Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The design, construction and tests of an in situ capacitance moisture sensor and a portable capacitance… Kra, Eric Y. 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1992_spring_kra_eric_y.pdf [ 1.42MB ]
Metadata
JSON: 831-1.0086704.json
JSON-LD: 831-1.0086704-ld.json
RDF/XML (Pretty): 831-1.0086704-rdf.xml
RDF/JSON: 831-1.0086704-rdf.json
Turtle: 831-1.0086704-turtle.txt
N-Triples: 831-1.0086704-rdf-ntriples.txt
Original Record: 831-1.0086704-source.json
Full Text
831-1.0086704-fulltext.txt
Citation
831-1.0086704.ris

Full Text

THE DESIGN, CONSTRUCTION AND TESTS OF AN IN SITU CAPACITANCE MOISTURESENSOR AND A PORTABLE CAPACITANCE MOISTURE METER FOR ORGANIC SOIL ANDSAWDUSTByEric Y. KraB.Sc. (Agricultural Mechanization), University of Ghana, 1986A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Bio-Resource EngineeringWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJuly, 1992© Eric Kra, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her repres&tatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of R,O-p9$0c4frc6 eAf4-,Af’ILrThe University of British ColumbiaVancouver, CanadaDate__________DE-6 (2/88)AbstractA capacitance chamber was constructed to measure the capacitance and hence the moisturecontent of a sawdust sample placed within it. Five readings of capacitance were taken for eachsawdust moisture content level. The readings varied widely for each moisture content level suggestingsome conditioning of the sample will be necessary if this type of chamber is used as an accuratealternative to the gravimetric oven method of moisture content determination.A simple in situ capacitance moisture meter was also designed and constructed. The deviceconsisted of a 2.0 x3.0 xO.6 cm sensor and a hand-held digital multimeter with a capacitance range.The sensitivity of this moisture meter was compared to that of a commercially available fibreglassresistance type. In a series of experiments, two sensors (one of the capacitance type and one of thefiberglass resistance type) were installed in a different saturated soil and sawdust samples and thereadings of the meters and weights of the samples were recorded at regular intervals. Of the twomoisture meters, the capacitance moisture meter was found to be more sensitive to small changes inmoisture content. With a few modifications (discussed in Chapter 5) to the sensor design, to improveaccuracy, it is possible to monitor small changes in small volumes of organic soil and sawdust with thissimple in situ capacitance moisture meter.TABLE OF CONTENTSAbstract I’TABLE OF CONTENTS iiiList of Figures viiList of Tables xList of Symbols xiiAcknowledgements xlvChapter 1- INTRODUCTION 11.1 Introduction 11 .2 Justification for study 51 .3 Objectives 7Chapter 2- THEORETICAL BACKGROUND 92.1 Capacitors 92.1 .1 Dielectric (or insulator) 102.1.2 Capacitance and Charge 102.1 .3 Calculation of capacitance from dimensions 112.2 Relative permittivity or dielectric constant 132.2.1 Variation of dielectric constant with a-c current frequency 162.3 Charging a capacitor 192.4 Discharging a capacitor 202.5 Capacitance Measurement 232.5.1 Bridge Circuits 232.5.2 Digital Capacitance Meter 24IIITable of Contents2.5.3 Sensitivity of a Meter 242.5.4 Resolution of a Meter 25Chapter 3 - LITERATURE REVIEW 263.1 Principles of measuring soil water content by soil dielectric properties 263.2 Earlier Experiments with Capacitance Moisture meters 273.3 The Time-Domain Reflectometry (TDR) Method of Measuring Dielectric Constant . 313.4 Capacitor-type Moisture Sensors in Industrial Process Control 333.4.1 Capacitance-moisture Measurement in Slurries, Pastes and Emulsions 333.4.2 Capacitance-moisture Measurement in Solids 343.5 Porous Material Conductivity Measurement Methods 353.5.1 Direct Soil Conductivity Measurement Methods 353.5.2 Porous Block Conductivity Methods 353.5.3 The Fibreglass Resistance Instrument 36Chapter 4 - MATERIALS AND METHODS 394.1 PART 1 - Design, Construction, and testing of a capacitance chamber 394.1 .1 Objectives 394.1.2 The design of the chamber 404.1.3 Materials 434.1.4 Procedure 444.1.5 Method of Analysis of Data 454.2 PART 2 - Design, construction and testing of the capacitance moisture sensors . . . 464.2.1 Objectives 464.2.2 Materials 464.2.3 Construction 474.3 Comparison of the two types of in situ moisture sensors 474.3.1 Apparatus 47ivTable of Contents4.3.2 Procedure 494.3.3 Method of Analysis of Results 51Chapter 5 - RESULTS AND DISCUSSION 535.1 Abbreviations and Terminologies Used in the Discussion 535.2 Important Notes 545.3 In situ Sawdust Moisture Content Measurement by Capacitance Sensors 555.4 Likely Factors Causing Differences in Calibration Curves 585.4.1 Differences in Dielectric Sawdust Packing Density 585.4.2 Contraction of Dielectric Sawdust During Drying 595.5 In situ Organic Soil Moisture Content Measurement by Capacitance Sensors 605.6 Mathematical Relationships between Capacitance of Capacitance Moisture Sensorsand Sawdust and Organic Soil Moisture Content 635.6.1 Linear Regression Models For the capacitance of Sensor Cl in Sawdust 645.6.2 Linear Regression Models For the capacitance of Sensor C2 in Sawdust 685.6.3 Sudden Change in Slope of Linear Regression Models in Sawdust 695.6.4 Linear Regression Models for Sensor C3 in Organic Soil 735.6.5 Linear Regression Models for Sensor C4 in Organic Soil 765.7 Sensitivity of the Resistance and Capacitance Sensors Compared 795.8 Moisture Content versus Capacitance of Chamber 83Chapter 6 - CONCLUSIONS AND RECOMMENDATIONS 86List of References 88APPENDIXA.1 Calculation of SensitivitiesA.3 Calculations of Mass of Soil MoisturevTable of ContentsA.4 Calculations of Moisture Contents (% dry weight basis) 19.’)-A.5 Calculations of Moisture Contents (% dry volume basis) tO 2viList of FiguresFigure 2.1 Some geometrical shapes of capacitors; (a) parallel plate (b) parallel conductor and (c)cylindrical 9Figure 2.2 Physical dimensions and capacitance; (a) parallel plate capacitor, (b) concentric spheres,and (c) cylindrical capacitor 12Figure 2.3 A typical frequency dependency of dielectric constant. (Source: Blech, 1989). . 18Figure 2.4 Simple RC capacitor charging circuit 19Figure 2.5 The voltage across a charging capacitor 21Figure 2.6 Simple RC capacitor discharge circuit 21Figure 2.7 Voltage across a 1 F discharging capacitor 22Figure 2.8 A bridge-type circuit for measuring an unknown capacitance 23Figure 2.9 A schematic diagram of a digital capacitance meter using the 3905 Timer 25Figure 3.1 Relationship between the frequency, f, in MHz and the soil moisture content 0 in percentby volume for bess and silty sand. (Source: Kurã± et al, 1970) 28Figure 3.2 Relationship between moisture content and capacitance reading, clay loam, Hughenden,N. Queensland. (Source: De Plater, 1955) 29Figure 3.3 Simplified fibreglass moisture sensor 37Figure 3.4 Construction of fibreglass soil moisture cell. (Source: Colman and Hendrix, 1949). 38Figure 4.1 (a) cross-sectional view of capacitance chamber, and (b) with lid in placeviiList of Figuresand clamped 42Figure 4.2 Dimensions of chamber; (a) side view and (b) plan of lid or base 43Figure 4.3 Construction of capacitance moisture sensor 48Figure 4.4 Placement of a fiberg’ass resistance sensor and a capacitance sensor in thesame pot of organic soil or sawdust 50Figure 5.1 Four calibration curves for Cl in sawdust 57Figure 5.2 Four calibration curves for C2 in sawdust 57Figure 5.3 Four calibration curves for C3 in orgainc soil 62Figure 5.4 Four calibration curves for C4 in orgainc soil 62Figure 5.5 Linear regression lines for test 1 of Cl in sawdust. R2 = 0.951 for line 1 and 0.994for line 2 66Figure 5.6 Linear regression lines for test 2 of Cl in sawdust. R2 = 0.971 for line1, and 0.995 for line 2 66Figure 5.7 Linear regression model for test 3 of Cl in sawdust. R2 = 0.986 67Figure 5.8 Linear regression line for test 4 of Cl in sawdust. R2 = 0.998 67Figure 5.9 Linear regression lines for test 1 of C2 in sawdust. R2 = 0.984 for line 1 and 0.998 forline 2 71Figure 5.10 Linear regression model for test 2 of C2. R2 = 0.983 for line 1 and 0.996 for line2 71VIIIFigure 5.12Figure 5.13Figure 5.14Figure 5.15Figure 5.16Figure 5.17Figure 5.161.0727474757577777878List of FiguresFigure 5.11 Linear regression model for test 3 of C2 in sawdust. R2 = 0.943 forfor line 2line 1 andLinear regression model for test 4 of C2 in sawdust. R2 = 0.981 72Linear regression model for test 1 of C3 in organic soil. R2 = 0.994Linear regression model for test 2 of C3 in organic soil. R2 = 0.990.Linear regression model for test 3 of C3 in organic soil. R2 = 0.980Linear regression model for test 4 of C3. R2 = 0.958Linear regression model for test 1 of C4 in organic soil. R2 = 0.994Linear regression model for C4 for test 2 in organic soil. R2 = 0.998R2 = 0.988Figure 5.19 Linear regression model for test 3 of C4 in organic soil.Figure 5.20 Linear regression model for C4 in organic soil. R2 = 0.974.Figure 5.21 Variation of meter reading (in “resolution” units) with sawdust moisture content, whenusing capacitance and resistance moisture instruments with Ri and Cl connected. 81Figure 5.22 Variation of meter reading (in “resolution” units) with sawdust moisture content, whenusing capacitance and resistance moisture instruments with R2 and C2 connected. 81Figure 5.23 Variation of meter reading (in “resolution” units) with organic soil moisture content, whenusing capacitance and resistance moisture instruments with R3 and C3 connected. 82Figure 5.24 Variation of meter reading (in “resolution” units) with organic soil moisture content, whenusing capacitance and resistance moisture instruments with R4 and C4 connected. 82Figure 5.25 Capacitance of moisture content chamber versus sawdust sample moisture content. Thesingle line is the average value of the five replicates 84ixList of TablesTable 2.1 Dielectric constants: Typical values (T= 20°C;atmospheric pressure; f< 1 MHz) . . 16Table 5.1 Changes in moisture content (i.e. s in equation (4.1)) that produced a unit change incapacitance and current meter reading. Calculations are based on the differencebetween meter readings at 10 and 40 % moisture content 80Table 5.2 Data obtained from tests of capacitance chamber with sawdust 83Table Al. Calibration data for moisture sensors #1 and #2 placed within two different saturatedsawdust samples being air-dried in the laboratory - Test 1Table A2. Calibration data for moisture sensors #1 and #2 placed within re-saturated sawdustsamples (from Table Al.) being air-dried in the laboratory - Test 2Table A3. Calibration data for moisture sensors #1 and #2 placed within two new saturated sawdustsamples being air-dried in the laboratory - Test 3Table A4. Calibration data for moisture sensors #1 and #2 placed within the fresh sawdust samples(see Table A3.) being air-dried in the laboratory after re-saturation- Test 4Table A5. Calibration data for moisture sensors #3 and #4 placed within two different saturatedorganic soil samples being air-dried in the laboratory- Test 1Table A6. Calibration data for moisture sensors #3 and #4 placed within re-saturated organic soilsamples (from Table A5.) being air-dried in the laboratory - Test 2Table A7. Calibration data for moisture sensors #3 and #4 placed within two new saturated organicsoil samples being air-dried in the laboratory- Test 3xList of TablesTable A8. Calibration data for moisture sensors #3 and #4 placed within the fresh orgnaic soilsamples (see Table A87.) being air-dried in the laboratory after re-saturation - Test 4 /Table A7. Calibration data for moisture sensors #3 and #4 placed within two new saturated organicsoil samples being air-dried in the laboratory - Test 3Table A8. Calibration data for moisture sensors #3 and #4 placed within the fresh orgnaic soilsamples (see Table A87.) being air-dried in the laboratory after re-saturation - Test 4 IJOxiList of Symbols= the slope of the capacitance-05curve of a particular sensor within a specified O, range.= intercept on the Ow-axis of the O,-C graph (102m3water/rn3 sawdust or organic soil).Ekc = the average change in capacitance of the capacitance sensor over a specified range ofmoisture content (nF).Al = the average change in the current through the fiberglass resistance moisture sensor, overa specified moisture content range (pA).= dielectric constant or permittivity of a dielectric (F/m).= permittivity of vacuum (F/rn).£r = relative permittivity.9 = volumetric moisture content.volumetric water content (m3/rn).p = bulk density of soil.0dc = the direct current conductivity of a dielectric.= frequency of an electromagnetic wave (radians/sec).a.c. = alternating electric current.A,B = empirical constants.c = propagation velocity electromagnetic wave in space.C = capacitance (farads).Cl = Capacitance sensor #1C2 = Capacitance sensor #2C3 = Capacitance sensor #3C4 = Capacitance sensor #4C3 = capacitance in vacuum (F).Cs = capacitance of a standard capacitor (farads).C = capacitance of the unknown capacitor (farads).d = distance between parallel plates (m).D = electric flux density in vacuum (F/rn).D0 = electric flux density in dielectric other than vacuum.e = natural logarithim (2.71 83).XIIList of SymbolsE = electric field (v/rn).Err = absolute error of estimation of m.c. (m3/)f = frequency of electromagnetic wave or electric current (Hz).k = relative dielectric constant.K = overafl dielectric constant.K’ = the real component of the dielectric constant, K.K” = the imaginary component of the dielectric constant.K1,K2 = emperical constants.L = inductance of a circuit (henrys).= moisture content.P = polarization (C/rn2).r = the resolution of the particular range of the meter used for measuring capacitance orcurrent (nF or pA).R = resistance (ohms,c).r1 = radius of inner sphere.Ri = Resistance sensor #1= radius of outer sphere.R2 = Resistance sensor #2.= regression coefficient as computed by Quattro-Pro Spreadsheet.R3 = Resistance sensor #3.R4 = Resistance sensor #4.RB = resistance of balancing resistor (ohms).= resistance of ratio resistor (ohms).t = time (seconds).T = temperature.= time constant.= the amount of time it takes the capacitor to charge to a reference capacitance.V = aplied voltage or maximum voltage (volts).V = voltage across capacitor.s = the sensitivity index for a capacitance sensor, as defined in equation (4.1).xmAcknowledgementsThe works of the Lord are great, studied by all who have pleasure in them (Ps. 111:2). I am gratefulto God for the pleasure of researching into this small portion of His vast and great creation. I thankHim for giving me such a wonderful Professor.I am very thankful to Dr. S.T. Chieng, my Professor, for allowing me the freedom, and encouragingme to explore those areas of this research of greatest interest to me, while at the same time continuingto offer invaluable suggestions. I am also very grateful to Professor L.M. Staley and Dr. A.K. Lau fortaking time off their very busy schedules to make great contributions to this thesis.I deeply indebted to the Faculty of Agriculture of the University of Ghana for awarding me thegraduate scholarship, enabling me to study here at the University of British Columbia.I am very grateful to all those wonderful friends of mine like Catherine, Selena, Edgar, Els, Akosua,Dawn, Marlene, Sam, Isaac, Ikuko, Guangxi, Redi, Shipra, Moez, Jun, Awal and all the others who Icannot mention here not because their contribution was of little value, but because of the limitation ofspace. I thank you all for the ways you contributed directly and indirectly to my work here over the pastthree years.My special thanks to Neil and Jourgen for their invaluable technical contribution in setting up theapparatus for the experiments.xivChapter 1INTRODUCTION1.1 IntroductionAn accurate knowledge of water content in soil is very important in agriculture. In order to designan efficient irrigation system, the appropriate crop water requirements and the available water in thesoil must be known. The crop water requirement data is usually obtained by monitoring changes in thesoil water content or by estimation from environmental variables such as temperature, radiation, relativehumidity, etc. But even when estimation methods such as the Penman Evapotranspiration (ET)formula, are used, their accuracy must be verified by measurements made directly in the soil.There are numerous growth media in use for agricultural crop production, particularly in thegreenhouses. Some of the more common ones are organic soil, sawdust, and compost. In modellingefficient irrigation systems for these growth media, an essential step is experimentation with the actualmedium. The quantification of the hydraulic properties of the medium can only be made throughmeasurement of its moisture content at various points within its profile. In this respect, methods whichdo not entail destructive sampling of the medium are particularly more desirable.There are several devices and procedures today for obtaining soil moisture measurements foragricultural irrigation purposes. The principles behind these methods have long been established.Haise and Hagan (1967), classified them broadly under these headings - sampling and drying (or thegravimetric method); electrical resistance; neutron scattering; gamma-ray absorption. Hillel, (1980)broadens this to include techniques based on soil thermal properties on water content and the use ofultrasonic waves, radar waves, and dielectric properties.Each method has its own set of advantages and disadvantages. The gravimetric method is simple,Introduction 2but time-consuming and can therefore be prohibitively expensive. Although this method is usually thestandard by which the accuracy of most other methods are judged, it is not very convenient in manystudies because of destructive sampling (i.e. the sample whose moisture content is determined cannotcontinue to be part of the experiment). The spatial variation of moisture content within even a smallarea of soil also limits the usefulness of this method.Tensiometers are relatively simple and inexpensive. However they are limited to potentials up toabout 0.8 bar and require frequent servicing for proper functioning, and because they measure potentialonly in the immediate vicinity of the unit, several tensiometers are needed to give a reliable spatialaverage (Campbell and Campbell, 1982).Gypsum resistance blocks, because they are inexpensive, are often a very popular option formeasuring soil moisture content. But one of their main drawbacks is that they require frequentcalibration to give measurements that are better than qualitative (Topp and Davis, 1985). The gypsumblocks contain suitable electrodes separated by the gypsum. The resistance between the electrodesvaries with the moisture content of the gypsum.The electrical conductivities of these blocks when they reach moisture equilibrium, are regarded asan index of soil water content. But the water content of the porous block (i.e. gypsum) depends uponthe energy status of the water rather than upon the water content of the soil with which it is in contact(Gardner, 1986).The electrical conductivity of most porous blocks is due primarily to the permeating fluid rather thanto the solid matrix. Thus it depends upon the electrolytic solutes present in the fluid as well as uponthe volume content of the fluid. Blocks made of such inert meterials such as fiberglass, for instanceare highly sensitive to even small variations in salinity of the soil solution. An undesirable consequenceof the solubility of gypsum is that these blocks eventually deteriorate in the soil. Hence the relationshipbetween electrical resistance and moisture suction varies not only from block to block, but also fromeach block as a function of time, since the gradual dissolution of the gypsum changes the internalIntroduction 3porosity and pore-size distribution of the blocks. For these and other reasons (e.g., temperaturesensitivity) the evaluation of soil wetness by means of electrical resistance blocks is likely to be oflimited accuracy (Hillel, 1980).Nevertheless, blocks are often used to indicate water content of a soil even though precision in suchuse is rather low. However, the popular use of porous blocks likely stems from their utility as indicatorsof water conditions favourable for or unfavourable to plant growth rather than their ability to indicate soilwater content (Gardner, 1986).Another electrical resistance moisture content device is the fibreglass Electrical Soil MoistureInstrument (Colman and Hendrix, 1949). One model from SoilTest* has moisture sensors made up oftwo metal plates separated by a fibreglass binding which provides a coupling that varies with soilmoisture content. But the accuracy of this instrument may also be affected by the salinity of the soil,since the conductivity between the two electrodes is governed primarily by the ions in the permeatingfluid rather than on the solid fiberglass matrix itself (Hillel, 1980).The use of conventional in situ soil moisture sensors such as the tensiometer or gypsum block maybe hampered by properties of the test materials which differ considerably from those of mineral-derivedsoils for which these sensors are designed (Baliscio and Lomax, 1989). Materials such as sawdust,peat and compost have typical values of moisture content much greater than what these instrumentsare used to measure. Secondly, water incorporated in organic matter such as the straw componentof compost may not be detected by some sensors. Thirdly, poor contact between the sensors and thetest material may result in misleading readings (Baliscio and Lomax, 1989).The neutron scattering and gamma-ray absorption methods are also highly accurate, and the resultsare available immeditely in volumetric units. But their high costs and possible health hazards whichmight result from inappropriate use of the radioactive probe, limit their widespread use. Although thisinstrument is expensive, it provides the opportunity of repeated soil water measurements at the samelocation in a representative volume of soil within the field (Heermann et al, 1990, Hillel, 1990).Introduction 4Moisture content measurement methods which depend upon the effects of moisture content on thedielectric properties of soil and other media have always attracted the attention of many researchers(Gardner, 1987). A dielectric is an electrical insulator or a non-conductor of electricity. The dielectricproperty of the medium that is normally measured is the capacitance. Capacitance (symbol, c, unit,farad) is the property exhibited by two conductors separated by a dielectric, whereby an electric chargebecomes stored between the conductors. A dielectric is a non-conductor of electricity eg. glass, wood,plastic (Turner and Giblisco, 1991).Capacitance is a function of dielectric constant which changes with moisture content. Relative toother substances, pure water has a high dielectric constant. When water is present in any material itsquantity modifies the capacitance of the material in which it is found compared to the dry material.Topp et al, (1980 a), found a strong dependence of the dielectric constant on volumetric water contentfor a number of soils with varying grain sizes.Various methods have been developed for measuring the capacitance of soil and other media formoisture content determination. Many of these employ some form of capacitor so arranged that atypical soil sample would be located between the plates of such a capacitor (Holmes et al, 1982).Gardner (1987), related that the greatest discouragement encountered by researchers on this techniquehas been the difficulty in achieving a suitable design for the electrodes which does not introduceunacceptable and extraneous capacitance. He however pointed out that, nevertheless, this approachhas so much to commend that it has never been totally forsaken.The capacitance has been measured by applying an alternating electrical current (the test signal)to an electrical circuit which includes the capacitor. By manipulating the frequency of this test signal,while monitoring the current through a certain point of the circuit, the unknown capacitance can bedetermined. One such method is the Wein Bridge circuit which is described in the Literature Review.Another method of measuring soil water content based on the soil dielectric properties is called thetime-domain reflectometry (discussed further in Chapter 2- Theoretical Background), and mayIntroduction 5eventually supplant the neutron meter, but at this stage it is still largely experimental and thecommercially units currently available are very expensive (Topp and Davis, 1985; Baliscio and Lomax,1989).1.2 Justification for studyThis present study is the result of an initial experiment to study the effects of temperature, emitterflow rate, and other variables on the moisture distribution pattern in the wetted cone (zone) of organicsoil and sawdust under trickle irrigation. A detailed study of that nature required the placement of manysmall, very sensitive moisture sensors in the irrigated medium.The IRAMS moisture meter, which works on the principle of time domain reflectometry was initiallyselected for the study. The experiment was modified to use a single probe, due to the ratherunaffordable cost of the probes. But the readings from this modified setup were unsatisfactory, andso work on the study was suspended until a suitable moisture meter could be acquired. It wastherefore decided to direct attention towards the development of an inexpensive and yet accuratemoisture meter. This new moisture sensor when constructed will be evaluated against anothercommercially available moisture meter (less expensive than the IRAMS moisture meter), with afibreglass resistance type of moisture sensorNelson (1973), showed that the dielectric constant which determines capacitance is itself dependentupon the frequency of the electrical test current. Baliscio and Lomax (1989) also found that thedielectric constant and hence the capacitance was highly sensitive to measurement frequency, theformer decreasing with increasing frequency. This frequency dependence is a major disadvantage ofthe Wein bridge device since frequency is manipulated as part of the measurement procedure.The effect of the frequency-dependence of the dielectric constant can be minimized by using testfrequencies in excess of 10- 30 MHz (Nelson, 1973, Wobschall, 1978 and Schmugge et al, 1980).Introduction 6However Baliscio and Lomax (1989) point out that the disadvantage of working with frequencies thathigh is that greater care must be taken in dealing with extraneous effects due to such factors as theinherent capacitance or inductance of circuit conductors and components. They therefore recommendthat improvements in the design of the capacitance measurement circuit need not focus on increasingthe frequency of the measurement signal, but could instead be directed towards developing a circuitthat would work at a single frequency or a very small range of frequencies.Recent advances in Electronics have made available at very affordable costs, high quality digitalcapacitance meters which work at a single frequency. One such meter, the EMCO Model DMR- 2012Digital multimeter with capacitance range (having a resolution of 1 pf (pico-Farad)), uses a testfrequency of 400 Hz. At such a low test frequency, extraneous effects due to such factors as theinherent capacitance or inductance of circuit conductors and component, are very much minimal.Higher capacitance values will also be obtained at such a low test frequency, increasing the potentialof miniaturizing the capacitors; small moisture sensors are important in many experiments such asstudies of moisture distribution within the wetted cone of a trickle-irrigated soil.Baliscio and Lomax (1989) suggested that an in situ sensor small enough to measure the moisturecontent of a small surrounding volume of growth medium would have a very low value of capacitance(less than one pico-Farad (pF)), that it might be difficult to reliably measure such small capacitances.However, as explained in section 2.3, the frequency of the a-c signal of the capacitance test circuitinfluences the dielectric constant of the dielectric and hence the measured value of capacitance. Thatis the dielectric constant of any dielectric is not really a constant, but is dependent upon the a-c signalfrequency to which it is subjected. The capacitance of a capacitor is directly related to the dielectricconstant. Therefore the capacitance of a capactitor measured with a circuit operating on a powerfrequency will be higher than that obtained using a test circuit operating with a higher frequency in theUHF range, for example. Now if the effective area of a given parallel plate capacitor is reduced, thecapacitance (measured with a UHF test signal frequency) would be reduced accordingly. However fromthe above discussion it can be concluded that the reduction in capacitance resulting from the reductionIntroduction 7in effective capacitor plate area will be lower if the reduction in plate area is combined with a switchto a lower test signal frequency. Therefore even though Baliscio and Lomax (1989) suggested theimpossibility of small in situ capacitance sensors, it might be possible if a capacitance test circuitoperating at a power frequency is used to measure the capacitance of the small in situ sensor.Verification of this conclusion is one of the objectives of this study.Given the above-mentioned facts there is therefore a great potential for the development of amoisture meter with the following desirable features:(i) Miniature size sensors;(ii) High accuracy;(iii) Inexpensive;(iv) High sensitivity to changes in moisture content;(iv) Fast response time to moisture content variations.1.3 ObjectivesThe main objectives of this study are therefore:(i) to design, construct, and test a capacitance chamber whose capacitance would be measuredby a digital capacitance meter employing a constant, low frequency (400 Hz), test signal.(ii) to determine the potential of the use of the chamber as a viable alternative (the gravimetricoven method) of determining the moisture content of sawdust samples placed inside thechamber;(iii) to design, construct, and test miniature in situ capacitance moisture content sensors (forIntroduction 8sawdust and organic soil) for use with a low-cost, low-frequency, and single test frequencydigital capacitance meter, and(iv) to compare the sensitivities (to changes in moisture content) of the capacitance moisturesensors in (iii), to those of a commercially available fibreglass resistance moisture meter -The SoilTest MC-3 13 Model moisture meter.The EMCO Model DMR -2012 Digital multimeter with a capacitance range was selected for measuringthe capacitances of the chamber and the in situ sensors because it is one of the least expensive onesand readily available from most electronic hardware dealers.Sawdust and organic soil media were chosen because they are widely used as growth media ingreen houses. Accurate, miniature and inexpensive moisture sensors for use in soil and sawdust wouldgreatly facilitate hydrological studies involving such media. For example in trickle irrigation studies, itis desirable to have many inexpensive moisture sensors to monitor the moisture content in various partsof the wetted cone of the irrigated growth medium.Chapter 22.1 CapacitorsTHEORETICAL BA CKGROUNDA capacitor is a passive electronic-circuit component consisting of, in basic form, two metal electrodes,or plates separated by a dielectric (or insulator) (Turner and Giblisco, 1991). Figure 2.1 shows threedifferent types of capacitor geometry.IflsuItorconductorIflSuItQr(c)conductor(b)InsuIutQrFigure 2.1 Some geometrical shapes of capacitors; (a) parallel plate (b) parallel conductor and (C)cylindrical.conductor9Theorectical Background 102.1.1 Dielectric (or insulator)A dielectric (or an electrical insulator) is a material that ideally, conducts no electricity; it cantherefore be used for isolation and protection of energized circuits and components. Although suchcommon nonconductors as glass, wood and plastic might first come to mind, dry air too is a dielectric,as is pure water.2.1.2 Capacitance and ChargeWhen a voltage is applied to the plates or electrodes of a capacitor, lines of electric flux form in thedielectric between the electrodes. The amount of flux developed is a measure of the capacitanceformed by the conductors and the dielectric (Levine, 1988). Capacitance (symbol c, unit farad) is theproperty exhibited by the capacitor, whereby an electric charge becomes stored between its electrodes.(Turner and Giblisco, 1991).A body becomes electrically charged when charge carriers are transferred to it or from it. A chargecarrier is a mobile particle whose movement constitutes electric current e.g. a mobile electron e or holep’ in a semiconductor. The negative charge U acquired on one body is equal to the positive chargeQ lost from the other, so that 0’ = Q. The magnitude of any charge is equal to an integral numberof n transported elementary charges e or p, so that U =ne or U = np’ (Dudley, 1989).When a dc voltage is applied to the terminals of a capacitor, a charge is developed on the platesof the capacitor. Theoretically, capacitance is expressed as a ratio of electric charge to the applied dcvoltage (Gore, 1989).c—2. (2.1)where,Theorectical Background 11C = capacitance (farads),Q = charge (coulombs), andV = applied voltage in volts.Unlike a resistor which dissipates energy, a capacitor stores energy and returns it to the circuit inwhich it is connected (Joerg, 1988) When a capacitor is charged, it stores energy. The energy wstored in a capacitor is given by (Levine, 1988).w I Cv2 (2.2)2where,W = energy (joules),C = capacitance (farads), andV = applied voltage (volts).When an alternating current is applied to a capacitor, the capacitor is alternately charged anddischarged. The charging and discharging of the plates through the external circuit makes it look asthough charges (current) flow through the capacitor. Actually since the plates are separated by aninsulator (or dielectric), the charges cannot flow between the plates through the capacitor, hence thereis no physical transfer of charges between the plates (Wilson, 1988 b).2.1.3 Calculation of capacitance from dimensionsFor systems of simple geometry (see Figure 2.2), the capacitance between two conductorsseparated by a single, homogeneous isotropic dielectric may be calculated in terms of the physicaldimensions of the conducting electrodes, and the perrnittivity or dielectric constant of the dielectricTheorectical Backgroundsubstance (see Table 2.1).12Figure 2.2 Physical dimensions and capacitance; (a) parallel plate capacitor, (b) concentric spheres, and (c)cylindrical capacitor.For a parallel plate capacitor, the capacitance is given bywhere,C=.EA_dC = capacitance (farads),A = the effective area of the plates (m2),d the distance between the plates (m), and= permittivity of dielectric (farads/meter).(2.3)(a)(b)(C)Theorectical Background 13The capacitance the system of concentric spheres (b) isc = 4ltE (2.4)(1/r —1/r2)where,C = capacitance (farads),r1 = the radius of the inner sphere (m),= the radius of the outer sphere (m), and= permittivity of dielectric (farads/meter).The capacitance of a system of coaxial cable is given byc = 2ItEL (2.5)In(r2/r1)where,C = capacitance (farads),r, = the radius of the inner conductor (m),r2 = the radius of the outer conductor (m), and= permittivity of dielectric (farads/meter).2.2 Relative permittivity or dielectric constantThe relative dielectric constant k compares the flux in a vacuum (k = 1) with the flux in the dielectricmaterial. The electric flux is the lines of force which are believed to extend in all directions from anTheorectical Background 14electric charge (Turner and Giblisco, 1991).The electric flux density, D, of an electric field, is the number of lines of force per unit area (C/rn2). Itis proportional to a property of space called the permittivity,= (2.6)where,D0 = electric flux density in vacuum (C/rn2),= permittivity of vacuum (F/m), andE = electric field (V/rn).Similarly the flux density in a dielectric is defined by Blech (1989), asD=EE (2.7)where,D = electric flux density in dielectric (C/rn2),= perrnittivity of dielectric (F/rn), andE = electric field (V/rn).Frorn the definition of k above,(2.8)D0 E0 EFrom equation (2.3), the capacitance of a capacitor in a dielectric, relative to that in a vacuum is givenbyTheorectical Background 15.2 =.. (2.9)C0 d d Ewhere,C, = capacitance in vacuum (F),C = capacitance with dielectric (F),d = distance between parallel plates (m), andEr = relative permittivity.Thus the dielectric constant is also defined, for a dielectric material as relative to the ratio of thevalue of a two-plate capacitor using the dielectric material to the value of the equivalent capacitor withfree space as the dielectric. Table 2.1 lists the Approximate values of some common dielectrics.Theorectical Background 16Table 2.1 Dielectric constants: Typical values (T = 20°C;atmosphericpressure; f< 1 MHz)Dielectric type kAir 1.0059Amber 2.9Asphalt 2.7Bakelite 6 (3.5 to 8.5)Beeswax 2.7Celluloid 6.2Ceramic 5.5 x io (4k to 7k)Distilled water 78Ebonite 2.8Ethyl 26Glass (window) 6Glycerin 56Mica 5 (6 to 7.5)Mylar 3Paper 2.5 (2 to 4)Paraffin 4 (3 to 5)Petroleum 4 (2 to 6)Polyethylene 2.3Polystyrene 2.6Porcelain 6.5 (6 to 7.5)Quartz 3.8Pyrex (glass) 4.8Rubber 3 (2 to 3.5)Slate 6.8Soil 2.9Teflon 2Vacuum 1.0Vaseline 2.2Water 81Wood 5.5 (2.5 to 8.5)*Approximate values are given. (Source: Joerg, 1988)2.2.1 Variation of dielectric constant with a-c current frequencyThe electric polarization is the addition in electric flux density in a dielectric material to the densityin free space, i.e.Theorectical Background 17P=D—E0E (2.10)where,P = polarization (C/rn2),D = electric flux density (C/rn2),= permittivity of free space (F/rn), andE = electric field strength (V/rn).The polarization is the total dipole moment induced in a unit volume of dielectric.From equation (2.10), and equation (2.6),P=EE—E0 (2.11)and= P + E p (2.12)E ETherefore, the relative permittivity or dielectric constant is proportional to the electric polarization,... =.......+1 (2.13)E E0The polarization within the dielectric material is determined by the displacement of charges. Foursources for polarization exist in the material: electronic polarization, due to displacement of electroniccharges; dipole polarization, due to reorientation of permanent dipoles; ionic polarization, due todisplacement of ions; and polarization by space charges, due to macroscopic displacement.Since the polarization process takes place in a finite time at any given temperature, it is expectedthat the relative permittivity, E, will be frequency dependent. A typical frequency dependence of theTheorectical Backgrounddielectric constant is seen in Figure 2.3.(a)Figure 2.3 A typical frequency dependency of dielectric constant. (Source: Blech, 1989).18The series of inflexions in the curve occurs at the relaxation times for the various polarizationprocesses. For example, the dipoles of the material represented in Figure 2.3 can follow audiofrequencies (20 Hz - 20 KHz), but cannot follow infrared frequencies. The dielectric constant at infraredfrequencies is decreased by the dipole component and contains only the ionic and electroniccomponents. Therefore the dielectric constant and therefore capacitance is higher at audio frequenciesis higher than at infrared frequencies.Dip10ElectronPower Audio UHF Intrared UV X-raysFrequencyTheorectical Background 192.3 Charging a capacitorA simple circuit for charging a capacitor is shown in Figure 2.4FI pV____I LJ1CFigure 2.4 Simple RC capacitor charging circuit.Here the capacitor, the charge power source (i.e. the battery), and a resistor are connected in series.When the circuit is open, the capacitor does not charge. But once the circuit is closed, the voltageacross the capacitor begins to rise, rapidly initially. The rate of voltage increase across the capacitorslows down with time until there is no apparent increase (Wilson, 1987). The voltage across thecapacitor (please see Figure 2.5) is described byv =vii — 1 (2.14)where,C = capacitance in farads,R = resistance, in ohms,= time, in seconds,Theorectical Background 20V,, = voltage across capacitor,V maximum voltage (battery voltage), in volts, ande = 2.7183.Practically, the length of time required to fully charge a capacitor is 5 7 (i.e. five time constants),whereT = RC (2.15)where,T0 = time constant,R = resistance (in ohms), andC = capacitance (in farads).Figure 2.5 shows the charging curve of a 1 F capacitor, with R = 1 ohm, and V = 1 volt.2.4 Discharging a capacitorAfter charging a capacitor, the acquired potential across its plates is retained, even when removedfrom the charging circuit, until it is discharged. An RC (Resistance-Capacitance) discharging circuit isshown in Figure 2.6.When the switch is closed, a discharge current will flow through resistor R. The voltage across thecapacitor t seconds after closing the switch is given byTheorectical Background0.90.80.7Cg 0.6C.0.50.400.300.2>0.10Figure 2.5 The voltage across a charging capacitor.214— — — — — —jCFigure 2.6 Simple RC capacitor discharge circuit.= V(e t/RCJ (2.16)4 5 6TTme (seconds)where,Theorectical BackgroundC = capacitance (farads),R = resistance (ohms),= time (seconds),V = voltage across capacitor (volts),V = maximum voltage or battery voltage (volts), ande = 2.7183.22The voltage across a 1 F capacitor in an RC discharging circuit, with R = 1 ohm, and V = 1 voltis shown in Figure 2.7. Practically, it requires 5 T to discharge a capacitor in an RC discharge circuit.0.9Co0.80.700.60.4C)De 0.30)a0.2>0.104 5 6Time (seconds)Figure 2.7 Voltage across a 1 F discharging capacitorTheorectical Background 232.5 Capacitance Measurement2.5.1 Bridge CircuitsBridge circuits are commonly used to measure the value of unknown capacitors. Bridge-typeinstruments work by comparing an unknown value of a component to a known value of anothercomponent of the same type (Lewis, 1988). An a-c type bridge set up for measuring capacitance isshown in Figure 2.8.Figure 2.8 A bridge-type circuit for measuring an unknown capacitance.A known standard capacitor C is connected in one arm, and the unknown C is in the other. Ana-c source with a suitable frequency is connected to the bridge, and the bridge is balanced with RA andRB, the phasing control R, and the Wagner ground control R, as required. The Wagner ground controlis adjusted with S2 in position 2 for the best null. Then S2 is returned to position 1. The Wagner groundcircuit is used to balance out any stray capacitance. When the bridge is unbalanced, a potentialdifference exists between the terminals of the head set, and a tone is heard. (The pitch of the tone isA-C SOURCETheorectical Background 24determined by the frequency of the a-c source connected to the bridge). RB is adjusted until the bridgeis balanced i.e. no sound is head on the head set, at which pointC,=C R8where,C = capacitance of the unknown capacitor (farads),C = capacitance of a standard capacitor (farads),RB = resistance of balancing resistor (ohms), andRR = resistance of ratio resistor (ohms).2.5.2 Digital Capacitance MeterFigure 2.9 shows the schematic diagram of a digital capacitance meter. The trigger input pulsestarts the capacitor on a charge cycle. The amount of time it takes the capacitor to charge to areference capacitance is marked T on the output signal. The smaller the capacitance the less time itwill take to complete the output pulse. Likewise a larger capacitor will produce a broader output pulse.The crystal oscillator signal, counted down to produce very accurate timing pulses, is also delivered tothe enable gate. The number of timing pulses that pass through the enable depends on the width ofT, which, in turn depends on the capacitance of C, (Wilson, 1988 b).2.5.3 Sensitivity of a MeterSensitivity should not be confused with instrument accuracy. Sensitivity is the ratio of the linearmotion along the scale by the instrument pointer, or indicator to the change of the physical quantity theinstrument is intended to indicate. Thus the sensitivity represents the output response per unit changeTheorectical Background 25Figure 2.9 A schematic diagram of a digital capacitance meter using the 3905 Timer.of input (Ambrosius et al, 1966).According to the above definition then, the sensitivity of a capacitance moisture meter is defined asthe change in capacitance per unit change in soil or sawdust moisture content. Similarly the sensitivityof the fiberglass resistance moisture instrument is the change in the current through the sensor or thechange in its resistance per unit change in moisture content.2.5.4 Resolution of a MeterThe resolution of a measuring system is defined as the smallest increment of the measured quantitywhich can be distinguished. The resolution of an indicating instrument depends on the deflection orincrement per unit input (Thompson, 1989). The resolution of the EMCO capacitance meter is 1 pF.TR I GGERHTx_IIENABLE111111111Chapter 3LITERATURE REVIEW3.1 Principles of measuring soil water content by soil dielectric propertiesMeasurements which depend upon the high dielectric constant for water have long been appealing.This is an intrinsic property of water, which serves to distinguish it quite readily from other materials inmany contexts (Gardner, 1987).Topp et al., (1980 a,b and c) showed that water content was the factor mainly determining thedielectric constant of soil material. Factors such as temperature, soil type, density of sample, and saltcontent had essentially insignificant effects (Topp and Davis, 1985).According to Kuraz et al., (1970), the dielectric constant E of a mixture of different materials maybe obtained, for example by the theory of Odelevskii (1951) by the relationE,—Ev.=o (3.1)i-i £1+2Ewhere V, is the partial volume of the i-th phase, and is the dielectric constant of the i-th phase.Equation (3.1) cannot be used for the direct calculation of £ of soils because of neglecting the geometryof the soil phases (i.e. solid, liquid and gaseous phases). However, it demonstrates a generaldependence of £ upon the moisture content.From Table 2, the dielectric constant of air (soil air) is roughly 1, that of soil water is 81 and that ofthe soil (solid) phase is 2.9. The E of soil actually fluctuates according to the composition of the soilin rough ranges from 2 to 10 (Kuraz etal., 1970). The partial volume of the water contained in the soil26Literature Review 27is the soil moisture content. Therefore changes in dielectric constant of the soil-air-water mixture willresult from changes in the soil moisture content.3.2 Earlier Experiments with Capacitance Moisture metersKuraz et a!., (1970) described a method of measuring soil moisture content, by measuring theresonant frequency of an oscillator circuit of which the soil capacitor is apart (please see their resultin (Figure 3.1). They pointed out that an empirical expression (Kaspar, 1969) originally found fordetermining the moisture content of porous building materials could be used:A (3.2)(B - 0)2where,A and B are empirical constants0 = volumetric moisture content, and= die’ectric constant of the materialNow the resonant frequency, 1, of an LC (i.e. Inductor-capacitor) circuit (Wilson, 1988 a) is given by1 1 (3.3)2it/Twhere,L = inductance of the circuit (henrys), andC = capacitance of the circuit (farads).Since the capacitance, C = EK, where K is a constant expressing the geometric arrangement of theLiterature Review 28electrodes, and since the circuit inductance, L = constant, then from equantions (3.2) and (3.3), it canbe proved that(3.4)where K1 and K2 are empirical constants, which have to be found by calibration.Thus they showed that there exists a linear relationship between frequency f, of the resonant circuitand the moisture content of the porous material. They tested this using the capacitance probe theydescribed as consisting of two semi-circular, rigidly connected electrodes placed on the periphery ofa glass tube. The electrodes had the form of a 2 cm. high cylinder sliding freely on the tube filled withthe soil. Their calibration curves for two soil samples- bess (bulk density p = 1.41 g cm3) and siltysand (p = 1.58 g cm3), is shown in Figure 3.1.Their test frequencies were of the order of 100 MHz. According to them, the use of such a highfrequency reduces errors due to ion conductivity in the soil.De Platar (1955), described a capacitance meter suitable for field or laboratory investigations. Hiscapacitance meter utilized a bridge circuit with a Hartley oscillator providing a 1000 Hz a.c. signalsource. His probe consisted of two stainless steel plates mounted on a suitable insulating handle; theplates were not insulated from the soil. For convenience in measuring moisture near an exposed soilsurface, plates 2.54 cm wide at a separation of 2.54 cm and penetrating into the soil for 1 .27 cm wereused. The calibration curve presented in their report is shown in Figure 3.2. The experimental datawas not published. The break in the calibration curve was reported to occur after field capacity isreached which was approximately 30% moisture content for the soil used. He noted that althoughAnderson (1943) found the relationship between capacitance and moisture content to be independentof soil type, Slyter (1954) found that variations can occur in some Australian soils, particularly at highmoisture content levels. There was however nothing reported on the repeatability or accuracy of thecalibration curve for the same soil sample.Literature Review 29160150140(Mhz)130120110100Figure 3.1 Relationship between the resonant frequency, f, in MHz and the soil moisture content 0 (% byvolume) for bess and silty sand. (Source: Kurá± et al, 1970).Halbertsm et al (1987), described a capacitive method of soil water content measurement basedon the measurement of capacitance of a capacitor with the soil-water-air mixture as the dielectricmedium. Thefr probe consisting of conductive plates or rods surrounded with soil, formed the capacitor.This capacitor was then connected into a resonant circuit of an oscillator. Changes in the watercontent, and thus changes in the capacitor capacitance, will change the resonance frequency of theoscillator. In this way the water content is indicated by a resonance frequency shift.They noted that an important source of error with the capacitive technique (employing conductiveplates) is the sensitivity for the electrical conductivity of the soil which influences the resonancefrequency of the oscillator. Their instrument was developed by the Technical and Physical Engineering0 10 20 30moisture content (% by volume)Literature Review 30capc tunc(mm F)007006• 005001• 003.002.001Figure 3.2 Relationship between moisture content and capacitance reading, clay loam, Hughenden, N.Queensland. (Source: De Plater, 1955).Research Services in Wageningen, Netherlands. No details of the design were however disclosed.Each probe had to be calibrated individually and for each soil. They reported an accuracy of betterthan 0.02 m3/m, provided that no changes in the matrix density occurred as is the case in swellingsoils. After installing the probes, no further disruption of the profile occurred. They found that theresponse of the instrument was almost instantaneous and even small water content changes can bemeasured.Baliscio and Lomax (1989) experimented with a parallel plate capacitance sensor. This sensor wasdesigned as a chamber into which the material to be tested is placed. it measures the dielectricconstant of any material placed between the plates. Compost and peat were used as test materials.5 10 15 20 25 30 35 40% Moisture (volume busTs)Literature Review 31The capacitance was measured using a modified Wein bridge circuit similar to one described byLayman (1979). A predetermined weight of air-dried compost or peat was placed in plastic bags, andwater was added to bring the batch to the desired moisture content. Uniform distribution of water withinthe sample was achieved by thorough mixing and by use of a microwave procedure reported by Hortonetal. (1982).Baliscio and Lomax (1989), found volumetric water content to be correlated with two circuitparameters: capacitance of the filled chamber, and frequency of the measurement a.c. signal. Thefollowing quadratic surface, with R2 = 0.999, was fitted to their peat sample data points for moisturecontent up to 0.591 m31m= 0.004736— 0.03075C + 0.04182f + 0.03125C2— 0.05596fC + 0.1418f2 (3.5)where,= volumetric water content (m3/),C = capacitance (nF), and= frequency (MHz).The following linear model with R2 = 0.965 was also fitted to their set of data points for the compostsample for 0 up to 0.71 04 m3/m= 0.200949 C — 0.0436 (3.6)where the terms are as defined in Equation 3.5. It was however not made clear if the data were fromrepeated testing of the moisture sensor or from just one test. They however pointed out that thecapacitance of the filled chamber was found to be highly sensitive to the test frequency — the testfrequency was manipulated as part of the capacitance measurement procedure.Literature Review 323.3 The Time-Domain Reflectometry (TDR) Method of Measuring Dielectric ConstantTime-Domain Reflectometry(TDR), a technique operating over a range of radio frequencies whichcan be used to measure the high-frequency electrical properties of materials, has been applied tomeasuring soil water content, both in the field and laboratory (Davis, 1975; Davis and Chudobiak, 1975;Topp eta!., 1980 a,b, and c).In soil application, TDR is used to measure the dielectric constant of the soil. In the TDR technique,a step voltage pulse is propagated along a transmission line. The signal’s propagation velocity and thepolarity of the reflected signal are dependent upon the electrical properties of the materials making upthe transmission line. Parallel pair transmission lines are usually used for measuring soil moisturecontents. The parallel rods or wires serve as conductors and the soil in which the rods are insertedserves as the dielectric medium. The pair of rods acts as a waveguide and the signal propagates asa plane wave in the soil. The signal is reflected from the end of the transmission line in the soil andreturns back to the TDR receiver. The signal propagation velocity is dependent on the volumetricmoisture content of the soil (Topp and Davis, 1985).3.3.1 Propagation Velocity and Water ContentThe dielectric is related to the signal propagation velocity byv = cI{K’ [1 + (1 + tan2)°5]I2}112 (3.7)where c is the propagation velocity of an electromagnetic wave in free space (3 x 108 mIs). The lossLiterature Review 33tangent is given by[K”+(3.8)tan8=K’where K’ is the real component of the dielectric constant, K. K” is the imaginary component of thedielectric constant, 0d, is the direct current conductivity of the dielectric, o is the angular frequency(2itf), and e,, is the free space permittivity (8.854 x1012 F/rn). In soils, tan S is usually much less than1 so that K’ K; thusv — (3.9)K112where Kis the overall dielectric constant. According to Topp etal., (1985) equation (3.9) is satisfactoryfor all soils sampled.Topp and Davis (1980) found that in general, the dielectric constant is given byI 2K=12!I (3.10)L. 2 L)where K is the dielectric constant, t is the propagation time, c is the velocity of light in vacuum, and Lis the waveguide length. Topp eta!., (1980 a), fitted the following empirical equation to their data=-5.3x10 + 2.9 x 10K -5.5 x 104K2 + 4.3 x 10K3 (3.11)where 0,, is the volumetric water content.In further studies they found that the empirical equation (3.11), above applied not only for coaxialtransmission lines, but also for parallel pair transmission lines placed in the soil (Topp and Davis, 1982).Literature Review 343.4 Capacitor-type Moisture Sensors in Industrial Process Control3.4.1 Capacitance-moisture Measurement in Slurries, Pastes and EmulsionsCapacitance sensors are used in moisture measurement in slurries, pastes and emulsions. Thecapacitance electrodes can be placed in the walls of the pipes and bases of chutes and belts, andhence the technique is applicable to this range of materials. However, most gases, including air havevery low dielectric constants compared with solids, and if entrained in the material, cause major errorsin the measured moisture content. Correction for entrained gas may be possible with slurries bymeasuring the capacitance at two different pressures. This technique is good for determining relativelyhigh water contents of slurries, pastes and emulsions (Carr-Bion, 1986).3.4.2 Capacitance-moisture Measurement in SolidsThe dielectric constant of water is considerably higher than that of most other materials and thisfactor is used to determine the moisture content of solids such as natural vegetable products which donot vary widely in composition. The dielectric constant also depends on the bulk density and chemicalcomposition of the solids, (especially varying concentrations of ionic conductors such as salt) and it isfor this reason that the method is mainly of value with materials of roughly constant composition.The capacitance is determined in most cases by high precision bridge techniques, with built-incompensation for variations in material temperature, bulk density and electrolytic conduction. In suitableapplications this method has a limit of ±0.2 per cent moisture over the range up to 30% moisture. Inthe 30 - 60% moisture content range, repeatability is between 0.5 and 2.0 % moisture, (Carr-Bion,1986).This use of capacitors requires electrically isolated conductors very close to the material beingmeasured. This method has been used to determine the moisture content of a wide range offoodstuffs, tobacco, chemicals, sugar beet pulp, fertilizers, drugs, soap flakes, powdered coal, sands,Literature Review 35wood and textiles. The capacitance is normally measured at radio frequencies: 2 - 12 MHz beingtypical (Carr-Bion, 1986).3.5 Porous Material Conductivity Measurement Methods3.5.1 Direct Soil Conductivity Measurement MethodsElectrical conductivity of porous materials varies with water content. Electrical resistance ofmaterials can ordinarily be measured with great precision; and if a reliable correlation with water contentexisted, moisture content measurements based upon this principle would have considerable utility.Unfortunately, such measurements made directly in the soil have not resulted in unique correlations withwater content and have therefore not come into general use. The major obstacles to the successfuluse of direct electrical resistance methods, seems to be soil heterogeneity, which prevents uniform flowof current in the soil mass, and uncertain electrical contact between electrodes and soil (Gardner,1986).Gardner (1987), also points out that the problem with this method is that is not the water itself thatis conducting the electricity, but the ions dissolved within the water and the double-layer ionsassociated with surfaces. He noted too, the fact that conductivity is also used as a measure of soilsolution concentration.3.5.2 Porous Block Conductivity MethodsElectrical conductivity methods made in porous blocks inserted in the soil yield far more dependableresults than those made directly in the soil. The electrodes are embedded in the porous blocks a fixeddistance apart. Porous blocks now available are made of a variety of materials ranging from nylon clothand fibreglass to casting plasters and numerous others, the most common being some form of gypsum.Literature Review 36Blocks are often used to indicate water content of a soil even though precision in such use is ratherlow. However, the popular use of porous blocks likely stems from their utility as indicators of waterconditions favourable for, or unfavourable to plant growth rather than their ability to indicate soil watercontent (Gardner, 1986).According to Topp and Davis (1985), although gypsum blocks are the least expensive option in soilmoisture measurement, they require frequent calibration to give measurements that are better thanqualitative. There are numerous sources of error involved in the measurement of moisture content byresistance blocks and it appears that precisions better than ±2% water content should not be expectedand errors as great as 100% are possible (Gardner, 1986).3.5.3 The Fibreglass Resistance InstrumentA simplified design of a fibreglass resistance soil moisture sensor is shown in Figure 3.3. Basicallyit consists of a fibreglass (or nylon cloth) sandwiched between two conductors. The electricalresistance of the fibreglass changes with its moisture content. Thus moisture content can be calibratedagainst resistance or current flow. The main difference between this and the capacitance method isthat the capacitance chamber method measures the capacitance rather than the electrical resistancebetween the two plates.A more elaborate design of a fibreglass resistance soil moisture sensor is shown in Figure 3.4. Afibre-glass soil-moisture meter consists of two parts: the soil unit (or the soil cell) and the meter unit.The soil unit is intended to be buried permanently at the point where moisture measurement is required.The meter unit is used to measure the electrical resistances of those elements of the soil unit withwhich soil moisture and temperature can be determined.The sensor (Figure 3.4) consists of a moisture-sensitive element enclosed in a monel case. Themoisture sensitive element is a sandwich composed of two monel screen electrodes separated by twoLiterature Review 37metal platesFigure 3.3 Construction of a simple fibreglass moisture sensor.berg I asslayers of fibreglass cloth (or nylon) and wrapped around with three layers of the same material. Theelectrical resistance between screens (electrodes) through the fibreglass cloth, varies in response tochanges in moisture content of the soil in which the unit rests.The monel case of the sensor (or soil cell), consists of two identical half-shells of 25-gage metalwhich are die-formed to ensure dimensional accuracy and rigidness. When assembled and spot weldedalong the flange-edges, the case serves to compress the fibreglass uniformly and good capillary contactbetween soil and fibreglass is ensured by the thinness of the metal (Colman and Hendrix, 1949).electricalconnectorto plate38casef berg as3 layermonel screenfiberglas2 layermone screenelectricalconnectors/holesfor contactbetween soiland fiberglassfiberglass3 layersFigure 3.4 Construction of fibreglass soil moisture cell. (Source: Colman and Hendrix, 1949)Chapter 4MATERIALS AND METHODSThe experiments for this research were divided into two main sections as follows:(a) PART 1 - Design, construction, and testing of a capacitance chamber.(b) PART 2 - Design, construction and testing of the capacitance moisture sensors and theircomparison with the fibreglass resistance cells.The objectives of each part, the materials used, and the respective experimental procedures areoutlined below.4.1 PART 1 - Design, Construction, and testing of a capacitance chamber.4.1.1 Objectives(i) To design and construct a chamber which would measure the capacitance and hence themoisture content of any sawdust sample placed in it, as a rapid alternative to the laboriousand time-consuming gravimetric oven method.(ii) To determine whether the use of a low test frequency (400 Hz) significantly increases themagnitude of the capacitances obtained (i.e. compared to the similar experiment by Baliscioand Lomax, 1989).(iii) To test the reliability of the capacitance chamber in predicting moisuture content by39Materials and Methods 40(a) determining the variation of the chambers capacitance with increasing moisturecontent of the sawdust sample inside it.(b) determining whether repeatability of the moisture content-capacitance data pairs isvery sensitive to changes in bulk density of the sample, and to unavoidable spatialvariation of density within each test sample.4.1.2 The design of the chamberThe capacitance chamber will indirectly measure the moisture content of a sawdust sample that isplaced within it, by measuring the capacitance of the chamber containing a sample. The box-like shapeof the chamber was selected in order to have a valid basis of comparison of the results with thoseobtained by Baliscio and Lomax (1989); the capacitance readings using a very low test frequency (400Hz in this experiment) will be compared to the results from using a higher frequencies (.02435 MHz -8.366 MHz in their experiment). Their capacitance measuring chamber had two copper plates,dimensions 30 cm by 30 cm at 1 .0 cm apart i.e. the effective area of the plates is 900 cm2. Sincehigher capacitance readings were expected, it was decided to use an effective plate area of at mosthalf of 900 cm2 (30cm x30 cm). Therefore 18cm x25 cm (i.e. 450 cm2) was selected. Even thoughthe effective area of the plates of this chamber is half the size of that used by Baliscio and Lomax(1989), a valid basis of comparison still exists because the capacitance of a parallel plate capacitor isdirectly proportional to the effective area of the plates. Therefore, all other conditions being equal, thecapacitance of the chamber used in this experiment should be half that obtained in theirs. Values muchlarger than half of what they obtained would suggest that the combination of a low test frequency andappropriate insulation of the plates can make possible small in situ capacitance moisture sensors.Since the water-soil/sawdust-air mixture will conduct electricity, the capacitance readings obtainedwould be inaccurate and lower than if there was no conduction between the plates. To avoid using acomplex circuit to compensate for the conductivity of the dielectric between the capacitance plates,Materials and Methods 41electrically insulated aluminium plates were used. Insulation was achieved by wrapping each plate withelectrical insulation tape. A cross-sectional view of the chamber is shown in Figure 4.1. A non-conductive coating (or paint) like that used for the in situ sensors could also be used, but there was notenough time to test one, given that a lot of time had already been spent on the original experimentwhich was suspended in order to solve some of the problems encountered in the course of it. Howevera paint could be used in place of the electrical insulation tape, in a future experiment.The chamber has a lid which consists of an insulated aluminium plate attached to a rigid plasticblock. The sides of the chamber are also made of 1 .2 cm thick perspex. The base of chamber, likethe lid, consists of an insulated aluminium plate fixed to a rigid plastic block. Rigid perspex glass wasselected because non-rigidity of the chamber will give unstable readings as the distance between theplates will vary at the application of the least external pressure, thus affecting the capacitance of thechamber.The sample is placed in a sample bag (dimensions 25cm x18 cm), before placing in the chamber.This procedure will ensure that all the water added at each wetting of the sample can be convenientlymixed into all of the sample by shaking the bag. Given the very small depth of chamber it would bevery difficult to mix up the sawdust after each wetting without losing some of the sample. To furtherensure that no deflections of the plate could occur, each time, before closing the lid, the sawdustsample would be shaken and conditioned to ensure that the lid of the chamber does not touch it duringmeasurements.Figure 4.1. shows the cross-sectional view of the capacitance chamber with sawdust samplebetween the lid (the upper plate) and the base (the lower plate). The dimensions of the lid various partsthe capacitance chamber are shown in the Figure 4.2.cJna,——a,a,-o—-a—a,a,a,—aaa,a,a,I—a,c-,0—-—a—a,—‘-,=.a,=—00a,-oa,-,0a•0-oa,a,a,a,0a_____-0-óa)ECu0-DCCua)0CuC-Q-DCCua)-QECu0a)0CCuC-)Cu0Cu00a)>CuC0C)a)U)C,)C’)0C)CuC)ILIa,a,a0Materials and Methods 430.5cm1,2cm0.5cmT1.5cmT18cmt1.5cmT28 cmFigure 4.2 Dimensions of chamber; (a) side view and (b) plan of lid or base.0.1cm‘— k—i,OcmHlast iinsulatedI am in i umplate4.1.3 Materials(i) Capacitance meter: EMCO Model DMR- 2012 Digital multimeter with capacitance range.Resolution is 1 pP Measuring frequency is 400 Hz. Measuring range is 1 pF- 21 pP(a)I. 25cm(b)Materials and Methods 44(ii) Mist Bottle: 1 litre bottle with mist sprayer.(iii) Weighing scale: Electronic Beam balance.The dimensions of the capacitance chamber are shown in Figure 4.2.4.1.4 Procedure(i) The capacitance chamber was electrically connected to the digital capacitance meter.(ii) About 40 g of air-dry sawdust was weighed into an empty pre-weighed plastic sample bag,and placed in the capacitance chamber. The lid of the chamber was put in place andclamped as shown in Figure 4.1.(iii) After reading the capacitance of the chamber, it was unclamped, the lid was removed andthe moisture content of the sample increased by spraying some water onto it with the aid ofthe mist bottle. The sample bag (with the wet sample) was weighed for its moisture contentcalculations at the end of the experiment. After that, the sample was shaken together in anattempt to distribute the added water evenly within the sample;(iv) The sample bag was sealed and placed in the chamber, arranging it in such a way that thelid of the chamber did not touch the sample when the chamber was covered. After placingthe lid and clamping it in place, the capacitance of the chamber was again measured withthe digital capacitance meter;(v) The sample was then removed from the chamber and shaken vigourously (in order to rearrange the particles in a different way) and returned to the chamber for another reading ofthe capacitance of the chamber. In this way five readings of chamber capacitance weredetermined for each moisture content point. The capacitance readings were repeatedbeacause the main objective of this experiment is to determine whether the capacitance ofMaterials and Methods 45the chamber will remain essentially constant, irrespective of how the sample is arranged init (i.e. the capacitance of the chamber is dependent only on the quantity of water betweenits plates).(vi) After taking the five readings of capacitance, the moisture content of the sample was raiseda little, and steps (iii) to (v) repeated untill the sawdust was compltetly saturated. Saturationwas assumed at the point when after adding more water to the sample and having shakenit to mix the water in, droplets of water remained visible on the inner surface of thetransparent sample bag.4.1.5 Method of Analysis of Data(i) A scatter graph of capacitance versus moisture content (% of dry weight) was plotted for theall the data pairs obtained. A graph of the average capacitance-moisture content curve wasalso plotted on the same graph. The deviations of moisture content measurements from theaverage line were examined for some of the points as an indication of the possible errorsof measuring sawdust moisture content with the capaciance chamber.(ii) The minimum and maximum determined capacitance of this chamber were compared withthose obtained by Baliscio and Lomax (1989) in their experiments to determine whether theuse of a lower signal test frequency significantly increases the capacitance of the chamber.(iii) Determine from the graphs whether capacitance of the chamber increases with increasingmoisture content, as expected from theory and other experiments reported in the literature.Materials and Methods 464.2 PART 2 - Design, construction and testing of the capacitance moisture sensors4.2.1 ObjectivesThe objectives of this experiment are,(i) To design, construct and test a capacitance-type of moisture sensor for in situ moisturecontent determination of small surrounding volumes of moist organic soil and sawdust.(ii) To determine if easily measurable capacitances can be obtained from the capacitancesensors in (i).(iii) To determine the smallest changes in moisture content that can be detected by thecapacitance sensors and compare this to the moisture content resolution (in sawdust andorganic soil) of a commercially available fibreglass resistance moisture meter - The SoilTestMC-3 13 Model moisture meter.(iv) To compare the sensitivities (to changes in moisture content) of the capacitance moisturesensors in (i), to those of commercially available fibreglass resistance moisture meter - TheSoilTest MC-3 13 Model moisture meter.4.2.2 Materials(i) An ample supply of 1 mm thick aluminium plate;(ii) 16 #4-40 screws and nuts;(iii) About 5 meters of AWG 24 insulated copper wire.(iv) Slab of 0.5 cm thick perspex, about 10 cm x10 cm.(v) A metal-plastic bonding glue, e.g. Lepage”Lepage is the manufacurer of a variety of commercial epoxy glues that bond metals, ceramic, glass, plastics etc.Materials and Methods 47(vi) Varathane24.2.3 ConstructionVarious shapes and designs, using a variety of materials were tried out in the laboratory, but arenot recorded here because they are considered to be outside the scope of the objectives of this paper.The procedure that gave the best results during preliminary tests, is described below.Eight pieces of 2.5 cm x3.0 cm aluminium plate pieces were cut out of a larger piece. This sizewas selected as reasonable for in situ moisture content determination, since the commercially availablefibreglass cells measured 3.8 cm x2.7 cm on the largest side.Eight plastic blocks measuring 0.5 cm x0.5 cm x2.5 cm were cut from the original larger piece.Two holes, (used for attaching the cable and connector), were drilled in each small piece of aluminium.The aluminium plates were insulated with varathane and each sensor put together as shown inFigure 4.3.4.3 Comparison of the two types of in situ moisture sensors4.3.1 Apparatus(i) Four plastic pots with perforated sides, dimensions: base: 7.4 cm x7.4 cm, height: 9 cm,top: 10cm xlO cm(ii) Sawdust;(iii) Organic Soil;2 Varathane is a registered trade name of a wood finishing system manufactured by Flecto Coatings Ltd. Richmond,B.C. Canada.Materials and Methods 48/0.1cme I ec tr I caconnectortopIansu ateda um n I umplate(c)Figure 4.3 Construction of capacitance moisture sensor.(iv) Four Fibreglass resistance moisture cells;(v) Four Capacitance moisture cells;(vi) One Digital capacitance meter (range: 1 pf - 20 sit)(vii) Fibreglass moisture meter;(a) [420.6cmcopperplastic blockMaterials and Methods 49(viii) Digital multimeter (for use with fibreglass moisture meter).4.3.2 Procedure(i) Two of the plastic pots were filled with soil to a uniform depth of about 1 cm;(ii) The gap between the electrodes of each capacitor cell will be filled with soil, taking care topack the soil to about the same density as the soil into which the cell is to be placed;(iii) One filled capacitor cell and one fibreglass cell were then positioned near the center of a soilpot while filling the rest of the pot with soil to a total depth of about 7.5 cm as shown inFigure 4.4;(iv) Step (iii) wiH be repeated with another pair of cells and steps (i) - (iii) for sawdust;(v) The soil in each pot will be wetted with water to saturation and then allowed to drain to fieldcapacity (i.e. allowed to drain for 2 days) in a moisture chamber(vi) The pots are taken out of the moisture chamber and the capacitance or resistance of eachcell will be determined, and the pot weighed at least once every 24 hours as thesoil/sawdust dried out, until the medium becomes fairly dry (i.e. very little changes in totalpot weight with time);(vii) When the medium sample in each pot had dried (i.e. no more changes in weight with time),it will be re-saturated and steps (v) - (vi) repeated;(viii) The cells in each pot were then taken out and weighed while each sample will betransferred to a moisture can for overnight oven-drying at 105°C;(ix) Steps (i) - (viii) were then repeated with fresh samples of each media.Materials and Methods 50cdble tocpc tncemetersoi orw d u scpc I tncesensorFigure 4.4 Placement of a fiberglass resistance sensor and a capacitance sensor in the same pot oforganic soil or sawdust.cable toml cromrnetermeterELZE7L i berg I assres stancesensorp1 ast IcpotMaterials and Methods 514.3.3 Method of Analysis of Results(i) For the cells in each pot, a graph of soil moisture content versus capacitance or current (orresistance) were plotted for all the different experimental conditions, i.e.(a) Experiment I: Using the same sample in each pot until sample became dry;(b) Experiment II: Repeating experiment I by re-saturating the sample when it became very dry;(C) Experiment III: Repeat of experiment I with a fresh sample of soil sawdust in each pot;(d) Experiment IV: Repeat of experiment III using the re-saturated sample of experiment Ill.(ii) In order to have a valid basis of comparison of the sensitivities to changes in organic soil andsawdust moisture content of the two types of sensors, graphs of counts of meter (or “resolutionunits”) versus moisture content would be plotted. The count of a digital meter is used here to referto a stepwise change in the meter indicator in response to a change in the variable being monitored.The above graphs would be compared on the basis of changes in sensitivity (as defined below),repeatability, and any other observable trends over specified ranges of moisture content.The sensitivity of the capacitance and resistance sensors to changes in moisture content will becompared by the following sensitivity index:LoS = __:.Au (4.1)Au = .L =r rwhere,s = Index of sensitivity (% moisture content),0,, = moisture content (% of volume of oven-dry soil or sawdust),Materials and Methods 52= the average change in capacitance of the capacitance sensor over a specified range ofmoisture content (nF),= the average change in the current through the fiberglass resistance moisture sensor, overa specified moisture content range (pA),r = the resolution of the particular range of the meter used for measuring capacitance orcurrent (nF or pA), and= changes in capacitance or current meter readings, in units of “resolution” (dimensionless)i.e. a change of 3 pA measured on a scale with a resolution of 0.1 pA, is equavalent to30 units.This sensitivity index, s, could be defined as the reciprocal of the resolution of each moisturemeasuring system as a whole (i.e. the indicating meter plus the moisture sensor). It indicates thesmallest change in moiture content that can be detected by either the capacitance moisture contentinstrument or the reisistance type. Therefore a higher s implies a lower sensitivity to changes in soilor sawdust moisture content.Chapter 5RESULTS AND DISCUSSION5.1 Abbreviations and Terminologies Used in the DiscussionTest 1: Two fresh samples (one sawdust, and the other organic soil) were used.Test 2: The samples used in test 1 were re-saturated and used for this test.Test 3: Two new samples (one sawdust, and the other organic soil) were used for this test.Test 4: The samples used in this test were the re-saturated samples from test 3.Cl: Capacitance sensor #1C2: Capacitance sensor #2C3: Capacitance sensor #3C4: Capacitance sensor #4Ri: Resistance sensor #1R2: Resistance sensor #2R3: Resistance sensor #3R4: Resistance sensor #4Dielectric sawdust or soil: Used to refer to that part of the soil or sawdust sample containedbetween the plates of the capacitor.53Results and Discussion 54m.c = moisture content.Err = absolute error of estimation of m.c. (m31)R2 = regression coefficient as computed by Quattro-Pro Spreadsheet.Moisture exchange area = the total area of contact between the dielectric soil, or sawdustand the surrounding soil or sawdust.C Capacitance, in nano-farads= moisture content (% of dry volume)I = current in micro-amperes.a = the slope of the capacitance-Ow curve of a particular sensor within a specified e range.s = the sensitivity index for a capacitance sensor, as defined in equation (4.1).5.2 Important Notes1. The 200 nF range of the digital capacitance meter was used for measuring the capacitance of thecapacitance sensors in sawdust. The resolution of this range is 0.1 nF;2. The 2000 nF range of the digital capacitance meter was used for measuring the capacitance of thecapacitance sensors in organic soil. The resolution of this range is 1 nF;3. The 100 pA range of the digital multimeter was used to measure the current through the resistancesensors. The resolution of this range of the meter is 1 pA.4. The SOILTEST soil-moisture instrument (Model MC-305B) is fitted with an ohmmeter scale reading0 to 1 .5 M2 with its resolution ranging from a minimum of 5 K1 at the zero end of the scale to 500Mi2 towards the end of the scale. According to the manufacturer, an alternating current of 93 Hzis generated in the solid-state circuitry, passed through the soil cell (or sensor) and rectified forResults and Discussion 55indication on a d.c. micro-ammeter. Thus the manufacturers readout of this instrument is actuallya micro-ammeter re-calibrated to read resistance of the fibreglass sensors in ohms (ci). Duringpreliminary testing of this instrument in sawdust, it was observed that the pointer of the analogmeter (manufacturers) was 0 Q at saturation point of the sawdust. During a typically calibration itincreased slowly to about 10 Kc2 and remained there until the sawdust was almost completely dry,at which point the reading suddenly drops to K The Department electricians thereforeconnected a digital micro-ammeter at an appropriate point in the circuit to measure directly thecurrent through the moisture sensors, in an attempt to monitor whatever minute changes there mightbe in current when the sawdust is nearly saturated. The digital meter was easier to read with lesslikelihood of reading errors.5. Temperature is not mentioned in the discussions because the temperature throughout theexperiments was virtually constant at about 24 °C.6. In Figure 5.1 to Figure 5.20, it should be noted that the linear regression lines are fitted only for them.c. data in the range 10% to 45%, since this is the range of moisture content normally ofmonitored in drip/trickle irrigation.5.3 In situ Sawdust Moisture Content Measurement by Capacitance SensorsFigure 5.1 and Figure 5.2, show the graphs of sawdust moisture content versus capacitance ofcapacitance sensor Cl and C2 respectively. For both sensors Cl and C2, capacitance decreased withdecreasing sawdust moisture content.Test 1 and Test 2 of both sensors (Cl and C2 in sawdust) produced essentially repeatable datapairs up to about 30 % m.c. Beyond 30 % m.c., the capacitances obtained in test 2 in both cases wassignificantly higher than those for test 1. The difference between the two curves was greatest at about40 % m.c. and gradually decreasing to almost zero as m.c. approached 50 % m.c.Results and Discussion 56The capacitances of Cl and C2 in test 4 were at all m.c. points greater than those I or test 3, exceptfor the crossover of the curves in the case of Cl (please see Figure 5.1). Apart from that, the curvesfor test 3 and test 4 are similar in shape. The main difference therefore seems to be a an upward shiftin the calibration curves from test 3 to test 4.The observed general pattern of decreasing capacitance with decreasing moisture content agreeswith the results of similar experiments reported by Kuraz eta!. (1970) and De Plater (1955) (please seeChapter 4- Literature Review). They however did not describe the repeatability of the data pairs,except that Halbertsma et a!.(1 987) reported an accuracy of better than 0.02 m3/m, provided that nochanges in matrix density occurred as is the case in swelling soils. But the sawdust and soil used inour experiment are both swelling types.The shifts and differences in the curves in Figure 5.1 and Figure 5.2 could be explained bydifferences in the packing density of the dielectric sawdust in each test as explained below.Results and Disc55100cJC,0o 20100Figu, 5.1C0C-)0C)000c 201005760Four caIibratjo Curves for Ci Ifl SawdustFigure 5.2 Four calibration curves for C2 in sawdustResults and Discussion 585.4 Likely Factors Causing Differences in Calibration CurvesThe capacitance moisture meter measures the capacitance of the composite dielectric (made upof sawdust or soil, air, water and the capacitor plate insulation) between the plates of the capacitancesensor. Theoretically therefore in order to obtain the same capacitance at a given moisture content,then the dielectric constant of the dielectric soil or sawdust must be the same at each moisture levelduring all tests of each sensor. This condition was not satisfied during the tests of the capacitancesensors due to the following factors:5.4.1 Differences in Dielectric Sawdust Packing DensityCare was taken to pack the dielectric sawdust into the gap between the plates of the capacitor,without compressing it. No special technique or instrument was employed for this. It is thereforeexpected that differences in the degree of compaction (packing or matrix density) would occur from testto test. Since the volume of the dielectric gap is a constant (3.5 cm3), a more highly compacteddielectric sawdust in the gap implies a smaller partial volume available to the air but more to sawdustand water.According to the theory of Odelevskii (equation 4.1), the partial volumes of the air, water andsawdust determine the overall capacitance of the capacitor. Sawdust (woOd) has a dielectric constantof 2.5 to 8.5, that of air is 1 .0 and that of water is 81 (please see Table 2.1). Since the moisturecontent of the sawdust is mostly the water contained within the particles, changes in packing densitywill result in more wood particles and therefore more water per unit volume of the dielectric gap,increasing the overall dielectric constant by a certain factor of 81. Also more sawdust per unit volumewould also increase the dielectric constant by a certain factor of 2.5 - 8.5. Therefore changes in thedielectric sawdust compaction influence its dielectric constant and hence the capacitance of the sensorcapacitor.Results and Discussion 595.4.2 Contraction of Dielectric Sawdust During DryingBoth the sawdust and the organic soil used expand with increasing moisture content and vice versa.During the preliminary experiments, it was observed that the tightness of the contact between thedielectric (i.e. soil or sawdust) and the insulated plates of the capacitor greatly affects the value of themeasured capacitance. A saturated dielectric organic soil or sawdust is at its maximum volumetherefore presses firmly against the capacitor plates. The contraction that occurs as it dries out,reduces the tightness of the contact between the plates and dielectric. The coefficient of expansiondepends in part on the initial bulk density of the dielectric. This could partially account for the slightdifferences in slope of the various m.c. versus sensor capacitance curves.5.4.3 Moisture Equilibrium between Capacitor and Ambient SawdustThe moisture exchange area is the total surface area of contact between the dielectric sawdust andthe surrounding sawdust. The area of contact between the dielectric soil and the surrounding soil isthe major limiting factor of the rate of moisture exchange between the capacitor and the surroundingsoil. The resistance to flow of water due to a moisture content differential, is directly proportional to thetotal moisture exchange area.The volume of the dielectric sawdust for each capacitance sensor is 3.5 cm3 with a moistureexchange area is 2.8 cm2. That means 0.8 cm2/c3of dielectric sawdust. This is an indication of therate of moisture migration in and out of the capacitor in response to moisture content changes in itssurrounding medium. Increasing this exchange area per unit volume would increase the rate ofattainment of moisture equilibrium between the dielectric sawdust and its surrounding. Thus themoisture content of the dielectric sawdust will be more nearly equal to that of the surrounding sawdust.This proposed modification is supported by the fact that at the time of removing Cl and C2 fromthe samples (for ovendried weight determinations) the surrounding soil was visually dry and also hadResults and Discussion 60a dry feel, while the dielectric sawdust was visibly wet compared to the surrounding sawdust. Thereforethe moisture content of the dielectric sawdust was not in equilibrium with that of its surroundings. Thuswhile in Figure 5.1 and Figure 5.2, the capacitance axes represents the capacitance corresponding tothe wetter dielectric sawdust, the moisture content axis essentially, represents the moisture content ofthe drier surrounding sawdust.This explains in part, the differences in the calibration curves shown.Due to lack of time, the effect of this modification could not be tested. Some proposed changes to thesensor design for the next experiment can be found in Chapter 6 -Conclusions and Recommendationsfor further research.5.5 In situ Organic Soil Moisture Content Measurement by Capacitance SensorsFigure 5.3 and Figure 5.4 contain the curves of the variation of capacitance with organic soilmoisture content for capacitance sensors C3 and C4. Each curve was obtained using a differentsample of organic soil.In both Figures the curves for C3 and C4 lie closer together for m.c. up to about 40 %, indicatinga high repeatability of the data pairs within this m.c. range. The high level of repeatability of thecalibration curve for organic soil in this region for C3 is indicated by the linear regression analysisperformed (using the Quattro Pro Computer Programme) on all the data points generated for all thetests of C3. The linear model is= 0.246C—7.621 (5.1)with R2 = 0.932 and error of estimation, Err = ± 0.024 m3/m.A similar model fitted for C4 iswith R2 = 0.897 and Err = ± 0.030 m3/m.Results and Discussion 61= O.282C—1O.735 (5.2)Equations (5.1) and (5.2) are similar to equation (4.7) obtained by Baliscio and Lomax (1989) using peatin their capacitance chamber.From the above discussion it is obvious that the capacitance moisture sensors in their currentdesign, give much more reproducible results when used in organic soil than they give in sawdust.However eliminating the effects of the differing packing densities of the dielectric organic soil (as in thecase of sawdust discussed above) is expected to improve the accuracy of estimation for organic soiltoo. The discussions of the dielectric sawdust in section 5.2 apply to organic soil too, since it alsoexpands upon absorbing water.Results and Discussion 62500400LC300a,Ci0C)200C-)1000Moisture Content (% volume)Figure 5.3 Four calibration curves for C3 in organic soil.350300250a, 200• 150a0C)100500Moisture Content (% volume)Figure 5.4 Four calibration curves for 04 in organic soil.Results and Discussion 635.6 Mathematical Relationships between Capacitance of Capacitance Moisture Sensors andSawdust and Organic Soil Moisture ContentExamining the calibration curves individually, it was found that the relationship between themeasured capacitance of the capacitance sensors and the sawdust or organic soil moisture contentswas a linear one. This is proved by the very high linear regression coefficients (R2) obtained in allcases ranging from 0.9 to 1 .0 for both media.All the linear regression models described below for the particular tests of the individual capacitancesensors, are of the form(5.3)where,= volumetric moisture content (102m3 water / m3 soil)= change in moisture content per unit change in O (102m3water / nF)13 = a constant (102m3water / m3 sawdust)C = capacitance (nF)The sensitivity of the capacitance moisture instrument to changes in 0, is the rate of change of thecapacitance of its sensor (or probe) to changes in 0, i.e. dC/d0dC = d0 = 1 (5.4)d0 dCwhere c is the slope of the capacitance-Ow curve of a particular sensor within a specified 0,, range.Therefore c is the same as the coefficient of C in the following discussion of the linear regressionmodels for the capacitance sensors in sawdust and organic soil.Results and Discussion 645.6.1 Linear Regression Models For the capacitance of Sensor Cl in SawdustThe regression lines through the data points for Cl in the four tests are shown in Figure 5.5 toFigure 5.8.The equations for the regression model shown in Figure 5.5 for test 1 of Cl is= 4.821 C—0.723 (for O<26%) (5.5)with R2 = 0.951 and Err = ± 0.013 m3/m; and= 1 .032 C +20.378 (for 26%<O<40%) (5.6)having R2 = 0.994 and Err = ± 0.003 m3/m. A sudden change in the slope of the regression lineoccurred at about 26 % m.c. The regression model for test 2 of Cl shown in Figure 5.6 is0, = 6.600C—8.853 (for 9<26%) (5.7)with R2 = 0.971 and Err = ± 0.009 m3/m; and0, = 2.252C+13.791 (for 26.<0<40%) (5.8)with R2 = 0.995 and Err = ± 0.003 m3/m. The equation of the single regression line fitted to the datafor test 3 shown in Figure 5.7, is= 2.114 C—8.363 (for 0<45%) (5.9)with R2 = 0.986 and Err = ± 0.013 m3/m. The test data for test 4 also followed a single linearregression line as shown in Figure 5.8. Its equation isResults and Discussion 65= 1.740C-10.894 (for 8<40%) (5.10)with R2 = 0.998 and Err = ± 0.004 m3/m.The ratio of the slope of the linear regression line 2 to that of line 1 decreased from 4.9 in test 1to 6.6 in test 2, while the slope of the single regression lines decreased from 2.114 in test 3 to 1.740in test 4.C)0a,cia0a0aC-)LiC)0a)0Ca00aaC-)15 20 25 30 35 40Sawdust Moisture Content ( volume)Results and Discussion 6620161284010Figure 5.5 Linear regression lines for test 1 of Cl in sawdust. R2 = 0.951 for line 1 and 0.994 forline 2.121086424510 15 20 25 30 35 40Sawdust Moisture Content (% volume) 45Figure 5.6 Linear regression lines for test 2 of Cl in sawdust. R2 = 0.971 for line 1, and 0.995 for line 2.Results and Discussion3025C0 20‘4-0ci,UC215UCaa010510 15 20 25 30 35Sawdust Moisture Content (% volume)Figure 5.7 Linear regression model for test 3 of Cl in sawdust. R2 = 0.986.3025I-ICo 20‘4-0a)UCaUaaa01 51054545674010 15 20 25 30 35 40Sawdust Moisture Content (Z volume)Linear regression line for test 4 of Cl in sawdust. R2 = 0.998.Figure 5.8Results and Discussion 685.6.2 Linear Regression Models For the capacitance of Sensor C2 in SawdustThe equations for the linear regression models shown in Figure 5.9 for test 1 are:= 10.9100—8.026 (for O<29%) (5.11)with R2 = 0.984, and Err = ± 0.008 m3/m ; and0, = 1.546C+23.082 (for 0<40%) (5.12)which has an R2 = 0.998 and Err = ± 0.002 m3/m.The linear regression model shown in Figure 5.10 for test 2 of C2 is= 10.541 0—4.295 (for 0<29%) (5.13)with R2 = 0.983, and Err = ± 0.008 m3/m; and= 2.447C—21.562 (for 29%.c0<40%) (5.14)with R2 = 0.996 with Err = 0.003 m3/m.Figure 5.11 shows the linear regression model for test 3 of C2:= 3.592C—26.132 (for O<25%) (5.15)with R2 = 0.943, and Err = ± 0.012 m3/m; and= 0.974C—10.181 (for 25%<0<40%) (5.16)with R2 = 1 .0 and Err = ± 0.001 m3/m.Results and Discussion 69The equation of the line shown in Figure 5.12 is= 0.982C+3.377 (for (5.17)with R2 = 0.981 and Err = ± 0.013 m3/m.The slope of the capacitance versus moisture content lines for test 1 and test 2 changed sharplybetween about 27 % and 29 % O,, becoming much steeper above 29 % m.c. Thus according toequation (5.4), sensor C2 was most sensitive to sawdust 9, changes in test 3, in the range 29 % - 40%, when sensitivity cC’ = Q974i nF/m3 or 1 .027 m3/nF. The lowest sensitivity of cC’ = 10.91 0.1 nF/m3or .092 m3/nF occurred during test 1 when 0, < 29 %. In all cases the sensitivity of C2 to changes in0, was very high at high sawdust moisture contents, ranging between 0.409 nF/m3 to 1 .027 nf/m3.Except for test 4 the sensitivity of C2 was generally lower at lower moisture contents.Even though the linear regression models for the various tests were different, the lowest R2 valueobtained for any of the tests was 0.951 with an accuracy of ± 0.013 m3/m or 1.3 % in test 1 of Cl.These figures are encouraging, and suggest that there is a great potential in these sensors for a highlyaccurate and affordable in situ moisture sensor for sawdust growth medium. The fact that therelationship between C and 0,, was found to be linear is encouraging since that suggests ease ofcalibration and errors due to calibration, would be held to a minimum. Also the need for calibrationcould be eliminated all together by simply initializing the capacitance meter with its sensor in asaturated sample of the sawdust whose moisture content is to be monitored. It is expected that morerepeatable calibration curves would be obtained after modifying the design of the capacitance sensorsaccording to the recommendations outlined in Chapter 6.5.6.3 Sudden Change in Slope of Linear Regression Models in SawdustThe break in the calibration curves for the capacitance sensors in sawdust was observed in allResults and Discussion 70cases except for tests 3 & 4 of Cl and test 4 of C2. The sudden change in the slope of the linearcalibration curves can be attributed to the expansion of sawdust with rising moisture content. Thisexpansion and contraction affects the tightness of the contact between the plates of the capacitanceand the sawdust dielectric.In preliminary experiments with the capacitance chamber, the pressure of the dielectric sawdust onthe plates was observed to increase in the measured capacitances. This observation could explain thesudden change in the fitted linear calibration curves of the sensors in sawdust. It was observed withthe capacitance chamber tests that beyond a certain limit increasing the pressure did not result in anyfurther increases in the measured capacitance at a given sawdust moisture content. Now the dielectricsawdust expands with increasing moisture content. At the start of the experiment, dry sawdust waspacked into the dielectric gap of the capacitance sensors. Now this dielectric sawdust expanded uponwetting. The expansion increases the pressure it exerts on the plates of the capacitor. The criticalpressure (i.e. above which capacitance does not increase) occurs at about 26 %. Below 26 % moisturecontent, the dielectric sawdust is no longer under pressure, resulting in a reduction of the slope C, ofthe capacitance-moisture content line. There could be other explanations of this observedphenomenon, but a full discussion of this is outside the scope of the objectives of this current study.Results and Discussion 7111109cJC-) 70a, 60CSawdust Moisture Content (Z volume)Figure 5.9 Linear regression lines for test 1 of C2 in sawdust. R2 = 0.984 for line 1 and 0.998 forline 2.87nCC%1(,‘—50ci0C.)000021Sawdust Moisture Content (Z volume)Figure 5.10 Linear regression model for test 2 of C2. R2 = 0.983 for line 1 and 0.996 forline 2.Results and Discussion 72CC%JC-)‘IaUaUaaC-)3530201520 30 35Sawdust Moisture Content (% volume)Linear regression model for test 3 of C2 in sawdust. R2 = 0.943 for line 1 and1.0 for line 2.105010Figure 5.115045-c:- 40C35‘4-0a) 30C)Ca25a0o 20151010 15 20 25 30 35 40Sawdust Moisture Content (% volume)Figure 5.12 Linear regression model for test 4 of C2 in sawdust. R2 = 0.981.45Results and Discussion 735.6.4 Linear Regression Models for Sensor C3 in Organic SoilUnlike the tests with the sawdust samples, the lines of best fit for the variation of moisture content withcapacitance of C3 in organic soil were all single lines, for the m.c. range 0 to 45 %.The linear regression model for test 1 of C3 is shown in Figure 5.13. Its equation with R2 = 0.994 andErr = ± 0.009 m3/m, is0, 0.253C+5.600 (for (5.18)The model for the data from test 2 (shown in Figure 5.14) is0v = 0.2830+7.466 (for (5.19)with R2 = 0.990 and Err = ± 0.011 m3/mThe model for test 3 is shown in Figure 5.15. Its equation with R2 = 0.980 and Err = ± 0.016 m3/mis= 0.240C—5.784 (for 9<45%) (5.20)The linear model for test 4 shown in Figure 5.16 has the lowest R2 of 0.958 and Err = ± 0.021 m3/m.Its equation is= 0.2060—9.647 (for 0<45%) (5.21)As can be seen, the R2 values for all the tests were very high ranging between 0.959 and 0.994.The differences between the slopes of the individual fitted calibration lines were not as much as wasobserved in the case of the tests with sawdust. This is probably due to the fact that the dielectricorganic soil, because of the smaller size of its particles made better hydraulic contact with thesurrounding organic soil. The slopes of the lines ranged between 3.5 to 4.6. The error of estimationof moisture content from the linear models was also very low, with the highest being 2.1 % for test 4,and the lowest 0.9 % obtained for test 1.Results and Discussion 74C4•)C-)‘I0a,UCU0C-)1601401201008060402015 20 25 30 35 40Organic Soil Moisture Content (2 vol.)Linear regression model for test 1 of C3 in organic soil. R2 = 0.994.45010Figure 5.13140120CC—)._800a,U60C,0c3 40C-)20020Organic Soil Moisture Content (% vol.)Figure 5.14 Linear regression model for test 2 of C3 in organic soil. R2 = 0.990.Results and Discussion 75CC-)00C0C)00C)C)‘I00C0C)000C)180160140120100806020010 15 20 25 30 35Organic Soil Moisture Content ( vol.)Figure 5.15 Linear regression model for test 3 of C3 in organic soil. R2 = 0.980.4016014012010080604020020 25 30 35Organic Soil Moisture Content (% vol.)Figure 5.16 Linear regression model for test 4 of C3. R2 = 0.958.Results and Discussion 765.6.5 Linear Regression Models for Sensor C4 in Organic SoilThe linear regression model for the data of test 1 of C4 shown in Figure 5.17 is= 0.267C—9.468 (for 0<45%) (5.22)with R2 = 0.994 and Err = ± 0.009 m3/m.The data for test 2 (please see Figure 5.18) had a slightly better linear model with R2 = 0.998 and Err= ± 0.005 m3/m. Its equation is= 0.384C—9.386 (for (5.23)The data for test 3 was fitted with a model (see Figure 5.19) which had a slightly lower R2 = 0.988 andErr = ± 0.011 m3/m. The linear regression model for test 3 of C4 is= 0.279C-f8.023 (for °<O%) (5.24)The linear regression model for test 4 of C4 is shown in Figure 5.200, = 0.291 C+11.273 (for 0<40%) (5.25)with R2 = 0.974 and Err = ± 0.014 m3/m, the lowest for all four tests.Comparing the linear regression models, the regression coefficients obtained for organic soil weresimilar to those obtained for sawdust. However the performance in organic soil appears to be betterin one respect - In all tests, it was possible to describe the relationship between the sensorcapacitance and organic soil moisture content by a single regression line. In both cases however thevalues of capacitance obtained were much higher than those obtained by Baliscio and Lomax (1989),thus proving that the use of small in situ capacitance moisture sensors in small surrounding volumesof soil or sawdust or other growth media is possible.Results and Discussion 770SI0a)U00a0C)C-1.C)SI0a,0C0000CC)1 20100806040200Organic Soil Moisture Content (% vol.)Figure 5.17 Linear regression model for test 1 of C4 in organic soil. R2 = 0.994.10080604020010 15 20 25 30 35Organic Soil Moisture Content (% vol.)Figure 5.18 Linear regression model for C4 for test 2 in organic soil. R2 = 0.998.40 45Results and Discussion 78C-)0a,U0C-)0t2.aC.)C-)‘I0a,0aC-)a0.aC.)1 2010080604020020 25 30 35Organic Soil MoTsture Content (% vol.)Figure 5.19 Linear regression model for test 3 of C4 in organic soil. R2 = 0.988.10080604020025 30 35Organic Soil Moisture Content (% voL)Figure 5.20 Linear regression model for C4 in organic soil. R2 = 0.974.45Results and Discussion 795.7 Sensitivity of the Resistance and Capacitance Sensors ComparedTable 5.1, shows a comparison of the sensitivities of the capacitance and resistance moistureinstruments when used with the indicated sensors. The 200 nF range (a with resolution of 0.1 nF) ofthe digital capacitance meter was used for measuring the capacitance of the capacitance sensors insawdust. The 2000 nF range (having a resolution of 1 nF) of the digital capacitance meter was usedfor measuring the capacitance of the capacitance sensors in organic soil.The pA range of the digital micro-ammeter was used to measure the current through the fibreglassresistance sensors. In order to have a common basis of comparison the scales of the digital meterswere reconverted to their uncalibrated units, where a unit is used to mean the smallest stepwise changein the meter indication. Please refer to equation (4.1) for further explanations. According to thissensitivity index, the higher the sensitivity index value of a particular moisture content instrument, theless sensitive the instrument is to small changes is moisture content.It can be seen from the table that the capacitance sensors were able to detect smaller changes insawdust and organic soil moisture content, than the fibreglass resistance sensors. Also, the resistancesensors were very insensitive to changes in sawdust moisture content. A graphical comparison of thetwo instruments is shown in Figure 5.21 to Figure 5.24. The calibration curve for the resistance metersusing Ri and R2 was unusable in the useful range of 10 to 40 % m.c. since it was virtually a straightline, parallel to the moisture content axis (please see Figure 5.21 and Figure 5.22). Thus the resistancemeter is not useful in monitoring sawdust moisture content. As can be seen, the calibration curves forthe capacitance sensors were more useful except for the fact that they were not very reproducible, dueto the factors explained earlier in section 5.2.Figure 5.23 and Figure 5.24 show a comparison of the calibration curves for the two instruments.As can be seen, the shapes of the calibration curves for both types of instruments are very similar toeach other up to about 40 % m.c. when the calibration curve for R3 and R4 become nearly parallel tothe moisture content axis. But the slope of the calibration curves for C3 and C4 become slightlyResults and Discussion 80steeper as m.c. increased beyond 40%. Thus the capacitance sensors are more sensitive to high m.c.changes at even at high organic soil moisture contents, though the differences between their predictedmoisture contents become wider.These values for the fiberglass resistance sensor R2 are correct. They are rather high because of the very poorresponse (low sensitivity) of this sensor to changes in sawdust moisture content.in the Appendix.Table 5.1 Changes in moisture content (i.e. s in equation (4.1)) that produced acapacitance and current meter reading. Calculations are based onbetween meter readings at 10 and 40 % moisture content.SENSITIVITIESunit change inthe differenceSensor Test 1 Test 2 Test 3 Test 4 AverageCl (sawdust) L2 c2 o______ c2C2 (sawdust) 03 c2____ 0.1 O3C3 (organic soil) 0.2 0.3 0.2 0.2 0.2C4 (organic soil) 0.3 0.4 0.3 0.3 0.3Ri (sawdust) 5.9 4.0 6.3 12.5 7.2R23 (sawdust) 14.7 33.3 25.0 14.3 21.8R3 (organic soil) 0.4 0.3 0.4 2.4 0.9R4 (organic soil) 0.7 0.4 0.4 0.5 0.5To verify please refer to Tables Al to A4Results and Discussion 81Variation of meter reading (in “resolution units) with sawdust moisture content,when using capacitance and resistance moisture instruments with Ri and Clconnected.Variation of meter reading (in ‘resolution” units) with sawdust moisture content,when using capacitance and resistance moisture instruments with R2 and C210 20 30 40Sawdust Moisture content (% of volume)450400C’,350300250200-oo 150- 100500Figure 5.21600C’,C.2 400CU,cC’-‘•- 300C)- 200a)Ic1a’ 1000Figure 5.2220 30 40Sawdust Moisture content (% of volume)connected.Results and Discussion 82400350C’,C300C02500C’,200150100500Figure 523 Variation of meter reading (in “resolution’ units) with organic soil moisturecontent, when using capacitance and resistance moisture instruments with R3300g 250C.2 200011)a)-.‘- 150C-Do 100a)a)a) 500Figure 5.24and C3 connected.Variation of meter reading (in “resolution” units) with organic soil moisturecontent, when using capacitance and resistance moisture instruments with R4and C4 connected.6020 30 40Soil Moisture content (% of volume)0 10 20 30 40 50Soil Moisture content ( of volume)Results and Discussion 635.8 Moisture Content versus Capacitance of ChamberOne of the main objectives of constructing and testing the capacitance moisture chamber was totest the possibility of instantly determining the moisture content of sawdust samples without having todry it in an oven to first determine the mass of water contained. To be useful for this purpose, thechamber will have to be fairly accurate. The results of the tests of the moisture content chamber aresummarized by Table 5.2 and the graph in Figure 5.25. It shows the variation of capacitance of thecapacitance chamber with increasing sawdust moisture content. The scatter of data points about theaverage of the five replicates for each moisture content value, is too large to be considered accurate.Baliscio and Lomax (1989) in their experiment with the moisture content chamber, hinted that thereadings of capacitance might be sensitive to spatial variations in density of the sample in the densityof the sample. But they did not show how serious this could be as a source of error this could be inpredicting moisture content with the chamber.Table 5.2 Data obtained from tests of capacitance chamber with sawdust.Moituie Capacitance Readings (pF) Average Standard‘° f d Capaci- Deviation‘‘eht)’ 1st 2nd 3rd 4th 5th reaings(pF)0.0 116.0 124.0 118.0 115.0 116.0 117.8 3.69.4 186.0 197.0 150.0 168.0 159.0 172.0 19.319.9 208.0 179.0 186.0 166.0 174.0 182.6 16.033.0 270.0 212.0 262.0 184.0 222.0 230.0 35.843.1 267.0 209.0 217.0 254.0 307.0 250.8 39.855.1 260.0 281.0 341.0 262.0 237.0 276.2 39.468.5 424.0 337.0 360.0 282.0 242.0 329.0 70.478.6 450.0 328.0 348.0 358.0 448.0 386.4 58.293.5 592.0 435.0 297.0 330.0 300.0 390.8 125.7114.1 1065.0 900.0 937.0 650.0 700.0 850.4 172.3133.7 1061.0 1092.0 900.0 930.0 804.0 957.4 118.8146.7 1062.0 870.0 655.0 770.0 730.0 817.4 157.2159.4 1029.0 845.0 980.0 1053.0 1074.0 996.2 91.5177.2 821.0 915.0 770.0 660.0 640.0 761.2 114.3189.9 749.0 660.0 670.0 660.0 760.0 699.8 50.3204.3 602.0 580.0 680.0 650.0 640.0 630.4 39.6Volumetric moisture content is not used for the analysis of the results of the capacitance chamber because of theinevitable changes in volumetric moisture content of the sample each time it is shaken before replacing it in the chamber.Results and Discussion 84Figure 5.25 Capacitance of moisture content chamber versus sawdust sample moisturecontent. The single line is the average value of the five replicates.The large variations of capacitance between the replicates at each moisture content, especially athigher moisture contents shows that the chamber is very highly sensitive to variations of density withinthe sample. At about 100 % moisture content (by weight) for example, the difference between thehighest and lowest readings is as large as 300 pF. Similar figures were obtained at all the othermoisture contents. This implies that, in order to use the chamber to predict accurately moisturecontent, the spatial variations of density within each sample would have to be conditioned to more orless a fixed standard. In other words, each sample has to be packed in exactly the same way withthe same pore space distribution, and other physical11009000700cDUC-IU5000C-)3001000 50 100 150 200Moisture content (% of dry weight)250dimensions, a task which is just impossible to accomplish, without special tools. It is also noted thatthe error in estimating moisture content from the average capacitance versus moisture content line istoo great for the chamber to be considered accurate. For example, at 300 pF the maximum error ofestimation of moisture content is more than 50 %.Results and Discussion 85In general capacitance increased with increasing sawdust moisture content up to about 140 % m.c.when it became unpredictable and eventually reduced with increasing sawdust moisture content. Thevalues of capacitance obtained were much higher than those obtained by Baliscio and Lomax (1989),with their twice larger chamber.The average capacitances recorded by Lomax and Baliscio (1989) ranged between 0.2 nF and 5.0nF. The capacitances recorded from our experiment (see Figure 5.25), ranged between 0.1 to about1.0 nF. This translates to a range of 0.2 to 2.0 nF when compared to the chamber used by Baliscioand Lomax since their chamber size was twice the size of ours. The order of magnitude of capacitanceis about the same as that obtained by Baliscio and Lomax (1989). It appears then that the use of alower test frequency does not in practice yield higher values of capacitance as expected from theliterature (Blech, 1989). This might be due in part to the differing materials used. They tested theirchamber with peatmoss and compost, while ours was tested with sawdust. However, a more likelycause of these results is to the kind of insulation (electrical insulation tape) used for the plates of thecapacitance chamber.This is obvious when the results for the capacitance moisture chamber are compared with thoseobtained from the miniature capacitance sensors which were insulated with a different material(Varathane). The effective area of the plates of the in situ capacitance sensors was about seventy fivetimes that of the capacitance chamber. And yet the average maximum capacitances measured for thein situ capacitance sensors was about 30 nF, which is more than 2000 times the maximum obtainedwith the capacitance chamber.Due to the sever limitation of time (the experiments started in November, 1991 and had to becompleted as soon as possible due to the impending cut in funding. Please see section 1 .2 for furtherexplanations), if further modifications (such as the use of a different insulator) could not be implementedand tested. Future experiments on the chamber should look into the effect on the measuredcapacitances by the use of different and thinner plate insulations. All the same it can be concludedfrom the results of these experiments that it is very feasible to use miniature in situ capacitancemoisture sensors to monitor the moisture content of small surrounding volumes of sawdust and othergrowth media.Chapter 6CONCLUSIONS AND RECOMMENDA TIONSThe capacitance moisture sensors respond very well to changes in soil and sawdust moisture content.Connected to the digital capacitance meter they were capable of detecting organic soil moisture contentchanges as small as 0.2 %, and sawdust moisture content changes as 0.1 %. The reproducibility ofthe calibration curves in organic soil is comparable to those obtained for the commercially availablefibreglass resistance moisture sensors. The capacitance-moisture content calibration curve for thecapacitance sensors was linear for all the sensors tested. This linearity is very important in reducingerrors in moisture content estimation resulting from calibration errors.The main problem with the capacitance sensors was shifts in the calibration curves, especiallywhen used in sawdust. This was not a serious problem in the case of organic soil. Nevertheless theperformance could be improved in organic soil too. The main cause of this problem of the capacitancesensors can be attributed to the difficulty of attaining moisture equilibrium between the dielectricmedium and the surrounding medium due to the small area of contact between the dielectric sawdust(or soil) and the surrounding media. This was more easily noticeable in the case of sawdust. Toovercome this problem, the following modifications to the design of the capacitors are proposed forfuture research. (Modifications to the existing design could not be carried out and tested due to lackof time, but they will be made and tested in a future experiment):1. The width of the capacitance sensors should be reduced from the present 2.0 cm to 0.6 cm. Thiswould reduce the effective area of the capacitor to one-third of its present value. Theoretically,since capacitance is proportional to the effective plate area, the value of the capacitancesmeasured will also be reduced by the same factor. The average capacitances would thenbetween 1 nF - 4 nF, values which would be easily measurable by the digital capacitance meterwhich has a minimum resolution of 1 pF. By this modification the moisture exchange area per unit86Conclusions and Recommendations 87volume of sawdust will increase by a factor of three, allowing for faster rates of equilibriumattainment.2. Instead of reducing the size of the sensors as suggested in (1) above, circular holes could bedrilled on the plates to allow for more contact between the dielectric soil or sawdust and thesurrounding medium.3. The distance between the plates of the capacitor could be reduced, and the inter-electrode gapfilled with a suitable material which easily absorbs water. At the same time holes would be drilledon the sides of the plates to increase the total area of moisture interchange between the capacitorand the surrounding soil or sawdust.The fibreglass resistance sensors were virtually useless in monitoring changes in sawdustmoisture content. The indicated moisture content remained essentially unchanged between saturationand just before the sawdust goes completely dry. Then the readings suddenly drop to zero at about0 % moisture content. In organic soil, the indicated moisture content varied with decreasing moisturecontent. The main problem was that the calibration curve was different with each test of the samesensor. It was also not possible to indicate differences in organic soil moisture content above about40%.The capacitance moisture content chamber gave greatly varying values of capacitance for thesame sawdust moisture content. The readings were found to be too sensitive to differences incompaction within the sample. Therefore, in its present design, it cannot be used to accuratelydetermine the moisture content of samples placed in it unless a reliable method can be found toprecondition each sample to the same density. Further research on this could be directed atdetermining the effect on the capacitance readings of compacting each sample to a specified volume.Another recommendation for further study is the replacement of the electrical insulation tape on theplates, with a thin coating of a non-conductive coating. This would greatly increase the capacitancesobtained, which might possibly improve the stability of capacitance values obtained at a given moisturecontent.List of ReferencesAnderson, A.B.C. (1943). A method of determining soil moisture content based on the variation ofelectrical capacitance of a soil at low frequency, with moisture content. Soil Science 56: 29 - 41.Ambrosius, E.A., Fellows, R.D., Brickman, A. D. (1966). Mechanical Measurement and Instrumentation.pp 41 -42. The Ronald Press Co. New York.Baliscio, C.C., and Lomax, K.M. (1989). A Comparative Study of Moisture Sensors for Use inMushroom Beds. Transactions of the ASAE vol 32(3), pp 928 - 933.Blech, l.A. (1989). “Properties of Materials”. In: Electronic Engineers’ Handbook (D.G. Fink and D.Christiansen, eds) 3rd Edition, pp. 6-1 to 6-119. McGraw-Hill, New York, U.S.A.Campbell, G.S., and Campbell, M.D. (1982). Irrigation Scheduling using soil moisture measurements.In: “Advances in Irrigation” (D. Hillel,e, vol. 3, pp 107- 127. Academic Press, New York.Carr-Bion, K. (1986). Moisture Sensors in Process Control. Elsevier Applied Science Publishers,London and New York. pp 42 - 54Colman, E.A., and Hendrix, T.M., (1949). “The fibreglass Electrical Soil-Moisture Instrument.” SoilScience (67)6: 425 - 438, June, 1949.Davis, J.L. (1975). Relative permittivity measurements of a sand and clay soil in situ. Geol. Surv. Can.,Ottawa, Paper. 75-iC, 361-365 (Report of Activities).Davis, J.L., and Chudobiak, W.J. (1975). In situ meter for measuring relative permittivity of soils.Geol. Surv. Can., Ottawa, Pap. 75-lA, 75-79 (Report of Activities).De Platar, C.V., (1955). A portable capacitance-type soil moisture meter. Soil Science vol 80, pp 391- 395.Dudley, B. (1989). “Basic Phenomena of Electronics”. In: Electronic Engineers’ Handbook (D.G. Finkand 0. Christiansen, eds) 3rd Edition, pp. 1-1 to 1-58. McGraw-Hill, New York, U.S.A.Gardner, W.H. (1986). Water content. In: Methods of soil analysis. Part I. American Society of8889Agronomy, pp 493 -544. Madison, WI.Gardner, W.R. (1987). Water Content: An Overview. In; Proceclings of International Conference onMeasurement of Soil and Plant Water Status, volume 1- soils, pp. 7-9. In commoration of theCentennial of Utah State University, July 6-10, 1987, Logan, Utah, U.S.A.Gore, Jr. T.S. (1989). “Capacitors”. In: Electronic Engineers’ Handbook (D.G. Fink and D. Christiansen,eds) 3rd Edition, pp. 7-8 to 7-13. McGraw-Hill, New York, U.S.A.Haise, H.R. and Hagan, R.M. (1967). Soil, plant and evaporative measurements as criteria forscheduling irrigation. Agronomy 11, pp. 557 - 604.Halbertsma, J., Przybyla, C., and A Jacobs (1987). Application and accuracy of a dielectric soil watercontent meter. In: Proceedings of International conference on measurement of soil and plant waterstatus. vol. 1. Utah State University, July 6 -10 1987.Heermann, D.F., Martin, D.L., Jackson, R.D., Stegman, E.C. (1990). Irrigation Scheduling Controls andTechniques. In: Irrigation of Agricultural Crops - Agronomy Monograph no. 30 (R.R. Bruce, E.T.Kanemasu, M.H. Niehaus, J.R. Gilley, eds). pp. 531 - 536.Hillel, 0. (1990). Role of Irrigation in Agricultural Systems. In: Irrigation of Agricultural Crops -Agronomy Monograph no. 30 (R.R. Bruce, E.T. Kanemasu, M.H. Niehaus, J.R. Gilley, eds). pp. 5 - 30.Hillel, 0. (1980). “Fundamentals of Soil Physics”. pp 126 - 134. Academic Press, New York.Holmes, J.W., Taylor, S.A., and S.J. Richards (1982). Measurement of Soil Water. In: “Irrigation ofAgricultural Lands (R.M. Hagan, H.R. Haise, T.W. Edminster (eds)). pp.275 -303. American Society ofAgronomy Publisher, Madison, U.S.A.Horton, R., P.J. Wierenga and Nielsen, D.R. (1982). A rapid technique tor obtaining uniform watercontent distributions in unsaturated soil columns. Soil Sd. 133(6):397-399.Joerg, E.A. (1988), AC circuit analysis, In: Handbook of Electronics Calculations for Engineers andTechnicians -2nd Edition, (M. Kaufman and A.H.Seidman, 1989, eds) pp 3.1 -3.26.Kaper, I. (1969). Mè?enI vlhkosti stavebnIch materiãlu (Techniques of measurement of the moistureof building materials.) Thesis, Fac, Civ. Eng., Technical University Prague.Kuráz, V., KutIlek, M., and Kaper, I. (1970) Resonance-capacitance soil moisture meter. Soil Sciencevol. 110, No.4, pp 278 -279.Layman, R.W. (1979). Use of a modified Wien bridge to measure soil moisture in sand, silts and90clays. Dartmouth College Plasma Laboratory, DCPPL-SM-4.Levine, M.E. (1988) Selecting R,L, and C components. In: Handbook of Electronics Calculations forEngineers and Technicians - 2nd Edition, (M. Kaufman and A.H.Seidman, 19, eds) pp 4.1-4.27Lewis, D.T. (1988) Bridge-Type Instruments. lntext Inc. Norwalk Connecticut, U.S.A.Nelson, S.C. (1973). Electrical properties of agricultural products (A critical Review). ASAE SpecialPublication No. SP-05-73. St. Joseph, Ml: ASAE.Odelevskii, V.1. 1951 Razchet obobshchennoi provodimosti geterogennykh sitem (Calculation of ageneralized permeability of heterogeneous systems.) Zhurn. Tekhn, Phys. 21:678 - 685.Schmugge, T.J., Jackson, T.J., and McKim, H.L. (1980). Survey of Methods for soil moisturedetermination. Water Res. 16(6): 961 - 979.Slatyer, R.O. (1954) Private communication (Land Research and Regional Survey Section, C.S.I.R.O,Canberra, A.C.T). 1954.Thompson, F.T. (1989). Measurement and Control Circuits. In: Electronic Engineers’ Handbook (D.G.Fink and D. Christiansen, eds) 3rd Edition, pp. 17-1 - 17-62. McGraw-Hill, New York, U.S.A.Topp, G.C., Davis, J.L., and Annan, A.P. (1980 c). Electromagnetic determination of soil water contentusing TDR: I. Applications to wetting fronts and steep gradients. Soil Sd. Soc. Am. J. 46, 672-678.Topp, G.C., Davis, J.L. and Annan, A.P. (1980 a). Electromagnetic determination of sol water content:Measurements in coaxial transmission lines. Water Res. Res. 16(3):574-582.Topp, G.C., and Davis, J.L. (1982). Measurement of soil water content using time-domainreflectometry. Can. Hydrol. Symp. Assoc. Comm. Hydrol., Nat!. Res. Council Can., Ottawa pp. 269-287.Topp, G.C., Davis, J.L., and Annan, A.P. (1980 b). Electromagnetic determination of soil water contentusing TDR: II. Evaluation of installation and configuration of parallel transmission lines. Soil Sc!. Soc.Am. J. 46, 678-684.Topp, G.C., and Davis, J.L. (1985). Time-Domain Reflectometry and its application to IrrigationScheduling. In: “Advances in Irrigation”. (D. Hillel,et, vol. 3 pp 107 -933.Turner, R.P. and Giblisco, S. (1991). “The Illustrated Dictionary of Electronics”. 5th Edition. TAB Books,Blue Ridge Summit, PA U.S.A.91Wilson, S., (1 988). Resonant Cicuits. Intext, Inc. Norwalk, Connecticut, U.S.A.Wilson, S. (1987). “Resistive, Capacitive and Inductive Components”. Intext Inc. Norwalk Connecticut,U.S.A.Wilson, S. (1988). “Inductance and Capacitance”. Intext Inc. Norwalk Connecticut, U.S.A.Wobschall, D. 1978. A frequency shift dielectric soil moisture sensor. IEEE Transactions on GeoscienceElectronics, GE-i 6(2):1 12-118.APPENDIXAppendix ATable Al. Calibration data for moisture sensors #1 and #2 placed within two different saturated sawdust samplesbeing air-dried in the laboratory - Test 1Sensor mass of mass of mois- mois- Current Capa- Mass of Oven- Weighttmoist sawdust ture ture flowing citance sensors dried ofsawdust + content content in R# of C# & weight drying+box + moisture (% dry box of cansensors (% by wt sample(g) (g) volume) basis) (g) (g) (g)80.3 69.9 6.1(pA) (nF)Cl and RI 3385 1822 492 2856 1000 169320.6 164.3 44.3 257.5 100.0 17.3304.8 148.5 40.1 232.8 100.0 17.1291.6 135.3 36.5 212.1 100.0 15.3286.2 129.9 35.0 203.6 100.0 14.4273.0 116.7 31.5 182.9 99.0 11.1266.7 110.4 29.8 173.0 98.0 9.1258.5 102.2 276 160.2 97.0 6.8248.7 92.4 24.9 144.8 96.0 5.5236.4 80.1 21.6 125.5 95.0 4.7229.1 72.8 19.6 114.1 95.0 3.8218.3 62.0 16.7 97.2 95.0 3.5211.3 55.0 14.8 86.2 95.0 3.1202.9 46.6 12.6 73.0 95.0 2.9193.3 37.0 10.0 58.0 95.0 2.5184.7 28.4 7.7 44.5 95.0 2.1175.6 19.3 5.2 30.3 94.0 0.8169.9 13.6 3.7 21.3 93.0 0.2C2andR2 3148 1630 504 2926 1000 122296.3144.5 44.7 259.4 100.0 12.2280.6 128.8 39.8 231.2 100.0 10.9267.5 115.7 35.8 207.7 100.0 8.1262.0 110.2 34.1 197.8 100.0 7249.5 97.7 30.2 175.4 99.0 4.7243.4 91.6 28.3 164.5 98.0 3.4236.1 84.3 26.0 151.3 98.0 3.1227.7 75.9 23.5 136.3 98.0 2.8217.2 65.4 20.2 117.4 98.0 2.5210.5 58.7 18.1 105.4 98.0 2.5201.4 49.6 15.3 89.0 98.0 2.1195.3 43.5 13.4 78.1 98.0 2188.2 36.4 11.2 65.4 98.0 1.8179.7 27.9 8.6 50.1 98.0 1.4172.3 20.5 6.3 36.8 98.0 0.9164.4 12.6 3.9 22.6 20.0 0.2159.6 7.8 2.4 14.0 96.0 0.1159.2 7.4 2.3 13.3 6.0 082.5 62.5 6.8Appendix A ILTable A2. Calibration data for moisture sensors #1 and #2 placed within re-saturated sawdust samples (from TableAl.) being air-dried in the laboratory - Test 2mass of mass of moisture moisture Current Capa- Mass of Oven-d Weight ofSensornumber moist sawdust content content flowing in citance sensors ed dryingsawdust + moisture R# of C# & weight of canbox + box samplesensor (% dry(% by(g) (g) volume) (g)6.1weightbasis) (VA) (nF) (g) (g)CI and RI 3473 1910 515 2994 ThO 93 803 9328.0 171.7 46.3 269.1 100 13.7318.0 161.7 43.6 253.4 99 12.9299.0 142.7 38.5 223.7 99 10.8289.6 133.3 36.0 208.9 99 10.0274.6 118.3 31.9 185.4 98 8.2269.3 113.0 30.5 177.1 98 7.3264.3 108.0 29.1 169.3 96 6.8258.5 102.2 27.6 160.2 96 6.1245.5 89.2 24.1 139.8 95 5.0238.5 82.2 22.2 128.8 94 4.7225.4 69.1 18.6 108.3 94 4.2220.1 63.8 17.2 100.0 94 3.9215.1 58.8 15.9 92.2 93 3.6204.6 48.3 13.0 75.7 93 3.2200.2 43.9 11.8 68.8 92 3.4196.5 40.2 10.8 63.0 92 3.0188.3 32.0 8.6 50.2 92 2.3178.5 22.2 6.0 34.8 90 1.0C2:d:R2: 323.0 171.2 52.9 307.4 100 12.4 82.5 62.5305.6 153.8 47.5 276.1 100 10.2296.3 144.5 44.7 259.4 99 9.2278.5 126.7 39.2 227.5 99 7.1270.3 118.5 36.6 212.7 99 6.3256.3 104.5 32.3 187.6 99 4.3251.4 99.6 30.8 178.8 97 3.7247.0 95.2 29.4 170.9 99 3.3241.7 89.9 27.8 161.4 99 3229.8 78.0 24.1 140.0 99 2.6223.5 71.7 22.2 128.7 99 2.5211.7 59.9 18.5 107.5 98 2.3206.8 55.0 17.0 98.7 98 2.1202.4 50.6 15.6 90.8 98 1.9192.8 41.0 12.7 73.6 98 1.6188.5 36.7 11.3 65.9 98 1.5185.2 33.4 10.3 60.0 98 1.3177.9 26.1 8.1 46.9 98 0.7169.0 17.2 5.3 30.9 96 0.26.8Appendix A(pA) (nF) (g) (g)100.0 25.8 80.6 62.1100.0 24.5100.0 20.6100.0 20.5100.0 17.198.0 14.898.0 12.998.0 12.497.0 11.997.0 10.796.0 9.690.0 9.116.0 7.40.0 5.2100.0 31.5 83.2 65.1100.0 28.2100.0 26.7100.0 23.8100.0 20.199.0 13.899.0 12.998.0 11.998.0 11.798.0 11.298.0 11.1Table A3. Calibration data for moisture sensors #1 and #2 placed within two new saturated sawdust samples beingair-dried in the laboratory - Test 3Sensor mass of mass of moisture moisture Current Capa- Mass Oven-dr Weightnumber moist sawdust content content flowing citance of ied ofsawdust moisture in R# of C# sensor weight drying÷box+ s& of cansensor box sample(%by (%drywt(g) (g) volume) (g)6.2:C1:ridR1:: 311.2287.0267.6257.5245.8222.6215.1207.2198.7195.1185.2179.0172.6168.3C2afldlR2 3236298.0277.4268.1255.7232.1224.0217.2208.9205.3196.1190.0183.5178.5162.3138.1118.7108.696.973.766.258.349.846.236.330.123.719.4168.4142.8122.2112.9100.576.968.862.053.750.140.934.828.323.350.042.536.533.429.822.720.418.015.314.211.29.37.36.049.842.236.133.429.722.720.318.315.914.812.110.38.46.9basis)290.3247.0212.3194.3173.3131.8118.4104.389.182.664.953.842.434.7289.3245.4210.0194.0172.7132.1118.2106.592.386.170.359.848.640.06.999.099.099.010.210.18.7Appendix ATable A4. Calibration data for moisture sensors #1 and #2 placed within the fresh sawdust samples (see Table A3.)being air-dried in the laboratory after re-saturation - Test 4.Sensor mass of mass of moisture moisture Current Capa- Mass of Oven- Weightnumber moist sawdust content content flowing in citance sensors dried ofR# of C# & weight of dryingbox sample cansawdust moisture+ box +sensor(%by (%drywt(g) (g) volume) basis) (g) (g)62.1 6.2(pA) (nF) (g)Cl and Ri 3326 1837 566 3286 100 418 806• 316.2 167.3 51.5 299.3 100 37.9299.6 150.7 46.4 269.6 100 35.5271.0 122.1 37.6 218.4 100 27.8262.0 113.1 34.8 202.3 100 25.9244.8 95.9 29.5 171.6 100 23.5237.3 88.4 27.2 158.1 100 22.1225.2 76.3 23.5 136.5 99 19.9220.0 71.1 21.9 127.2 99 19.1213.8 64.9 20.0 116.1 98 17.7209.8 60.9 18.8 108.9 98 17.0200.4 51.5 15.9 92.1 98 15.2195.6 46.7 14.4 83.5 98 14.4188.6 39.7 12.2 71.0 98 13.3173.2 24.3 7.5 43.5 98 6.2C2and R2 3290 1738 514 2986 100 510 832313.8 158.6 46.9 272.5 100 49.5296.3 141.1 41.7 242.4 100 45.1269.9 114.7 33.9 197.1 100 39.5261.6 106.4 31.5 182.8 100 36.6245.2 90.0 26.6 154.6 100 30.4238.3 83.1 24.6 142.8 100 27.3227.3 72.1 21.3 123.9 99 23.9222.4 67.2 19.2 115.5 99 22.6216.7 61.5 18.2 105.7 98 21.1213.1 57.9 17.1 99.5 98 19.4204.6 49.4 14.6 84.9 98 18.9200.2 45.0 13.3 77.3 98 18.8193.0 37.8 11.2 64.9 98 16.7178.2 23.0 6.8 39.5 98 13.165.1 6.9Appendix ATable A5. Calibration data for moisture sensors #3 and #4 placed within two different saturated organic soil samplesbeing air-dried in the laboratory - Test 1.Sensor mass of mass of moisture moisture Current Capa- Mass of Oven- Weight ofnumber moist soil content content flowing in citance sensors dried dryingsoil + moisture R# of C# & weight of canbox + (% dry box samplesensor (% by weight(g) (g) volume) basis) (pA) (nF) (g) (g) (g)458.2 203.5 76.2 125.7 100.0 194.0 78.6 169.0 7.1439.6 184.9 69.2 114.2 100.0 195.0424.8 170.1 63.7 105.1 100.0 199.0412.8 158.1 59.2 97.7 100.0 190.0407.8 153.1 57.3 94.6 100.0 189.0396.1 141.4 52.9 87.3 99.0 186.0390.3 135.6 50.8 83.8 99.0 186.0382.8 128.1 48.0 79.1 98.0 176.0374.0 119.3 44.7 73.7 96.0 166.0365.4 110.7 41.4 68.4 91.0 147.0352.6 97.9 36.6 60.5 76.0 122.0345.0 90.3 33.8 55.8 71.0 109.0333.0 78.3 29.3 48.4 63.0 90.0324.7 70.0 26.2 43.2 57.0 77.0313.7 59.0 22.1 36.4 46.0 64.0301.5 46.8 17.5 28.9 36.0 49.0292.0 37.3 14.0 23.0 28.0 36.0282.2 27.5 10.3 17.0 12.0 20.0276.7 22.0 8.2 13.6 1.0 9.0C4and R4 4572 2064 789 1301 1000 2070 800 1647 61439.2 188.4 72.0 118.8 98.0 206.0424.2 173.4 66.3 109.3 96.0 192.0412.4 161.6 61.8 101.9 95.0 180.0407.7 156.9 60.0 98.9 93.0 177.0396.3 145.5 55.6 91.7 91.0 172.0390.9 140.1 53.5 88.3 93.0 169.0384.1 133.3 50.9 84.0 93.0 164.0375.8 125.0 47.8 78.8 88.0 154.0367.7 116.9 44.7 73.7 86.0 141.0355.3 104.5 39.9 65.9 82.0 118.0348.2 97.4 37.2 61.4 76.0 106.0337.4 86.6 33.1 54.6 72.0 85.0329.2 78.4 30.0 49.4 64.0 72.0318.9 68.1 26.0 42.9 61.0 59.0306.7 55.9 21.4 35.2 57.0 44.0296.6 45.8 17.5 28.9 51.0 31.0285.9 35.1 13.4 22.1 47.0 17.0279.3 28.5 10.9 18.0 42.0 7.0Appendix AnumberTable A6. Calibration data for moisture sensors #3 and #4 placed within re-saturated organic soil samples (fromTable A5.) being air-dried in the laboratory - Test 2.Sensor mass of mass of moisture moisture Current Capa- Mass of Oven-d Masscontent content flowing citance sensors ned ofin R# of C# & box weight dryingsensor of cansample(g) (g) (g)78.6 169.0 7.1moist soilsoil + moist-box + uresensor________________(g)______(g):::Cafld:R3::: 415.9 161.2413.8 159.1397.5 142.8388.5 133.8370.9 116.2362.7 108.0348.9 94.2344.1 89.4339.3 84.6334.9 80.2322.0 67.3315.4 60.7303.9 49.2299.3 44.6295.1 40.4286.8 32.1285.0 30.3281.6 26.9275.5 20.8269.7 15.0:4ø4 417.3 166.5416.5 165.7415.0 164.2400.2 149.4391.6 140.8376.0 125.2368.7 117.9356.2 105.4351.8 101.0346.8 96.0343.1 92.3331.1 80.3324.7 73.9312.5 61,7307.5 56.7302.2 51.4293.0 42.2289.9 39.1286.5 35.7279.4 28.6272.0 21.2(% byvolume)60.359.653.550.143.540.435.333.531.730.025.222.718.416.715.112.011.310.17.85.663.663.362.757.153.847.845.140.338.636.735.330.728.223.621.719.616.114.913.610.98.1(% dryweightbasis)99.698.388.282.671.866.758.255.252.349.541.637.530.427.525.019.818.716.612.89.3105.0104.5103.594.288.878.974.366.563.760.558.250.646.638.935.832.426.624.722.518.013.4(pA)1001001009998968681737264584843382710600100100100100999898959389856369595449402925110(nF)2202401951781391229688837760534336312010620149169181143133107978176736753483832281813115080.0 164.7 6.1Appendix ATable A7. Calibration data for moisture sensors #3 and #4 placed within two new saturated organic soil samplesbeing air-dried in the laboratory - Test 3.Sensor mass of mass of moisture moist- Current Capa- Mass of Oven-dri Weightnumber moist soil content ure flowing citance sensors ed ofsoil + moist- content in R# of C# & weight dryingbox + ure box of cansensor sample(% dry(% by weight(g) (g) volume) basis) (pA) (nF) (g) (g) (g)C3andR3 3915 1618 715 1180 1000 2470 782 1443 72364.8135.1 59.7 98.5 100.0 208.0345.2 115.5 51.1 84.2 100.0 189.0336.4 106.7 47.2 77.8 98.0 178.0325.6 95.9 42.4 69.9 97.0 161.0299.8 70.1 31.0 51.1 68.0 96.0290.7 61.0 27.0 44.5 64.0 80.0282.3 52.6 23.3 38.4 58.0 69.0272.4 42.7 18.9 31.1 49.0 57.0268.2 38.5 17.0 28.1 45.0 50.0258.5 28.8 12.7 21.0 24.0 32.0253.2 23.5 10.4 17.1 6.0 24.0248.2 18.5 8.2 13.5 0.0 13.0244.6 14.9 6.6 10.9 0.0 4.0C4andi4 4005 1705 751 1239 1000 2450 802 1437 61376.7 146.7 64.6 106.6 100.0 213.0360.2 130.2 57.3 94.6 100.0 198.0351.8 121.8 53.6 88.5 99.0 190.0341.1 111.1 48.9 80.7 99.0 178.0316.0 86.0 37.9 62.5 88.0 114.0307.6 77.6 34.2 56.4 81.0 91.0298.6 68.6 30.2 49.9 77.0 76.0288.2 58.2 25.6 42.3 71.0 60.0283.6 53.6 23.6 39.0 67.0 53.0272.3 42.3 18.6 30.7 56.0 36.0266.3 36.3 16.0 26.4 52.0 30.0260.0 30.0 13.2 21.8 42.0 21.0255.4 25.4 11.2 18.5 24.0 15.0Appendix AlaDSensornumberTable A8. Calibration data for moisture sensors #3 and #4 placed within the fresh orgnaic soil samples (see TableA87.) being air-dried in the laboratory after re-saturation- Test 4.mass of mass of moist- moisture Current Capa- Mass of Oven- Weightmoist soil ure conteflt flowing citance sensors dried ofsoil + moist- content in R# of C# & weight dryingbox + ure box of cansensor (% dry (% dry samplevolume weight(g) (g) (g)78.2 144.3 7.2(g) (g) basis) basis) (pA) (nF)03 and R3 3713 1416 626 1033 100 3980355.1 125.4 55.4 91.5 100 339.0327.6 97.9 43.3 71.4 99 200.0318.3 88.6 39.2 64.6 98 156.0298.9 69.2 30.6 50.5 96 92.0290.8 61.1 27.0 44.6 96 71.0277.4 47.7 21.1 34.8 95 48.0272.0 42.3 18.7 30.9 95 40.0266.3 36.6 16.2 26.7 94 31.0262.9 33.2 14.7 24.2 94 25.0255.4 25.7 11.4 18.7 93 17.0252.2 22.5 9.9 16.4 86 14.0247.4 17.7 7.8 12.9 70 5.0239.9 10.2 4.5 7.4 55 1.0C4and.R4 3810 1510 665 1097 100 3160•363.9 133.9 59.0 97.3 100 251.0337.6 107.6 47.4 78.2 99 172.0329.5 99.5 43.8 72.3 99 145.0313.4 83.4 36.7 60.6 96 94.0305.8 75.8 33.4 55.1 94 75.0292.5 62.5 27.5 45.4 88 50.0286.7 56.7 25.0 41.2 84 43.0280.1 50.1 22.1 36.4 79 34.0276.3 46.3 20.4 33.6 75 28.0267.9 37.9 16.7 27.5 65 20.0263.7 33.7 14.8 24.5 50 16.0258.2 28.2 12.4 20.5 43 10.0247.7 17.7 7.8 12.9 31 2.080.2 143.7 6.1Appendix A/01A.1 Calculation of SensitivitiesThe sensitivities of the various sensors were calculated using the capacitance and current valuescorressponding to approximately 10 % and 40 % soil or sawdust moisture content levels. For examplefor Cl and Ri (please see Table Al), the moisture content levels used were 40.1 and 10.0. Thecorresponding values of capacitance are 17.3 nF and 2.5 nF respectively. The change in capacitance,AC is, therefore 14.8 nF. The total change in moisture content, 9,,, corresponding to this range is 30.1%. The resolution of the capacitance meter used for this measurement is 0.1 nF. Therefore thesensitivity index, s, for the capacitance sensor, according to equation (4.1), is calculated as=______= (40.1 -10.0)= 0.2(AC /r) (17.1 -2.5)0.1The s value for the fiberglass resistance sensor Ri, and all the other sensors were calculted in a similarway.A.2. Spreadsheet CalculationsAll calculations were performed with the QUATTRO PRO computer spreadsheet programme. Thespreadsheet name for each of column headings for Table Al to A8 are shown below: A copy of thecomputer program is kept in the Bio-Resource Engineering Department.A Sensor number8 Mass of moist sawdust or soil + box + sensorsC Mass of sawdust or soil + moistureD Moisture Content (% by volume)Appendix A/o.E Moisture Content (% by weight)F Current flowing through R#G Capacitance of C#H Mass of sensors and boxOven-dried weight of sampleJ Mass of drying can.The above letter designations are used in the explanations of the spreadheet calcultions which follow.The explanations are based on Table 1, in which the other ordinate of the first line of figures is 13 i.e.the cell address of the weight of drying can 6.1 g is J13.A.3 Calculations of Mass of Soil MoistureE = B13-H13-l13-J13AA Calculations of Moisture Contents (% dry weight basis)E = C13*1OO/(l13J13).A.5 Calculations of Moisture Contents (% dry volume basis)Both the soil and sawdust expand on absorbing water. The volumes of the samples weretherefore not constant throughout the experiment. It was therefore decided to base the volumetricmoisture content calculations on the average volume of the oven dried sample. The average specificvolume of the oven-dried sawdust sample was 5.81 g/cc, and that of the organic soil was 1 .65 g/cc.Therefore D is E13 / 5.81 for sawdust and E13 / 1.65 for organic soil.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086704/manifest

Comment

Related Items