Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Measurement of copolar attenuation through the bright band at 4 & 7 GHz Van der Star, Jack A. 1982-04-13

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


831-UBC_1982_A7 V36.pdf [ 15.12MB ]
JSON: 831-1.0095325.json
JSON-LD: 831-1.0095325-ld.json
RDF/XML (Pretty): 831-1.0095325-rdf.xml
RDF/JSON: 831-1.0095325-rdf.json
Turtle: 831-1.0095325-turtle.txt
N-Triples: 831-1.0095325-rdf-ntriples.txt
Original Record: 831-1.0095325-source.json
Full Text

Full Text

MEASUREMENT OF COPOLAR ATTENUATION THROUGH THE BRIGHT BAND AT 4 & 7 GHz by JACK A. VAN DER STAR BASc, The University of British Columbia, 1977 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE DEPARTMENT OF ELECTRICAL ENGINEERING We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1982 • Jack A. Van der Star, 1982 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date DE-6 (.3/81) ABSTRACT This thesis describes an experiment designed to measure microwave propa gation through the bright band at 4 and 7 GHz. The path is located approxi mately 100 kilometers east of Vancouver, British Columbia, Canada, forming part of the Trans Canada Telephone System microwave network. The path is coastal and mountainous in nature, 41.3 kilometers in length and experiences an average annual rainfall of 1600 mm/year. Due to these factors and an ele vation differential of 1227 m between transmitting and receiving sites, the 0°C isotherm and hence bright band effect normally exist along the path from November to April. A measurement system based on remote telemetry is used to obtain high resolution and accurately time-correlated data. Received signal levels are taken from five selected 4 and 7 GHz microwave channels which are sampled at a rate of 10 Hz. Meteorological information is obtained from five locations along the path and sampled at a rate of 1 Hz. The data thus collected are then time-correlated as it arrives at the University of British Columbia (Vancouver) recording site where it is analyzed using high-level language routines developed as part of a propagation data base management system. A detailed description of both the measurement system and the data management system are provided in the thesis. Results from several precipitation systems indicate that bright band attenuation can be many times (in dB per kilometer) greater than attenuation due to equivalent amounts of rain. This is described by an Excess Attenuation (ii) Ratio (EAR) defined as the ratio of the excess attenuation in dB/km calculated using the Laws and Parson distribution at 0°C. The experimental results com pare favourably with those predicted by the theoretical model of Matsumoto and Nishitsuji. A scintillation type fading phenomenon superimposed on the broad-band fade has also been observed during bright band propagation conditions. From the preliminary results this phenomenon appears to be correlated with sudden changes in differential temperature between transmitter and receiver sites and thus a corresponding change in the thickness of the bright band. (ii±) TABLE OF CONTENTS Page ABSTRACT ii TABLE OF CONTENTS v LIST OF ILLUSTRATIONS viLIST OF TABLES xiACKNOWLEDGEMENTS xv I. INTRODUCTION 1 1.1 The Importance of Microwave Propagation in the Design of Microwave Systems1.2 Factors Affecting Microwave Propagation 2 1.2.1 Path Factors 3 1.2.2 Rain Attenuation 5 1.2.3 Multipath Fading 10 1.2.4 Other Propagation Factors 12 1.3 Improving Reliability in Path Design 3 1.4 Bright Band Effects 15 1.5 Scope of Thesis 21 1.5.1 The Research Program 21.5.2 Thesis Objectives 3 1.5.3 Thesis Outline 4 II. THE EXPERIMENT 26 2.1 Introduction2.2 The Path2.3 Received Signal Monitoring 26 III. METEOROLOGICAL INSTRUMENTATION 35 3.1 System Design 33.1.1 Measurement Criteria3.1.2 Site Selection 35 3.2 Meteorological Measurements 9 3.2.1 Rain3.2.2 Temperature Transducer 41 3.2.3 Wind Velocity and Wind Direction Transducer .... 42 3.3. Meteorological-Data Sampling(iv) Page IV. DATA ACQUISITION SYSTEM 45 4.1 Design Criteria for a Real Time Data Acquisition System . . 45 4.2 Site Selection 7 4.2.1 Received Signal Site 44.2.2 Meteorological Sites 8 4.3 Data Collection Network Design 9 4.3.1 Data Statistics 44.3.2 Link Capacities4.3.3 Node Considerations 50 4.3.4 Implementation of the Network Topology 51 4.4 Real Time Data Storage 2 4.4.1 Microprocessor Considerations . . 52 4.4.2 Data Storage Formats ,_4.5 Allowance for Future Data Requirements 4.6 An Alternate Data Acquisition System Using Chart Recorders. V. DATA BASE MANAGEMENT SYSTEM 55 5.1 Specifications 5.2 Design Considerations for DBMS Relative to Existing and Future Systems 5 7 5.3 A System Description of DBMS 59 5.3.1 Data Transfer and Handling 59^ 5.3.2 Estimate of DBMS Data Volumes 61 VI. RESULTS 62 6.1 Some Initial Results Obtained Using Chart Recording .... 62 6.2 Events Measured Using Remote Telemetry Showing Bright Band Propagation 6 6.2.1 January 23, 1982, 7:30-11:30 p.m 66.2.2 January 23, 1982, 2:00-4:30 p.m 70 6.2.3 February 19, 1982, 7:30-9:00 a.m 9 VII. CONCLUSIONS AND DIRECTIONS FOR FUTURE RESEARCH 84 7.1 Conclusions . . . 84 7.2 Directions for Future Research 85 (v) Page 89 APPENDIX A Automatic Gain Control (AGC) Calibrations APPENDIX B Details of the Data Acquisition System Layout 95 APPENDIX C Equipment and Site Layouts 106 APPENDIX D Meteorological Transducers 122 APPENDIX E Signal Conditoning Units 7 APPENDIX F Analog to Digital Convertor 136 APPENDIX G Digital to Analog Convertor 142 APPENDIX H Modem Units 145 APPENDIX I Microprocessor Units 150 APPENDIX J Microprocessor Software 169 APPENDIX K The Bright Band Propagation Experiment's Data Base Management System . 190 REFERENCES 20(vi) LIST OF ILLUSTRATIONS Figure Page 1.0 Path Factors 3 1.1 Specific attenuation as a function of frequency for coherent wave propagation through uniform rain. The curves are based on the Laws and Parsons dropsize distribution and the ter minal velocities of Gunn and Kinzer. Rain tempera ture of 20°C. Rain temperature of 0°C 3 1.2 A microwave system diagram illustrating space diversity, equipment diversity and frequency diversity 14 1.3 The character and classification of snow as it passes through the bright band as seen on water-blue paper 18 1.4 Falling velocity vs. radii of raindrops and snowflakes ... 19 1.5 Relative bright band geometries between slant path terrestial and earth space links 22 2.0 Geographical layout of the bright band propagation experiment 27 2.1 Path profile: Ryder Lake to Dog Mountain 28 2.2 Path photographs 9 2.3 Frequency selection plan and receiver equipment used at receiver site 31 2.4 4 GHz receiver microwave transmission block diagram 32 2.5 7 GHz microwave transmission system block diagram 33 3.0 Measurement system layout 36 3.1 Path cross-section showing relative locations of the weather station sites 38 3.2 Weather station inter-site distances 40 3.3 Photograph of the U.B.C. Weatherlog microprocessor and meteorological signal conditioning unit 43 (vii) Figure Page 4.0 The data acquisition system block diagram with component areas identified 46 4.1 Data collection network topology 51 4.2 Photograph of the Video Terminal Displaying Incoming Data . . 53 5.0 The bright band experiment data system flow chart 56 5.1 DBMS in relation to other propagation data management systems 58 5.2 User flowchart to process time series data on the DBMS software system 60 6.0 Recordings of received signal at 7 GHz during bright band propagation: a) Event "A" (January 11-12, 1980) b) Event "B" (February 2, 1980) 63 6.1 The Agassiz temperature and 7.142 GHz signal level versus time 67 6.2 Agassiz rain rate and the 7.496 GHz signal level versus time 8 6.3 The 3.550 and 7.496 GHz signal levels versus time 71 6.4 The Dog Mountain transmitter site temperature and the 7.496 GHz reveiver levels versus time ....... 72 6.5 The Ryder Lake receiver site temperature and the 7.496 GHz receiver level versus time 73 6.6 An expanded view at approximately 80 minutes into the event showing the Ryder Lake temperature and 7.496 GHz receiver level versus time 74 6.7 Ryder Lake differential temperature and 7.496 GHz received signal level versus time showing the fade discontinuity 80 minutes into the event . . . . 75 6.8 Ryder Lake differential temperature and 7.496 GHz received signal level versus time showing discontinuities 30 minutes into the event 76 (viii) Figure Page 6.9 Ryder Lake windspeed, temperature and the 3.550 GHz received signal level versus time 77 6.10 An expanded view of the Ryder windspeed and 3.550 GHz received signal level versus time approximately 90 minutes into the event 78 6.11 Ryder Lake and Dog Mountain Temperature, Ryder Lake Windspeed and the 7.496 GHz receive signal level versus time 80 6.12 Agassiz rain rate and the 7.496 and 4.010 GHz receive signals versus time 81 6.13 Ryder Lake differential temperature and the 7.496 GH receive signal level versus time 83 7.0 Proposed system configuration to incorporate the digital radio monitoring system 88 A-1 3550 MHz receiver AGC calibration 90 A-2 3790 MHz " " " 1 A-3 4010 MHz " " " 92 A-4 7142 MHz " " " 3 A-5 7496.5 MHz receiver AGC calibration . 94 B-l Path system layout for U.B.C. microwave propagation experiment . 97 B-2 Path profile Ruby Creek to Ryder Lake 98 B-3 Path profile: Bear Mountain to Ryder Lake 9B-4 Circuit layout from Dog Mountain to Ryder Lake 99 B-5 Circuit layout from the Agassiz Experimental Farm to Ryder 99 Lake B-6 Schematic for the Ryder Lake to U.B.C. data circuit 10° B-7 The RS232 interface for the Ryder Lake to U.B.C. data circuit 101 (ix) Figure Page C-l Ryder Lake site photograph 102 C-2 Ryder Lake site equipment configuration 103 C-3 Ryder Lake site layout 104 C-4a Dog Mountain site photograph 105 C-4b A photograph showing damage to the anemometer caused by severe icing conditions at Dog Mountain 106 C-5 Dog Mountain site equipment configuration 107 C-6 Dog Mountain site layout 108 C-7 Agassiz Experimental Farm site photograph 109 C-8 Agassiz Experimental Farm site equipment configuration . . . 110 C-9 Agassiz Experimental Farm site layout Ill C-10 Ruby Creek site photograph 112 C-ll Ruby Creek site equipment configuration 113 C-12 Ruby Creek site layout 114 C-13 Bear Mountain site photograph 115 C-14 Bear Mountain site equipment configuration 117 C-15 Bear Mountain site layout 118 C-16 The University of British Columbia site photograph 119 C-17 The University of British Columbia equipment configuration . 120 C-18 The University of British Columbia site layout 121 D-l A photograph of the anemometer 122 D-2 Anemometer circuit and wiring diagram 123 D-3 A typical rain bucket and tipping assembly 124 D-4 Photograph of the temperature probe 125 (x) Figure Page E-1 Top view Photograph of the meteorological signal conditioning unit 127 E-2 Circuit schematic for the meteorological signal conditioning unit 128 E-3 Front and rear views of the meteorological signal 129 E-4 Circuit schematic for the Bear Mountain signal conditioning card 132 E-5 Circuit schematic for the receiver signal conditioning card . 134 F-l Photographs showing the A/D convertor separately and installed 137 F-2 Circuit schematic and physical layout for the Weatherlog analog to digital convertor 139 F-3 Control signal timing diagram for the A/D convertor 140 F-4 Sample oscilloscope traces of the A/D output during calibration 14F-5 A/D calibration program 141 G-l Circuit schematic for the D/A convertor 143 G-2 Listing of the program to provide chart recordings from the receiver data using the D/A converter 144 H-l Photograph of an installed modem transmit unit 145 H-2 Interface schematic for a Weatherlog modem transmit unit . . 147 H-3 Physical drawing of the Ryder Lake receiver unit showing top, front and rear views 148 H-4 Interface schematic for one of the modem receiver units . . . 149 1-1 Physical drawing of the UBC data formatting ( P//1) micro processor showing front, top and rear views 151 1-2 Physical drawing of the UBC data processing ( P#2) micro processor showing front, top and rear views 152 (xi) Figure Page 1-3 Photograph showing the interior layout of the data formatting (uP#l) and data processing (uP#2) microprocessors 153 1-4 .. Circuit schematic and physical layout of the asynchronous interface (ACIA) card 156 1-5 Physical drawing of the Ryder Lake microprocessor unit showing front, top and rear views 160 1-6 Physical drawing of the Weatherlog microprocessor showing front, top and rear views 162 1-7 Photograph showing the interior layout of a typical UBC Weatherlog microprocessor unit 163 1-8 Physical drawing of the low power microprocessor unit used at Bear Mountain 165 1-9 Photograph showing the interior layout of the Bear Mountain microprocessor unit 166 1-10 Equipment configuration to transfer data from the cassette tape drive to the NOVA 840 magnetic tape drive 167 1-11 Circuit diagram of the RS232C to current loop interface . . . 168 J-la Program flow chart for uP#l; the UBC data formatter .... 171 J-lb Data formatter buffer memory organization 172 J-2 Program flow charts for uP//2; the UBC data processor micro processor unit 173 J-3 Diagram showing the structure of the time series queue (TSQ) 176 J-4 Data acquisition flow chart for the Ryder Lake datalog 6800 microprocessor 183 J-5 Program flow chart for the UBC Weatherlog 8085 data acquisi tion microprocessors 185 J-6 Data acquisition and control program flow chart for the remote Bear Mountain 1802 microprocessor unit 187 J-7 Program flow chart to transfer data from cassette tapes to magnetic tape using the NOVA 840 189 (xii) gure Page K-1 DBMS functional flow chart showing completion status .... 192 K-2 DBMS main program flow chart 193 K-3 Sample Plotting Run 199 (xiii) LIST OF TABLES Table Page 1.0 Attenuation Multipliers Due to Watery Snow 20 2.0 Microwave Transmission Calculations . 34 3.0 Geographical and Functional Site Details 7 3.1 Inter-Site Distances as a Function of Path Length 40 4.0 Data Acquisition System Link Capacities 54 5.0 DBMS Data Volume Estimates 61 6.0 Preliminary Results6.1 Bright Band Excess Attenuation Ratio (EAR) Results 64 6.2 January 23, 1982 Results (7 GHz) 69 6.3 Bright-Band Excess Attenuation (EAR) Results 66.4 February 19, 1982 Results 82 6.5 February 19, 1982 EAR Results 8A-1 Receiver Frequencies Polarizations and Associated AGC Curves. 89 B-l VHF Radio Path Transmission Calculations 96 U E-1 The Resistor Values Used in the Differential Gain Block for Optimum Gain 133 F-l A/D Convertor Channel Assignment Table 138 H-l Modem Center Frequency Assignments 146 1-1 I/O Port Address assignments for the Data Formatter Unit uP#l 154 1-2 I/O Port Address assignments for the Data Processor Unit uP#2 155 1-3 6800 Interrupt Vectors 157 1-4 Analog to Digital (A/D) Convertor Channel Assignments .... -j.59 1-5 1/0 Port Address Assignments for the Ryder Lake Unit .... 161 J-l RAM Time Assignments (uP#l) 178 J-2 Time Series Block Format 9 J-3 Data Format for the Distribution Buffer I80 K-1 Data Directory for the Time Series Format 195 (xiv) ACKNOWLEDGEMENTS I would like to acknowledge my appreciation to Dr. M.M.Z. Kharadly who has provided me with much needed support, supervision and suggestions throughout the course of this research. A grateful acknowledgement is also extended to Mr. Neville Owen of the British Columbia Telephone Company for his much appreciated assistance, his field coordination and his advice throughout my thesis work. I would also like to thank the following people at the British Columbia Telephone Company for their invaluable contributions made during this research: Mr. Bill Squans Mr. Stan Dahl Mr. Red Matthews Mr. George Gatt Mr. Dwight Chan Mr. Peter Claydon In the same way I would like to thank the following staff at the Communications Research Centre in Ottawa for their numerous suggestions, for their continued support and for providing a stimulating working environment during my stay: Dr. Stewart McCormick Dr. Rod Olsen Dr. Dick Butler Dr. Ben Segal Dr. John Strickland Mr. Neville Reed Mr. Hassen Kheirallah Mr. Joe Schlesak (xv) My gratitude is also extended to the following people at the Agassiz Experimental Farm for generously providing valuable back up meteorological data from their weather station, for maintaining the UBC chart recordings and looking after the UBC Weatherlog computer: Mr. Frank DeZwaan Ms. Moira Jewell I would like to give many thanks to Dr. Terry Enegren and Mr. David Michelson, for implementing the integration phase of the remote telemetry and DBMS systems , making the results possible. I would also like to express my gratitude to the following staff at the Department of Electrical Engineering for their contributions: Mr. James Johnston Mr. Eric Minch Mr. Edwin Lee Mr. Jun Lee Thanks are due to Ms. Sherry Lashmar and Mrs. Kathy Brindamour for typing the manuscript, Mr. Ben Van der Star for his help in assembling the elec tronics and to Mr. Len Smart for the development of the entry procedure and plotting routine packages. I would also like to thank all my friends and colleagues, particularly Dr. Basil Peters, Mr. Peter van der Gracht, Mr. Konrad Mauch and Mr. Frank Peabody for creating an enjoyable and stimulating working environ ment. I would like to greatfully acknowledge, as well, the Agassiz Experi mental Farm and the Canadian Broadcasting Corporation for their cooperation in allowing the use of their facilities at intermediate valley sites along the path. (xvi) This work was supported by the British Columbia Telephone Company through contract number 0007Q8 (NO/KO), the Communications Research Centre in Ottawa through contract numbers OSU 79-00061, OSU 80-00146 and 81-00112 and The National Science and Engineering Research Council of Canada through grant number A-3344. Finally, I would like to thank my wife Kathy for her continued patience, encouragement and support during my master's program. (xvii) 1 CHAPTER 1 INTRODUCTION 1.1 The Importance of Microwave Propagation In the  Design of Microwave Systems Microwave propagation parameters affect the reliability (availability) of terrestrial and satellite communication systems and thus have a major in fluence on their economics [1]. Therefore, techniques to confidently estimate this reliability are essential before system planners and designers can imple ment cost effective and economically viable microwave transmission systems [2]. Microwave transmission systems below 10 GHz include both satellite and terrestrial networks with common carriers and CATV operators being the main users. The least demanding from an availability viewpoint is the CATV opera tor. His performance requirements are deemed satisfactory if the system is available for more than 99.8% of the time, which represents an outage of approximately 1000 minutes per year [3]. On the other hand, the most demand ing user of communication systems are the common carriers who presently operate at 4, 6 and 7 GHz and propose to start operations in the 8 GHz band [4]. An example of a typical common carrier system specification is the one proposed for the Vancouver to Halifax, 8 GHz digital circuit. The system is specified to be available for more than 99.98% of the time [5], where the unavailability of less than 0.02% has been allocated with an allowance of 1/2 (.01%) for equipment failures, 1/4 (.005%) for rain outages and 1/4 (.005%) for multi-path fading outage. On an average per-hop basis this represents an 2 unavailability requirement of less than .0002% (63.2 seconds) of the time, where the combined propagation outage factors of rain and fading may not account for more than .0001% (31.6 seconds). Therefore, accurate estimation of the propagation reliability is essential before installation since propagation factors not accounted for in the original design could prevent meeting the total system availability" objective. In this light, the work described by this thesis was started to study factors associated with bright-band propagation at 4 and 7 GHz, for which no account has yet been made. 1.2 Factors Affecting Microwave Propagation Microwave propagation for frequencies below 10 GHz are largely affected by a path's geometry and associated weather characteristics. Microwave propa gation is affected by the path's geometry in the way of free space attenua tion, curvature changes due to variations in the refractivity profile and obstruction losses. Propagation factors related to meteorological conditions along a microwave path result in rain attenuation, multlpath fading, and reductions in received signal levels due to the presence of the bright-band. Other factors such as gaseous absorption by water vapour oxygen and fog also exist to a minor extent below 10.GHz. A discussion of these factors and their determination is the subject of this next section. 3 1.2.1 Path Factors Figure 1.0 shows an example of a line-of-sight microwave path which would form a segment in a common-carrier's back-bone microwave system. It would typically have a hop length of 40 to 60 km, a fade margin of 40 dB and an operating frequency between 4 and 10 GHz [5]. Figure 1.0 Path Factors 4 The Effect Due to Changes In Refractivity Although a microwave path is often termed to be "line-of-sight", the beam does not travel in a straight line through the atmosphere from the transmitter to the receiver but rather bends as a result of a slight decrease in the atmospheric refractive index with height. This bending is described by a "K" or equivalent earth radius factor which, if multiplied by the actual earth radius rQ, gives the ficticious earth curve, r, travelled by the propagating microwave beam. Therefore, variations in atmospheric refractivity cause cor responding changes in the K factor and are described as a function of atmos pheric temperature and water vapour content, as follows [6]: N = (n-1) 106 = 77.6 | + 3,73 105 6 (1-1) T2 dry term wet term where N: refractivity (N units) n: refractive index P: atmospheric pressure (m bar) T: absolute temperature (°K) e: water vapour pressure (m bar) If a point is taken at a fixed elevation the radius of curvature of the beam, r, relative to the earth's radius rQ can be expressed as a function of the vertical index of refraction gradient dn/dh to give the following expres sion for the K factor [7]: K - (1 + r |M 1 = (1 / 157) 1 r v o dh' v dh 1 o , (1-2) r = 6370 km o Published maps on world atmospheric radio refractivity and refractivity 5 gradients are available [8,9,10]. Calculation of the Free Space Attenuation The attenuation to a microwave signal emitted from an isotropic radiator can be determined as a function of frequency and distance by the following expression, [2] : AdB = 92,4 + 20 1Og10 f + 20 lQgio D (1_3) where, A : free space attenuation (dB) OLD f: frequency (GHz) D: distance (km) Obstruction Losses Obstruction losses vary according to a path's characteristic profile and operating frequency which allows their estimation, eg. Bullington [11]. 1.2.2 Rain Attenuation Early theoretical estimates of rain attenuation on microwave propagation were made by Ryde and Ryde [12,13] during World War H and were based on a first-order forward-scattering model that used a uniform distribution of equi-diameter spheres. The resulting expression for attenuation for a plane wave is given in equation (1-4): a = 0.4343 xNxTrxD2xf (D/A, m) dB/km (1- 4) Si where, a: attenuation (dB/Km) N: density (m-3) D: drop diameter (mm) 6 m: . n-jnx, the complex refractive index of water ffl: ratio of energy absorbed and scattered to that incident upon the projected area of the drop. Medhurst, in his review [14], corrected the numerical calculations derived by Ryde using this model but found a large variation, greater than expected, in experimental measurements [15,16,17,18]. Medhurst suggested that some of this variation could be attributed to multiple scattering processes but further theoretical analysis by Crane [19] and Rogers & Olsen [20] deter mined that this effect would be insignificant and the "single scattering" model for rain attenuation was, in fact, valid for frequencies lower than 20 GHz. The discrepancy between theory and experiment which Medhurst observed has since been attributed to inaccurate rain rate measurements. Most researchers at present acknowledge that good agreement between theoretical estimates and carefully obtained experimental observations are possible [21]. In order to allow the calculation of the attenuation coefficient using path-average rain rate data, a relation between the rain rate, the density and the drop diameter is required. This is given in terms of the terminal velocity of a falling drop in (1-5): R = 1.885 x v x N x D3 (mm/hr) (1-5) where, R: rain rate (mm/hr) v: terminal velocity (m/sec) N: density (m-3) D: drop diameter (mm) The terminal velocity is related to the drop diameter, as measured by Gunn and Kinzer [14] , and therefore the calculation of the attenuation coefficient is 7 possible for uniform rain for a single drop size using (1-4). Physically, rain is composed of drop-sizes exhibiting a continuous range of diameters from 0.5 mm to 7 mm. Integrating the specific attenuation relation for a single drop size (1-4) over the whole drop size distribution for various rain rates gives the attenuation by actual rain vs. rain rate. Figure 1.1, taken from [22], provides a graph of the specific attenuation versus frequency at various rain rates using this approach. The applicability of the results derived using the Ryde model is dependent on the diameter of the spheres used and, therefore, the proper selection of the drop size distribution is important. For most temperate-continental rainfall types the Laws and Parson distribution provides a good correlation between the Ryde model and experiment [23]. Other drop size dis tributions are available for differing applications which include the Marshall and Palmer distribution [24] and the Joss et al. distributions [25] for drizzle, wide spread rain and thunderstorms. A Simplified Empirical Model for Rain Attenuation  as a Function of Rain Rate Further improvements to the calculation of specific attenuation as a function of rain rate have been made by Olsen, Rodgers and Hodge [22], who have developed the simplifying empirical formula in (1-6), known as the A - R relation: A = (1-6) where, A: attenuation (dB/Km) 8 FREQUENCY (GHz) Figure 1.1 Specific attenuation as a function of frequency for coherent wave propagation through uniform rain. The curves are based on Laws and Parsons dropsize distribution and the terminal velocities of Gunn and Kinzer. Rain temperature of 20°C. Rain temperature of 0°C. 9 R: rain rate (mm/hr) a,b: frequency and rain temperature dependent parameters tabulated in [26] Rain Rate Measurement The determination of path attenuation due to rainfall requires the accur ate measurement of path average rain rate. This is complicated by the fact that rain is non-uniform in nature and often consists of rain cells of limited extent [21]. Therefore, to obtain an accurate path average rain rate, the individual rain rates must be sampled at as many locations along the path as possible [26,27] or be estimated using synthetic storm techniques [59-62]. In a distributed system of rain guages, the path can be treated in seg ments where the total path attenuation due to rain is the sum of the individual segment attenuations, as follows: n n A = I Ai = I ax (R± n± (1" 7) i=l i=l where, R^: rain rate at guage i (mm/hr) Aj_: length of segment i as a percentage of path length (R^): specific attenuation in dB/km at rain rate R^ An approximation to (1-7) is usually adopted, (1-8), which uses the path average rain rate directly through the assumption that variations in rain rate segments are sufficiently small that the inter-segment specific attenuations are linearly related. 10 n ^ x £ A = ax{ I 1 L *) x L (1-8) J * where, A: total path attenuation a^: specific attenuation for a determined path average rain rate (dB/km) R^: rain rate for segment i (mm/hr) £^: length of segment i L: total path length (km) *: path average rain rate 1.2.3 Multipath Fading Under normal atmospheric conditions a line-of-sight hop provides one propagation path between transmitter and receiver antennas. If certain changes in the refractivity profile occur, additional propagation paths can result parallel to the main beam [6,28]. When these arrive together at the receiver, they add together according to their phase relationship producing variations in the received signal known as multipath fading. For N multiple propagation paths of amplitude and delay the transfer function describ ing this phenomenon can be expressed as follows [6]: N -j irf T H(f) = I a e n (1-9) , n n=l 0 11 Measurement Techniques and Physical Models Investigations of multipath have either used a time-domain radar tech nique where the arrival of pulses are monitored by number, amplitude and time delay or used a frequency-domain technique where the path is swept in fre quency and Fourier transformations are used to obtain the time-domain responses [29,30,31,32]. The conclusion from these experiments was that the number of propagation paths depends strongly on the links' meteorological conditions and that the frequency of a multipath fade increases with fade depth. Sanberg, using results from frequency swept measurements concluded that no multipath events occurred which could not be characterized by four or fewer rays [33]. Other swept measurements conducted by Martin [34] concluded that fades of the order of 20 dB are primarily due to two path propagation while deeper fades, of the order of 40 dB or more are due to the existence of at least three paths. Models based on these physical parameters have been verified experi mentally for ducting and for specialized types of atmospheric layering [35] but a general model which can be applied empirically from directly measured multipath parameters has yet to be developed [6]. r Estimation of Multipath Propagation Outages In order to characterize multipath propagation in a manner which allows path availabilities to be estimated, a general formula has been developed by Barnett [36] to calculate multipath outage probabilities as follows: U = a x b x 6.0 x 10~7 x f x D3 x 10t_F/10] (1-10) 12 where, U: fade probability below fade margin a: path roughness factor (4 for very smooth, 1 for average with some roughness and 1/4 for very rough mountainous terrain.) b: factor to convert worst month probability to annual probability (1/2 for hot humid, 1/4 for average inland and 1/8 for very dry mountainous) f: frequency (GHz) D: path length (km) F: fade margin under normal operation Availability For propagation factors such as multipath the availability is given by (1-11) [2]. A = (1-U) x 100% (1-11) where: A - Availability defined as the percentage of time the received signal is useable U - fade probability defined as the fraction of time the received signal is not useable. 1.2.4 Other Propagation Factors Gaseous Absorption by Water Vapour, Oxygen and Fog A minor effect on radio wave propagation near the earth's surface for frequencies lower than 10 GHz is due to the absorption by water vapour and oxygen. This effect increases with higher frequencies. At 10 GHz the attenu ation is .007 dB/km at 20°C for oxygen absorption and .0045 dB/km at 20°C 13 for water vapour absorption [10]. The attenuation, therefore, amounts to less than one dB for an average 50 km microwave path. Fog also attenuates micro waves by the same scattering mechanism as rain but due to the much smaller drop diameters involved the amount of attenuation is minimal. Measured attenuations of 1 dB/km at 90 GHz are reported [37] which means for an average link the attenuation is considerably less than one dB for frequencies lower than 10 GHz. Bright Band Effects There is increasing evidence to suggest that the bright band causes attenuation in excess of what is normally predicted by a rain model. This is discussed in detail in section 1.4. 1.3 Improving Reliability in Path Design Path reliability (availability) is a combined function of the propagation and equipment reliabilities and therefore both need to be considered in the design of a microwave link [2]. Improvements in propagation reliability can be achieved through careful path selection and the use of frequency and space diversity to minimize the effects of multipath and rain attenuation, while improvements in equipment availability are accomplished by using reliable system components and redundant configurations. Figure 1.2 illustrates these techniques. . 14 Figure 1.2 A Microwave System Diagram Illustrating Space Diversity, Equipment Diversity and Frequency Diversity. For a typical non-diversity path of given path distance, fade margin, frequency and rain statistics the outage probabilities due to rain and multi-path fading can be calculated from equations (1-6) and (1-10) respectively. If further propagation reliability improvement is required space diversity and/or frequency diversity can be used on the link. The relationship used to calculate the space diversity improvement factor is given by Vigants [38] as follows: I__ = (1.2 x 10~3 x f x S2 x 10[F/1°])/D (1-12) where, T-SD: sPace diversity improvement factor 15 S: vertical antenna spacing in meters D: path length in kilometers F: fade margin associated with the second antenna f: frequency in GHz. Frequency Diversity Similarly, a relation to calculate the frequency diversity improvement factor is given by Barnett . [36]: IFD = a x [Af/f] x 10[F/10] (1-13) where, IFJJ: frequency diversity improvement factor a: frequency band factor (3 for the 890 - 960 MHz band 1 for the 2 GHz band 1/2 for the 4 GHz band 1/4 for the 6 GHz band 1/8 for the 7 & 8 GHz bands and 1/12 for the 12 GHz band) f: frequency (Hz) Af: frequency spacing (Hz) F: fade margin (dB) 1.4 Bright Band Effects The bright band is the transition region immediately below the 0°C iso therm where falling snowflakes melt and are turning into rain. Thus, bright-band propagation occurs when a microwave beam passes through precipitation in this region. It was named during World War II when high radar returns result ing from this melting layer region caused a "bright band" on the radar screens. 16 From a propagation point of view, there is increasing evidence which suggests that, during wet snow or sleet events, low angle microwave beams in temperate marine climates experience attenuation in excess of values predicted due to rain [39,40,41]. It has been postulated that the excess attenuation is a result of increased absorption and scattering upon transmission through the 0°C isotherm or bright band. Accurate direct measurement of this excess attenuation is difficult and there have been only few reported cases where quantitative results are given. Oomeri and Aoyagi [40] from propagation tests carried out Sapporo and Hokuriku, Japan, indicate that sleetfall attenution (in dB) to be six to seven times as large as the attenuation that can be pre dicted for the equivalent amount of rain. Watson [39], in his survey, cites propagation studies carried out in the USSR [Al] which found similar results. Measurements taken using radar also shows excessive attenuation in the presence of bright band [42-47]. In addition to these published results, there are some reports in Canada, Scandanavia and the United Kingdom of exces sive attenuation in the presence of sleet or wet snow [39]. Recently, by using stallite beacon signals to measure direct attenuation together with proven radar prediction methods for rain attenuation, more accurate measurements of bright band attenuation have been taken [44,47,48]. In this technique the rain attenuation is calculated to the base of the bright band and then this value is subtracted from the total attenuation to give an amount attributable to bright-band attenuation. Measurements taken by Hendry Antar, Schlesak and Olsen [47] using a dual-channel polarization diversity radar show a correlation of increased excess attenuation ascribed to the melt-17 ing layer with an increase in the vertical thickness of the bright band layer for a given precipitation rate. They also found that the percentage of pre ferred orientation of the particles in the melting layer to be typically 20% as opposed to 65% to 85%, for rain. An attempt to model the bright band has been made in a series of paper by Nishitsuji and Matsumoto [49,50,51] using the Ryde and Ryde approach [12,13]. Their first paper [49] establishes a set of Nrs density distribution tables of snowflake-size distributions for various classifications of snow similar to the tables of Laws and Parson. These distribution tables were prepared for four snow classifications: namely, dry snow, moist snow, wet snow and watery snow of which the last three represent snow types present in the bright band. Figure 1.3 taken from this paper illustrates these progressive changes in the character of snowflakes as they would appear on water-blue paper for snow as it passes through the bright band. In addition, Nishitsuji and Matsumoto measured the falling velocity corresponding to each snowflake diameter and for each snow classification. A graph of these results is presented in Figure 1.4 taken from [49]. The precipitation rate (P) can then be related to the density (Nrs), fall velocity (Vrs) and the radius (rr) to allow the calculation of the snow attenuation for each snow classification as follows [51]: P=47T^r3V N (1-14) 3 r rs rs The greatest attenuation for equal precipitation rates was then found to be due to watery snow in the frequency range between 4 and 7 GHz. In terms of attenuation relative to rain of the same precipitation rate the attenuation % • Large Small Rain drop or watery snow (2) • •-large Small Wotery snow (3) 0 Watery snow or wet snow (4) # Wet snow (5) wet snow or moist snow (6) Moist Snow (.7) Moist snow dry snow (B) Dry snow (9) • Graupel Figure 1.3 The Character and Classification of Snow as it passes through the Bright Band as seen on Water-Blue Paper. 10 0 .5 1.0 Radius ol ro in drop and snowfloke {cm) Figure 1.4 Falling Velocity vs. Radii of Raindrops and Snowflakes 20 due to watery snow in this range was calculated to be 15 times in dB's per kilometer. Table 1.0 presents these multiplier's in dB's per kilometer for various frequencies as derived from the model for wet snow [51]. There are several supporting physical reasons why the attenuation at a fixed precipitation rate for watery and wet snow is greater than that due to rainfall [50]: a) for a raindrop and watery snowflake of the same weight the latter has the greater radius; b) the rate of fall of watery snow as compared to a raindrop is smaller so that the number of snowflakes in a unit volume is greater than that of rain; c) snowflakes do not break up during their fall through the melting layer since the drop size spectrum just above the melting layer is similar in shape to that just below it [53]; Table 1.0 Attenuation Multipliers Due to Watery Snow. Frequency (GHz) Attenuation Multiplier (dB/km) 1 2 4 7 11 24 35 50 0.2 4.6 15.1 15.0 12.1 7.3 7.0 6.5 21 d) the distribution of raindrop radii is semilogarithmic while that of watery snow is the sum of the same semilogarithm due to aggregation at the top of the melting layer [52-57], which means there are many large sized snowflakes in the bright band. This greatly impacts attenuation since the increase is proportional to the cube of the diameter of a snowflake or raindrop. Possible Impact on Microwave Transmission Systems For paths such as earth-space links and slant paths in temperate marine climates the bright band could have a significant effect which suggests that allowances should be made for this type of attenuation when predicting propa gation reliability. Excess attenuation as a result of propagation through the bright band appears to be increasingly more important to account for as frequencies greater than 8 GHz are used. Examples showing the geometries associated with an earth-space link and a slant-path terrestial link relative to the bright band are shown in Figure 1.5 1.5 Scope of Thesis 1.5.1 The Research Program The objectives of the present propagation research program which is being carried out in association with the Canadian Research Centre in Ottawa and the British Columbia Telephone Company in Vancouver are as follows: 1) To provide the facilities to perform research into the various aspects of both analog and digital microwave propagation. 2) To select and fully instrument a suitable path(s) for the monitoring of a number of factors that affect microwave propagation. Figure 1.5 Relative Bright Geometries Between Slant Path Terrestial and Earth-Space Links 23 3) To establish the necessary data collection and data analysis infra structure to efficiently store, retrieve and analyze large amounts of propagation data. 4) To establish the relationship, if any, between the occurrence of bright band and multipath fading phenomenon and the performance of certain microwave links. 5) To be able to determine a statistical model which takes into account bright band attenuation factors that would enable improved availability design for both terrestrial and earth space paths. 1.5.2 Thesis Objectives The main objectives of the work in this thesis may be stated as follows: 1) Identification and measurement of the parameters affecting bright band propagation. 2) Development and implementation of a working telemetry based data col lection system which can accurately measure, time correlate and pre-process large amounts of received signal and meteorological data. 3) Development of a data base management system to store and analyze the data collected. 4) Analysis of the data collected to determine bright-band propagation effects in the 4 and 7 GHZ range and provide a comparison to the theoretical model 24 1.5.3 Thesis Outline A description of the path selected is the subject of Chapter II. Included is a path profile, a summary of the path's previous propagation history as well as the calculated and measured system performance charac teristics. Chapter III deals with the criteria for meteorological site selection and includes a description of the wind direction, wind speed, temperature and rain transducers through which the meteorological parameters affecting excess path attenuation can be measured. Chapter IV deals with the data handling aspects of collecting the data from received signal and meteorological sensors and the network design for doing this in real time. Included are sections dealing with the data statis tics, the link capacities, the microprocessors and the preprocessed time series and distribution series formats. Chapter V deals with the specification, the design and the implementation of a data base management system. This is discussed in relation to software systems presently in existence, to the current bright band system and to future propagation research software requirements. Included is an illustra tion of the total data processing and handling system and the economics associated with using this system. Chapter VI presents results of copolar attenuation through the bright band at 4 and 7 GHz. One set of these results has been taken from chart recordings during January-February 1980. Another set of results taken during January-February 1982 using the telemetry based data collection system as described in Chapters III, IV, and V, are also presented in this chapter. 25 These results represent a limited set to convey the character of the data measured and the nature of bright band propagation and to illustrate possible correlations to meteorological parameters. Conclusions based on both the chart recorder and the telemetry based system are presented in Chapter VII as well as suggestions to further develope the measurement system and the analysis techniques used in this thesis. 26 CHAPTER II THE EXPERIMENT 2.1 Introduction This chapter provides a description of the path selected, the equipment used and the system fade margin parameters. 2.2 The Path The path under investigation is part of the Trans Canada Telephone System microwave network and is located approximately 100 kilometers east of Vancouver, B.C., Canada. It is coastal and mountainous in nature, 41.3 kilo meters in length and lies between site elevations of 236 m and 1436 m above mean sea level. Because the path has an elevation differential of 1227 m and an annual average rainfall of 1600 mm/yr. , there is a high probability that the 0°C isotherm (and hence bright band) will exist at an intermediate eleva tion between the transmitter and receiver sites. The geographical layout and profile of the path are shown in Figures 2.0 and 2.1 respectively. Photo graphs are also provided looking directly down the path from the transmitter site (Figure 2.2(a)) and directly up the path from the receiver site (Figure 2.2(b)). 2.3 Received Signal Monitoring At the receiver site, signal levels from five microwave channels are monitored and sampled at a rate of 10 Hz. These were selected on the basis of obtaining maximum information with respect to broadband and narrowband multi-Figure 2.0 Geographical Layout of the Bright Band Propagation Experiment. Figure 2.1 Path Profile: Ryder Lake to Dog Mountain. 29 a) From Transmitter Site: (Dog Mountain) Looking South indicates location of Receiver Site b) From Receiver Site: (Ryder Lake) Looking North "+" indicates location of Transmitter Site. Figure 2.2 Path Photographs. 30 path fading, copolar attenuation and cross-polar effects. Four of the chan nels are of horizontal polarization and were selected at each end of the 4 and 7 GHz bands. The remaining channel is at 4 GHz and has a vertical polariza tion. Specific information on the frequencies selected and the radio equipment used are given in Figure 2.3. A sampling rate of 10 Hz has been selected as a compromise between the capacity of the high speed data link and maintaining the integrity of the re ceived signal data during fast fade events. Fade rates of 50 to 60 dB/sec have been observed on a similar experimental link in British Columbia [57], Therefore, a 10 Hz receiver signal sampling rate has been chosen to avoid losing this fade information and yet maintain the data flow within the capa city of the data link. The block diagrams are given for both the 4 and 7 GHz transmission sys tems used in this experiment as shown in Figures 2.4 and 2.5 respectively. These are used to calculate the received signal levels, fade margins and resulting annual outage probabilities in the transmission calculations of Table 2.0. In both the 4 and 7 GHz cases the measured and calculated received signal levels agree to within 1 dB. The resulting outage probability due to propagation events is estimated to be 4.4 minutes/year for the 4 GHz system and 31.6 minutes/year for the 7 GHz system. The received signal level is monitored from the Automatic Gain Control (AGC) feedback voltage which is proportional to the magnitude of the incoming signal. The variation of the AGC voltage versus receiver input signal level is given in Appendix A for the five microwave receivers monitored. 31 PARABOLIC HORIZONTAL HORN REFLECTOR FROM DOG MOUNTAIN * - SELECTED RECEIVER CHANNELS HORIZONTAL 878 C3 RADIO EQUIPMENT * 7142 MHz 7496.5 « MHz 7 GHz FREQUENCY PLAN MHz *• <!0!0 4090 4170 FROM DOG MOUNTAIN VERTICAL HORIZONTAL RADIO EQUIPMENT TO 2S TO 2S ^J TO 2 J RA 3 RA 3 TD 2 TD 2S TO 2 MHz 3550 * 3630 3710 3790 * 3870 4 GHz FREQUENCY PLAN Figure 2.3 Frequency Selection Plan and Receiver Equipment used at Receiver Site. HORIZONTAL 40/0 MHZ Figure 2.4 4 GHz. Microwave Transmission Block Diagram tV/? /37 <S\ 7/42. O AfMZ o 0 TO ^SC£/t/£/ZS +30 ^5, 7/42.0 M/-/Z. Figure 2.5 7 GHz. Microwave Transmission System Block Diagram 34 Table 2.0 Microwave Transmission Calculations 1) Locations Dog Ryder Dog Ryder Mtn. Lake Mtn. Lake 2) Latitude °N 49°24'35" 49°06'52" 49°24,35" 49°06'52 3) Longitude °W 121033'28M 121°54'07" 121°33'28" 121°54'07 4) Elevation (meters) 1463m 235m 1463m 236m 5) Antenna Height (meters) 20m 20m 9m 8m 6) Azimuth (°T) 217.5 217.5 8) Frequency (MHz) 4010 7496.5 9) Path Length (km) 41.3 41.3 10) Path Attenuation (dB) 136.8 142.3 11) Misc. Losses (dB) 1.0 1.0 12) Transmission Line Type WR229 WR229 WR137 WRI 3 7 13) Transmission Line Loss 0.8 0.8 1.8 3.8 14) Circulator Loss 0.6 0.6 0.8 0.8 15) Filter Loss - - -16) Connector Loss 0.4 0.4 0.6 0.6 17) Radome Loss 1.0 1.0 1.0 1.0 18) Total Loss (dB) -143.4 -153.7 19) Antenna Type Horn Horn Reflector Reflector Parabolic Parabolic 20) Antenna Gain (dBi) +39.5 +39.5 +42.9 +42.9 21) Transmitter Power (dBm) +33.0 +30.0 22) Received Signal Level -31.4 (-31.0 Measured) -37.9 (-38.5 Measun (18+20+21) 23) Receiver Threshold (dBm) -68.5 -69.5 24) Fade Margin (22-23) -37.1 (-37.5 Measured) 31.6 (-31.0 Measun 25) Propagation Availability % (Annual)* 99.9992 99.994 26) Annual Outage Estimate 4.4 Min/Year 31.6 Min/Year *The propagation availability on line 25 has been calculated using the following formula [2,36,58]: Propagation Availability % =• 100[l-ab(6.0xl0~7xfxD3xl0(~F/10))] where a • Terrian roughness factor (1 « average) b • Climate Factor (1/4 •» normal temperate) f » frequency in Gigahertz D • distance in kilometers F • fade margin in dB 35 CHAPTER III METEOROLOGICAL INSTRUMENTATION 3.1 System Design 3.1.1 Measurement Criteria In order to measure the effects of copolar propagation through the bright band it is important to be able to monitor the presence of the 0°C isotherm along the path, to determine the thickness of the bright-band region and to measure the precipitation rates through the band. The presence of the bright band can be detected by establishing the temperature differential between the transmitter and receiver sites to see if the 0°C isotherm is included. The thickness of the bright-band region, once detected, is then determined indirectly by establishing the temperature gradient through the 0°C isotherm. The precipitation rates can be monitored directly through a network of rain gauges or indirectly using synthetic storm techniques [59-62]. The latter requires the measurement of windspeed and wind direction, preferably at several points along the path to determine precipitation cell locations as a function of time. 3.1.2 Site Selection In order to meet the measurement criteria of obtaining detailed temper ature gradient, point rainfall rate information and wind information, five weather stations have been selected between transmitter and receiver sites. These sites are shown in Fig. 3.0 with their geographical and functional site 36 DOG MOUNTAIN (V ) TRANSMIT SITE RECEIVE SITE ELEV' 256 in Figure 3.0 Measurement System Layout. information provided in Table 3.0. Detailed site plans, equipment configura tions and site photographs are available for each site in Appendix C. The temperature gradient can be determined from the three sites located at beam elevation, with one at the transmitter site (V), one at mid-path (III) (at suitable elevation) and one at the receiver site (I). The point rainfall rate information is obtained from these as well as from the two remaining weather stations (II and IV) situated at intermediate sites along the valley floor beneath the path. These intermediate sites are valuable since they allow the measurement of the melted precipitation rates directly under the Table 3.0 Geographical and Functional Site Details SITE Coordinates & Elevation Data Collected I. Ryder Lake 49 06* 52" N. Lat. 121 54' 07" W. Long. Elevation: 236 m Receiver Signals 2-4 GHz. Hor. Pol 1- 4 GHz. Ver. Pol 2- 7 GHz. Hor. Pol Meteorological II. Agassiz Experi mental 49 14' 40" N. Lat. 121 47' 18" W. Long. Elevation: 15 m Meteorological III. Bear Mtn. 49 18' 25" N. Lat. 121 41' 30" W. Long. Elevation: 945 m Meteorological IV. Ruby Creek 49 21' 15" N. Lat. 121 36' 45" W. Long. Elevation: 31 m Meteorological V. Dog. Mtn. 49 24* 35" N. Lat. 121 33' 28" W. Long. Elevation: 1463 m Meteorological Primary Importance Multiplex's Field data Receiver Site Receiver Signals Rain Rate Temperature Rain Rate Intermediate Site on Valley Floor Temperature Intermediate Site at Path Elevation Rain Rate Intermediate Site on Valley Floor Temperature Transmitter Site 38 path during bright band activity. This is possible due to their much lower and hence warmer site elevations. Fig. 3.1 provides a cross-section view of the path showing the relative locations of the weather station sites selected. 0 5 10 15 20 25 30 35 40 41.5 • DISTANCE IN KILOMETRES Figure 3.1 Path Cross-section Showing Relative Locations of the Weather Station Sites. 39 3.2 Meteorological Measurements 3.2.1 Rain The rain gauges used in this thesis are of the tipping-bucket variety capable of measuring point rain rates of up to 400mm/hour. Accurate measure ment above this rate is not important because of the low probability of such events in the path area [63,64]. The rain bucket tips after each 0.318mm of rain which generates a pulse by momentarily closing a glass encapsulated reed switch relay. The pulse thus generated is latched using a signal conditioning circuit for sampling by the microprocessor after which time the latch is cleared to await the next bucket tip. For a detailed description of the signal conditioning circuit with respect to circuit schematics, physical specifications and photograph refer to Appendix E. With the path being located approximately 100 kilometers east of Vancouver, it was essential that the tipping buckets require low maintenance. For this reason a molded plastic unit which is inherently corrosion resistant was acquired. Appendix D-3 provides a detailed description and photograph of this unit. Five rain gauges were installed along the path at each of the weather station sites. Figure 3.2 and Table 3.1 show the inter-site spacings and the spacings as a function of total path length. 40 RYDER AGASSIZ BEAR RUBY DOG LAKE EXPERIMENTAL MOUNTAIN CREEK MOUNTAIN FARM . I 1 1 1 1 0.0 17.83 24.20 33.71 41.25 DISTANCE IN KILOMETRES Figure 3.2 Weather Station Inter-Site Distances. Table 3.1 Inter-Site Distances as a Function of Path Length Ryder Agassiz Bear Ruby Dog Lake Exp.Farm Mtn. Creek Mtn. Ryder Lake - .43 .59 .82 1.0 Agassiz Exp. Farm - .15 .39 .57 Bear Mountain - .23 .41 Ruby Creek .18 Dog Mountain -The total rain rate for the path is determined using a distance weighted average from the rain rates at each of the sites. The rain rate between any two interval bucket tips at a working gauge is first determined by equation 3-1. RR = tJ-P slze n ^\ ^site 1 AT(3_1) where tip Tip Size = 0.318mm AT^p = time interval between two tips. 41 Then "path averaged rain rate" (RR path) calculated using equation (3-2). (RR _ , + RR . .) x (% Path) ,„ .. . DD _ site 1 site 2' 'site 1 - site 2 path ' 2 (RR . „ + RR _) x (% Path) . . , site 2 site 3 site 1 - site 2 2 H~ • • • • + (RR .k .. n + RR „) x (% Path) , site N-1 site N site N-1 - Site N It is important to calculate this rain rate with only, those gauges which are operational and not affected by accumulations of snow. An attempt has been made to minimize errors due to these factors by rejecting rain information from gauges where no activity is observed during the passage of a precipitation event. 3.2.2 Temperature Transducer Temperature transducers were placed at all the sites to measure the out side ambient temperature. As a system, these transducers determine the pres ence of the 0°C isotherm and hence the bright band along the path as well as indirectly determining the thickness of the bright band region by establishing the temperature gradient between transmitter and receiver site elevations. Their design is based on the linear temperature coefficient of a semi conductor junction when forward biased with a constant current source. The transducer is conditioned to provide a switch selectable output of 0.01 volts per degree centigrade or per degree fahrenheit allowing for single point cali bration. Further details are given in Appendix D describing the unit as well as providing the calibration procedure. 42 3.2.3 Wind Velocity and Wind Direction Transducer All the sites, except Ruby Creek and Bear Mountain, are equipped with a propeller type anemometer with the purpose of monitoring both wind velocity and wind direction. The velocity is derived from the output of the unit's propeller-driven dc generator which exhibits a linear windspeed-voltage char acteristic. The azimuth, on the other hand, is provided by the linear output voltage which is proportional to the angle relative to true north based on the output of a gear-driven potentiometer. The wind velocity and wind speed information obtained from these trans ducers is required to be able to apply synthetic storm techniques [59-62] in the processing of the propagation data for research purposes. More detailed information on the anemometer unit is given in Appendix D. 3.3 Meteorological-Data Sampling The basic ac powered weather sites of the Agassiz Experimental Farm, Ruby Creek and Dog Mountain are each equipped with one U.B.C. Weatherlog Micropro cessor and one Meteorological Signal Conditioning Unit. A photograph of these units is presented in Figure 3.3. Details on these units is given in Appendix I for the weatherlog processor and in Appendix E for the signal conditioning unit. 43 Figure 3.3 Photograph of the U.B.C. Weatherlog Microprocessor and Meteorological Signal Conditioning Unit. The weatherlog has been programmed so that it samples the meteorological variables of wind direction, wind velocity, temperature and rainfall on a one-second software-determined interval. The output of the unit is a voice fre quency FSK modulated signal. The sampling of the meteorological variables at the Ryder Lake and Bear Mountain sites follow the same "weatherlog" system design, but use other microprocessor configurations at these sites due to differing monitoring requirements and power availability constraints. In choosing the optimum sampling rate, a trade-off is made between losing information of fast varying variables, transmitting data within the maximum link capacity and minimizing the design complexity of the signal conditioning unit. The compromise reached was to use a 1 Hz sampling rate, to transmit the data at 110 bps and design the conditioning unit to latch bucket tips until 44 sampled. This 'is a reasonable compromise since the information for all the variables except the fastest changes in wind velocity and wind direction are retained and little information is lost in the most important variables of temperature or rainfall. This leaves a growth allowances of 120% for future data requirements as additional transducers are needed. 45 CHAPTER IV DATA ACQUISITION SYSTEM 4.1 Design Criteria for a Real Time Data Acquisition System The data acquisition system must be capable of acquiring in real time the sampled meteorological and received signal level data and then be able to store this data in a format compatible for processing on a general purpose computer. To do this the data must be routed from the sites, shown in Figure 3.0, through a series of data links to arrive at the University of British Columbia for real time correlation. The design of the data acquisition system comprises three basic areas: a I) The on-site data acquisition II) The data collection network III) The real-time storage, formatting and coordination of the data. Figure 4.0 shows the data acquisition system block diagram with each of the three component areas identified. Component area I, the on-site data acquisi tion, deals with the type of data, the analog to digital (A/D) sampling rates and the interface to the outgoing data link. Component area II, the data-collection network, is concerned with minimizing system cost and delay time under certain link-capacity, link-flow and routing contraints. Component area III, the real-time storage, formatting and coordination of the data, is con cerned with the time correlation of the incoming data and the processing of the data into suitable time series and distribution series storage formats. 46 SITE NAME RECEIVED SIGNAL LEVELS (AGO CHANNEL TYPE MICROWAVE PATH UNIVERSITY OF BRITISH COLUMBIA ( U.B.C.} RAIN RATE RYOER TEMPERATURE LAKE WIND SPEED l— WIND DIREC. RR —-WS WD RR —— BEAR T —* MOUNTAIN MICRO PPO;ESSOR| (yuP) SIGNAL LEVEL CABLE DATA METEOROLOGICAL CABLE 100 tt>j DOG MOUNTAIN AGASSIZ EXPERIMENTAL FARM RUBY CREEK Z1 P H M MICROWAVE ^ 1 M MWTx VHF M ws-wo-RR -T -H M VOICE CIRCUIT M P MM VHF H M RYDER LAKE TO UBC DATA LINK VIDEO TERMINAL AREA I -AREA II AREA rr U.B.C. COMPUTING CENTRE ASEA m Figure 4.0 The Data Acquisition System Block Diagram With Component Areas Identified. 47 4.2 Site Selection 4.2.1 Received Signal Site The Ryder Lake data acquisition equipment, which monitors the received signal levels, must provide sufficient samples and quantization resolution to accurately reconstruct the received signal levels from the outgoing data stream. This applies particularly to fast fades which have previously been observed at rates up to 50-60 dB per second on similar paths in the region [57]. The received signal data transmitted must also allow easy formatting, must be sent at a rate which is equal to or less than the outgoing link capa city and must maintain overall system timing. To meet these design objec tives, an accurately determined sampling rate of 10 Hz was chosen, the received signal levels were conditioned to +0-5 volts matching the input range of the A/D and the number of outputted quantization bits were limited to eight bits to meet system data formats. The resulting data rate of 600 bits per second from the five monitored received signal levels is calculated as follows: (One - 8 bit synch byte + five - 8 bit receiver samples + 1 start bit/byte + 1 stop bit/byte) all times 10 sampling cycles per second = 600 bps. For more information concerning the receiver signal conditioning units refer to Appendix E, and for a detailed description of the Ryder Lake received signal level data acquisition microprocessor refer to Appendix I. 48 4.2.2 Meteorological Sites The purpose of the data acquisition equipment at the meteorological sites is to sample the weather variables of wind direction, wind velocity, tempera ture and rainfall at one second intervals and send these back via the site's outgoing communication channel into the data collection network. The data acquisition equipment consists basically of a microprocessor which coordinates the sampling (see Appendix I), a meteorological signal conditioning unit which provides a full range, +0-5 volt, input to the A/D (see Appendix E) and a modem which encodes the data as a frequency shift keyed (FSK) output to the data channel (see Appendix H). The outgoing data rate of 110 bps was chosen keeping future data growth requirements in mind. At present, the data sent for a one second sampling cycle includes four - 8 bit bytes, one - 8 bit synchronization byte as well as one start bit and one stop bit for each of these bytes to give a net data rate of 50 bps. This leaves an excess capacity of 60 bps for future use. The type of communications channels used from the meteorological sites are shown in the system block diagram of Figure 4.0 and includes VHF radio channels from the Bear Mountain (III) and Ruby Creek (IV) sites, a microwave radio channel from Dog Mountain (V), a telephone circuit from the Agassiz Experimental Farm (II) and a cable at Ryder Lake (I). The organization of these communication channels into a data collection network is the subject of the next section. 49 4.3 Data Collection Network Design 4.3.1 Data Statistics From section 4.2 it is evident that the data collection system must handle the statistically-different receiver and meteorological data packets. A packet of receiver signal level data is sent from site I every one-tenth of a second, consisting of 5 samples (the 3550, 3790, 4010, 7142.0 and 7496.5 GHz AGC voltages), followed by a synchronization byte. The net arrival rate for this receiver amplitude data including the start and stop bits is 600 bps. A data packet of meteorological data, on the other hand, is sent from each of the five weather station sites, consisting of four samples (wind direction, wind velocity, temperature and rainfall), followed by synchronization byte. The net data rate for the meteorological data on a per site basis including start and stop bits is 50 bps. 4.3.2 Link Capacities In order to define a network topology, the link capacities and cost for each of the outgoing communication channels were considered. Several alterna tives exist for each site but the link and associated communications channel providing for the best cost-capacity trade-off are as follows: i) Ryder Lake to UBC (telephone circuit) This link was restricted to an order-wire type telephone circuit chan nel available from the telephone company's existing network. This link corresponds to a conditioned telephone circuit with a data capacity of 2400 bps. 50 ii) Other Telephone Circuits Other telephone circuit channels were easily obtained from the Agassiz Experimental Farm and from the Dog Mountain site (via microwave) to the Ryder Lake site. The circuit from Agassiz is a dedicated unconditioned telephone circuit through the regions switched network with a capacity of 1200 bps (see Figure B-5). The Dog Mountain circuit uses the lower frequency portion of a network alarm channel and has maximum capacity of 300 bps (see Figure B-4). iii) Radio Channels Two VHF radio licenses were obtained to provide the remaining two com munication channels from Ruby Creek and Bear Mountain to Ryder Lake. These utilize the voice frequency pass band of the VHF radios providing a maximum link capacity of 600 bps. For detailed information on all these data links please refer to Appendix B where their circuit schematics, path profiles and transmission calculations are given. 4.3.3 Node Considerations Individual node design must be consistent with the overall data acqui sition objective to coordinate and time correlate all collected data. This implies that no data storage or buffers at any of the collector or acquisition nodes can delay the arrival of data at the storage node (UBC) and that the sampled data should be sent immediately (as it is monitored). 51 4.3.4 Implementation of the Network Topology The final data acquisition topology was largely determined by the con straints of cost, capacity and availability of the communication channels be tween nodes. The resulting topology is presented in Figure 4.1. Figure 4.1 Data Collection Network Topology 52 4.4 Real Time Data Storage 4.4.1 Microprocessor Considerations The acquisition system data storage microprocessor is located at UBC and handles the administrative tasks of tagging the data with real time as it arrives, of organizing the data into useful data storage formats and of out-putting the processed data onto a magnetic tape and onto a video monitor. This processor's main objective is to reduce the volume of data to manageable levels by selecting only desirable time series intervals as well as compress ing the data using hourly distributions. This technique reduces the stored data volume to approximately 10% of the total incoming data so that under normal operating conditions a 4.5 megabyte data cassette will record approxi mately 10 days of system output. To make it effective, the UBC computer is interfaced using RS 232 ports and uses two CPU's to distribute the tasks of processing, formatting and data selection. A video monitor has been added to allow the real time display of incoming data as shown in Figure 4.2. Refer to Appendix I for a complete description of the UBC microprocessor hardware and to Appendix J for a dis cussion of the UBC processor programs. 4.4.2 Data Storage Formats The data which arrives at UBC is stored in two basic formats; one for time series data and another for distribution data. The time series data includes all the arriving data whereas the distribution data format provides a cumulative hourly total of each variable for the number of samples taken at a given value. The time series data is only stored when an event occurs and |INE 63 1] 15 57 51 ws ei ei MI <w ee MY 44 44 SA 44 06 •W 55 53 45 <N ee MSZ M> 61 ta 94 61 n.n. Cl 16 81 <n ei wt.n. C7 02 D7 13 14 C4 05 N 13 14 C5 02 05 13 14 Cb 05 03 13 14 CJ 02 08 13 14 C5 03 05 13 14 C4 05 03 13 14 C8 02 03 13 14 C4 05 Ob 13 14 C4 02 07 13 14 Figure 4.2 Photograph of Video Terminal Displaying Incoming Data therefore time series dumps are random in nature and vary in length whereas the distribution buffer is dumped hourly and has a fixed length of 3890 bytes. The time series data is organized into a queue comprised of 12 complete one second blocks where each one second block is composed of 80 bytes. The dis tribution data is comprised of 5 types of data (wind direction, wind velocity, temperature, rainfall and receiver amplitude) with 5 double byte distributions for each meteorological data type and 5 double byte distributions for each receiver amplitude. Of these, each meteorological distribution has 64 double byte "bins" and each receiver signal amplitude distribution has 128 double byte "bins". Each bin represents a certain sampled value. The format used for the time series queue is given in Table J-2 and the format for the dis tribution series queue is provided in Table J-3 of Appendix J. 54 4.5 Allowance for Future Data Requirements The present data acquisition system design has left a substantial amount of data capacity unused as shown in Table 4.0. This was intended to provide the capability of adding additional propagation experiments to the network as the program progresses. If the assumption is made that the network capacity is largely determined by the Ryder Lake to UBC link then there is 65% or 1550 bps of unused data handling capacity. This is calculated based on using effective total data rates since a statistical data multiplex is used on the Ryder Lake to UBC link. Table 4.0 Data Acquisition System Link Capacities CAPACITY CAPACITY TOTAL USED UNUSED LINK CAPACITY bps % bps % 1) Ryder to UBC 2400 850 35 1550 65 2) Dog to Ryder 300 50 17 250 83 3) Ruby to Ryder 600 50 8 550 92 4) Bear to Ryder 600 50 8 550 92 5) Agassiz to Ryder 900 50 6 850 94 4.6 An Alternate Data Acquisition System Using Chart Recorders Prior to the implementation of the automated data acquisition system a system of chart recorders were used to monitor two 7 GHz microwave channels at the receive site and meteorological variables from four stations located along the path. This system produced the first preliminary results which were pre sented in two papers [65,66] and are given in this thesis in section 6.1. 55 CHAPTER V DATA BASE MANAGEMENT SYSTEM 5.1 Specifications The data base management system (DBMS) has been developed as part of this research to provide an efficient means, through the use of a general purpose computer, to handle the analysis and storage of large volumes of meteoro logical and propagation data. Within the overall data handling problem, as far as the bright-band research is concerned, DBMS takes over data handling at the point where the 800 bpi, 9-track raw data tape is transfered from the NOVA 840 to the UBC computing center. This interface is illustrated in the data system flow chart of Figure 5.0. In developing the specifications for DBMS, several general issues were considered. First, DBMS must be easy to use so that a researcher with minimal familiarity to MTS can utilize it. Second, DBMS must be flexibile to allow different propagation data series formats to be handled and yet require a minimum amount of software revision to incorporate the necessary changes. Third, DBMS must make maximum use of existing software packages developed for the 74 GHz experiment [66,70] and the library routines' resident on the MTS system. Finally, DBMS must use "time" as the universal field of access to system data records since it is inherently a common variable for all propagation data. > In addition to these general specifications DBMS must satisfy four func tional requirements: 1. It must enter new data onto the system (ENTER), fmt fro'-ofty .co / \ fmtt<or*/*g><*J^\ fTf***er0f<yco7\ Aw,'M^/«lfflA J V aarv J V data * > \^ ^ 56 *y</Cr Lit. (eeoc) freceiver ^ eft mux ^ PATLOG UBC (eeoo) DATA ACOU/3ITIOA/ DATA SASE / . DBMS CA m daM'/A4 TS) Jo* STAT/ST'tCAl €r GRAPHICAL A/VALfS'S ure 5.0 The Bright Band Experiment Data System Flow Chart 57 2. It must be capable of limited perusal and data selection from summarized results (SCAN), 3. It must be able to extract specified data records from the data base for independent analysis (EXTRACT), 4. It must allow easy application of graphical and statistical analy sis routines (PLOT). For a detailed description of the ENTER, the SCAN, the EXTRACT and the (PLOT) routines, refer to Appendix K. 5.2 Design Considerations for DBMS Relative to Existing and Future Systems In the design of DBMS, several other experiments and data base storage systems (in progress or proposed) were considered in relation to each other as shown in Figure 5.1. At the present time the 74 GHz and the bright band experiments contribute data to the UBC data base while the B.C. Telephone Company monitoring system develops its own data base. In the future, the proposed digital radio propagation experiment will contribute data to both these data bases as well as to the one at the Communications Research Centre, Ottawa, Ontario. Therefore, it would be desirable to transfer data between the systems in order to allow researchers at one location to have access to data from the other. There are several methods which can be used to transfer these data between systems, as illustrated in Figure 5.1. The most reliable method, which is also the most universal, is to transfer the raw data cassettes between systems. This method is desirable since the data processing would be under the control of the researcher who wants the results. Another reason, adding to this method's universality is that the cassette recording formats 58 74 RAW GHZ P/ITA ENTER SAIN I 1 I CRC P/OITAL | I &AIY R4TA ac TCL RAW DATA PP£SE/VrAT/OA/ /POL/r/A/fS PPfSEA/TAT/OA/ POur/A/ES Figure 5.1 DBMS in Relation to Other Propagation Data Management Systems. 59 are the same as other Canadian propagation experiments. The most efficient method to interface the two data bases would be to transfer the data on magnetic tapes between computing centers and use a bridg ing routine to reformat the data. This method is workable provided a large number of bridging routines are not required and the number of tapes to be transferred is not excessive. 5.3 A System Description of DBMS 5.3.1 Data Transfer and Handling DBMS starts inputting data by evoking the "ENTER" mode. This inputs raw data of either time series or distribution series formats for processing from an unlabelled magnetic tape, previously transferred on the NOVA system at a density of 800 bpi in half word hexidecimal format from the cassettes. Once inputted the data is entered on a disc file. For the distribution series data one preprocessing scan is made and for the time series data two preprocessing scans are made. In the latter case the first scan processes the bucket-tip information to determine rain rates and the second scan converts the data using scaling factors and look-up tables to provide the engineering units. The converted disc files can then be outputted to a labelled 6450 bpi data tape using the MTS "FILESAVE" system or further processed by evoking the "PLOT" mode to display the data in one of several options. These include time series plots of windspeed, temperature, temperature gradient, differential temperature and receiver signal strength. Figure 5.2 gives a block diagram of the DBMS software system to process the time series data. At the present time the routines to process the dis tribution series data have yet to be implemented. R DBMS OPTIONS 1, ENTER 2, PLOT 3, SCAN 4, EXTRACT PLOT OPTIONS 1, lRx, RAIN 2, 1 TEMP, lRx 3, lRx, 1 WIND 4, 2Rx 1,3 SCALE FACTORS SELECT VA~IA3LT PLOTSEE 3,4 • TEMP OPTIONS 1,1 TEMP, lRx 2, dT/dt, lRx 3, T:SITE, lRx 4, dT/dH, lRx Rx OPTIONS 1, N. RES, 1 SEC. AVE. 2, HIGH RES. lflO SEC. ENTER OPTIONS 1, TAPE 2, PERM FILE 3, PERM FILE 4, TEMP FILE 5, TEMP FILE "EAD DATA CONVERT DATA OUTPuT TO-DAT/. OUTPUT Figure 5.2 User Flowchart to Process time series data on the DBMS software system. 61 5.3.2 Estimate of DBMS Data Volumes Size estimates for the bright-band data base are of interest. The amount of data acquired is determined mainly by the event criteria set to record time series data. Assuming that the event criteria is set so that on an average, 5% of the time series data is recorded the corresponding yearly data volume estimates are presented in Table 5.0. It is estimated that two 6250 bpi 2400 foot magnetic tapes per year is necessary to store the distribution data and six 6250 bpi 2400 foot magnetic tapes per year are required to store the time series data. Table 5.0 DBMS Data Volume Estimates Distribution Volumes (in bytes)  Time Series Volumes (@ 5% in bytes 3,890/hr. 93,360/day 653,520/wk. 2,800,800/mo. 33,609,600/yr. 14,580/hr. 349,920/day 2,449,440/wk. 10,497.600/mo. 125,971,200/yr. 62 CHAPTER VI' RESULTS 6.0 Introduction Results are presented from two series of measurements. The first series were taken in early 1980 using chart recorders from which two events are analyzed exhibiting opposite dynamics in the movement of the 0°C isotherm. These results are presented in section 6.1. The second series of measurements were taken in early 1980 and uses the remote telemetry system as described in Chapters II, IV and V. Results from three events are analysed in section 6.3 showing various aspects of bright band propagation phenomenon. 6.1 Some Initial Results Obtained Using Chart Recordings Prior to completion of the telemetry based system data was collected from January 3, 1980 to March 30, 1980, using chart recorders. During this time a number of events were recorded showing the presence of bright-band effects. Table 6.0 shows two sets of data taken during separate bright-band events. The two events selected show opposite dynamics associated with the verti cal movement of the 0°C isotherm. Charts of the received signal data are shown in Figure 6.1 along with the corresponding sampling intervals that were used in Table 6.0 and 6.1. Event "A", recorded on January 11-12, 1980, shows the 0°C isotherm rising in elevation from below the path while event "B", recorded on February 2, 1980, shows the 0°C isotherm descending in elevation from above the path. The presence of the bright band is verified by tempera ture recordings and rain gauge activity at the transmitter and receiver sites. Figure 6.0 Recordings of Received Signal at 7 GHz During Bright-Band Propagation a) Event "A" (January 11-12, 1980) b) Event "B" (February 2, 1980) Table 6.0 Preliminary Results Average Average Attenuation in dB Event & Rain Rate Excess Sample No. (mm/hr) Meas. Rain* Atten. "A" 1 13.4 16.6 4.3 12.3 2 9.5 13.5 2.3 11.2 3 13.4 20.0 4.3 15.7 4 3.8 9.0 1.0 8.0 5 13.3 15.5 4.3 11.2 "B" 1 8.1 2.9 2.4 0.5 2 8.0 3.8 2.4 1.4 3 17.7 6.9 6.2 0.7 4 19.1 10.9 6.8 4.1 5 9.3 15.0 2.5 12.5 6 7.2 16.6 2.1 14.5 7 5.5 17.4 1.6 15.8 8 4.3 10.4 1.1 9.3 *Derived using the Laws and Parson Distribution for 0°C Table 6.1 Bright-Band Excess Attenuation Ratio (EAR) Results Average Rain Bright-band Excess Attenuation Event No. Attenuation Attenuation Ratio dB/km dB/km (EAR) "A" 1 0.104 0.851 8.2 2 0.056 0.775 13.9 3 0.104 1.086 10.5 4 0.024 0.553 23.1 5 0.104 0.775 7.5 "B" 1 0.058 0.036 0.6* 2 0.058 0.097 1.8* 3 0.150 0.048 0.3* 4 0.164 0.291 1.7* 5 0.060 0.865 13.1 6 0.051 1.003 19.7 7 0.039 1.093 28.0 8 0.027 0.643 23.8 *The 0°C isotherm is above the transmitter site for these data points. 65 These results show that the measured attenuation is considerably above that predicted by a rain model and, at this frequency, could only be attributed to excess bright-band attenuation. The radar reflectivity profiles presented by Dissanayake and McEwan [44] show that the bright band can be ex pected to occur for approximately 35% of the total path length. In this thesis, the excess attenuation has been defined in terms of a ratio of the excess attenuation in dB/km attributable to bright band divided by the rain attenuation in dB/km using the Laws and Parson distribution at 0°C [23], The ratio thus defined will be referred to as the Excess Attenuation Ratio (EAR). The EAR's for the corresponding sampling intervals used in Table 6.0 have been calculated and are presented in Table 6.1. The excess attenuation ratios for event "A" are generally lower than for event "B", both being determined assuming rain along the whole path. This may be explained as follows. In event "B" the 0°C isotherm started from above the transmitter site with rain attenuation occurring along the whole path. Since wet snow or sleet would have been falling at the transmitter site during event "B", a contribution to the excess attenuation may also have come from accumu lations on the radome. Taking these factors into account, the EAR's for event "A" would be lower and more in line with the EAR's found in event "A". Radome accumulations for event "A" would have been minimal since "dry" snow was present at the transmitter site and rain was falling at the receiver site for the sampling intervals selected. In both events, rapid scintillation type fluctuations of 5 to 10 Hz appeared on the received signal recordings. These fluctuations coincided with the occurrence of heavy bright-band fading and were estimated to be up to 15 66 dB in depth. It has been suggested that these fast fades are due to refrac tive multipath as a result of propagation through the bright band region.* Although these spikes were present in the recordings, the measured attenuation in Table 6.0 was averaged along a baseline assuming that the spikes were not present. In order to study the cause of these scintillations in more detail, to determine the bright-band EAR's with greater resolution and to minimize errors due to radome snow accumulation, an improved experimental design based on remote telemetry was implemented. A description of this improved measure ment system has been presented in Chapters II, IV and V and some preliminary results from this system are given in Section 6.2. 6.2 Remote Telemetry Results Showing Bright Band Propagation 6.2.1 January 23, 1982, 7:30-11:30 p.m. These results were taken after a major storm front moved easterly through the path during a period of evenly distributed wide spread precipitation. An increase in attenuation .is evident as the 0°C isotherm moves into the path as shown in Figure 6.1. The precipitation rates were measured for the decaying portion of the event as shown by the rain rate and attenuation plots in Figure 6.3. Excess Attenuation Ratios (EAR's) were calculated for the sampling points shown in Figure 6.2 and the results are presented in Tables 6.2 and 6.3. These range from 10.4 to 19.8 and agree favourably to the EAR of 15 predicted by theory [50]. The presence of the bright band was verified by temperature recordings at the transmitter and receiver sites. * Private discussions with R.K. Crane at the URSI, Commission F, Symposium held at Lennoxville, Quebec, May 1980. 67 Figure 6.1 The Agassiz Temperature and 7.142 GHz Signals Versus Time 68 Figure 6.2 Agassiz Rain Rate and The 7.496 GHz Signal Level Versus Time 69 Table 6.2 January 23, 1982 Results (7 GHz) Average Average Attenuation in dB Event & Rain Rate Excess Sample No. (mm/hr) Meas. Rain* Atten. 1 4.4 8.8 1.5 7.3 2 4.4 7.8 1.5 6.3 3 3.5 7.1 1.1 6.0, 4 4.4 6.9 1.5 5.4 5 1.3 3.0 0.4 0.2 *Derived using the Laws and Parson Distribution for 0°C Table 6.3 Bright-Band Excess Attenuation Ratio (EAR) Results Average Rain Bright-band Excess Attenuation Event & Attenuation Attenuation Ratio Sample No. dB/km dB/km (EAR) 1 0.036 0.51 14.0 2 0.036 0.44 12.1 3 0.027 0.42 15.4 4 0.036 0.37 10.4 5 0.009 0.18 19.8 *The 0°C isotherm is above the transmitter site for these data points. These results show that the measured attenuation is considerably above that predicted by a rain model and again as in the chart recordings could only be attributed at this frequency to excess bright-band attenuation. Although rain information was taken at only one site the uniform and widespread nature of this event's precipitation gives consistent results with theory. 70 In this event which is similar to the later part of event B in Figure 6.0, no rapid scintillation type fluctuations were observed. This corre sponded to gradual changes in temperature over this period. 6.2.2 January 23, 1982, 2:00-4:30 p.m. The system front for the event on January 23, 1982, passed through the microwave link between 2:00 and 4:30 p.m. producing fades of up to 30 dB on the 7 GHz receivers and up to 19 dB on the 4 GHz receivers, as shown in Figure 6.3. The bright band was present along the path as indicated by the tempera ture recordings for the Dog Mountain transmitter site and the Ryder Lake receiver site in Figures 6.4 and 6.5 respectively. As shown on Figure 6.5, there appears to be a correspondence between the passage of the temperature through 0°C at Ryder Lake site with the discon tinuities in received signal level recordings at 7 GHz. Figure 6.6 shows an expanded view of Figure 6.5 around 80 minutes into the event showing details of this correspondence. Expanded views of the differential temperature around the discontinuities shows a corresponding increase in the rate of change of temperature at these points. Figure 6.7 shows the increase in differential temperature for the fade discontinuity 80 minutes into the event and Figure 6.8 shows discontinui ties earlier in the event around 30 minutes. A similar correspondence is seen in the Ryder temperature and the rapid changes in the 3.550 GHz signal as shown in Figure 6.9. These changes show that as the 0°C isotherm moves into and out of the path there are corresponding rapid changes in received signal attenuation. 71 Figure 6.3 The 3.550 GHz and 7.496 GHz Signal Levels Versus Time PRELIMINARY BRIGHT-BAND RESULTS BASED DN JANUARY 23. 1982 7ELEHETRT DfllA Figure 6.4 The Dog Mountain Transmitter Site Temperature and The 7.496 GHz Receiver Level Versus Time 73 Figure 6.5 The Ryder Lake Receiver Site Temperature and The 7.496 GHz Signal Reciever Level Versus Time PRELIMINARY BRIGHT-BAND RESULTS BASED ON JfiNURRY 23. 1982 TELEMETRf DflTfl jAPPRox. so MI:TLTES INTO EVENT  ^ i-* 2D J'.D 4^ ?o ^ ?T TIME (MINUTES) 8-B Figure 6.6 An Expanded View At Approximately 80 Minutes Into The Event Showing The 7.496 GHz Receiver Signal Level Versus Time 75 Figure 6.7 Ryder Lake Differential Temperature and 7.496 GHz Receiver Signal Level Versus Time Showing The Discontinuity 80 Minutes Into The Event 76 Figure 6.8 The Ryder Lake Differential Temperature and 7.496 GHz Receiver Signal Level Versus Time Showing Discontinuities 30 Minutes Into The Event. 77 Figure 6.9 Ryder-Lake Windspeed, Temperature, and The.3.550 GHz Reciever Signal Level Versus Time 78 Bh a 4-K-5 tun UJ £L CD PRELIMINARY BRIGHT-BAND RESULTS BASED ON JflNURRT 23. 1982 TELEMETRY DflTfl RYDER ULW WINDSPEED 11 3.550 GHz 1 1— 1 1 1 1 1 1 T 1 D.Q 3.0 6-0 9.0 12.0 15.0 IB.a 21.0 34. D 77.0 30.0 TIME (MINUTESI Figure 6.10 An Expanded View Of The Ryder Lake Windspeed and 3.550 GHz Receiver Signal Level Versus Time Approximately 90 Minutes Into The -Event 79 Figure 6.10 shows an expanded view of the Ryder windspeed and 3.550 GHz signal versus time starting before the 90 minute discontinuity. The sudden reduction in received signal level was preceded by a sudden drop in the wind-speed. Rain information for this portion of the event was not available since the rain buckets were obstructed by ice. 6.2.3 February 19, 1982, 7:30-9:00 a.m. These results were taken during the easterly passage of a storm front through the path on February 19, 1982. The presence of the 0°C isotherm dur ing this event is indicated by the temperature recordings for the transmitter and receiver site in Figure 6.11. Excess Attenuation Ratios were calculated for both 4 and 7 GHz and are presented in Tables 6.4 and 6.5 according to the sample points shown in the received signal level and rain rate plots in Figure 6.12. These results generally show that the measured attenuation during the presence of the bright band is significantly greater than that due to pure rain. Although there is a great variability in the Excess Attenuation Ratios for this event this could be attributed to rain rate for the path being taken at only one site (Agassiz) and to the turbulent nature of the storm front as it passed through. Evidence of the turbulent make-up of this event is observed in the rapid fluctuations of the Ryder Lake temperature and windspeed plots shown in Figure 6.11. Scintillation like fading phenomenon which appeared first on the chart recordings in section 6.1 also were observed in this event in the 7.496 re ceived signal levels. More observations and analysis needs to be undertaken Figure 6.11 Ryder Lake and Dog Mountain Temperatures, Ryder Lake Windspeed and The 7.496 GHz Receiver Signal Level Versus Time 81 Figure 6.12 Agassiz Rain Rate, The 7.496 GHz and 4.010 GHz Receiver Signals Versus Time Table 6.4 February 19, 1982 Results Frequency Average Average Attenuation in dB & Rain Rate Excess Sample No. (mm/hr) Meas. Rain* Atten. 4.010 1 0.6 0.6 .04 0.56 2 4.0 1.6 .20 1.40 3 11.2 2.5 .57 1.93 4 9.0 2.4 .45 1.95 5 4.0 1.0 .20 0.80 6 1.2 0.7 .06 0.64 7.495 1 0.6 1.5 0.17 1.33 2 4.0 6.0 1.36 4.64 3 11.2 8.5 4.50 4.00 4 9.0 6.5 3.50 3.00 5 4.0 3.0 1.36 1.64 6 1.2 3.7 0.36 3.31 Derived using the Laws and Parson Distribution for 0°C Table 6.5 February 19, 1982 Results Frequency Average Rain Bright-band Excess Attenuation & Attenuation Attenuation Ratio Sample No. dB/km dB/km (EAR) 4.010 1 .00103 .044 42.7 2 .0048 .097 20.2 3 .0137 .134 9.8 4 .0110 .135 12.3 5 .0048 .056 11.7 6 .00145 .044 30.3 7.496 1 .0041 .092 22.4 2 .0329 .321 9.8 3 .1110 .277 2.5 4 .0860 .208 2.4 5 .0329 .114 3.5 6 .0088 .229 26.0 82.5 before definite conclusions as to the cause of these scintillations can be determined. This data does suggest, however, that these rapid fades corre spond to rapid changes in temperature and that these could result in changes to the thickness of the bright band along the path. The spikes in the temperature plot change at rates of up to 2°C per second at each side of the fade transition as shown in Figure 6.13. The feasibility of differential temperature physically changing at these rates is shown by Thompson et al. in their paper on atmospheric turbulence measurements [71]. In this particular event the temperature fluctuations decrease to just above 0°C which would suggest that the thickness of the bright band would increase due to a reduction in temperature gradient for these periods. Further research is being carried out to find the cause of this scintillation phenomenon. 83 Figure 6.13 Ryder Lake Differential Temperature and The 7.496 GHz Receiver Signal Level Versus Time 84 CHAPTER VII CONCLUSIONS AND DIRECTIONS FOR FUTURE WORK 7.1 Conclusions Previous observations using chart recordings [64,65] showed that excess attenuation at 7 GHz due to the bright band is considerable with the magnitudes being consistent with those predicted by Nishitsuji and Matsumoto for watery snow. Based on these preliminary observations, the excess attenuation ratios ranged from 7 to 28 times (in dB) while the model for watery snow predicts an excess attenuation ratio of 15 at 7 GHz. The large variance between the measured values from those predicted can be attributed to several factors; first, the inaccuracies which can be attributed to a low density rain bucket network for determining rain rates in turbulent events, second, deviation of the thickness of the bright band from the assumed 400 m and third, differences between the path averaged precipitation rate from that of a precipitation rate averaged solely within the bright band. The results, however, suggested that the Nishitsuji and the Matsumoto model for watery snow was applicable and therefore variations between measure ment and that predicted by theory would most likely be attributable to the experimental methods used. The remote-telemetry-based measurement system provided accurate correlations between meteorological phenomenon and received signal levels, confirmed the scintillation phenomenon superimposed on the broad band fade previously observed with the chart recorders and showed that bright band propagation was affected by the turbulent meteorological phenomenon during the 85 first phases of a storm system which became more stable as the storm decayed. Excess Attenuation Ratios consistent with theory were obtainable during the decaying phases of a storm where widespread uniform precipitation rates were occurring. Under these conditions (unfortunately only one was operating) a single rain bucket near the center of the path generally provided sufficient information. However, when the path was influenced by a storm front the rain information from a single site was not enough to obtain consistent results. The scintillation phenomenon during fading appeared to correspond with rapid fluctuations in temperature around 0°C but further measurements and analysis are required to draw firm conclusions regarding the physical mechanisms resulting in these observations. 7.2 Directions for Future Research This thesis lays the groundwork to evaluate the proposed 8 GHz digital radio with respect to its availability under various propagation conditions. During this intial bright-band study it has become clearer that factors which can be attributed to the bright band may also significantly impact the propa gation of the 8 GHz digital radio on certain slant paths. For example, the rapid scintillations observed during deep bright band fading may affect the bit error rate performance [5,6] of the radio. Another factor of considerable concern is that the EAR is expected to be 15 at 8 GHz [51] and since rain attenuation starts becoming a significant factor at this frequency it makes bright band attenuation even more important. As shown by the recordings in Figure 6.0 under heavy widespread rain conditions a bright band fade can last for several hours making it an important design variable. Therefore 86 to derive maximum benefit from the next phase in this research program, the digital radio and the bright band experiment should be run concurrently. Figure 7.0 gives two alternate system block diagrams to accomplish this, the first uses the data link to send the real time clock as data and the second option uses a WWVB time recorder at each site. In terms of improving the present system, the following suggestions can be made: (a) The temperature measurements at Ryder Lake (I) should be made to provide greater resolution around 0°C and be sampled at 10 Hz to retain information on its variations. (b) A sturdier anemometer and radiant heater should be installed at the Dog Mountain site. (c) A time-lapse camera should monitor the Dog Mountain transmit antenna radomes for accumulations during precipitation events. (d) A heating system should be developed to eliminate accumulations of snow in the rain buckets. (e) The code for the SCAN and EXTRACT features of DBMS needs to be implemented. (f) A graphical routine could be developed to plot temperature contours on a path cross section. (g) A statistical routine needs to be incorporated in DBMS to determine fade probabilities from the distribution series data. Long term objectives in the area of bright band propagation research should include: 87 A detailed evaluation of other slant paths such as earth-space links with respect to bright band effects and their potential effect on sys tem availability. The Nishitsuji and Matsumoto models describing snow attenuation by snow classification should be extended to the bright band. This extension should determine attenuation for the bright band by first establishing its characteristic snow profile in terms of moist, wet and watery snow and then calculating the attenuation as a function of temperature gradient (or bright band thickness) and precipitation rate. A model needs to be developed to account for the scintillation type which corresponded to rapid changes in temperature. 88 PATA (5 &4PJO SIGNALS P/G/T/ll ty/?oc£Ssoe.\ &4/VP PAT* P/OfTAL \T/M£ S/e/Grr~r J&fP/O P/G / Ts4L M/CXO • Y'/eocf-ssoK WKJI/B r/MS UBC M/C&O-# • • Bs4/VP /&4P/0 Py4T>* /eyptzy^ Figure 7.0 Proposed System Configurations to Incorporate the Digital Radio Monitoring System 89 APPENDIX A AUTOMATIC GAIN CONTROL (AGC) CALIBRATIONS To obtain accurate received signal level values the AGC feedback voltage must be accurately calibrated. This is achieved by inserting a known signal level of the correct frequency into the input of the microwave receiver and monitoring the AGC voltage output. In this experiment the AGC voltages were monitored at the output of the receiver signal conditioning circuit (see Appendix E) and plotted against the receiver input signal level to produce the AGC curves shown in Figures A-l to A-5. These curves then form the basis for the look up tables and interpolation routines used as part of the entry procedure software in the data base management system as described in Chapter V and Appendix K. The receiver frequencies, polarizations and associated AGC curves used in this experiment are given in Table A-l. TABLE A-l Receiver Frequencies Polarizations and Associated AGC Curves Receiver Frequency (MHz) Polarization V = Vertical H = Horizontal Associated AGC Curve by Figure # 3550 3790 4010 7142.0 7496.5 H V H H H A-l A-2 A-3 A-4 A-5 FIGURE A - I 3550 MHZ RECEIVER AGC CALIBRATION RECEIVER INPUT LEVEL dBm RECEIVER INPUT LEVEL dBm 95 APPENDIX B DETAILS OF THE DATA ACQUISITION SYSTEM LAYOUT Figure B-l gives a 1:250,000 topographical map showing the system layout for the UBC microwave propagation experiment including all the sites and associated data links. These include the Dog Mountain to Ryder Lake microwave path, the outgoing high speed data link to UBC and the end-links to the inter mediate sites of Ruby Creek, Bear Mountain and the Agassiz Experimental Farm. A more detailed description for all these links is given in Figure B-2 to Figure B-7. Figures B-2 and B-3 provide detailed path profiles on 4/3 earth paper for the VHF radio links from Bear Mountain to Ryder Lake and from Ruby Creek to Ryder Lake, respectively. Table B-l presents the VHF transmission calculations for these two paths using techniques developed by Bullington [11,67] and Okumura et al. [68]. Detailed circuit layouts, showing levels, entry points and routing, are provided in Figures B-4 and B-5 for the links from Dog Mountain to Ryder Lake and from the Agassiz Experimental Farm to Ryder Lake, respectively. Finally, a detailed circuit and routing diagram for the Ryder Lake to UBC data link is shown in Figure B-6 and a detailed illus tration of the associated RS 232 interface configurations between the micro processors and the statistical multiplex units is given in Figure B-7. 96 TABLE B-l VHF Radio Path Transmission Calculations 1) Locations Bear Mountain Latitude °N 49°18'25" Longitude °W 121°41'30" Elevation (Meters) 945 Antenna Height (Meters) 7 DOC Call Sign Azimuth (°T) Frequency (MHz) Path Length (km) Path Attenuation (dB) Shadow Loss (dB) VGK 927 Ryder Lake 49o06,52• 121o54,071 236 15 VGK 928 218°T 160.11 26.1 -104.7 0.0 Transmission Line Type RG 58 Line Loss (dB) -1.0 Connector Loss (dB) -1.0 Misc. Loss (dB) -0.5 RG 58 -2.5 -3.5 -0.5 Total Losses (dB) Antenna Gain (dBd) -113.7 9.0 +11.0 Transmitter Power (dBw) -5.2 Received Signal Level (dBw) -98.9 (16+17+18) Threshold @ 12 dB SINAD (dBw) -149.0 Fade Margin (dB) SINAD (dBw) 50.1 Availability % Annual 99.9998 [67,68,69] Ruby Creek 49°21'15" 121°36'45" 31 10 VGK 926 Ryder Lake 49o06,52• 121o54'07" 236 15 VGK 928 212°T 151.79 33.7 -102.4 26.0 RG 58 -5.0 -1.0 -0.5 RG 58 -2.5 -3.5 -0.5 -141.4 +9.0 +11.0 + 4.8 116.6 -149.0 32.4 99.989 97 FIGURE B-l PATH SYSTEM LAYOUT FOR UBC MICROWAVE PROPAGATION EXPERIMENT l_jf >-Cultus Lake Chitttark Iwtr /y?™^^p.rk\ I'-i^r^.^ »A PROVINCIAL • :»72 FOREST^°°nl_ i 49°00-, ~ i N1TED STATES • 122°00' OF AMERICA 650 600 Figure B-2 Path Profile: Ruby Creek to Ryder Lake DISTANCE IN KILOMETERS Figure B-3 Path Profile: Bear Mountain to Ryder Lake 99 DOG MOUNTAIN DOG MOUNTAIN MICRO PROCESSOR V.f. DATA OUTPUT 2.0V P-P LEVEL - l6dB RYDER LAKE 34A 3 4A GROUP GRP. SHELF SH. UBC DATA ON GROUP N9 2 CHANNEL Ns 5 OF SIG ALARM SYSTEM BNC V.f. DATA INPUT 16V P-P RYDER LAKE MODEM BOX Figure B-4 Circuit Layout from Dog Mountain to Ryder Lake AGASSIZ EXPERIMENTAL FARM MICRO -PROCESSOR a < Ml a IE BNC Z v.f. DATA OUTPUT 10V P-P CUSTOMER INTERFACE EXCHANGE ROUTING RYDER BNC LAKE MODEM 0 >> BOX T " J V.f. DATA INPUT 8 V P-P Figure B-5 Circuit Layout from the Agassiz Experimental Farm to Ryder Lake RYDER LAKE R52J2 INTERFACES 9 C TEL RYDER LAKE DTI •*1 OATA CHANNELS FROM UBC < TELEMETRY ) SYSTEM ' C.I • « iHen IYHCHRONOUI MUX tOI C MOMM W M r-4112 LOOP UCI UNIT PAP 4 It «• I—I o C X R DROUPt CHANNELt VANCOUVER RYDER LAKE 14 A BROUP t CHANNEL I r L. CUTU WK III UBC ELECTRICAL ENOINEERINB BUILOINS ROOM 44B I •** I <1>. 1 II... {-•) (•«•• »•«• •Ma uec RSE» INTERFACES 4111 LOOT BACK UNIT w 201 C MODEM SYNCHRONOUS MU> PN4I00II TO UBC OAYA LOG DATA COLLECTION MICROPROCESSOR t CNAMMIl. VANCOUVER RYDER LAKE 34 A CROUP 2 CHANNEL t SCHEMATIC FOR THE RYDER LAKE TO UBC DATA CIRCUIT O 101 PIN I Rx 2 Tx 3 4 5 6 7 6 20 22 25 RYDER LAKE PIN <— —• PIN DESIGNATIONS U.B.C. PIN RS 232 DATA TERMINAL EQUIPMENT (DTE) DATA COMMUNICATIONS EQUIPMENT (DCE) DCE PIN n I 2 3 4 5 6 7 8 20 22 25 DTE RYDER MICROPROCESSOR OR MODEM BOX RYDER SUPER MUX 480 UB.C. SUPER MUX 480 U.B.C. DATA LOG MICROPROCESSOR Figure B-7 The RS232C Interface for the Ryder Lake to UBC Data Circuit 102 APPENDIX C EQUIPMENT AND SITE LAYOUTS C-l Ryder Lake This site is the data concentration node where all the experimental field data are collected before being statistically multiplexed onto a telephone circuit for transmission to the University of British Columbia. The on-site data acquisition system serves two main purposes: first, to collect received signal level data from five selected 4 and 7 GHz microwave channels and, second, to monitor the wind speed, wind direction, temperature and rain rate variables at the site. Figure C-l is a photograph of the site, Figure C-2 shows the equipment configuration and Figure C-3 illustrates the site layout. Figure C-l Ryder Lake Site Photograph ANEMOMETER HAJN tUCKET imp omecriwi KETtOlltV COMCMTIOMIM I T »»ov DC •ana ymo imp • IKO DIHCCT. TI4I MMi »«C-M»i MNl. AtC -•ISO MMl. *«C • ITSO MMl. MC -4010 MMl. A»C" TII.HK7.tH0. 9RL 816 fUl ANTENNA MIN' DO AM TO 000 MOUNTAIN 4 VIA 37A OROUP t CHANNEL 9 (SEE FIO 0-4) TO AGASSI! 4 EXPERIMENTAL FARM (SEE FIO 8-51 ninr •Ptirrta vttr RADIO HCCCIVI* IHCCIIVIII U.B.C. CATALOG MICRO PROCESSOR (6800) RYDER LAKE UBC MODEM UNIT RYDER LAKE "TIT" 73 •C-l NC—U NC * RSIS2 INTERFACE (SEE FIO »-7) fc=d no VAC _L_ OC POWER SUPPLY • TO UB.C VIA J4A OROUP t CHANNEL t (SEE FIOO-OI OC POWER SUPPLY RYDER LAKE SITE EQUIPMENT CONFIGURATION FIGURE C - 2 MAR. 12,110! w.l c. nfli IOI •tola LUI o BAT4 LM ITICI LUI O OC POWER SUPPLY FOR VHF RADIOS BEAR MTN. VHF RECEIVER RUBY CREEK VHF RECEIVER -METEOROLOGICAL CONDITIONING UNIT M»u mix 410 mm* uui 4«o > ASYNCRONOUS MULTIPLEXER 201C -MODEM - SYNCHRONOUS MULTIPLEXER 4112 LOOPBACK MODULE FRONT VIEW OF THE UBC EQUIPMENT BAY AT RYDER LAKE 1 VHF RECEIVE ANTENNA ANEMOMETER TELEPHONE CIRCUIT PLAN VIEW OF THE RYDER LAKE SITE RYDER LAKE SITE LAYOUT FIGURE C- S MAR 12,196?. o 105 C-2 Dog Mountain The Dog Mountain site which is the transmitter site and the northern terminus of the microwave link under investigation, has an elevation of 1463r. A weather station located there senses meteorological information and sends It back to Vancouver via a spare data channel on the 51G alarm system to the Ryder Lake site. At Ryder Lake this information is statistically multiplexed for retransmission to UBC. Access to the site is by cable car as shown in the site photograph, Figure C-4a. Figure C-4a Dog Mountain Site Photograph The equipment configuration is given in Figure C-5 and a site layout (relevant to this work) is shown in Figure C-6. 106 This site experiences severe icing conditions which have caused the destruction of the experiment's initial anemometer, as shown in Figure C-4b. To prevent this destructive ice build-up in the future it is recommended that a radiant heater be installed directed at the anemometer*. The weather micro processor could be used to control the heater during icing conditions by turning it on only during a -5°C to +3°C temperature range. Figure C-4b A Photograph Showing Damage to the Anemometer Caused by Severe Icing Conditions at Dog Mountain ANEMOMETER KfTtOM-LMICAi. IJtUl r 110 VAC 1 OC POWER BUPPLV HPO PIPICTIOW »'»» •»» UBC. WEATHERLOO MICRO PROCESSOR (BOSS) DOS MOUNTAIN VF. DATA OUTPUT IV p-p IjooJ tiit PAD • WAT «PO WU MIO0I 34 A OROUP I CHANNELS 0 no NE ?N 910 LARM CHANNEL TO U.B.C. VIA RYDER LAKE UE FlOt C-t 00 •-• TO II t AIMN tTPTIM PROPOSED PROPOSEO CAMERA HEATER CONTROL CONTROL DOG MOUNTAIN SITE EQUIPMENT CONFIGURATION •'• FIGURE C-S ISSUE I FEB. 26,1981 TO 34* CHANNEL (ONI CHS) \ V/// TEMPERATURE PR06E OC POWER SUPPLY -BC TEL EOUIPMENT »U «e«TNU LM DO# MOUNTAIN r PROM METEOROLOOICAL SENSORS use. EQUIPMENT -BAY -METEOROLOOICAL INTERFACE UNIT BATTERY AREA CABLE FROM METEOROLOOICAL SENSORS THE PLAN VIEW OF DOG MOUNTAIN SITE ///////;/ / /"/ / -ANEMOMETER FRONT VIEW OF THE UBC EOUIPMENT BAY AT DOG MOUNTAIN DOG MOUNTAIN SITE LAYOUT FIGURE C-8 FEB. 28,1981 o 00 ln9 C-3 Agassiz Experimental Farm This site has an elevation of 15m, is 17.8 kilometers from the Ryder Lake receive end and is 200 meters off beam-center. Its primary purpose is to collect meteorological data with particular emphasis on the measurement of rainfall. This site is co-located with the experimental farm's weather station which provides a source of valuable back-up data, as well as gives information on other variables such as pressure and humidity which are not monitored in this experiment. Being situated 610m below beam elevation, melted precipitation rates can be measured using a tipping bucket rain gauge even though the freezing level may be at or below path elevation. The data thus collected are sent to Ryder Lake via a telephone circuit and from there on to UBC, using the statistical multiplex. Refer to Figure C-7 for a site photograph, Figure C-8 for the equipment configuration and Figure C-9 for the site layout. Figure C-7 Agassiz Experimental Farm Site Photograph 8—r-<3 ANEMOMETER TEMPERATURE WHO OmiCTIM W'WD imp IIO VAC _J POWER SUPPLY BNC WEATHERLOQ MICRO PROCESSOR (8085) AOASSII EXPERIMENTAL SITE BNC =13 vr DATA OUTPUT lOVp-p • C TtL CUITOIttA IMUNrtCI TO U.B.C. POC— VIA RYDER LAKE IMS n«* e-i A AGASSIZ EXPERIMENTAL FARM SITE EQUIPMENT CONFIGURATION FIGURE C-8 h-1 O j- BC TCUPHONf / OUTLET I T*»LI ro* THfl lUIIII •lr>CHII1II1T«L MM HITCOItOLOflCAL (MTKUM UTI y WALL _1 U.i.C •TIATHItLM mini o 0 I1HIMH1IL WMMO HETEOBOLOOICAL SIGNAL CONDITIONING UNIT FRONT VIEW OF THE UBC EOUIPMENT AT THE AGASSIZ EXPERIMENTAL FARM PLAN VIEW OF THE AGASSIZ EXPERIMENTAL FARM SITE AGASSIZ EXPERIMENTAL FARM SITE LAYOUT FIGURE C-9 FEB. 2S.I9SI 112 C-4 Ruby Creek The Ruby Creek Site has an elevation of 31m, Is 300m off beam-center and 33.9 kilometers from the Ryder Lake receive end. The site provides for power, a tower and a shelter. It also lies 1158m below beam elevation, making it well suited for use as an intermediate weather station for monitoring melted precipitation rates. The sampled data are sent to Ryder Lake via a VHF radio link and from there it is assigned a data multiplex channel to complete the routing to UBC. For detailed information, refer to Figure C-10 for a site photograph, Figure C-ll for the site's equipment configuration and Figure C-12 for the site layout. Figure C-10 Ruby Creek Site Photograph NO CONNECTION TO ANEMOMETER Tiff IN* RAIN tUCXET TEMPERATURE ~ It ItETEORO-LORICAL •IQRAL CONDITION!* UNIT •»— . RAIN CLEAR <^HC WIHO Ol RECTI ON MC^ j . NC VINO •Mtft NC, IIOVAC I OC POWER SUPPLY U3.C. WEATHERLOG MICRO PROCESSOR (SOBS) RUBY CREEK -o 191.76 MHi GAIN 9.0 4B* 2I7.4*T POWER SUPPLY TO UBC VIA RYDER LAKE (SEE rio'i Ct • Ml VHF RADIO TRANSMITTER torn TO RMC RUBY CREEK SITE EQUIPMENT CONFIGURATION «0I> IIKN UK WCATHCRLOO DC POWER SUPPLY FOR VHF RADIO FRONT VIEW OF THE UBC EQUIPMENT AT THE RUBY CREEK INSTALLATION PLAN VIEW OF THE RUBY CREEK SITE RUBY CREEK SITE LAYOUT ISSUE f 115 C-5 Bear Mountain The Bear Mountain site is located at beam elevation on a mountaintop near mid-path. It has a site elevation of 945m and is situated 26.3km from Ryder Lake and 3km to the northwest side of the microwave beam. The Bear Mountain site is important for investigating the bright band region since it allows an Intermediate temperature point to be monitored thus enabling a more accurate temperature gradient profile to be established allow ing a more accurate determination of the bright-band thickness. In the final site selection, special consideration has been given to finding a sheltered Figure C-l3 Bear Mountain Site Photograph 116 location that would minimize misleading readings due to local convection effects. A photograph of the site is illustrated in Figure C-13. The weather data will be monitored by a battery-powered data acquisition system and transmitted to UBC via a 0.3 Watt VHF radio link to Ryder Lake. The equipment configuration for this system is shown in Figure C-14 and the site layout in Figure C-15. C-6 University of British Columbia Recording Terminus The University of British Columbia (UBC) site is located in Room 448 of the Electrical Engineering building and is the terminus for all the data col lected. The data which arrive via a telephone circuit, are demultiplexed, coded with the sampling time and condensed for storage on magnetic cassette tape. Facilities are provided for the real time viewing of the incoming data using the video terminal and the chart recorder together with the digital to analog convertor. A photograph of the UBC terminus is shown in Figure C-16. The equipment configuration is shown in Figure C-17 and the site layout in Figure C-18. •am CIIH BNC 0 \) (00 tl MH] u 0AINSOI rt tH.6«T ; TO U.B.C •y RYDER LJ U.B.C. REMOTE WEATHERLOG MICRO PROCESSOR (1802) BEAR MOUNTAIN JUL. DC POWER SUPPLY (BATTERY! CAUSTIC POTASH 1000 A-MR CAPACITY BEAR MOUNTAIN SITE EQUIPMENT CONFIGURATION •' FIGURE C-14 ISSUE I MAR IC.ISS2 SIDE VIEW OF EQUIPMENT TREE PLAN VIEW OF THE BEAR MOUNTAIN SITE FIGURE C-15 •MR. 12.198 J 00 119 Figure C-16 University of British Columbia Site Photograph FROM RYDCR LAW AND OTHER _ , METEOROLOOICAL 300C* SITES I Ml Flit • •I, C-l, O-l, c-i, c-uae-tll IBM 3010 COMPUTER VIDEO TERMINAL • fK> D/A CONVERTOR i ~M DC POWER SUPPLY CHART RECORDER MAONETIC TAPE RECORDER UNIVERSITY OF BRITISH COLUMBIA EOUIPMENT CONFIGURATION FIGURE C-17 MAR. 12,1982 WINDOW IBM 3101 VIOCO TERMINAL /lP N» I OATA a FORMATTER. MICRO PROCESSOR RACK Ml OATA PROCESSOR STNC MU> M^D'EM SUPER MUX 480 SUPER MUX 480 ^ WALLS] \ \ \ \ \ FRONT VIEW OF THE UBC EQUIPMENT TABLE I POWER STRIP IBM 9101 VIDEO TERMINAL OA CONVERTOR rtPt mcm. j^j RECORDER PROCESSORi CHART RECORDER ROOM 448 HECTOR MCCLOUO BUILDING SUPER MUX 4*0 PLAN VIEW OF THE UBC SITE UNIVERSITY OF BRITISH COLUMBIA SITE LAYOUT FIGURE C - 18 MAR. .It, 1982 122 APPENDIX D METEOROLOGICAL TRANSDUCERS D-l The Anemometer The anemometers used are of the propeller-vane variety capable of measur ing both wind speed and wind direction. A photograph of an anemometer is shown in Figure D-l. Figure D-l A Photograph of the Anemometer To be able to monitor wind speed and wind direction using the anemometer, three separate input and output signals need to be Interfaced to the unit, as shown in the circuit and wiring diagram of Figure D-2. These include the potentiometer input excitation voltage for azimuth, its output voltage and the wind speed output voltage derived from the propeller-driven DC generator. 123 Figure D-2 Anemometer Circuit & Wiring Diagram These units have proven to work reliably at all sites except for Dog Mountain where severe icing conditions damaged the initial anemometer instal lation. A radiant heater has been recommended to resolve this problem, as indicated in Appendix C-2. D-2 The Tipping-Bucket Rain Gauge The rain gauges used are of the tipping-bucket variety and are capable of measuring point rain rates up to 400 mm/hr. Accurate measurement above this rate is not important due to the low probability of such events in this area. The rain gauges have a collecting area of 383.6 cm2 and a nominal tip size of 12.2 grams or 12.2 cubic centimeters of water so that a tip occurs 124 after each .318 mm of rain. This calibration is done by first pouring a known quantity of water through the bucket and counting the number of tips to give the volume of water per tip. Then, by dividing the volume of water per tip by the collecting area of the bucket, the rainfall per tip, is calculated. Electrical sensing is accomplished by generating a pulse as the bucket tips by passing a permanent magnet, attached to the bucket's tipping arm, in near proximity to a chassis-mounted reed switch. Rain bucket maintenance has been minimal during the last year of operation, largely due to their plastic corrosion proof construction. Figure D-3 shows a photograph of the rain bucket and its tipping assembly. Figure D-3 A Typical Rain Bucket and Tipping Assembly. 125 D-3 The Temperature Transducer The operation of the temperature transducer is based upon the linear temperature coefficient of a semi-conductor junction when forward biased with a constant current. The forward voltage which results across the junction varies linearly with temperature and is amplified to produce a DC output volt age of 10 millivolts per degree with separate gain circuits to provide indica tions In degrees Centigrade or degrees Fahrenheit. A photograph of the temperature transducer appears in Figure D-4. The temperature sensor Is located on the tip of the probe. The tri-position switch allows for selection between an off position, or either the Fahrenheit or centigrade gain circuits. In this experiment the Fahrenheit gain circuit is selected in order to maintain unipolar operation into the conditioning cir cuits over the desired temperature range of 0°F-64°F, (Appendix E). Figure D-4 Photograph of the Temperature Probe 126 Calibration of the temperature transducer can be accomplished by the following one-point calibration procedure. First the probe is immersed in a cup of semi-melted snow (temperature 32.2°F or 0°C). Then the tri-position switch is used to select both the Fahrenheit and Centigrade gain scales and the output voltages are noted. These voltages represent the freezing point for each scale and the difference between the two readings represents 32.2°. Thus if the Fahrenheit scale is used for monitoring, the lower of the two voltage readings is 0°F and, similarly, if the centigrade scale is used the higher of the two voltage readings is +32.2°C. (This calibration assumes that both the °C and °F gain circuits have been accurately calibrated according to the instruction manual.) Care must be taken to ensure that the temperature probe is installed with an isolated DC power supply since the probe loses calibration when a non-isolated source is used due to a ground loop condition. Response Time The response time of the temperature transducer was measured to be 60 degrees Centrigrade per second. This was done by quickly immersing the temperature probe into a cup of cold water after being in a room temperature enviroment and observing the response on a storage oscilloscope. 127 APPENDIX E SIGNAL CONDITIONING UNITS E-1 Meteorological Signal Conditioning Units The -eteorological signal conditioning units provide the interface between the various weather transducer outputs and the + 0-5 volt range analog to digital convertor inputs. The circuit, as presented in Figure E-2, is op-amp derived. Adjustable potentiometers at the front edge of the printed cir cuit board (see photograph in Figure E-1) provide gain control adjustments for each of the conditioned signals. The input/output connections for the unit are shown in Figure E-3. Figure E-1 Top View Photograph of the Meteorological Signal Conditioning Unit. K> CO 129 G o G TIP 0 BKT G © CNT ® RAIN G TEMP ® TEMP o POWER G G G G G G SPARES ® ws © WD METEOROLOGICAL INTERFACE UNIT h-ACCESS HOLES TO GAIN CONTROL POTENTIOMETERS 0) FRONT VIEW RAIN 0 TEMP 0 WS 0 WD 0 0 0 0 0 RAIN CLEAR -12 V • 5 V + 12 V b) REAR VIEW Figure E-3 Front and Rear Views of the Meteorological Signal Conditioning Unit. 130 The set-up procedure is best completed in the laboratory before installa tion and is done through the adjustment of the gain control potentiometers. The procedure to calibrate each meteorological variables is as follows: i) Rain: Connect the counter and rain bucket leads to the front panel inputs using 'harwin' connectors and observe the output from the rear panel using an oscilloscope. Connect a pulse generator into the rain clear port on the back panel and set it for a several second interval between pulses. Now adjust the gain of the rain circuit until a bucket tip pulse latches high at the output, indicated by the green LED turning on. Next, observe to see if the pulse from the generator clears the latch. If not, ease off the gain until the latch is cleared. When both a tip of the bucket sets the latch and a pulse from the generator resets it, the rain circuit is ready for operation, ii) Temperature: Connect the output of the temperature probe into the front panel input of the meteorological conditioning unit using 'harwin' connectors. Ensure that the temperature probe is powered from an iso lated -9.0 volt DC power source so that the accuracy of the calibration is maintained. Now place the probe into an ice water mixture and adjust the gain until 32.2°F. on the Fahrenheit scale gives an output of 2.5 volts. The circuit is now ready for operation. This procedure provides a temperature dynamic range of 0°F (0 volts) to 64.4°F (5.0 volts). 131 iii) Wind Direction: Connect both the +5 volts azimuth exciter voltage and the wind direction input voltage to the input molex connector (See Figure E-2 for details), iv) Wind Speed: Connect the windspeed output of the anemometer to the input of the molex connector (See Figure E-2 for details). Now adjust the potentiometer for a gain of 1/2.. This is easily accomplished by using a dual trace oscilloscope where the first channel at 1 volt per division monitors the input voltage and the second channel at 0.5 volt per divi sion monitors the output voltage. At ground position align both traces. Now apply a fan to the anemometer propeller and adjust the gain so that the two traces remain superimposed. This procedure allows for wind speeds in excess of 100 kilometers per hour to be monitored and allows the factory calibration graph for output voltage versus wind speed to be read directly by simply dividing the "oltage scale in half. The Bear Mountain Conditioning Card The circuit schematic for the Bear Mountain signal conditioning card is presented in Figure E-4. As shown, the circuit is basically two parts, the first being a single op-amp differential amplifier to condition the tempera ture voltage and the second is a voltage follower, S-R latch to condition the bucket tip voltages. The total current consumption is less than 5 ma for this configuration making it well suited for this application. 132 GAIN OUT ^2 V = R, VALUES USED R, =22K R2 = IOOK a) TEMPERATURE CONDITIONING CIRCUIT b) RAIN CONDITIONING CIRCUIT Figure E-4 Circuit Schematic for the Bear Mountain Signal Conditioning Card 133 E-2 Receiver Signal Conditioning The receiver signal conditioning unit provides an optimum input range of + 0-5 volt signal to the analog to digital convertor from the 4 and 7 GHz receiver AGC outputs. The circuit schematic for one of the differential amp lifier gain blocks used on the receiver conditioning card is shown in Figure E-5. A differential gain block was used for each of the five AGC voltages monitored and all five gain blocks required were incorporated on one card which was mounted on the 6800 mother board. The same card was also used at Ryder Lake to condition the meteorological signals. Table E-1 provides the resistor values used to achieve the optimum gains for each receiver and meteorological signal monitored. Table E-1 The Resistor Values Used in the Differential Gain Block for Optimum Gain SIGNAL Rl»Ri+ R2,R5 R6'R7 R3 GAINQpT 3550 GHz 20K 5 OK 25K lOOKpot - .93 3790 GHz 20K 50K 25K lOOKpot - 1.26 4010 20K 50K 25K lOOKpot - 1.40 7142.0 10K 10K 50K 50Kpot -35.0 7496.2 10K 10K 50K 50Kpot -50.0 WS 10K 50K 10K lOOKpot + 0.5 WD 10K 50K 22K lOOKpot + 0.8 TEMP 50K 5 OK 50K lOOKpot + 5.0 RAIN 10K 50K 22K lOOKpot + 1.0 134 IF: R | = R 4 ; R2 = R5 i R6=R? Figure E-5 Circuit Schematic for One of the Gain Blocks for the Receiver Signal Conditioning Card 135 Set Up Procedure Each receiver conditioning circuit is set-up in the same manner. First insert a signal +10dB higher than the normal clear weather received signal levels into the input of each microwave receiver using a microwave signal generator. This provides a safe margin for overfades. At this level, tune the gain of the receiver conditioning unit to match the highest allowable input voltage to the analog to digital convertor. After doing this the receiver signal conditioning unit is now ready to be included as part of the calibration procedure to obtain the AGC curves in Appendix A. 136 APPENDIX F ANALOG TO DIGITAL CONVERTOR Two types of analog to digital (A/D) convertors have been used in this work. The first unit is a 16 channel A/D with 12 bits resolution developed for use with the 6800 system by the Communications Research Centre, Ottawa, and is currently being used in the Ryder Lake microprocessor (see Appendix I). The second unit is a low power consumption 16 channel A/D with 8 bits resolu tion which has been especially designed for this research, to be used in con junction with both the RCA 1802 and the INTEL 8085 microprocessors. The remainder of this appendix will deal with this second A/D convertor. Figure F-l provides photographs of this A/D convertor viewed by itself and as an installed unit. F-l Functional Description The second A/D convertor has the capability of addressing one of sixteen + 0-5 volt analog channels and converting the voltage appearing on the select ed channel to a latched hexi-decimal output with values ranging from 00 to FF. Although the 0816 A/D convertor module has the capability of 16 input chan nels, so far only four conditioned meteorological outputs from the condition ing unit have been connected. These A/D channels are assigned according to the following table: Figure F-l Photographs showing the A/D Convertor Separately and Installed. 138 TABLE F-l A/D Convertor Channel Assignment Table A/D Channel Meteorological Variable 3 Wind Direction 4 Wind Speed 5 Temperature 6 Rain 7 Not Defined When this A/D convertor is used with an 8085 system, control of the chan nel selection and variable conversion is done through the selection of appro priate software output instructions. Figure F-2 provides a schematic of the A/D convertor board with the con trol logic associated with specific output instructions identified. Figure F-3 provides the timing between each of the control signals. In the design of this A/D board special care has been taken in selecting components with low power consumption to allow direct application of this unit in battery powered sites such as Bear Mountain. F-2 A/D Convertor Calibration Procedure Calibration of the A/D convertor is accomplished by sweeping the 0-5 V range of a selected A/D channel with a known voltage and monitoring the value of the hexidecimal output. This procedure is simplified if the test program in Figure F-5 is used. It allows the A/D's hexidecimal output to be viewed directly on an oscilloscope by monitoring the serial output pin (SOD) of the 8085 microprocessor. Figure F-4 provides a sample oscilloscope trace of SOD using the calibration test program. RAIR cum A/D BOARD SCHEMATIC ICI •Dean PHYSICAL LAYOUT /I 40M \i CIRCUIT SCHEMATIC a PHYSICAL LAYOUT FOR THE WEATHERLOG ANALOGUE TO DIGITAL CONVERTOR FIGURE F-2 MAR. 12.1962 140 OUT 01. 0UT"4F OUT 2F START CONVERSATION (SO ADDRESS LATCH ENABLE (ALE) END OF CONVERSATION (EOC) JI I I i I I TYPICALLY lOQx/sec Figure F-3 Control Signal Timing Diagram for the A/D Converter RS232 OUTPUT (SOD) •12V -12V OUTPUT = 29 HEX 0 0 1 0 1 0 0 1 7 MSB 6 5 4 3 2 1 RAIN CLEAR OUTPUT •5V 0 LSB PULSES PRODUCED IN SOFTWARE BY THE "OUT IF" INSTRUCTION Figure F-4 Sample Oscilloscope Traces of the A/D Output During Calibration 141 Tektronix 8000/8085 ASM V3. 1 P«9C 00001 00002 -00005 00004 00005 oooo: i t THIS PROGRAM READS ONE CHANNEL OF THC A/D AND OUTPUTS THE I RESULT IN C0NTIN0U3 LOOP FASHION SO THAT THE DATA CAM DE READ FROM Or : I LLC A RAIN CLEAR PULCE—IT I '".'CD 00007 00003 -©000?-00010 00011 •O0012• I EACH PIT AND A DOUBLE PULSE IS USED TO INDICATE THE START OF THE I DATA BYTE. ON THE SCOPE THE DATA APPEARS WITH THE LEAST SICVIIFIO* -: : BIT—AS—THE-THE FIRST-BJT--AT-THE-LEFT SIDE-OF-THE DATA BYTE. - -00013 00014 -OO015. 00016 00017 •ocuais oooo 0100 0102 01 fi-1 oooo C30001 ORG OOOH JMP MAIN -ORG—lOOH-D300 MAIN MVI A.31H OUT 00 MVI A.03H 00019 OOOl'O -0002-SIM MVI H.3FH SPHL DATA CHANNEL ON A/D IS SELECTED 00040 00041 00012 IN 02 IN 02 00043 00044 00045-0004 & 00047 000.1;: DATA BYTE IS SENT OUT IN A CONTINOUS LOOP FASHION 012A 01 7'T OEOS ?#####**##***#*#*#*#**«•#*#####**#»****#****#** CC2 MVI COSH 0004? 00050 —00051-00052 00053 C12E 0130 0131 0132 0134 D31F OF 47 Fi.^O E6C0 CUT IFH CC1 RRC MOV..D*-A— ORI 40H ANI OCOH 00054 00055 -00056 O0057 0005S 0136 0133 0139 013A 013B OOOS'-' 010C D31F 30 78 OD C23001 OUT IFH SIM DCR C JN2 CC1 «.'MP CC2 00060 00061 -OO062 00063 00064 000^5 0150 0151 0152 0154 Oir-5 0150 00 •C9 1E50 ID 025101 DELAY ORG 150H NOP -SET DL1 CC? MVI E DCR E JNZ ceo 50H 00066 00067 0153 C? RET Figure F-5 A/D Calibration Program 142 APPENDIX G DIGITAL TO ANALOG CONVERTOR The digital to analog (D/A) convertor has been used in the first phase of this experiment to generate chart recordings of the variables sampled by the data acquisition system and to verify the accuracy of the results processed by the data base management system (Chapter V). The circuit schematic for the D/A unit and its interface to the UBC 6800 microprocessor is presented in Figure G-l. In the first phase of this experiment the D/A was used in conjunction with the cassette recorder. This was done by first using the recorder to store the time series data as it arrived at UBC from one selected RS 232 data channel. The data thus taken was then read back from the cassette recorder into the UBC 6800 microprocessor which selected a particular variable from the time series data to be outputted through the D/A. This selection process is easily accomplished since each variable arriving on the RS 232 port is assigned a specific position with respect to the sampling-interval synchro nization byte. The program for this selection is given in Figure G-2. TYPE RS232 CONNECTOR TO D/A PORT ON MICRO PROCESSOR MSB Bl 7 Bz 6 6 B3 7 5 B4 8 4 Bs 9 3 B 6 10 2 B7 11 | LSB B8 12 9 _ 1 r DAC-08 3 16 13 Aj>F r X H .01/JF Eo = IFS"RL TO CHART RECORDER H r -IOV. +IOV. Figure G-l Circuit Schematic for the D/A Converter THIS PRnpftAM IS USED TO SF|_ FCT TTMF—SFRTFS ! AS IT ARRIVES AT UBC FROM THE DATA LINK j A VARIABLE RELATIVE TO A SYNCHRONIZATION i BYTE ISL.SELEC.TEEI_FCR_ OUTPUT. TO.. ! AND THE APPROPIATE VARIABLE IS CHOSEN BY '. EDITING THE COMMENT STATEMENTS j ORG OEOOOH I VARPTR EQU OOOl ?VARPTR EQU 0002 i 1 ?VARPTR EQU 0003 VARPTR ?VARPT EQU ^ EQU 0004 0005 7 ACTIVATED FOR RAIN OR RX#4 STK'PTR EQU 0AO50H PIACRA EQU 0801 IK • • PIADRA EQU OSO1 OH : PIA REGISTERS CTR EQU CLR 0A040H PIACRA LDA A #OFFK CLEAR PIA CONTOl RESIST ER STA A PIADRA SET PIA AS OUTPUT LDA A #04H ; NON INTERRUPT PORT STA A PIACRA LDA A #03H 7 RESET ACIA STA A OSO1 EH LDA A #015H STA A OSO1 EH ; COMMAND REGISTER SET UP LDS #STKPTR 5 STACK POINTER LOOP LDA A OSO1 EH LOAD STATUS FROM PORT 7 LSR A BCS DTOA RDRF YES GET BYTE •JMP LOOP DTOA LDA B 0S01FH 7 LOAD CMP B ttOFFH ; IS IT SYNCH BEQ ZERO YES- RESET COUNTER ! LDA A CTR CMP A #VARPTR IS THIS THE VARIABLE BEQ OUTPUT YES: SEND OUT D/A I NC CTR •JMP LOOP NO: INPUT NEXT VARIABLE OUTPUT STA JL PIADRA 5 SEND OUT D/A I NC CTR JMP LOOP 7FRi"! i":.' R CTR RESET COUNTFR I NC CTR JMP LOOP LOOK FOR NEXT BYTE „ „. END _ Figure G-2 Listing of the Program to Provide Chart Recordings from the Receiver Data Using the D/A 145 APPENDIX H MODEM UNITS H-l Modem Transmit Units A modem transmit unit is included in each of the Dog Mountain, Ruby Creek and Agassiz Experimental Farm Weatherlog Microprocessors. Each unit takes a ±12 volt RS 232 compatible 110 bps serial data stream and encodes this to output a voice frequency FSK modulated signal. The modulated signal is then carried by way of either a VHF or a telephone communications channel to the modem receivers at the Ryder Lake Site. The receive modems are described in section H-2. Figure H-l Photograph of an Intalled Modem Transmit Unit 146 The centre frequencies for the modems that were available in this project are given in Table H-l as follows: Table H-l Modem Centre Frequency Assignments Frequency (Hz) Code Links Used 480 DA 720 B Dog Mountain 960 DC Agassiz Experimental Farm^Bear Mountain 1200 D Ruby Creek 1440 DE Figure H-l shows a picture of a typical modem transmit unit as installed in a Weatherlog Microprocessor and Figure H-2 gives the interface schematic between the modem transmit unit and the microprocessor. H-2 Modem Receive Units All four modem receiver units are located at the Ryder Lake Site (see Appendix C) and have been integrated into one 19" rack mounted unit known as the "UBC Modem Unit". Each modem receiver inputs a voice frequency FSK modu lated signal and decodes it into a 110 bps RS 232 compatible output to dupli cate the input of the transmit unit (see Section H-l). The RS 232 output from each RS 232 unit is then relayed onto UBC for storage via separate channel on the statistical multiplex. The physical layout of the "UBC Receiver Modem Unit" depicting front, back and top views is shown In Figure H-3 and a schematic of a typical inter face between a modem receiver unit and the RS 232 output connections is given in Figure H-4. 147 MODEM TRANSMIT UNIT AO B e o 0* E e F • HO i J o KO LO MO NO + 12 - 12 INDEX VF OUT GND. RT S TRANSMITTED DATA HALF DUPLEX HUE CTS DSR INTERLOCK LOCAL COPY 12 PIN MOLEX CONNECTOR (RED) (GRAY) (ORANGE) (BLACK) (WHITE) (YELLOW) (BLUE) (BROWN) (GREEN) © (2) 0 © © (2) 0 © © 0 0 © © © © 0 © © © 0 0 © + 12 -12 POWER SUPPLY ,BNC 1 N.C. " SERIAL DATA IN N.C. + 12 • +12 N.C. N.C. TERMINAL -±? STRIP Figure H-2 Interface Schematic for a Weather Log Modem Transmit Unit in ui PORTS DOB MOUNTAIN RECEIVER MOOCH CARO RUBY CRCCK RECEIVER MCOtM CARO TERMINAL STRIP S 0 9 9 ® 0 o e s a e s S 0 0 s e 0 0 e •CAR MOUNTAIN MOOCH RECEIVER CARO DC POWER SUPPLY AGASSI! EXPERIMENTAL FARM RECEIVER MODEM CARD DUMMY RESET SWITCH POWER LIGHT SWITCH TOP VIEW SHOWING INTERIOR LAYOUT OF THE MODEM UNIT |~) UBC MODEM UNIT ^ © RESET O FRONT VIEW VF INPUTS r © © © © AGASSIl BEAR MTN. RUBY CK. 000 MTN J RS 232 OUTPUTS REAR VIEW PHYSICAL DRAWING OF THE RYDER LAKE .i RECEIVER UNIT SHOWING TOP, FRONT 8 REAR VIEWS FIGURE H-S MAR 12.1912 CO 149 Y MODEM RECEIVER UNIT O o o o o o o o o o o o + 12V. -I2V VF INPUT GND NC RCV LOOP (-I2V1I RECEIVED DATA RCV LINE SIG DET NC REMOTE RCV DISABLE LOCAL COPY INPUT NC NC NC "\ TO SPOWER -) SUPPLY BNC < 20 GND' GND RTS CTS DTR Tx FEMALE RS232 CONNECTOR Figure H-4 Interface Schematic for One of the Modem Receiver Units 150 APPENDIX I MICROPROCESSOR UNITS 1-1 The UBC Microprocessors i) General Description The two UBC microprocessors are both Motorola 6800 based units and are used for the purpose of time correlating the data as it arrives at UBC from the sampling processors along the path. The data formatter unit (uP#l) col lects the incoming data, constructs the time series queues and outputs the results each second to the data processor unit (yP#2). The data processor unit then reduces the data to manageable levels by constructing hourly distri butions and choosing selected time series intervals for storage onto a 4.5 megabyte capacity cassette recorder. The software routines to do this are discussed in Appendix J. Two methods have been employed to prevue and test the incoming data stream. The method currently used takes the time series data outputted from the data formatter (uP#l) and resends it using the data processor (uP//2) to a video monitor. A photograph of the video display and discussion of this method is given in Chapter IV. A second method was employed in the first phases of the system integration by reading selected variables from the arriv ing data streams and outputting these through a digital-to-analog convertor, as described in Appendix G. The physical layouts for the UBC microprocessors are shown in Figure 1-1 and 1-2 for the data formatter and data processor respectively. These draw ings show the rear termination assignments, the front panel display and the WK tn dew tn at »-< a *t< CD O >" K tL>- Out DO U 3 KO OCO I LI I I o I v 1/0 SLOTS MPU ( WITH EPROM) RE0ULAT0R jBOARD DISPLAY BOARD S800 MOTHER BOARD TOP VIEW POWER SWITCH IIOVA PCV~ RUBY AGASSIZ SPARE ' • • CD O 2A SA OO FUSES .(=• CZD 1—II—1 OUT DOG BEAR RYD. SWT. BUG REAR VIEW • UT MA. RCV. o o o o o o g FRONT VIEW UBC DATA LOG © PROSRAM RE a 3 PHYSICAL DRAWING OF THE UBC DATA FORMATTER (uP N*2) MICROPROCESSORj SHOWING FRONT,TOP AND REAR VIEWS FIGURE l-l MARCH It.lSSZ (—1 Ln h-1 < a: - £ < F 1° _• .> KQ. Q a. Til K m ta. O O o o O ID <D CD CO O t «D O O o Q O o CD « CO CD \ v—— I/O SLOTS MPU(WITH EPROM) TOP VIEW •rt'mt- na OC. POWER CONNECTOR FUSES TT] +} +12 -12 ii] ooo PROCESSOR DIGITAL I 1 'i^'A- VIDEO TERMINAL PORTS | ^ TAPE UNIT REAR VIEW UBC NI ! OATA PROCESSOR RYDER RESET RESET P 3 <= FRONT VIEW PHYSICAL DRAWING OF THE UBC OATA PROCESSINC tyiP N« 2 ) MICROPROCESSOR SHOWING FRONT.TOP ANO REAR VIEWS -FIGURE 1-2 ISSUE I MARCH 12,1962 a) The Data Formatter (uP#l) Figure 1-3 Photographs Showing the Interior Layout of the Two UBC Microprocessors 154 processor mother board layouts. A photograph of both microprocessors is given in Figure 1-3 illustrating their interior layouts. An integral part of the UBC unit are the ACIA cards which provide the communications interface between the serial RS 232 input/output ports and the parallel data bus of the microprocessors. Each ACIA card provides an inter face to two RS 232 ports according to the assignments shown in Table 1-1 and 1-2 for uP 1 and 2. Its circuit schematic and physical layout are given in Figure 1-4. Table 1-1 I/O Port Address Assignments for the Data Formatter Unit yP//l I/O PORT it PORT ADDRESS FUNCTION ASSIGNMENT 0 8000 8001 8002 8003 Not assigned 8004 PIA I/O Reg. 8005 DDR 8006 PIA I/O Reg. 8007 DDR Terminal Port PIA 8008 8009 800A 800B Not assigned 800C 800D 800E 800F 8010 8011 8012 8013 ACIA CMD ACIA I/O Reg. ACIA CMD ACIA I/O Reg. Not assigned Spare 1/0 Port Ryder Lake Meteorological Table 1-1 cont'd. 8014 ACIA CMD Reg. 8015 ACIA I/O Reg. 8016 ACIA CMD Reg. 8017 ACIA I/O Reg. 8018 ACIA CMD Reg. 8019 ACIA I/O Reg. 801A ACIA CMD Reg. 80IB ACIA I/O Reg. 80IC ACIA CMD Reg. 801D ACIA I/O Reg. 80IE ACIA CMD Reg. 80IF ACIA I/O Reg. Agassiz Experimental Farm Bear Mountain Ruby Creek Dog Mountain Tape Unit RS232 Output Port Receiver Channel from Ryder Lake Table 1-2 I/O Port Address Assignments for the Data Processor Unit uP#2 PORT ADDRESS FUNCTION 8018 ACIA CMD Reg. 8019 ACIA I/O Reg. 801A ACIA CMD Reg. 80IB ACIA I/O Reg. 80IC ACIA CMD Reg. 801D ACIA I/O Reg. 801E ACIA CMD Reg. 80IF ACIA I/O Reg. ASSIGNMENT Output Port to Video Screen and Tape Unit Receive from uP#2 Future Digital I/O Ports BAUD RATE SELECTION JUMPERS OUTPUT CONNECTORS «3 NCRttBXaCTSr TXtTXiRTS GND MOLEX CONNECTOR TO 6800 MOTHER BOARD J tt ii i4 it >*• \fi ii io 9 ir -mrt. | 4CI.4 USO-J , IJ H •*-tJa'-sSSjl lis DATA BUS ti *CM8» <TX) jTtrti j j h 1 rn. * s 5 S U 5 5 i IH 13 H is /* ir im HUHIIDM »sv REG. DATA BUS I/O SELECT RESET !' IIO 300 600 1200 9600 »8V R/W DT 06 D5 04 P' D2 DI DO RSI RS2 IRO NMI INDEX GND tl2V -I2V NC NC INTERFACE CONNECTOR TO 6800 MOTHER BOARD CIRCUIT SCHEMATIC ASYNCHRONOUS INTERFACE CARD 7BI54 68S0-2 I I •TO SELECT OPTIONS PINS MUST BE JUMPERED |_ CHAN2,II0BANDJ CHI, IIO t_CHAN2'300 - CHI. 300 :HAN 2,600 . - CHI, 600 j—CHAN 2,1200 - CHI, 1200 IH AN 2,9600 -CHI, 8600 I NC —CTS: EXT . CONTROL CTS". ALWAYS v ENABLED \ OPTION SELECTION CONNECTOR • 8V NC CHI RX CH2RX CTS NC CHI TX CH2TX RTS GND MOLEX(F) >OUTPUT CONNECTOR TORS 232 PORTS I/O CONNECTIONS PHYSICAL LAYOUTS ASYNCHRONOUS INTERFACE CARD CIRCUIT SCHEMATIC AND PHYSICAL LAYOUT OF THE ASYNCHRONOUS INTERFACE (ACIA) CARD FIGURE 1-4 ISSUE I MAR 12 ,19B2 Ul 157 Specifications a) MPU board • modified to produce data rates of 110, 300, 600, 1200 and 9600 bps. (uP//l). • unmodified in uP#2 to give data rates of 110, 300, 600, 1200 and 2400 • if the SWTBUG monitor is to be used switch off both "PROM" dip switches and switch on the "SWT" and "MONITOR" dip switches. • to use a program in the EPROM, switch on the "LO PROM" or the "HI PROM" dip switch and turn off the "SWT" and "MONITOR" dip switches. The EPROM program then starts in location E0001& of the memory space with the interrupt vectors being assigned the following addresses. Table 1-3 6800 Interrupt Vectors Location Type of Vector E7F8 E7FA E7FC E7FE Interrupt Request (IRQ) Non Maskable Interrupt (NMI) Software Interrupt (SWI) Restart b) ACTA board • The boards must be strapped according to Figure 1-4 to determine their data rates. 158 c) Regulator board Input Voltage: ±24V Output Voltages: +12V -12V +14.6V -14.6V UD1 (uP//l only) UD2 (uP//l only) d) Total AC power required 500ma at 110V. A.C. 1-2 The Ryder Lake Microprocessor Unit The Ryder Lake Microprocessor unit is based on the Motorola 6800. Its purpose is to sample both the received signal levels and the meteorological information from this site and send it to UBC via two separate data channels. Every 1/10 of a second the received signal levels are sampled and on every tenth sampling interval (i.e. each second) the weather information is also sampled and sent. The analog-to-digital convertor in this unit is able to resolve to 12 bits which is more than required. Therefore, each sample is first reduced by stripping the four least significant bits before the 8 bit samples are output-ted. The channel assignments for the analog-to-digital convertor are provided in Table 1-4, as follows: 159 Table 1-4 Analog to Digital (A/D) Convertor Channel Assignments A/D Channel Channel Assignment F Unassigned E Unassigned D Rain C Temperature B Wind Speed A Unassigned 9 Wind Direction 8 Unassigned 7 Unassigned 6 Unassigned 5 Unassigned 4 7496 GHz Rx 3 7142 GHz Rx 2 4010 GHz Rx 1 3790 GHz Rx 0 3550 GHz Rx • Figure 1-5 provides a drawing showing the hardware assignment slots for the CPU, the ACIA card and the analog-to-digital convertor. Table 1-5 pro vides the I/O port address assignments for the Ryder Lake Unit, as follows: 5 J! 5°si « s 9 S K o o OF t CO u. 5 m o « 1 o m | § i s 1 o CD o CO 1 o 1 « o i u Q 1 o 1 » 1 CO 1 u « o « 5 • s • 8 5 CD A/D BOARD |8t;»-8823) METEOROLOGICAL CONDITIONING CARD RECEIVE SIGNAL CONDITIONING CARD MPU BOARD (WITH E PROM I TOP VIEW WEATNCR CHANNELS RECEIVER CHANNELS G«O 8 & bM r ^ i t i C o o o o vie LH SPEED *KM TL4« Q-I2V. © PORT OC POWER OUTPUTS REAR VIEW UBC DATA LOO • RESIT O c. MO 1200 o o o na MET (CV RYDER LAKE . ft POWER m FRONT VIEW PHYSICAL 0RAWING OF THE RYDER LAKE MICROPROCES SOR UNIT SHOWING FRONT, TOP AND REAR VIEWS FIGURE X- S (MARCH 12 ,1982 o 161 Table 1-5 I/O Port Address Assignments for the Ryder Lake Unit I/O PORT # PORT ADDRESS FUNCTION ASSIGNMENT 8008 ACIA CMD Reg. 8009 ACIA I/O Reg. 80OA ACIA CMD Reg. 800B ACIA I/O Reg. High speed Receiver Data Low speed Meteorological Data 8018 to 801B DC Power Supply Interface All other I/O port addresses are unassigned. 1-3 The Weatherlog Microprocessor Units The weatherlog microprocessor units used at Dog Mountain, Ruby Creek and the Agassiz Experimental Farm are based on the INTEL 8085. Their purpose is to sample the meteorological variables of wind speed, wind direction, tempera ture and rain at one second intervals. The inputs to the microprocessor's A/D channels are 0-5 volts which are provided at the output of the meteorological signal conditioning unit (see Appendix E). The weatherlog microprocessor incorporates a low speed frequency shift keyed modem between the RS 232 output of the 8085 internal processor to the voice frequency 600 ohm unbalanced out put of the weatherlog. The FSK output thus derived is then applied to the corresponding channel back to Ryder Lake. From the Ruby Creek location this is via VHF radio, from Dog Mountain this is via the 37A message circuit on the A/D CONVERTOR j, h i MOOCH TRANSMIT CARD [A/0 TO AO»a INTERFACE | MOOIM INTERFACE | POWER INTERFACE r-i-ri •taHnninnnniBiiii_ XNTIL SIC RO/M •ORt COMPUTIR •OARD TOP VIEW MAIN - TERMINAL STRIP ©•'« WIND SPEED © WIND DIRECTION o IIO VAC ©•. O-.t VOLTAOC OUTPUTS o MOOIM OUTPUT © RAIN o RAIN CLEAR o TEMPERATURE WPUSI REAR VIEW L UBC WEATHER LOO © RESET DOG MOUNTAIN 3 5 9ITE NAME' FRONT VIEW PHYSICAL DRAWING OF THE WEATHER LOG MICRO PROCESSOR SHOWING FRONT, TOP AND REAR VIEWS FIGURE 1-6 MARCH IS , 1981 163 7 GHz radio and from the Agassiz Experimental Farm this is via a telephone link. A drawing indicating the positioning within the weatherlog unit of the 8085 computer board, the A/D convertor and the modem transmit card is given in Figure 1-6. A photograph of this interior layout is shown in Figure 1-7. Figure 1-7 Photograph showing the Interior Layout of a Typical UBC Weatherlog Microprocessor Unit. 1-4 The Bear Mountain Microprocessor Unit The purpose of the Bear Mountain Microprocessor is to sample the meteoro logical variables of temperature and rain once per second and output the data onto the radio link to Ryder Lake. 164 In designing the microprocessor unit for this site, special care was taken to minimize the power consumption and to choose a technology which could withstand large temperature variations (±40°C). CMOS technology was selected for the microprocessor, the A/D and the signal conditioning unit. The micro processor is an RCA 1802 and the circuits for the A/D and signal conditioning units are described in Appendices F and E respectively. The total power con sumption of the CMOS electronics was 70 ma. The radio used is a hand-held unit from Motorola specially designed for low power consumption applications. Optimization of the radio's power con sumption was achieved through the use of a high gain transmitting antenna and eliminating the power consumption of the radio's final transmitter stage, thus reducing the RF output power from 1.5 watts to 300 mwatts and the radio's current draw to 225 ma. The expected operational life-time for the Bear Mountain site based on 2000 A-HR caustic potash cells is therefore 1.01 years for the radio battery and 3.23 years for the electronics battery. It is recommended that battery replacement is done by helicopter since the road access is rugged (four wheel drive only) which could damage the cells. A drawing of the equipment configuration showing the microprocessor, the radio, the A/D and the signal conditioning unit is given in Figure 1-8 and a photograph in 1-9. £91 166 Figure 1-9 Photograph showing the Interior Layout of the Bear Mountain Microprocessor Unit 1-5 The NOVA 840 Minicomputer The purpose of the NOVA 840 minicomputer is to take the data recorded on the cassette tapes and then transfer these data in serial fashion to IBM for matted 800 bpi magnetic tape. Figure 1-10 illustrates the equipment arrange ment needed to transfer the data. In order to enter data on the NOVA, an RS 232 to current loop interface was constructed as the NOVA had no RS 232 input ports. The RS 232 to current loop interface schematic is given in Figure I—11. 167 IO0 4.5 MEGA BYTE CASSETTE TAPE COLUMBIA RS232 DATA 300 D TO CASSETTE 1 1 CURRENT RS232 LOOP TAPE UNIT INTERFACE | 1 NOVA 1CURFcENT 840 LOOP 800 bpi MAGNETIC TAPE Figure 1-10 Equipment Configuration to Transfer Data from the Cassette Tape Drive to the Nova 840 Magnetic Tape Drive. 6 168 ro /vow 64o TO ojsssrrr *?S 232 l£l/£LS Figure 1-11 Circuit Schematic of the RS-232C to Current Loop Interface 169 APPENDIX J MICROPROCESSOR SOFTWARE J-l The UBC Microprocessors i) Functional Description The main function of the UBC 6800 microprocessor software in the data formatter (uP//l) and the data processor (yP#2) is to time-correlate all the field data as these arrives on the data link, output the data for real time viewing on a video monitor, compress it into statistical distributions, select time series segments for storage and finally store the data on magnetic tape using suitable formats. Due to timing constraints and the complexity of in corporating all these tasks under the control of a single CPU, these tasks were distributed between two processors. The first processor, the data for matter or uP//l, has been assigned the tasks of inputting the arriving field data, coding this data to real time, formatting it into buffers and outputting the formatted data once per second to the second CPU. The second processor, the data processor or uP//2 has been assigned the tasks of outputting the data for real time viewing on a video monitor, generating statistical distributions from the data, choosing time series segments and storing these records onto a magnetic cassette recorder in a recoverable format. The interfaces between both UBC microprocessors, from the statistical multiplex and uP#l, from both microprocessors to the video terminal and from uP#2 to the cassette recorder are all RS232C. The input/output ports are all configured in the software. 170 The program flow chart for the data formatter (yP#l) is given in Fig. J-la, a diagram of its buffer memory organization is shown in Figure J-lb and the program flow chart for the data processor (yP#2) is presented in Fig. J-2. ii) Input/Output Software Design Specifications I. Received Signal Data The signal amplitude data arrives at the input of the data formatter (uP//].) in rotational order at one-tenth of a second intervals starting from the lowest frequency receiver. A synchronizing byte (FF) is received after the highest frequency receiver sample arrives and is used to update the software clock on the basis of 10 incoming synch bytes equalling 1 second. After each second the received signal data is dumped to the data processor (yP#2). The data processor checks the incoming levels for multipath and fading by determining if they exceed a specified range. If this is the case, a time series dump is initiated. The data processor finishes by taking each byte of receiver signal data to update an hourly distribution buffer, to store it in a rotating time series queue and to send it for viewing to a video monitor. II. Meteorological Data Each second,.four successive bytes arrive on the five input meteoro logical channels, and immediately following their arrival a synchronizing byte is sent. Their order of arrival is wind direction, wind speed, temperature and rainfall. As each of these samples is received, their position in the data formatter unit's (uP#l) input buffers are updated with the new values and before they are outputted to the data processor |» ACIA* • BUFFER PO i NT E ft [• CL C BUFFERS |« BIT FIABB 10 99 POLL IK»UT AMD OUTPL/T POUT a «C POL BHD POL 008 POL HUB POL mc POL IND POL • CII POL ASA POL MC POL • MO POL WTO POL -*® —0 -K£> -KiD -*® -MD -KD -KD -KD -KD _-KD ^ PCVW ) SCT DATA*« I 1 BND POL ^ BEl SVNCH FLAB •CT BFLAB BIT BUFPtn ADDPEtS EQUAL TO DUMP AOON, fOP n»o »/W PROGRAM FLOW CHART FOR uP N« I,THE ue: DATA FORMATTER MICRO PROCESSOR UNIT FIGURE J-lo MARCH ti , JeJ| BUFFER Ns-I BUFFER Ngl 2 BUFFER 3 Figure J-lb Data Formatter (yP#l) Buffer Memory Organization INPUT DATA _rROM_«_P Nt" I OUTPUT DATA TO VIDEO DISPLAY BESET TIME __SEHIES FLAO SET TIME SERIES DUMP FLAG BEGIN T1 ME SERIES DUMP . 1 r RESET TtME SERIES FLAO SET DISTRIBUTION DUMP FLAft START OlSTRlBUTToiT DUMP . 1 r UPDATE ACIA 2-« T S_ I GET MINUTES «T LAST COUNT •12 RESET EVENT FLAG ™SET TIME SERIES WAIT FLAO COPT TIME SERIES QUEUE r I SIT HOUR FUt ^ MAIN ) PROGRAM FLOW CHART FOR uP N*2 THE UBC DATA PROCESSOR MICRO PROCESSOR UNIT FIGURE J-2 MARCH 121ISS2 174 (yP//2). The data processor then takes each byte of meteorological data updating its hourly distribution buffer, storing it in a rotating time series queue and outputting it for viewing to a video monitor. III. Headers and Trailers Every second the time series queue block is sent by the data formatter (uP#l) after being terminated by two "FF" trailer markers. Immediately following this a header for a new time series queue block is initialized, comprise^ of data bytes to identify record type, month, day, hour, minute and second. IV. Criteria to Dump the Distribution Buffer Every hour the buffer containing the distribution data in the data processor (uP//2) is dumped to the magnetic cassette recorder unless a  time series dump is in progress. For this special case, the dump of the distribution buffer waits until the time series queue dump is completed before transferring the distribution data. The distribution buffer is comprised of five types of data (wind direction, wind speed, temperature, rainfall and receiver amplitude) with five single byte distributions for each meteorological variable type and five double byte distributions for each receiver amplitude. Each meteorological distribution has 4016 = 641Q double byte bins and each receiver amplitude distribution has 128 double byte bins to give the following total memory required for one distribution buffer: 175 5 x [7 + 4 x 128 + 256 + 3] = 3890 bytes This includes a header of 7 bytes and a trailer of 3 bytes for each variable type. For two distribution buffers twice this number of bytes is required to give a total processor buffer requirement of 7780 bytes. Therefore, at 9600 bps the hourly dump will take 3.2 seconds. V. Criteria to Dump the Time Series Data Whenever an event is detected the time series queue is dumped from the buffer in the data processor (uP//2) to the cassette recorder unless a distribution dump is in progress. In this case the time queue is trans ferred into an output buffer to await completion of the distribution dump. The time series queue (TSQ) is comprised of: 12 x [7 header + 5x4 meteorological + 10x5 receiver + , 1 event + 2 trailer] = 960 bytes At 9600 bps this will take a minimum of 1.0 seconds to output for storage. The receiver event flag is tested in the MAIN program. If it is found to be set, the TSQ dump is initialized and the TSQ is copied to an output buffer to first allow completion of other I/O in progress. When a propagation event has been detected, the time series dump flag over the next 12 seconds is disabled so that a dump is not being requested every tenth of a second during the duration of the event. Another consideration is that a flagged event will not, in general, occur conveniently at the end of an integral second which means when making, the TSQ copy to the output buffer typically only 11 complete seconds of data would be transferred. The twelfth second block of data 176 would normally be incomplete and would consist on ,the average of five one-tenth of a second receiver sample sets and an incomplete meteorologi cal set. The buffer must therefore be designed to hold 13 seconds to, ensure a twelve second record. The resulting TSQ is given in Figure J-3, as follows: TSQ POINTER Figure J-3 Diagram Showing the Structure of the Time Series Queue (TSQ) The TSQ pointer is used to find the first block for dumping and the next block for filling. When the event flag is set, the TSQ starting from the TSQ pointer is dumped. The dump pointer cycles around from the 177 high address to the low address and then upwards until the most recently completed block is encountered. VI. Storage Requirements The question arises as to how much data will be generated for stor age. There are basically two types of data series; the first being dis tribution data recorded at 3890 bytes per hour and the second being TSQ data with a maximum rate of 288,000 bytes per hour but more typically 14,400 bytes per hour at an anticipated event probability of 0.05. Thus the total number of bytes expected to be recorded hourly is 18,290 bytes. At this rate the 4.5 mega byte magnetic cassette cartridges can be expected to last an average of 246 hours, or 10.25 days. This is expect ed to vary from 15.4 hours where all TSQ data are recorded to 48.2 days when no TSQ data are recorded, iii) Software Clock Real time is maintained through a software clock that is incremented upon the arrival of each synch byte and is kept on the basis of a 10 synch bytes equalling one second. The time is stored in RAM locations, 200-204 of the data formatter with the initial time being entered through the "SWTBUG" moni tor (see Appendix 1-1). Table J-l gives the RAM time assignments which must be entered in hexedecimal format when initiallized. 178 Table J-l RAM Time Assignments (pP#l) Location Assignment Month 0200 0201 Day Hour 0202 0203 Minute 0204 Second iv) Data Formats There are two basic data formats outputted by the data processor (yP//2); a time series format and a distribution series format. The time series format is given in Table J-2 and is repeated for each one second record dumped. A minimum of 12 one second records are dumped preceding an event with new time series records being dumped as long as the event persists. The distribution series format, on the other hand, is given in Table J-3 and is dumped hourly even if no propagation events occur. 179 Table J-2 Time Series Block Format Byte Description Specific Comments 1 Record Type FE = Time Series Block 2 Months BCD 3 Days BC4 Hours BCD 5 Minutes BC6 Seconds BCD 7 Gauge Status HEX 8 Wind Direction Dog Mountain 9 Wind Speed Dog Mountain A Temperature Dog Mountain B Rainfall Dog Mountain C Wind Direction Ruby Creek D Wind Speed Ruby Creek E Te-perature Ruby Creek F Rainfall Ruby Creek 10 Wind Direction 11 Wind Speed 12 Temperature 13 Rainfall 14 Wind Direction 15 Wind Speed 16 Temperature 17 Rainfall 18 Wind Direction 19 Wind Speed IA Temperature IB Rainfall IC Receiver 1 ID Receiver 2 IE Receiver 3 IF Receiver 4 20 Receiver 5 Bear Mountain Bear Mountain Bear Mountain Bear Mountain Agassiz Experimental Farm Agassiz Experimental Farm Agassiz Experimental Farm Agassiz Experimental Farm Ryder Lake Ryder Lake Ryder Lake Ryder Lake 3550 MHz First 100 msec. 3790 MHz First 100 msec. 4010 MHz First 100 msec. 7142 MHz First 100 msec. 7496.5 MHz First 100 msec. Table J-2 Time Series Block Format (cont'd.) Byte Description Specific Comments 21-25 Second 100 msec. 26-2H Third 100 msec. 2B-2F Fourth 100 msec. 30-34 Fifth 100 msec. 35-39 Sixth 100 msec. 3A-3E Seventh 100 msec. 3F-43 Eighth 100 msec. 44-48 Ninth 100 msec. 49-4D Tenth 100 msec. 4E EVENT INDICATOR cc = Fade 99 = Rain 4F-50 EOB TRAILER Two consecutive FF's Table J-3 Data Format for the Distribution Buffer Byte Description Specific Comments 0 Record Type (FD) FD = Distribution Series 1 Month BCD 2 Day3 Hour BCD 4 Minute5 Second BCD 6 Data Type (01) Data Type Control Word 7-86 Dog Mountain Wind Direction Distribution 87-106 Ruby Creek 107-186 Bear Mountain 187-206 Agassiz Exp. Farm Wind Direction Distribution 207-286 Ryder Lake 287 FF Trailer 288 FF 289 FF Table J-3 Data Format for the Distribution Buffer (cont'd.) Byte Description Specific Comments 28A Record Type (FD) Header 28B Month 28C Day 28D Hour 28E Minute 28F Seconds Header 290 Data Type (02) 291-310 Dog Mountain Wind Speed Distribution 311-390 Ruby Creek 391-410 Bear Mountain 411-490 Agassiz Exp. Farm Wind Speed Distributions 491-510 Ryder Lake 511 FF Trailer 512 FF 513 FF 514 Record Type (FD) Header 515 Month 516 Day 517 Hour 518 Minute 519 Second Header 51A Data Type (04) 51B-59A Dog Mountain Temperature Distribution 59B-51A Ruby Creek 51B-69A Bear Mountain 69B-71A Agassiz Exp. Farm Temperature Distribution 71B-79A Ryder Lake 79B FF Trailer 79C FF 79D FF 79E Record Type (FD) Header 79F Month 780 Day 781 Hour 782 Minute 182 Table J-3 Data Format for the Distribution Buffer (cont'd.) Byte Description Specific Comments 783 Second Header 784 Data Type (08) 7A5 824 Dog Mountain Rainfall Distribution 825-8A4 Ruby Creek 8A5-924 Bear Mountain 925-9A4 Agassiz Exp. Farm Rainfall Distribution 9A5-A24 Ryder Lake A25 FF Trailer A26 FF A27 FF A28 Record Type (FD) Header A29 Month A2A Day A2B Hour A2C Minute A2D Second Header A2E Data Type (10) A2F-B2E Receiver 1 (3550 MHz) Received Signal Distribution B2F-C2E Receiver 2 (3790 MHz) C2F-D2E Receiver 3 (4010 MHz) D2F-E2E Receiver 4 (7142 MHz) Received Signal Distribution E2F-F2E Receiver 5 (7496 MHz) F2F FF Trailer F30 FF F31 FF J-2 The Ryder Lake 6800 Microprocessor Software The purpose of the Ryder Lake software is to coordinate the sampling for both the meteorological variables and receiver signal amplitudes and transmit these on two links; a low speed (110 bps) link and a high speed (1200 bps) link respectively to UBC. The software flow chart to do this is given in Figure J-4. The program starts by initializing the ACIA I/O ports and A/D and 183 ( MAIN ") i INITIAIZE ACIA A/D STACK POINTER SEND SYNCH. = FFH SEND MET DATA SEND SYNCH. = FFH DEC MCOUNT l/IOSEC (" MASTER ) SAMPLE EACH VARIABLE 8 TRANSMIT Rx^CH^OOH Rx2=CH# I OH Rx. =CH#30H 4 SEND SYNCH. =FFH SAMPLE MET VARIABLES a STORE WD =CH#90H WS =CH#80H TEMP=CH#COH RAIN =CH#DOH IF DATA=FFH RESTORE AS F EH Figure J-4 Data Acquisition Flow Chart for the Ryder Lake Datalog 6800 Microprocessor. 184 then waits to sample data. The basic sampling interval is 1/10 of a second at which time the receivers are sampled and their data sent. Every tenth inter val, that is for each second, the meteorological variables are sampled and the data transferred to an output buffer for transmission. In this way the received signal levels are sent as soon as they are sampled, whereas the meteorological data are sent, one at a time, from an output buffer. The ACIA output ports at Ryder Lake are configured for one start and one stop bit which are matched by the UBC microprocessor data formatter input ports. J-3 The Weatherlog 8085 Microprocessor Software The purpose of the weatherlog software is to coordinate a one second A/D sampling interval of the four meteorological variables: wind speed, wind direction, temperature and rain, and then send these data on the communication channel to UBC at a rate of 110 bps. The weatherlog software flow chart is given in Figure J-5 and is the same for the Dog Mountain, the Ruby Creek and the Agassiz Experimental Farm site locations. As shown, the program starts by initializing the stack pointer, the A/D channels and the software delay loop. Then after each second the variables are sampled and outputted through the SOD pin of the 8085 microprocesor to the communications link. This cycle is repeated every second which is determined by a program delay loop. If any software changes are made to the main program software the delay loop must be recalibrated. This is facilitated by monitoring the rain clear output con nector of the microprocessor (see Figure 1-6) and using a frequency meter to measure the sampling interval. 185 VARIABLE a STORE AS  FOLLOWING a) CH N9'3 = WD b) CH N9 4:ws OCH N?5=TEMP d)CH NS«=RAIN t_ IF OATA IS «FFH RESTORE DATA AS=FEH _ _SEND_ WD_ SEND WS SEND TEMP SEND SYNCH « FFH Figure J-5 Program Flow Chart for the UBC Weatherlog 8085 Data Acquisition Microprocessors 186 J-4 The Bear Mountain 1802 Microprocessor Software Functionally, the Bear Mountain RCA 1802 microprocessor software shown in the flow chart of Figure J-6, is the same as the 8085 weatherlog. The only differences are in the method by which the timing for the one second sampling interval and the 110 bps data stream are derived. Instead of using software derived delay loops as is done in the 8085 system the 1802 software relies on externally derived interrupt pulses to establish sampling intervals and trans mission timing. J-5 The NOVA 840 Minicomputer Data Transfer Software The purpose of this software is to transfer data from the magnetic cas sette recorder to the magnetic reel-to-reel tape on the NOVA 840. The soft ware flow chart for this data transfer operation is given in Figure J-7. After initialization, the program first requests a block of data (2k bytes) from the cassette recorder by issuing an ASCII control command. The NOVA then converts the data to EBCDIC characters and outputs the block to the magnetic tape. After transferring one block another is requested for transfer and the process is repeated. 187 C BEAR MTN. INITIALIZE A/D REGISTERS WAIT FOR I SEC INTERRUPT 6 SAMPLE/SEND SAMPLE EACH VARIABLE a STORE-a) TEMP. = CH. N2 5 b) RAIN=CH.N26 IF DATA=FFH RESTORE DATA AS= FEH SEND 99H SEND CCH SEND SYNCH • FFH Figure J-6 Data Acquisition and Control Program Flow Chart for the Remote Bear Mountain 1802 Microprocessor Unit 188 J-4 The Bear Mountain 1802 Microprocessor Software Functionally, the Bear Mountain RCA 1802 microprocessor software shown in the flow chart of Figure J-6, is the same as the 8085 weatherlog. The only differences are in the method by which the timing for the one second sampling interval and the 110 bps data stream are derived. Instead of using software derived delay loops as Is done in the 8085 system the 1802 software relies on externally derived interrupt pulses to establish sampling intervals and trans mission timing. J-5 The NOVA 840 Minicomputer Data Transfer Software The purpose of this software is to transfer data from the magnetic cas sette recorder to the magnetic reel-to-reel tape on the NOVA 840. The soft ware flow chart for this data transfer operation is given in Figure J-7. After initialization, the program first requests a block of data (3k bytes) from the cassette recorder by issuing an ACSII control command. The NOVA then converts the data to EBCDIC characters and outputs the block to the magnetic tape. After transferring one block another is requested for transfer and the process is repeated. 189 c NOVA 8 40 INITIALIZE = CURRENT PORT MAG TAPE READ BLOCK FROM CURRENT PORT CONVERT FROM HEX TO EBCDIC STORE BLOCK ON MAGNETIC TAPE c STOP Figure J-7 Program Flow Chart to Transfer Data from Cassette Tapes to Magnetic Tape Using the Nova 840. 190 APPENDIX K THE BRIGHT BAND PROPAGATION EXPERIMENT'S DATA BASE MANAGEMENT SYSTEM K-1 Introduction This appendix gives the documentation to facilitate program development and maintenance of the data base management system (DBMS) used in this bright band propagation experiment. As described in Chapter IV, the data acquisition system generates sufficiently large volumes of data to make the bright band research project impractical unless a computerized data reduction, storage and analysis system is utilized. Thus the data handling function performed by DBMS is an integral part of this research project since it is responsible to provide the data analysis and graphical outputs which are used to draw the ex perimental conclusions. K-2 The System DBMS is a sequential system where the input is raw experimental data and the output is given as a presentation of graphical results via plotting rou tines. A functional flowchart relative to the bright band experiment showing their developmental status to the completion of this thesis work is given in Figure K-1. In terms of DBMS development, the entry procedure is complete, and an integrated plotting package which includes routines to plot attenuation with path average rain rate, individual rain rates, site temperature, site dif ferential temperature, differential temperature, temperature gradient and wind 191 £Xf>£.&/A>/£-A/rs4L £>ATA (FtfOAf A/OM A4//\Jf COMPUTER. 800 3 FORMAT COXPESPOAVDS TO DD ' /MPUT FORMAT . V. / \ TAPE P^EPAPED 3Y°THE EXTPACT MODE DBMS DATA ppocessor. DATA tapes PLOTT/NG RO(JT/NES_, PIOTTED OUTPUTS Figure K-1 DBMS Functional Flow Chart Showing Completion Status 192 speed as a function of time, have been finished. Several DBMS packages are still under development and these include the "SCAN" function, the "EXTRACT" function and several "PLOT" routine options. A detailed description of the major system software packages which com prise DBMS are dealt with in the following sections. These include the main DBMS program, the bright band "ENTER" function, the data directory, and the "PLOT" routine options. All these system packages use the Fortran IV-G compiler. K-3 DBMS Main The DBMS main program initializes the system and delegates program con trol to user requested operating modes as shown in its software flow chart given in Figure K-2. At the present time the "ENTER" and "PLOT" modes are fully integrated into DBMS, however, it is anticipated that the "SCAN" and the "EXTRACT" mode will be incorporated into the system during the next phase of this research program. K-4 "ENTER" "ENTER" is a conversion algorithm designed to convert recorded hex values to floating point engineering units for analysis. When the "ENTER" option is envoked the disc file or tape containing the hex information is read, convert ed to engineering units and then outputted to a temporary disc file, "-DATA". These data can then be analyzed using "PLOT" or stored to permanent disc files or on magnetic tape using MTS "FILESAVE". 193 I smerevec. Figure K-2 DBMS Main Program Flow Chart 194 K-5 The Data Directory Each of the formats into and out of the DBMS entry procedure are speci fied by a data directory. Each item in the directory is separately and para-metically defined so changes can be made to it without changing the executable software. Since the data directory defines the DBMS input and output formats, it is central to the development of new entry procedures for other experi ments . A listing of the data directory used in the entry procedure for the bright band experiment is given in Table K-1. The lettered labels for the column headings can be defined in more detail, as follows: A. Record sequence number; a label (not entered as part of the data directory), B. Field Name; descriptor of the data elements, C. Tape Type Number; 1 for output and 2 for input tape, D. Variable Type Number; "1" for 2 byte or 1 half-word size and "2" for 4 byte or 2 half-word size data fields, E. Starting byte position of this field in the data record, F. Ending byte position of this field in the data record, G. Field extraction parameter; gives the optional default specification, H. Data variable encoding format. Table K-1 Data Directory for the Time Series Format A B C D E F G H 1 Record Type 2 1 1 2 BCD 2 Month 2 1 3 4 BCD 3 Days 2 1 5 6 BCD 4 Hours 2 1 7 8 BCD 5 Minutes 2 1 9 10 BCD 6 Seconds 2 1 11 12 BCD 7 Gaugestatus 2 1 13 14 HEX 8 Dog-Wind DN 2 1 15 16 00 HEX 9 Dog-Wind SP 2 1 17 18 00 HEX 10 Dog-Temp 2 1 19 20 00 HEX 11 Dog-Rain 2 2 21 22 00 HEX 12 RBY-Wind DN 2 1 23 24 HEX 13 RBY-Wind SP 2 1 25 26 00 HEX 14 RBY-Temp 2 1 27 28 00 HEX 15 RBY-Rain 2 2 29 30 00 HEX 16 Bar-Wind DN 2 1 31 32 00 HEX 17 Bar-Wind SP 2 1 33 34 00 HEX 18 Bar-Temp 2 1 35 36 00 HEX 19 Bar-Rain 2 2 37 38 00 HEX 20 AGZ-Wind DN 2 1 39 40 00 HEX 21 AGZ-Wind SP 2 1 41 42 00 HEX 22 AGZ-Temp 2 1 43 . 44 00 HEX 23 AGZ-Rain 2 2 45 46 00 HEX 24 RLK-Wind DN 2 1 47 48 00 HEX 25 RLK-Wind SP 2 1 49 50 00 HEX 26 RLK-Temp 2 1 51 52 00 HEX 27 RLK-Rain 2 2 53 54 00 HEX 28 01-RX1:3550 2 1 55 56 HEX 29 01-RX2:3790 2 1 57 58 HEX 30 01-RX3:4010 2 1 59 60 HEX 31 '01-RX4:7142 2 1 61 62 HEX 32 01-RX5:7496 2 1 63 64 HEX 33 02-RXl:3550 2 1 65 66 HEX 34 02-RX2:3790 2 1 67 68 HEX 35 02-RX3:4.010 2 1 69 70 HEX Table K-1 Data Directory for the Time Series Format (cont'd.) B H 36 02-RX4:7142 37 02-RX5:7496 38 03-RX1.3550 39 03-RX2:3790 40 03-RX3:4010 2 2 2 2 2 71 73 75 77 79 72 74 76 78 80 HEX HEX HEX HEX HEX 41 03-RX4:7142 42 03-RX5:7496 43 04-RX1.3550 44 04-RX2:3790 45 04-RX3-.4010 2 2 2 2 2 81 83 85 87 89 82 84 86 88 90 HEX HEX HEX HEX HEX 46 04-RX4:7142 47 04-RX5:7496 48 05-RXl:3550 49 05-RX2:3790 50 05-RX3:4010 2 2 2 2 2 91 93 95 97 99 92 94 96 98 100 HEX HEX HEX HEX HEX 51 52 53 54 55 05-RX4:7142 05- RX5:7496 06- RXl:3550 06-RX2:3790 06-RX3:4010 2 2 2 2 2 101 103 105 107 109 102 104 106 108 110 HEX HEX HEX HEX HEX 56 06-RX4.-7142 57 06-RX5:7496 58 07-RXl:3550 59 07-RX2:3790 60 07-RX3:4010 111 113 115 117 119 112 114 116 118 120 HEX HEX HEX HEX HEX 61 62 63 64 65 07-RX4:7142 07- RX5:7496 08- RXl:3550 08-RX2:3790 08-RX3:4010 121 123 125 127 129 122 124 126 128 130 HEX HEX HEX HEX HEX 66 08-RX4:7142 67 08-RX5:7496 68 09-RXl:3550 69 09-RX2:3790 70 09-RX3:4010 2 2 2 2 2 131 133 135 137 139 132 134 136 138 140 HEX HEX HEX HEX HEX 197 Table K-1 Data Directory for the Time Series Format (cont'd.) A B C D E F G H 71 09-RX4:7142 2 141 142 HEX 72 09-RX5:7496 2 1 143 144 HEX 73 10-RX1:3550 2 1 145 146 HEX 74 10-RX2:3790 2 1 147 148 HEX 75 10-RX3:4010 2 1 149 150 HEX 76 10-RX4:7142 2 1 151 152 HEX 77 10-RX5:7496 2 1 153 154 HEX 78 EVENT FLAG 2 1 155 156 HEX 79 FF END OF BLOCK 2 1 157 .158 HEX 80 FF 2 1 159 160 HEX K-6 "PLOT" "PLOT" is an integrated interactive plotting package specifically designed for presenting the recorded measurements in a graphical time series format. The plotting options developed in work are as follows: 1. Rain rate and one receiver signal versus time; 2. Temperature -a) site temperature and one receiver signal versus time, b) site differential temperature and one receiver signal versus time, c) site to site differential temperature and one receiver signal versus time, d) temperature gradient and one receiver level versus time; 3. Site windspeed and one receiver level versus time; 4. Two receiver signals versus time (one second average); 5. High resolution two receiver levels versus time (all 1/10 second sample points). 198 In all these plot options the scales and offsets are interactively adjustable. The "PLOT" package takes source data in engineering units from the "-DATA" tempory file, processes it for the plotter into another temporary file called "-PLOT#". The plot can then be previewed using PLOTSEE and routed for output to the CALCOMP plotters or to a printronix printer (RMPROUTE=PTRXPLOT for the best output). A sample plotting rum is given in Figure K-3. 199 PROPAGATION DATA PLOTTING SYSTEM SELECT OUTPUT OPTION REQUIRED (ENTER FUNCTION NUMBER) (1) TIME SERIES PLOT OF ONE RECEIVER SIGNAL STRENGTH AND ONE RAIN RATE (3) TIME SERIES PLOT OF ONE RECEIVER SIGNAL STRENGTH ANO ONE OF SEVERAL TEMPERATURE DISPLAY OPTIONS (3) TIME SERIES PLOT OF ONE RECEIVER SIGNAL STRENGTH AND WIND SPEED ENTER CODE FOR SOURCE FILE DESIRED: t »> TEST SCRATCH FILE "-DATA" 2 •> PERMANENT DISK FILE "DO01.DAT" I . ENTER STARTING RECORD NO. TO BE PLOTTED: (16) 1 . ENTER NO. OF RECOROS TO BE PLOTTED: (16) 80O0. PLOT WILL DISPLAY BCOO RECORDS. STARTING WITH NO. t ARE DEFAULT VALUES OF FIRST RX . SIGNAL LEVEL DISPLAY OK? ORIGIN VALUE • -40 00 INCREMENT • 5.00 ENTER "<CR>" DR "NO * NO ENTER ORIGIN VALUE: (F5 0} -65. . ENTER INCREMENT VALUE: (F5.0) 5. . ARE OEFAULT VALUES OF SECOND RX. SIGNAL LEVEL DISPLAY OK? ORIGIN VALUE - -40.00 INCREMENT • 5. OO ENTER "<CR>" OR "NO" NO ENTER ORIGIN VALUE: <F5.0> -50.. ENTER INCREMENT VALUE: (F5.0) 5. . ENTER DISPLA* RESOLUTION COOE: t -> ONE PLOTTED POINT PER SECOND 2 -> TEN PLOTTED POINTS PER SECOND 1 . ARE DEFAULT VALUES OF TIME AXIS DISPLAY OK? ORIGIN VALUE • 0 0 INCREMtNT - ».OO ENTER •<CR>* 0« "NO" NO ENTER ORIGIN VALUE: (F5.0) O. . ENTER INCREMENT VALUE: (F5.0) • O. . FOR FIRST RECEIVER CHANNEL: ENTER RECEIVER NUMBER TO BE SELECTED: 1 •> 3550 MHZ. 3 •> 3990 MHZ. 3 >> 4010 MHZ. 4 •> 7143 MHZ. 5 •> 7496 MH?. 3. FOR SECOND RECEIVER CHANNEL; TYPICAL RESULTS: (INCHES) STRENGTHS SEQ. NO. X RX . 0 1 RX. #3 PLOT 1 1 OO 5.88 1 83 1301 3.00 5.83 0.37 3401 5.00 5.50 0.44 3601 7.00 5.B6 1 .45 4801 9.00 5.69 0.91 6001 11 .00 5.74 6003 11.00 5. 74 0.92 SUCCESSFULLY REAO 8001 DATA RECORDS PLOTTING WILL TAKE APPROX. 3 MIN. 53 SEC. MAXIMUM Y VALUE IS APPROX. 10 INCHES. 0 MIN 51 SEC, OR 33% OF TOTAL PLOT TIME SUCCESSFUL PLOT. fExecutIon laminated 0 % 16. 13, $4 1 87T #SET RMPROUTE-PTRXPLOT * $01. J41.B9T #R •PLOT SE E 0--PLOT* •Execution Begins AND 14 INCHES OF PAPER IS WITH PEN UP. I-LII--I-;v„v . i- j. -j --J j— • - -i i- i IG? PX •RMPRINT* assigned Job number 463548 Prlntronlx plot generation done. IG? PX Prlntronlx plot oeneration done. IG? PX Prlntronlx plot generation done. IG? End of plots. Last chance to use *IG. IG? $•RMPRINT• RM462548 released to PTRXPLOT 6 pages PRIORITY-NORMAL OEVICETYPf'PTRX *E«ocutlon Terminated Figure K-3 Sample Plotting Run 200 REFERENCES Van Trees, H., Hoversten, E., and McGarty, T., "Communications Satellites: Looking into the 1980's", IEEE Spectrum, pp. 43-51, Dec. 1977. GTE Lenkurt, Engineering Considerations for Microwave Communication  Systems, 1970. Silverthorn, D. , and Tetarenko, R., "Microwave Radio Spreads TV", Telesis, vol. 5, pp. 209-213, 1976. HervieUx, P., "RD-3: an 8 GHz Digital Radio System for Canada", Telesis, vol. 4, pp. 53-59, 1975. Anderson, C. , Barber, S. , and Patel, R. , "Propagation Experiments Show Path to Success", Telesis, vol. 6, pp. 180-185, 1977. Stephansen, E. , "Review of Clear-Air Propagation on Line-of-Sight Radio Paths", URSI Symposium on Effects of the Lower Atmosphere on Radio Propa gation at Frequencies Above 1 GHz", Lennoxville, Quebec, Canada, 26-30 May, 1980. Dougherty, H., "A Survey of Microwave Fading Mechanisms Remedies and Applications", ESSA Technical Report ERL69-WPL4, Boulder Co., NTIS Access No. COM-71-50288, Mar. 1968. Segal, B. , and Barrington, R., "The Radio Climatology of Canada - Tropo-spheric Refractivity Atlas for Canada", Communications Research Centre, Ottawa, CRC Rpt. #1315, Dec. 1977. Bean, B., Cahoon, B. , Samson, C., and Thayer, G., "A World Atlas of Atmospheric Radio Refractivity", Boulder Co. ESSA Monograph 1, NTIS Access No. AD648-805, 1966. International Radio Consultive Committee (CCIR), XlVth Plenary Assembly, Koyoto, Japan, vol. V, 1978. Bullington, K., "Radio Propagation Fundamentals", BSTJ, vol. 36, pp. 593-627, May 1957. Ryde, J., and Ryde, D. , General Electric Corporation Report,' #7831, 1941; #8516, 1944; #7831, 1941; #9670, 1945. Ryde, J.W., "The Attenuation and Radar Echos Produced at Centimetre Wave-Lengths by Various Meteorological Phenomena", Rpt. of the Physics Society, London and the Royal Meteorological Soc. entitled "Meteorologi cal Factors in Radio-Wave Propagation". 201 14. Medhurst, R.C, "Rainfall Attenuation of Centimeter Waves: Comparison of Theory and Measurement", IEEE Trans. Ant. & Prop., vol. AP-13, vol. 4, pp. 550-564, 1965. 15. Robertson, S., and King, A., "The Effect of Rain Upon the Propagation of Waves in the 1 and 3 Centimeter Regions", Proc. IRE, vol. 34, pp. 178P-180P, 1946. 16. Okamura, S., Funakawa, K., Uda, H. , Kato, J., and Oguchi, T. , "On the Measurement of Attenuation by Rain at 8.6 mm Wave Length", J. Radio Research Labs (Tokyo), vol. 6, pp. 255-267, April 1959. 17. Robinson, N. , "Measurements of the Effect of Rain, Snow and Fogs on 8.6 mm Radar Echos", Proc. IEE, vol. 102, pt. B, pp. 709-714, Sept. 1955. 18. Funakawa, K. , and Kata, J., "Experimental Studies of Propagation Characters of 8.6 mm Wave on a 24 km Path", J. Radio Research Labs (Tokyo), vol. 9, pp. 351-367, Sept. 1962. 19. Crane, R. , "Propagation Phenomena Affecting Satellite Communication Systems Operating in the Centimeter and Millimeter Wavelength Bands", Proc. IEEE, vol. 59, pp. 173-188, Feb. 1971. 20. Rodgers, CV. , and Olsen, R.L., "Calculation of Radio Wave Attenuation Due to Rain at Frequencies up to 1000 GHz", Communications Research Centre (Canada), Rpt. #1299, 1976. 21. Crane, R.K. , "Attenuation Due to Rain - A Mini Review", IEEE Trans. Antennas and Propagation, AP-23, pp. 750-752, 1975. 22. Olsen, R.L., Rodgers, D.V., and Hodge, D.B., "The aRb Relation in the Calculation of Rain Attenuation", IEEE Trans, on Antennas and Propaga tion, AP-26, pp. 218-239, 1978. 23. Laws, J.O., and Parsons, D.A., "The Relation of Drop Size to Intensity", Amer, Geophys. Union Trans., vol. 22, pp. 452-460, 1943. 24. Marshall, J. , and Palmer, W. , "The Distribution of Raindrops with Size", J. Meteorology, vol. 5, pp. 165-177, 1948. 25. Joss, J. , Thams, J., and Waldvogel, A., "The Variation of Raindrop Size Distributions at Lacarno", in Proc. of the Intl. Conf. on Cloud Physics, Toronto, Canada, pp. 369-373, 1968. 26. Fedi, F. , and Mandarini, P., "Influence of Spacing and Integration Time of the Raingauges on the Attenuation Rain Rate Relationship", J. Rech. Atm., vol. 8, #1-2, 1974. 27. Semplak, R., and Keller, H., "A Dense Network for Rapid Measurement of Rainfall Rate", BSTJ, pp. 1745-1756, July-Aug. 1969. 202 28. GTE Lenkurt, "Antenna Decoupling Due to Non-Selective Power Fading", Demodulator, vol. 27, #3, pp. 1-6, Mar. 1978. 29. DeLange, 0., "Propagation Studies of Microwave Frequencies by Means of Very Short Pulses", BSTJ, vol. 31, #1, pp. 91-103, Jan. 1952. 30. Meadows, R. , Lindgren, R., and Samuel, J., "Measurement of Multipath Propagation Over a Line-of-Sight Radio Link at 4 Gc/s Using Frequency Sweep Technique", Proc. IEE, vol. 113, #1, pp. 41-48, Jan. 1966. 31. Rummler, W., "Extentions of the Multipath Fading Channel Model", Intl. Conf. on Commun., vol. 2, Boston, Mass., pp. 32.2.1-32.2.5, June 1979. 32. Sasaki, 0., and Akiyama, T., "Multipath Delay Characteristics on Line-of-Sight Microwave Radio System", IEEE Trans, on Commun., vol. COM-27, #12, pp. 1876-1886, Dec. 1979. 33. Sanberg, J., "Extraction of Multipath Parameters from Swept Measurements on a Line-of-Sight Path", IEEE Trans. Ant. and Prop., vol. AP. 34. Martin, L. , "Etude de la Selective des Evanouissements dus aux Trajets Multiple", URSI Comm. F. Symposium, Lennoxville, Quebec, Canada, pp. 3.5.1-3.5.5, May 1980. 35. Bello, P., Pickering, L., and Boardman, C., "Multipath Over LOS Channels Study", Rpt. by CNR INC for Rome Air Development Center, Air Force Sys tems Command, Griffiss Air Force Base, N.Y. 13441, Aug. 1977. 36. Barnett, W. , "Multipath Propagation at 4, 6 and 11 GHz", BSTJ, vol. 51, #2, pp. 321-361, Feb. 1972. 37. Weibel, G., and Dressel, H., "Propagation Studies in Millimeter-Wave Link Systems", Proc. IEEE, vol. 55, pp. 497-513, Apr. 1967. 38. Vigants, A., "Number and Duration of Fades at 6 and 4 GHz", BSTJ, vol. 50, #3, pp. 815-841, Mar. 1971. 39. Watson, P., "Survey of Measurements of Attenuation by Rain and Other Hydrometeors", Proc. IEE, vol. 123, pp. 863-871, 1976. 40. Oomorl, T. , and Aoyagi, S., "A Presumptive Formula for Snowfall Attenua tion of Radio Waves", Elec. & Comm. in Japan, vol. 54-B, #9, pp. 34-39. 41. Kolosov, M.A., Poshidaer, V.N., Sokolov, A.V., and Sukhoniu, E.V., "The Prediction of Millimetre Radio Wave Attenuation in Precipitation", IVCRM Colloquim, Nice, France, Preprints paper IV-12, Oct. 1973. 42. McCormick, K.S., "A Comparison of Precipitation Attenuation and Radar backscatter Along Earth-Space Paths", IEEE Trans. Ant. and Prop., AP-20, #6, pp. 747-755, 1972. 203 43. Goldhirsh, J., "Attenuation of Propagation Through Rain for an Earth-Satellite Path Correlated with Predicted Values Using Radar", IEEE Trans. Ant. and Prop., AP-24, pp. 800-806, 1976. 44. Dissanayake, A.W., and McEwan, J.J., "Radar and Attenuation Properties of Rain and Bright Band", IEE Antennas and Propagation Conf. Proc., London, Eng., Nov. 1978. 45. Goldhirsh, J., "A Review on the Application of Nonattenuating Frequency Radars for Estimating Rain Attenuation and Space Diversity Peformance", IEEE Trans, on Geoscience Elec, vol. GE-17, #4, pp. 218-239, 1979. 46. Goldhirsh, J. , "Prediction of Slant Path Rain Attenuation Statistics at Various Locations", Radio Sc., vol. 12, #5, pp. 741-747, 1977. 47. Hendry, A., Antar, Y., Schlesak, J., and Olsen, R., "Melting Layer Attenuation at 28.56 GHz from Simultaneous Comstar Beacon and 16.5 GHz Diversity Radar Observations", presented at IEE Second Intl. Conf. on Antennas and Propagation, U. of York, U.K., Apr. 13-16, 1979. 48. Watson, P., McEwan, N. , Dissanayake, A., and Haworth, D. , "Attenuation and Cross-polarization Measurements at 20 GHz Using the ATS-6 Satellite with Simultaneous Radar Observations", IEEE Trans. Ant. & Prop., AP-27, pp. 11-17, 1979. 49. Nishitsuji, A., and Mats umoto, A., 'The Character of Falling Snow", from SHF and EHF Propagation in Snow Districts, Momograph Series of the Research Inst, of Applied Electricity, Hokkaido U. , Sapporo, Japan, vol. 19, pp. 63-78, 1971. 50. Nishitsuji, A., and Matsumoto, A., "Calculation of Radiowave Attenuation Due to Snowfall", ibid., pp. 63-78, 1971. 51. Nishitsuji, A., and Matsumoto, A., "Problems on the Radio Wave Attenua tion in Millimeter and Microwave Regions at Snowfall", ibid., pp. 79-91, 1971. 52. Mason, B., The Physics of Clouds, Oxford University Press, Amen House, London, 1957. 53. Pruppacher, H. , and Klett, J., Microphysics of Clouds and Precipitation, D. Reidel Publishing Co., Dordrecht, Holland, 1978. 54. Hallgren, R. , and Hosier, C. , "Preliminary Results on the Aggregation of Ice Crystals", Physics of Precipitation, Geophysical Monograph #5 NAS-NRC #746, pp. 257-263, 1959. 55. Rogers, R., A Short Course in Cloud Physics, Pergamon of Canada Ltd., Toronto, Canada, 1976. 204 56. Ludlarn, F., Clouds and Storms, The Pennsylvania State U. Press, 1980. 57. Owen, N. , "Digital Radio Test at 8 GHz, Big Sicker-Vancouver", B.C. Telephone Co. Rpt., Exp. Proj. 000723, Jan. 1981. 58. Dept. of Communications, "Application Procedures for Planned Radio Sta tions Above 890 MHz in Terrestrial Fixed Service", RSP-113, Issue 2, Sept. 1975. 59. Kheirallah, H., Segal, B., and Olsen, R. , "Application of Synthetic Storm Data to Evaluate Path Average Rain Rate Techniques for Predicting Rain Attenuation Statistics", Proc. URSI Commission F Open Symposium, Lennoxville, Quebec, Canada, pp. 2.10.1-2.10.6, 1980. 60. Drafuca, G., J. Rech. Atmos., vol. 8, pp. 399-441, 1974. 61. Bertook, E., DeRenzis, G., and Drufuca, G., Proc. URSI Commisssion F Open Symposium, La Baule, France, pp. 295-300, 1977. 62. Watson, P., Ahmed, H., and Papaioannou, G., IEE Conf. Proc. 169, part 2, pp. 92-96, 1978. 63. Annual Meteorological Summary for 1978, 1979, 1980: Agassiz Experimental Farm, Atmospheric Environment Service, Environment Canada. 64. Van der Star, J., and Kharadly, M., "Measurements of Copolar Attenuation Through the Bright Band at 4 and 7 GHz", Annales des Telecommunications, vol. 36, #1-2, Feb. 1981. 65. Van der Star, J. , and Kharadly, M. , "Measurement of Copolar Attenuation Through the Bright Band at 4 & 7 GHz", Proc. URSI Commission F Open Symposium, Lennoxville, Quebec, Canada, pp. 4.8.1-4.8.3, 1980. 66. McNicol, J., Rain Attenuation at 74 GHz, M.A.Sc. Thesis, Dept. of Electrical Engineering, The University of British Columbia, Mar. 1977. 67. Bullington, K., "Radio Propagation for Vehicular Communications", IEEE Trans, on Behicular Technology, vol. VT-26, #4, pp. 295-308, Nov. 1977. 68. Okumura et al., "Field Strength and it's Variability in VHF and UHF Land-Mobile Radio Service", Review of Tokyo Elec. Comm. Lab., vol. 16, pp. 835-873, Sept.-Oct. 1978. 69. Shepherd, N., "VHF and UHF Propagation", General Electric Datafile Bulletin #10003-1, July 1962. 70. Peters, B., An Experimental Investigation of Dual-Polarized Atmospheric  Propagation at 73 GHz, Ph.D. Thesis, Dept. of Electrical Engineering, The University of British Columbia, April 1982. 205 Thomson, D., Coulter, R. , and Warhaft, Z. , "Simultaneous Measurements of Turbulence in the Lower Atmosphere Using SODAR and Aircraft", Fourth Symposium on Meteorological Observations & Instrumentation of the American Meteorological Society, Denver, Colorado, April 10-14, 1978. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items