' and 3'. Applying equation (8) and assuming all strains on the individual gages are equal; net: (2' + e4. - (-¢1I) - (- (3') 4.. 4e where ‘3‘1v3‘2F‘3i3 e4: 64 If strain gages are improperly placed equation (9) does not hold, in which case gages 1, 2, 3 and 4 will sense some of the load LV (cross-sensitivity). FIGURE 16: Strain Gage Circuitry (Wheatstone Bridge) 65 4.4 System Configuration For both calibration and field tests, the whole data acquisition system and three point hitch dynamometer system were fully mounted on the tractor. The signal conditioners, data acquisition unit and microcomputer were carried on specially cushioned wooden structures and held in place with rubber straps. The computer and data acquisition unit was positioned on the operator's left side while the signal conditioners were on the right side, directly above the rear wheels. For these units to fit properly, tractor side-windows had to be kept open. Polythene sheeting was used to cover over the open windows - this kept dust, rain and mud out of the tractor. Continuity between the left and right-placed components was maintained through cables connected to the interface box which was secured on the window glass behind the operator. Signal cables from the transducers were held together in a bunch with tape and then directed through the slightly open back window for connection to the signal conditioners. Power supply cables leaving the tractor battery were bundled with those on the input side of the power inverter. The power inverter was positioned on the tractor floor under the operator's legs. A single cable leaving one of the plug-in positions on 66 the inverter directed power to a socket box to which the various data acquisition system units' power cords were connected (see Figure 15). The whole data acquisition system could be operated by one person, who also operated the tractor during a typical field test. However, if external data like slippage or fuel consumption data was required, then a second person was necessary. 4.5 The Computer Program and Data Acquisition Procedure Computer programs for data collection, analysis and storage were written in BASIC (see Appendix). It is apparent from the flow diagram shown in Figure 17, that five of the eight available specially defined soft keys were utilized in controlling the various data collection operations. Before each run, whether in calibration, verification or field test, the channels were balanced to read zero voltage. However in the event of a shift from zero, resulting off-zero readings were sampled and averaged so as to be subtracted later from the readings obtained. Soft key No.1 was pressed whenever the operator was ready to start data sampling while key No.2 was pressed to stop data sampling for whatever reason. On pressing key No.2 all data obtained at a STATISTICALLY ANALYSE THE DATA ‘rv—i OBTAIN MINIMA AND MAXIMN i FOR EACH CHANNEL PRINT ANALYSIS Rm 67 START ;* l KEY No. 5 SAMPLE AND AVERAGE ZERO READINGS ‘1 KEY No. 1 START DATA SAMPLING STOP DATA SAMPLING ? KEY No. 2 STOP DATA SAMPLING FIGURE-I7: Multichannel sampling flow-chart diagram. V NAIT WANT TO PLOT DATA ? KEY No. 3 PLOT ON CRT 5 COPY ON PRINTER - KEY No. 4 STORE DATA ON TAPE ‘! _ TO STORE DATA. - 7 A SET FOR NEXT RUN ? YES 68 1 KEY No. 5 SAMPLE AND AVERAGE ZERO READINOS KEY ND. 1 ENTER CHANNEL TO BE SAMPLEO 1 ENTER SAMPLING RATE 'RETURN' KEY TO INITIATE SAMPLI WISH TO STOP DATA FIGURE 18: One channel 5””le I l sampling f ow-chart diagram. "5" KEY No. 2 STOP SAMPLING , ‘ I STATISTICALLY ANALYSE DATA if OBTAIN _ . MIN. a MA}. FORCE ‘5' “°- 3 A * , PLOT ON CRT a. COPY ON PRINTER PRINT . ms" ANALYSIS RESULTS To STORE DATA 7 . SET . 5°“ . KEY No. 4 ' A NEXT RUN STORE DATA ON TAPE ? 69 given time t from all channels measuring horizontal force were summed. The same occured for data from channels measuring vertical forces at time t, t+ t, t+2 t etc. This added an additional two arrays of data for storage to the other five, (three for horizontal force and two for vertical force). To check the nature of data collected, and utilizing key No.3, a plot was made (force vs time), for each channel, sum of horizontal and sum of vertical forces. After a look at the plots a decision was made on whether to store and analyze the data. To store the data, key No.4 was pressed. The statistical analysis to obtain the mean, standard deviation and coefficient of variation for each channel's sampled data was done in a subroutine. This analysis was also done for total horizontal and total vertical force. In a second subroutine the minimum and maximum obtained readings were sorted out. This was done for each channel, and for the total horizontal and total vertical force. A printout of results was obtained. It was important that data were stored before sorting it for minimum and maximum values. This was because the sorting process left data in order of size from smallest to largest and it might have been necessary to reproduce a run at a later time. Storage was done on a tape catridge and after this the memory was cleared for another run. Two separate data acquisition programs were 70 _Used (see Appendix). The multichannel sampling program (see Figure 17) sampled 5 readings per second for five channels simultaneously at a given time t. At this rate a four and half minute typical field-test run could be made before the available memory was filled. The single-channel sampling program (see Figure 18) was developed to check the nature_ of data obtained with a variety of data sampling speeds, equal or higher than the standard 5 readings per second of the multichannel program. Sampling speeds upto a sample every 0.1ms (maximum possible) were tried with the single channel sampling. program. At these higher sampling speeds the number of channels sampled needed to be reduced, in order that the available 32k memory space was sufficient. Hence the single channel program. 4.6 System Calibration The calibration process required application of forces on the various transducers individually. Forces and the Wheatstone bridge imbalance voltages they produced were recorded and plots of force vs voltage (calibration curves) were made (see Figures 23 to 25). The gradients of these graphs (F Newtons/Volt) were the calibration factors, (different for different channels) used in the data acquisition computer .program. Draft forces measured 71 were obtained from the equation: P (Newton) = F (Newton / Volt) * (V - Z) Volts where P = Force to be measured F a calibration factor V = resulting off-balance voltage 2 = zero-balance voltage During calibration the pin sensitivity was adjusted using the signal conditioner sensitivity resistors. Sensitivity for each channel was adjusted to give an output reading of about 6 to 7V (60 to 70%) at maximum expected operation load. The 0 to 10 output range was used , with a 10V steady strain gage excitation voltage. After optimal sensitivity setting, the sensitivity and calibration-resistor voltage readings were recorded for each channel. Before each test process these values would then be reset for reproducibility of operation along original calibration curves. The calibration resistor voltage output readings also helped check on possible acquired permanent strain (hysteresis) on gages due to yield or repositioning of the sensing pins under load. Quick system checks for shifts in original calibration were possible even between tillage operation runs during field tests. 72 munch Hmofluuo> Cc COAHMHSSHmI ADV . .oopoa Anacoufihoc Ho nowadaSEHmH Adv unison wCAUMOH coflpmhnaamo .mu mmDOHm Anv Cwmno Haves «my mason wcflcmca Any once; Hao oflasnsca: Aav soaoEoem:>u scam: gain an. LopmSoemcho awasmuchz ANV Loocflaho wcaumoq AHV 73 A test on effect of hysteresis on repeatability of loading curves was made for all channels. Calibration forces were applied in a line as close to horizontal as practically possible. As one bridge was loaded during calibration of lower link sensing pins, the effect of the load on the orthogonally placed bridge was monitored. The loading mechanism was set-up as shown on Figure 19. The force was applied by a retracting a two way hydraulic cylinder. The cylinder had a 7.6cm internal and 3.2cm rod diameter and was capable of applying forces upto 45000N during retraction. At loads of magnitude within the range 0 to 17000N used in calibration, leakage around the piston was insignificant and reasonably steady loads could be applied over a significant period of time. Each sensor-pin was loaded to at least 15568N (3500 lb), in each loading direction. The loading was in steps ranging from 890N (2001b) to 2670N (6001b). A Chatillon hydraulic tensiometer, type HLC (John Chatillon and Sons, New York), capable of handling a maximum load of 44482N (10000 1b),was used to measure forces. The tensiometer was put in the loading-line between the loading cylinder and sensing-pin. 74 Chains were used to couple the loading-mechanism together. Heavy blocks of wood placed behind tractor wheels and locked parking brakes, together with tractor static weight, supplied the required resistance to pull effectively. The loading cylinder was remotely operated with oil from the tractors internal hydraulic oil reservoir. Readings of the load force magnitudes were obtained on an analog scale coupled to the tensiometer by a hydraulic hose. At the adaptor joining the hydraulic hose there was a dampening valve which was utilized in adjusting scale-arrow speed of response. 9: Loading angle‘ VF: Vertical component (bidirectional) HF: Horizontal component R: Resultant force (Applied load) * see Fig. 20 (1) Ti ‘ b—A' my “L (2LT (HF) . J L . _. FIGURE 21: Vector diagram for verification force 75 4.7 System Verification The calibration process produced data in form of forces and the imbalance voltages they produced. Plots of these data produced calibration curves (straight lines through the origin) whose gradients were calibration factors F (N/Volt). The calibration factors were incorporated into the computer program. At this point, the program was ready to control a field test run. However, to be sure measured forces were correct verification tests were necssary. With signal conditioners set at original sensitivities and bridges balanced, a single line pull was made on the three point hitch dynamometer (see Figure 20). The program was run sampling the voltages produced on all channels. Summations of vertical and horizontal forces were carried out and their listings printed out. As the loading took place, a visual monitor of the load applied was made and recorded. This was displayed on the Chatillon tensiometer scale. Comparisons of applied and sampled loads were made. The computer printout of loads separated horizontal and vertical loads, the resultant of which was load R (as read on the tensiometer scale- see Figure 21). Figure 21 is a force triangle derived from figure 20. (R) was the applied load of which the 76 computer sampled the magnitude of horizontal and vertical components separately. The vertical component was bidirectional and could have been positive or negative depending On (R). The loading angle was approximated by measuring distances a, b, and c shown on Figure 20. Then the horizontal and vertical loads could be computed (components of (R)) and compared to those on the computer output as a verification check. Vertical forces in the direction of vector (2) (see Figure 21) were sampled as negative while those in direction of vector (1) (same load direction as used in calibration), came out positive. During verification, loads were raised consecutively in increments from zero to forces varying in magnitude between 2224N (5001b) and 30136N (70001b). Before each consecutive loading, the force was allowed to drop back to zero. Meanwhile plots and data storage of the current loading were made. The chain (used to load the system was adjusted so that the relatively large verification forces were evenly distributed between the two lower pins and that only about 20% of the load was carried by the top-link pin. After the verification process the sensing pins were recalibrated - a check on possibly attained permanent strain. The recalibration was done by switching the calibration-resistor (switch S on 76a .quxoon mewomma umpoSoEmcac soapmoamapm> .om mmoon k\é\\(\ . \u . . r z 7‘ LL r . t v t . >~ a .1. r N! r k r __ z . .wm — r. I l 2‘ Eh .I‘ . III” I» . J . u _ _ c _ _ _. .H _ =;**ueé& JsAILal AW _ \\\ _ a — a l ..--.5. ...... -- ---..._- . . - a a3 _c|& umpmaoamcac sop“: Pain any. Hmong nouomus Aav AU -14/ as getawaao unwumoq any nopoaoENCAG owdzmuuaz ANV 598 A: 77 Figure 16) into the circuit and comparing the output voltage displayed to that recorded during the calibration process. With the loading mechanism used, it was not possible to simulate the expected tillage shock loads. This was because the hydraulic loading cylinder could only retract at a uniform and relatively slow rate. FIGURE 22: Recording-penetrometer probe (seen being forced steadily into the ground) 78 4.8 Field Tests The verified data acquisition system was assembled on a Ford 8700 tractor, ready for field work. Tractor windows and force transducers were carefully covered with polythene sheets. The transducers were also covered as added protection against rain and water finding their way through the silicon rubber coating, onto the delicate strain gage wiring. Field tests made in the middle of summer were arranged to be done in the early morning and late afternoon. Temperatures at these times were bearable and safe for the data acquisition system which operated in a non-airconditioned cab. Any hardware failures that required the disassembling of sensing pins and a gage reinstallation, required a recalibration and reverification. A reassembled pin required a series of at least six "load and unload" operations. This eliminated the obvious hysteresis and as the pin was retightened it showed a better repeatability of output signal for the same load. The most satisfactory test results were obtained in early fall 1984. When conditions were 79 similar to those in mid-Summer 1984, the data acquisition system never showed any weaknesses as a result of dusty conditions or mid-summer temperatures. As with any new system, "bugs" had to be worked out. In the absence of the Vanner power inverter, at one point, a gasoline generator was used as the power source. This operated the system satisfactorily until some mechanical failure caused it to stall. The stalling reduced the voltage supply causing the computer to draw excessive current resulting in considerable damage. This re-enforced the need for dependable power supply. During the course of the field tests, several hardware failures on the dynamometer occured, which were rectified as they took place. The most severe one was a deformation of the top pin support beam. This occured as a result of excessive load on the top sensing-pin during implement transportation between test runs. The pin was not damaged but its support was deformed. A relatively level field was selected for field tests. It had been left fallow for a year. The soil was a sandy loam type and a 15 to 80cm growth of grass and other weeds was growing on it at the time of the test (whose data is reported here). The growth did not cause much implement clogging as the first implement tested was a moldboard plow. Field work was carried out to observe and study the reliability and capabilities of the 80 developed system. Draft maesurements with two different tillage implements, a 3-bottom White moldboard plow and a cultivator were made. Data were also obtained on distribution of moisture, penetrating pressure and tillage depth along the length of the 20mx100m selected portion of the field which was used for the testing. Soil moisture samples were obtained using a cylindrical cone sampler. Samples were taken at an approximate depth of 5 to 9cm, below the soil surface. Penetrating pressure measurements were made close to the same randomly selected spots where moisture samples were taken. A recording soil penetrometer developed by Robertson L.S. and Hansen C.M. (Departments of Crops and Soil Science and Agricultural Engineering, Michigan State University, 1950) was utilized (see Figure 22). This mechanically actuated functional instrument of adequate performance was considered accurate within 9%. The penetrometer results were made up of the pressure required to force a selected probe into a soil, recorded on the abcissa of a graph whose ordinate was the depth penetrated by the probe. Four probes of various sizes were available and the one used depended on the soil conditions. The graph chart was standardized from calibration and factors for penetrating pressure were supplied for the various probes. 81 In determining the complete data acquisition system's limitations, error sources and generally to observe performance, areas of particular interest were: a) Draft force distribution between the three tranducers as affected by type of implement and tillage depth. b) Influence of moisture content and penetrating pressure on draft magnitude and distribution. c) Possible electrical noise generation by tractor tillage operation or surroundings. d) General reliability of the system hardware and ability to maintain calibration and repeatability. 4.8.1 Test Procedure The tillage plot was divided along its length into five sections each 20m long and 6m wide. An additional five sections were measured parallel to the first five. In each section four soil samples were randomly taken. Close to the spots where these samples were obtained, penetrometer measurements were also made. The tillage operation was done along the length of the plot, one section at a time. After an implement was attached to the three point hitch dynamometer the, various channels were 82 balanced and calibration resistor voltages checked. This was done with the implement lowered to the ground surface. The system was then ready for a test-run. The White moldboard plow's cross shaft needed some adjustment to fit properly on the lower hitching points of the dynamometer. ‘The adjustment was necessary because of the shaft's offset design. Also, to avoid any play of the shaft fit to the dynamometer, the implement hitch pins were fitted with adaptors for both lower hitch points. A test run outside the test plot was made after any implement change to be sure implement was well levelled. This was done by checking performance and assuring that none of the transducers was excessively loaded. Working from one section into the next, a 15 to 20m tillage run was made in each. For all tests, engine speed was maintained at 1400 rpm and 4th gear which maintained an average speed of 6.4 Kph. After each run within a section the data were summed (to obtain total vertical and horizontal forces). Plots of force variation with time for each channel and the sums were made. Data were stored on tape after statistical analysis computations were obtained and results printed out. A listing of both horizontal and vertical forces was also printed out. From observation of the plots and statistical analysis results, any hardware failures or excessive loads on 83 one of the pins could be detected. Each run took about 10 to 20 seconds while the summing, plotting, analysis and storage processes took about 5 to 7 minutes. During the latter time, measurements were made of tillage depth. Four measurements of tillage depth were made of which an average was computed and recorded. A total of 10 runs with the moldboard plow were made. Without collecting data the rest of the selected plot was moldboard-plowed to provide enough tilled land for the test using the cultivator. After this, an implement change was made to work with the cultivator over the same testing plot. Data sampling process was the same but fewer runs were made with this much wider implement. The cultivator required no modification to sit well on the dynamometer. Forces were therefore expected to be distributed better between the pins with the cultivator than with the moldboard plow. After the test with the cultivator was completed, the implement was removed from the tractor, the small bridge imbalances removed (using signal conditioner balance resistors) and the system was ready for a noise-signal-test. For this test a variety of sampling rates other than the 5 readings/second, (used in tillage tests) was needed so the single channel sampling program was utilized. Channel No. l was selected for the test. To check 84 whether environmental signals other than those from the bridge imbalance were sampled into the output, a run over the relatively rough tilled ground was made with bridge excitation voltage switched off. The data sampled underwent the same analysis process as that of tillage test. Two different ground speeds averaging 6.6Kph and 10.1Kph (at l400rpm engine speed) were used. A second set of data were then obtained, this time with bridge excitation voltages switched on. Same sampling and ground speeds were maintained and the data underwent the usual analysis. 5. RESULTS AND DISCUSSION To complete the performance test of the three point hitch dynamometer developed in this study, it was necessary to make close observations of performance in field tests and in the earlier processes of transducer calibration and verification. The main objective of this study was to determine the system's capabilities and limitations and these are reported and discussed in this chapter. Thus the dynamometer's potential and qualities as a future research tool can be easily judged. 5.1 Calibration Results Figures 23 through 25 show typical calibration curves obtained. Very high (see Figures 23 through 25) correlation coefficients were obtained, for the linear regression of a load (input)/voltage (output) relationship. Calibration curve data for all five channels are summarized on Table 1. Table 1 includes calibration factors used in the data acquisition program. Note that calibration factors were different for each channel. This is because each channel was calibrated individually. 85 86 see." nutcarm m msslmmmm .m In. mas+mmmm ... no xs+oa> S 35:28 333:: «33833 no 263 322338 .3 "$5on Awhqo>v HDQHDO Ba 84 88 8d 84 is B— (N) 0V0? 87 see.“ noroawm m mss+mmmm .m In. msu+mm- ... no xs+on> afl Au «053:3 Canaan: cannon—moon mo 95:0 coda-.333. ”3N mmaoum Amhqo>v HDQFDO : 844 8.8 Jae. 8M a.“ - . 88 - - Bum - . I-u - . m8: - 1 8.: L . b . It (N) UVUW 88 SSE .H «atonrm m mas-.83 .m .... . $5.83 .- ... xb+ou> 8d .3.~occa:ov 323:: 033.303 no 3:50 scion—3:5 ”mm mmDUHL Amhqoa PDQ-SO 34 8d . ad 84 ad q u 1 q n 1 an (N) DVD—I 89 .2252 39:33 an...» 8 39C made :33 53933-3 n Hosanna no pea on: pawn-mung manmanooomca £30m Mano: Ami-.53 ._.Dn_._.DO Ba Ba 84 88 in I; I“. 1 q q a # m8. ...-aim”. mealmwz .N IA Nsu+mmm¢ .v u xfbu» (N) GVOH 90 .aaasoauo can manacoo acuaoaaou usuvaoa uuuuu a so: 3 33 .m «05.35 you 33 cone-3:3 no no: 02.5333 .Anvom mmauHm Awkqo>v HDAHDO Es BM 84 Ba Ba 84 Ba 89 tsiam .. a! $1.me .m ...n. . IO 89.3mm m . II coma-745.; J g 4 l.— .x L L . t . a" (N) UVOW 91 .moogsdow umoacshioa no noise a sauna c0583... .n Hosanna you aoan :oflpananmO hawawnaamonoh oanmanooo< .mw mmaunm ..Amedo>v Haleao and. mud" nu.v nusm nHJN n-.— u.“— u .. a 89_l§idm msslmmm .N .... r . I. NED?” .m .... Xfou> r. - ... a r ... a c L . . -b: all“— (N) UVO‘] 92 Figures 26 and 27 are calibration curves for channel 3 for 'the first recorded calibration loading after the sensing pin was assembled, and after several load/unload operations respectively. Calibration curves of a correlation coefficient (R Square) less than .97 were considered unacceptably linear (see Figure 26). The linear regression on Figure 26 had a correlation coefficient of .983 but considering that most of the non-linearity occured at low loads it was considered unacceptable. Therefore the magnitude of the correlation coefficient alone was not enough to consider calibration-curve acceptability. Localized non-linearity magnitudes were also considered. The curves on Figures 23 through 25 were therefore considered acceptable and in order to obtain the adequate linearity shown several loading cycles were required. This seemed to reduce the hysteresis effects considerably (see Figures 26 and 27). The transducers were to undergo considerable numbers of load cycles in the work situation. One of the features that determine the quality of a transducer is repeatability. This feature is very dependent on a characteristic called hysteresis. In the loading operation all the energy put into straining the transducer material is not recoverable upon unloading. (second law of thermodynamics). The non-coincidence of loading and unloading curves is due to internal friction (hysteretic dumping of 93 stressed parts. A measure of the amount of hysteresis effects present in the transducers was made. Figures 29 and 30 respectively show the greatest and least cases of hysteresis encountered. They were obtained from plotting data for a complete loading cycle. The numerical value of hysteresis can be measured in terms of either input or output and is given as a percentage of full scale. The largest value encountered was 35%.The transducers used to obtain data in this study had a maximum hysteresis of 3%. Any value above 10% was considered unacceptable. ~Table 1 Calibration Curve Data CHANNEL FORCE Y CALIBRATION CORR. MAX % STABILITY NO. COMPONENT INTERCEPT FACTOR COEFF. ORTHOGONAL MEASURED (N) (N/V) R LOAD (e) % l LRH -558 3626 .999 6.3 0.4 2 LRV -129 3568 1.000 -1.2 0.2 3 LLH 33 2238 .999 12.7 3.1 4 LLV -35 3550 .998 -l.8 8.3 5 UH -598 3280 .980 - 1.3 LRH = Lower right (from operator's seat) horizontal LRV = Lower right vertical force 94 LLH = Lower left horizontal force LLV = Lower left vertical force UH = Upper horizontal force A comparison of Figure 26a to Figure 27 shows the sensing pin was unstable the first time it was loaded. This is attributable to the 'settling' of the transducer support hardware, transducer cement and the strain gage conditioning which must take place before acceptable linearity and consequential repeatability is acquired. Figure 26b is a logarithmic plot of calibration data used on Figure 26a. The plot shows that it did not fit the desired straight line. The correlation coefficient of the two fits compare well (0.961 for logarithmic plot compared to 0.983 for a straight line). Clearly evident is the fact that the more the hardware 'settled down' the more repeatability of the transducer improved, unless excessive strain took place. Figure 28 shows the effect on channel 3 as channel 4 was loaded from zero to maximum load. Channel 3 sensed the lower left horizontal (LLH) draft components while channel 4 (a bridge on the same transducer) sensed the lower left vertical (LLV) draft component. As channel 4 was loaded for calibration, Channel 3 output was also monitored. This observation gave either the amount by which the applied force was not totally horizontal (and therefore had a small 95 vertical component) or error resulting from off-axis placement of strain gages on the transducer. While the load was made as horizontal as practically possible. Figure 28 is supposed to be a measure of extent of the latter error source.The orthogonal load effects for other channels on the lower link transducers are recorded as e on Table l. The plotted relationship between channel 3 and 4 was the worst case encountered (12.7% orthogonal sensitivity). e is the maximum percent of the horizontal component of draft measured as a vertical component or vice versa. In some cases during the tillage operation these cross sensed components may be oppositely directed and therefore cancell. Although in the final analysis total horizontal and total vertical forces add up separately, the orthogonal force sensing was considered an error because various channels were calibrated separately. In other words a channel supposed to measure the horizontal component may have a different calibration factor from that measuring the vertical component and does therefore not qualify to measure part of the vertical component being sensed horizontally or vice versa. Magnitudes involved as can be seen on Table l were fairly negligible. Immediately after strain gages are attached to make a transducer there's a heat-up that results as excitation voltage is fed through the bridge circuitry. The gage cement and soldered joints etc., respond to the resulting heatup in a way that determines the stability 96 .Hmccmzo cmcmoa adamsuom mo usmasonap Hosanna Hmcomosuuo mo usnpsouam cowumowhwuo> you commod aw 0:0 mm mamccmco coomad zHHmcoMozpuo mo mpzmuso mo node comwumneoo < .om mmaon Apr-.05 HDnE-DD nth mama. and" nu.v nUJm uIJN -.u n.“— r I. a Re "3r .. J 'u mmm. a} 3 a. Xfon> I . I l L . b b u ! (N) DVD-1 97 . 3:30am manopmooomc: cw canopcsooco no mwmonovmzn cdoacs\cmoa mcasonm wade < .mw mmaumm 8.5.56 m 4. m N u a. J u u d u 1 IIIN ncoqza L lllv ...-m w L! 939 l!— :— 98 .nucsoam mHQMpnmoom CH nonopcsooco no mummuowwh: udodcs\cmoa mcwsonm Hoag c .on mmaunm SE no» «not: Macaw-u.) IODJIQIUIothu'Ov-Hv— -1-u~ammv.n-uwumu (o ‘J‘DthCOU‘G'mJ-bnm J I . 0 ‘ GHOO‘\JMCDNOJI‘OHO’IOJKDUIID‘DHOuMlomm\D’~10)'~£'IL-lh «BOIO'J‘OHFUIG‘OEUIM NM‘JNAPFMOobl‘Ifiwmiifl-la 0103i OII‘JPOJBV‘DODOJW" Ulb‘fibG'UIJ-U‘OFIVOUWU'WH ”PF“HNf00101f003‘1‘dV‘J'JONOI'JIUHO‘IUHleUIUIUIJL J N N N N to h) N no N m N to N N N r.) h: r.) m N N m M N r.) m m m h 0) (.0 03 co 0) (It 0) 0) (CI co '4 "J "I‘J “-1 K] “I -l "I N ‘\l «1 'J K) “J ‘1'") "I“ 4398. VfiFT. FORCE a. a. m 2 V n-- ”4.30. -22? ~225. ~217. -299. -209. '197. -194. -198. ~189. -1€6. -185. '183. -181. -182. -181. '179. -181. -i? -131. -182. -179. '179. -l73. -179. -179. ~183. ~179. -l?9. -1?3. -1?8. -188. -1?6. -177. -1?8. -1??3 -1?7. -176,‘ -181. ~135. -187. -;87. -184. -182. -1?8. -1?6. -1?5. -l7?. ~177. -1?5 -1?6. -1?S. -1?8. -166. -149. -136. -112 VOOIF‘N‘DNCO'J‘IMHOIOIHQF‘m‘NKDQU-‘CND‘O O'ICO (D'JINOINmMO-‘NHLI'IGJ‘J manuawnwm 44 Q -9e.éé3 -82.245 -72.993 ‘PQ A h - .f_ __ —— - Z V I. " 13») 0.1 I U , a 13 , O J’ _ IL a'gfr _ c. f F- 4 ’ u '— 5 -2 ~— -—-«—-" Q -.7 A l A A 1 g . 12 24 35 4e 56 DRTR NUMBER FIGURE 31(3) Typical verification output at medium size loads 1(30 ilHTIS'IlIJiL UFiTR FDR CHHNNEL 1 RUEPEGE -268 3 MINIMUH? ~287.1 HHXIHUH -167.7 STD DEUN= 25.080 COEFFV 0F VERTN (Z)‘ -9.32 16E 983.2 . .HUH. -287.1 HRXINUN 1864.9 STD DEUN7 186.178 COEFF. 0F UHRTN (Z)= 16.88 STHTISTICRL DRTR FOR CHHNHEL 3 RUEPRGE 1114.8 MINIMUM- 642.8 MRXIHUH 1205.5 STD DEUN‘ 198.960 COEFF. 0F UHRTN (7.); 9.77 STHTISTICHL DHTR FOR CHfiNNEL- 4 RUERHGE 91.1 MINIHUH$ 58.2 HRXIHUH' 98.9 STD DEVH= 9.754 COEFF. 0F URRTN (Z)= 18.76 STRTISTICRL 08TH FOR CHRNNEL: S RHERRCE; 591.0 MINIHUH= 352.9 HHKIHUH- 626.? BTU UEVN= S COEFF. 0F URRTH (X HUERHGE= 2689.6 MENIHUH= 1530.8 HRKIHUH: 2838.5 STD DEUN= 268.971 COEFF. 0F URRTH (X) 18.88 STAT. DHTR FDR TOTRL UERT. FORCE RUEPHGE -177.2 HINIHUH= -237.8 HRXIHUH‘ -72.9 STD DEUN= 28.394 CDEFF. 0F UHRTH (Z)= ~16.B3 FIGURE 31(b): Typical statistical analysis of output from verification at medium size loads. 101 HORIZ.FURCE VEPT. FORCE MlN 596% .)2 1(88SN6)4 5944.6 183.85 5939.5 184.3-1 5935.2 185.72 5932.2 183.88 5938.3 185.55 5924.5 184.35 5925.1 184.25 5917 4 184.59 5916.5 185.15 5913.4 185.35 5989.1 185.76 5986.1 187.75 5981.3 185.25 5898.3 185.43 5898.? 186.81 5893 188.84 5888 3 188.98 588’ 1 186.68 5394.5 186.95 5883.3 188.26 5877.9 187.21 5874.3 185.38 5874.3 189.16 5873.9 189.31 -E (445w) E 0-5 (‘0me l H 0) DRHFT FOR A —L- L‘13‘27 25 13919 Nunee% Ulr FIGURE 32(a): Typical verification output at large sii loads. (Sampling bequn after load had stabiliudfi‘ 102 STRTISTICHL 08TH FOR CHRNNEL 1 RVEPRGE 2397 3 MINIMUH- 2382.9 HRXIHUH 2428.6 STD DEUN: 9.698 COEFF. OF UHRTN (Z): 48 STfiTISTICfiL 0878 FOR CHRNNEL= 2 8062866 '68.? HINIHUH' 2382.9 flRXIHUHI -67.8 8T0 OEVN= 1.188 COEFF. OF VHRTN (Z)’ -1.68 RHERRGE‘ 2743.8 MIHIHUH= 2726.4 MRXIMUM; 2768.2 8T0 DEUN= 11.586 COEFF. OF UHRTN (X): .42 STRTISTICHL 08TH FOR CHRHNEL= 4 RUERRGE- 174.7 MINIHUH= 173.6 HRXINUM= 176.3 8T0 DEUN= .795 COEFF. 3F VERTN (Z): .46 STRTISTICRL 08TH FOP CHRNNEL= 5 RUEPHGE' 767.5 MINIMUM“ 768.9 HRXINUH‘ 776.4 STD DEUN= 3.997 COEFF. OF URRTN (21' .52 FIGURE 32(b): Typical statistical output of large size load verification output STRT.DRTR FOR TOTRL HORIZ. FORCE fiUERfiGEi 5987.8 MINIHUH= 5873.9 HRXIHUH= 5965.2 8T0 DEUH= 24.952 COEFF. 0F UHRTN (Z) .42 STRT. 08TH FOR TOTHL vERT. FORCE RVERRGE= 186.8 NIHIMUH= 183.1 HRXIMUH= 189.3 3T0 DEUH= 1.792 COEFF. OF UHRTN (X) 1.69 103 .Mumu :oMumonwuo> HazaOM wcflm: mouom ummuc noanadm cam umuammms mo «can comwhmnaoo < Amvnn mmaoum Gaga—mi as 3“ 88¢ a . .- . . 4 i u 104 vof the bridge output. This stability is measured by monitoring bridge output over an extended period of time following gage installation. Stability depends on quality of transducer construction. Externally induced electrical noise is also a stability determining factor. In this study bridge stability was determined by recording output approximately every three hours over a 24 hour period. Voltage (output) shift away from the balanced zero reading over the 24 hour period was the measure of stability. The last column on Table 1 gives percentage shift of output voltage (shift volts/10 volt-maximum output range). From Table 1 it can be seen that the worst case of stability encountered was 8.3%. In the field test measurements made were over a maximum period of 30 seconds. Therefore assuming a linear output change it can be concluded that instability error was very negligible. Column 1 of Table 1 shows the y-intercept values (output at no load). The existing outputs at no load can be attributed to both hysteresis and the nature of the loading mechanism used. The loading mechanism (see Figure 19) support was rather unstable and at the end of each loading operation the chain work and hydraulic cylinder took positions that tended to put some load on the hydraulic dynamometer. However as long as the linearity of calibration graphs was at acceptable levels, it did not matter that y-intercepts were 105 .Masu cowpmoflhwno> mafia: woman vumud uoaaemm can cousudoa no wean somwnmqeoo wcwppnu o>u=o < Anvnn Mancum C. . sz women. omm3m nu“: aqueouuuo: no souuusaloo "awn adamam 3 mos“. guhmm> §>< ’N g in . a” s as: q d oi. .....Srm a glans.” In. . vaa+mmmmg Io xn+ol> (N) 33803 'IVJMJZIW WV 119 £32533 a non cowom 13:2, :3: ~333qu «a sour—cg 3cm gum... .. . 3 HEP... .2359, wo<¢u>< I :3 g .30 an {n X r I an *8 . .823 m x x uaa+mmm~.- .3 eas.mm~e.a no xn+o.»_ (N) 3:180: 'IVLWZIW BOWBAV 120 negative vertical force implies a vertically downward directed force. For analysis and comparison purposes absolute values of the vertical forces were used.A rise on the negative scale is actually a reduction in magnitude of the vertical force. Figures 36 a and b and 37 a(i) through d(ii) are graphs of the data on Table 3. The plots are for comparison purposes and they show variation between horizontal and vertical components of draft force and the influence on these forces of some soil physical properties. Only very general tendencies were sought in the analysis because farm soil is a complex medium to work with and much more data than was collected in this study would be necessary if conclusive remarks were to be made. Data collected in this study is therefore only illustrative and not conclusive. For conclusive studies it would also be necessary to measure effecct of other physical properties like cohesion, friction shear strength etc. It would also be necessary to measure side draft especially for a complete comparison of draft force distribution. To make the general comparisons and using the data on Table 3, linear regressions were made as a means of checking the general tendencies of various parameters to influence the horizontal and vertical components of draft. Logarithmic (y= a+ngx,exponentia1 (yaaex' ) and power (ysaxb) regression curves were also tried for a comparison of correlation coefficients between the various regressions. Compairing correlation 121 aw“. Auoroawm a. Nas+mhmm .N In. vss+mm~v A no xn+ou> .30“; oncogene! a new: «ensue: can: ucoucou ousuodol saw) ouuou anusouwuca mo douuanuc> « any PZMPZOU wmnhmao: Jnom can can wan man J- d adv nun muacuh v4: (N) liVHO WVLNOZIUOH BSVUBAV 122 .33.— uuaonvaos - suns mud—.33 no.5 33:3 33:8 5.3 ouuou 139:; an .3313: n :5 can :8: av ...zmhzou wghmmo: .38 all. actuarm a Nss+wmms.~ In mss+momm.l Io xn+on> 9.8 94.6. udm adm 9.3 . 3.3 d 4 4| u vwvw (N) 3380:! 'IVDILHEA BOVEBAV 123 coefficients they varied in difference from that of linear regression by 2% (for Horizontal draft versus Moisture Content) to 54% (for vertical draft versus moisture content). Although 54% might seem a large difference graphical appearances were visually very similar. The linear regression was therefore selected as a representative basis of general comparisons.Correlation coefficients (R Square - see Figures 36 and 37) for linear regression varied between.006 and .749. Figure 36a (Average Horizontal Force versus Average Vertical Force) shows a tendency for horizontal force to increase as vertical force increased when working with the moldboard plow. The corresponding graph (Figure 36b) for a cultivator shows a decreasing tendency of horizontal force as vertical force increased. As mentioned earlier the hook-up of the moldboard plow on the quick-hitch needed cross-shaft modification. Moldboard adjustments therefore might not have allowed for a uniform force distribution. However as the resultant draft force increased, the components (measured) were expected to increase. This was the tendency observed on Figure 36a. Adjustments for the cultivator were non-problematic although at times it would tend to clog-up. Distribution and magnitude of force components when working with the cultivator were expected to be more random (than with moldboard plow) l4 2 1 all. noroarm m m8+w-m .m In. mglmmmm .v ..o xn+ou> .33.— vusoa 30. a :3: 9.333 can: 3:33 .33 53 «8.8 1.8329. so 52325 u 3 an... 9:8... I." on.— J A83 £528 58 #8 3.— J 8.— 4 . l .Ngruz g— g— g— (N) lavas "lVlMZIHG-C WV 125 mg . :93er m magic 3. .N In. ¢8+mmh~ ... Io xa+o.» as." .33.. ...—no.— 33. a :3: 9.3.33 5:: 5:32. :3 fit. «83 1332, no 82325 u $3 in 9:62 38mg >2sz 53m 48m 8..“ B.” ,3; 8." a." r _ L! (N) 3380:! "IVO 1183A 39V83AV 126 was. ..goawm m ~8+wmmm .I In vss+m~mm .u no xn+ou> .33.. ...—.338 a new) 95.333 nos: cog-a3.— nouuuuuoaoa a»? .38 5333.3.— uo 333.5? n 63 3n 5:5: «0&5 mu2 . r X X X (N) mun 'IVDILUSA SOVHBAV 128 and that seems evident when Figures 36a and b are compared. The slightly negative (linear regression) slope of Figure 36b may be only a coincidence. The data is relatively randomly distributed. Again much more data would be required to be able to make conclusive remarks. Too many variables associated with the tractor, terrain and soil physical properties determine this distribution. Figures 37a (i) through 37d (ii) show plots of variation of both horizontal and vertical forces compared to variation of other parameters mentioned above. Figures 37a (i) and (ii) shows that moisture content of the soil has an effect upon both the horizontal and vertical draft components. The magnitude of horizontal draft component depends upon the coefficient of friction between the plow material (steel) and the particular type of soil. The magnitude of the coefficient rises to a peak and falls as moisture content is increased from a relatively low percentage value. Between moisture contents of 12 to 25%, soil goes through an adhesion phase. It reaches a peak coefficient of friction and then begins to decline as the higher moisture begins a lubricating phase. Regarding the vertical force variation with moisture content, several investigators have found that soil-metal coefficients of friction decrease as normal loads become larger, particularly in moist clays and clay loams (Kepner et al., 1980 pp121). 129 8mm. uOLODTm m Nss+mvam.m um vss+mmom.~ no xn+on> n.o~ .33..— vuooagol I :33 3.3.33 62.: gauge ou-Hduu so“: .uuou fluuuouaua: uo aoaunauas u «My can museum Axuv Ihmmn mo<44~h mo<¢m>< ~.s— v.v~ - o.u~ J unsh— L auunw lSVBO WVLNOZIHOH 39V83AV (N) 130 as“. ..Loawm m Nss+wmum A In. mss+mmmm .... no xg.ou» ndm .335 23338 a no“: wads—no: can: . £300 09333 £35 uouou H3332, no souudnuflr u :3 2n 55H.— 99 5.me ”2.3:. wo< vJN odN Nd; . 9.2 ad— 1 J d a x . X 1 (N) 3380:! "IVDILHBA SEWER/W 'v.‘ “’0 N w hJ (pl N to h) 0) (d to [‘0 h) It) 01le lo h) to h) rt to IV N to h)" 31".! 01 he to (S- 01 ID “I o- 0) to (7" “J ('1 In 4!) ”1'00! 1’8 v- 01 h.- U| I.) H~ H mmuumoza-r-ans h) 5 (uh-«Ho 'DO‘O'DOIOJOKJUI$-UI‘O'fl‘OU‘IQOJO‘ut-IQMH“JVGHJI cutouts)‘IWUFOO‘W'fi'Oh‘L-IHO-‘JCO‘HOUIO-‘DU'QUI (. I w '20 .- 0) N‘DHOJNIUtJHHHlf-QNIJBHH'ANO)\DOJOI l D) UERT (445M3 -63! -628.3 -746. -?32. -784. -?68. -”51. -?76. -588. -979. I D 0] UI (u 45 ' “Mlutu N (4 u) w (n 91.9 DRHFT FORCE (445M) “WW W {3.0 M "J ijfi. min'flooC"J‘C'Uo"‘0(0 ‘lNOtv-‘OJO.‘JOIHU| ' Jet—"J- AA . A . A. f ‘ 18 “‘6 :5 . Darn‘uungép *4 go 5 FIGURE 38(a) Typical ti11age force output showing evidence of hardware failure. (Allforces in 445N) 1132 STRTISTICHL DRTH FOR CHRNNEL- 1 RUERRGE' 818 4 MININUH= 235.3 HHXIHUH- 1793 4 STD DEUN: 381.364 COEFF. 0F URRTN (Z) 46 68 RUERRGE .8 HINIHUH= -.1 HRXIHUH= .2 STD DEVN= .0?? COEFF. 0F URRTH (2) 444.78 STHTISTICHL BETH FOR CHRNNEL= 3 RUERRGE- 431.6 MINIMUM -396.2 HRKIHU‘ 926.2 STD 06;- 295 631 COEFF. 0F UHRTN (Z)‘ 68.49 STRTISTZCHL ORTR FOR CHRNNEL: 4 RUERRGE' -812.7 MIHIHUH= -1133.6 HRXIHUMz -454.9 570 DEUN= 155.274 COEFF. 0F URRTN (Z)= -19.11 STHTISTICRL DRTR FOR CHHHNEL: 5 RUERRGE 1615.8 MINIMUH= 1861.5 HRXIMUM4 2652.0 STD OEUN= 233.227 ’) CDEFF. 0F URRTN (L srar.oara FOR TOTRL HoeI:_ FLRCE AUEPRGE 2865.1 rlulnbm 3116 3 M— “MM" 3869.? 5‘: 2” ~ 434.?61 226:: J' URPTN '). 14 1: 579* BETH FOP TOTHL UEFT FGFCE fiwEPfiGE -812;? Nihtfiur -1 33 7 HRHZfiUH -4S4.9 STD JEUH= 155.233 COEFF. 0F UHFTN <25: 713:11. FIGURE 38(b): Statistical analysis results for output with evidence of hardware failure. (All forces in 445M) 133 1. . 3 . S ._ ? L. E _ 1... 1 O. M" %1 3.: — ... 9. 8 M H . 337... I. C. 8095 s“ . 8683 ..r R. ...6- 2 0. EEE.) R F. 56812 T - . . .Albi H a. 8683 .50 T. TSéSN 1 .H . 87.64T n...65499 . 63226 L _ 111)....v 9. ... 8 C. -F 01. - . .. . N0 .- . EHMU 5 . GUUE . I . .fiHNDF I. ..I W" HNHOU 9.3 3 3 3.33.? T. UIHTO 3.2 1 1 . _ 3. RMNSC 23$ 3%... Emma E C F 0 F )346 293 21.- 64154 9565 863... 263 94 88 ...4 “:38536898985611.198349681461134STS49311 F42122111688£44111123999a98883.56..6387?? 4444544444 4544 4. £11.97: qs.~0~.uqd.jq3q.4 4e:u-.c.:u::5c.5c.4 J. U0 . ._o.:.s.:.a.:.:.e.:.:.:.o.:a:_o.1.9.7.o.o_o.saa.:.2.c_o.w.o.¢.a_:.o_:.:.:.1.:_ 1.5111.1-“QIL‘.1.1...11111 01“110"0111lololotilbiqoouao.ou.1a EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEESEEEE \I.:.4.... ...;44. ...; ;.;._.;...4:.:.a.¢.9.¢.scp_a._. ..aas_:_e..£a:.c.:...2.:. ... ._ ZN 6666. 666. c666666666666€666. o. 66666666.66.95.26 IMTT?T?TTTT .....?..~T???T?r?TTT?TTTTT.T P nylx...4.:.:.o.....s.......,_:.J.;.o.:.a.o.:.s.s.3.....:.o.s.:_:.:.4...J.._:_:_... . . . . _ __ FORCE FIGURE 38(c) Typical output showing evidence of short-circuiting problem in the wiring of channel 1. 134 While no coefficient of friction data were obtained in this test, no conclusive discussion can be made from plots of Figure 37a (i) and (ii). A look at variation on Figure 37a (1), however would imply an adhesion phase as horizontal draft seems to increase with increased moisture content. Figures 37b (i) and (ii) show that as penetration pressure increases both horizontal and vertical force components decrease. It would seem reasonable to expect that there is reduced implement penetration and hence lower horizontal resistance at the implement/soil interface as penetration pressure (resistance) increases. This also results in a lower vertical force. In the most part this is supported by the graphs obtained. It would also seem reasonable to expect that penetrating pressure should vary in an equivalent manner as the bulk density. The same argument would then hold for variation in vertical and horizontal draft. However soil of a large bulk density is necessarily heavier and more difficult to not only turn over but also pass an implement through (due to large cohesive forces). The tendency observed of both horizontal and vertical draft components to increase with increasing bulk density supports this theoretical argument. Figures 37d (i) and (ii) both show a general increase in force with increase in tillage depth. From theory it would be expected that increased depth should _FnRCE(445N) -238 53 -233 S: - 63845 -.881}§ I h‘lb$"fl‘.f’aj&AHHP‘me‘OsHHHO-‘O-‘Nm‘w' beeecficeeeceécmememmmmmmsawasasaaaaa»as comm «J~-.1-Jwwrowoocouwromum-umwwoomoowmloumcocooooorowuw gr.SUSHSHSHSIGISIQQHH@QQSOGIIS‘QGISIISII‘HHD‘HQQHHHPHGIHix. Awnwmmmehmbwwe h." to rr- 0;: I33 0:. (0 m he 0: l"-.' I“... M N N GO N N to [no 03 ru m N h." N to he 03 [U m m f‘o.’~ ['0 w 02- Iv N 02 f. .32623 -.03642 .6?893 -.68112 -.38112 .02628 .8?B98 62828 .62823 -.63842 .62623 E (445M) -- 'd FUR DRHFT 135 STRTISTICHL MINIMUM“ MRXIHUH' STD DEVN. CUEFF. 0F unRrN (zjg‘ "‘ H r.: r I "A III I.-J (I) '33 r DJ 09TH FDR CHHNNEL- 1 -4 s -233.5 .1 3' ‘68 I U L. l ‘1 r l l’ FIGURE 39: Noise-error test without strain gage excitation. FORCE(445N) ago-5pm. -!0~%Hflw ll «BOP bio-h Mr“ N- 1 ‘5 SIP-30101 OJ H '3) UI #- 03'1" 01011.0 0‘- ' "OJ CC! O-I‘ 4.1.9.!” .‘3 r0. HW‘UIU‘WU'W‘JUL-IUH l . a]. '3' h) 01" '1) 91% J“ 0‘: UI Ul 1'0) 3- '-I' W pr, —.J -.J ‘-.j . H '30 com .19985 -4.4654 -4.1185 -16.?85 -1.B1?8 -5.1752 -6.?9?6 -51.819 -18.154 136 BTRTISTICHL 09TH FOR CHRNNEL= 1 HUERHGEI :4.6 n1u1nun= -g1.e MHXIHUH‘ 163.? q STD DEUN= :58.1_9 _ ¥fi COEFF. 0F UHRTN t4)‘ ‘Ols.9t 33- % 3 1 g ‘- '- *’13~ U] U ,. g 13 u. ’3 i h 3- u. . ‘---«—-—_.—---—____.______ ... _______-- - ¢ -21 m C1 .. q- k .l 1 _._ A14 K“). p... IDP N FIGURE 40: Noise-error test with strain gage excitation. 137 tend to increase draft. Consequently therefore both horizontal' and vertical components of force should increase with increased tillage depth. At larger depths there's more soil/tool interaction hence more frictional resistance and soil weight. The resultant force angle also increases (see Figure 21) as depth increases. Figure 38 demonstrates how easily evidence of hardware failure was detected in the output. Output graphs with excessive readings, (shown by vertical curves - Figure 38c), or very close to perfectly horizontal line graphs (zero - reading) Figure 38b, implied system misfunction. Excessive unexpected outputs, were evidence of short-circuiting in the bridge wiring. Non-varying zero voltage implied discontinuity in the wiring. This was also evidenced in the statistical output (see channel 2 statistical output, Figure 38b). A distinctly different coefficient of variation is also recorded for this channel. 5.3.1 Noise-Error Sources Test Results Figures 39 and 40 show results of test for error due to electrical noise. Figure 39 is a typical output of data sampled in the field with the bridge circuit power cut off. The corresponding single channel statistical output data is shown. Figure 40 shows the output for the noise test in the field with bridge excitation switched on. The test 138 with gages excited is a little more irregular, a response by the transducers to the tractor vibration during travel over the rough ground. At the beginning of the test there appeared to be a trauma in the system which resulted in a noise signal. This signal is recorded at the beginning of the graph of Figure 40. A look at the. list of forces sampled, on this same figure shows a relatively large first value which makes the mean appear to be relatively large. The averages with and without gage excitation seem close and~ approximately zero except for the noise error due to starting trauma. While several sampling speeds were tried, no significant difference was evident between the outputs. Thus the error due to noise, except at the beginning of a particular test, was considered to be absent. At higher sampling speeds some noise from the surrounding engine electrical systems and hardware vibration, may affect the data. 5.3.2 Hardware Performance At the end of the field test, the dynamometer and its related hardware had been exposed to vigorous shock loads and vibration. The transducers stood up to the torture quite effectively. The only weakness was the top transducer support. While this beam performed well in tillage it showed weakness in the implement transport position. There was excessive load arising from the 139 implement overhang. This load was more severe to the support than to the pin itself. The transducer stayed relatively straight and maintained calibration, but the support was deformed. At the end of the field test, a check was made to determine how much the transducer outputs had changed from those at calibration time — a test for the amount of permanent strain acquired. This was done by rebalancing all channels, switching the ‘calibration resistors into the circuits and comparing outputs with those obtained at calibration time. Table 4 shows the resulting percentage differences for the various channels. The values indicate the amount of permanent strain acquired during the tillage process. Table 4 Permanent Strain Test Results. Channel Force component Calibration measured change (%) 1 LRH 6.3 2 LRV 2.2 3 LLH 5.4 4 LLV 1.8 5 UH 14.0 LRH a lower right horizontal force LRV a lower right vertical force LLH 8 lower left horizontal force LLV a lower left vertical force A count was made of number of times sampled loads exceeded the maximum calibration loads. The load 140 that exceeded the calibration by the greatest amount was that of lower right horizontal component when working with the moldboard plow. These loads exceeded maximum calibration loads 27% of the time. There was no detrimental effect of these excess loads on the dynamometer. 141 6. SUMMARY AND OBSERVATIONS 6.1 Summary jTillage consumes a major part of the energy input in the farm system. To increase knowledge of energy required by the tillage operation, forces between the implement and tractor. need to be measured. As fully mounted and semimounted implements have increased over the years, the three point hitch dynamometer has become necessary. In development of strain gage transducer measurement systems, placement of the strain gage transducers has varied. Of the various alternatives, a frame to carry the transducers between the tractor and implement has become the most convenient. A three point hitch dynamometer was developed, which used a quick coupler to support the strain gage transducers. This dynamometer was calibrated, verified and field tested for capabilities and limitations as a future tillage energy research tool. A micro-computer based data acquisition system was utilized for obtaining data at all stages of the development and testing process. A computer program was written to control data collection. The same program was capable of doing a statistical analysis of the data making the whole process capable of giving real time performance feedback. Two implements were used in the on-farm tillage 142 test. Data for tillage forces, tillage depth and soil physical properties were obtained. Results were analyzed to see if relationships between variation of the forces and variation in soil physical properties were as expected. This test confirmed quality of performance of the dynamometer. Presence of electrical noise in the measurement system was also examined. A reasonably good demonstration of the system's performance capabilities and limitations was accomplished. With the integration of other energy-related parameters of the tillage systems, into the data acquisition system, a full input/output energy utilization study is possible. 6.2 Observations The dynamometer developed, proved to be relatively simple and economically feasible to build. After the weaker points were corrected, the dynamometer's performance was excellent. The data acquisition‘ system and power supply proved very satisfactory and convenient. The cantilever type force transducer support will require strengthening because this is currently a weak point. The following are general observations about the dynamometer performance: a) The transducers maintained calibration effectively with the exception of a maximum permanent strain of 14% for the upper transducer. b) The dynamometer is a useful tool for a detailed measure of draft force magnitude, distribution and direction 143 among the three linkage arms. The effect on this distribution as influenced by type of implement, tillage depth and soil physical properties suggests many possible and interesting studies. c) At the signal sampling speeds used, no significant amount of electrical noise appeared to affect the force-initiated off-balance signal. Much higher sampling speeds would be required to detect any such signal, if it existed.' d) The ability of the data acquisition system to measure, store and make real-time plots and analysis of data provided great time saving and convenience. It was possible to enter tillage test data before each test. e) Errors and hardware failures were easily detected from the observation of the output. Software errors could be quickly corrected while still in the field. E) Performance accuracy of the dynamometer was better at higher loads. Transducer sensitivity was adequate for all data sampling speeds used and for the resolution required. Hysteresis on the transducers was insignificant. g) Total force on the tractor due to the implement soil interaction was not a function of soil alone. While hitch geometry and terrain may determine the distribution and magnitude of the force, there may be a dynamic load reaching the dynamometer and therefore getting measured as if it was part of the draft force. This load depends on parameters like type of field, length of plots tire pressure etc. It was not separable from the load due to soil alone. 144 6.3 Future Work The development and testing of this tillage force measuring system was part of an on-going effort to develop a fully instrumented tractor for complete input/output energy analysis. Although the tractor instrumentation system has experienced various tradeoffs among time, cost, accuracy and availability of equipment; (as it has undergone development), the three-point-hitch dynamometer was the part of tractor system that required major effort. Other parts needed for the instrumentation package will be required to measure side draft, drive wheel axle torque and rotating speed, fuel consumption, PTO torque and speed, ground and engine speeds. While the three-point-hitch dynamometer performed well, its main weakness was physical strength. To correct this the following should be done before more tillage work is conducted: a) Heat treat the transducer support beams and make them from a completely solid steel beam, with necessary framework to reinforce them. b) Locally manufacture sensing pins which will in turn be made long enough on the threaded side, hence enough room for more grip by the nut. The inner locking half-nut will be replaced with a collar hence supplying enough locking strength. This may need a minor modification of tractor links if the diameter of the pins is increased. c) A cross-bar should be made, for use across the two lower hitching points, for one and two point semi-mounted implements. d) A more convenient. calibration loading mechanism should be developed. A system that utilized static loading units (weights) would be excellent. This would eliminate the dynamics of the load encountered in the system used. e) A more moisture resistant strain gage cover (other than the unreliable silicon rubber) should be used. This will prevent spurious strain, swelling or contraction of cement, loss of insulation resistance within the wire grid plus electrolytic polarization that was possibly caused by moisture that leaked into the bonding area. f) An instrument with an analog electrical signal output to measure tillage depth would be a useful part of the instrumentation package. 9) A study on a means to separate the dynamic load 'on the tractor from loads due to soil/implement interaction alone is suggested (see Section 6.2) REFERENCES Barker, G. L., L. A. Smith, R. F. Colwick, 1981. Three point hitch dynamometer for directional force measurement. ASAE Paper No.81-1044. Barker, M. 1974. Unattended field measurement instrumentation. Society of Automotive Engineers, Paper No. 740940. Bedri, A. A., 1982. Computerized data acquisition system for tractor field performance. Unpublished Ph.D. dissertation. Library, Iowa State University, Ames, Iowa . Beppler, D. C. and M. D. Shaw, 1980. Tractor instrument package for field energy studies. ASAE Paper No. 80-1522. - BLH Electronics, 1980. Strain gages (SR-4) handbook-HDBK 103, BLH Electronics Waltham MA. BLH Electronics, 1982. Strain gages (SR-4), BLH Electronics Waltham MA. Carnegie, E. J., R. R. Grinnell and N. A. Richardson, 1983. Personal computer for measuring tractor performance. ASAE Paper No. 83-1065. Carr, J. J., 1982. Microprocessor interfacing. TAB Books Inc., Blue Ridge Summit, Pennsylvania. Carter L. M., 1981. Instrumentation for average draft. Transactions of ASAE 24(1):23-25. Chung Y. G., S. J. Marley and W. F. Buchele, 1983. A data acquisition system for tractor field performance. ASAE Paper No. MCR 83-120. Chung, Y. 6., S. J. Marley and W. F. Buchele., 1983. Development of a three point hitch dynamometer. ASAE Papeer No. 83-1055, 1066. Clark, J. 8., and J. R. Gillespie, 1979. Development of a tractor performance meter. ASAE Paper No. 79-1616. Clark, J. 8., and J. R. Gillespie, 1982. Field experience with a tractor monitor. ASAE Paper No. 82-3536. Culpepper, W. J. 1979. Description of a data-logger/analysis field test system. Society of Automotive Engineers, Paper No. 790522. Dally, J. W., and W. F. Riley, 1978. Experimental Stress Analysis (2nd edition), McGraw Hill Book Co. Inc. Devine, R. J., 1973. The effects of livestock waste on some soil strength characteristics and selected tillage operations. MS Thesis, South Dakota State University, Brookings S.D. Doeblin, E.O., 1975. Measurement Systems, Application and design. McGraw Hill Book Co. Dove, R. C., and P. H. Adams, 1964. Experimental Stress Analysis and Motion Measurement. Charles E. Meril Books Inc., Columbus, Ohio. Dowding, E., J. A. Ferguson and C. F. Becker, 1967. A comparison of four summer fallow tillage methods based on seasonal tillage energy requirement, moisture conservation and crop yield. Transactions of ASAE 10(1):1-4. Flis, T. J. and F. B. Cupp, 1974. Automotive high-density digital tape recording system. Society of Automotive Engineers, Paper No. 740219. Ford Tractor, Parts catalog and drawing for lower linkage system. Garner, T. H., D. Wolf and J. W. Davis, 1980. Tillage energy instrumentation and field results. Proceedings of 1980 Beltwide Cotton Production Research Conferences, pp117-120. Gill, W. R. and G. E. V. Berg, 1968. Soil dynamics in tillage and traction. USDA, Agriculture Handbook No. 316. Godwin, R. J., 1975. An extended octagonal ring transducer for use in tillage studies. Journal of Agricultural Engineering Research, 20:347-352. Greenwood, D. T., 1965. Principles of Dynamics, pp518. Prentice Hall, Inc., Eglewood Cliffs, N.J. Grevis-James, I. W., D. R. DeVoe, P. D. Bloome and D. G. Batchelder, 1981. A microprocessor based field data acquisition system for tractors. ASAE Paper No. 81-1578. Grevis-James, I. W., and P. D. Bloome, 1982. A tractor power monitor. Transactions of ASAE 25(3):S95-597. Gunderson, D. G., T. G. Kirk, J. N. Wilson and F. B. Dyck, 1981. Draft-Speed-Depth characteristics of cultivators and diskers and their effect on fuel consumption. ASAE Paper No. 81-1603. Harter, D. D. and K. R. Kaufman, 1979. Microprocessor based data acquisition system for tractor tillage measurements. ASAE Paper No. 79-5026. Hendrick, J. G., C. E. Johnson, R. L. Schafer and J. D. Jarrell, 1981. A microprocessor based field data acquisition system. ASAE Paper No. 81-1577. Herron, M. M., M. D. Shrock, J. A. Kramer, D. K. Kuhlman and S. J. Clark, 1977. Instrumentation of a farm tractor using off the shelf components. ASAE Paper No. MC 77-501. Hetenyi, M., 1950. Handbook of Experimental Stress Analysis, 1950. John Wiley and Sons Inc. Hoag, D. L., and R. R. Yoerger, 1974. Designing load rings for measurement. Transactions of ASAE, 1974. Jensen, J. K., 1954. Agricultural Engineering, Vol. 35, No. 9, pp 615. Johnson, C. E., and W. B. Vorhees, 1979. A force dynamometer for three point hitches. Transactions of ASAE 22(2):226-228. Joint Task Force of Southern Region Agricultural ‘ Experiment Station and ARS, USDA, 1976. Energy in Agriculture, University of Kentucky, Lexington, ky. Kepner, R. A., R. Bainer and E. L. Barger, 1978. Principles of Farm Machinery. AVI Publishing Co. Inc., Westport, CT. (3rd Edition Ch. 5, 8). Lambert, W. R. and G. E. Miles, 1981. Intelligent control of multiple servos with a single board computer. ASAE Paper No. 81-5029. Lal, R., 1959. Measurement of forces on mounted implements. Transactions of ASAE 2(1):109-111. Lin, T. W., R. L. Clark and A. H. Adsit 1980. A microcomputer based instrumentation system for field research on a tractor. ASAE Paper No. 80-5525. Luth, H. T., V. G. Floyd and R. P. Heise, 1978. Evaluating energy requirements of machines in the field. ASAE Paper No. 78-1588. Muhtar, H. A., 1982. An economic comparison of conventional and conservation tillage system in the South East Saginal Bay coastal drainage basin. Ph.D. dissertation, Agricultural Engineering Dept., Michigan State University, East Lansing, MI. 48824. Murray, W. M. and P. K. Stein, 1956. Strain gage technics lectures and laboratory exercises. Massachusetts Institute of Technology, Cambridge Massachussets. Neubert, H. K. P., 1967. Strain gages - kinds and uses. St. Martin's Press. Neuholt, A. L., 1959. Measuring forces in two or more numbers with one instrument. Agricultural Engineering 40:456-457. Neville J. H., 1956. Unpublished MS thesis. University of Durham. Orme, R. W., 1976. A full-portable computer installation for radio telemetry data acquisition and reduction. Society of Automotive Engineers. Paper No. 760680. Perry, C. C. and H. R. Lissner, 1962. The strain gage primer. McGraw Hill Book Co. Inc. Prather, O. C. and R. C. Schafer, 1969. A travel reduction analog computer with automatic initial balance. Transactions of ASAE 12(3):336-337, 341. Reece, A. R., 1961. A three point linkage dynamometer. Journal of Agricultural Engineering Research, 6(1):45. Reynolds, W. R., G. E. miles and T. H. Garner, 1982. Microprocessor system for data acquisition and processing in the field. ASAE Paper No. 82-5510. Reznicek, R., A. Kment and L. Fojtik, 1957. Use of strain gages to determine the slip characteristics. National Institute of Agricultural Engineering, Tranlation 138. Rogers, 1. C. and G. M. Johnston, 1952. Measuring forces in the tractor linkage. Paper presented at a meeting of North Atlantic section of the ASAE at Oreno Me., August 1952. Scholtz, D. C., 1964. A three point linkage dynamometer for mounted implements. Journal of Agricultural Engineering Research, 1964 9(3):252. Scholtz, D. C., 1966. A three point linkage dynamometer for restrained linkages. Journal of Agricultural Engineering, 1966 ll(1):33-37. Shelton, D. P. and L. L. Bashford, 1977. Tractor instrumentation for practical demonstrations and data acquisition. ASAE Paper No. 77-1523. Shigley, J. E., 1977. Mechanical Engineering Design. McGraw Hill Book Co. Inc. (3rd Edition). Skalweit, H., 1964a. Field measurements on tractor with three point hitch mounting and controlling power lifters. Landtechnische Forschung l4(1):l-S. Skalweit, H., 1964b. Measurement of forces on the three point linkage in field tests on ploughs with control system operated through the tractor hydraulic lift. National Institute of Agricultural Engineering, Translation 272. Stanley, W. D., G. R. Dougherty, and R. Dougherty, 1979. Digital Signal Processing, (2nd Edition). Reston Publishing Co. Inc., Reston, Virginia. Stephens, L. E., A. D. Spencer, V. G. Floyd, and W. M. Brixius, 1981. Energy requirements for tillage and planting. ASAE Paper No. 81-1512. Summers, J. D. and J. C. Frisby, 1980. An economical on-board digital monitor for tractor-operation variables. Transactions of ASAE 23(1):57-59, 64. Thiel, R., 1958. The measurement by electrical means of the tractive effort at three-point attachments to agricultural tractors. Landtechnische Forschung. 8(5):132-134. Timoshenko, 1940. Elements of Strength of Materials. Tompkins, F. D. and L. D. Wilhelm, 1981. Instrumentation for measuring energy inputs to implements. ASAE‘ Paper No. 81-1575. Upadhyaya, S. K., L. J. Kemble and N. E. Collins, 1983. Draft prediction of mounted implements from strain gage readings. ASAE Paper No. 83-1037. Van Nostrand Reinhold Co., 1976. Encyclopedia of Computer Science. Van Nostrand Reinhold Company, New York. Volkov, B. G. and L. A. Klochev, 1958. Dynamometers for recording the drought resistance of mounted implements. National Institute of Agricultural Engineering, Tlanslation No. 69. Wendte, K. W. and H. Rozebom, 1981. Data acquisition for tillage energy evaluation. ASAE Paper No. 81-1045. Wilhelm, L. R., F. D. Tompkins and M. W. Cantrell, 1981. Software for a tractor-mounted data acquisition system. ASAE Paper No. 81-1576. Zoerb, G. C. and J. Popoff, 1967. Direct indication of tractor wheel slip. Canadian Agricultural Engineering, 9(2):91-92, 108. Zoerb, G. C. and R. B. Howse 1976. An integral drawbar transducer. ASAE Paper No. NCR 76-101. A P E N D I X LISTINGS OF THE DATA ACQUISITION PROGRAMS. INCLUDED ARE THE TWO COMPUTER PROGRAMS USED. THE FIRST ONE IS FOR FIVE CHANNEL (SIMULTANEOUS SAMPLING) AND THE SECOND ONE IS THE MODIFIED VERSION USED TO SAMPLE ONE CHANNEL AT ELEVATED SAMPLING SPEEDS WHICH THE THE FIRST ONE COULD NOT HANDLE. SEE SECTION 4.5 OF THE TEXT. MULTICHANNEL SAMPLING DATA COLLECTION CONROL AND ANALYSIS PROGRAM DISP "1 HM sunnINS FORCES" PRINT "H0912 FORCE“,Tfiaq153- "‘UIERT. FORCE" 15 CLEHF £8 OPTION 683E 1 38 CON J 71.53. u,mz NE 48 REHL ‘f? 5. 58 9 EN HBLE LEO 1+é+32+128 68 SHOFT 8 554.-.B c554 C(5543.D '554) E' 554 “(554 0(554: ?8 DIN 93(32] 88 OISP "ENTER TEST 88TH e s n; BOHFO F8351 JUN6 98 INPUT 8: 188 RS: READY TO TRKE 08TH" 118 k 8 S N 8 138 FEM THE F'S REE SPLISRRTION FHCTORS 1‘ F 11:88? 8 Fv23=848 Q Fri39=8 1: 8 Ff4h=632 8 Ff5“=739 148 SOTO 3588 158 ON KEYfi 1."ORTR" SOTO 488 168 ON KEY# 2.“5TOP“ SOTO 248 1?8 ON KEYh 3-"PLOT" SOTO 3?8 188 ON KEYS 4."3TORE" SOTO 2348 198 UN RE?! 5-" ZIS “ SOTO 2588 £88 OFF KEYh 6 £18 OFF KETfi ? 228 OFF Kth 8 338 SOTO 358 248 OUTPUT F89 5"T88" 258 N=N~1 368 CLEHE 2?8 288 298 PRINT 388 FOR I=1 TO N 318 F IE=HQIfi+CCIF+ELI3 328 8£15=BIID+OCIb 338 PRINT PklfiiTHBIIEEIOQI‘ 348 NEHT I 358 CLEHR 368 BEEP 3?u KEf LHBEL 338 013? 399 OISP "oars COLLECTED HND 3T0 REE 2v PRESSINS USER DEFINED K v." 488 leP 418 815 3? "KETO I-STHRT OHTH COLL E117 DH" 42 OISP "rETfi 2 STOP 89TH INPUT 438 816? "uE-h 3 FLQIT LATEST DR 75 EPMFLEO 448 CZ:P 'sErfi 4 SToRE 39TH" 458 OISF 468 OISP H: 4?8 SOTO 478 488 ' ........................ 298 ' Oa's sampling Subroutine I 830 848 858 868 378 388 399 988 950 CLEHF OISP USING "3x63.288" ; “OHT H BEINS COLLECTED" kEY LHBEL 83:"REROY TO PLOT 8 STORE on T8“ SLERR T89 OUTPUT T89 ,"T02888E18F18LSU ‘ I OUTPUT ?09 5"USIUF1UNSUOSURI N=8 WRIT 1888 N=N+1 OUTPUT T89 s"RCIJT2$Ol“ OUTPUT 789 ;"V9“ ENTER T89 5 8(NJ-BcN),C(N),O 2NJsELN! HxN)=(R(N)-Z(1))XF<1) 8(N3=£EKNJ'Z(2)E$F(2} C(N)=EC(N)-Z(3)}¥F(3) URN)=(U§N)'2(4)}XF(4J ExN)=(E(NJ-Z(S)JXF(SJ IF N>399 THEN 248 SOTO 688 I ! Samellne Explained I T0288-TTL Pulse output ewe r7 .2 seconds | HEl-External channel incre ament ' HFI-Let f1rst channel be I ! HLS-Let last cnannel be 5 ! UR3-Oata signals are in 18 volt range ‘ ! ”Fl-Format is 83011 I HHS-Five readlnes Per trie eer ! HOS-O o? dieits N01e=HORE DIGITS IMPLIES LONER CONVERS ION SPEED ! URI-flute Zero UR 1s"DN" No teIREHDING RRTE IS HHLVED BU T RSCURECY IMPROVED” 9 USl-Storaee in 83011 ! ROI-Close channel 1-(Start sampllne all channels) ! VTZ-Triaeer externally !0801-System wazt for output .. ' H PRINT E PRINT PRINT 83 PRINT GRRPH PEN 1 E GCLERR PLOTTER IS 1 §0 LOCRTE 25,133,2a.98 9 9(8 980 FAD SCHLE BIN+‘I-?'- 0,0 990 RE" XXLRBEL ENG 098“ RXESXXX 1000 1010 1020 1030 1040 1050 1060 1070 1030 0‘0 0W3 (SQ HFH‘PhH‘HFHHHhH‘ NHHF‘HHHF‘HH Q‘DOD‘JOIU-bbthH SGHDSGMSQGNSG LHXES -(N/10),S.O,-?,2;1 FOR J=1 TO P HOUE 8,8 k=8 E T1=O 3 32:3 FOR I=1 TO N SLEHR SRRPH k=K+1 ON J SOTO 1093~11’O.1238,13 86,13?6,1440,1518 LINETYPE 1 H=H(1) HQI)=R«I)/IBO PLOT K,H(I) HKI)=R(IJXIOO SOSUB 2928 SOTO 15?B LINETYPE S N=B(I) BaI)=B(I)xIOO PLOT K,BtI) BXIOB SOSUB 2929 SOTO ISFB LINETYPE 6 N=S(I‘1 SLI)=S(I)XIOB PLOT K,C(I) CKI)=CQI)*IBB SOSUB 2926 SOTO 1576 LINETYPE 7 N=D(I) OaI)=OkI)z188 PLOT K,D(I) o<15=0c13x1aa SOSUB 2920 SOTO 15?8 LINETYPE 3 N=EfIJ EKI)=E(I)/IOB PLOT K,E(I) E=E HO‘HF‘PHHHPHH HHHHprOHroHv-a 0". 0.000500010111000! W J-NNHCHDOJ‘JO‘UI 0 0000000000 00000303000 ‘xl‘l‘u‘l‘l’d'd’slfl‘l *10‘0145 (ANN-‘0 WW‘IOIUIJHDOJNN-‘Q 00000000 0000000000 Hthfidepfi-s 80 1890 1900 1910 1920 1930 1940 1950 1960 1970 1930 1990 2000 2010 2020 2030 2040 2050 2060 20?0 2030 SKIP=O SOSUB NEXT I RLPHR T2=T1A2 S=&(SZ-T2/N:x(N-1))A.5 V1=T1fh S3=SXIBa/vl SOSUB 3810 ON J SOTO 1650,1678,1698,1? 10,1?30,1750;17?O P=1 SOTO 1?83 1730 P=2 1780 1?39 178a 1?36 R L=1 TO 5 ON L SOTO 1886,1828,1840,18 56,1886 H(P,L)=Ul SOTO 1899 HaP,L)=H1 SOTO 1896 HéPTL>=M2 SOTO 1898 H(P,L)=S SOTO 189B HtP,L)=03 NEXT L SRHPH REM UI=RUERHSE O M1=HINIHUM e M2=HRXIHUH REM S=STO OEUIHTION E 83:00 \1)1100 2920 EFF. OF VHRIRTION NEKT J MOVE N22 5:-12 LDIR 0 LHBEL "DHTR NUMBER” HOUE -(Nf8);-? LDIR 90 LHBEL “DRRFT FORCE 100LB“ COPY REM 1* END OF PLOT XX HLPHH FOR U=1 TD ? IF U=6 THEN GOTO 2030 IF U=7 THEN GOTD 2100 PRINT USING 2230 J U Q GOTO 2120 2240 E PRINT US PRINT USING INC 2260 2 PRINT USING 2260 ION H0 0‘1 00 [U to w mmm mm 0» o-b H HHHHH ‘0 CC. “I 0'. UN: um.— a 0 c: 00006 [‘0 H 0 In N N to w h? [u lo OHUHG| 0000 N [u Iu [u N to 01 U! «I‘— 0 0 0 [U ‘5 o 0 GOTO 2130 PRINT USING 2250 E PRINT US ING 2260 0 PRINT USING 2260 GOTO 2130 PRINT USING 2260 PRINT FOP U=1 TO 5 IF ”=1 THEN PRINT USING 227 0 5 H(U)\’) IF ”=2 THEN PRINT USING 228 0 : H R=0(I) DII)=0(H) 0(H)=R GOTO 3410 IF E(I)<=E(N) II) EII)=E(M) (H)=R GOTO 3410 m m m IF PRI)<=P(M) R=P(I) Pg1>=P IF Q(I){=Q(H) E=0€I‘ Q£I3=Q(H) Oin3=R NEXT H THEN THEN THEN THEN THEN THEN 3410 3410 3410 3410 3410 3410 ON J GOTO 3430,3460,3490,35 20’3550’3580;3610 "1:0(1) HZ=H£N> SOTO 3630 N1=B(1) N2=B(N) SOTO 3630 n1=c<1> n2=C (TOIOJ-HUNORED SRH PLES R SECOND BEEP INPUT T DISP USING “5f6XI200" : "DRT R BEING COLLECTED" OUTPUT 709 USING "RRPDDDH “TO“,T OUTPUT F09 I"REIUR3" OUTPUT F09 ;"U31vF1UNIVOSVRl N=0 WRIT 1000 N=N+1 gUTPUT F09 USINS "BRIO" 5 "8 .“IO OUTPUT F09 5"UT28O1" OUTPUT 709 ;"US" ENTER 709 5 9(N) RfiN)=(R(N)-Z)*F ! B(N)=(B(N)-Z(2))*F(2) ! C(N)=(C(N)-Z(3J)XF(3) ! 0(N)=(D(N)-Z(4))*F(4) ! EIN)=(E(N)-Z(5))*F(5) IF N>4000 THEN 230 GOTO 590 I I ! 70200-TTL'Pulse outeut eve r? .2 seconds i HEl-External cnannel incre ament ! RFl-Let first channel be 1 I HLS-Let last channel be 5 ! UR3-Data signals are in 10 volt range ! UPI-Format is HSCII ! UNI-One readinss Per tries e“ ! JDS’N of disits Note=HORE OISITS IMPLIES LOHER CONVERS ION SPEED ! URI-Huto Zero UH is"ON“ No te=REROINS RRTE IS HRLUEO BU T HSOURHCY IHPROUEO" I USl-Storase in 03011 I ROI-Close channel 1-(Start samplins all channels) ! UTZ-Trxsser externally GI p—r-Hr-Hh-Hr—lnwkfltflwliv'LIlD‘DWUJ 014030;- 06'006'06'0'00'40’10$0le0‘00‘10. "JU‘IUI-F OJNH0G'0G'006'013'6'6'06'6013'0' t¢HHHHHHHHr~HHHFHH “-J’dmm-fimmm-bHHr-‘HHGIG' H001!—NNHSI'I'GINHHGIKDCO 000000000000 0000000000000000 00000000 NFL'HG'U'NVOIUIJ-OJPO HHHHr—HHHO—HHH (OCOCOCO‘J "-1'~J‘\I'~l"‘-l ’4 "-1 III PRINT 0 PRINT PRINT 83 PRINT SRHPH PEN 1 2 SSLERR PLOTTER IS 1 LOCRTE 251130120190 SCHLE 0,N+1,-7130 PRO 0.0 REM *XLRBEL RND ORHH HHESXK$ LHNES -uN/10),5,0,-7;2Il ! FOR J=1 TO F MOVE 0,0 K=0 S T1=0 O 82=0 FOR 1=1 TO N SLERR SRRPH K=K+1 I ON J SOTO 1080,1140,12009 12601132011380,1440 LINETYPE 1 N=HCI§ HII)=H(I)/100 PLOT Klfiil) Hi1)=H=M2 SOTO 1810 HiL)=8 SOTO 1810 HCLJ=S3 NEXT L SRRPH REM V1=RVERHSE M1=MININUN 2 M2=NRXIMUM REM 8=8TO OEUIHTION S CB=SO EFF. OF VHRIRTION MOVE N/2.SI-12 LOIR 0 LHBEL "08TH NUMBER" .5 l! MOVE -KN»8J,-7 LOIR 90 LNBEL "ORPFT FORCE 100LB“ COPY ! REN xx END OF PLOT tx RLPHR PRINT USING 2150 3 S PRINT USING 2180 FOR V=1 TO 5 IF ”=1 THEN PRINT USING 219 0 ; H(V) IF U: 2 THEN PRINT USING 220 0 j H(U) IF U: 3 THEN PRINT USINS 221 0 . HéV) .7 8:4VTHEN PRINT USING 222 t‘ J 'I 13 IF 0:5 THEN PRINT USING 223 0 3 HQU) NEXT U PRINT 0 PRINT INRSE "STRTISTISHL DRTR FOR OHHNNEL=1NIOO IMRSE "F'FF .............. INRSE "RUERNSE=“-1XIODOODDO .O INRSE “HININUN:”,1X,OODDDDD .O INRSE "HHXIMUM=“,1XIODODDDO .O INNSE "STD OEUN=",1XIDODODO OO0.000 IMRSE "COEFF. OF URRTN (K): “IIXJODOOOO.OD OISP USINS "8/6X9280“ "NI SH TO STORE ORTRP-Ker #4" 5010 453 I Storing Subroutine R$=“REHDY FOR NEXT IMPLEHEN TI. OLERR OISP “ENTER FILE HOME FOR P RESENT STORHSE" INPUT 0: KEY LRBEL SRERTE OS-II9XN RSSISN8 1 TO U: PRINT! 1 5 N FOR I=1 TO N H(I)=R(I)1100 .‘ 0-4 u—a v—a r-a v-4 t-d VIC“ V4.6 .V..'- II II II II II II 001710er al'x Ir- .I'x Ir-I 4r- s- 2450 2460 24? 0 2480 2490 2500 2510 2520 2530 2540 2550 30 IO [‘0 N N 5.] "sj K] a] To «I "J mm A 0 -uM I‘uI‘O uaI NH 115 [U '1 ‘4 '4 “-1 -uw «I'd m U! h (51 NM PRINT# 1 5 0(1) NEXT I RSSISNQ 1 TO x SOTO 340 ENO ! Zero and CaIIDration sub LLERF’ 709 OUTPUT 709 ze scanner. Instr CLERR ! FOR 1=1 TO 5 2:0 NEWT I 018 -P "PRESS HO“! TO SRNPLE SHL. BEEP PRUSE FOR I=1 TO 5 FOR J=1 TO 10 : OUTPUT 709 USING JR!“ ’1 . ENTER 709 5 C ' C(I)=C(I)+C ! NEXT J I C(I)=C(I)/10 I 5"SI" I In1tia11 See HaInIrame aSONT) NHEN R RERDING ! 9 E '5' i "Ham" 5 NEXT I . CLERR DISP "PRESS BY TO SRNPLE BEEP PRUSE FOR J=1 TO 10 ON C GOTO 277052772 ' 7652778 USING "9850“ 5 " {CO 2E: 3.11 CI: OUTPUT 709 01' 1 GOTO 2780 OUTPUT 709 USING "BRIO" 5 " HI"52 GOTO 2780 OUTPUT 709 USING "9850" 5 " F11 II GOTO 2780 OUTPUT 709 RI"54 GOTO 2780 OUTPUT 709 HI"55 ENTER 709 2=2+Zl NEXT J Z=2/10 SOTO 140 I USING "0050" 5 " USING "8050" 5 " 21 I 0 III U! 45‘ 2860 2370 2800 2890 2900 2910 2920 2930 2940 2950 2960 2970 2990 3000' 3010 3020 3330 3350 3360 3550 3560 ! Sum 2 Sum of Square; Subr outxne Minfflax Subroutlne FOR I=1 TO N-l FOR H=I+1 TO N IF Ril)€=fi(fl) R=R€IJ RKI)=R(M) fi