LIBRARY OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 AT URBANA-CHAMPAIGN 
 
 510.84 
 
 Cop. Z 
 
The person charging this material is re- 
 sponsible for its return to the library from 
 which it was withdrawn on or before the 
 Latest Date stamped below. 
 
 Theft, mutilation, and underlining of books 
 are reasons for disciplinary action and may 
 result in dismissal from the University. 
 
 UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN 
 
 ;'•>■■ 
 
 V 
 
 ?> 
 
 
 L161 — O-1096 
 
Digitized by the Internet Archive 
 in 2013 
 
 http://archive.org/details/digitalradioaids633facc 
 
f UIUCDCS-R-74-633 
 
 )U-6>33 
 
 /tyaX4 
 
 *? 
 
 DIGITAL RADIO AIDS FOR A FLIGHT SIMULATOR 
 RADIO MEMORIES - DISPLAY - CODE KEYER 
 
 by 
 
 Lucien Isidore Facchin 
 
 May, 19 J k 
 
 IH£ LIBRARY OF TH£ 
 
 DEPARTMENT OF COMPUTER SCIENCE 
 UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN 
 
 URBANA, ILLINOIS 
 
UIUCDCS-R-7^-633 
 
 DIGITAL RADIO AIDS FOR A FLIGHT SIMULATOR 
 RADIO MEMORIES - DISPLAY - CODE KEYER 
 
 by 
 
 Lucien Isidore Facchin 
 Ingenieur E. S. I. E. E. , Paris (France), 1972 
 
 May, 197U 
 
 Department of Computer Science 
 University of Illinois 
 Urbana, Illinois 6l801 
 
 This work was supported in part by the Institute of Aviation and 
 was submitted in partial fulfillment of the requirements for the degree 
 of Master of Science in Electrical Engineering, May, 197*+ • 
 
Ill 
 
 ACKNOWLEDGEMENTS 
 
 The author wishes to express his sincerest gratitude to his 
 advisor, Professor W. J. Kubitz for his guidance, timely suggestions and 
 continuing interest. The author also wishes to thank Mr. Lynn Staples, 
 from the Institute of Aviation, who suggested the digital Radio Aids. 
 
 Further, the author wishes to thank Ms. Evelyn Huxhold, who 
 typed the final version of the paper. 
 
IV 
 
 ABSTRACT 
 
 The research described here is part of a joint study between the 
 Department of Computer Science and the Institute of Aviation of the 
 University of Illinois. This study is a result of the need to consider 
 alternatives to using general purpose digital computers for the solution 
 of the entire aircraft simulation problem. The software costs of such 
 systems have sky-rocketed in recent years while hardware costs continue to 
 fall. Thus, serious consideration must be given to direct functional 
 mechanization through the use of special purpose computers. The use of 
 small dedicated computers may greatly simplify interfacing to the central 
 control computer in addition to reducing the amount of software required 
 by difusing the processing throughout the simulator. This document 
 presents an implementation of some of the functions of such a system, the 
 Radio Aids System. 
 
TABLE OF CONTENTS 
 
 Page 
 
 1. INTRODUCTION 1 
 
 1.1. Definition of the Basic Simulator ----------- l 
 
 1.2. Present Systems -------------------- 2 
 
 1.3. Problems Appearing in the Present Systems and a Possible 
 Way to Decrease Some of Them -------------- 2 
 
 1.3.1. Software Cost 2 
 
 1.3.2. Wiring from Cockpit to Interfaces ------- 1+ 
 
 1.3.3. Proposed Solution to These Problems (Special 
 Purpose Computer or Direct Functional Mechani- 
 zation). -------------------- k 
 
 1.3-H. The Radio Aids and a System Using CORDIC 6 
 
 1.3.5. Radio Memories, Display and Code Keyer ----- 8 
 
 2. THE RADIO MEMORIES 9 
 
 2.1. Radio Memories --------------------- 9 
 
 2.2. Memory Elements -------------------- 9 
 
 2.3- Scanning Device --------------- •_---- io 
 
 2.3.1. A First Approach 10 
 
 2.3-2. A Second Approach 15 
 
 2.3.2.1. First Solution 15 
 
 2.3.2.2. Second Solution IT 
 
 2.3-3. A Third Approach 22 
 
 3. THE DISPLAY 25 
 
 3.1. Conversion to BCD 25 
 
 3-1.1. Conversion Using Counting -----------25 
 
 3.1.2. Conversion Using Combinatorial Logic ------ 27 
 
Page 
 
 3.1.3. Conversion Using BCD Adders 29 
 
 3.1.^. Conversion Using ROMs and a Counter -------29 
 
 3.2. Solution Implemented ------------------32 
 
 3.2.1. The Conversion Circuits -------------33 
 
 3.2.2. The Control Logic 36 
 
 k. THE CODE KEYER ' 39 
 
 k.l. Definition of the Problem - • 39 
 
 k.2. First Solution kO 
 
 k.3. Second Solution U2 
 
 k.k. Multiplexing and Morse Decoder ------------- k6 
 
 It. 5- Code Keyer Card ---------------------1+6 
 
 k.6. Code Keyer Circuits -------------------52 
 
 k.6.1. Clear Device 52 
 
 k.6. 2. The 1 kHz and 10 Hz Clocks 5^ 
 
 U.6.3. Shift Register - Data Enable Flag - Clock 
 
 Inhibit Flag 5U 
 
 h.6.h. End of Character, End of Code Detectors - 
 
 Letter Counter ----------------- 56 
 
 U.6.5. Multiplexer Address Selector ----------58 
 
 h.6.6. Jump if No I 58 
 
 U.6.T- Sequence Device ----------------- 60 
 
 5. CONCLUSION 62 
 
 LIST OF REFERENCES 63 
 
VI 1 
 
 LIST OF FIGURES 
 
 Figure Page 
 
 1. Block Diagram of a Typical Flight Simulator -------- 3 
 
 2. System Under Study 5 
 
 3. Block Diagram of the Radio Aids -------------- 7 
 
 k. Memories Configuration ------------------ n 
 
 5. Functional Diagrams of Radios ---------------12 
 
 6. First Approach of the Scanning -------------- il+ 
 
 7. Case of Stations Using the Same Frequency --------- lU 
 
 8. First Solution l6 
 
 9- Second Solution (First Scheme) -------------- 18 
 
 10. Second Solution (Second Scheme) --------------21 
 
 11. A Third Approach 23 
 
 12. Block Diagram of the Display --------------- 26 
 
 13. Conversion to BCD Using Counting (Basic Operation) - - - - 26 
 lU. Conversion to BCD Using Counting (Extended Operations) - - 28 
 
 15. Conversion Using Combinatorial Logic ----------- 28 
 
 16. Conversion Using BCD Adders -----------.-----30 
 
 17. Conversion Using ROMs and a Counter ------------31 
 
 18. Converter Used for Multipurpose Conversions -------- 32 
 
 19- Block Diagram of the Display --------------- 3U 
 
 20. Conversion Circuits --------------------35 
 
 21. Control of the Display 37 
 
 22. First Solution Ul 
 
 23. Block Diagram of the Code Keyer (Solution l) 1+3 
 
 2k. Block Diagram of Solution 2 - - 1+5 
 
Vlll 
 
 Figure Page 
 
 25. Block Diagram of the Letter Select and Morse Decoder - - - - kf 
 
 26. Card CK1: Implementation of the Letter Select and Morse 
 Decoder -------------------------- U8 
 
 27. Block Diagram of the Code Keyer ' 1+9 
 
 28. Circuit Diagram of the Code Keyer -------------51 
 
 29. Clear Device 53 
 
 30. 1 Khz and 10 Hz Clocks ■ 53 
 
 31. Shift Register-Data Enable and Clock Inhibit Flags ----- 55 
 
 32. End of Character Detector-End of Code Detector- Letter 
 Counter --------------------------57 
 
 33. Multiplexer Address Selector ----------------59 
 
 3U. Jump if no I 59 
 
 35 • Sequence Device ---------------------- 6l 
 
1. INTRODUCTION 
 
 1.1. Definition of the Basic Simulator 
 
 Since the thirties, when general aviation underwent a large expan- 
 sion, it has been suggested that simulators he used in order to ease the 
 flight instruction process, at least in the initial stage. One of the 
 first trainers was the Link Trainer. 
 
 This trainer is a relatively simple device. It is composed of 
 a series of pneumatic pumps and pushrods that can turn and bank a plywood 
 box, whose inside is made to look like an aircraft cockpit. The pilot sits 
 in the box and moves it around using appropriate stick or control wheel 
 movements. The pilot learns the basic means for making the aircraft respond 
 by using the control system, as well the rate with which an aircraft will 
 respond to the pilot's commands. A trainer gives a pilot a good idea of 
 how the instruments on the aircraft panel react to each movement of the 
 control wheel. The pilot quickly learns how the speed and altitude of the 
 aircraft are affected by different power settings and control wheel move- 
 ments. The pilot learns which instrument to watch for precise monitoring 
 of the behavior of the aircraft. With this knowledge, a pilot is better 
 able to fly the first time he pilots an aircraft. In addition, there 
 is the distinct advantage of being able to make mistakes in the Link 
 Trainer during initial attempts without worrying about damaging the air- 
 craft. In some cases, flight simulation is the only possible way to teach 
 flying; for example, the landing on the moon or the training for air-to- 
 air combat. 
 
1.2. Present Systems 
 
 Since the first Link Trainer, aircraft have become more complex 
 as have the simulators, which rely heavily on computers. Figure 1 shows a 
 block diagram of a typical flight simulator system. It represents a cock- 
 pit which contains the aircraft instruments delivering analog signals. In 
 order to he processed by a computer, these signals must be converted to 
 digital form through an analog to digital interface (commonly called a 
 linkage in the jargon of the flight simulator specialist). The data is 
 then processed in a computer which contains data and programs; in fact, in 
 several cases we need more than a single computer since aircraft have 
 become more and more sophisticated. Associated with the computer is a 
 console for the instructor who gives orders, decides changes of data in the 
 simulated flight and checks the reactions of the trainee. The data coming 
 from the computer is fed back to the cockpit through a digital to analog 
 interface. 
 
 1.3. Problems Appearing in the Present Systems and a Possible Way to 
 Decrease Some of Them . 
 
 Several problems have arisen in current flight simulators; two of 
 them will be discussed here. 
 
 1.3.1. Software Cost 
 
 Flight simulators typically need several computers. Each time 
 these computers change, the software must be rewritten, usually at great 
 cost. During the past two decades, hardware and software have exchanged 
 roles. At the beginning of the '60s, the hardware costs accounted for more 
 than 50 per cent of the total computer system costs. Now the situation is 
 reversing in favor of hardware, which currently accounts for less than 50 
 
COCKPIT 
 
 ,** 
 
 * r 
 
 3 
 
 A/D 
 INT. 
 
 5 
 
 COMPUTER 
 
 71 
 
 3 
 
 D/A 
 INI 
 
 INSTRUCTOR K> 
 CONSOLE 
 
 Figure 1. Block Diagram of a Typical Flight Simulator, 
 
per cent of the total computer system costs. In future years, it seems 
 that the share of costs due to hardware will continue to decrease. Due 
 to the low cost of LSI hardware, more and more complex functions, usually 
 performed "by software, are "being implemented in hardware form. Another 
 argument for direct hardware implementation is reducing the time needed to 
 perform the required computation. 
 
 1.3.2. Wiring from Cockpit to Interfaces 
 
 Due to the large number of airplane instruments, the amount of 
 wiring connecting the cockpit to the interfaces is very large. In addi- 
 tion, the distance between them may reach several tens of yards making the 
 cost very high. 
 
 1.3.3. Proposed Solution to These Problems (Special Purpose Computer or 
 Direct Functional Mechanization) 
 
 Among the other functions to be simulated, the Radio Aids, which 
 are the main navigation instruments, may be implemented without going 
 through a classical computer but by using a special purpose computer as 
 shown in Figure 2. The data coming from the radio instruments are 
 directly coded in binary form, thus eliminating any analog to digital and 
 digital to analog interfaces. This is made possible by using an algorithm 
 called CORDIC which implements trigonometric functions (sin, cos, arctg,... 
 log) using only a small amount of hardware. The idea of CORDIC was origi- 
 nally proposed for realizing the transformation (X, Y) +- (R, 0) to compute 
 the range of the plane. The use of Read Only Memories allows one to imple- 
 ment several functions using the same basic elements (adders, shift regis- 
 ters) without duplication, by writing new codes in those ROMs. 
 
 
ANM.06 
 
 COCKPIT 
 
 Kr 
 
 j^A/D 
 
 
 INI 
 
 GENERAL 
 PURPOSE 
 COMPU- 
 TER 
 
 (REDUCED) 
 
 ~7V 
 
 3f a 
 
 INT 
 
 SPEC AT 
 PURPOSE 
 COMPUTER 
 
 J- 
 
 '_^L 
 
 \/ 
 
 INSTRUCTOR 
 CONSOLE 
 
 Figure 2. System Under Study. 
 
I.3.U. The Radio Aids and a System Using CORDIC 
 
 Radio navigation is now the most common form of navigation. In 
 addition to weather information, radios help the pilot to determine his 
 position from specific radio stations. In the cockpit, the pilot has 
 available a set of radios which are used for different purposes. One 
 class of radios are those used for the airway system and are themselves 
 divided into two groups: the VOR (Visual Omni-Range) which informs the 
 pilot on which radial he is flying, and the Low Frequency Radio Beacon 
 (Distance Measurement Equipment) which determines the distance by computing 
 the time required for a pulse of radio energy to travel from the plane to a 
 ground station and return. For landing, the radio instruments used are the 
 ILS (instrument Landing System) which provide the information a pilot needs 
 for landing in poor visibility conditions. 
 
 A very simplified block diagram of the radio aids simulator is 
 given in Figure 3. The system monitors the radios in the cockpit. At the 
 request of the pilot, the Radio Aids simulator must provide to the cockpit 
 the information that. the pilot would receive if he were in a real flight 
 (such as Morse code, range, ...). 
 
 The system has a set of Radio Memories (PROMs) which can be pro- 
 grammed and which contain the data characterizing the stations within 
 range of the aircraft. These memories are scanned so as to find all sta- 
 tions receivable by the radios. The Cordic Unit uses data from the memories 
 in conjunction with the aircraft position coordinates to perform several 
 calculations such as the range computation. A particular function of the 
 Radio Aids is the Code Keyer which must provide the identifying code for 
 each radio station which is received. 
 
~1 
 
 RADIOS 
 
 =Fh 
 
 I I 
 
 CONSOLE 
 
 COCKPIT -| |. 
 
 DISPLAY 
 
 3 
 
 SCAN- 
 NING 
 
 3 
 
 ~i 
 
 M EMORI ES 
 
 AlflCRAFT 
 POSITION 
 
 3 
 
 CORDIC 
 UNIT 
 
 RA.NGE . 
 
 > 
 
 i^i: 
 
 CODE 
 KEYER 
 
 BUS 
 
 J 
 
 Figure 3. Block Diagram of the Radio Aids. 
 
8 
 
 1.3-5. Radio Memories, Display and Code Keyer 
 
 The following pages present solutions which have "been proposed 
 and the implementation of one of them for the following functions: 
 
 a) Radio Memories 
 
 b) Display 
 
 c) Code Keyer. 
 
2. THE RADIO MEMORIES 
 
 Introduction 
 
 In the introduction, a system using the Cordic algorithm was 
 discussed in its general configuration. One of its functions will be 
 considered here and a design is proposed for the Radio Memories system. 
 This is described in the following pages. 
 
 2.1. Radio Memories 
 
 A flight simulator is a completely self-contained device, having 
 no communication with its environment. Therefore, we must store the radio 
 stations parameters in some way. By radio station parameters, we mean: its 
 X, Y coordinates (latitude and longitude), its altitude, the type of station, 
 the magnetic variation, the power of emission, and the morse code letters 
 which there are three. In this system we can differentiate two functions: 
 the PROMs which are the memories themselves, and the scanning device which 
 selects one address of the memories in order to load or check its contents. 
 
 2.2. Memory Elements 
 
 The first factor to consider for the memories is the number of 
 words they should contain (a word corresponds to one station). This is 
 determined by the autonomy of the plane used. In the case of small planes, 
 a radius of 300 miles has been considered adequate. An investigation of 
 the number of radio stations within a 300 mile radius has been done and it 
 indicates a number on the order of 200. Therefore a possible configuration 
 is to use 256 word memories (e.g. INTEL 1T02A). The data to be stored are: 
 X: the first coordinate of the station with respect to a given 
 
 origin; X is either positive or negative, coded in two's 
 
 complement . 
 

 10 
 
 Y: second corrdinate of the radio station. It has exactly 
 the same format as X. 
 
 Z: altitude in feet and essentially a positive number, 
 
 magnetic variation which indicates the difference between the 
 
 geographic and magnetic longitudes. 
 
 radiated power 
 
 type (VOR, radar, DME, ...) 
 
 Morse code letters which are three five bit words. 
 X and Y are the variables which require the largest number of bits {2h) 
 and thus require 3 PROMs. Therefore, the use of a 2k line bus has been 
 considered. For the altitude Z (l6 bits) and the magnetic variation (8 
 bits), the data is spread from the first to the 2^th lines. A set of con- 
 trol lines provide signals which enable the PROMs containing the desired 
 data. See Figure k. 
 
 2.3. Scanning Device 
 
 A radio is an element which has two aspects (see Figure 5)- One 
 is the tuning system. When a frequency is selected a binary number is 
 generated and placed on the bus in order to select corresponding data in 
 the PROMs. The other aspect of a radio is the visual display which dif- 
 fers with the type of radio. This visual information is obtained after 
 processing of the data contained in the PROMs for that radio. 
 
 The analysis of the scanning device has progressed through 
 several stages and in the following pages , several approaches will be 
 presented. 
 
 2.3.1. A First Approach 
 
 A first approach considers a limited number of radios (up to 8) 
 and supposes that each frequency corresponds to only one radio station. 
 
11 
 
 2 4 
 
 2 4 
 
 16 
 
 8 3 5 
 
 MV 
 
 POWER MORSE 
 TYPE CODE 
 
 I c £ 
 
 ■ i 
 
 i CE 
 
 CE 
 
 
 
 
 ! 
 
 \ 
 
 
 t H 
 
 
 • 
 
 • 
 
 
 
 
 
 - 
 
 
 
 ' 
 
 
 
 
 
 
 
 
 
 , 
 
 DATA 
 BUS 
 
 CONTROL 
 
 ' BUS 
 
 Figure h. Memories Configuration. 
 
RADIO 
 
 I 
 I 
 
 VISUAL OR 
 
 audio IT LINING 
 
 INFORMATION 
 
 DATA IN 
 
 fftCQUENCy 
 BUS 
 
 i 
 i * 
 
 12 
 
 FREQUENCY 
 
 RADIO 
 
 Setecl- 
 
 BUS 
 
 RADIO 2 
 
 fi 
 
 Seled 
 
 
 Figure 5. Functional Diagrams of Radios, 
 
13 
 
 In addition we do not consider the interaction with the computing element 
 (CORDIC unit). Each frequency may he represented "by 8 hits. Refer to 
 Figure 6. 
 
 We have a set of radios each delivering an 8 hit number. Each 
 radio must he considered individually as it addresses the memory. This is 
 implemented "by using eight 8 hit multiplexers. Each multiplexer selects 
 one hit among 8 which belong to all radios. A clock drives a 3 hit binary 
 counter which sequences the multiplexers selecting the radio frequencies 
 individually and consecutively. 
 
 This is a very simple approach to the system which must be modi- 
 fied for the following reason. Since the number of radio stations is much 
 greater than the number of frequencies available, there exist frequencies 
 used by more than one radio station. Since we are addressing the PROMs by 
 the radio frequency, the preceeding system must be modified. 
 
 Figure 7 indicates, in an X, Y plane, the position of the aircraft 
 and two stations noted 1 and 2 which are using the same frequency f . 
 When a frequency is set on one of the radios, we are to determine which of 
 the stations is the one we are receiving (in simulation). In addition, we 
 suppose that the plane cannot receive the signals emitted by more than one 
 station having the same frequency. The selection of the right station is 
 realized as indicated. First, the distance dl from the plane to the first 
 station found in the PROM (let it be station l) for the given frequency is 
 determined. Then, distance d2 from the plane to the next station with 
 same frequency (let station 2) is computed. If they are the only stations 
 using that frequency, we compare dl and d2 and the station having the 
 smallest distance is the right one (for dl < d2 it is station l). 
 
lU 
 
 AADIO I 
 
 RADIO 2 
 
 up to 
 8 
 
 M 
 P 
 X 
 
 1 
 
 M 
 P 
 X 
 
 8 
 
 clock 
 
 output 
 
 -^counter 
 
 1 
 
 PROM 
 address 
 
 Figure 6. First Approach of the Scanning. 
 
 station 2'* 
 
 station 1 
 
 > X 
 
 d'l < d2 
 
 Figure 7. Cane of Stations Using the Same Frequency. 
 
15 
 
 2.3.2. A Second Approach 
 
 At this point, we will use the above constraint. We still sup- 
 pose that there are only 8 bit frequencies, and as before we do not con- 
 sider the interaction with the computing unit. Several solutions have 
 been proposed which will be roughly sketched. 
 
 2.3.2.1. First Solution 
 
 This solution uses one additional ROM which contains the frequen- 
 cies of the stations. Figure 8b shows this PROM which contains the 
 frequencies. For a frequency fl, we have two stations and therefore the 
 frequency fl will appear twice in the PROM (addresses 8 and 91 for example). 
 When the PROMs are selected at these addresses, the data corresponding to 
 these stations are available on the data bus. 
 
 The system works in the following manner: one radio is selected 
 using multiplexers and its frequency enters a comparator. During this 
 time, an 8 bit counter scans the entire radio memory in order to test 
 the frequency which is contained in each word. This frequency is the 
 second input of the comparator. When both frequencies are identical, the 
 comparator generates a signal A = B indicating that we have found a sta- 
 tion having the same frequency as the one corresponding to the radio con- 
 sidered. At this time we can use the data presently on the data bus. The 
 counter scans all the addresses. When the last address has been selected, 
 we again start the cycle and a signal is generated in order to select the 
 next radio (this is implemented by incrementing a 3 bit counter which add- 
 resses the multiplexers). The same procedure as indicated above is repeated 
 for the new radio and so on. See Figure 8a. 
 
 An analysis of this system shows a great advantage to this method 
 of scanning: any station data (frequency and parameters) can be placed 
 
16 
 
 PROM ADDRESS 
 
 PROM 
 (frequency) 
 
 Id. 
 
 DATA PROMS 
 
 >A= B 
 
 data bus 
 
 ci<Ure«S 
 
 confente 
 
 
 
 9 
 
 
 8 
 
 f 1 
 
 - 
 
 ■ 
 
 
 
 91 
 
 f 1 
 
 
 
 
 (or 9 1 
 
 Figure 8. First Solution. 
 
IT 
 
 at any location in the PROMs, which means great ease of maintenance. 
 When new radio stations are constructed, the parameters of these are loaded 
 into the PROMs at the next location without any removal or rearrangement of 
 the data already inside. However, the system has an important drawback in 
 this form: for each radio, we have to scan the whole 256 words among which 
 one or two or occasionally more exist. In the next approach, a solution is 
 proposed in order to avoid this drawback. 
 
 2.3.2.2. Second Solution 
 
 Another solution (with two forms) is now presented which attempts 
 to provide faster scanning of the stations in which we are interested. 
 
 Compared to the preceding one, it does not use a PROM for storing 
 the frequencies and therefore implies that the parameters of the radio sta- 
 tions are ordered in a particular way. 
 
 Two schemes of ordering the data are presented: the first one 
 provides automated scanning while the second is semi-asynchronous. 
 
 First Scheme 
 
 This scheme assumes 8 bit frequencies and that, for a given fre- 
 quency, there exists no more than four stations. Presently this assump- 
 tion proves to be close to the facts if we consider an area of 300 miles 
 radius; using Champaign as center it was found that only one frequency 
 was used by more than four stations. An additional supposition is that no 
 more than 256 stations exist in the zone. A block diagram of the system 
 is shown in Figure 9- 
 
 The 8 bit radio frequency is divided into two parts. The first 
 one, consisting of the 6 most significant bits, is used to address the 6 
 most significant bits of the "address PROMs" (to be defined later). The 
 second part (two least significant bits) is decoded through a 2 to h line 
 
18 
 
 data 
 PROMs 
 
 
 <3<Jir«i 
 
 Same 
 Co<ie 
 
 " 2 
 T 
 
 I 
 
 5 
 
 CHIP CNAS 
 
 u_y 
 
 Figure 9. Second Solution (First Scheme) 
 
19 
 
 decoder. Each of the four lines is used to enable one of the four address 
 PROMs. Now, let us define the address PROMs. They are up to h in number 
 and each of them is selected as indicated above. For addressing, the two 
 least significant bits are generated by the two least significant bits 
 of a 5 bit counter. The three most significant bits of this counter are used 
 to select one of the radios. Consider the case where the counter begins to 
 select one radio. Its corresponding 8 bit frequency is therefore selected 
 by the mean of multiplexers. Bits 7 and 8 determine the address PROM of 
 that frequency. Meanwhile, the two least significant bits of the counter 
 vary from 00 to 11, therefore providing an automatic scanning of four con- 
 secutive words of the enabled address PROM. Each address PROM is loaded 
 with 8 bit words which are the addresses of the PROMs containing the para- 
 meters of the radio staions. The words are organized in groups of four, 
 each group containing the addresses for stations using a given frequency. 
 If only one station is using a frequency, three out of four words will be 
 loaded with dummy data corresponding to a dummy address of the data PROMs. 
 Let us analyze the efficiency of such a system. Its main advan- 
 tages are automation and rapidity. If we consider 8 radios, these are 
 completely scanned in 8 x k = 32 time units compared to the preceding sys- 
 tem which requires 256 x 8 = 20U8 time units. However, the system seems 
 more complex since it requires more integrated circuits. One drawback is 
 the inefficient use of the address PROMs. We have decided to assign four 
 words for each frequency but the average number of stations using one fre- 
 quency is between two and three. This means that only 60% of the words 
 can be used. Assigning three words for each frequency would increase 
 this efficiency, but the system would require more hardware since we are 
 using binary representation. Another drawback is the fact that we have 
 assigned a fixed number, four, for the maximum possible number of stations 
 
20 
 
 using a given frequency. This is the main limitation of the system. How- 
 ever, it has the advantage of simple maintenance since, when we have to 
 load data for new stations , the data already inside is not removed or re- 
 arranged. 
 
 Second Scheme 
 
 Another scheme, which is as fast as the preceding one, but which 
 avoids the limitation on the maximum number of stations using one frequenc 
 is now presented. Refer to Figure 10. 
 
 We suppose, in addition, that we may need up to four address PROMs. 
 Therefore we choose a 10 bit frequency. Eight of them are used to address 
 directly the address PROMs; the other two, using one 2 to h line decoder, 
 provide an individual selection of these PROMs. In addition, the 8 bits 
 provide a code which points to the word containing the first station using 
 a given frequency. When a radio is selected by a set of multiplexers, the 
 8 bit frequency presets an 8 bit counter addressing the "address PROMs" and 
 at the same time enables the clock which is monitoring this counter. The 
 outputs of the address PROMs are fed back into a device which detects a 
 dummy variable used to indicate that there exist no more stations at that 
 frequency. When such a dummy code is detected, the 8 bit counter is 
 stopped while enabling the 3 bit counter which selects one radio among the 
 eight. 
 
 This method has several advantages: it is very fast (faster than 
 the preceding ones); it has good storage efficiency which improves as the 
 number of stations using the same frequency increases (we loose one word 
 for each frequency). However several drawbacks can be pointed out. The 
 first one is the fact that the system is not very automated, needing more 
 hardware and controls. But the major problem with this scheme is its 
 
 , 
 
21 
 
 radio 
 1 
 
 M 
 P 
 X 
 
 M 
 P 
 X 
 
 8 
 
 ourp 
 
 r pu t- 
 
 M 
 P 
 X 
 
 9,10 
 
 counter 
 
 CLK 
 
 JL 
 
 C 
 
 8 
 
 N. 
 
 2to4 
 line 
 
 PROM . 
 
 PROM 
 2 
 
 r 
 
 dat a 
 PROMs 
 
 AdJrto 
 
 "1 
 
 firjr word 
 
 r 
 
 Consent* 
 
 
 
 
 
 
 
 
 //////, 
 
 
 
 SFarions wirh 
 same fr^Q'icncy 
 
 dummy dat J 
 
 Figure 10. Second Solution (Second Scheme) 
 
22 
 
 impractical maintenance. To load a new station at a frequency which is 
 already used, we must load the address in place of the dummy word; this 
 dummy word is written in the next location and the rest of the contents 
 must he shifted down in the address memory, which requires substantial 
 effort. This rearrangement of the contents may he avoided by using the 
 following procedure. Instead of having a dummy code we load the address 
 to which the 8 hit counter has to jump in order to find the next set of 
 stations using the same frequency. This avoids removal of data everytime 
 new stations are loaded in, hut requires more control logic, which makes 
 the system even more complex since it is already not very automated. 
 
 2.3.3. A Third Approach (Figure 11 ) 
 
 In the first approach, a very simple system was proposed. In 
 the second approach, we took into account the fact that more than one sta- 
 tion could use the same frequency. Now, we consider a system which is one 
 more step toward a final design. The system uses the first solution indi- 
 cated in the second approach (the method using a comparator). It was 
 noted that the main drawback of this method was its long cycle time (while 
 possessing ease of revision). A method which avoids this problem, while 
 using the same principle, consists of scanning the entire memory (256 x 8) 
 to determine all the addresses required by the CORDIC unit. These add- 
 resses are loaded into an 8 word RAM one after another until completion. 
 When all radios have been processed, the RAM is cycled, each time provid- 
 ing a station address for each of the radios tuned. When a new radio is 
 tuned (or the frequency changed), we set a flag which indicates that we 
 want to repeat the scanning in order to refresh the addresses contained in 
 the RAM. Let us analyze the operations involved in an entire cycle. When 
 the flag indicating that the operator has selected a radio,,, the first radio 
 
23 
 
 13 
 
 radio 
 1 
 
 flag 
 
 ClotW 
 
 c 
 
 
 
 u 
 N 
 
 T 
 C 
 
 PROM 
 
 Counter 
 
 T 
 
 La 
 
 13 
 
 CPl 
 
 compare k 
 
 13 
 
 3to8 
 
 Counter 
 
 flag 
 
 flag 
 
 RAM 
 
 8x 8 bir 4 
 
 data 
 P ROMs 
 
 data bus 
 
 control bus 
 
 Figure 11. A Third Approach. 
 
2k 
 
 is selected and scanning of the frequency PROMs "begins. When the comparator 
 indicates equality, a flag is set which stops the scanning and activates a 
 control line which causes the address to be saved and signals the CORDIC 
 unit that it must compute the plane to station distance. When finished, a 
 control line clears the flag and the scanning proceeds and so forth. When 
 one cycle has been completed, CORDIC determines the right station (smallest 
 distance) and loads its address into the RAM. When finished, the frequency 
 PROMs are scanned again to determine the next station and so forth until 
 all radios have been taken into account; at that time, we are ready to use 
 the RAM for the cycling which is now of only 8 words. 
 
25 
 
 3. THE DISPLAY 
 
 A single display will be used for the entire system. It is 
 capable of displaying any of the data coming from the 2h bit data bus. 
 The data to be displayed is determined by the operator by means of a 
 selector. This selector provides lines which feed the control bus as 
 indicated in Figure 12. 
 
 The data on the bus is in two's complement representation. In 
 order to drive the display, it must be converted to BCD, with decimal 
 point and sign. 
 
 Let us define the types of data to be displayed: 
 
 - data corresponding to a positive binary number, without any 
 decimal point 
 
 - distances, either positive or negative, involving a decimal 
 point 
 
 - distances to be converted in feet 
 
 - data to be read in angles (37768 _ = 180.00 degrees) 
 
 - the three letters of the Morse code which must appear simul- 
 taneously. 
 
 Thus there are five types of data to be converted to BCD. 
 
 3.1. Conversion to BCD 
 
 The main problem to be solved is the conversion of a binary num- 
 ber, coded in two's complement when negative, into BCD. Several possibili- 
 ties are offered and others have been considered. 
 
 3.1.1. Conversion Using Counting (Figure 13) 
 
 This is the simplest conversion. It uses two sets of counters; 
 The first one is a set of binary counters; the second one uses BCD counters. 
 
26 
 
 selector of data 
 to display 
 
 Figure 12. Block Diagram of the Display. 
 
 binary number 
 
 load 
 
 control 
 
 BINARY COUNTER 
 
 J* up/ down 
 
 zero s 
 
 down 
 
 clear 
 
 to display 
 
 Figure 13. Conversion to BCD Using Counting (Basic Operation) 
 
27 
 
 The conversion procedure works in the following manner: the number to be 
 decoded is loaded into the binary counter while the BCD counter is cleared. 
 Then a clock common to both counters is enabled causing the binary counter 
 to count down and the BCD counter to count up. When all zeros are detected 
 in the binary counter, we stop counting and the number coded in BCD appears 
 at the outputs of the BCD counter. 
 
 The same hardware may also be used to convert a negative number 
 (two's complement) into BCD; in this case, the binary counter counts up 
 instead of down and stops counting when all zeros are detected. 
 
 This system has the advantage of being available for other kinds 
 of operations. For example, as noted, some of the data corresponding to 
 distances expressed in miles may be converted into feet. In the PROMs , 
 the data is stored in binary and corresponds to miles x 100 (e.g. 
 100100001 = 2.89 miles). If we desire to display the distances in feet 
 (a mile is 6000 feet in aviation), we must either multiply by 6000 or by 
 .6 and move the decimal point four places to the right. The factor .6 is 
 chosen so that we are able to use a binary rate multiplier (SN 7^+97) which 
 may be cascaded to obtain better accuracy. Therefore, in order to per- 
 form the previous transformation in feet, the BCD decoder will be provided 
 with a rate multiplier which modifies the clock frequency which operates 
 the BCD counter. This type of converter may be used to display degrees, 
 which also involves a rate multiplier. See Figure Ik . 
 
 This system, although using a small number of chips, has the 
 great disadvantage of being slow, especially if we desire to convert 2h bits 
 
 3.1.2. Conversion Using Combinatorial Logic 
 
 In order to use a converter having no clock and thus very fast 
 conversion speed (determined by the propagation rate through TTL logic), 
 
28 
 
 c loc k 
 
 binary 
 
 load 
 
 number 
 
 up/down 
 
 co'ntrol 
 
 — 4 
 
 BINARY COUNTER 
 
 <- 
 
 I 
 
 all Os 
 
 clear 
 
 rate 
 
 mutiplier 
 7^7— 
 
 r a t e (s) 
 
 dow 
 
 i 
 
 BCD COUNTER 
 
 Figure lU. Conversion to BCD Using Counting (Extended Operations) 
 
 _ binary 
 B 5 B, B, B 4 y B, B, 
 
 SN 74185 
 binary-bed decoder 
 
 f V V v V V V 
 
 BCD 
 
 MSB L S B 
 
 Figure 15. Conversion Using Combinatorial Logic. 
 
29 
 
 the add/subtract-three algorithm implemented with ROMs (SN 7^185) may "be 
 used. However, the number of packages required grows rapidly with the 
 number of bits: k for 9 bits, 8 for 12 bits, l6 for l6 bits, and for 2k 
 bits, the number of packages becomes prohibitive. Figure 15 shows one 
 SN 7^185 ROM which can convert up to 6 bit binary numbers. 
 
 3.1.3- Conversion Using BCD Adders 
 
 Another possible implementation of binary to BCD conversion 
 using combinatorial logic is the following: use of ROMs and BCD adders. 
 Figure l6 indicates the implementation for a l6 bit number. The first 8 
 bits provide the numbers from to 255 » thus requiring 2 ROMs each for 
 the units, tens and hundreds. The next 8 bits address 3 ROMs since we 
 need units to ten thousands. Five BCD adders are required to perform 
 the additions of the units, tens and hundreds provided by the two sets of 
 ROMs. This system can be extended to 2k bits but it requires a large 
 number of ROMs and BCD adders. In addition it is not flexible. 
 
 3.1. U. Conversion Using ROMs and a Counter ( Figures 17 and 18 ) 
 
 This scheme has some similarities with the preceding one, except 
 that it requires more time and can be used for many kinds of transforma- 
 tions. Its principle is the following: the binary number is loaded into 
 a shift register. A set of ROMs (U for 2k bits) contains the BCD code for 
 2,2 ,...2,2. The clock shifts the shift register right and, at 
 the same time, decrements a counter initially set at 2k. For each digit, 
 we use a BCD adder whose inputs come from the ROMs, and whose outputs are 
 one input of a 2 input multiplexer with storage. The second input of this 
 device is 0. The output of the serial output shift register selects either 
 or the data coming from the BCD adder. When the counter reaches 0, the 
 clock is inhibited, and the operation is complete. The system has the 
 
30 
 
 0° 
 
 1 
 
 £ 6 
 
 I' 
 
 il 
 
 212 
 
 2i3 
 2i4 
 ois 
 
 2 ROM 
 
 256X8 
 
 units (4) 
 
 tens (4) 
 
 hundreds 
 
 1 
 
 3 ROM 
 
 256X8 
 
 units 
 
 tens 
 
 hundreds 
 
 thousa 
 
 ten tho 
 
 bed 
 adder 
 
 82S83 
 
 ZZTI 
 
 ds 
 
 usands 
 
 units 
 
 >- 
 
 (4) 
 tens 
 
 (4) 
 hundreds 
 
 (4) 
 
 thousands 
 
 (4) 
 
 ten 
 
 (4) . 
 thousands 
 
 Figure l6. Conversion Using BCD Adders, 
 
2* 
 
 serja 
 out 
 
 1__ 
 
 binary 
 
 shift register 
 
 .clock 
 
 clock 
 
 KD- 
 
 start 
 
 flag 
 
 [ _ 
 clear 
 
 -J 
 
 c 
 o 
 u 
 n 
 
 t 
 e 
 r 
 
 n word s el ecj. 
 
 
 c'l 
 
 31 
 
 2" 
 
 serial in 
 
 ri 
 
 d 
 r 
 
 ^ s 
 
 ROM 
 
 outp uts 
 
 r-=CL 
 
 bed adder 
 
 82 S 83 
 
 turn ouT 
 
 
 
 ock, :Li — ! 
 
 carry 
 
 1 inpiirt •£ 
 
 accumulator 
 
 74 29S 
 
 BCD 
 
 TO DISPLAY 
 
 Figure IT. Conversion Using ROMs and a Counter. 
 
32 
 
 block 
 
 
 select 
 
 
 
 2 5 
 down 
 
 \ 
 
 / 
 
 > 
 
 :> 
 
 z / / -^_ 7^~z : 
 
 :zzzzzzzz^z: 
 
 ^i 
 
 P fl O M 
 
 ■ 1st block 
 2nd block 
 
 Figure 18. Converter Used for Multipurpose Conversions, 
 
33 
 
 great advantage of being useful for binary numbers which are negative and 
 coded in 2's complement. For this case, a 1 appears in the 2Uth "bit. 
 Figure IT indicates an Exclusive OR circuit whose inputs are the outputs of 
 a latch containing the 1 of the 2Uth bit and the current bit out of the 
 shift register. In the ROMs, we use 25 words instead of 2k and the last 
 one contains a one. When the 25th word of the ROMs is addressed, if the 
 number is positive, is added to the accumulator; if negative, the 25th 
 word is added. Another interesting aspect of such a circuit, is the fact 
 that it can be used for the conversion to feet or the angular representa- 
 tion as indicated above. The angular representation is as follows: 
 180° = -180° which corresponds in binary to 100... 0; 90° is obtained by 
 shifting right: 010... and so on. Therefore when 100... appears, we 
 have to display 180.00 (using two decimal digits for greater accuracy). 
 
 3.2. Solution Implemented 
 
 Figure 19 shows a block diagram of the solution which has been 
 implemented. It is a hybrid system, consisting of a set of binary and 
 BCD counters for most of the conversions and of three T^ 185 ROMs when 
 it is desired to simultaneously display the three Morse code letters (the 
 ROMs are used in order to avoid additional logic). 
 
 The system consists of two parts, the conversion circuits and the 
 control logic. 
 
 3.2.1. The Conversion Circuits (Figure 20) 
 
 The circuits for the Morse code letters consist of a set of 
 latches: 2 hex latches (SN 7 1 +17 1 and one quadruple latch (SN 7^ 175). 
 They store the data coming from the data bus. The outputs of the latches 
 are in groups of five and energize the 7^ 185 ROMs used for the binary to 
 BCD conversion. 
 
3k 
 
 MSB 
 
 clock 
 
 IE 
 
 up_ 
 
 fountt 
 
 control logic 
 
 input" Cortt-rols 
 
 ITT 
 
 i 
 
 , Speci<H J 
 I clock 
 
 rr 
 
 tk 
 fare 
 
 Morie 
 
 i letter 
 deaf dfjplay 
 
 down 
 
 carry 
 
 borrow clock 
 
 clear 
 
 down 
 
 i I 
 
 ;i 
 
 bed counter 
 
 MPX8 
 
 I 
 
 LL 
 
 MPX7 
 
 (4) 
 
 47 
 
 Ji 
 
 - MPX6 
 
 (4) 
 
 *4> 
 
 L 
 
 - EI HI 33 
 
 3Q4J — [TI L 3Q2[ - 
 
 bin. 
 bed i 
 decoder 
 
 (4) 
 
 MPX2 
 
 _J 
 
 h MPX1 
 
 (4) 
 
 (4) 
 
 EB 
 
 bed 
 
 7segment 
 
 decoder 
 
 display 
 
 Figure 19. Block Diagram of the Display. 
 
35 
 
 up 
 
 down 
 
 SN 2*193 
 
 data 
 bus 
 
 SN 74193 
 
 load 
 
 I 
 
 clear 
 
 binary counter 
 
 bed counter 
 
 down 
 
 17 
 
 I 
 
 SN 74192 
 
 I 
 
 SN 74192 
 
 2' 
 
 I 
 
 SN 74193 
 
 SN 
 latchels ^74 
 
 decoders 
 
 K 
 
 SN 
 '174 
 
 SN 
 "185 
 
 LI 
 
 Q 
 
 lQ. 
 
 SN 
 '185 
 
 Li 
 
 
 
 Ji 
 
 157 
 
 Ul 
 
 
 
 ii 
 
 157 
 
 47A 
 
 47A 
 
 SN 74192 
 
 u 
 
 '157 
 
 47 A 
 
 multiplexers 
 
 bcd/7 segment decoders 
 
 SN 
 '175 
 
 '185 
 
 
 display 
 
 TIL 302 
 
 Figure 20. Conversion Circuits. 
 
36 
 
 The circuits used for all the other conversions consists of a 
 set of binary counters (SN 7^ 193) and a set of BCD counters (7U 192). 
 
 Quadruple 2 line to 1 line multiplexers (SN T^ 157) choose 
 either the data coming from the BCD counters, or that coming from the 
 binary to BCD ROMs. The multiplexers outputs are fed into BCD to 7 seg- 
 ments decoders and drivers (SN "jh ^-7A). The decoder outputs drive the 
 final stage, the 7 segment numeric display (type TIL 302). An additional 
 circuit, the TIL 30^4 is used in order to indicate the sign. It uses a 
 buffer driver. 
 
 3.2.2. The Control Logic (Figure 21 ) . 
 
 When data are to be displayed, the instruction (a pulse on a given 
 line) is decoded in order to perform particular functions which are of two 
 types depending on whether the counters or the ROMs are to be used. 
 
 If we use the counters, three control lines are needed: 
 
 - counter start : this pulse loads the data into the binary 
 counter, while clearing the BCD counters. Then it enables 
 either the up or down line for the binary counter (for the BCD 
 counter, only the down line is used). 
 
 The carry and borrow lines of the binary counter are used to 
 determine when it is empty. When this is detected, a pulse 
 is produced which clears the flag and therefore disables the 
 counting. 
 
 - special clock : this line is used when the data to be displayed 
 requires a special conversion for angles or conversion of 
 distance to feet). The flag is reset to zero when the binary 
 counter is empty. 
 
37 
 
 Courier 
 start 
 
 Jl 
 
 O 
 
 flag 
 
 II 
 
 clock 
 
 clear 
 
 Special 
 ClocK 
 
 JL 
 
 flag 
 
 V 
 
 Rale 
 
 seled- 
 
 A 
 
 flag 
 
 k 
 
 Morse 
 letter 
 
 line 
 
 JL 
 
 © 
 
 *-flag 
 
 TL 
 
 ® 
 
 flag 
 
 Gear Display 
 
 "*" (5 
 
 MSB 
 
 up/ 
 down 
 
 clock 
 
 i binary 
 3 counter 
 
 i_ 
 
 load 
 
 r— CZP G 
 
 down 
 
 carry 
 
 ,borrow 
 
 single 
 shot 
 
 | BCD counter 
 
 clear 
 
 J* 
 
 JL 
 
 rSZe — 
 nnultipl ier|— ' 
 
 -— i 
 
 multiplexer 
 
 ratel 
 
 rate2 
 
 u 
 
 morse latches 
 '"multiplexers 
 
 ■»j DISPLAY 
 
 Figure 21. Control of the Display. 
 
38 
 
 - rate select line : this line determines which rate must be 
 set into the rate multiplier. The selection between rate 1 
 and rate 2 is done through 2 to 1 line data selectors. 
 
 The flag is cleared when the binary counter is empty. 
 If the Morse code letter circuits are used, one line is required: 
 
 - Morse letter line : this line feeds the clock input of the 
 latches and sets a flag which maintains the select line of 
 the multiplexers which transmit either the outputs from the 
 BCD counters or the outputs of the ROMs at 1. 
 
 A fifth line is used for turning off the display by clearing a flag whose 
 output is fed into the Bl/RBO line of the 7 segment display circuits. 
 
 
39 
 
 k. THE CODE KEYER 
 
 U .1. , Definition of the Problem 
 
 Each radio station has an identifying code consisting of three 
 letters (for Champaign it is CMl) which is transmitted using the Morse 
 code. Using the same example, CMI is coded: 
 
 C MI 
 For the Morse code, the unit time used will he 1/10 second. The dash is 
 coded 3/10 of a second and the dot as 1/10 of second. Within one letter, 
 there is a blank of 1/10 of a second after each dash or dot. A blank sep- 
 arates two consecutive letters and lasts 3/10 of a second. Finally, bet- 
 ween each complete station code, we have to provide a blank of approxima- 
 tively k seconds. Example: 
 
 3/10 1/10 3/10 1/10 
 In addition, we must take into account that this three letter Morse 
 code presents an additional letter, I, which is the same for all stations 
 and which preceeds the 3 letter code (ICMI). The letter I appears when the 
 pilot is approaching the runway. The code is generated by a narrow beam 
 antenna which covers the runway. 
 
 The problem, then, is to generate an audio signal corresponding 
 to the Morse code for a given station when the pilot tunes to its frequency. 
 Since several stations use the same frequency (there are 200 of them avail- 
 able for radio stations at the present time), the C0RDIC unit must deter- 
 mine which address in the PROMs corresponds to the most probable station 
 which is emitting the signal we are receiving. Having done this, it will 
 generate a signal indicating that we must load the code from the PROMs. 
 
ko 
 
 A major consideration for the Code Keyer is the choice of the code 
 used to represent the Morse code. We know that the longest letter is Q 
 which requires 13/10 of a second. 
 
 k.2. 
 
 First Solution 
 
 A possible way of coding is the following: 
 
 00 for 
 
 01 for - 
 
 10 for end of a letter 
 
 11 for end of the whole code 
 
 In addition to this, if after a dot or dash we find either 00 or 
 01 we automatically create a 1/10 of a second "blank. 
 
 Now, we have to look for the longest possible word which contains 
 h letters, the first of which is I and the others with a maximum number of 
 h dashes or dots: 
 
 
 bits 
 
 
 1st letter I 
 
 2 
 
 
 blank 
 
 2 
 
 
 2nd letter 
 
 2 x 
 
 k 
 
 blank 
 
 2 
 
 
 3rd letter 
 
 2 x 
 
 k 
 
 blank 
 
 2 
 
 
 Uth letter 
 
 2 x 
 
 k 
 
 end of code 
 
 2 
 
 
 Therefore, we need 10 + 2k = 3*+ bits or 5 PROMs. (h 1/2). The system 
 would appear as indicated in the Figure 22 for the first solution. This 
 coding seems to be the one which requires the least amount of memory if 
 we want to load the complete code in only one time, without using any 
 other kind of memory such as a ROM. 
 
Ill 
 
 scan 
 
 r 
 
 "V 
 
 ; 
 
 \* 
 
 S 
 
 From 
 
 CORDIC 
 
 A<l«4re« P R M S 
 
 "^ 
 
 Out-puts 
 
 BUS 
 
 code keyer 
 
 1 
 
 ^ 
 
 code keyer 
 2 
 
 audio 
 
 Figure 22. First Solution. 
 
1+2 
 
 The outputs of the PROMs are connected to an 8 x k 1/2 "bus. For 
 each desired radio, we use the same system (noted "code keyer radio i" ) 
 which is enabled by the Cordic unit. The output of each of them delivers 
 the audio signal in an unique loud speaker. Figure 23 presents a more 
 detailed description of each code keyer. When a radio frequency is selected 
 and Cordic has determined the address at which the corresponding data is 
 located, a "data enable" signal is generated for this radio. This signal 
 indicates that the right data is available at the outputs of the PROMs, and 
 therefore on the bus. A enable device loads the information into a 3^ bit 
 shift register and disables the inputs after loading. Then, a decoder 
 analyzes the first two bits of the shift register and generates either a 
 3/10 s or a 1/10 s 5V signal. Next, the decoder acts on the clock 
 inhibit which allows two clock pulses to the shift register, therefore 
 shifting right twice. The decoder analyzes these new right most bits, 
 generating the corresponding signal. If these two bits represent either a 
 dot or a dash, the device automatically generates a one unit time blank. 
 When the code indicating the end of the whole Morse code occurs, the enable 
 device is activated with a delay of h seconds in order to separate the 
 repetitive Morse code. The output of the decoder is fed into an oscillator, 
 then into an amplifier which drives a loud speaker. 
 
 This solution has the advantage of loading the entire Morse code 
 for a given station which helps simplify the amount of control needed. But 
 it has the disadvantage of needing k and a half PROMs which are still quite 
 expensive (about 60 dollars each). 
 
 U.3. Second Solution 
 
 A cheaper way, which uses more logic is now presented. This solu- 
 tion requires the smallest possible number of PROMs. Each letter of the 
 
1+3 
 
 
 BUS 
 
 
 
 
 
 
 
 enable 
 device 
 
 
 SHIFT REGISTER ; 2 1 ! 2° 
 
 i I 
 * • 
 
 
 
 
 
 * 
 
 
 
 
 M 
 
 1 
 
 \ 
 
 1 i 
 
 i 
 
 
 
 
 clock 
 inhibft 
 
 
 
 ' 
 
 
 
 
 ^"" — -^^ c 
 
 Jecoder 
 
 OIL 
 
 oscillator 
 
 
 
 
 — 
 
 
 
 
 
 
 
 , 
 
 i 
 
 
 CLOCK 
 1 krz 
 
 \ A 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Figure 23. Block Diagram of the Code Keyer (Solution l) 
 
kk 
 
 code is assigned a binary representation. Since there are 26 possible let- 
 ters, the required number of bits is 5. Then we need 1+3x5 bits = 
 l6 bits which requires two PROMs. In addition to this we need a Morse de- 
 coder. Such a system is shown in Figure 2k. This system has several 
 drawbacks. One of them is the fact that it is not possible to load the 
 entire code in one operation from the two PROMs. This is due to the fact 
 that the length of the coding of a Morse letter is variable. Thus we 
 have to take the largest size (13 bits), which leaves unacceptable blanks. 
 To avoid this we load only one letter at a time, send it to the code keyer 
 and, when the corresponding audio signal has been generated, a flag enables 
 the load of the next letter. The time to load it is at most the cycle 
 time to scan the PROMs. This is on the order of one microsecond, which is 
 negligible compared to a 1/10 of a second. The system works as follows: 
 the enable signal indicates that data is available at the outputs of the 
 PROMs. The letter select enables only one letter (5 bits). This letter 
 addresses a Read Only Memory which contains the Morse alphabet whose cod- 
 ins is: 
 
 111 
 
 for 
 
 dash 
 
 1 
 
 for 
 
 dot 
 
 
 
 for 
 
 blank 
 
 For example, C would be coded as: 11101011101. 
 
 The resulting output is loaded into the right shift register, which 
 is shifted right every 1/10 of a second, therefore producing a 3/10 of a 
 second signal when three consecutive ones are found or a 1/10 of a second 
 blank when zero is found. When the shift register is emptied of its useful 
 information, we select the next letter which is loaded when data is deter- 
 mined to be present. This explains the operation of the system. In the 
 following pages the various functions will be presented in more detail. 
 
U5 
 
 scan 
 
 ■£-^ 
 
 V 
 
 2 
 
 P R O M s 
 
 address 
 
 multiplexer and 
 morse decoder 
 
 CORDIC 
 
 audio 
 
 Figure 2k. Block Diagram of Solution 2. 
 

 he 
 
 k.k. Multiplexing and Morse Decoder 
 
 Figure 25 shows the block diagram of the hardware which is shared 
 by all the radios. It contains a selector which enables only one five bit 
 letter among four. In fact, the first letter is coded 1 if there is a I, 
 if there is not, requiring only one bit in the PROMs. This bit is con- 
 nected to, the first input of the five multiplexers. When the multiplexer 
 address is 00, the outputs are either 11111 or 00000 depending on whether 
 there is an I or not. For the other letters, each bit is connected to the 
 same input of each of the five multiplexers. Their outputs address two ROMs. 
 Since the maximum information required to load a letter using the code as 
 indicated above is 13 bits and the number of letters being 26, the use of 
 two 32 x 8 bit ROMs satisfies the requirements (INTEL IM 5600). The out- 
 puts are open collector type. See Figure 26. 
 
 U.5. Code Keyer Card 
 
 For each radio, we need one card of this type which is indepen- 
 dent from the others and has its own clock. Figure 27 shows the various 
 parts of the code keyer card: 
 
 - shift register, data enable, clock inhibit 
 
 - 10 hz clock, divider 
 
 - end of letter detector 
 
 - letter counter 
 
 - end of code detector 
 
 - multiplexer address selector 
 
 - jump if no I 
 
 - sequence device 
 
 - initial clear device 
 
ini letter 
 
 PROM 
 2 
 
 3rd lerr«r 
 
 1 LETTER SELECT 
 
 (MULTIPLEXER) 
 
 MORSE 
 alphabet 
 
 (ROM) 
 
 16 bits 
 
 SHIFT REGISTERS 
 
 hi 
 
 Figure 25. Block Diagram of the Letter Select and Morse Decoder. 
 
H8 
 
 FROM 
 CK2 
 
 Ycc 
 
 ROM 1 
 
 IM 5600 
 ROM 2 
 
 TO CARD CK2 
 
 v4 , \J/\l'\J/\l/\i'v 
 
 Figure 26. Card CK1: Implementation of the Letter Select and Morse Decoder, 
 
U9 
 
 ready 
 
 DATA 
 enable 
 
 CLOCK 
 inhi- 
 bit 
 7T" 
 
 CLEAR 
 
 from &ON{ 
 
 16 bit SHIFT REGSIER 
 
 10Hz 
 
 clocK 
 
 end of letter 
 detector 
 
 end of 
 code 
 
 sequence 
 device 
 
 1Kz 
 
 ^F 
 
 etter 
 1 — Hcounter 
 
 data 
 out 
 
 multiplexer 
 address 
 
 Iorrol 
 
 Figure 27. Block Diagram of the Code Keyer. 
 
50 
 
 Before giving details concerning each of the above functions, the general 
 processing of the data through them will "be discussed. See Figure 28. 
 The input signals are: 
 
 - l6 data hits coming from the Morse ROMs. 
 
 - the first bit of the PROMs which indicates the presence or 
 absence of the letter I. 
 
 - two control signals : 
 
 "data ready," which indicates that the data at the outputs of 
 the PROMs are those corresponding to our radio station. 
 
 "key on," which indicates that a radio has been set at a 
 given frequency and that the pilot desires to hear its corres- 
 ponding Morse code. 
 The output signals are: 
 
 - "data out," which is the output of the shift register, modulated 
 with a 1 khz digital clock. 
 
 - the two line address of the multiplexers, for selecting the next 
 letter to be loaded in the shift register. 
 
 Suppose that the pilot tunes a radio. The signal "key on" goes to 
 zero. This makes the Clear Device clear the general flipflops contained in 
 the system. When the Data Ready pulse arrives, it causes the Data Enable 
 circuit to load the data coming from the ROMs into the shift registers. 
 When assured of the loading, the Clock Inhibit enables the clock to the 
 shift registers at a frequency of 10 Hz. - The right most bit of the shift 
 register is fed into a device clocked at the same frequency which generates 
 a signal when all the code has been shifted out for a given letter (this cir- 
 cuit is called the "end of letter detector"). Each of these pulses increments 
 a binary counter which counts the number of letters which have been pro- 
 cessed. Its outputs are also used for other purposes. The first is to 
 
 1 
 
51 
 
 1- 
 
 1- 
 
 data ena 
 
 [ -j— |cl»ek(iOHQ 
 
 register 
 
 clock 
 inhibit 
 U flag 
 
 audio 
 
 end of character detector 
 
 -^^—J '16 7 
 
 clocks 
 
 to 
 
 • 1 • 
 
 123 123 
 
 fclrih. 
 
 letter 
 counter 
 
 end of 
 code 
 
 167 
 
 l|OH, 
 
 t- _ _ 
 
 rpy^*- 
 
 B 
 
 27 
 
 Ktf i 
 
 u-J 'multiplexer 
 
 naif 
 
 address 
 
 3**> 
 
 i __ 
 
 HI 
 
 =p 
 
 W>- 
 
 sequence device 
 
 I • 
 
 tnJ»'< 
 
 jumpif no I 
 
 ^ 
 
 Figure 28. Circuit Diagram of the Code Keyer. 
 
52 
 
 address the one letter select multiplexers. Another function is to inform 
 the 'jump if no I" circuit when the next letter to be loaded is 00 or the 
 first bit of the PROMs. If this bit is zero, meaning that there is no I, 
 the letter counter will be set at 01 in order to be ready to address the 
 next letter (first letter in this case since there is no I). One pecularity 
 of the repetition of the identification Morse code for a station is the 
 following. If there is an I, we generate the code in a continuous manner. 
 If there is no I, we send the code four times followed by a blank whose 
 time length is the same as the code itself. All these required sequences 
 are determined by the "sequence device" which uses the data from the 
 first bit of the PROMs, the "jump if no I" circuit and the outputs of 
 the letter counter. In addition to this, a circuit (end of code) disables 
 the loading during k seconds, after a complete code has been generated. 
 
 In the following, the different devices which constitute the Code 
 Keyer are discussed. 
 
 k. 6. Code Keyer Circuits 
 
 U. 6.1. Clear Device (Figure 29) 
 
 When the radio has been tuned, the signal "key on" is set at 
 zero. At this time a narrow pulse is generated which resets the flipflop 
 devices. The element cleared at this time are: 
 
 - letter counter 
 
 - end of code detector 
 
 - clock inhibit flag 
 
 - data enable flag 
 
 The circuit used is anSN 7^ 123 monostable multivibrator with clear, an 
 external resistor, a diode and a capacitor. 
 
53 
 
 1Ka. .-° 1 ' F 
 
 m 
 
 SN 74123 
 
 Key on 
 1 
 
 if 
 
 £=D .5 
 
 v« r 
 
 Clear 
 
 Figure 29 . Clear Device 
 
 Vcc 
 
 i 
 
 05^>f 
 
 '05^ 
 
 H 
 
 6^ 
 
 C-.54^F 
 
 Vcc 
 
 '05 
 
 J74l67 e „ 
 
 CUck output- 
 
 74I67 
 
 CI«cK 
 
 Yo-t>«r 
 
 lOHa 
 
 Figure 30. 1 Khz and 10 Hz Clocks. 
 
5* 
 
 U.6.2. The 1 kHz and 10 Hz Clocks (Figure 30) 
 
 Each code keyer card is equipped with its own clock. This must 
 he a low frequency, 10 Hz . Since we need an audio signal of 1 kHz to 
 modulate the signal coming out of the shift register, we design a 1 kHz 
 clock and obtain the 10 Hz by frequency division. The circuit used to 
 generate the 1 kHz is the one indicated on Figure 30. It uses three cas- 
 caded open collector inverters connected together to form a loop. The 
 signal obtained is such that the off to on time ratio is 5/6 and the fre- 
 quency is given by: 
 
 1 1 
 
 f = 
 
 t l +t 2 V 1+6 /5) 
 
 V - V - V 
 
 t - PR T..f CC BE ° ESat ^ 
 t - CR^ In( ) 
 
 CC BE t 
 
 ' t 2 = C Rl ln( y Vt _V CE -. a v } 
 CC BE t 
 
 This circuit has the drawback of being dependent on the source voltage V . 
 For a capacitor C = . 5yF we obtain the desired frequency. In order to im- 
 prove the shape of the output signal, an additional open collector inverter 
 is used. This yields very short transition times. 
 
 To obtain the 10 Hz clock, we divide this frequency two consecutive 
 times by 10. The circuits used are two SN "jh 167 synchronous and program- 
 mable counters which are easily cascaded by connecting the enable output to 
 the enable and strobe inputs of the next stage. 
 
 k.6.3. Shift Register - Data Enable Flag - Clock Inhibit Flag (Figure 31 ) 
 The data coming from the Morse ROMs are connected to the parallel 
 inputs of two serial shift registers. Loading is controlled by the load/ 
 shift input of the shift register. 
 
55 
 
 from 
 ROMs 
 
 DATA ENABLE FLAG 
 
 CLOCK INHIBIT 
 FLAG 
 
 Clear 
 
 Figure 31. Shift Register-Data Enable and Clock Inhibit Flags 
 
56 
 
 When a radio is tuned, the clear device sends a negative pulse 
 which clears flipflop 1 making the Q output 0. When "data ready" is 
 present (meaning that the code corresponding to our radio is available at 
 the outputs of the PROMs), a positive pulse, whose length is the same as 
 "data ready," appears at the output of the three input NOR gate 2. This 
 pulse, through an inverter, holds the load/shift line of the shift regis- 
 ters to zero for a period equal to the length of the pulse, causing the 
 data at the parallel inputs to be loaded in. The falling edge of this 
 same positive pulse, sets the output of flipflip 1 to 1, disabling the 
 load line prior to the next "data ready pulse." At the same time (falling 
 edge), the output of flipflop 2 changes state and enables the shift register 
 clock, making it start. Flipflop 2 is initially cleared, disabling the 
 shift register clock. It is also cleared when the end of character signal 
 is one, indicating that the data in the shift registers has been unloaded 
 and that shifting should stop to prepare for another load. Similarly, the 
 loading is enabled by the falling edge of the end of character signal or by 
 the end of code signal which occurs after an entire Morse code has been 
 generated. 
 
 h.6.h. End of Character, End of Code Detectors - Letter Counter 
 
 The "end of character detector" detects when the end of the data 
 contained in the shift register occurs. Since the data is either 111 or 1 
 or a single 0, end of data occurs when two consecutive zeros follow a one. 
 Figure 32 shows this logic. It consists of three cascaded flipflops whose 
 common clock is the 10 Hz one. The input is the serial output of the shift 
 register. In order to determine the sequence 100 we enter the Q input of 
 the first two flipflops and the complemented Q of the third one into a NOR 
 gate which generates a positive pulse when the desired sequence occurs. 
 
57 
 
 letter counter 
 
 J Q 
 '111 
 
 c 
 
 K Q 
 
 from Clear 
 
 shift register -jq 
 
 J O 
 '111 
 
 ic 
 
 K O 
 
 JL 
 
 J O 
 '111 
 
 c 
 
 K O 
 
 J 
 
 27 
 
 JL 
 
 end of 
 character -H K O 
 
 Hz 
 
 end of character detector 
 
 HJ o- 
 
 '111 
 c 
 
 '00 
 
 ~L 
 
 1 1 
 
 end of 
 code 
 
 end of code detector 
 
 J O 
 '111 
 
 c 
 
 K O 
 
 t-'-Qj 
 
 Pi 
 
 00 
 
 Figure 32. End of Character Detector-End of Code Detector-Letter Counter. 
 
58 
 
 The falling edge of this pulse triggers a two hit counter which memorizes 
 the next letter taken from the Morse code. When the outputs Ql and Q2 are 
 one (indicating that the letter on the way is the last one) and when the 
 end of character signal is one (meaning the code is completed) a falling 
 edge triggers a 20 ns single shot (monostable mult ivi Drat or) . The falling 
 edge of this short pulse triggers another single shot which has a pulse 
 length of k seconds which is the blank required between two consecutive 
 codes. This pulse, called the "end of code," disables the "data enable" 
 circuit for h seconds, as well as the clock inhibit. 
 
 k.6.5. Multiplexer Address Selector (Figure 33) 
 
 As indicated by its name, this circuit provides the selection of 
 the address of the desired letter to be loaded into the shift registers. 
 
 This address consists of two bits which are 1 when one or more 
 radios are not set. The selector uses open collector components so that the 
 address select for all radios are wire ORed. 
 
 The following signals are required to load the letter address onto 
 the bus : 
 
 - a "data ready" pulse corresponding to the desired radio, 
 
 - "key on" set, indicating that this radio is used. 
 
 The result, when these two conditions are met, is to transfer to card CK1 
 the values which are at the outputs of the letter counter. The two inver- 
 ters used to complement the outputs of the NOR gates are open collector 
 type. If either "data ready" or "key on" or both are 1, the outputs are 1. 
 
 h.6.6. Jump if No I (Figure 3*0 
 
 When the letter counter is at 00, we are ready to load the first 
 letter of the Morse code (I or no I). When the signals "data ready" and 
 "key on" are set, we check the first letter: 
 
59 
 
 , CARD CK1 
 
 Daha ra»4y Key on 
 
 Figure 33. Multiplexer Address Selector. 
 
 Q,~ 
 
 b -- 
 
 IT 
 
 30 
 
 r 
 
 
 jump 
 if nol 
 
 Figure 3^. Jump if no I. 
 
6o 
 
 - if it is (no I) we set the letter counter to 01 in order to 
 load the next letter (therefore, only Ql is changed). 
 
 - if it is 1 (there is an I), no particular action is done and 
 we load it as any other letter. 
 
 U.6.T. Sequence Device (Figure 35) 
 
 The purpose of this circuit is to transmit the Morse code in a 
 continuous manner if there is an I. If there is no I, the code is trans- 
 mitted four times followed by a blank which lasts the same time as the 
 code itself. 
 
 In order to realize this function, we use the output of the logic 
 noted "jump if no I"' which generates a negative pulse when no I is found. 
 The circuit is shown in Figure 35. The flipflop shown is initially cleared. 
 Its Q output is fed back to the clock input together with the jump if no I 
 pulse. When this pulse arrives, its rising edge changes the output of flip- 
 flop 1 and, due to the feedback, this output will stay at one as long as we 
 do not change the frequency of the radio. When the outputs of the letter 
 counter go to 00, the NOR gate output changes to 1. When Ql goes to 1, 
 the output of the NOR gate falls to and this transition increments an 
 SN 7^+93 three bit counter. When its Q Q~ 0_ outputs are 100 (meaning that 
 the Morse code has been sent four times), a NOR gate which has these sig- 
 nals as inputs disables the output of the shift register and a blank is 
 generated. After detection of the 100 output, this counter is reset to 000. 
 
6i 
 
 SEQUENCE DEVICE 
 
 Figure 35- Sequence Device. 
 
62 
 
 5. CONCLUSION 
 
 Some operations of the Radio Aids Simulator have been discussed 
 in the preceding pages. A prototype Code Keyer system was fabricated and 
 proved operational. 
 
 The implementation of the system presents several unique features. 
 The main one is the large use of ROMs, not only for storing data, "but 
 also for reducing the amount of combinatorial logic in the control systems. 
 This is made possible by the constant decrease in the price of integrated 
 circuits, while they are simultaneously becoming faster and smaller. In 
 addition, each year new integrated circuits become available which imple- 
 ment even more complex functions. 
 
 For each function of the Radio Aids discussed in this paper, 
 several solutions using digital circuits have been presented. This estab- 
 lishes that the Radio Aids can be implemented in digital form using a 
 special purpose computer, and provides a partial solution to one of the 
 major problems in flight simulation, the need for large digital computers. 
 
 
 
63 
 
 LIST OF REFERENCES 
 
 1. Texas Instruments Incorporated, "The TTL Data Book for Design Engineers," 
 
 hy the Engineering Staff of T. I. 
 
 2. Kohavi , "Switching and Finite Automata Theory," McGraw Hill Computer 
 
 Science Series, 1970. 
 
 3. Poppelbaum, W. J. , "Computer Hardware Theory," Macmillan, 1972. 
 
 h. Institute of Aviation - University of Illinois, "Fundamentals of 
 Aviation and Space Technology," Institute of Aviation of the 
 University of Illinois, 1971- 
 
 5. D. A. Hodges, "Semiconductors Memories," IEEE Computer Society, 1972. 
 
 6. Millman and Tauh, "Pulse, Digital and Switching Waveforms", McGraw Hill, 
 
 1965. 
 
IOGRAPHIC DATA 
 
 T 
 
 1. Report No. 
 
 UIUCDCS-R-7U-633 
 
 3. Recipient's Accession No. 
 
 C DIGITAL RADIO AIDS FOR A FLIGHT SIMULATOR 
 RADIO MEMORIES - DISPLAY - CODE KEYER 
 
 5. Report Date 
 
 May 197 ^ 
 
 hor(s) 
 
 Lucien Isidore Facchin 
 
 8. Performing Organization Rept. 
 No - UIUCDCS-R-7U-633 
 
 funning Organization Name and Address 
 
 Department of Computer Science 
 
 University of Illinois at Urbana-Champaign 
 
 Urbana, Illinois 6l801 
 
 10. Project/1 ask/Work Unit No. 
 
 11. Contract /Grant No. 
 
 insonng Organization Name and Address 
 
 Department of Computer Science 
 
 University of Illinois at Urbana-Champaign 
 
 Urbana, Illinois 6l801 
 
 13. Type of Report & Period 
 Covered 
 
 technical 
 
 14. 
 
 jblcmemary Notes 
 
 The research described here is part of a joint study between the Department of 
 mter Science and the Institute of Aviation of the University of Illinois. This 
 Ly is a result of the need to consider alternatives to using general purpose digital 
 niters for the solution of the entire aircraft simulation problem. The software 
 .s of such systems have sky-rocketed in recent years while hardware costs continue 
 'all. Thus, serious consideration must be given to direct functional mechanization 
 >ugh the use of special purpose computers. The use of small dedicated computers may 
 itly simplify interfacing to the central control computer in addition to reducing 
 amount of software required by difusing the processing throughout the simulator. 
 ; document presents an implementation of some of the functions of such a system, the 
 .0 Aids System. 
 
 y Words and Document Analysis. 17a. Descriptors 
 
 .tal radio 
 
 ;ht simulation 
 
 .o aids 
 
 lojititier.s 'Open-Kndcd Terms 
 
 <*A I ' |-i,-|d/(.r,.i 
 
 111 i lilt; Statement 
 
 unlimited distribution 
 
 'tlV IS ( 10- 70) 
 
 19. Security ( lass (This 
 Report ) 
 
 UNCI.ASSIHKD 
 
 20. Security (lass (Thi 
 
 Page 
 UN( LASSIFIED 
 
 21. No. ot Page 
 
 ges 
 
 12 
 
 22. Pric 
 
 USCOMM-OC 40329-^7 1 
 


 to 
 

 
 JVS* 
 
 M*