mi 553% OF SEMPLEFEED EQUJPMENT FOR THE RAPE?) DEERMZRATION OF THE MGiST'EJRE CGNTENT Ct? GRAN AND FGRAGE CROPS Thai: (‘6: H10 Dem: a! Pit. D. MECWGAN STATE COLLEGE fiaraid Wéiiiam isaacs W54 *urcfi This is to certify that the thesis entitled "The DeSign of Simplified Equipment for the Rapid Determination of the Moisture Content of Grain and Forage Cro " presented by Gerald.W. Isaacs has been accepted towards fulfilllnent of the requirements for PhoDo degree in AgI‘iCUltural Engineering 445» 1’" 1%% Major professor Date May 171 1951-1 0-169 ll ‘- .-_.u,_‘i_____——..._.. ‘ _.i‘ 1;" THE DESIGN OF SIMPLIEIED EQUIPMENT FOR THE RAPID DETERMINATION OF THE MOI TU; CONTENT OF GRAIN AND FOhAGE CROPS By Gerald William Isaacs M A THESIS Submitted to the School of Graduate Studies of Michigan State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Engineering Year l95h GERALD w. ISAAcs , ABSTRACT Moisture content is commonly used as an index of the keeping quality of farm grain and forage creps. Farmers need a means of determining the moisture content of such crops in order to obtain the fullest benefit from the use of modern equipment and procedures for harvesting and curing farm crOps. Equipment for determining the moisture content of farm crOps on farms should be accurate, simple to operate, low in cost, and applicable to all the common hay and forage crops within a wide range of moisture contents. Moisture content varies widely within individual lots of farm crops. This is particularly true of hay which is curing in the field. This non-uniformity in moisture content makes it important to obtain a representative sample. Detailed studies were made of two methods of measuring moisture content, a direct oven method and an electrical re- sistance method. Exhaust Oven Method. An oven which utilizes the heat from the exhaust gases of an internal combustion engine to dry samples of hay or grain was designed and tested. This oven mixes ambient air with the exhaust gases, which lowers the tempera- ture of the gases and prevents burning of the sample. Samples of hay weighing 100 to 200 grams are dried in about seven min- utes. Samples of grain are dried in about fifteen minutes. Electrical Resistance Method.’ Studies were made of the effects of electrode pressure, electrode design and frequency of the Arnm-__~.——-—~, GERALD W. ISAACS ABSTRACT measuring current upon the electrical resistance measurement of hay and grain. Tests of resistance-type moisture meters indicated that a single reading taken on hay or grain is not a good index of the moisture content of the whole lot. When testing hay with a resistance-type meter, at least 25 readings should be taken and the results averaged in order to obtain a reliable indica- tion of the moisture content of the whole field. Recognizing that most farmers would not carry out such a laborious computation, an automatic averaging device was de- veloped which gives a single indication of the average mois- ture content of a large number of samples. This device meas- ures the average electrical conductivity of the samples and thereby indicates a close approximation of their average mois- ture content. Other Methods Studied. Preliminary investigations were made on several other methods of moisture determination: a method involving the measurement of the pressure of the acetylene produced in the reaction between calcium carbide and the water in the sample, a specific gravity method, an oil-distillation method, a method involving the measurement of the shearing force required to cut hay samples, and a method using wet and dry-bulb thermometers to measure the equilibrium relative hu- midity of grain. None of these methods were found to be as applicable to the problem as the resistance and oven methods. -- w...__' I.“'~w— ACIC‘IOWLHDGEHENTS The author wishes to acknowledge gratefully the inspiring guidance and many helpful suggestions contributed by Professor D. B. Wiant. under whose direction this research was undertaken. He also wishes to thank the other members of the guidance committee. Doctors 3. J. Benne. S. T. Dexter. E. A. Hiedemann. and K. T. Payne. Professors I. B. Baccus and D. J. Renwick. and Dean 0. R. Megee. for their suggestions and guidance during the investigation. Special recog- nition is due Doctors Benne and Dexter who drew upon their long experi- ence in.moisture determination for much.welcome advice to the author. The writer also appreciates the efforts of Professor A. W. Farrell. Head of the Agricultural Engineering Department, in obtaining funds for this investigation through a research contract with the 0.5. Department of Agriculture. Special recognition is also due Doctor T. E. Hienton. administrator of this contract for the U.S. Department of Agriculture, for his continuing interest in the activities of this project. The author also recognizes the contributions of the following: Messrs. W. J. Delmhorst and R. S: Hart. manufacturers of moisture meters. for making their meters available for study; Mr. R. B. Hopkins. for his aid in coordinating the hay studies of this project with those of the J. I. Case—Michigan State College Cooperative Research Project; and Mr. J. B. Cawood. mechanical technician. for his help in constructing eduipment. The author is especially grateful to his wife. Phyllis, who typed this manuscript and gave much encouragement during the investigation. 338665 ii Gerald William Isaacs candidate for the degree of Doctor of Philos0phy Final examination: May 17, 1951+, 1:00 P.M.. Agricultural Engineering Building Dissertation: The Design of Simplified Equipment for the Rapid Detera mination of the Moisture Content of Grain and Forage Crops Outline of Studies Major Subject: Agricultural Engineering Minor Subject: Mechanical Engineering Biographical Items Born: September 3. 1927. Crawfordsville. Indiana Undergraduate Studies: Purdue University. 1943-b5. cont. 19h6-u7 BSEE. 1947 Graduate Studies: Purdue University. l9h7-49. MSEE. 1949; Michigan State College. 1952-1954 Experience: U.S. Navy. 1945-h6; Graduate Research Assistant. Purdue University. 19b7; Instructor and Research Assistant in Agricultural Engineer— ing, Purdue University. l9h8—52; Special Graduate Research Assistant. Michigan State College. 1952-54- Honorary Societies: Pi Mu.Dpsilon Society of Sigma Xi Professional Societies: American Society of Agricultural Engineers‘ man-as :J-ldJ-WLI 5313 0? Direct H Air 0v Water ‘ Veer: Ewing Eistil‘ Inc‘irect Elect: Dielec Ge: Pr. specif‘ 53mm heme "when Literatu m Pasxc. Forages Grains ‘ Pmperti at en uteretu 5111351}. AL I TABLE OF CONTENTS I NTRODUCTI ON 0 O O I O I O O I O O O O O O O O O O O O O O O REV I m OF LI TERATUR-E I O O O O O O O O O O O O O O O O O O 0 Direct Methods . . . . . . . . . . . Air Oven Method. . . . . . . . . . . . . . . . . . . . . Water Oven Method. . . . . . . . . . . . . . . . . . . . Vacuum Oven Method . . . . . . . . . . . . . . . . . . . Drying with Dessicants . . . . . . . . . . . . . . . . . Distillation Methods . . . . . . . . . . . . . . . . . . Indirect Methods . . . . . . . . . . . . . . . . . . . . . Electrical Resistance Method . . . . . . . . . . . . . . Dielectric Method. . . . . . . . . . . . . . . . . . . . Gas Pressure Method. . . . . . . . . . . . . . . . . . . Specific Gravity Methods . . . . . . . . . . . . . . . . Hygrometric Methods. . . . . . . . . . . . . . . . . . . Thermal Conductivity Method. . . . . . . . . . . . . . . Mechanical Methods . . . . . . . . . . . . . . . . . . . Nuclear Methods. . . . . . . . . . . . . . . . . . . . . Literature Cited 0 e e e e e e e e e e e e e e e e e e e, e THE PHYSICAL NATURE OF FORAGE AND GRAIN CROPS WHICH AFFECTS MOI STUBE DETERMINATI ON C O O O I O O I O O I O O 0 I O . O O Forages I O O O O O O O O O O C O O O O O O O O O C O I C . Grains e e e e e e e e e e e e e e e e e e e e e e e e e 9 Properties of Forages and Grains Which vary with Moisture Cont ent e e e e e e e e e 0 e e e e e e e e e e O O O 0 0 Literature Cited e e e e e e e e e e e e e e e e e e e e e PRINCIPAL DESIGN REQUIREMEHTS OF FARM MOISTURE METERS. . . . Accuracy . . . . . . . . Rapidity e e e e e e e e e e e e e e e e e e e e e e e e e Simm11C1ty Of Operation. e e e e e e e e e e e e e e e e e coat 0 e e e e e e e e e e e e e e e e e e e e e e e e e e Versatility. e e e e e e e e e e e e e e e e e e e e e e e INVESTIGATION OF EXHAUST OVEN METHODS. . . . . . . . . . . . Principle of Operation of the Exhaust Oven . . . . . . . . iii 0\ \O\O(D(D‘Q‘Q 10 10 20 22 24 25 28 30 30 33 38 38 #1 #2 43 to 1+5 46 a? a? TABLE OF CONTENTS Continued Apparatus e e e e "e e e e e e e e e e e e e 0 Test Procedure. . . . . . . . . . . . . . . . . . . . . . . Test Results. . . . . . . . . . . . . . . . . . . . . . . . . Field Tests . . . . . . . . . . . . . . . . . . . . . . . . Laboratory Tests. . . . . . . . . . . . . . . . . . . . . . .Residual Moisture Determinations. . . . . . . . . . . . . . Dry Matter Loss . . . . . . . . . . . . . . . . . . . . . . Deposit of Combustion Products. . . . . . . . .'. . . . . . Scale Studies 0 O O O I O I O O O O O O O O O O I O O O O 0 Moisture Content Calculator . . . . . . . . . . . . . . . . 00110111810118 e e e e e e e e e e e e e e e e e e o e e e e e 0 Literature Cited. . . . . . . . . . . . . . . . . . . . . . . INVESTIGATIONS OF THE ELECTRICAL RESISTANCE METHOD FOR HAY AND GRAIN 0 e 0 e e e e e 0 e e e e 0 0 e I e e e e e e e e o e e O Investigations of the Resistance Method for Bay . Resistance Measurements of Hay. . . . . . . . . . . . . . . Apparatus . . . . . . . . . . . . . . . . . . . . . . . . Procedure . . . . . . . . . . . . . . . . . . . . . . . . ResultSOOOOOO...OOOOOOOII~OOOOOOO Field Performance Tests of Two Commercial Moisture Meters forWOOOOOOOOOOOOOOOOOOOOOOOOO Apparatus.............oo.._......o Procedure . . . . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . Investigations of the Resistance Method for Grain . . . . . . Laboratory Investigations on Experimental Designs . . . . . Apparatus........................ Procedure . . . . . . . . . . . . . . . . . . . . . . . . lmsuts.’OOOOOOOOOOOOOOOOOIOO... Test of a Commercial Moisture Meter on Grain. . . . . . . . Procedure for Wheat Tests . . . . . . . . . . . . . . . . Procedure for Corn Tests . . . . . . . . . . . . . . . . ResultS......................... Literature Cited. . . . . . . . . . . . . . . . . . . . . . . iv 51 q .4’ 57 57 66 71 75 ' so 83 85 87 9o 96 98 99 108 108 108 109 111 117 117 118 119 125 . - lira-.90“. '4‘5'Jo q VA'J‘1‘1‘ FF: vu‘..._. ‘beu '§ 8: p ,Q ,- V: ‘1‘le LIN “rub TABLE OF CONTENTS Continued DEVELOPMENT OF AN AVERAGING MOISTURE METER Design of the Averaging Device Tests of the Averaging Device. . . . . . Field Tests. . . . . . . . . . . . . . Apparatus. . . . . . . . . . . . . . Procedure. . . . . . . . . . . . . . Results. . . . . . . . . . . . . . . Laboratory Tests . . . . . . . . . . . Apparatus. . . . . . . . . . . . . . Procedure. . . . . . . . . . . . . . Results. . . . . . . . . . . . . . Conclusions. . . . . . . . . . . . . . . Literature Cited . . . . . . . . . . . . pmmrmr INVESTIGATIONS or OTHER METHODS or MOISTURE MEASURE- MmT. O O O O O O O O O O O O O C O C O 0 Gas Pressure Method. . . Oil Distillation Method. . . . . . . . . Specific Gravity Method. . . . . . . . . Mechanical Shear Studies on Bay. . . . . Wet and Dry-bulb Thermometer Methods SUMMARY AND CONCLUSIONS. . . . . . . . . . Exhaust Oven Method. . . . . . . . . . . Electrical Resistance Method . . . . . . Wet and Dry-bulb Thermometer Method. . . APPENDIX I Data from Tests of Injector-type Exhaust APPENDIX.II Analysis of the Average Indicated by the oven 0 O O I O O O O O Averaging Device. . . 126 127 1&5 1M5 ms 146 146 151 151 151 153 165 16? 168 168 172 176 178 182 192 192 193 191; 196 199 I‘J "\ \(J 10 ' 'J kn I h ~;__C ‘Pl “‘d. ’n' v «1.5-: 7.31 Figure 10 11 12 13 lb 15 16 17 LIST OF FIGURES Diagram of the injector—type exhaust oven . . . . . . . . . The injector-type exhaust oven installed on a farm tractor. Basic equipment necessary to carry out exhaust oven determination of moisture . . . . . . . . . . . . . . . Apparatus for testing the exhaust oven in the laboratory. Ekhaust oven.performance curves with alfalfa at 51.8% moisture content. Engine speed 750 r.p.m.. . . . . . . Exhaust oven.performance curves with alfalfa at 51.8% moisture content. Engine speed 1250 r.p.m. . . . . . . Exhaust oven performance curves with alfalfa at 70.5% moisture content. Engine speed 1000 r.p.m. . . . . . . Exhaust oven performance curves with alfalfa-brome hay at 28.7%'moisture content. Engine speed 1000 r.p.m.. . Exhaust oven.performance curves with corn.samp1e at 16.5% moisture content. . . . . . . . . . . . . . . . . Exhaust oven performance curves with wheat sample at 14.7% moisture content. . . . . . . . . . . . . . . . . Color photograph of hay samples dried by exhaust oven , . Color photograph of grain samples dried by exhaust oven . A.mechanical nomograph for computing percentage wet basis moisture content. . . . . . . . . . . . . . . . . The Delmhorst Model F hay moisture meter and the Model 11E probe . . . . . . . . . . . . . . . . . . . . The hydraulic press used to apply pressure to hay samples contained in a b" x b" plastic box for resistance measurement . . . . . . . . . . . . . . . . . . . . . . Resistance-pressure charateristics of hay measured with Various types of electrodes . . . . . . . . . . . . . . Pinptype electrodes used in Test E of resistance measure— mentonhayftegooeeeeeeeeeoeeoeeee vi Page #9 52 52 54 59 61 62 . 65 69 69 79 86 86 88 89 F i gure l8 19 2O ‘0 ”O 23 $26 27’~ 2&3 3() 31. LIST OE FIGURES Continued Detailed resistance—pressure characteristics of an sample of alfalfa hay . . . . . . . . . . . . . . . . Arrangement of apparatus for determining the depth of penetration of current in measuring the resistance of hay with co-planar electrodes. . . . . . . . . . . Delmhorst Model RC—l moisture meter and Model 283 probe for hay and grains. . . . . . . . . . . . . . . . . . Hart Model CUQQ moisture meter for use on hay . . . . . Data from tests of Hart meter on alfalfarbrome hay. Average of five meter readings vs. oven-determined moisture content. . . . . . . . . . . . . . . . . . . Data from tests of Hart meter on alfalfa-brome hay. Individual meter readings plotted against oven- (xetifirfflinfid mOiSt‘LlI‘e contents. 0 o e e e e e e o e e e Tests of the Delmhorst meter on second-cutting alfalfa- I’rome Ian-5:; . . g e o o e o e e e e o e e e e o e e e 0 Comparison of the moisture contents of alfalfa samp1es taken simultaneously in pairs one foot apart in the' V31 Ila-row o e e e o e o e e e e e o o 0 O O O 0 O O O 0 Test cell for resistance tests on grain - - ° ° ° ° ° ' Constant-spacing electrodes for grain test cell . . . . \ Modification of the Delmhorst Model 28E probe for resis- tance tests on hay and grain. . . . . . . . . . . . . . Electrode pressure vs. electrical resistance of corn measured in the test cell with variaole- pacing Clectl‘OdGB. e o e e e e e e e e e O 0 O o 0 e I 0 O 0 Electrode pressure vs. electrical resistance of corn measured in the test cell with constant-spacing eleCtrOdesfl O O O O O O O O C I O O O C C O O O O O 0 Electrical resistance vs. moisture content at constant electrode pressures. . . . . . . . . . . . . . . . . . 100 102 103 107 109 109 110 112 113 114 viii LIST OF FIGURES Continued Figure Page Tests of Delmhorst meter on wheat samples drawn at country elevator (Test A). . . . . . . . . . . . . . . . . 120 \J) N :33 Test of the Delmhorst meter on recently-harvested corn samples (Test D). . . . . . . . . . . . . . . . . . . 123 31% Tests of the Delmhorst meter on ground corn samples (Group I and II of Test E) . . . . . . . . . . . . . . . . 125 :15 Schematic diagram of averaging device for use with resistance-type moisture meter . . . . . . . . . . . . . . 129 :36 The experimental averaging device for windrowed hay in use in the field with the Tag-Eeppenstall meter. . . . . . 131 37' Drawing of the averaging resistance-type moisture meter for hay. . . . . . . . . . . . . . . . . . . . . . . 13“ \ .1) 0') Detail sections of the averaging resistance-type moisture meter for hay . . . . . . . . . . . . . . . . . . 135 35? Simple series ohmmeter circuit . . . . . . . . . . . . . . . 137 “(3 Relaxation oscillator circu't. . . . . . . . . . . . . . . . 137 I. o n]. An operating cycle of a relaxation osc1llator circuit connected to a constant resistance . . . . . . . . . . . . 140 ' o O O “d2 An Operating cycle of a relaxation osc1lletor circuit used with an averaging device. . . . . . . . . . . . . . . le bf3 Field date from the :veraging resistance—type moisture meter for hay. . . . . . . . . . . . . . . . . . . . . . . 1L7 ..ta from averaging resistance-type moisture meter or hrs. Logarithm of resista.ce vs. the average moisture t of three oven samples. . . . . . . . . . . . . . . 150 “E5 Arrangement of ecuipment for the latorntory tests of the averaging device . . . . . . . . . . . . . . . . . . . . . 152 146 Components of the relaxation oscillator circuit of the Shafcr moisture meter. . . . . . . . . . . . . . . . . . . l); 47 Test cell for trials of the gas pressure method on grains. . 169 ix LIST OF FIGURES Continued Figure Page #8 Rate of pressure rise in a receptacle containing ground corn and calcium carbide or calcium hydride. . . . . . . . 171 49 Apparatus for dehydration tests in oil at below—atmos- pheric pressures . . . . . . . . . . . . . . . . . . . . . 173 50 ste of weight loss of corn and alfalfa samples heated in corn 011 at 325.Foo o o o O o o o o o o o o I 0 O O O 0 175 51 Weight in water vs. moisture content of 25 gram samples Of corn. 0 O O C O O O O O I O I O O O O O O O O O O O O O 179 52 The Tenderometer used in the mechanical shear studies onmyoooooooooooooooooooooooooo180 53 The Master Moisture Meter, 8 commercial moisture meter operating on the wet and dry-bulb principle. . . . . . . . 183 5h Experimental wet and dry-bulb thermometer apparatus for testing ground grain . . . . . . . . . . . . . . . . . . . 188 :0 Q.. 'u m. “a Ki :1 VQ’;~:: "bod . s O ' 5 -u»\. " u -'~)c.‘ .9 45“.. U,":_H, " Gran-an... F a N-‘ d- N \~. ~ “~, 'K.l 7'- ‘A U y 'L h ‘ v‘ .1 L. .‘ 9 ’. ~k'~..'. :‘ ‘1 7. u" “(HM Table II III IV VI VII VIII II XI XII XIII XIV XV XVI XVII LIST OF TABLES Methods used by 52 farmers to determine the moisture content of hay. . . . . . . . . . . . . . . . . . . . . Summary of moisture residual data for hay and grain . . . Results of protein and sugar analyses of alfalfa Samples. Maximum effects of a scale error of 1.00 g. on the accuracy of moisture determinations . . . . . . . . . . Effects of current penetration on the measurement of the electrical resistance of hay with co—planar electrodes. Summary of data from commercial moisture meters tested on has, 0 O O I C O O O I O O O O O O O O O O O I O O O . Repetitive resistance data from corn taken with the Delmhorst Model 28E probe before and after modification Summary of the results of investigations of the Delmhorst meter in testing grain. . . . . . . . . . . . Location and values of test resistances (Test A). . . . . Summary of date from Test A of the averaging device . . . Location and values of test resistances in Test B. Odd- numbered resistances were disconnected for Test C . . . Summary of data for Test B and C of the experimental averaging device. . . . . . . . . . . . . . . . . . . . Location and values of test resistances in Test D and EOftlleaveragingdBViceooooooooooooooo Summany of data for Tests D‘and E of the averaging device Data from Tenderometer tests on alfalfa stems . . . . . . Lata from Baldwin moisture tester . . . . . . . . . . . . Data from tests of injector—type tractor exhaust oven . . 70 71+ 95 107 115 121 154 156 159 160 163 16k 181 185 196 .cci, 40.... 1:1 up“ . 1 V0“. .O-a. 'o-A .r. . L. n,,’ 5&5“ .Fu‘ 'I“ v‘.‘ ’a u ‘U '7) 1 . INTRODUCTION Moisture content has long been an index of the keeping quality of grain and forage crops. Prevention of spoilage has been a great problem to mankind since he first learned to store for his future needs what he could not eat immediately. In spite of reported surpluses in certain farm commodities, the modern American farmer can no better afford waste through spoilage than his forefathers. Not only do wasted products mean wasted money to the individual concerned, but waste by spoilage means useless depletion of our nation's greatest physical asset, its natural resources. Changes in farm harvesting and storage procedures have brought about changes in the need for moisture measuring equipment. The need for equipment for the quantitative measurement of moisture or storage quality in farm crops on the farm parallels the development of the combine and the modern crop drying and conditioning equipment. . Smell grains used to be cut with a binder and placed in the shock to dry before threshing. There was no great need to rush the threshing Procedure, since the grain was reasonably safe from weather damsgp in the shock. There was little need for a quantitative measurement of moisture content, since the grain could be dryed beyond the critical Safe storage moisture content. Today's combine harvesting methods require that the harvesting be done during a certain rather short period in the maturity of the grain. Grain which is combined too wet will obviously spoil in the granary. 1'.- 1 av x. I .302! e _ AI ,. it h “‘2 e .4» 1 H h,,,_f ,‘e ‘C I“ Combining grain when too dry results in greater losses of grain at the cutter bar due to shattering. Furthermore, leaving the grain in the field longer than necessary increases the risk of loss through lodging and shattering due to heavy rains, wind or hail. Some farmers prefer to harvest their grain before it has field cured to the extent that it would keep safely in an unventilated storage, thereby avoiding the risks of leaving it in the field at a time when weather losses could be great. Final curing is accomplished by arti- ficial drying. Being able to determine the moisture content of the grain is a valuable asset in the management of a crap drier; but whether or not artificial drying is employed, a means of measuring the moisture content of small grains on the farm is needed to indicate the Optimum time for harvest. Like the combine, the corn picker has greatly alleviated the labor problem in corn harvesting, but under some conditions has created new problems in grain storage. An increasing number of farmers today are following the practice of harvesting corn before it is dry enough.for, safe storage, then artificially drying it to an allowable moisture content. By picking the corn at high moisture content, the pidting efficiency of the picker is improved and the risk of loss through bad weather is reduced. In some years, particularly in the northerly States such as Michigan, weather conditions are such that the corn does not dry properly in the field and artificial drying becomes necessary to obtain a high grade product. Where artificial drying of grain is practiced, some convenient means of determining the moisture content of the product is needed to t v I... u . e p. b. up. b. "v .. a u 5 s w u n ”I! P ~ 0: 9‘ .3.» er“ I..‘ D u - u \' pa I. u. allow the most efficient management of the drying equipment. On many modern farms now practicing artificial drying of grains, moisture tests are obtained by taking samples to the local grain elevator for testing. For those farmers who are not located very close to an elevator, this practice is both inconvenient and time-consuming. Furthermore, certain types of moisture meters used in commercial elevators are not accurate when testing grain recently dried in a heated air dryer. New practices in the harvest, curing, and storage of forage crops have also created an increasing need for a means of moisture measurement. Bay has been cured for centuries by leaving it in the field until it was adjudged dry enough to be stored without molding or excessive heat- ing. However, farmers have long been aware of the importance of the losses of nutrients incurred by over-curing and the dangers of spoilage and heating caused by under-curing. The need for care on the part of the farmer in curing hay was indicated by Columella, e.Roman.who wrote as follows about 50.1.D. (1): “There is a measure to be observed in drying hay, that it be put together neither oven dry nor yet too green: for, in the first case, it is not a whit better than straw for it has lost its Juice: and in the other, it rots in the loft if it retains too much of it: and after it is grown hot it breeds fire, and sets all in a flame. They do not put it up in mews, before that they suffer it to heat, and concoct itself, and then grow cool, after having thrown it loosely together for a few days“. Modern agricultural science has deve10ped new curing and storage procedures which greatly retains the nutritional value of forage crops and reduce the risk of loss due to rain. These modern practices can be carried out most efficiently when a quantitative estimate of the mois— ‘ture content of the crop is available to indicate when to rake, when to r..- ‘ Av“ 0 “Ila. V‘s-u He“ start baling or chopping, and when to start or stop the crop dryer. Most experienced farmers have to a great extent developed a sense of judgment for determining when hay has been field cured sufficiently to keep safely in an unventilated storage. However, modern haying practices require that the farmer be able to recognize stages of curing which were not previously of specific interest to him. For example, he may now wish to rake his hay at 50% moisture content instead of 25% and he may wish to bale or chop it at 35% instead of 20$ moisture content. Further- more, if he is to realize the maximum benefit from modern haymaking practices, he should be able to determine moisture contents more accue rately than when the hay was left in the field until 1t “.1. "good and dry“. Host farmers now using barn dryers for mow-curing hay use rather unquantitative methods of indicating when to place the hay on the dryer. A recent survey of the experience of farmers who use barn dryers was made by Vary (2). The answers that 52 farmers gave when asked what method they used to determine moisture content are given in Table I. TABLE I METHODS USED BY 52 names TO DETERMINE THE MOISTURE CONTENT or HAY (2) Method ‘Number Method Number Reporting Reporting None 10 Break 1 Guess 6 Rub 2 Feel 15 Oven l Twist ' 6 Leaves curl ’ 1 Battle 1 Experience 7 Squeeze 2 fl i .1 n i. vl~a~ s- V‘- on ‘ ah. . \ .Illl'r The fact that farmers resort to such indefinite methods as those indicated in Table I reflects the unavailability of simple, inexpensive, and reliable means for quantitatively determining the moisture content of forage crops. As the acceptance of modern forage harvesting and curing practices advances, the demand for moisture measuring equipment will likewise increase. Literature Cited (1) Ahlgren, G. H., “Forage Craps“, ed.1, Mc-Graw-Hill, New York, 1:3, 1949. (2) Vary, Karl A., ”Statistical Summary of Farmers Experience with Barn Driers'.e Unpublished mimeo. ll pp. Agricultural Economics Department, Michigan State College. February 2, 195“. REVIEW OF LITERATURE The determination of moisture was probably one of man's earliest acts of chemical analysis. Procedures for moisture determination may be found among the oldest of scientific papers. The following review of literature is not intended to include all significant work in the subject of moisture determination. but will include some of the papers which apply most directly to the experimental work to be discussed subsequently. The author has compiled a bibliograpkw of rapid methods of moisture determination (with abstracts) which should be of use to future investigators of this subject (1). Although this work contains 450 references. it does not necessarily include all of the research on rapid moisture determination of the last sixty years. All methods of moisture determination may be classified generally as either direct or indirect. The direct methods include the oven melihods. the distillation methods, and the methods involving drying with deSSicants. They are called direct methods because they require the fieParation of the moisture from the dry matter and the amount of mois- ture removed is either measured directly or calculated from the loss in "Oight of the sample. The indirect methods involve the measurement of 8(”no property of the material which is dependent upon moisture content. IIldirect methods require calibration with one of the direct methods. 1.." '- oh: a... ‘ l'.. --~ ~ Direct Methods Actually all methods of moisture determination are somewhat empiri- cal or indirect, as indicated by Zelew (2) and other investigators. Even the so-called direct methods cannot measure with great accuracy the actual amount of water in a sample. First of all. one cannot be certain that all of the water was removed from the sample. In organic materials. some of the water is present in forms which appear to be more or less tightly "bound“ by strong physical forces to proteins. high molecular weight carbohydrates. and other colloidal materials. Further- more. when testing organic materials with the oven or distillation methods. heat must be applied which may lead to decomposition of dry matter con- stituents and the production of water which was not originally present as such. Volatile constituents of the dry matter may also be driven Off and result in errors in the moisture measurement. ' It is not the purpose of this studyto discuss at great length the 8hortcomings of the direct methods of moisture determination. Of greater valias is to point out the importance of indicating the procedure by which moisture determinations were made when presenting moisture data so that the results therefrom. may be reproduced with reasonable accu- racy. Wm There are several air oven procedures for the moisture determination of various materials. Standard procedures have been drawn W3 by organizations interested in each material. In the case of grain the U.S. Department of Agriculture has set forth air oven procedures in the Official Grain Standards of the United States (3). Basically. this method requires heating the finely ground grain in an air oven for one hour at 130%. Another air oven procedure has been set forth by the Association of Official Agricultural Chemists (it). This method re- quires heating a finely ground portion of the sample for two hours at 135° C. The latter procedure usually gives higher moisture content. as would be expected from heating the sample at a higher temperature for a longer period of time. There have been a great number of devices developed which heat Samples in air to dry them rapidly for purposes of moisture determination. These devices generally use higher temperatures in order to reduce the time required by the official methods. Heating is usually produced by electric heating coils or infrared lamps. Brabender (5) developed an oven which has a built-in scale for accurately weighing the samples While still in the oven. This removes the necessity of cooling the samples in a dessicator before weighing. which is required in the official method. W The water oven method is recommended for corn and requires heating the whole grain at 99'to 100.6 for 96 hours in a water Oven (3). A water oven is essentially similar to an air oven except it 13 heated by circulating water which is boiling at atmospheric pressure through a water Jacket in the walls of the oven. The temperatures at Various points within a water oven are usually more uniform than in an E‘~1 r oven. W The vacuum oven method for grain as set forth by the Association of Official Agricultural Chemists. (’4) requires heating the finely ground grain at a temperature of 98 to 100.0 in a vacuum chamber ItlL at a pressure of 25mm or less of mercuijr. Heating is continued until essentially constant sample weight is observed (about five hours). Christie (6) has developed a special type of vacuum oven which uses the dielectric heating effect of a high-frequency electric field to dry samples of various organic materials. Since such heating effects occur uniformly throughout the sample. heating time is only about ten minutes. The cost of this equipment is rather high. W This method of drying moisture samples requires Placing the sample over an efficient dessicant in a closed container. Since the dessicant maintains a very low vapor pressure within the container. the sample gradually dries to a constant weight. The Associ- ation of Official Agricultural Chemists (4) has set forth a standard Procedure which requires maintaining the ground sample in a vacuum with anhydrous sulphuric acid until constant weight is attained. This method 13 not widely used. since'it requires a great length of time to complete. HOwever, it is sometimes used on materials containing dry matter which decomposes easily whenheated. W. In the distillation methods. the sample is heated in some non-aquecus liquid and the moisture determined either by direct vOilumetric measurement of the water distilled and condensed from the grain or by determination of the weight loss of the sample during the heating process. The official toluenedistillation method for grain as described by the Association of Official Agricultural Chemists (Lt) requires heating the finely-ground sainple of grain in toluene. An apparatus is used which . r... , nv-r , , C b vu- . .. .l ..'. '.A ID: A. II 10 condenses the water vapor Cistilled from the toluene and collects it in a graduated tube for measurement. The Brown-Duvel distillation method for grain was first described by Brown and Duvel (7) in 1907. This method is somewhat similar to the toluene distillation method except that the grain is unground and heated in oil instead of toluene. The Brown-Duvel apparatus has been standard- ized and marketed commercially as a rapid method moisture determination. It is still used to a limited extent. ' A simple distillation method was devised by Dexter (8) for use by farmers in testing forage and grain crops. In this method. the moisture content is determined by heating the-sample in vegetable oil at 190°C. and determining the weight loss of the sample and the oil. The method may be applied in the field using an alcohol stove, tractor exhaust or Other means as a heat source. Indirect Methods W. The dependence of the electrical. resist— Eon-Ce or conductivity of materials upon their moisture content has been the basis for a number of successful moisture meters. , There is some dilegreement among investigators as to the eiact form of the relation- ship between moisture content and resistance; however, it is generally agreed that resistance varies inversely with the moisture content and t1lat relatively small chances in moisture content produce large changes in resistance. Therefore. a relatively inaccurate measurement of re- sistance may give a fairly accurate indication of moisture content. Moisture meters operating on the electrical resistance principle are ll usually calibrated against one of the standard oven or distillation methods. as are all indirect methods). The first report of the application of the resistance method found 111 the literature was by Whitney (9), who applied it to soils in 1897. iniitney used a direct measurement of the conductivity of the soil as a measure of the moisture content. Many investigators have published results on this type of measurement applied to soils during the past fdrfty'years. but the general conclusion drawn today is that the results arms too dependent upon the type of soil and its salt content. For this reason. the modified electrical resistance method developed byr Bouyoucos (10) has largely replaced.methods involving direct measure- ment of resistance. The method developed by Bouyoucos requires the measurement of the resistance of a nylon. glass fiber. or plaster-of- Paris element which is buried in the soil. To avoid the effects of PCfilarization. the resistance is measured with an alternating voltage 3‘Olln'rce. An attempt to apply the Bouyoucos method to the moisture test- ing of hay in storage was attempted by Kline (11). The method was un- stuccessful mainly because of poor contact between the hay and the resist- ance elements.. The first application of the resistance method to organic material ffrund in the literature was reported by Briggs (12) who adapted the meth- Ofl for testing grain in 1908. Briggs ured a Wheatstone bridge to measure the resistance between twmelz-inch electrodes placed in a glass jar of Wheat. He found a linear relationship between the moisture content and the logarithm of the resistance between the ranges of 11 and 16 percent moisture content. Data taken on wood by Stamm (13) (1h) indicate that v O‘uv o. Ins ‘5 s‘. , Is.. If 12 the logarithm of the electrical resistance decreases in direct proportion to an increase in moisture content up to the point of fiber saturation. «or about 30 percent moisture content. Above the point of fiber satura- tixmn there is very little change in resistance with increased moisture content. The variation below the fiber saturation point is such that the resistance increases 1.8 times for a decrease of one percent abso- lute moisture content. In other words. a decrease to 56.8 percent of the original resistance corresponds roughly to a moisture increment equal to one percent of the dry weight of the birch wood tested. . The simple logarithmic relation.put forth.separately by Stamm and Haseelblatt (15) indicates a linear relation when the logarithm of resist-— ance is plotted as a function of moisture content. The degree of lin- earity was found in later experimentation by Suits and Dunlap (16) to be not 'so great and more closely approximated by a relation of the form: Loam (Loglo R) 0C ( M ) % wherein R represents resistance and M represents moisture content. Whether or not the simple relation determined separately by Stamm and Hasselblatt or the more complicated one of Suits and Dunlap is true. the important fact is that resistance changes very rapidly with moisture content. The resistance method of moisture determination would be handi- calapped indeed if the same percentage of accuracy in resistance measurement were necessary to obtain a given percentage of accuracy in moisture deter- Initration. The above empirical relations between resistance and moisture cone tent do not rigidly hold true in all cases. Various other factors affect the relationships in varying degrees. According to Stamm (13) the 13 factors which affect the accuracy of the resistance method for moisture testing wood are as follows: 1. The amount of pressure applied and the type of contact between the electrodes of the device and the wood. 2. Dimensions of specimens 3. intent of specimen beyond electrode surfaces. A. Direction of flow of the current in the wood with respect to the grain. 5. Species and density of the wood. 6. Temperature 7. Salt content 8. Distribution of moisture in the specimens. 9. The presence of moisture on the surfaces parallel to the flow of current. 10. Polarization. electrolysis and electroendosmose. The work by Stamm in the evaluation of these various factors was forund to be very comprehensive and basic. Since the work was done on VOOd. which has many physical properties in common with forages and grains. an extensive review of Stamm's work follows. Three different kinds of electrodes were used in Stamm's earlier teats (13): (a) plain copper plates 7.5cm. by 7.5cm. made from 1mm. plate. “0 7.5cm. by 7.5cm. heavy brass plates with rantings m. wide and 2mm. deep cut in both directions so as to produce a cleated surface to be Pressed into the surface of the sample. and (c) 5.0cm. by 5.0cm. brass ‘Dlates with soft lead faces which. under pressure. conform to the sur- face of the board. The above electrodes were tested with a one-inch 11.1 Great deviation in readings was experienced in the specimen of wood. The use of the plain copper plates held firm by a wooden screw clamp. other two types of electrodes gave not more than a 20 percent deviation in the resistance reading. The routed electrode required tighter clamp- ing; to obtain consistent readings; therefore. the lead faced electrodes were employed in most of the subsequent experiments studying the other factors. The tests determining the effect of thickness on resistance in.- dicate that the resistance varies directly with the thickness. The Affect of specimen cross-section was studied by using two one-inch Specimens. one 5cm. square and the other 7.5cm. square. The former. when fitted with the 5cm. lead-faced electrodes. showed a resistance Of 82 megohms. The 7.5cm. piece tested similarly gave a resistance of 142 megohms. The unit resistance (ohms per sq. cm. of cross-section) for the Sam. square specimen was 2050 megohms and for the 7.5cm.speci- men was 1970 megohmh These results indicate reasonable verification of the inverse law of areas for the purpose of this study. (Inverse Law of Areas - The resistance of a mass of material is inversely Proportional to the cross-sectional area). The effect of the direction of current flow with respect to the grain of the wood was found to have a pronounced effect on the measured I‘eeistance of the sample. The resistance in the longitudinal direction (Parallel to the direction of the grain) is about half of the resistance in the cross-grain direction. Tests to determine the effects of species of wood on the resist- ance-moisture relation show that for the range of moisture content 15 determined (6 to 16 percent). the resistance-moisture relation for the six species tested was essentially the same. Again it is stressed that a rather wide variation of actual resistance is allowable to yield approximately the same calculated moisture content because of the log- arithmic relation between the two quantities. At the same time the species of the samples was varied. density was necessarily changed. It is indicated that appreciable changes in density of the samples (from .336 to .581»). as species were varied. failed to produce any error in determining the moisture content beyond that attributable to experimental error. The fact that a change in species does not change the resistance— moisture relation appreciably is a favorable condition indeed; however. it is certainly not to be expected that a change in the material would ha‘Ve the same negligible effect. It is of interest to determine the effect of the sample extending indefinitely beyond the area actually covered by the electrodes. Ex- Periments were conducted to indicate the amount of decrease in resist- anCe with an increase in cross-sectional area beyond the surfaces of the electrodes. A basswood specimen 21.5 x 19.5 x 2.7cm. gave 450 megohms with the 5cm_ lead'faced electrodes. while the section from the piece 5cm. square 11Iciicated 800 megohms. Increasing the cross-section of the piece 16.4 times caused a 141! percent decrease in resistance. Further experiments Showed that a decrease in thickness compared to the area will lessen t"he effect of cross-section on resistance. Other sections 5cm. square a11d 18 by 21mm” cut from 0.2cm. thick veneer. gave 1800 and 1600 meg- ohms respectively. A decrease in resistance of only 11 percent was 16 experienced with an increase of 17.3 times the original cross-section. Thus it appears that for very thin sections of material. such as veneer. no correction need be made for an extension of the cross-section. HOW- ever, for sections of l l/LI—inch stock, it appears that a correction of about one percent of moisture content would be necessary. The actual relation between resistance and moisture in sections of considerable thickness and extent is an exceedingly complex one; therefore. it is suggested that a complete tabulation for any one size of stock to be tested be worked out empirically. The effect of temperature on the moisture-resistance relation is not too damaging in practice if the ordinary range of working temperatures is maintained. Tests show that the resistance of wood falls off with an increase in temperature. producing a temperature coefficient negative to that of ordinary solid conductors. Tests show that for a decrease in temperature of 11°C.. the resistance is doubled; however. this change in resistance changes the calculated moisture content only about one Percent. The effect is rather small for ordinary operating conditions or may be easily compensated for when necessary. The salt content of wood and other materials has a definite effect on the electrical resistance. being inversely proportional to the resist- alice. There is little variance between the salt content of species of "God; however. the salt contents of various materials differ greatly. The salt content of a specimen of wood was varied artificially by soak- ing in sodium chloride solution for the purpose of demonstrating the effect of the salt content. The specimen decreased in resistance from “90 to 120 megohms after it had taken on about one percent of sodium -u.e', ,_. . Etch-6:3.C. ( - ...:.x t‘ i: 1.16 .. £01.. - to 'f‘:r‘l as»: 'u‘ CL“. u..; Q ‘eflavt- ,. V: I- Air‘s... . .~ 0 at with veg- ‘ ” ‘v x; ' h" “- 5:4. ' ‘. Ctr ‘ S...e ' "“C i - ‘ ‘ at :fi‘ehifi . 4C... A: .3. :Q ‘. H. “d ~s,‘ “u.{ r 'oee ‘V- 1? chloride. The salt content of wood is ordinarily less than one percent. Furthermore. a large part of the natural ash in wood is insoluble. It is probable that the variations in the calculated moisture content due to variations in the natural ash content in wood are much less than one percent; however. the variations are appreciable in other materials.“ particularly in soils where this method of moisture determination has met with very limited success because of the large variety of salt con.- tents in soils. Somewhat closely related to the effect of salt content on measure- ments is that of polarization. .or what is generally attributed to polar- ization. J'ew researchers have attempted a really rigorous explanation Of the phenomenon other than to say it is present when d—c voltage is used to measure a resistance where electrolysis may Operate. The effect Of polarization is to build up the measured d—c resistance to as high as twice the a—c resistance. The effect is attributed to two different actions: electrolysis and electroendosmose. Electrolysis has the effect 01' “building up an opposing emf or effective resistance in addition to the true resistance. Electroendosmose tends to create an uneven dis- tribution of free water within the sample due to electroendosmotic motion of the water. Both of these effects are somewhat dependent on time to come into effect and are probably not very significant as long as mois- ture contents below that of the fiber saturation point are studied. Surface moisture has a definite effect that cannot be denied. In ext"rams cases in the wood tests. where a dry Specimen of 2000 megohms Was Btreaked with a wet cloth from electrode to electrode. the resistance measurement indicated a short circuit. The resistance was reduced to 18 half the dry resistance when a man's breath was blown on the surface. These tests indicate that the resistance method is highly inapplicable to wood. the surfaces of which have been wet recently. However. this is a special condition and should. not apply to the general case of wood specimens in storage. The effect of unequal moisture distribution in the sample definitely limits the use of the resistance method as discussed thus far. Further research on the subject by Stamm (ll-'r) revealed that the plate-type electrodes were not suited to applications where non-uniformity in mois- ture content of the specimen is encountered. His research revealed that pin-type electrodes could be so arranged as to measure the average moisture content of a sample of wood containing non-uniformly distribut- ed. moisture. Electrode pins penetrating to a depth of one-fourth the thickness of the sample were found to indicate average moisture contents ' Where normal. drying gradients existed. The effects of surface moisture and surface finish on woods were much less with the pin-type electrodes than with the plate types. Additional tests with the pin-type electrodes confirmed that the species 6-116. density of the wood had little effect on the readings of moisture content. The width and length of the board tested has no effect: on I'13aclings with pin-type electrodes. as it did with the surface contact electrodes. A number of investigators have published results of tests conducted °n Commercial resistance-type moisture meters for grain. Such work °°naisted mainly of calibrating the meters with oven determinations and determining statistically the accuracy of the meters. One of the more 19 comprehensive studies was made by Cook, Hopkins, and. Geddes (17). In reviewing the test results of four commercial resistance-type meters for grain. it is evident that there is a linear relationship between the logarithm of the resistance of wheat and the moisture content in the range of 11 to 17% moisture content. Above 1779 moisture content. the relation between moisture content and the logarithm of resistance is parabolic and is described by a relationship of the following form: Moisture content 2* 53(log R)“: (log R)2 where A. B. and C are constants. Bylnka. Martens. and Anderson (18) made a study of ten electrical meters which operated either on the electrical resistance or the di- electric principle. The Tag-Heppenstall meter. which measures the electrical resistance of grain with an ohmmeter circuit. gave the most accurate results. This meter is still the only electrical meter approved for use in the inspection of grain under the Official Grain Standards Of the United States. Aranovsky and Sutcliffe (19) tested a commercial resistance-type met er on baled straw. A “satisfactory" comparison was reported between the average of 214 meter readings per bale and the oven-determined moisture content of the bale. The results were somewhat erratic at higher moisture contents; however. the tests were conducted on bales of “raw which had been stored outdoors and were wet in spots. An increase in the reading of the moisture meter with an increase in bale density Was reported. 20 Dielegtrig Mgthoda. The dielectric constant of water is 81. while the dielectric constant of such good insulators as dry wood ranges from ?.5 to 7.7. accox ding to data taken from the Smithsonian Physical Tables (2(3). Materials which contain greater amounts of moisture have higher dielec- tric constants. Since capacitance increases with the dielectric con- stant. the capacitance between two plates of condenser having a hygro— scopic material as a dielectric will increase with an increase in the moisture content of the material. Thus by measuring capacitance. an indirect measurement of moisture content is obtained. This has been the operating principle of a number of moisture meters. Lampe (21) in 1929 used the dielectric method to determine the moisture content of cereals. fruits and vegetables. The results obtain- ed varied less than 0.5% from those obtained by regular drying methods. Balls (22) determined the moisture content of baled cotton by measuring the capacitance between the nine iron hoOps on the bales. Balls also applied the dielectric method to the measurement of soil moisture. In this work. he found that the capacitance bridge method he had used previously on cotton would not work well because of the Power loss in the soil when it was used as a dielectric. By using a resonance method he avoided this difficulty. A number of dielectric-type moisture meters have been described by various authors since the early work of Lamps and Balls. The meters described vary chiefly in the details of the capacitance measuring cir— cuit and in the product to which the meter is applied. Variations in the meter readings-due to rather small variations in sample density have been of much concern to those developing dielectric-type meters. 21 A patent by Stein (23), the originator of the well-known Steinlit‘e moisture tester for grain. has as its principal feature a design for an arrangement to let the sample fall into the test chamber in such a manner that a constant density is obtained. The sample is held by a trap door at a fixed height above the test chamber. The sample is dropped an, m into the chamber instead of being poured in at a vary- ing rate. Hartmann (24) is the author of a British patent for a design with essentially the same purpose as the Stein patent. Concerning the comparative accuracy of dielectric-type meters. Zelerw (2) states the following: "----- it has never been demonstrated that the capacitance- type meters are capable of greater over-all accuracy in testing the normal run of commercial grain than are con- . ductance meters. Capacitance meters should be particular- ly useful. however. in testing freshly dried or tempered grain. and grain of very high or low moisture content”. Zeleny further states that capacitance meters have been difficult to keep in adjustment and that the results of capacitance meters are less consistent. particularly when the results of several meters are Compared. Hart (25) and Sherwood (26) have applied the dielectric method to 1She determination of moisture in hay. Hart used a capacitance bridge Circuit operating at 1000 cycles per second to measure capacitance of hay samples. A rigorous accuracy test was not made with this instrument; hm"’GVer. the accuracy was estimated at 5% moisture. Considerable drift in the indication of the meter was observed under certain conditions. Sherwood used a capacitance substitution method to measure the caPacitance of hay samples with a resonant circuit. The standard error or eatimate of this meter was 3.0 moisture percentage points when testing 22 one pound samples of alfalfa. He avoided excessive power loss in the sample by insulating the plates of the test capacitor with sheets of plastic. Tuning the circuit for each detemiration is necessary with this meter. W. This method involves mixing the material to be tested with a chemical which will react with the water in the material to produce a gas. The volume of the gas or itspressure at constant vo lume may be measured to determine the quantity of gas evolved and. to indicate indirectly the amount of moisture in the material. The first reference to this method found in the literaturevwas made by Roberts and Frazer (27) in 1910. They used calcium carbide to determine the moisture content of crude petroleum. The volume of the acetylene produced was measured by a method of liquid displacement. The equation for the resulting reaction is as follows: Ca02+ 2320 -—> Ca(OH)2 + 6232 (acetylene) About 580cc. 02112 (at 760mm. Hg and O C.) were obtained for every gram of water in the sample. A 11.5. patent on a commercial device employing this method was issued to Stanworth (28) in 1932. This device is manufactured in E“gland and is now being sold in this country principally for use by ca«Sting foundaries for testing the moisture content of sand. The appa— ratug consists principally of a steel receptacle fitted with a Bourdon- tube pressure gauge and safety valve. a small scale for accurately weighing the test sample. and a quantity of reagent. This apparatus is rec"Olllxnended for use by farmers on hay and grain. An operating time of three minutes is claimed for grain tests when the sample is ground. Fukunaga (29) used a steel receptacle and a steam gauge (0-30 psi range) to employ this method to soil moisture determination. The re- sults obtained were found to be reproducible with great accuracy. Masson (30) used this method on several organic substances and claims that there are no side reactions which.would produce error. He used benzene to dissolve such fatty substances as butter and cocoa. He determined the yield of acetylene to be 10.50c. of gas at O C. and 760mm. of pressure for each 18mg. of water. The theoretical yield is 11.2cc. Masson used a mercury manometer to measure the gas pressure. Perkins and Kipping (31) state that acetylene can be spontaneously explosive above pressures of two atmospheres and should not be stored even at atmospheric pressures in metallic containers.since explosive metallic derivatives may be formed. It can be safely handled in solue tion with acetone. Calcium hydride was proposed as a reagent in this method by Rosenbaun and wanton (32) in 1930. The reaction is described by the following equation: C832 + 2320 —- Ca (0102 + 232 .An.advantage over the use of calcium carbide is claimed.since twice the Volume of hydrogen.is produced as compared to the volume of acetylene jproduced in the carbide reaction. Rosenbaum and Walton (32) further State that hydrogen is not as soluble in other materials as acetylene; however. it cannot be used if calcium hydride reacts with some material other than.water to produce hydrogen or some other gas. thevarp (33) also used calcium hydride to determine the moisture content of various solids and highpboiling liquids. 74 Larson (34) reported the use of Grignard reagent (methyl magnesium iodide)to produce methane as a moisture indicator. He used this method to determine moisture in mineral oil. Dietrich (35) reported the use of the reaction of magnesium nitride with water to produce ammonia for the determination of water in alcohol. §n§§1112_fizaxitx_flhmhpda A.method for determining the moisture content of sand for concrete work was suggested by Waters (36). This method de- pends upon knowing the dry specific weight of the sand tested, which is determined.previously. A.two-liter cylindrical container and scale are used to take the data necessary to determine the moisture content. The .wet sand sample is weighed and immersed in water inside the container. The container is filled to overflowing with water and weighed. The value of the total weight of sand and water is referred to a table which gives the moisture content for the corresponding dry specific gravity of the sand. Bauer (37) discussed this method further and related unpublished Work by the late Professor Dunagan of Iowa State College toward the de- Velopment of equipment for weighing samples in air and suspended in water. Lee (38) used a Specific gravity method to determine the maturity Of frozen vegetables, namely corn and peas. The method involved weigh- ing in air and then in a solution of xylene and carbone tetrochloride of specific gravity 1.000. A.weak brine solution.was also used. The Spe— cific gravity of the sample is determined by the following relation: "-4 Weight in air - Weight in liquid Specific Gravity_' Weight inLair x Specific gravity of liquid Batson (39) used a specific gravity method in the determination of y 5) kn moisture in sweet potato starch. The density of the dry matter was de— termined. Moisture content was based upon the difference in.weight be- tween five hundred ml. of water and five hundred ml. of an aqueous SOlUF tion of one hundred g. of starch and water. Conversion tables were pre- pared. By weighing to 0.01g. accuracy. a moisture determination could be made to an accuracy of 0.7% moisture. The method is reported to be applicable to materials insoluble in water, wettable by water. and mark- edly denser than water_when in the dry state. Materials tested should also preferrably be of a nonpswelling or an non-hydrating nature. homo— geneous or of constant density, and non-colloidal. Wm When grain or'arw other hygroscopic substance is placed in a closed container. the interstitial air becomes stabilized at a relative humidity dependent upon the moisture content of the grain. This is due to the equalization of the partial pressure of the water vapor in the air with that of the grain or other material. Thus a measurement of the relative humidity of the interstitial air 1n.equilibrium with a material may be used as a indication of moisture Content of the material. The relation between the equilibrium relative humidity and moisture content is different for each product and has also been found to be somewhat dependent upon the history of the particular Sample. Although the hygrometric method may not be considered a highly accurate method of moisture content determination under field conditions, the measurements of relative humidity are actually better indices of storage quality than are moisture measurements. This is due to the fact that the growth of molds and other microorganisms which cause “J C\ grain spoilage is dependent largely upon the relative humidity. Milner and Geddes (40) have shown that rapid mold growth occurs at ordinary _ temperatures in grain.when the relative humidity of the interstitial air is greater than 75%. Below 75% relative humidity, mold growth is greatly retarded. The variations in the method of application are about as numerous as the methods of indicating relative humidity. There have been articles published and patents issued on the application of almost all known | methods of humidity measurement to the indication of moisture in many different materials. Dexter (#1) has developed a‘ simple method for use on grains which is based on.wet-bulb and dryebulb temperature readings of the air and ggradn inside a closed container. The wet bulb is wetted with a saturated solution of a metallic salt which maintains about the same equilibrium relative humidity as the grain tested, thereby reducing the wet-bulb de- Pression and time to reach a stabilized temperature reading. Agitation 18 £1180 used to decrease the time to come to equilibrium and to produce a1rmovement around the wet bulb to aid vapor transfer. The experimental aPIQaratus consists of two thermometers mounted in a rubber stopper, salts, and..g container such as a milk bottle wrapped for insulation. Agitation- is provided by shaking. Gauss e1 31.. (’42) developed a device which withdrew air from the center of bales of cotton and; passed it over wet and dry-bulb thermometers f°r humidity measurement. In terms of moisture content. the standard deviation of the device was 1.115% moisture. The time required to run a 8811191413 was two to three minutes. Balls (43) employed similar methods on 97 baled cotton, but he indicated a standard error of 2.5% moisture. He attributed this high error to hysteresis effects caused by variations in the history of the samples with respect to past wettings from rain. Ives (4b) develOped a simple device which indicates the dew-point temperature of the air inside a closed container containing the sample. The dew-point temperature is determined by observing the temperature at which a polished metal surface collects dew inside the container. The metal surface is the outside of a small tube which contains and is cooled by an evaporating liquid of low vapor pressure such as acetone.. The rate of evaporation is increased by bloWIng air through the acetone. The dew-point temperature is assumed to be the temperature of the acetone as determined by a thermometer. Another thermometer determines the dry— ln11b temperature. The relative humidity is determined in the usual nuznner from the dry bulb and dew-point temperatures. Dexter (45) developed a very simple hygrometric method of deter- misiing whether hay or grain crops will keep in storage. A salt which when saturated with water has an equilibrium humidity of about 7575 is Inilted with the sample in.a closed container. If the sample is too wet t0 Iksep ( has an equilibrium humidity above 75%) the salt crystals will 1tunp together because they are saturated. If the sample is dry enough to’ Store, the salt will not lump. The property of certain salts to change color when exposed to vzr- iCN1E relative humidities'has been used as an indication of moisture con- tent. Bother and Grau (46) were issued a British patent in 1926 for the apPlication of this method to wood. textiles, and other substances. Toddand Gauger (#7) describe the application of various cobalt salts to (‘0 CO this method. Severe color shifts with temperature were noted. Dexter (#8) used a mixture of ferric ammonium sulphate and potassium ferrocycnide in a carrier of sodium chloride as a colorimetric humidity indicator. In the dry state this mixture is blue. but in the presence of water. ferric thioqyanate which is red is formed. The use of this type of indicator eliminates the serious temperature effects encountered with the cobalt salts. The indicator is applicable in the range of 8— 12% for wheat and the accuracy of observation is approximately 1% mois- ture. Miloslavskii's and Palant's (49) attempts to apply the hygrometric method of moisture measurement to tobacco were unsuccessful. They atfiributed the failure of the method to the fact that tobacco is neither chemically or biologically definite. W. Most organic materials when dry are relatively Poor conductors of heat, but as the moisture content of such materials 18 increased. their thermal conductivity likewise increases. Powley EIH1 Algren (50) found that the thermal conductivity of wood varies es- Senntially in a linear fashion.with moisture content. They also found tTuat thermal conductivity varies linearly with density. Shaw and Baver (51) designed a soil moisture meter based on a ins direct measurement of thermal conductivity of the soil. A Wheatstone briJLge circuit was used to compare the electrical conductivity of two resistors having a high coefficient of thermal resistivity. One resis— tOIP‘was placed in the sample of unknown moisture content and the other "”18 placed in a dry sample of the same soil. Sufficient current was Passed through the resistors to cause significant temperature rise. 2’9 The temperature of the resistors at a given time after the current start— cd to flow is a function of the ability of the resistors to dissipate heat. which is determined largely by the thermal conductivity of the materials surrounding them. The resistor in the dry soil will have a higher temperature than the resistor in the wet soil because the thermal conductivity of dry soil is less. The temperatures of the two resistors are compared by comparing their electrical conductivity with the bridge circuit. The meter developed by Shaw and Baver gave a fairly reliable in- dication of moisture content; however. from 30 minutes to 2 hours were required for it to yield a stabilized reading. By taking a reading at a pro-determined time after the current flow through the resistors was started. an indication of moisture content was obtained from previous calibration for that period of time. By this procedure. a reading could be obtained in 10 to 15 minutes. One of the greatest probable sources of error in this method is caused by the fact that when the temperature of the material is increased next to the resistor in the wet sample. the partial pressure of the water vapor at this point will be increased and water vapor will be transferred away from the resistor to the colder parts of the sample. This means that the moisture content of the sample changes at the point of measurement. This effect can be reduced by limiting the temperature of the resistors and the time of heating. - The principal advantage of the thermal conductivity method is that it 18 not affected by the salt content of the material tested. 30 Mechaniggl Mgthgd. Mechanical means of determining moisture content have not received a great amount of research in a formal way; however, these means are probably the oldest method man has used to estimate moisture. particularly with regard to farm forage and grain crops. The bite test on grain and the twist or thumbnail test on hay are methods which depend upon the mechanical properties of the product tested. Nichols (52) patented a mechanical device for determining the moisture content-of raisins and such materials. His apparatus consisted of a cylinder in which the sample was compressed by a known weight. The percentage of compression was determined by a suitable mechanical linkage which was calibrated to indicate the moisture content of the product. Brough g]; a]... were granted a British patent for a device which in- dicated the moisture content of webs of paper or cloth. This device transmitted sound waves on one side of the web and received them on the other. The exact variable measured was not indicated in the literature a-‘Vailable. W In recent years there have been two methods proposed which detect the presence of hydrogen nuclei in material as an index of m01sture content. the neutron scattering method and the nuclear resonance absOrption method. Belcher g1 3.1. (51+) measured the moisture content of soil by a neutron scattering method. Fast neutrons emitted from a radium-beryllium 01‘ a polonium—beryllium source are allowed to pass into the soil to be tested. In passing through the material. the neutrons collide with the at"Omic nuclei of the material. In these collisions. they lose some of 31 their energy and are scattered in all directions. When a neutron col— lides with a heavy neuclei such as that of silicon, much less kinetic energy is transferred from the neutron than if it hits a liynt nuclei such as that of hydrogen. (Analogy: A billiarl ball hitting another billiard ball head-on loses all of its kinetic energy. If it hits a heavier object such as a bowling ball, it loses little of its kinetic energy). - Thus if more hydrogen nuclei are present in the material. a greater number of slow neutrons will be scattered back to the vicinity of the source. 3y locating a radiation counter sensitive to slow neutrons near the source, a count of slow neutrons may be obtained which is a function of moisture content. The counting rate of slow neutrons was foumd by Belcher to be essentially in direct proportion to the moisture content. The accuracy of the method when used on soils in 5.1M). was estimated at 1%.moisture. Non-homogeneities such as rocks in the soil. salt content. or the hummus content of organic soils did not greatly effect moisture readings by’ this method. The radiation hazards involved in using this method are recognized by th authors. A source of about 8 millicuries activity 18 ‘used in this work. Since the cost of the apparatus is quite high and the equipment rat”'l’ler complex, this method does not show great promise as a field Inethod for use by farmers. A highequclity radiation counter. including aGeiger—Muller tube, counting circuit, and scaling circuit,_as well as a I2Ither expensive radioactive source are required. Shaw and Elsken (55) showed a linear relation between the moisture- content of several vegetable materials, gig, apple. potato. and maple wood, and measurements of the nuclear magnetic resonance absorption of the hydrogen nuclei. The sample was placed in a unidirectional field of 3500 genes and also in a second field of constant radio frequency. With a favorable relationship between the-strength of the unidirectional field and the frequency of the radio—frequency field, the hydrogen nuclei absorb energy from the radio-frequency field at a maximum rate; i.e.. resonance occurs. The greater the number of hydrogen nuclei present, the greater is the absorption of energy from the field, as indicated by an increased current flow through the radio-frequency coil producing the field. Shaw used a third 60—cycle field to sweep the effective value of the unidirectional field through the resonance value and to facilitate the indication of the magnitude of the peak of the radio-frequency cur— rent on an oscilloscope. A sample 0.3cm. by Sen. long was used in the tests by Shaw and I31sk:en. The accuracy of the method as a means of moisture determination buss feund to be 1% moisture. Correction for shrinkage of the sample Vans necessary when a series of tests were conducted on a single sample vdrich.was allowed to dry between readings. (1) (h) (5) (6) (7) (8) (9) (10) (11) (12) Literature Cited Isaacs, G. w., A Bibliography of Rapid Methods of Moisture Deter- mination (with Abstracts). unpublished research report. Mich. State College. Agricultural Engineering Department Library. 1953, 63 numb. leaves. #50 references. Zeleny, Lawrence. Methods for Grain Moisture Measurement. Agri- cultural Engineering. v 35. pp. 252-256. (195“). U.5. Department of Agriculture. Air Oven and Water-Oven Methods Specified in the Official Grain Standards of the United States for Determining the Moisture Content of Grain. Service and Regulatory Announcements no. 1&7. (revised) 3 pp. (19#l). Association of Official Agricultural Chemists. Official Methods of Anaxysis. Association of Official Agricultural Chemists. Washington. D.C.. 7th ed. (1950). Brabender. C. V.. Apparatus for Determining the Moisture Content of Materials such as Foods. Wood. and Coal. U.S. Patent no. 2.047.765. July 14. 1936- Christie. Alfred. High Frequency Dessicator. U.S. Patent Eb. 2.360.108. (1944). Brown. Edgar and Duvel. J. W. T. A., A Quick Method for the Deter- mination of Moisture in Grain. U.S. Dept. of Agriculture Bureau of Plant Industry Bull. 99. (1907). ‘ Dexter. S. T.. An Oil-distillation Method for Determining the Moisture Content of Farm Crops. Mich. Agr. Expt. Sta. Quarterly Bull. v 31. pp. 275-286. (1949). Whitney. M.. Gardner. F. D. and Briggs. L. J.. An Electrical Method for Determining the Moisture Content of Arable Soils, U.S. Dept. of Agriculture,soils Division Bull. to. 6. 26 pp.. illus.. (1897). Bouyoucos. G. J.. An Electrical Resistance Method for the Contin— uous Measurement of Electrical Resistance under Field Conditions. Mich. State College. Agr. Expt. Sta., Tech. Bull. no. 172. (19uo), Kline. C. K.. Methods and Equipment for Determining the Moisture Content and Rate of Drying of Hry in the Field and in the How, Unpublished M.5. thesis. Michigan State College. 1929. 116 numb. leaves. . Briggs. L. J.. An Electrical Resistance Method for the Rapid Deter- mination of the Moisture Content of Grain. U.S. Dept. of Agri- culture. Bureau of Plant Industry Circular no. 20. 8 pp.. (1908). (13) (14) (15) (16) (17) (22) (23) 68+) has) (26) 34 The Electrical Resistance of Wood as a Measure of steam, A. Ute, Indus. and Eng. Chem. v 19. pp. 1021-5. Its Moisture Content, (1927). Stamm. A. J.. An Electrical Conductivity Method for Determining the Moisture Content of Wood. Indus. and Eng. Chem. (Analytical Edition). v 2. pp. 240. (1930). Hasselblatt. M.. The Aqueous Vapor Tension and Electrical Conduc— tivity of Wood in Relation to Its Water Content. Z. anorg. allgem Chem. v 154. pp. 375-85. (1926). Suits, C. G. and Dunlop. N. E., Determination of the Moisture Content of Wood by Electrical Means. General Electric Review. 7 3“. pp. 706-13. (1931). Cook. W. H.. Ebpkins. J. W. and Geddes. V. J.. Rapid Determination of Moisture in Grain, Part II. Calibration and Comparison of Electrical Moisture Meters with Vacuum Oven for Hard Red Spring wheat. Canad. Jour. 398.. v 11. pp. hog-u? (193s). Hylnka. 1.. Martens. V. and Anderson. J. A.._ A Comparative Study of Ten Electrical Meters for Determining the Moisture Content of Wheat. Canad. Jour. Res.. Section E. v 27. PP. 382-97 (1999). Aranovsky. 5. I. and Sutcliffe. H. M.. Rapid_Estimation of Moisture of Straw in Bales, Paper Trade Journal. January 15. 19GB. Standard Handbook for Electrical Engineers. 'McGraw-Hill Co.. New York. 7th ed. pp. 343. . Iempe. B.. Rapid Moisture Determination, Z. Spiritusind. v 52. PP- 337-89 (1929). (Seen in abstract only) Chem. Abs., v 2t, pp. 2965. Balls. W. L.. Rapid Determination of Water Content in Undisturbed Soil and in Bales of Cotton, Nature. v 129. PP. 505-6 (1932). Apparatus for Testing the Moisture Content of Cereal, Ste-1n, F. W. 9 U.5. Minerals. Foods, Gases. Plastics, Papers. Explosives, etc.. .Patent no. 2.251.541. August 5. l9hl. Apparatus for Determining the Moisture Content of Hartmann. S. H., British Patent No. 440.966, January 9. 1936. Granular Substances, Hart. W. 3., Mow Curing Baled Hay. Unpublished M.S. thesis. Michigan State College. 19h8. 99 numb. leaves. Sherwood. E. M.. The Rapid Determination of Bay Moisture Content. Unpublished M.S. thesis. Michigan State College. 1951. 118 numb. leaves. (27) (28) (29) (30) (31) (32) (33) (34) (35) (36) (3?) (38) (n9) ' (#0) v‘ ' V a 4' , ,. t. 2 , h L 2 y __ . ’ 1, ‘ '~ ~ . . . - b k L '1 9-0; $.1LU $-Lt— "va [481‘ Roberts, R. W. and Frazer. A.. Lrsy Prrc in Petroleum, Jour. Soc. Chem. Ind.. v 29. pp. 197 (1910). Shanworth, 5.. Moisture Determination Apparatus, U.S. Patent ;b. 1.976.752. October 16. 193M. Funkunaga. E. T. and Dean. L. A.. The Carbide Method for the De- termination of Moisture in Soil, Hawaii Agr. Exp. Sta. Report 1938. pp. 50. Masson. Irvine, The Use of Calcium Carbide for the Determination of Moisture, Chemical News, v 103,-pp. 37—38. (Seen in abstract only), Chem. Abs.. v 5. pp. 1377 (1911). Perkin. W. H. and Ripping. F. 5.. Organic Chemistry. J. D. Lippincott Company. Philadelphi (1916)- hn-enbaum,‘C. K. and Walton. J. H., Use of Calcium Hydride for the Determination of the Solubility of water in Benzene. Carbon Tetrachloride. and Toluene. Jour. Am. Chem. Soc.. v 52. pp. 3568 1930). Notevarp. 0.. Determination of Mater Content by Means of Calcium Hydride. Z. anal. Chem.. v 80, pp. 21 (1930). (Seen in abstract only). Chem. Abs.. v 2h. pp. M731. Larson. R. 6.. Application of Grignard-Reagent to the Study of Mineral Oils; Jour. Ind. Eng. Chem. (Ana1.). v 10. pp. 195 (1938). Dietrich. K. R. and Conrad. C.. Determination of Water in Alcoholic Motor Fuels, Z. Angew.-Chem.. v Eh. pp. 532 (1931). (Seen in abstract only), Chem. Abs.. v 25. pp. 4491. Waters. C. R.. New Method to Determine Moisture Content in Sand. Eng. News Record. v 132. pp. 3u1 (1944). Bauer. E. E... Moisture Content in Sand. Eng. News Record, v 13?. pp. 726e7. (1949). Lee. E. A.. de Felice. D. and Jenkins. R. R.. Determining the Maturity of Frozen Vegetables. Ind. Eng. Chem. Anal. ed.. v 14. pp. 240-1 (19M2). Batson, D: M. and Hogan. J. T.. Estimation of the Moisture in Sweet Potato Starch. Anal. Chem.. v 21. pp. 718-21 (1949). Milver. M. and Geddes. W. F.. Grain Storage Studies, Part III. The Relation Between Moisture Content, Mold Growth. and Heepirc— tion of Soybeans. Cereal Chem.. v 23. pp. 225-247 (19“6). MD (42) on (44) up (46) aw) (48) (49) (50) (51) (52) (53) Dexter. S. T.. The Wet and Dry-bulb Method of Moisture Content Determination. Mich. Agr. Exp. Sta. Quarterly Bu11.. v 31. Gauss. G. 3.. Shaw. Charles S. and Diever. Wilda R.. A Practical Seed-Cotton Moisture Tester for Use in Gins. U.S. Dept. of Agriculture Circular no. 261. Novenber. 1941. Balls. W. L.. Air Test for Moisture in Cotton. Jour. of the Textile Institute. v 40. T757-66. 1949. (Seen in abstract only). Chem. Abs.. v 44. pp. 6133c. Ives. N. C.. A Dew—point Moisture Indicator. Agr. Eng. v 33. pp. 85-87 (1952). Dexter. S. T.. A Method for Rapidly Determining the Moisture Content of Bay or Grain. Mich. Agr. Exp. Sta. Quarterly Bu11.. v 30. pp. 158—166 (1947). Bother. Paul and Grau. George. Apparatus for Estimating the Water Content of Hood. Textiles. etc.. German Patent No. 635,750. September 23. 1926. (Seen in abstracts only). Chem. Abs.. v 31. musah Todd. R. C. and Gauger. A. I.. On the Measurement of Mater Vapor in Gases. Proc. Amer. Soc. Test. Mat.. v 41. pp. 1134 (1941). Dexter. S. T.. A Colorimetric Test for Determining the Percentage Moisture or the Storage Quality of Farm Products or Other Dry Materials. Mich. Agr. Exp. Sta. Quarterly Bu11.. v 30. no. 4. pp. 422-26 (May. 1948). Miloslavski. N} M. and Palant. A. 1.. Determination of Moisture in Tobacco By Means of a Hygrometer. Ukrainskii Kham. Zhur. v 5. Tech. Pt.. pp. 117—25. 1930. (Seem in abstracts only). Chem. Abs. v 25. PP. 2521. Rowley. F. B. and Algren. A. 3.. Thermal Conductivity of Building Materials. Univ. of Minnesota Eng. Exp. Sta. Bull. no. 12. 1937. Shaw. B. T. and Baver. L. D.. Heat Conductivity as an Index of Soil Moisture. Jour. Amer. Soc. of Agron.. v 31. pp. 886—891. 1939 . ' ' Nichols. P. F.. Apparatus for Determining the Moisture Content of Raisins and Like Substance by Compression Tests. U.S. Patent 1b. 1.6239263! April 5. 19270 Brough. A. R.. Puleston. P. R. and the Watford Engineering works. Ltd.. Method for Measuring the Moisture Content of Webs of Paper Cloth. etc.. British Patent No. 486.005. June 28. 1937. \d ‘J (54} Belcher. D. J.. Cuykendall. T. B. and Sack. H. 5.. The Measure- ment of Soil Moisture and Density by Neutron and Gamma-Rey Scattering. U.S. Civil Aeronautics Administration. Technical Development Report no. 127. October. 1950. (55) Shaw. T. M. and Elsken. R. R.. Nuclear Magnetic Resonance Absorp— tion in Hygroscopic Materials. Jour. of Chemical Physics. v 18. pp. 113-4 (1950). THE PHYSICAL NATURE OF FORAGE AND GRAIN CROPS WHICH AFTECTS MOISTURE DETERMINATION Forages The greatest problem in the determination of the moisture content of hay curing in the field is that of the non-uniformity of its mois- ture content. Obtaining representative samples for moisture testing is very difficult. The moisture content of all the hay of a given variety in.a field is essentially the same at the time of cutting. However, flhe rate of drying after cutting is dependent upon a number of factors and.differs at various points in the hay mass. Some factors which affect the drying rate of cut, hay plants are: l. The parts of individual plant involved. 2. The relative opening of the stomates or pores in the leaves. 3. The atmospheric conditions to which the cut plants are exposed. 4. The species of hay. 5. The rankness of growth. 6. Topography of the field. The rate of drying of cut plants takes place mostly through trans— Pil‘ation, a natural process which continues after cutting. The rate of ‘tr‘unspiration of moisture from the leaves is much greater than from the “one. In fact. it has been shown that the leaves aid in the drying of “19 Items by transpiring not only the moisture which is contained in the leaves when the plant was cut. but also part of the moisture of the stems (1). As the leaves become drier, moisture moves into the leaves from the stems by cohesive and capillary action. Transpiration from the leaves is greater because of the greater surface area exposed and the greater moisture permeability of the leaf surfaces. The rate of transpiration from the leaves is to a great extent de- pendent upon the degree of contraction of the stomates, the very numer- ous small openings in the leaf surface through which most transpiration occurs. The stomates of uncut plants have been found to open and close periodically throughout the day. The opening of the stomates does not remain fixed after cutting. They close at a rate dependent upon the atmospheric conditions to which the cut plant is exposed. Jones and Palmer (l) found that the stomates of cut hay curing in the windrow stayed open longer than the stomates of hay in the swath. They also found that hay raked into a windrow two hours after cutting was drier at the end of the day than hay left in the mower swath. for any given opening of the stomates, the rate of transpiration from curing hay is also dependent upon the partial pressure of the water vapor in the atmosphere. Periods of highly humid weather result in reduced rates of transpiration, since the atmospheric water vapor pressure is more nearly at equilibrium with the pressure of water vapor inside the plant. From the standpoint of hay quality, windrcwing soon after cutting 1! desirable. When hay is windrowed, the conditions to which individual Plants are subjected are highly variable. These conditions range from OJttensive exposure to sunlight and wind at the tap of the windrow to rather close confinement at the bottom of the windrow. For this reason, hay at the top of the windrow dries much faster than hay at the bottom. The species of hay also affects the drying rate to some extent. In hay mixtures, this factor contributes considerably to the non-uni- formity of moisture content. For example, in alfalfa-bromegrpss mix» tures, the alfalfa usually dries at_a slower rate than the-bromegrees. due to the slow drying rate of the alfalfa stems. However, in grasses having very heavy stems, such as a rank growths of Johnson grass, the drying rate is much lower than in alfalfa (l). Rankness of growth is in turn affected by soil conditions which may vary widely in a given field and result in non-uniformity in the moisture content of the hay when curing. Grains Considerable non-uniformity of moisture content exists in grain, particularly in that which has been recently harvested. However, it is much more uniform in moisture content than hay and is therefore easier to test for moisture. Soil conditions may affect the rate of maturing of grains. Since soil conditions frequently vary throughout a given field, the maturity and the moisture content of the grain may likewise vary. - Non-uniformity of moisture content within the individual kernels ¢3f grain also exists and affects the reading of some types of moisture Ineters. The moisture content is normally unevenly distributed within the grain kernel, the germ having the highest moisture content. Mois- tttre may also be abnormally distributed in grain which has been recently 41 subjected to drying by rapid artificial means. In this case, the outer surfaces may be abnormally dry, while the inner parts are still rela- tively wet. Non-uniformity of moisture content is usually more pro- nounced in the larger grains such as corn than in the smaller grains such as wheat, since the smaller grains equalize in moisture content more rapidly. Properties of Forages and Grains Which Vary with Moisture Content Moisture content affects nearly all of the physical properties of forages and grains or any other hygroscopic material. Among the phys- ical properties which are affected by moisture content are: 1. 2. 3. Electrical resistance or conductivity Dielectric properties Equilibrium relative humidity of the air in a closed vessel containing the material. . Mechanical strength Thermal conductivity Specific gravity Each of the above factors have been used as bases for rapid indirect Inethods for determining the moisture content of some materials. The Principal source of error in all of these methods lies in the fact that there are other factors besides moisture content which also affect the Ffinvsical properties listed. For example, temperature as well as mois- tttre content affects most of the physical properties of materials. 42 The rather non-exclusive effects which variations in moisture con- tent have upon the properties of materials and the great non-uniformity Aof moisture content in forages and grains makes the problem of develop- ing simple, rapid methods of moisture determination a challenging one. Literature Cited (1) Jones, T. N. and Palmer, L. 0., ”Field Curing of Bay as Influenced by Plant Physiological Reactions“, Part I,Agr. Eng., v 13, pp. 199-200, 1932. Part II,Agr. ing., v 1h, pp. 156-158, 1933. Part III,Agr. Eng., v 15, pp. 198—201, 193“. 43 PRINCIPAL DESIGN REQUIREMENTS OF FARM MOISTURE METERS Accuracy High accuracy is one of the most obviously desirable attributes of a successful farm moisture meter or any measuring device. The accuracy required in farm applications seems quite variable with the crop tested, the range of moisture content, and the purpose for which the test is made. It should be generally assumed that a moisture meter developed for farm use would not be used in buying and selling grain on a moisture content basis. Therefore, the accuracy required of a moisture meter in a given situation depends mainly upon the tolerances of the specified. safe storage moisture contents for the products involved. No such tolerances have been found set forth in the literature available: there- fore, no very sound basis can be given for estimates of the required accuracy of moisture meters. A maximum error of one moisture percentage point seems allowable for testing grain and field cured hay to determine its storage quality. Bay which is to be partially mow-cured need not be moisture tested to accurately, since observations of the hay temperature are probably the most practical way of determining when the drying fan may be shut off. For testing hay to determine when to place it in the mow for mow our 136. an error of moisture determination of three moisture percentage 9°11”. should be allowable. For testing hay for silage or for deter- mining when to start raking high moisture hay, an error of f1" moisture #4 percentage points should be allowable. The above estimates of the allowable error in determining the moisture content of farm products are intended to include the sampling error. An exact indication of the moisture content of one sample does not necessarily tell the farmer the moisture content of the whole field of hay or the whole lot of grain. Rapidity Since the moisture content of agricultural products is so nonauni- form, a farm moisture meter should be rapid enough to take a large number of samples within the length of time which a farmer would want to spend on moisture determination. This is particularly true of hay where non-uniformity of the moisture content is the greatest. It is.doubtful that many farmers would be willing to spend longer than fifteen minutes in determining the moisture content of a field of hay; therefore, no more than one to two minutes should be required to determine use mois- ture content of one sample. Simplicity of Operation The ideal farm moisture meter might be described as a device which reads directly the average moisture content of any amount and any type of farm product when a probe is placed in the product. It is very doubtful that this ideal device will be realized for some time; however, there are number of features which may be incorporated in a farm mois- ture meter to keep its operation as simple as possible. The use of loose conversion charts in the field is undesinable. whflnever'possible, moisture measuring equipment should be calibrated vi directly in moisture content. In order to allow for different cali— brations for different products, a removable scale fze; made of a trans— parent material could be used for each product for which the meter is calibrated. Tuning or frequent standardization of electric meters is objectionable. The necessity for these operations should be avoided in the design of electric meters. Weighing samples, although not impossible, is not a very desirable Operation to be performed in the field. Arithmetic computation shouhi not be necessary to determine the moisture content from the indication of the equipment. Cest The most efficient of moisture meters will contribute little te the solution of thg,problem if the farmer does not buy it because it costs too much. ' It is very difficult to say what the typical farmer can afzgzd to pay for a moisture meter and it is even more difficult to say what he actually :11; pay for one. One might consider what he can afford to pay for a moisture meter in the light of the cash value of the craps he would test with it each year. In such a consideration, the cost of the meter might be considered as insurance against spoilage of grain. In this light, a cost of a few hundred dollars might not seem unreasonable for a large farm operation. From a more realistic viewpoint. one observes what a farmer will pay for a moisture meter? It is doubtful that a moisture meter costing the farmer more than fifty dollars would find wide acceptance in the near future. However, one small manufacturer is selling his output of electric farm moisture meters which have a retail cost of $135.00. Since the number of these meters in use is less than 500. the success of this manufacturer does not necessarily indicate the present farm market for moisture meters. However, the success of this manufacturer does give an indication of the future market for such equipment. For this reason, it does not seem desirable to limit the work of this study to methods which would sell for less than fifty dollars. In evaluating a method of moisture determination from the stand- point of cost, one should consider that most items of small equipment such_as moisture meters have a retail price of about four times the cost of labor and materials in manufacturing them. Versatility A farm moisture meter should be applicable to most farm uses with a minimum of necessary changes in the apparatus. Not only should it be applicable to the principal forage and hay crops, but it should also be capable of testing each crop in the ranges of moisture content which are of interest in the various harvesting and curing procedures. A Successful moisture meter should be. capable of determining the storage quality of grain which has been'recently dried as well as grain which has been recently harvested. A moisture meter should also be capable of giving a satisfactory test on forage crops of widely different moisture contents, from forage for silage at 60 to 75% moisture to hay for un- vent11a.ted storage at less than 25% moisture content. 47 INVESTIGATION OF EXHAUST OVEN METHODS A new type of exhaust oven was developed which is capable of rapidly drying forage and grain samples without. detrimental burning and without special attention from the operator. The exhaust oven method was tested in the field and in the laboratory to determine the following operating characteristics: 1. 2. Time to dry samples of various crops Drying temperatures at various tractor speeds and, exhaust temperatures Dry matter loss of the sample Collection of products of combustion by the sample Drying effectiveness Effect of scale accuracy Principle of Operation of the Exhaust Oven The first use of an engine exhaust for drying moisture determination 8amPlea was reported by Dexter (1). His oven, for which a patent was granted (2), consists of an Open-ended cylinder which contains the ”“1310 as the' exhaust gases are passed directly through it. The procedure outlined by Dexter includes selecting a representative ““919. weighing it in the cylinder on a platform-type spring scale, “tactung the oven to the exhaust pipe, and heating the sample until a °°nltqnt weight 1. attained. The initial and final net sample weights ar 9 use; to calculate the moisture content. 1+8 The sample temperature is measured by a thermometer and controlled at about lbOOC by adjusting the engine speed. To prevent burning during the heating process, the sample is turned end-for-end once per minute. Six minutes were required to dry a sample of mixed grass of 15.8% mois- ture content and ten minutes were required to dry a sample at 38% mois- ture content. Ten minutes were required to dry a sample of cats of 27.3% moisture content. Dexter published a diagram for the construction of the exhaust oven by farmers. The oven was later manufactured and marketed commercially. A similar design for an exhaust oven for home construction was reported by Hore and Ayers (3). They also developed plans for the home construction of a balance for use with this method. Tests of the Dexter exhaust oven showed that a successful moisture determination can be made with this oven; however, a considerable amount of attention from the operator is required to prevent burning of the Sample, especially when testing forages. Reversing the position of forage samples in the oven must be done frequently and attention must be given to the engine throttle setting in order to control the oven “aparature as indicated by a thermometer. These general observations °f the Dexter oven were confirmed by conferences with users of the device. 8'.iniplification of the procedure used with the exhaust oven method Seemed possible if a means could be found to reduce the temperature of ”1° 63! passing through sample and, at least to some extent, to make ”1° t9-perature less dependent upon the throttle setting. The diagram 0f "sure 1 shows a new type of exhaust oven developed under this pro- J°cts An injector pump assembly (00100111? “110d 8 "JG" PWP) 1' 2,9 7.9604 ,Wllll 8:62.00 mEEomll TI ml. N. a a _ a Q t lifimuhi .. Ills. .Zm>0 Fmadfxw mOkodm... mn;._. $05.09.? D h a QEaa cmosom 53.2.. m 33:01 poo; cmosom cums. - w m llltTNitIn all: ._ mmDoC 552:. T BE. :4 m M / .25 m8 3:25 50 located between the engine exhaust and the sample container, which is similar to that used by the Dexter oven. The engine exhaust gas is directed into the nozzle of the injector pump where its velocity is increased by a reduction in the cross—sectional area of the passage.. The high-velocity jet of exhaust gases thusly directed into the throat of the pump induces a flow of ambient air which is mixed with the exe haust gas and passed thrdugh the sample container. As a result, the mixture flowing through the sample is lower in temperature and flows at a higher rate than if the exhaust gas alone were passed thrqwgi the sample. Both of these characteristics are desirable for obtaining rapid drying of the sample without burning. Kravath (h) described applications of venturi ejectors where air under pressure is used as the primary fluid passing through the jet and the secondary fluid, the fluid to be transported, is also air or pre- dominantly so. He presented experimental and theoretical analyses of a low pressure air ejector of the type which might be used to withdraw large volumes of air containing corrosive or errosive materials from factory space. This type of ejector is driven by outdoor air from a centrifugal fan. Since the ejector described by Kravath operates princi- pally on air at low pressures, the analyses presented is partially appli- cable to the design of an injector-type exhaust oven. Further analyses and design data are presented by Keenan (5) and Xastner (6) for air ejectors with driving pressures from 20 psi. upward. 51 Apparatus Figure 2 shows the injector-type exhaust oven installed on a farm tractor. The muffler of the tractor has been removed to make the sample container more accessible. Connection of the oven to the exhaust pipe was accomplished in the test model by means of a tapered fitting. This allows a degree of versatility in the application of the oven to various tractors with different exhaust pipe diameters. The oven can be attach- ed to the tractor by placing the tapered fitting over the exhaust pipe and pushing downward firmly with a twisting motion. This arrangement is similar to that used on a commercial model of the Dexter exhaust oven. Figure 3 is a photograph of the basic equipment necessary to deter- mine the moisture content of hay and grain with the ejector-type exhaust oven. Shown with the ejector assembly are a long sample container for testing hay, a short container for testing grain, and a Hanson dietetic scale. Wooden handles have been provided on the sample containers to allow handling them without the use of gloves. The sample container for_hay is enclosed at either end with 8-mesh screen to confine the sample and the sample container for grain is similarly enclosed with 16-mesh screen. Both sample containers are made from sheet aluminum in order to keep their weight at a minimum and thereby maintain better weighing accuracy. The Hanson dietetic scale was used in the field tests of this apparatus. An O'haus laboratory balance was used in some field tests and in all laboratory tests to check the accuracy of the less costly Hanson scale.' The rated sensitivity of the Hanson dietetic scale is 1.0 gram. The rated sensitivity of the O'haus scale is 0.1 gram. 52 Fig. 9. The injector-type exhaust oven installed on a farm tractor. \ Fig. 3. Left to right: The injector—type OVen, the long sample container for hay. the short sample container for grain. and the Hanson dietetic scale. 53 Figure u shows the apparatus used in testing the performance of the exhaust oven in the laboratory. Iron-constantan thermocouples were used to measure temperatures at various points in the oven and sample container. These thermocouples were connected to the potentimeter by means of a manual selector switch. A tachometer was used to measure the speed of the engines. Test Procedure The injector-type exhaust oven was tested by two general procedures: by field tests where the Operation of the method was observed under farm conditions and by laboratory tests where more complete instrumentation could be employed . 3421.5; Tgst Prggedurg: Tests of the ex'haust'oven applied to forage test- ing were conducted using the following procedure: 1. Representative samples were selected by making up each sample from several small “pinches“ of foragegathered from several locations which seemed representative of the field. ’ 2. The empty container and lid were placed on the scale and the Scale was balanced or the rotating dial was set to zero when the Hanson 8Gale was used. The scales remained so adjusted until the conclusion of the test. 3. The samples were folded and cut with a knife into two or three len€ths to fit the length of the sample container. Lt. 150 to 200 grams of the sample were placed in the container and weighed. The net weight of the sample was indicated directly by placing the container and' sample on the scales. 51+ F18. 4. Apparatus for testing the perfomztmics of the exhaust oven in the laboratory. ‘J \h (n In the field tests of the exhaust oven, numerous recording; of heating time and oven temperature were also made. The oven was given further field trials by Hr. Bruce Hopkins who used it throughout the summer season of 1953 as a guide for his field operations in forage processing research. WW: Direct determination of the accuracy of this method is difficult. particularly in the case of forages. One cannot readily employ the usual methods of taking duplicate samples for testing by standardized oven methods. since two samples of forage taken simultaneously and adjacently in the field will vary considerably in moisture content as shown in other work of this project; For this reason. the-eva1uation of the accuracy of this method was based upon separate determinations of the effects of various factors which could conceivably bring about errors 'in the moisture determinations. These factors are enumerated as follows: 1. Incomplete drying of thesanmle 2. Loss of dry matter 3. Deposit of combustion products on the sample 1:. Weighing inaccuracy The extent to which the tractor exhaust oven is capable of drying ”19 Sample was determined by running standard laboratory oven tests on 8a'mples previously heated by the exhaust oven. Any further loss in weight in the sample during the laboratory‘oven test was assumed to be r°31<111a1 moisture content that the exhaust oven did not remove. Hay SamPies were heated at 220°}I for 1+8 hours in a laboratory air oven. G ra111 samples were heated at 220°? for 96 hours in an air oven. One 56 series of hay samples was also heated in a vacuum oven for 2h hours at 212°F. Dry matter loss could taken place through mechanical loss of leaves or kernels and by decomposition due to high temperatures. Mechanical losses can be directly observed and prevented by the proper selection of screen in the sample chamber. Forages are generally considered to be more subject to dry matter loss due to decomposition at high temperatures than are grains. To evaluate the extent of this loss. protein and sugar analyses were made on a series of forage samples which had been dried in the exhaust oven. Similar analyses were made on undried samples and the results were compared. Further qualitative evaluation of dry matter loss was made during field and laboratory tests by observations of color Changes and other signs of burning the sample. The weight of the deposit of combustion products on the sample in the exhaust oven was determined by means of a simple tests.. One-hun— dred grams of Ho.'00 steel wool was placed in the long sample container, heated in the oven for three minutes and weighed. This period of heat- ing was intended to drive off any moisture entrained in the steel wool. The steel wool and sample container were then placed in the oven and ”eiehed again at 5. 10. and 15 minutes later. An increase in weight observed was assumed to be due to the products of combustion. To simi- late an extreme condition such as a very smoky‘ exhaust, the carburetor choke was engaged until the engine mis—fired badly and was visibly smok- ing at the exhaust. The steel wool was again placed in the oven and r°"91ghed after 10 and 20 minutes. During a large number of the above laboratory tests. duplicate 57 weight determinations were made. first with the Hanson dietetic scale and secondly with the Ohaus triple-beam laboratory balance. At approxi- mately every tenth weight determination, the Hanson scale was standards ized at a point within 50 grams of the weight to be measured. This was done by means of standard scale weights and was done in addition to the initial zero adjustment of the scale. The corresponding weight deter- minations made by the Hanson and Ohaus scales were then statistically analyzed to determine the standard error of estimate of the Hanson scale in comparison to the Ohaus scale. Test Results ‘1;31g_231;g .application of the new exhaust oven in the field demons strated that moisture determinations could be made on forage in the field with minimal attention required of the operator during the drying process. 'It is possible for the operator to go about other chores such as machinery lubrication while the sample is drying. Errors in determining the moisture content of a field of hay by this method are most likely to occur from nonprepresentative sampling. There is a tendency for operators to select their sample from but one or two locations and not to test enough samples. Inaccuracies in weighing in the field can be caused by wind and poor adjustment and failure to level the scale. The aforementioned application of the oven by Mr. Hopkins for guiding forage processing research required that the method be accurate and rapid enough to follow the drying progress of cut forage in the 50 field. Moisture data from the oven method was used as a guide for raking and baling operations. After baling. core samples of the hay were taken and tested in a laboratory oven. Agreement between the field and laboratory determinations was generally satisfactory. Use of the exhaust oven in future research on this project is anticipated. égpgggpgzy_gg§13, Results of some of the laboratory tests of the exhaust , oven are shown in Figures 5 to 8. In these curves. temperatures and sample weights are plotted against time from the start of drying. The points at which the temperatures were taken are shown in the inserted diagram of Figure 5. The subscript designations of temperatures are consistent through Figure 8. These tests were conducted using a Case Model VAC tractor which has a rated engine speed of 1425,rgP-m- and ‘ displacement of 231 cubic inches. The data of Figure 5 was taken.with the engine running at 750 r.p.m.. a normal idling speed. The drying was considered complete at the end of 9 minutes. The maximum sample temperature was 255°I. A temperature difference (T1 - T2) of approximately 350°? was observed between the exhaust gas temperature T1 and the temperature T2 of the exhaust gas and air entering the sample. This temperature difference is a measure of the effectiveness of the indector in reducing the sample temperature and increasing the air flow through the sample. Figure 6 shows data taken with.a sample similar to that of Figure 5. The tractor was run at 1250 r;p.m. or at-about two-thirds of full throttle. The drying time was reduced to five minutes. The sample temperature did not exceed 310°F and the temperature difference (T2 - T1) was approximately uoo°r. LL 0 fl Temperature flare 5 700. 600 500 400 300 200 IOO 59 EXHAUST OVEN PERFORMANCE Engine Speed — 750 RPM. n” Second Cutting Alfalfa 'L-J ‘ M.C. 51.8% 1 I75 _ :50 o 125 S (D :5" ,4 [00 .9 CD a .S.’ r 75 3' C U) r 50 p 25 l l l l O W 2 4 6 8 IO Tame, Minutes 60 WW EXHAUST OVEN PERFORMANCE ‘w E- Englne Speed -l250 RPM ,J L I i Second Cutting Alfalfa- M.C. 5|.8°/o Tl T2 T3 T4 . 800 F l n ___V__—_.O 700 L. T..—Mi ° o d l g I50 “- U) 0 r l25 g 9’ e 3 s E.- 2’. q IOO g E B a 2 o 75 ‘5‘ U (D _ 5O '00 - _ 25 O 1 l l l 1 O O I 2 3 4 5 6 Time, Minutes ¥-/- /\ o Fmaqxxm a: Ego; 00m 0 o ‘ O FmDo Hm3m 0.2 o\o On 0? O¢ on On — p q q _ Rainmém .393 :33 335% eEoem-e:o:< @5230 3803 N5 0 .252. $5.92 mmahgos. :3: Nu earn. \\ a: I \ \ \ \ om NON u .m.m Sm. ...> ON m. -0. In. low now now mm sfiunpoaa Jetew HDH 6M3 lo afioianv V represents the moisture content of the sample by oven determination. The standard error of estimate was determined for the above data and found to be 2.07 Hart meter scale units. Figure .23 presents the same experimental data shown in Figure 22 except that all of the meter readings were plotted individually instead of the average of five readings per sample. This data was analyzed in the same manner as that in Figure 22, A straight line with the following equation was fitted to the data by the Method of Least Squares: ;,~ 2 0.628 + 3.0 The standard error of estimate for this data is 3.11 meter scale units. Figure 24 presents data taken with the Delmhorst meter on alfalfa- brome hay. The average of five readings per sample are plotted against the oven-determined moisture content for that sample. .The data is not linear throughout its full range and is probably parabolic'at the upper range. In order to facilitate comparison with other data having a linear characteristic in the range below 4035 moisture content, the data of I"ig‘llre 214» was divided and analyzed in two sections, that above 40% and that below 140%. I From the data for samples below 14070 moisture content, the following regression line was determined: 2’: 1.17x - 8.6 The stSandard error of estimate for this data was th meter scale units. For samples above 14033 moisture content, the equation of the regression line was: at: 0.6011: + 13.6 Th e B"-‘artidard error of estimate for this data was 5.0 meter scale units. 102 :35 foaouonom ram 02 o\o me 0v mm Om mm ON 0. O 4 a _ new. tmmém .263 :33 3358 3:95 - 3.3.4 9.530 ecooom Nuao $002 KNEE). meHQO—z .531 mu ensur— q 41 q __.m "mm on + xmmm. u» 1mm 10m mm buupoaa Jaiaw ,uoH 103 Fig. 214. Tests of the Delmhorst meter on second—cutting alfalfa-brome hey (Test 35‘). \ o \ W N \ \ \ _ 0 ° to 5 o (D - to > «S O 1' .: X 4 o (D _ o 3‘ 0' <1- " o ‘— )_ O C o\ 3 C . 8 > - s V c L 3 ‘— .‘L’ O _. o 2 N s ‘9. _ °° <1- ; u a _O_ N [LI 2 a) II )— co 0 o o o g o In c '0 N sbu,posa 1943“] 0M5 40 sbmuv 101; An increase in the standard error of estimate‘at the higher moisture range is typical of resistance-type moisture measuring methods. This can by attributed to the decrease in the change of resistance with moisture content at higher moisture contents and to a greater possibility of un- even drying of the hay in the windrow in the early stages of the drying process. A summary of the results obtained in trials of the Hart and Delmhorst meters on hay are shown in Table VI. This data indicates a generally lower standard error of estimate for the Hart meter than the Delmhorst meter; however. the methods of sampling with the two meters were very different. The sample which was oven-tested in the tests of the Hart meter was small and much of it had actually been contacted by the meter. In the Delmhorst tests. the oven samples were much larger and very little 01' each one had actually been contacted by the meter. 121’ tests of both meters. an average of five meter readings was a m°r e accurate determination of the moisture content of the oven sample than was an individual meter reading. More than five meter readings per Sample would probably give still more accurate results. In. order to evaluate the extent of the variability of the moisture content at various points in a windrow. the moisture content of hay samples which were taken in tests of the Delmhorst meter were compared with the moisture content of corresponding samples taken at the same time in tests of the Hart meter. The resulting pairs of samples originated from points 1°“ than one feet apart in the windrow. The samples from the Delmhorst t "We Were designated Group A and the samples from the Hart tests were (1 9‘1mted Group B. .Cx/ U .2038 0.3338 desagopedlnouro 23 .8» awnacmoa Hence 03% mo omdaosrm on» we engage 0.33 «3.38 heaps Had . essence 0.28.368 assessment... $96 on» pudwemm mwdadmma umuoa Handguns.“ em consensus as: same age _.. sin 0.0 + £3.01 mwmnfl .. _ .. m .m m; + awmma u a menu: e332. ._ CA 0.3+ H85: ensue: .. .. _. H5 06 I as; «h $343 maoamlmmamuq pmsomaaoa em.m N2: u Nommd «a when? a a a ma mega + 03.0 "a “3.2 $33... .. Kim o.m + Named "gm fiudlmm .. .. .. Him mwé + N536 n.» finanwm oaonmdmaegq «new Hosea 2H3 scammeamom owned @333 mo soauesdm magmas: hem some: E rlnlllllll H> "H.349 a no gamma manna“ adamHoz fluommzzoo 20mm 4949 3.0 29m 100 Oven-determined moisture data from pairs of hay samples taken simultane- ously were compared by plotting the moisture content of each sample in Group A against the moisture content of each correslionding sample of Group B. The samples in Group A ranged from 75.1 to 341:.1 grams in weight. while Group B ranged from 2h.2 to 70.9 grams in weight. The Group B samples were those taken in the above tests of the Hart meter. Data from samples of alfalfa are plotted in Figure 25. The equation of the line fitted to this data is: b: 1.108. + 1.2 The standard error of estimate between the moisture content of samples in Group B with respect to corresponding samples in Group A, is 3.38 moisture percentage points. Similar data from 25 pairs of alfalfapbrome 391313168 has a standard error of estimate of 14.147 moisture percentage points. The data taken in the tests of the Hart and Delmhorst meters was taken over a period of 1310 or more days for each group of data such as that .hown on each of the curves Just presented. Data taking was resumed ”Ch morning as soon as most of the dew had disappeared or at a time when a farmer might reasonably expect to start beling hay. It was expected that the readings taken first in the morning might be unusually high in comparison to oven determinations due to the presence of surface moisture. However. examination of the data with regard to the time of day taken did n , 0t reveal any significant number of high meter readings which were not i :1 8“(39111 with the oven results. 107 Adaaonov :30 305.630... xm 05.. 0\0 ON 9 Om. m¢ Ov mm Om m N . J a . a a mam. .3 -mm 3.63 39653 2.: E Zone 82 see when 5 33032.36 $3.2 3.3.3 8.3.4 no o\\ 22.50 056.85 of no comtanoo \\\ $.3an mm.mu.mm \ ~._+ 00:2 fi On (a anJS) uaAo “woman-1 K8 0w °/. 9v 108 Studies of the resistance method applied to grains included deter- mination of the resistance-pressure characteristics of two types of electrode design. and an evaluation of the reproducibility of two types of constant-spacing electrodes. Comparison of the moisture readings of a commercial meter intended for use on farms was also made with reference to air oven determination of moisture. Laboratory Investigations on Experimental Designs W. A test cell. shown in l'igure 26. was constructed for con- ducting resistance tests on grain. The cell consisted of a steel cylin- der lined with acrylic plastic tubing (Lucite) for insulation. A steel Plunger 1%- inches in diameter was fitted to the inside of the plastic tubing. The bottom of the plunger served as one electrode while the ”991 plate at the bottom of the cylinder served as the other. {the cell as shown in Figure 26 was used first for tests with vari- able electrode spacing. The cell was later modified for tests with con- stant electrode spacing. For these tests. two brass electrodes were attached to the bottom of the plunger as shown in Figure 27. Resistance was measured with the same calibrated Tag-Heppenstall resistance meter used in the work on hay. Pressure was applied to the tag ' t cell by means of a small hydraulic press equipped with a Bourdon- tub 63 gauge eight inches in diameter. Tests were also conducted using the Delmhorst Model 28E electrode 109 /- PLASTIC ELECTRODE |,i._.—_.r .— Fig. 27. Constant-spacing electrodes F13. 26. Test cell for resis- for grain test cell. tance tests on grain. for hay and grain (Figure 3))and a modification of this device (Figure 28). "“1111 cation of the Model 28E probeyconsisted of replacement of the electrode pins with two aluminum plates as shown in Figure 28. The plates Iare Parallel to each other and spaced 1 1/8 inches apart. Each plate is 1 1n<=h wide and protrudes into the sample 5/8 inches. Penetration into the sample by the electrodes is controlled by the Same plastic guard plate used with the standard probe. The sample is contained in a standard No. 8 grain sample can while being tested, as “ho"n in Figure 20. Pressure is applied to the sample by hand. The e orreCt pressure is indicated by the compression of a spring in the handle w hich allows the handle to telescope to a red mark. The resulting pres- s lire was measured and found to be 25 pounds. 110 Modification of the Delmhorst Model 283 probe for resis- Fig. 28. tance tests on hay and. grain. 2%. In tests with the experimental test cell. a weighed quantity of corn was placed in the cell and pressure applied with the hydraulic press. Seven gram samples were used for the tests with variable electrode Spacing and 25 gram samples for the tests with constant electrode spacing. I'} 18ctrdcal resistance readings were taken immediately after the desired pressure level was reached and one minute afterward. Constant pressure "8‘3 Carefully maintained during this time. fiter tests with the standard probe. the samples were tested with the 11'iCDdii‘Iied probe in a similar manner. The moisture contents of all Bamples was finally determined by heating for 96 hours'at 212°F. in an 8.11s oven. 111 Figure 29 shows the relation between electrical resistance and agenggg. pressure applied to the experimental test cell, lied and the seven-gram sample was comln‘essed. The electrode Spacing “It . varied as presume was egg, Similar data is presented in Figure 30 from tests of the experimental test cell with the constant electrode spacing. Both sets of curves are reasonably regular and tend to level off at the higher pressures. indicating a decreasing rate of change of resis- tance as pressure increases. There is slightly less change in resistance with pressure with the constant-spacing electrodes. The same data as that plotted in Figures 29 and 30 is presented in Fig'lJ-re ,31. wherein electrical resistance isplotted against moisture content with certain pressures as parameters. Moisture-resistance char- acteristics withvariable electrode Spacing and 980 pounds pressure applied is quite linear when resistance is plotted on a logarithmic scale. This indicates that moisture content varies directly as the logarithm of resis- t‘itllcze at a constant applied pressure. The moisture—resistance charac- teristics at lower constant pressures deviate more from linearity at the lower pressures. The curves plotted from data. taken with constant elec- trode Spacing are somewhat similar to those plotted for variable electrode spacing; however, the non—linearity persists to much higher applied pres- sureS- This non-linearity could be due to differences in the areas of the two types of electrodes used. The constant-spacing electrodes were smaller and would likely be more affected by variations in the actual area of contact with the individual kernels of corn. Table VII contains resistance data taken with thevstandard Delmhorst Dr ' Obe and with the modified probe shown in Figure 28. This data is 112 Fig. 29. Electrode Pressure vs. Electrical Resistance of Corn Measured in the Test Cell with Variable—Spacing Llectrodes. IOO Ijjln l |.2 °/o M.C. r7 1 U) E .C C) CD 0 2 0 {‘5’ l3.9 ‘7.» MC. ‘5 o 'as {E LO: :. . _ I527. MC. 0 _ O l6.0°/ . i 0.: r _ . MC. — . ‘0 0.05 ' l I L J o 200 400 600 300 nooo . Electrode Pressure , Pounds 113 Fig. 30. Electrode pressure vs. electrical resistance of corn measured IOO 'Irrr1 I in the test cell with the constant—spacing electrodes. 0 O l l.2 °/o MC. I42 °/o M.C. '0‘ IO :- g a, 8. 14.4 °/. MC. 0 2 m . 0 C I! .— E» g; ’ l6i£¥96 “‘13. a LG? E l6.8 °/. M.C. p- ().l l I 1 J lJ O 200 400 600 800 IOOO Electrode Pressure , Pounds 114 rig. 31 Electrical resistance vs. moisture content at constant electrode pressures. IOOE L t p F' \ lOt— \ in r \‘ \1- E .. \\ \ 3 ~ \\ x a: - \\ +98|bs. a: \+ 2 - \{A ' \ \r 3 _ \ \ 245 lbs. : . \ l g \ 490 lbs. 1’ l.O— \ E 3 +980 lbs. :' 98 lbs. _ 245 lbs. ' 490 lbs. O.l — Z. 980 lbs. :_ +-_-+ Constant Electrode Spacing _. .__‘4, Variable Electrode Spacing 0.0l I t 1 I J 8.5 I0.0 - l2.5 l5.0 l7.5 20.0 °/o Moisture Content 115 It. oo.m o.m o.mn o.Hn o.am H:.H o.N o.m: o.ms o.ms mm.N m.s m.a: 0.4: 0.5: am.o mm.ou mm.m o.n m.u ma.a m.mu m.ma o.m: o.am ma.e m.mn m.ms o.as o.nm mm.a Ne.ai mm.n a.» o.o an.o mm.ou mm.m m.m o.m use: scam new: Scam oonssuauom : .02 m .em nouumabmn aofipdabon can mo pudendum adaaxmz use: wagons: «umqaudsm ooqdpuauom ohspudo: c an c an o.ma g o.Hm o.ms m.ma : o.Hm o.w: o.#a = m.a m.a m.nH cadence: 0.0: o.mm m.ma s o.m¢ o.ud 0.:H : m.m n.e m.ma . m.m m.m m.mH ensequpm N .om H .02 m .snosnoo spasm aoHadloHQoz.MHah4 mad HMOhHm Hmomm Nam ANGUS ammomzafin mmH mHH3 amude amoo 20mm 4949 Hondymmmfim fibueHammfim Hm» mnmda 116 intended to compare the reproducibility of the two probes. Resistance rest dings for a. given sample seem to be more constant with the modified probe which has the two plates as electrodes instead of the two row of pins - Standard deviations from the mean are consistently higher with the pin-type electrodes when the data are compared at similar moisture c011t ents. Test of a Commercial Moisture Meter on Grain W. A Delmhorst Model RC-l moisture meter with a Model 28E elec- trode assembly was tested. The meter, probe and sample container are shovrn in Figure 20. W m Khan. Iagta, Wheat samples used in Test A were drawn at a commtry elevator from lots of wheat being sold by farmers. These samples varied greatly in variety, quality. and cleanliness (weed seed content). The mo at common weed seeds observed in those samples were co‘ckle and (Black grass seed. Records were kept of the relative amounts of foreign material in the samples. However. no significant effect of the foreign matter on the reading of the meter could be observed. Samples were tested with the Delmhorst meter immediately after being drawn. Five readings were taken from each sample. The Ballplefi were weighed immediately after testing to determine their wet weight for the OVen moisture content determination. The samples were then held in closed conmariners until the moisture content could be determined. conveniently in 8' 1e~‘boratory air oven. Test B was conducted with samples held in closed containers for 1+0 to ’48 hours. Samples in this test included the samples tested in T8815 A Plus additional samples which had been collected from farm harvests in Huron County. Michigan. Immediately after testing with the Delmhorst meter. the samples were reweighed for determination of their moisture con— tent by the oven method. Sample temperatures were also determined. In Test C all of the samples of Test B were held in open containers 118 for six hours in an oven at 130° F. The results of. these tests are in- tended to show the effect of artificial heated-air drying upon the in— dication of the meter. As in the previous tests. after the meter readings were taken, the samples were reweighed to determine their moisture con- tent by the oven. method. Sample temperatures were again taken. The samples were tested after they had reached the approximate room temperature of 80°F. Following Test C. the samples were dried in a laboratory air oven for 96 hours at 212°F. W £91. £0.23 lam. Samples for the first test on corn (Test 1)) were drawn from approximately 1/2 bushel of corn selected at random from corn being harvested. The corn was shelled and thoroughly mixed before six samples were selected by dividing with a Boomer sainpler. The samples were allowed to dry in Open containers in the laboratory where the temp- erature was 75°F. and the relative humidity was approximately 25%. Determination of moisture by the Delmhorst meter. weight reading. and sample temperatures were determined occasionally as the samples dried from an original moisture content of 3233 to 12% at the time of the meter reading. O Five separate meter readings were taken from each sample. The samples were mixed between readings by shaking. immediately following each meter determination. the samples were weighed in order to determine their mois- ture contents at the time of testing by a later oven determination. Test 3 was conducted to demonstrate the advisability of grinding corn samples. particularly under abnormal sample conditions. Two groups of samples were tested. Group I consisted of eight samples of corn with 119 an initial moisture content of approximately 227°. This corn had been rewet from an original moistureycontent of 114% and allowed to equalize in moisture content in a closed container for two weeks before the test. Group II consisted of five suples of corn having a moisture content of approximately 32%. This corn had a perceptible coating of mold on the kernels. . During Test E. the samples were dried at room air conditions which war-e approximately 80°F. and 20% relative humidity. Occasional readings with the Delmhorst meter on the whole grain were taken at intervals during the drying period. Five repetitive readings were taken on each sample with the sample being agitated by shaking between each reading. Imrn ediately following meter readings the samples were weighed and ‘ tanperatures were taken. At intervals between. the beginning and end of the drying period. 0118 Sample at a time was withdrawn from each group, tested with the meter, ground in a coffee mill. and finally retested with the meter. After grinding. the samples were weighed before placing in the oven for “Oi-Stine determination. Grinding was' done with a coffee mill which re- duced the grain to a size of which approximately 757% would pass through a standard 8 mesh screen. W. The results of all tests on grain are summarized in Table . VIII 3 which shows the equations of the regression line and the standard error of estimate for each test. ‘ The data obtained in Test A on wheat is shown in Figure 32. The average of five meter readings per sample are plotted against oven de- termined moisture content. These results demonstrate the ability of the Reading 3 0 Ole l2 120 its. 32. fists st Delmhorst lots:- os inst samples has st s mtg elevator (hot H. Y=I.OSX— 5.5 o ‘o/ 9 S.E.=O.65 ° 9 / L I I J I I IO N I2 l3 l4 I5 % MoisIure Content 121 in!!! Ill-I. .dqsonm one: moan: AHHVa and “Ham «page eons .oaaagm .noaaaa. eoqoasaeaoo aaaaaoamassq maasaoawupas nauseous «sagas. m woos .wqwuuop ouomop undom m: o» on new uuonasusoo decode nu damn noaaasm a I .moaz .uoaomo .oo nopspoaa assume as daoa madam noon: scum some» uoaassm (I. .Ill.‘ll amasaemi :m.a o.mm u Nmm.m » wmnma ma shoe AHH.HV m a .aoamaso adds: m:.m m.mm I wmm.m H mmnom mm M naoo Annvn Ham.o s.NN ..u-.~ » mm.ma N: H auoo lava .deaoobudn as nhoo mm.a m.w I_K:m.a M «muma w: shoe 9 .eoaue mm.o ao.a ummw.o » aa.m mm sacs: o ma.o oa.o wmoaao » mauaa an «son: m no.0 min I amo.a w. manma ea neon: 4 oussdunfi mafia soamuouwom asooaom «madman . causes no hogan on. no souessam .owqdm no .on mono cues cusp-«o: eauequsm 23¢ azHBmfiB 2H mafia: ammoan a ho mmoHadmuamEH ho 9.55% a .mo gm k HHH> flqm L . J 1 I” .IQ; f ..i.. .R’ I 3993 :l‘ n i p. a '1 4.7+ 04., ~54 Qt J 4.... H040 , t 0.. e ran . 0 cash.“ . r, aMNHQ. JWIrMfie. 40 t ’ V S. . 2, J 59‘ “JPH.u qunmfl. f 1 Card .39 momma maniacs samenmoscpmfinon manganese on» mo moofipomm ~3qu .wm .muw aw: i. a‘\ r‘\\\-w Lh' ‘ 5' -v--. 0-. ‘I 11¢ ‘11“ 136 ' electrodes (2?) which is at the lower-most‘position is connected to the resistance meter by means of the following circuit: conductor (3). to collector bar (14). to brush (6). to brush holder liner (7). to conductor (9). to the meter- One terminal of the resistance meter is connected to the frame of the device by the grounded conductor (23). The brush holder liner (7) is insulated from the brush holder (10) by an insulating sleeve (’20). A spring (8) holds the brush against the collector. A removable guard plate (13). attached with bolts (18) to the bearing assembly (11). is provided to protect the electrical wiring inside the wheel from mechan- ical damage. Any direct-reading resistance-measuring circuit calibrated. in mois- ture content may be used with the averaging device; however. the action of the galvanometer movement used must be heavily damped in order to main- tain a reasonably steady average reading as resistances of different values are connected to it by the averaging device. The well—known series ohmmetor circuit with its essential features is shown in Figure 39, The resistance to be measured. 3;. is connected in series with the galvanometer; an adjustable series current-limiting re- sistor. Ba; and the voltage source. V. A resistor. R.. is placed in parallel with the galvanometer to produce electrodynanio damping. 3, places an electricalload on the galvanometer coil which absorbs enery from it and prevents excessive oscillation of the needle. The current through the measured resistance. at! in the ohmmeter circuit of l'igure 39 is expressed by the following relation. V . . I 'm ‘5’ t where s. is the sun of all the resistances measured from the terminals ’4 RS fa U Golvanometer <1 JD Ra WW Fig. 5. Simple series ohmmeter circuit with damping resistor. Neon Tube ”IF U .——+ t——4 Fig. 6. Relaxation oscillator circuit for resistance measure eat. \7 1.38 of D.. with the terminals of V shorted; i.e.. the sum of It. and the parallel combination of R. and the galvanometer coil resistance. In order for the relation of Equation (14) to hold true for en ohm- meter circuit. B. must be negligibly small in comparison to Rt° This condition is not difficult to obtain in ohnmeters for high resistances such as are used for moisture measurement. However. in most of the com- mon laboratory or portable ohmmeters. 30 is not much smaller than 3t. since the value of n. is selected such that if the test leads are short- ed (it a O). the galvanometer deflects to full scale. This not only safeguards the galvanometer in case of accidental shorting of the test' leads. but also provides a convenient means of standardizing the circuit by adjusting the value of R.. The adjustment is frequently necessary in ohmmeter circuits in order-to allow for variation in the battery voltage. Another type of resistance measuring circuit applicable to the averag- ing device is the reluation oscillator circuit which has been described fully by Suits and Dunlop (1). In the circuit of Figure 1&0.the resistance to be measured. 3. is placed in series with a capacitor. 0. and a battery voltage. I. Across the capacitor is placed a neon glow lamp which has an ignition or starting voltage. 7. of about 80 volts. The battery vol- tags. m. is usually more than 5013 greater than the igiition voltage. v. When the circuit is completed by the measured resistance. a. the voltage across the capacihrs, e. increases with time..t. according to the rela- tionship. 0 = 3 (1 - 5‘ 't/Bc) (7) when)! equals the natural logarithmic base. 139 The operation of the relaxation oscillator circuit is explained graphically with the aid of Figurehl. The oscillation of the voltage across the capacitor and the lamp is represented by the cycle passing through the points 31. Di. 32. D2. 33. D3. etc.. When the voltage across the capacitor and the neon lamp reaches the value of the ignition voltage of the lamp. V. (Point 31). the lamp "fires“ and emits light. At the same time the internal resistance of the lamp drops to a very low value and quickly drains the electrical charge from the capacitor until the capacitor voltage dreps to the extinguishing voltage of the lamp. Y-a (Point 1);). At this voltage. the lamp ceases to glow and again becomes a very high resistance so it is again possible for the capacitor to re- charge according to Iquation (7) from Point D1 to 32. When the capacitor has charged to Point 32 the lmnp fires again and the capacitor voltage drops to D2. The above process repeats periodically and an oscillation occurs. Each time the lamp fires. a flash of light is produced. The rate of flashing gives a visual indication of the value of the measured resistance. R. The time elapse between flashes or the period ef the oscillation is expressed by the following relation: r . xlnc ' (8) where II is a constant of proportionality dependent chiefly upon the battery voltage. the ignition voltage of the leap. and the extinguishing voltage of the Imp. The frequency of the flashing of the lap is: : =3.le (9) T R where 12 is the reciprocal of ale. Thus. Equation (9) indicates a re- sponse of the relaxation oscillator circuit which is similar to the Vans ’ Copochor Voitaje 1L0 Fig. b1. Operating cycle of the relaxation oscillator I40 circuit. :5 _________ [20. R: LO Megohm C= 0.5 Microforod IOO- _Y_ _ - _ _ __B' ___B.2__ 83......83.-- _ 80 F ,o _V:9________ ___ .__._. .___ __ J.- D D D 60~ ' 2 3 D“ 40-» 20m- of __ J l J 4 J O 0.2 0.4 0.6 0.8 LO Time , Seconds ° 141 response of the ohmmeter circuit describedvby Equation (h). A plot of logarithm R 2. logarithm f is linear. In the application of the relaxation oscillator circuit to the aver- aging device. the operator would run the averaging device continuously for a measured length of time, e.g.. one minute. The number of flashes of the neon lamp observed during this time is an indication of the aver- age moisture content of the material tested. The operation of the circuit when connected to an averaging device can be readily explained by means of an example. A relaxation oscillator circuit having the following typical circuit constants is connected to an averaging device: Supply voltage. I. =3 135v Starting voltage of the lamp. V 8 85v Extinguishing voltage of the lamp. v-a :- 67v Circuit capacitance = 0.5 microfarad The circuit is connected to an averaging device such as that of l'igure 35 which contacts in sequence three resistances at the rate of 0.05 second per resistance. The values of the resistances are 31 c 1.0 megohms. 32 a 2.0 megohms. and R3 = 3.0 negohm. , While the circuit is in contact with any one of the resistors. the capacitor charges according to the relation of Equation (7). The relation;- ships between capacitor voltage and time for each of the three resistances are illustrated in Figure 1+2. Assuming that the capacitor has already been charged to 67v. the extinguishing voltage of the lamp. and that 21 has Just been contacted. the capacitor will charge for 0.05 second along the charging curve of 21. At the end of this time. 82 is contacted for 0.05 4L 1b? Fig. 42. An operating cycle of z. relaxation oscillator circuit used with the averaging device. I40 - I20 - {3 IOO— O > ' _-¥ _________ 8 2" 80 — 7 ‘ " " ’ ““““ _ :‘r‘.;:f:::. —6 ~0- - -— - — - - 4 5 E V-o | Z ’ 3 ------- e i 60 — a go“ § § “8: 9‘4 0 8 4o — r" 6'!" o Q~ If C=0.5 microfarod 4' '. 20 — O l r n 1 1 1 4 O 0.2 0.4 0.6 0.8 LO [.2 L4 Time , Seconds 143 seconds and the capacitor is charged for the same length of time accord- ing to the“ curve of Re. After R3 has similarly been contacted. R1 is again contacted and the capacitor charges to the firing voltage of 85 volts. At this time. the lasp fires and the capacitor discharges to 67 volts. the extinguishing voltage of the lamp. Since 32 is the next re— sistance to be contacted. the second charging cycle will start with He and proceed in a manner similar to the first cycle. The time required to charge the capacitor to the starting voltage of the lamp during the cycle indicated on Iigure 42 is 0.1 x 1r=.0..l+ sec.... The flashing frequency corresponding to this period is l/OJt :- 2.5 flashes per second. In the charging cycle Just described and outlined en Figure 1&2. it is apparent that the rate of charging of the capacitor was not as great dur- ing the second period of time El was contacted as it was during the first period of contact. This is due to a lower charging rate at higher ‘ voltages. Thus. it is apparent that whether a certain resistance is con- nected to the circuit early or late in the charging cycle has some bear- ing on its effect on the charging of the capacitor and the period of the circuit oscillation. However. it is very doubtful that this apparent source of error would be important where either a very large number of resistances of assorted values are contacted or where a small number of resistances are contacted in sequence over a period of time of one minute or more. Of greater importance in the application the relaxation oscillator circuit is the effect of fluctuation in the supply voltage upon the ac- curacy of resistance measurement. According; to information by Suits and 1141+ Dunlap (1). the circuit is not affected by changes in supply voltage un- til the simply voltage drOps below about 1.5 times the starting voltage of the lamp. The current drain on batteries supplying this type of cir- cuit is small; however. aging alone will reduce the terminal voltage of batteries. Therefore. some means should be provided in a meter of this type to indicate. when the battery voltage is too low. This could readily be done by providing an accurate resistor which could be placed in the circuit for. purposes of checking the accuracy by the substitution method. It is recognised that in such applications as in testing windrowed hay in the field. the rather dim flashes of light from the neon lamp would not be clearly visible in bright sunlight. An audible indication of the flashing of the neonlsunp can be obtained by connecting an ear- phone into the circuit in series with the neon lamp. A distant “thump' is heard each time the lamp flashes. The earphones for this purpose must be of relatively low impedance. e.g.. 2000 ohms. Most of the in- expensive electromagnetic types of earphones are suitable. Laboratory experience with the relaxation oscillatorcircuit has shown that it is easier for the operator to count accurately at high rates from an audible signal than from a visual signal. It is also easier to watch the timing device while counting from the audible signals than to attempt to watch both the timing device and the flashing .lamp simultaneously. 145 Messiihemmm Both field and laboratory tests were performed on the averaging de- vice for windrowed hay. The most desirable procedure would have been to conduct the laboratory tests before the field tests; however. the con- ception of the idea for the averaging device came very late in the 1953 haying season. In order .to avoid delaying the field tests until another year. the laboratory tests were delayed so that the field tests could be run while field hay was available. Field Te at 8 Two field tests of hay were tested with the averaging device for windrowed hay. one of alfalfa-brome mixed hay and another of practically pure alfalfa. The crop in both fields was from the second cutting and was very light in yield due to very dry weather. mustus. In the field tests. the wheel-type averaging device pictured in I'igure 30 was used and resistance readings were taken with a Tag-Hep- penstall Model No. 8003-T1l resistance meter. the sane motor which is used with the Tag-Heppenstall resistance-type moisture meter for grain. The movement of this motor is highly damped. Additional damping can be provided if needed by pressing a button on the meter face. The circuit of the Tag-Hoppenstall meter is basically that of a series ohmmeter with a low internal meter resistance. 30. (see Equation 6). Therefore. the response of the meter to changes in resistance is hyperbolic: i.e.. log- arithm of the meter deflection _v_s_.. logarithm of the measured resistance is linear. light ranges of resistance measurement are available on the meter. The overall range of the motor as determined from a previous 146 laboratory calibration is 1500 ohms to 100 megohms. l’rocedure. Approximately 75 feet of windrow was used for each test run of the apparatus. . Three marker stakes were layed out at points approx- imately l/h. 1/2. and 3/1l of the length of windrow to be tested. While running the wheel over the windrow. the operator read off aloud the meter reading as the wheel passed the marker stakes. These readings were re- corded by an assistant. Immediately after each test run. the operator also recorded an estimate of the average meter reading during all of the ‘ test run. As soon as possible after each test run with the wheel. cross-sections were cut out of the windrow at the location of the marker stakes to serve as sanples for moisture determinations by air oven methods (212°? for 118 hours). while cutting the cross-sections. the windrow was clamped be- tween two 5/8 x 2-1/2 x 14' boards mich were hinged together at one end. . A vertical cross-section of the windrow about 2-1/2 inches thick was cut with a hedge shears. The samples so obtained were placed in paper bags and weighed immediately in the field on a laboratory balance. Weighing was done in an enclosed trailer to avoid scale errors due to wind. level- ing. and other factors. Remit s. Oven-determined moisture contents of the samples cut from the windrow have been plotted in Figure 43 against the logarithm of resis- tance readings recorded as the wheel passed the point where the samples were taken. The data from the alfalfa and alfalfa-brome are plotted to- gether and identified. It is apparent from the data of figure 1&3 that the relation between moisture content and logarithm of resistance is not linear throughout the range of moisture contents tested. The direction I“? :25 .__< E 23:00 «.5562 oxo Ob om on oe om ON 0. H \d _ d _ _ yd e\o had "Md emm + 2 mmvmodum mo... o\om\..N u.w.m manmu o :o:<+ 0:33 . Field data from the averaging resistance-type moisture meter for hay. Fig. #3. a: r . mm.¢+zm_omo.o..&oo._ 0. 0. 0. K) on - Q ¢ sumo”). ‘ oouoisgsaa ;o wq4g40601 .148 of the curvature of the relation at the higher moisture contents suggests that it might be linear if plotted as logarithm moisture content 13.10- garithm resistance. A plot of this sort was made in analyzing the data and a somewhat greater degree of curvature in the opposite direction was found in the relation. Therefore. it was decided to divide the data in two groups. that above I40$ and that below ”0%. assuming linearity within these two ranges. The equation of the regression line for the data below 140% moisture content was determined by the Method of Least Squares and found to be as follows: . log a = 0.08016 14 + M955 (10) The standard error of estimate of the logarithm of resistance was 0.237. In terms of moisture content. the standard error was 2.79 moisture per- centage points (0.237/0.08016 = 2.79). The equation of the regression line for the data above W moisture content is as follows: log a =- 0.03h53 u + 2.936 (11) The standard error of estimate for this data is 6.57 moisture percentage points. The error of the moisture measurement with the wheel-type meter seems to be much greater in the higher moisture range. This is to be e1:b pected from all resistance-type meters. since the variation of the loga— rithm of the resistance with moisture content is less at higher moisture contents. It is also possible that in the early stages of drying in the windrow. the moisture content may be more non-uniform than in the later stages of drying. thereby increasing the sampling error involved in 149 Figure Luv, shows the average of the oven-determined moisture contents of the groups of samples taken following each run of the wheel plotted against the estimated average reading of the meter recorded by the opera- tor after the corresponding run. The average of the moisture contents of the samples taken after each run probably better represents the moie- ture content of the windrow.and the hay tested by the meter than do the single ample moisture contents of Figure 1:3 represent the ha from which. the individual meter readings were t aken. The equation of the regression line for the data below W moisture log n .. 0.0929 n + 5.38 (12) The standard error of estimate for this data is 2.11 in terms of oven- determined moisture percentage points. The equation of the regression line for the data above W moisture content is: log a = 0.014356 :1 + 3.1a (13) The standard error of estimate for this data was 3.22 moisture percentage points. Indicated in the right-hand scale of Figure 141: are the ranges of the eight scales on the Tag-Happenstall meter. Since, all of the scales over- lap below 20 scale divisions. only the ranges from 20 to 60 scale divi- sions are shown for each scale. The span of the meter is sixty scale divisions. Oompari sons of the data plotted in Figure #3 and Figure “Labor that the error of estimate of the meter is much less when the meter readings are plotted against the average of three simultaneously-determined samples. 150 Fig. 44. Field data from averaging resistance-type moisture meter for hay. Logarithm resistance vs. the average moisture content of three oven samples. 00 03 o |_ I I l l l I I l I <3 f~ 3 . -0 ION ' +a4 to 2m . 3|: ° :3“! 0." -° 0 In I u t on o -J _0 q- _0 (0 e) . -0 {3- * N 4-; g =. N g .. 0 LL! o as '- n. I: O 3. 1 1 L I $.53 0. Q 0. 0. 0° e- F) N _ sumo”): ‘eouotsgseg 1° wq4g40501 % Moisture Content by Air Oven 151 This suggests that much of the apparent error of the meter is due to sampling error in obtaining the oven samples. The standard deviation of the average of the moisture contents of each group of samples was deter- mined. The average of these standard deviations was found to be 11.75 moisture percentage points for all the alfalfa-brome samples. 2.77.for the alfalfa samples below 1409!: moisture content. and 3.15 for the alfalfa samples above 140% moisture content. These values are not intended to show quantitatively the extent of the sampling error in these tests; how- ever. the need is indicated for taking a much greater number of samples per run of the wheel in future tests of this device. Laboratory Tests Laboratory tests of the averaging device were conducted to demon- strate its effectiveness in determining the average conductivity of known resistances and to substantiate the theoretical analysis of its operation. Apparatus 33d Procedure. The averaging device for windrowed hay was mounted in a fixed position as shown in rigure 1+3. The wheel was belt- driven by an electric motor operating through a hydraulic speed regulator. By adjusting the speed of the drive. a complete range of practical ground speeds could be similated. Resistors of assorted values were attached to the electrode sets to similate the testing of hay samples of various moisture contents with the averaging device. During all of the tests. the same Tag-Eeppenstall resistance meter used in the field tests was connected in the usual manner to the averaging device. On some of the test s. resistance readings were also taken with the ohmmeter section of a Simpson Hodel260 portable ohmmeter and with a relaxation oscillator circuit manufactured by the Shafer Instrument Compaw of Loudenville. Fig. 1&5. Arrangement of equipnent for the laboratory tests of the averaging device. 1'13. 1+6- Components of the relaxation oscillator circuit of the shafer moisture meter. The earphone and the test leads with clips were added to the meter by this project. 153 Ohio. The circuit of the Shafer meter is similar to that shown in Figure no. The voltage supply is 135v (battery source); the value of the capaci- tor is 0.5 microfarad; and the neon-tube is a General Electric Type 1m- 36 with a starting voltage of 8h volts. An inside view of the relaxation oscillator resistance meter is shown in Figure 56. This meter is mar- keted commercially for use by farmers for the moisture testing of grains. Results. Table IX shows the arrangement of resistances connected to the 21 electrode sets on the wheel during Test A. The resistancevalues shown were determined with the Tag-Happenstall resistance meter after the resistances were connected to the electrodes. Conductance values have been calculated and are shown in Table n. Also shown in ”Table Ix ~ are mutual conductance values which are the values of conductance con-' nected to the meter during the time of transition between one collector bar and another when the brush is in contact with two electrode sets simultaneously. The mutual conductance is calculated as the sum of the conductances of adjacent resistors. Entries of‘mutual conductance in Table I)! are for a parallel combination of the resistor connected to the indicated electrOde number and that of the next electrode; e.g.. the mutual conductance opposite No. ’4 is equal to the sum of the conduct- ances of No. 11 and No. 5. The width of the brush contact area was about 25$ of the center to center distance between the collector bars- Therefore. in determining average conductances which the meter should indicate. the conductances of the connected resistors should be given a weighted value of three and the mutual conductances should be given a weighted value of unity. The average conductance. Ta. is a follows: TABLE IX LOCATION AND VALUES 01‘ TEST misrmons (TEST A) 154 WI Electrode Connected Uonnected Mutual 5” "°' 3323:??? °§fi§2°la€83 53:2“???- 1 7.8 133.3 1h1.h 2 128 8.1 38.1: 3 32.5 30.3 32.0 n 620 1.7 13.3 5' 92 11.6 12.3 6 1750 0.7 5.0 7 260 11.3 11.14 8 10.000 0.1 166.8 9 5.8 ' 166.7 173.5 10 150 6.9 36.3 11 36 29.11 31.0 12 7110 1.6 11.8 13 108 10.2 10.8 11h 21100 0.6 5.1 15‘ 250 14.5 11.6 . 16 10.000 0.1 1.14 17 950 1.3 12.9 18 90 11.6 12.2 19 2000 0.6 6.2 20 192 5.6 5.7 21 10.000 0.1 1 .u '2 YIn 155 Ya=-fi%—( BZY +2Ym = k (114) = Til-H (3 x h29.3 + 859.6) 2 25.56 x 10"3 mhos Re a m3—mfiifi = 39.12 Kilolnns or an indication of 26 on the '3' range of the Tag-Hoppenstall meter. A summary of the data from Test A is given in Table X.. Two obser- vations were made with the Tag-Heppenstall meter. one with the normal damping in the meter movement (designated "undamped') and one observation with extra damping introduced by depressing the Idamp" button on the face of the meter. This button introduces additional resistance in parallel with the micrometer. The meter readings are given in meter scale units with the meter range designation indicated as a prefix to the scale read- ings; e.g.. an indication of 26 scale units on the '3' range is indicated as 326. The scale units are proportional to the current through the sample. . The predicted meter reading as calculated using Equation (11+) is also shown in Table 1 for comparison with the observed values. The ob- served values indicated are a range of readings, representing the lower- most and uppermost travel of the needle during the test. As should be expected. the range of the observed readings is less with the additional damping than with only normal damping. The range of the readings also decreases with an increase in the speed of the wheel or the number of resistances contacted per unit of time. Comparison of the calculated and observed values of Table X is most favorable at the highest speed and with the greatest damping. A somewhat erroneous conclusion might be drawn that the average of the TABLE X 156 SUMMALY OF KATA FROM TEST A 05‘ Trill AVERRGING DIVIC‘; —‘ m Simpson Ohmmeter Peripheral Tag-Heppenstall Meter wheel Readings in Scale Units Readings. Kilohms‘: Speed Observed *' Calcui Observed‘_" Calculated __m.p.h. Undamped Damped lated _' verage Average ___ 1 33-38 313-33 326 h5-1500 87 151 1.5 312-36 319-31 60-1200 113 2 313-33 321-28 60-h5o 106 2.5 317-31 323-27 70—700 127 3 BIZ—28 BBM—Z? 120-h20 187 157 uppermost and lowermost readings of the meter at a given speed should agree closely with the predicted reading and that the average of these two observed scale readings should not vary with the speed of the wheel. This is not true. since the average of the uppermost and lowermost read- ings of the meter at a given speed is not a time average. as is the calcu- lated value. The observations do not reveal the proportionate length of time the meter needle dwelled on the highest or lowest reading. Com- parison of the observed and calculated values is most favorable at high speeds because the needle is steadier at higher speeds and does not indicate readings of short time duration which are widely variant from the average. The observed readings taken with the Simpson ohmmeter were taken immediately after the readings with the Tag-Happenstall meter. Since the Simpson meter is not highly damped. the ranges of the observed read- ings are much.greater. The averages of the observed readings with the Simpson meter shown in Table x are not arithmetic means of the resis— tances but are the reciprocals of the average of the conductances indi- cated. For example. the average of the observed readings at 3 m.p.h. are indicated as follows: Average conductance = _%_. (1%6. + 3&5) = 0.00535 x 10"3 mhos. Observed average resistance = "5—06535'_’ = 187 kilohms. The calculated average resistance value was determined by a proce- dure similar to that used for the Tag-Happenstall readings and explained previously except that the internal resistance of the meter was taken into consideration. This was not necessary with the Tag-Happenstall meter because its internal resistance is negligibly low in comparison to the measure resistances. This is not the case with the Simpson meter since its internal resistance is 118 kilohms. In calculating the average indicated resistance, the internal resistance is added to the observed . value of the individual resistors before the values of connected con- ductance and mutual conductance are detemined by the same procedure used with the Tag-Heppenstall meter. The average resistance as calculated by Equation (114) must be adjusted to the indicated average resistance by subtracting the internal resistance of the meter. The special procedure indicated above for analyzing the operation of the Simpson meter on the averaging device is necessary because the meter movement indicates the average conductance in' the whole meter cir- cuit and not Just that of the resistances connected to its terminals. Also for this reason the indicated values of average resistance of Table X are not the same for the Simpson meter as they are for the Tag-Hep— penstall meter, even though they were Operating on the same resistors. The Simpson meter does not indicate a value of resistance corresponding to the average conductivity of the resistances tested. However. it is , possible by the above procedures to predict within experimental error the reading of either of the meters on a given set of resistances. Further laboratory‘testing of the averaging device was conducted in Test 3 using the resistance values indicated in Table XL. In-pre- vious tests of the averaging device, a copperéimpregnated carbon. brush was used. In Test B. a brush was made from a piece Of 5/16 in. x 3/14 in. aluminum flat bar stock. In order to reduce the shorting effect as the brush passes from one collector bar to another. a blunt edge was ground on the brush so that the width of the area contacted on the collector . LOCATION AND VALUES 0]? TEST ‘RESISTANCES IN TEST B ~- ODD-NUMBEIED RESISTANCBS WERE DISCONLEICTED FOR TEST C - TABLEXI 159 W 5"” “°' 311328383 3:23“??? 3223??? 1 510 1.960 35.858 2 29.5 33.898 37.898 3 250 n.000 5.613 n 620 1.613 31.025 5 1 . 3n 29.h12 3h.967 6 i 180 5.555 6.963 7 710 1.h08 30.820 8 31+ 29.1112 311.620 9 I 192 5.208 6.712 10 665 1.5014 31.807 11 33 30.303 3u.651 12 230 M3148 5. 776 13 700 1.u28 31.731 1h 33 30.303 3h.h70 15 aka n.167 5.575 16 710 1.1t08 30.820 17 3h 29.h12 33.288 18 258 3.875 5- 337 19 510 1.961 31.373 20 3h 29.1412 35. 2911 21 170 '553f§§§*=1§ r fié‘én x. TABLE XI I 160 sum 0:? DATA roe Tess B m) c or THE mmmmrm AVERAGING DEVICE Peripheral TEHeppenstall Meter Readings in Scale Units Test Speed, served Talculated m_.p. h. "aidinped - giggled Ave rajgg 3 1 3111-22 1316—21 chi I 1.5 oh5-60 0&9-57 I 2 gnu 39M~ ' 2.5 036-38 037 ' 3 036.5 -c37.5 037 0 1 0111-32 018.5-26 ‘ 022.5 " 1-5 018-29 020.5-211 I 2 019-26 021-23 . v 2. 5 ' 019-23 020. 5-21. 5 " 3 c19.5-2o.5 020.5-2‘1.5 ' 3.5 020.5-22 ‘ 021.5 161 was approximately 109-5 of the center-to—center spacing of the bars. The aluminum was hard enough not to wear badly when ground to so sharp a point, but was not so hard enough to cut the copper collector bars. Calculation of the predicted meter readings was done by the same method used in Test A. except that 10% brush coverage was asswned instead of 25%. The test“ data and the calculated results are shown in Table VI Agreement between the calculated and experimentally determined re— sults is somewhat more favorable at the 2 m.p.h. than at 3 m.p.h.,'which could be due to 'Jmnping" of the brush as it passed over ,the groove be— tween the collector bars. The sharp aluminum brush would. be more apt to do this than the flats-faced carbon brush. If this phenomenon did occur, it would mean that contact with the collector bars was not continuous and would explain the lower indicated conductivities which resulted. In Test 0. the resistors were disconnected from the odd-numbered electrode sets. so only ten resistors remained connected to the averaging device. Since there were no two adjacent commutator bars to which re- sistors were connected, there was no mutual conductance entering into the circuit when the brush passed between bars. Therefore, the predicted average reading was calculated as follows: Ya = l/R. = “‘1! (Y2 + m «OI-Y6 --- Ygo) (15) = ‘1‘!" (1111.329) = 6.730 micro-mhos. Re a 103 = 118.6 kilohms 6.730 x 10-5 This is a reading of 022.5 on the Tag-Happenstall meter and it agrees favorably with the experimental results shown in Table XI. In Tests D and E, a relaxation oscillator circuit for resistance 162 measurement was utilized as well as the Tag-Heppenstall meter. The calibration of the relaxation oscillator circuit is described by the following relation: 11: 11.22 (16). where R is the measured resistance in megohms and f is the number of the flashes observed per minute. Resistors were connected for Tests D and I as indicated in Table XIII. The calculation of the predicted meter readings was carried out as in Test 0. The observed meter readings at various speeds of the averaging device are tabulated in' Table XIV. The data observed from the Shefer meter shows that there is little or no dependence on speed of the reading of this meter when used with the averaging device. This is a reasonable result, since there is no inertia involved in the indications of this meter as there is with e' D'Arsonvel galvanometer movement. Each resistance contacted provides a current path for charging the capacitor. The amount of charge passing to the capacitor is irrespective of the value of the resistors Just previously contacted. With the D'Arsonval galvanometer, if a single resistance of a value much different from the value of those resistances Just preceed- ing it is contacted. the effect on the galvanometer reading will be much greater at low speeds than at high speeds. LOCATION AND VALUES OF THE TEST RESISTANCES IN TEST D AND E OF THE AVERAGING DEVICE TABLE XIII l 63 Electrode h"ctEnnected esznnected Gonnect :csf fonnect ed Set Resi stance, Conduct ence. Resi stance Conduct ence, No. Mggohns Micro-mhos flows Micro-mho s 2 1. 7 0. 612 1 1. 7 0. 612 h 10.1 0.099 10.1 0. 099 6 111. 5 0.069 1.18 0. 677 s 1.55 0.6h5 _1.55 ‘ 0.5n5 10 9. 2 0. 109 9. 20 0. 109 12 13. 8 0.073 1. m: 0. 69h 1!: 1.611 0.610 1.611 0.610 16 10.1 0.099 10. l 0. 099 18 12. 5 0.080 1.50 0.667 20 1.611 0.610 . 1.6!; 0. 610 6 16L; .. .. 2.: 2. .. .. 99.0: mmummn m .. .. .. 5.: t .. .. «ALT: $.on m .. R5 32 5.: E. R5 n.8— Nddd inning H a c : mm.» m.m: : a m.m1:.~ Hzlsmw n s . .. . mm c. m .9 .. .. m 6.06 938 N .. on. T? .91. 0.9 0.» Re m.mum.2 Flue H n 7 mamowew .cmaxame flaws: .cmBKammeHh mfiomm: madam macaw naflomozmficbtmamom wQPnghfirflb HUR>9HQMAHO gmmmgumb‘ .CQPHQWPO egegee ”$08 mandate canoes steam nousafiomo soapsficaom ammsnm magnum nope: Swansonmemlwma Heaonmaaom MOHPMQ 93.033154 .8 a 979. Q memma mom <94d mo ggm >HK age in all—- F 165 W The field test data of Figure M4 indicate that a standard error of estimate of 2.11 moisture percentage points applies to the averaging meter when testing windrowed hey between 20 and 140% moisture content. The standard error of. estimate of the average of five readings per sam- ple with the Delmhorst moisture meter was l$.10 moisture percentage points when testing hay from the ewe field at the same time. The standard error of estimate of five readings for sample with the Hart moisture meter was 2.8 moisture percentage points when testing hay from the same field at the sane time. ' In comparing the above results, one should examine the sampling pro- cedures in the tests of each meter in regard to what degree the oven san- ples taken were representative of the hay actually tested. The Hart meter tested only a small bunch of hay about one inch in diameter and this hay was used as the oven sample. This procedure is probably more favorable to the meter than either of the procedures used on the other meters, but it does not indicate} the ability of the meter to determine the average moisture content’of the field or the windrow. Tests of the Delmhorst meter involved taking a cross-section of the windrow where five readings'had been taken with the probe. This procedure is certainly less favorable in indicating the accuracy of the meter than that used with the Hart meter. since only a very small percentage of the oven samples had actually been contacted by the probe. The procedure for obtaining oven samples used in testing the averag- ing meter was probably the least favorable of all to the meter. Considering the high standard deviation of the average moisture content of simul- taneously drawn samples. the number of samples taken was inadequate to supply reliable information as to the average moisture content of the hey tested in one run by the averaging meter. Future tests of the averag- ing meter should be conducted taking at least ten samples from the win- drow per pass of the wheel. Laboratory tests of the averaging meter substantiate the initial theory that it indicates the average conductivity of the resistances connected to it. However, this is strictly true only if the internal resistance of the measuring circuit is negligible in comparison to the measured resistance. ‘Therefore. ohmmeters of the most common portable type cannot be used, since they have very high internal resistances. The relaxation oscillator circuit has been found quite applicable to the averaging device. This circuit is inexpensive in comparison to other resistance measuring circuits. . Its methods of operation would probably be as acceptable to farmers as a direct-reading ohmmeter. Use of the relaxation oscillator circuit requires the counting of flashes or audible signals during a reasonably accurately determined period of time; however, operation of the ohmmeter circuit requires taking a meter read.- ing from a pointer which fluctuates considerably during a run of the averaging device. This variation in reading is of the order of the varie- tion encountered in the Tag-Eeppenstall resistance-type moisture meter for grain. This meter is widely accepted for testing grain for market purposes. but a degree of Judgment 'is usually required to decide upon 7 the exact indication of the meter. When the relaxation oscillator cir- cuit is used on the averaging device for windrowed hey, the reading taken 167 is dependent upon the moisture content of all the hay contacted during perhaps a.minute of time. A reading taken with the ohmmeter circuit is either a Judgement of the average reading of the meter during the test period or a single reading taken on only a few feet of windrow. The mechanical design of the averaging device for windrowed hay was generally acceptable. except for one feature. The insulating segment be- tween the collector bars should be wider than the brush width so that shorting together of adjacent collector bars is avoided. The immediate practical value of the averaging meter for windrowed hay as a tool for farmers is somewhat open to question. principally be- cause of its high cost of construction. A less elaborate simplification of this device might include a.crank-operated rotary switch connected to several electrode sets attached to a.fixed member. the arrangement which is shown schematically in rigure 35. An averaging device in this form could readily find application on baled hay and in testing other’materials of non-uniform moisture content. The application of the averaging device to the process industries such as paper making (from straw). cotton.mill- ing. and tobacco manufacturing should be investigated. In these indus- tries where cost is not such an important consideration. the instrument *would probably find more immediate application. Literature Cited (1) Suits. 0. G. and Dunlap. H. 3.. Determination of the Moisture Content of Wood by Electrical Means, General Electric Review. V 3“. PP. 706-13 (1931) 168 PRELIMINARY INVESTIGATIONS 0F OTHER.METHODS OF MOISTURE MEASURMENT Investigations discussed in this chapter are those which.were con— ducted in the early stages of the project to gain general information on the applicability of several methods of moisture measurement to the prob— lem. The information and experience obtained were used as a guide in selecting the methods on which to concentrate.attention. Gas Pressure Method épnargfina, figure Lfl7shows the steel receptable and gauge for testing the moisture content of grain in relation to the amount of gas released when mixed.with a chemical such as calcium carbide or calcium hydride. The receptacle was constructed from standard 1% inch.pipe fittings. A gasket union.with a neoprene gasket was used. A.sma11 rubber disc of the type used as a safety valve on.pressure cookers was installed in the end of the receptacle to reduce the danger of explosions. A small piece of steel wool was installed in the gauge port to prevent dust from enter— ing the gauge.s The total displacement of the receptacle was approximately nine cubic inches. The Bourdonstube gauge shown has a pressure range of O to 60 psi. Another gauge of similar size which has a range of 0 to 15 psi. and a minimum scale subdivision of 0.25 psi. was also used in these tests. Ezgggdnze,' A weighed sample of grain was placed in one half of the re— ceptacle and a weighed quantity of chemical in the other. (Weighing the chemical is not altogether necessary if an excess is used). The two 169 .Tig. #7. "Test cell for trials of the gas pressure method on grains halves of the receptacle were then closed tightly as quickly as possible with the aid of a pipe vise and a wrench. Care was taken to prevent the sample and the chemical from mixing before the receptacle was tightly closed. In the early tests. shaking was continued throughout the test to mix the grain and the chemical. but it was found that continuous agitation did not aid the reaction after the first minute. Later samples were shaken only occasionally after the first minute. Frequent checks of the tightness of the seal on the receptacle were made by immersing the test cell excepting the gauge in water and watching for gas bubbles. This could be done only at the conclusion of a test be- cause the water would affect the temperature of the receptacle and comp sequently change the pressure of the gas. Data from the tests in which. 170 leaks were encountered were discarded. The corn used in these tests was ground to the consistency of com- mercial corn meal in a burr mill. The lots of corn at each moisture con— tent were divided after grinding by means of a Boerner sampler. One portion was used in the tests and the other was moisture tested in.an air oven. figggltl, Representative data from these tests are plotted in Figure “43 Data are plotted for three different moisture levels for 3.0 gram samples of corn and 3.0 grams of calcium carbide. A.curve is also plotted from data taken from a 3.0 gram sample of corn at 5.2% moisture content tested with 3.0 grams of calcium hydride. These data were also plotted on a logarithmic time scale and the relation between.pressure and the logarithm - of time was found to be linear within the range of the data. .Although the rate of pressure rise decreased. no true equilibrium point could be definitely indicated. During the extended-time test run at 1h.0% moisture content. the pressure increased one pound per square inch between 12 and 18 hours after the test was started. This pressure data was taken on a four—inch gauge with i pound divisions and a range of O to 15 psi. As was indicated in the literature. the pressure of the hydrogen produced by the calcium hydride was much greater than the pressure of acetylene produced by calcium carbide under similar conditions. The re- lative danger of handling hydrogen and acetylene in the quantities in, volved is about the same according to consultations with chemists. No explosions were experienced with either chemical in these tests. Fig. 48. Rate of pressure rise in a receptacle containing ground corn and calcium carbide or calcium hydride. O «'3 N 2). “2 3‘6 an _o 9': ‘9 C o; a -8 2 0 \° 0' ‘6. E 2 5 ID o 3 8 O I d ‘0. e- N . N e I 2 - a} O O . | 0 o\ e e N 0‘ - 0 o S 3 ‘0' o I | W" N . 0 6“ ° 0 U '0 0 to: —S_’ L: 1 I ‘ ‘ ' o to o In 0 ‘0 o N N —' "' '61'53 ‘ amsseJd 171 Time , Minutes 172 anclusions. The gas pressure method as applied in these tests is not very suitable as a method of moisture determination for grain because of the following disadvantages: 1. Excessive time required to reach an equilibrium point. 2. Possible hazard from fire and explosion. 3. Very accurate weighing of a small sample is required. 4. Maintenance of a gas-tight seal on the receptacle would probably be difficult for a farmer under field conditions. 5. The equipment cost of the method is rather high because an accurate scale and pressure gauge are required. 011 Distillation.Method W. The first of these tests was intended to determine the weight loss of corn oil (Mazola) during extended periods of heating. For this pumpose an ordinary porcelain enameled cooking pan.was heated over a gas flame manually regulated to maintain the desired temperature. Five hunp dred grams of oil were used in this test. Temperature was measured with e.mercury thermometer. The oil distillation tests on hay and grain.were conducted using a four-quart pressure cooker over the gas flame. The cooker was loosely covered during the tests at atmospheric pressure. During the tests under vacuum. the cooker was closed and attached to a vacuum pump as shown in Figure #9. IPressure was measured.with a Bourdonptype gauge and regulated manually by an air bleed valve. Temperatures were measured by a mercury thermometer. A.weighted screen.was required to keep the hay submerged in the oil. 173 Fig. u9, .Apparatus for dehydration tests in oil at below-atmospheric pressures. Zrnggflurg. The stability test on corn oil was conducted by heating the oil at 212°?. for thirty minutes, 300°F. for thirty minutes. and 350°}. for thirty minutes. The oil was weighed occasionally during these tests. In the oil distillation tests on hay and grain. the container. oil. and thermometer were weighed at the beginning of each test. The oil was then heated to 325°F. and a weighed sample was introduced. The apparatus and the sample were weighed at intervals until the weight became constant. The loss in weight of the apparatus and the sample was assumed to be the moisture loss of the sample. A vacuum of twenty inches of mercury was maintained during the vacuum tests. This degree of vacuum was selected because it is normally available at the exhaust manifold of a tractor. 17a Results. The weight loss of the five—hundred gram sample of oil during the entire stability test was 0.h grams or 0.008%. This apparent loss is well within the range of expected experimental error in.weighing the oil and container. Observation of the rate of weight loss of alfalfa and corn samples when heated in oil are shown in Figure 50. The corn samples reached the equilibrium condition much more slowly than alfalfa hay. This cannot necessarily be considered typical of alfalfa in the field, since these samples were of rather light-stemmed hay and had been frozen in storage. The freezing may have damaged the cell structure of the hay so as to make it more easily dried. The use of vacuum pressures in the heating increased the drying rate as indicated by the curves of Figure 50. 'Qggglggigng, Stability tests on.Mazola oil indicate that a negligible weight loss of the oil occurs during the usual periods of heating which are required in the moisture testing of farm products by this method. Therefore. this method for moisture testing, which requires the assump- tion that the weight loss of sample. container, and oil is equivalent to the absolute moisture content. is not subject to errors due to weight loss of the oil. Drying grain and hay samples in oil is much faster than drying in air as in the exhaust oven. The drying process can be further hastened by the use of vacuum pressures. No further work was conducted on the method because of the following disadvantages in comparision to the exhaust oven method: 1. This method requires the weighing of a considerable quantity of Haul Original of °/o Weight , in corn oil at 395 F, . 50. Ra?e of weight loss of cor: and alfalfa samples heated IOO ! | 90 _Atmospheric Pressure i " CORN . . O - . 80 ' 20in. Hq.Vocuurn 70-- 60 '- Atmospheric Pressure 50 - . ALFALFA _ 20 in Hg. Vacuum 40— 30 ' I I 1 J 0 5 IO IS 20 25 Ti me , Minutes 176 oil in addition to the sample and the sample container. Since the accus racy of most scales decreases with the mass weighed, the net weight of the sample is determined less accurately by the distillation method than with the exhaust oven method. 2. Working with the distillation method is objectionable to the operator. since the handling of hot oil is necessary. However, the equip- ment can be design to reduce this objection. 3. The distillation method requires essentially the same equipment as the xhaust oven plus a supply of 011. Specific Gravity Tests éapeggfigg, A small brass wire basket was constructed for weighing a 25 gram sample under water. Weighing was done with a balance with a minimum scale division of 0.01 gram. Zzgggdurg, The 25 gram corn samples were first weighed in air. then im- mersed in water and reweighed. The samples were divided by means of a Boerner sampler. Half of each sample was used in the specific gravity tests and the other half was tested for moisture in an air oven. figfiultg, Weight in water of a 25 gram sample of corn is plotted against moisture content in Figure 51.. The water temperature was 70°F. A rela- tionship between moisture content and weight in.water (or specific gravity) was indicated; however, it does not seem very reproducible. The change in weight while the samples were suspended in water was observed and found to be negligible. Weight in Water ,Grams Fig. 51. 'v.’ei;;ht in water vs. moisture content of 95 gram samples of corn. 3.75" 31H3r 3.25 '- 3.00 - 2.75?- 2.50 - 2.25 I l l J 10 I5 20 25 3O % Moisture Contenf 178 Conclusions. The observations of this test indicate that the Specific gravity method does not show great promise as a method of moisture de- termination for grain due to its lack of reproducibility. The method also requires very accurate weighing and would not be applicable to hay or to some of the small grains such as oats which.would absorb water rapidly. Mechanical Shear Studies on Bay The objective of this work was to evaluate the relation between' moisture content and the shearing force required to cut alfalfa stems with a Tenderometer. This machine. shown in Figure 52, is a commercial device used in the canning industry for the numerical evaluation.of the tenderness of meats, fruits. and vegetables in terms of the force re- quired to shear them. Appargtna, The test cell in the Tenderometer consists of two sets of sparallel plates of l/B—inch in thickness and spaced l/8-inch apart. When the two sets of plates are brought together they mesh and cut the contents of the test cell into l/8—inch.pieces. The force to shear the sample is automatically applied and registered on a dial in terms of numbers proportional to the actual shearing force. Ezgggdnrg, The hay used in these tests were stripped of leaves and cut to h—inch lengths to fit the Tenderometer test cell. The moisture COD! tent of the test samples was assumed equal to the oven-determined mois-‘ ture contents of the hay mass from which they were taken. The samples were divided into three size classifications. 179 Fig. 52. The Tenderometer used in the mechanical shear studies on hay 180 Resultg. The data obtained from this test are indicated in Table XV. Although certain trends are indicated by the data, it is doubtful that this sort of measurement could be made a :eliable indication of the mois- ture content of hay. Generally, the data indicate that the higher the moisture content. the lower the force required to shear hay stems. How— ever. other variables such as stem size seem to affect the data greatly. Past experience of the Department of Horticulture with this machine in testing snapdragon stems. has indicated a very definite effect of the origin of the sample with respect to height on the growing plant, regard- less of stalk size. There were also indications of the effect of this variable in this test on hay. anglusigns. The shearing force required to cut hay stalks is affected by other factors besides moisture content. A moisture meter operating on this principle would have to be designed to take into consideration stem size and location of the test point along the stem. Maturity of the crop would also very likely affect the shearing force—moisture re- lationship. TABLE XV DATA FROM TENDEROMETEB TESTS ON ALFALEA STEMS Mm-a- Moisture no. of -Relative Tenderometer Content Stalks Stalk Scale (Percent) Tested Size Reading 61.2 2 Small 57 61.2 2 Small 53 61. 2 2 Medium 77 61.2 2 Large 175 43.7 2 Small 105 h3.7 2 Medium 105 #3.? 2 Medium 97 “3.7 2 Large 1u2 “3.7 2 Large 88 lh.3 2 Small 150 1n.3 2 Medium 158 1h.3 b Small Over 250* 14.3 2 Medium Over 250"I * Maximum reading of the machine was exceeded. 0 182 Wet and Dry—bulb Thermometer Method The modified wet and dry—bulb thermometer technique developed by Dexter for determining the moisture Content or storage quality of grains was investigated. The operatinr principles of this method were discussed in the Review of Literature. Dexter has shown thrt this method works satisfactorily on laboratory samples of grain. i.e.. samples of grain which have been held in a closed containers for extended periods of time. AdCitional work was undertaken by the author with the purpose of investigatin: the effects of recent drying and foreign matter content of the sample upon the accuracy of the method. A commercial device operat- ing on the wet and dry-bulb principle was also investigated.to determine its applicability to testing grains on the farm. Apparatu§. The apparatus for employing the Dexter wet and dry-bulb thermometer technique included an ordinary square-type glass milk bottle wrapped with an inner layer of p inch of glass wool insulation and an outer wrap of paper. Two mercury thermometers with a mimimum scale di- vision of 10F. were used and read with the aid of a reading glass. The wick on the wet—bulb thermometer was made of cotton tubing available commercially for this purpose. The thermometers were mounted in a rubter stOpper. A saturated sodium chloride solution was used in the wet bulb. The Master Moisture Meter, manufactured by the C. H. Baldwin Company of Lansing, Michigan, is shown in Figure 53. The base of the device, shown at left in the photograph, contains an electric motor with a rated speed of 1550 r.p.m. The sample container (center) fits on top of the Fig. 53. The Master Moisture Meter. 9. commercial moisture meter operating on the wet and dry-bulb principle. 183 I I]; (will, 184 base and stirrer blades in the bottom of the container are connected directly to the motor. A dry-bulb and a wet-bulb thermometer are attached to the lid of the samnle container bv means of rubber grommetS. - :1 Results. Early trials of the Baldwin tester revealed that the temperatures of the two thermometers did not reach a definite equilibrium point. This is apparently due to the fact that the electric motor continuously ir- a (J arts considerable energy to the sample in the form of mechanical work Iv and heat. The power input to the motor was found to be 100 watts. Table XVI shows temperature data from a typical test run with wheat at 13.1%'moisture content. Both the wet and dry-bulb temperatures increase steadily. but the wet bulb depression. which is the index of the moisture content or storage quality of the grain. does not stay constant. The ob- served decrease in temperature difference with time can be partialLy explained by the fact that equilibrium relative humidity of grain in_ creases with temperature at a given moisture content. which causes a lower wet bulb depression as the dry-bulb temperature increases. How— ever. the observed wet bulb depression is much greater than that observed in the original work by Dexter for similar changes in.dry-bulb temperature. A greater contributing factor to the decrease in dry-bulb temperature seems to be the grain dust which is produced in considerable quantities ly the high speed stirring blades. After approximately two minutes of operation. the wet bulb becomes noticably coasted with this dust. which no doubt reduces the rate of evaporation from the bulb and thereby re- duces the wet bulb depression. The milk bottle apparatus was used in tests on samples taken from lots of wheat being sold by farmers to the elevator company at Okemos. 185 TABLE XVI DATA FROM BALDWIN MOISTURE TESTER Time, ~ Dry—Bulb Wet-Bulb Wet-Bulb iinutes Temperature. Temperature, Depression, 0?. 0E. 0F. 0 72.5 79.5 0.0 1 74.2 73.0 1.2 2 75.2 7h.2 1.0 3 76.6 _ 75.6 1.0 5 t 79.2 78.1.L 0.8 7 80.4 81.0 0.6 9 82.? 82.6 0.4 Michigan. The samples were drawn from within the grain masses in the ‘wagons and tested immediately. The wheat was of different varieties and contained different amounts of foreign matter. Wet bulb depression plotted against moisture content in.Figure 5h (Group A) indicates a reproducible relationship between the wet bulb depression and the moisture content of the samples by an oven determinp ation. The variations in the results which did occur can be attributed largely to_experimental error in reading the temperatures. No effect of foreign matter in the wheat was observed. The data observed in this test agrees well with that presented by Dexter. particulany at moisture contents above 13;. The data designated as Group B in Figure 5; was obtained from wheat samples which were Spread in a thin layer and exposed to the air in the laboratory for a period of one hour before testing. The temperae ture in the laboratory was 85°F. and the relative humidity was uofl. This data indicates a greater wet—bulb depression at a given.moisture 186 Fig. 52. Data from tests of the wet and dry-bulb thermometer method on wheat. 7.— 6- + 5r Dry-Bulb Temperature 34 4.. 82-89°F. CF .9 u) .33 3- Q. Q) o E 2" a Group B-Portially dried 8 4- 3 IL- 0 L- Group A_- Freshly drown samples -I r- __ .0 J' 1 I I ll IO l2 l4 l6 IS 20 % Moi stu re Content 18? content in comparison to the data from Group A. This is very likely due to the drying experienced by the samples of Group B. Drying of the in— dividual kernels occurs first in the outer layers; therefore, air which is in moisture equilibrium with the outer layers will be lower in humidity than if it were in contact with the inner parts of the kernels. A.wet bulb depression is indicated which is too high and not an accurate meas- urement of the moisture content or storage quality of the grain. Grain samples which have abnormally dry kernel surfaces, such as might be drawn from a crop drying system, could be allowed to equalize in a closed container before testing, but this procedure requires several hours. Grinding the sample seems to be a solution to the problem; how- ever. when the samples were shaken in the milk bottle it was found that dust from the ground grain collected on the wet-bulb thermometer and reduced the evaporation from it. A design intended to allow the testing of ground grain without ap- preciable dust collection on the wet-bulb thermometer is shown.in Figure 55. The objective in this design is to pass air through the grain at low enough velocity that no appreciable dust is entrained and yet to pass the air over the wet bulb at a sufficiently high.velocity to sup courage evaponation. In the design presented. air is recirculated through the grain by means of an aspirator bulb. The air is injected into the grain mass at the bottom of the container through a perforated tube. The air passes through the grain at a very low velocity because of the large area of the cross—section of the can. The tube through.which the air leaves the can surrounds the wet bulb so that the velocity of the air passing over the salt solution on the wet bulb is high. .. I {III-Illa." 188 THERMOMETE * THERMOMETER >14 CHECK VALVE ASPIRATOR ' BLADDER GROUND GRAIN PERFORATED TUBE ,Fig. .55. Experimental wet and dry-bulb thermometer apparatus for testing ground grrain. 189 Tests with this device indicate that it reaches an equilibrium temperature much faster on both whole and ground grain than the milk bottle apparatus. particularly when high wet—bulb depressions are en- countered. Equilibrium has been reached in two minutes with wet-bulb depressions of 8°F. Since the wet-bulb temperature changes more rapidly, the equilibrium point is more easily determined. This feature would be important to a non—skilled operator. There was no noticeable amount of dust in the air in the can or collection of dust on the wet bulb when testing ground grain. This was true even for very finely ground wheat. As in the case of the milk'bottle apparatus. no measurable rise in div—bulb temperature occurred during teats with the experimental apparatus. Apparently the transfer of heat from the operator's hand on the aspirator bulb to the air was not significant under the conditions of the test. hble XVI shows data from a test-with the experimental dry and wet- bulb device cf Figure 55. A sample of wheat at 11.1% moisture content (determined by even at the conclusion of the test) was tested the first time immediately after being removed from a tight can where it had been stored for several months. The sample was then spread in a thin layer and allowed to dry at room conditions for a period of 20 minutes. The data for this test indicates a much higher wet-bulb depression than was observed before the grain was dried. The sample was then finely ground and tested again. The wet-bulb depression decreased to 6.5OF.. which is comparable to that observed before the sample was dried. Dexter also observed that the wet—bulb depression of ground grain was slightly lower with ground grain than with whole grain, particularly at lower moisture TABLE XVI 190 DATA FROM TEST OF THE EXPERIMENTAL WET AND DRY-BULB APPARATUS OF FIGURE 53 USING A SINGLE SAMPLE or WHEAT AT 11.1% MOISTURE CONTENT f Time. Dry-bulb Wet-bulb Wet-bulb minutes Temperature, Temperature. Depression Remarks °F °F ’F O 82 77.5 h.5 Whole wheat o . 5 82 74 . 5 7 . 5 1 82 714.5 7.5 2 §§:_ 7415 715, c-1- 0 80 77.5 2.5 Whole wheat 0.5 80 72.5 7.5 dryed in room 1 80 70.5 9.5 air for 30 min. 2 80 69.5 10.5 __.g1 _ A- 80 69.5 $1035 78 75.5 2.5 Ground fine 0.5 78 73 5 1 78 72 6 2 78 71.5 6.5 3 78 71.5 6.5 191 contentS. W. The equilibrium wet-bulb depression when using the Baldwin tester is difficult to observe. If an equilibrium condition does exist, it is not maintained for long. Furthermore. the fact that both.wet and dry—bulb temperatures are changing continually makes the equilibrium depression difficult to observe. The wet and dry-bulb method in general does not seem to be greatly affected by the presence of reasonable quantities of foreign matter in the sample. The method appears to be accurate for whole grain which has not been recently subjected to drying air. ‘Under such conditions grinding the sample seems very desirable. The experimental apparatus successfully tests ground grain without depositing dust on the wet bulb. Since the air velocity over the wet bulb is higher, this apparatus reaches temper- ature equilibrium more rapidly than the more simple milk bottle apparatus. 192 SUMMARY AND CONCLUSIONS Exhaust Oven Method The exhaust oven method of moisture determination is the most ver- satile of all the methods studied. It is applicable to all crops in any condition and at practicalLy any moisture content. Since the results of this method are not greatly affected by factors other than moisture con- tent. conversion charts are not necessary. The principle of the method can be readily understood by nonptechnical personnel; therefore. errors caused by mis-use or misunderstanding of the method should not ordinarily occur. The cost of the equipment for the exhaust oven method should not be prohibitive. The complete apparatus as previously described should cers tainly sell at retail for less than $50.00. including the indectorutype oven. the moisture calculator. a scale. and a carrying case. The principal disadvantage of the method lies in the fact that the time required to test a single sample is so great that testing many re— petitive samples is not feasible. Experimental data indicate that the moisture content of hay curing in the field is highly variable at any given time. Therefore. the validity of one or two samples as an indicap tion of the moisture content of the crop is very questionable. An injectorbtype‘tractor exhasut oven has been developed which is capable of drying samples of alfalfa hay in approximately 7 minutes and samples of grain in approximately 15 minutes. The injector principle 193 employed limits the temperature of the exhaust gas and air passing through the sample so that serious burning does not ordinarily occur. The Operator need not tend the sample while it is heating in the oven. Leaving the sample in the oven as long as twice the time necessary to reach constant weight does not appreciably affect the moisture deter— mination. A simple calculator has been designed for use in the field to de- termine the moisture content from the wet and dry weights without arithp metic computation. Electrical Resistance Method Laboratony studies of the resistance method indicated the importance of electrode pressure on the resistance measurement of hay and grain. This demonstrates the need for standardizing the electrode pressure for such measurements. Electrode configurations of constant spacing were found most desirable, since weighing of the sample is not required and the resistance reading is less dependent upon pressure. Of the types of constant-spacing electrodes tested, the pinstype electrode configuration seems most desirable, since it is applicable to both hay and grain and it applies the measuring current to a greater definite sample volume than the coplanar plate—type electrodes. Field investigations of the resistance method indicated that single moisture determinations made with resistance-type meters do not reliably indicate moisture content of the lot. particularly when applied to hay. A reasonably accurate measurement of quantities of hay or grain can be obtained if the average of a number of readings is taken. 194 Since it was anticipated that most farmers would not care to take many readings and average them arithmetically. an automatic averaging device was designed to give a single reading of the average moisture content of a large number of samples. The averaging device actualhy determines the average conductivity and a close approximation of the arithmetic mean moisture content of a large number of samples. The principal advantage of the resistance method is in its great rapidity and versatility. The use of this method with the averaging device appears to be the best solution to the problem of sampling error which is so prevalent in the moisture testing of hay. The cost of the equipment for the electrical resistance method is the highest of the methods studies. particularly if the averaging device is employed. This factor may delay the broad acceptance of this equip- ment until such time as it can be economically produced or the demand for moisture measuring equipment becomes well developed among farmers. The averaging resistance-type moisture meter for hay should find immediate application as a research tool. The rapid determination of the average moisture content of large lots of hay has always been a serious problem to those engaged in research on hay harvesting and curing methods. Wet and Dry Bulb Thermometer Method The wet and dry-bulb thermometer method of determining the moisture content or storage quality of grain.was found to give an accurate indica- tion of the moisture content of wheat which had not been subjected to drying air. However, when samples were tested.which had been exposed to 195 drying air for short periods of time, the method consistently indicated moisture contents too low. This is due to abnormally dry surfaces of the kernels which cause humidity readings that are too low. When testing grain from a dryer. grinding the sample is necessary in order to expose to the air surfaces of the grain.which are more rep- resentative of the moisture content of the mass. A wet and dry-bulb thermometer apparatus was developed which is capable of testing ground grain.without the collection of grain dust on the wet bulb. Dust on the wet bulb prevents accurate determination of the wet-bulb temperature. The wet and dry—bulb thermometer method requires less costly equip- ment than any of the methods tested. No weighing is required. The method is not as versatile as the resistance and exhaust oven methods. since it is applicable only to moisture contents in a narrow range above and below the safe storage moisture content. It is applicable to testing grain in most of the moisture range of interest to farmers. It would not be applicable to the testing of hay in the field for silage or now curing. Its applicability to testing hay for any purpose is somewhat limited by the fact that a large number of samples cannot be tested rapidly. r0 1.. TABLE XVII DATA FROM TESTS OF INJECTOR—TYPE TRACTOR EXHAUST 0V ’3 ‘l. «'0 III-l \OJ‘JNOCDOQ) HHHHNHNH $5.30: Firirirhfi \O H “\O\\OO\V\V\o amsmnxm dd cede wnwzwnx “wall! # .,~“_——.—.“ul' ‘Hh - Iv ~-a‘-" .UQ‘V. _-.-u - Q -~~.--‘——. 4-: .o‘. vu-v-‘a—-—-——. -—._ - mam “MM 0 can. mm: mm: mum mam cam can cam omw can omm mam 0mm mmm mmm omm mam mom can mmn can now mum omm omm mum mom m:m 0mm s.'l I ’l .h ..maoe oaaamm Jfiafig ——. _ u-.- .__..... __._.. _-_‘ —.—.__-....—. ”a- .. -....-—oo.o- ‘---—-—..-. .—-. -__ _ -~_...—_.. .. - I!I|< -a flmw can mam mwm nmm own mun mom can now owe oz“ mmw cow mmm one mmm mmm mow. mmm nan “cm can mno 0mm ,omu man mmw mau emu moo _.& ..mawa pmdmnxm nouowua . Nd: -A paoanoo onspufio: Husufimmm m I I O O O. O O O 0 0 O O \Oxoxn cfizrr4U\v\c>r4~¢ Siadr-GH HHMu-lmmmm, OQNCDb-Q MHmNa—Cmc-fl‘n O O O O ERQ$ Nd) [NI-1 O o o. “(‘5 HNN O Fcoxrc 05 P4 :riacu" HMMWGOWWOQQ (“NH nmfiiammk o \O I 5 E :3 .pmq mp Ill‘! ' N.Hm o.mm m.mm ll“ m.mm : w.om o.moa w.moa _ m.m:H a N.Hm N.:oa m.moa w m.omaw : m.mm H.m~a m.m~a M u.mma : m.:a n.5wa H.mNH _ w.oma . o.am W m.mo~ m.mo~ m.ona _ g m.:a H m.m~H m.mwa a m.m:a . H.~H H m.mm~ N.HmH _ «.maa name _ o.:a . 3.:m a.mw o.ooH . o.oa m.mw 0.3m o.ooa . n.3a J N.mm m.mm w o.ooa g m.~a m m.mm .H.mw o.ooa : _ m.mH H m.mnfl M m.mma :.oma : m.m w H.0mH m :.oma o.oom amen; , m.on W m.m: a 5.3: m.HmH w . m.ma ” m.mm W H.0m w.mmfl ~ g M m.mm _ 5.0afi W H.moa “.mna w s W H.5m M o.HHH m o.oo~ :.m:a W aufiamflq w m w . H.mm m a.moa M o.noa . m.m:a a m.mm m m.moa w c.0oH m.m¢a . _ m.mH ~ N.mNH _ a.mma m.mnH oaoap _ u M namammdq m.am M m.mma m.mmH m.mmH : ”.mm w «.50 n.3m o.ma g W.Hd W m.Hw H.0w N.HOH x m.ma . m.om o.mm n.5m g m.w: n.nm «.mm m.mm : o.m~ :.NHH H.0HH a.a:a : m.mn m.mu m.~m H.HmH g ®.Hm N.mn m.am o.mda s N.m~ m.mo~ o.:oH m.¢:~ . m.nm m.mo N.mw H.Nma . N.m: n.0m N.ow m.m¢a undamaq . 83.5 38.8 :83 capo .nopo .puq pusmgnm uaaaa ”mamg an pnmfloz an pnwfloa .pnmfloz lam an camamm mamamm mamswm dogmas .0 .2 m Ham m-¥ qufim HafipfiaH mono 197 TABLE XVII Continued -1. --n--!:;n.-4: g:,.:;-: : ---;:ae- - ma 0mm “mm oo.a m.mm 5.:N m.m:a «. n.0na a o.oom g mN mmm mmm um.a N.H~ o.ma o.mHH n.0NH N m.mza ; ma mmm 0mm m~.H m.HN H.om :.nm~ _ m.mn~ m o.oom . ma omm mnm o:.H m.mm H.3N m.HHH _ m.mHH _ o.omH g aH 0mm mom om.o o.n~ m.:~ o.omH _ e.gnfl “ o.oo~ : ma 0mm omw om.o m.n~ o.mm m.maa H.oma m o.oo~ . ma can one oo.o :.mm m.:m N.maa 3.0m” w o.oom namom brflz ma mam coo mm.o m.m H.m :.omH m m.HwH m o.oom = NH omm mom om.H m.m u.m w.omH w w.mmH m o.oow madam . whom om mmm mmm mm.o m.ma m.ma m.ooH _ o.moH 7 m.mma g mH com mon Ho.a n.5H m.ma u.am w m.mw o.ma : ma can own mo.m m.ma m.oa H.HNH w «.3NH m.m:H . g mfl mmm mam mm.H m.mH H.mH N.mmH . m.ama H.Hom v g cm nmm can nm.a m.mH o.mH m.ama 0.:mH o.oom m z 4H oom own mm.H m.ua «.ma o.mu _ m.om , «.00 w = ma aim cam o:.H m.om H.mH :.mHH m m.omH . o.¢¢a M : om can can :N.o N.om m.ma m.mmH m.oma m.oom m a om mmm emu mm.H o.~m 0.0m w.mma m.mma o.mma : mm can con mo.” m.ma :.ma a.moH w.moa o.mom g om own ,omm am.a :.Hm m.am mwuma 3.0:H m.oom 9 mm can m:m nH.~ m.wH m.ma m.~ma m.me o.oom naoo nopdnwz wanna newuo .:o>o no>o .m .h ..m§oe nope nope .pmq pudmgxm nsmua pmdnnxm ..m209 andmnufi anonnoo nmbo soda: hfi unwaos an unwaoa aAmaos a“ mafia camemm nouowns cascade: .pwq up tum mp camadm mamewm wamamw coaaoa mafihun; .xmz .wm: Hasuduom m .o .:.m .o .2 m Haqfim Hmafim HafiudnH mayo 198 APPENDIX II ANALYSIS OF THE AVERAGE INDICATED BY THE AVERAGIHG DEVICE The averaging device was found by analysis and experimention to indicate the average conductivity of n samples of resistance R1, R2, R3, —~- Ru and moisture content M1, M2. M3. --- Mn. The indication of the meter in terms of resistance is expressed by the following: .lsi J. .1: .J. --- l. (1) Re. n (31+R2+33+ Rn) If the moisture content is assumed to be inversely proportional to the logarithm of resistance, any sample of moisture content,M1,has a resistance which may be expressed by the following exponential relation: _ -aM; ' hi: b10 1 (2) where a and b are calibration constants of the meter. Substituting Equation (2) into Equation (1) and simplifying: 2313:; an. :1le an, all" (3) 10 n 10 +10 +10 +----1o Taking the logarithm of both sides of Equation (3) and simplifying: L ( EM. 8M: 8M; BMn)] Ma: log n 10 + 10 +10 +--- 10 (4) J a Thus an expression is obtained for the average moisture content inp dicated by the meter in the terms of the moisture contents of the indi- vidual samples and a calibration constant. a. The calibration constant, a, is equal to the rate of change of the logarithm of resistance with the moisture content. The true arithmetic mean moisture content, Me: of n samples is 199 expressed by the following relation: Me = M. + 142+ Mir ...... Mn (5) 11 Me and Ma are equal only when the moisture contents of the individual samples are equal. It can be easily shown.by substitution of moisture values into Equations (b) and (5) that the greater the variation of the moisture values, the greater is the difference between Me and Ma. Of concern in.the application of the averaging device is the mag- nitude of the difference between M, and Ma for a given standard deviation of the moisture contents of a large number of samples. The standard deviation of the moisture contents of hay samples in the field was found to be approximately three moisture percentage points. In order to demon- strate the magnitude of the inherent error of the averaging device in indicating the arithmetic mean of a number of samples. a hypothetical group of 20 samples. were averaged by Equations (4) and (5) and the results compared. The moisture contents of the 20 samples were normalLy distributed about an arithmetic mean of 30% moisture content with a standard deviation of 3.2%. A typical value of 0.1 was assumed for the calibration constant, a. The average. Ma, calculated by Equation (h) was 30.7%rmoisture. Thus the inherent error of the device in determining the arithmetic mean is 0.7% moisture. The averaging device will always indicate an average higher than the arithmetic mean . since the rate of change of conductivity with moisture content decreases as moisture con— tent increases. By assuming a constant standard deviation for all hay, most of the inherent error of the averaging device could be removed by calibration. 200 However, there is little reason to believe that the arithmetic mean noisture content is any better index of storrge quality of hay than the slightly higher value which is indicated by the averaging device. umfi... " —. -—. __.. *—-—~«- _4_. ._ ._. — w"-.. .-y. -y ROOM USE ONLY INTER-Llflnnifl LUAN FEB 1 9 '« ”Y 'Zu ,~. HER—LIBRARY 1.02".“ 55 Jfi‘. i ' m0 ‘ am 1 '5! 1 1‘ H ERSI ”'Tl’w'y‘fin‘l‘vfinfllmfmiimm 3 1293 030 HIHIWIWITVS L3 2186