A microprocessor is an information device that, according to a program specified by control signals, processes this information, that is, it implements input, output, storage information and performs arithmetic and logical operations.
Simplified structural scheme one of the microprocessors is shown in Fig. 10.37 and consists of a control unit (CU), an eight-bit arithmetic logic unit (ALU) and a set of n parallel registers of m general purpose bits (POH), designed to store binary numbers used in the calculation process . The microprocessor also includes two parallel buffer registers (BR) designed for short-term storage of numbers (A) and (B) during the operation (ALU).
The control device (CU) with firmware for individual operations sets the operating modes in all elements of the microprocessor. During the operation of the microprocessor, the numbers (A) and (B) on which the operation is performed are transferred via the highway from the registers (RON) to the buffer registers (BRA) and (BRV). Then, at the command from the control unit (CU), the arithmetic logic unit (ALU) performs the specified operation, and its result (F) according to
highway is transferred to registers (RON), in which the previously recorded number is erased. For example, the addition of three numbers is performed in this way: first, the first two numbers are added and the result is recorded in RON. Then the device (ALU) receives the result of the addition and the third number, as a result of their addition, the final result is written to the registers (RON).
Fig.10.37. Simplified block diagram of a microprocessor
In order for the microprocessor (MP) to perform its functions, additional devices are required, which are shown in Fig. 10.38 and constitute a microprocessor system or microcomputer.
Fig.10.38. microprocessor system
The microprocessor system contains a memory (ROM) and (RAM) designed to store information. (ROM) are read-only memories containing permanent information that can only be read using commands (K). (RAM) is random access memory that stores programs, that is, information that can be repeatedly written and read during the execution of the program, in the form of data exchange (D). Information in memory is located in cells, each of which has its own address (A). The data (D) is fed to the inputs of the information input device (IUV), and the information is read from the output device (Alas). The microprocessor (MP) is connected by addresses (A) with (ROM) and (UVV). The input of information in microprocessor systems is carried out from the keyboard, sensors of technological parameters with digital outputs, from photoreaders, and the output of information is carried out by means of registers. Input-output devices are a combination of registers, amplifiers and keys. The interaction of microprocessor systems with each other and with external devices is carried out using special hardware that obeys the commands of the central processor.
Digital microcircuits have by now achieved impressive performance at an acceptable current consumption. The fastest of the digital circuits have a switching speed of the order of 3..5 ns. (chip series 74ALS). At the same time, you have to pay for the speed of microcircuits with an increased current consumption. The exceptions are microcircuits built on the basis of CMOS technology (for example, microcircuits of the 1564, 74HC, 74AHC series). In these microcircuits, the current consumption is directly proportional to the switching speed of the logic gates in the microcircuit. Those. the microcircuit automatically increases the current consumption if it requires more speed, therefore, at present, the vast majority of microcircuits are produced using this technology.
Often digital devices do enough challenging tasks. The question arises - since the microcircuits have reached such a high speed, is it possible to use the same microcircuit repeatedly? Then it will be possible to exchange the speed of microcircuits for the complexity of the problem being solved. It is this exchange that microprocessors allow. These microcircuits repeatedly use the same device - ALU (arithmetic logic unit). Therefore, it is possible to exchange the maximum speed of the microcontroller for the complexity of the implemented device. It is for this reason that they try to maximize the speed of microprocessors - this allows you to implement more and more complex devices in the same volume.
Another reason for the widespread use of microprocessors was that the microprocessor is a universal microcircuit that can perform almost any function. Universality provides a wide demand for these microcircuits, which means mass production. The cost of microcircuits is inversely proportional to their mass production, that is, microprocessors become cheap microcircuits and thereby increase demand even more.
To the greatest extent, all of the above properties are manifested in single-chip microcomputers or, as they are more often called by their field of application: microcontrollers. In microcontrollers, all the components of a computer are combined on a single chip: a microprocessor (often called the core of a microcontroller), RAM, ROM, timers, and I / O ports.
Findings:
CMOS technology allows you to exchange the speed of operation for the current consumed (the faster the logical elements of the microcircuit are switched, the more current the microcircuit consumes);
Microcontrollers make it possible to implement a control scheme of almost any complexity on a single universal microcircuit;
Microcontrollers allow you to exchange the speed of their work for the complexity of the designed device.
Microcontrollers make it possible to implement equipment with minimal cost, dimensions and current consumption.
The term for the development of equipment on microcontrollers is minimal.
Modernization of equipment consists in changing the control program.
Principles of construction of MP-systems.
The core of any microprocessor system is a microprocessor or simply a processor (from the English processor). It is most correct to translate this word into Russian as a “processor”, since it is the microprocessor that is the node, the block that performs all the processing of information within the microprocessor system.
microprocessor we will call a program-controlled device designed to process digital information and generate signals that control this processing.
The remaining nodes perform only auxiliary functions: storing information (including control information, that is, programs), communicating with external devices, communicating with the user, etc. The processor replaces virtually all of the "hard logic" that would be needed in a traditional digital system. It performs arithmetic functions (addition, multiplication, etc.), logical functions (shift, comparison, code masking, etc.), temporary storage of codes (in internal registers), transfer of codes between microprocessor system nodes, and much more. The number of such elementary operations performed by the processor can reach several hundred. The processor can be compared to the brain of the system. But at the same time, it must be borne in mind that the processor performs all its operations successively, that is, one after the other, in turn. Of course there are processors with parallel execution of some operations, there are also microprocessor systems in which several processors work on the same task in parallel, but these are rare exceptions. On the one hand, the sequential execution of operations is an undoubted advantage, since it allows using just one processor to execute any, the most complex information processing algorithms. But, on the other hand, the sequential execution of operations leads to the fact that the execution time of the algorithm depends on its complexity. Simple algorithms run faster than complex ones. That is, a microprocessor system is able to do everything, but it does not work too fast, because all information flows have to be passed through a single node - a microprocessor (Fig. 1.3). In a traditional digital system, it is easy to organize parallel processing of all information flows, however, at the cost of complicating the circuit.
manager
information
(program)
Rice. 1.3. Information flows in a microprocessor system.
So, the microprocessor is capable of performing many operations. But how does he know what operation he needs to perform in this moment? This is what is determined control information, program.
A program is a set of commands (instructions) that is, digital codes, deciphering which, the processor will know what it needs to do. The program from beginning to end is compiled by a person, a programmer, and the processor acts as an obedient executor of this program, it does not show any initiative (unless, of course, it is in good order). Therefore, comparing the processor with the brain is not very correct. He is just an executor of the algorithm that a person has compiled for him in advance. Any deviation from this algorithm can only be caused by a malfunction of the processor or some other components of the microprocessor system.
All instructions executed by the processor form the command processor system. The structure and volume of the processor instruction set determine its speed, flexibility, and ease of use. In total, the processor can have from several tens to several hundreds of commands. The instruction system can be designed for a narrow range of tasks to be solved (for specialized processors) or for the widest possible range of tasks (for universal processors). Command codes can have a different number of digits (occupy from one to several bytes). Each command has its own execution time, so the execution time of the entire program depends not only on the number of commands in the program, but also on which commands are used.
To execute instructions, the processor structure includes internal registers, an arithmetic logic unit (ALU, ALU - Arithmetic Logic Unit), multiplexers, buffers, registers and other nodes. The operation of all nodes is synchronized by a common external clock signal of the processor. That is, the processor is a rather complex digital device (Fig. 1.4).
Rice. 1.4. An example of the structure of the simplest processor.
However, for the developer of microprocessor systems, information about the intricacies of the internal structure of the processor is not very important. The developer must consider the processor as a "black box", which, in response to input and control codes, performs one or another operation and produces output signals.
The developer needs to know the command system, processor operating modes, as well as the rules for the interaction of the processor with the outside world, or, as they are also called, information exchange protocols .
About the internal structure of the processor, you need to know only what is necessary to select a particular instruction, a particular mode of operation.
Microprocessor system - this is a computing, control and measuring or control system in which the main information processing device is the MP. The microprocessor system is built from a set of microprocessor LSI.
The construction of MPS systems is based on three principles: backbone; modularity; microprogram control.
Trunk principle determines the nature of the connections between the functional blocks of the MPS - all blocks are connected to a single system bus.
Modularity principle is that the system is built on the basis of a limited number of types of structurally and functionally completed modules. Each module of the MPS system has a third (high-impedance) state control input. This entry is called СS (Сhір Select) - choice of crystal or OE (Output Enabie) - exit permission.
The action of the CS signal for the trigger is shown in fig. 1.5. Initial trigger signal Q will appear on the output only when active (in in this case - zero) signal level CS. If CS = 1, the flip-flop is placed in a high-impedance state. The trigger output is tri-stable, that is, it can be in one of three states: logic one, logic zero, or high-impedance. At each moment of time, only two modules are connected to the MPS system bus - the one that receives and the one that transmits information. Others are in a high impedance state.
The principles of backbone and modularity make it possible to increase the control and computing capabilities of the MP by adding other modules.
Firmware control principle consists in the possibility of performing elementary operations - microinstructions (shift, information transfer, logical operations). With a certain combination of microinstructions, you can create a set of commands that will best meet the purpose of the system, that is, create a technological language.
Consider a generalized block diagram of the MPS (Fig. 1.6.) The MPS includes: the central processing unit (CPU), PZP, OZP; interrupt system, timer, UVV. I/O devices are connected to the system bus via I/O interfaces.
Read-only memory and random access memory make up a memory system designed to store information in the form of binary numbers. Read-only memory is designed to store programs, tables, constants.
Random access memory - for storing intermediate results of calculations. Memory is organized as an array of cells, each of which has its own address and contains a byte or a word.
The CPU module processes data and manages all other modules in the system. The central processor, in addition to the LSI MP, contains synchronization circuits and an interface with the system bus. It fetches command codes from memory, decrypts them, and executes them. During the execution of an instruction - an instruction cycle - the CPU performs the following actions:
Sets the address of the instruction on the address bus AB;
Gets the command code from memory and decrypts it;
Calculates operand addresses and reads data;
Performs the operation specified by the command;
Accepts external control signals (for example, interrupt request);
Generates the status and control signals necessary for memory operation
and UVV.
I / O devices or external devices are devices designed to enter information into the MP or output information from it. Examples of ICDs are displays, printers, keyboards, digital-to-analog and analog-to-digital converters, relays, and switches. To connect air-blasters to the system bus, their signals must meet certain standards. This is achieved using I/O interfaces.
The input-output interfaces perform the function of matching the air-blast signals with the signals of the MP system bus. They are also called controllers or adapters. The microprocessor accesses the interfaces using special I/O commands. At the same time, the MP puts addresses on the bus AB interface address, and on the data bus DB reads data from an input device or writes to an output device. On fig. 1.6 shows one input interface and one output interface.
The interrupt system allows the MPS to respond to external signals - interrupt requests, the sources of which can be: readiness signals from external devices, signals from generators, signals from sensor outputs. When an interrupt request occurs, the CPU interrupts the main program and proceeds to execute the interrupt request service routine. To build an interrupt system, MPCs contain LSI of special programmable interrupt controllers.
The timer is designed to implement functions related to timing. After the MP loads into the timer a number that specifies the frequency, delay or division factor, the timer implements the desired function on its own.
The use of microprocessor systems in almost all electrical devices is the most important feature of the technical infrastructure modern society. Electric power industry, industry, transport, communication systems are significantly dependent on computer control systems. Microprocessor systems are built into measuring instruments, electrical devices, lighting installations, etc.
All this obliges an electrician to know at least the basics of microprocessor technology.
Designed for automation of information processing and management of various processes.
The concept of "Microprocessor system" is very broad and combines such concepts as "electronic computer (computer)", "control computer", "computer", etc.
The microprocessor system includes Hardware or in English - hardware and software(software) - software.
digital information
The microprocessor system works with digital information, which is a sequence of digital codes.
At the heart of any microprocessor system is a microprocessor that can only perceive binary numbers (made up of 0 and 1). Binary numbers are written using the binary number system. For example, in Everyday life we use the decimal number system, in which ten characters or numbers are used to write numbers 0,1,2,3,4,5,6,7,8,9. Accordingly, in the binary system there are only two such symbols (or numbers) - 0 and 1.
It must be understood that the number system is just the rules for writing numbers, and the choice of the type of system will be determined by ease of use. The choice of the binary system is due to its simplicity, which means the reliability of digital devices and the ease of their technical implementation.
Consider the units of digital information:
A bit (from the English "BInary digiT" - a binary digit) takes only two values: 0 or 1. You can encode the logical value "yes" or "no", the state "on" or "off", the state "open" or "closed " etc.
A group of eight bits is called a byte, for example 10010111. One byte allows you to encode 256 values: 00000000 - 0, 11111111 - 255.
A bit is the smallest unit of information representation.
A byte is the smallest unit of information processing.A byte is a part of a machine word, usually consisting of 8 bits and used as a unit of the amount of information during its storage, transmission and processing on a computer. A byte is used to represent letters, syllables and special characters (usually occupying all 8 bits) or decimal digits (2 digits per byte).
Two interconnected bytes are called a word, 4 bytes are called a double word, 8 bytes are called a quadruple word.
Almost all the information that surrounds us is analog. Therefore, before information enters the processor for processing, it is converted by means of an ADC (analogue-to-digital converter). In addition, information is encoded in a specific format and can be numeric, logical, text (character), graphic, video, etc.
For example, to encode textual information, the ASCII code table (from the English American Standard Code for Information Interchange - American Standard Code for Information Interchange) is used. One character is written in one byte, which can take 256 values. Graphic information is divided into dots (pixels) and the color is coded and the position of each dot horizontally and vertically.
In addition to the binary and decimal systems, the MC uses the hexadecimal system, in which the characters 0 ... 9 and A ... F are used to write numbers. Its use is due to the fact that one byte is described by a two-digit hexadecimal number, which significantly reduces the digital code entry and makes it more readable (11111111 - FF).
Table 1 - Writing numbers in different number systems
To determine the value of a number (for example, the value of the number 100 for different number systems can be 42, 10010, 25616), at the end of the number add a Latin letter denoting the number system: for binary numbers, the letter b, for hexadecimal - h, for decimal - d. A number without an additional designation is considered decimal.
Translation of numbers from one system to another and basic arithmetic and logical operations on numbers allows you to produce an engineering calculator (a standard application of the Windows operating system).
The basis of the microprocessor system is the microprocessor (processor), which performs the functions of information processing and control. The remaining devices that are part of the microprocessor system serve the processor, helping it to work.
Mandatory devices for creating a microprocessor system are input/output ports and partly memory. I / O ports connect the processor with the outside world, providing information input for processing and output of processing results, or control actions. Buttons (keyboard), various sensors are connected to the input ports; to output ports - devices that allow electrical control: indicators, displays, contactors, electrovalves, electric motors, etc.
Memory is needed primarily to store the program (or set of programs) necessary for the processor to work. A program is a sequence of commands understandable to the processor, written by a person (more often a programmer).
The structure of the microprocessor system is shown in Figure 1. In a simplified form, the processor consists of an arithmetic logic unit (ALU) that processes digital information and a control unit (CU).
Memory typically includes read-only memory (ROM), which is non-volatile and designed for long-term storage of information (eg, programs), and random-access memory (RAM), designed for temporary storage of data.
Figure 1 - The structure of the microprocessor system
The processor, ports, and memory communicate with each other via buses. A bus is a set of conductors, united according to a functional feature. A single set of system buses is called intrasystem highway, in which there are:
data bus DB (Data Bus), through which data is exchanged between the CPU, memory and ports;
the address bus AB (Address Bus), used for addressing memory cells and ports by the processor;
control bus CB (Control Bus), a set of lines that transmit various control signals from the processor to external devices and vice versa.
Microprocessors
Microprocessor - a software-controlled device designed to process digital information and control the process of this processing, made in the form of one (or several) integrated circuits with a high degree of integration of electronic elements.
The microprocessor is characterized a large number parameters, since it is both a complex program-controlled device and an electronic device (microcircuit). Therefore, for the microprocessor, both the type of package and the processor instruction set. The capabilities of a microprocessor are defined by the concept of microprocessor architecture.
The prefix "micro" in the name of the processor means that it is performed using micron technology.
Figure 2 - Appearance Intel Pentium 4 microprocessor
During operation, the microprocessor reads the program instructions from the memory or input port and executes them. What each command means is determined by the processor's instruction set. The system of commands is embedded in the architecture of the microprocessor and the execution of the command code is expressed in the execution of certain micro-operations by the internal elements of the processor.
microprocessor architecture is its logical organization; it determines the capabilities of the microprocessor in terms of hardware and software implementation of the functions necessary to build a microprocessor system.
Main characteristics of microprocessors:
1) Clock frequency(unit MHz or GHz) - the number of clock pulses in 1 second. Clock pulses are generated by a clock generator, which is most often located inside the processor. Because Since all operations (instructions) are performed by clock cycles, then the performance of work (the number of operations performed per unit of time) depends on the value of the clock frequency. The frequency of the processor can be varied within certain limits.
2) Processor bit depth(8, 16, 32, 64 bits, etc.) - determines the number of data bytes processed per clock cycle. The bitness of the processor is determined by the bitness of its internal registers. The processor can be 8-bit, 16-bit, 32-bit, 64-bit, etc. i.e. data is processed in chunks of 1, 2, 4, 8 bytes. It is clear that the greater the bit depth, the higher the performance.
Internal microprocessor architecture
A simplified internal architecture of a typical 8-bit microprocessor is shown in Figure 3. The structure of the microprocessor can be divided into three main parts:
1) Registers for temporary storage of commands, data and addresses;
2) Arithmetic logic unit (ALU), which implements arithmetic and logical operations;
3) Control and timing scheme- provides a selection of commands, organizes the functioning of the ALU, provides access to all registers of the microprocessor, perceives and generates external control signals.
Figure 3 - Simplified internal architecture of an 8-bit microprocessor
As can be seen from the diagram, the processor is based on registers, which are divided into special (having a specific purpose) and general-purpose registers.
Program counter (PC)- register containing the address of the next command byte. The processor needs to know which instruction will be executed next.
Accumulator - a register used in the vast majority of commands for logical and arithmetic processing; it is both the source of one of the data bytes required for the ALU operation and the place where the result of the ALU operation is placed.
Feature register (or flag register) contains information about internal state microprocessor, in particular the result of the last ALU operation. The flag register is not a register in the usual sense, but is simply a set of trigger latches (flag up or down. Usually there are zero, overflow, negative, and carry flags.
Stack pointer (SP)- monitors the position of the stack, i.e. contains the address of its last used cell. A stack is a way of organizing data storage.
The command register contains the current command byte, which is decoded by the command decoder.
The external bus lines are isolated from the internal bus lines with buffers, and the main internal elements are connected by a high-speed internal data bus.
To improve the performance of a multiprocessor system, the functions of the central processing unit can be distributed among several processors. To help the central processor, coprocessors are often introduced into the computer, focused on the effective execution of any specific functions. Widespread mathematical and graphic, input-output coprocessors, unloading the central processor from simple, but numerous operations of interaction with external devices.
On the present stage the main direction of productivity improvement is the development multi-core processors, i.e. combining two or more processors in one package to perform several operations in parallel (simultaneously).
The leading companies in the development and manufacture of processors are Intel and AMD.
The algorithm of the microprocessor system
An algorithm is an exact prescription that uniquely specifies the process of converting initial information into a sequence of operations that allow solving a set of problems of a certain class and obtaining the desired result.
The main control element of the entire microprocessor system is the processor. It is he who, with the exception of a few special cases, controls all other devices. The rest of the devices, such as RAM, ROM and I / O ports are slaves.
Immediately after switching on, the processor begins to read digital codes from the area of \u200b\u200bmemory that is reserved for storing programs. Reading occurs sequentially cell by cell, starting from the very first. The cell contains data, addresses and commands. A command is one of the elementary actions that a microprocessor can perform. All the work of the microprocessor is reduced to sequential reading and execution of commands.
Consider the sequence of actions of the microprocessor during the execution of program commands:
1) Before executing the next command, the microprocessor contains its address in the PC program counter.
2) The MP accesses the memory at the address contained in the PC and reads the first byte of the next command from the memory into the command register.
3) The command decoder decodes (decrypts) the command code.
4) In accordance with the information received from the decoder, the control device generates a time-ordered sequence of micro-operations that implement the instructions of the command, including:
Retrieves operands from registers and memory;
Performs arithmetic, logical or other operations prescribed by the command code on them;
Depending on the length of the command, modifies the contents of the PC;
Transfers control to the next command, the address of which is again in the PC program counter.
The set of microprocessor instructions can be divided into three groups:
1) Data movement commands
Movement occurs between memory, processor, I / O ports (each port has its own address), between processor registers.
2) Data conversion commands
Any data (text, picture, video, etc.) are numbers, and only arithmetic and logical operations can be performed with numbers. Therefore, the commands of this group include addition, subtraction, comparison, logical operations, etc.
3) Transfer command
It is very rare for a program to consist of one sequential instruction. The vast majority of algorithms require program branching. In order for the program to be able to change the algorithm of its work depending on any condition, control transfer commands are used. These commands make the program run according to different ways and organize cycles.
External devices
External devices include all devices that are outside the processor (except random access memory) and connected via I/O ports. External devices can be divided into three groups:
1) devices for human-computer communication (keyboard, monitor, printer, etc.);
2) devices for communication with control objects (sensors, actuators, ADC and DAC);
3) external storage devices of large capacity ( HDD, drives).
External devices are connected to the microprocessor system physically - using connectors, and logically - using ports (controllers).
An interrupt system (mechanism) is used to interact with the processor and external devices.
Interrupt system
This is a special mechanism that allows, at any time, by an external signal, to force the processor to suspend the execution of the main program, perform operations associated with the interrupt-causing event, and then return to the execution of the main program.
Any microprocessor has at least one interrupt request input INT (from the word Interrupt - interruption).
Consider an example of the interaction of a personal computer processor with a keyboard (Figure 4).
Keyboard - a device for entering symbolic information and control commands. To connect a keyboard, the computer has a special keyboard port (chip).
Figure 4 - The work of the processor with the keyboard
Work algorithm:
1) When a key is pressed, the keyboard controller generates a numeric code. This signal goes to the keyboard port chip.
2) The keyboard port sends an interrupt signal to the processor. Each external device has its own interrupt number, by which the processor recognizes it.
3) Having received an interrupt from the keyboard, the processor interrupts the execution of the program (for example, the Microsoft Office Word editor) and loads the program for processing codes from the keyboard from memory. Such a program is called a driver.
4) This program directs the processor to the keyboard port and the numeric code is loaded into the processor register.
5) The digital code is stored in memory and the processor moves on to another task.
Due to the high speed of work, the processor performs a large number of processes simultaneously.
1.1 Microprocessor definition
In the early 1970s, advances in technology in microelectronics led to the creation of a new elemental base of electronics - microelectronic large-scale integrated circuits (LSI) (module 1 chapter 1.6.3). According to the degree of integration (the number of active elements: diodes and transistors), integrated circuits (ICs) are conditionally divided into ICs of a low degree of integration - up to 100 active elements, medium degree of integration (SIS) - up to 1000 active elements, LSI - over 1000 active elements, VLSI - over 10,000 items. Release of a new BIS at modern level design automation is a very complex and expensive process due to the high initial costs of developing it. logical structure and topology, production of photomasks and technological preparation of production. This is 0.5-1 year of work of a large team. Therefore, the manufacture of LSIs is economically justified when they are produced, estimated at tens or hundreds of thousands of pieces per year. It is almost impossible to produce specialized LSI for each specific application. As a result of the search for areas of mass application of microcircuits with a high level of integration, their developers proposed the idea of creating one universal LSI or some set of LSIs, the specialization of which for each specific application is achieved not by circuitry, but by software. This is how standard universal elements appeared - microprocessor LSI with a structure similar to that of a computer.
A microprocessor (MP) is a processing and control device capable of processing information, making decisions, inputting and outputting information under program control, and is made in the form of one or more LSI.
1.2 Manufacturing technology of MP LSI
There are two types of LSI manufacturing technology: bipolar - based on the use of bipolar transistors and MOS (metal - oxide - semiconductor) - technology based on the use of field-effect transistors.
LSIs manufactured using bipolar technology differ in their schematic implementation methods. Basically, transistor-transistor logic with Schottky diodes (TTLSh) and emitter-coupled logic (ECL) are used. The TTLSH logic uses bipolar n-p-n transistors, supplemented by Schottky diodes (DSh). DS is a rectifier contact at the Al-nSi metal-semiconductor interface. In metal and silicon, the majority carriers of the same type are electrons, and there are no minority carriers. LHs open at U=0.1-0.3 V and have a steep current-voltage characteristic. They are connected in parallel with the collector transition n-p-n transistor and form a Schottky transistor, manufactured in a single technological process. The use of LH significantly increases the speed of the transistor, since saturation of the collector junction is eliminated and there is no absorption of charges in it.
First generation
4004 - 1971
The history of the MP began in 1971, when INTEL (its name comes from the words Integrated Electronics) released the first MP i4004, manufactured using p-MOS technology with a resolution of 10 microns. It had a data width of 4 bits, the ability to address 640 bytes of memory, a clock frequency of f=108 kHz, and performed 60 kop/sec. Such a processor could already work as a computing core of a calculator. It contained 2300 transistors.
8008 - 1972
In 1972, the first improved eight-bit MP i8008 appeared, also made using p-MOS technology. It was housed in a 16-pin package. Executed 48 commands, addressed 16 Kb of memory, f=800 KHz. It had 7 internal 8-bit registers and a 7-level internal stack.
Second generation
8080 - 1974
In 1974, the i8080 MP appeared, manufactured using n-MOS technology with a resolution of 6 microns, which made it possible to place 6000 transistors in a crystal. The processor required three power supplies (+5, +12, -5 V) and complex push-pull synchronization at a frequency of 2 MHz. Its full analogue of the Russian production KR580VM80 is discussed in detail above. At the same time, Motorola released the M6800 MP, which differed from the i8080 in that it had one supply voltage, a more powerful interrupt system, contained two batteries, but did not have RON. Data for processing was retrieved from external memory and then returned there. The memory commands are shorter and simpler than in the BM80, but the transfer takes longer. No advantages have been revealed in the internal structure of the M6800 to date. There are two competing families of Intel and Motorola left. However, most of both the world and the Russian market is occupied by Intel products.
The next was the i8085 processor (f=5 MHz, 6500 transistors, 370 kop/s, 3-micron technology). It retained the popular i8080 register architecture and software compatibility, but added a serial interface port, clock generator, and system controller. The supply voltage is one: + 5V.
Z80 - 1977
Some of the Intel developers who did not agree with a number of management decisions moved to Zilog and in 1977 created the Z80 MP (the Russian analogue of the K1810VM80). This MP was used in the British computer "Spectrum" by Sincler, which was considered the best representative of the 8-bit MP of the 2nd generation.
third generation
8086 - 1978
This generation of Intel MPs laid the foundation for modern personal computers. In 1978, the 16-bit i8086 processor was released. His data: f=5 MHz, performance 330kop/s, 3µm technology, 29k transistors. It began to use memory segmentation and a new instruction encoding scheme.
8088 - 1979
However, the too complicated and expensive technology for the production of this processor forced Intel, since 1979, for some time to release a somewhat simplified version called i8088, the data bus of which was only 8 bits. It was this processor that IBM chose for its first personal computer, the IBM PC / XT model.
80186 - 1980
In 1980, MP i80186 was created. Compared to the i8086, it additionally includes two independent high-speed DMA channels, a programmable interrupt controller, and signals for selecting 7 peripheral devices are generated. There are 16 internal programmable timers, two of them have an output to the outside, the rest can create time delays. Command queue - 6 bytes (in i8088 - 4 bytes). There are 10 additional commands that speed up the execution of programs compared to the i8086. However, this processor was not widely used in computers.
fourth generation
80286 - 1982
In 1982, the i80286 processor appeared, which was used by IBM in the PC / AT computer (AT - Advanced Technology - promising technology). It already had 134 thousand transistors (1.5 micron technology) and addressed up to 16 MB of physical memory. It could work in two modes: real and protected. In real mode i80286 works like i8086 with increased performance(f up to 20 MHz). The memory is considered as a number of segments, each of which contains 2 16 bytes. Segments start at addresses that are multiples of 16 (lower 4 address bits are always 0). Segments can be set arbitrarily in programs. Segment addresses are stored in segment registers. In protected mode, the high segment address is not calculated by adding 4 leading zeros, but is extracted from tables indexed using segment registers. This allows you to work with large amounts of information, the volume of which exceeds the amount of physical memory. If the physical memory is fully loaded, then the data that does not fit is located on the hard drive. In addition, protected mode can support multitasking. For this purpose, a operating system OS/2.
In this mode, the processor can execute various programs in the allocated time slots allocated for each of the programs. It seems to the user that the programs are running at the same time.
Fifth generation
80386 - 1985
Its first representative was a 32-bit MP i80386DX containing 275 thousand transistors, 1.5 micron technology, 4 GB addressable physical memory. There are new registers, new 32-bit operations.
In order for the MP to execute programs written for previous generations, it has three modes of operation.
After resetting or applying the supply voltage, the MP goes into real mode and works like a very fast i8086, but, at the request of the programmer, with 32 bits. All actions: addressing, memory access, interrupt handling are performed as in i8086. The second mode - protected, is enabled by loading a certain status word into the control register. In this case, the MP works as i80286 in protected mode. Implemented multitasking, memory protection using a four-level privilege mechanism and its paging. The MP operates as multiple virtual processors with shared memory, each of which can be in i8086, i80286, or i80386 modes.
In the third, virtual mode, the advantages of this processor are fully revealed. Here, all 32 bits of the address are fully used and it is possible to work with virtual memory. It was only with the advent of the i80386 that the rapid introduction of Windows began, since the power of processors of previous generations was insufficient for Windows.
80386SX - 1988
In 1988, the i80386SX processor appeared, which filled the gap between the already obsolete i80286 processor and the very expensive i80386DX processor. Replacing the outdated i80286 processor on the motherboard with the i80386DX is not possible due to the larger data bus width of the latter. The i80386SX processor allows such a replacement. Internal processes in i80386SX are the same as in i80386DX, but communication with " external environment" is carried out only through a 16-bit bus. As a result, communication occurs in 2 steps of 16 bits, which slows down work by about 10%. Another limitation of the i80386SX processor is a 24-bit address bus, which limits the size of RAM to 16MB. Following With the i80386SX MP reviewed, Intel created and marketed the 33MHz i80386SL processor, built on CMOS structures that provide minimal power consumption.
sixth generation
80486 - 1989
It appeared in 1989 as MP i80486DX. In contrast to the MP of previous generations, this MP does not represent something fundamentally new. In it, the i80386 processor, the i80387 coprocessor and the primary cache with a capacity of 8 KB were copied in one chip.
Note.
Despite the 32-bit architecture inherited from the i80386 MP, as a result of combining the processor, coprocessor and cache on one chip and other improvements, the i80486 at the same clock frequency performs calculations 3-4 times faster than its predecessor.
Intel has been improving this processor all the time, and the i80486DX2 MP, in which the external clock frequency is doubled by the microcircuit's own quartz, and the i80486DX4, in which the frequency is multiplied by 3, were released. In these processors, all instructions that do not require data transfer to an external bus, performed 2-3 times faster. Only the time spent accessing RAM and slower peripherals reduce the speed of work. Besides, in i80486DX4 the cache memory is increased up to 16 KB.
Generations of Pentiums
Pentium P5 - 1993
In 1993, the i80586 appeared, which was given the name Pentium (P5). It was a 32-bit processor with an external clock frequency of 66 MHz, built using submicron technology with a CMOS structure (0.8 microns) containing 3.1 million transistors. The Pentium has two 32-bit address spaces (logical and physical), a 64-bit data bus, and 2 instruction processing pipelines operating in parallel. Two sets of commands are executed at the same time. The 16 KB cache is divided into 8 KB instruction cache and 8 KB data cache. Contains a new floating point unit that performs operations 4-8 times faster than the i80486.
P54, Pentium Pro - 1994
In 1994, the second generation Pentium processors (P54) appeared. With almost the same number of transistors, they were made using 0.6 micron technology, which made it possible to reduce power consumption. Supply voltage reduced to 3.3 V. Internal frequency multiplication applied. In this case, the interface circuits of the external system bus operate at frequencies of 50,60,66 MHz, while the processor core operates at a higher frequency (75,90,100,120,133, 150, 166, and 200 MHz). Frequency separation makes it possible to realize the achievements of MP manufacturing technology, which are significantly ahead of the possibilities of increasing memory performance. The multiplication factor (1.5;2;2.5;3) is set by a combination of signal levels at two control inputs. Processors with different f-values, indicated in the marking on the case, are manufactured according to the same templates. Frequency markings are applied after rigorous screening tests. Depending on the frequency at which the MP completely passed the output control.
In parallel with the Pentium, the Pentium Pro processor also developed. Its main difference in the principle of organization of calculation is dynamic execution. In this case, inside the processor, instructions may not be executed in the order that the program assumes. This improves performance without increasing f. In addition, a dual independent bus architecture has been adopted to increase the overall throughput. One bus is the system bus, it is used to communicate with the core of the main memory and interface devices. The other is intended exclusively for exchange with the secondary cache of 256 KB (512 KB) integrated in the MP case. To reduce the heating of the crystal, it is possible to instantly reduce the power consumption by approximately 10 times by stopping the clocking of most processor nodes. The MP switches to this state by a signal from the internal temperature sensor, as well as by executing the HALT command.
Pentium MMX - 1997
In 1997, the Pentium MMX processor (P55C) was released. MMX technology represents the most significant improvement in the Intel processor architecture since the introduction of the i80386. The Pentium MMX crystal has an area 50% larger than the classic Pentium. The buffer circuits of the output circuits of the microcircuit operate at a voltage of 3.3 V, the internal circuit is 2.8 V for desktop and 2.45 V for portable computer models.
MMX technology is focused on solving multimedia problems that require intensive calculations over integers. Similar tasks are solved by gaming, communication, educational and other programs that use graphics, sound, three-dimensional image, animation, etc.
The essence of MMX technology is the appearance in the processor of 8 new virtual 64-bit registers and 57 new instructions for solving multimedia problems. The eight new registers are virtual because these registers are physically coprocessor registers. This maintains compatibility with previous generations programs.
Pentium II - 1997
In May 1997, the Pentium II, manufactured using 0.3 micron technology, appeared on the market. It is a slightly stripped-down version of the Pentium Pro core with a higher internal clock speed, in which MMX support has been introduced. This processor uses new technology- a chip with a processor core and a set of static memory chips and additional circuits that implement the secondary cache are placed on a small printed circuit board - cartridge. All crystals are covered with a common lid and are cooled by a special fan.
The internal clock frequency is 233,266,300 MHz, the external one remains 66.6 MHz.
The processor has additional low power modes:
1. Sleep ("Sleep mode"), when it does not clock its internal nodes, except for the frequency multiplier circuit.
2. Deep sleep ("Deep sleep"). Occurs when an external clock is removed. In this mode, the processor does not perform any functions and the current consumption is determined only by leakage currents.
Pentium III - 1999
In 1999, the 600 MHz Pentium III processor appeared, containing 9.5 million transistors. According to Intel, this processor will allow you to receive audio and video information from the Internet, as well as three-dimensional graphics of the highest quality. According to the forecasts of manufacturing companies, further development of MP production technology will go in the direction of increasing the density of transistors on a chip, increasing the number of metallization layers and increasing the clock frequency, along with a decrease in the supply voltage and specific (per transistor) consumed electrical and released thermal energy. Currently, the Pentium IV processor is being produced, the clock frequency of which has reached 3000 MHz.
The technological limit of the linear dimensions of transistors on a chip, due to physical limitations, is about 0.05 microns. On the way to further minimization, in addition to physical limitations, there are also economic ones. For each next generation of chips, the cost of technology doubles. In 1986, the i80386 was produced at a $200 million factory. The Intel plant is currently worth $2.4 billion. Therefore, a plant producing chips using 0.25 micron technology will cost $10 billion. The production time for MP is increasing. So the Pentium processor is produced in 6 months, and the newer Pentium Pro - in 9 months. MP generations change every 2-3 years. With each generation, the linear dimensions of the elements decrease by about 1.5 times. In 2000, the width of the conductors was 0.2 µm, and in 2006 it reached 0.1 µm, the clock frequency has already exceeded 2000 MHz.
The above brief data on the development of MP using the example of Intel products show how rapidly the production of MP is developing and improving. No other branch of technology is developing so rapidly. Gordon Moore, the founder of Intel, expressed this very figuratively: “If the automotive industry developed at the speed of the semiconductor industry, then today a Rolls-Royce would cost $ 3, could drive half a million miles on one gallon of gasoline and it would be cheaper to throw it away than pay for parking.
In this review, only Intel processors are considered. It should be noted that the technology of other companies that produce processors, such as AMD, Cyrix, Motorola and others, also goes through a similar path of development. But the leading "trendsetter" in this struggle for quality remains Intel.
9 Microprocessors and microcomputers in information and measuring equipment
9.1 The main functions of the MP in the measuring equipment
The most commonly used built-in MP and MK. They significantly improve the characteristics of devices (accuracy, reliability, efficiency, etc.). The use of the built-in MP allows turning a single-functional device into a multifunctional one by combining several functional units together with switching devices in one unit. MP makes such a device programmable.
The MP improves the accuracy of the measuring device by automatically compensating for the zero setting before the start of measurements, automatically performing calibration (self-calibration, performing self-control), and performing automatic statistical processing of measurement results.
MT expands the measurement capabilities of instruments through the use of indirect and aggregate measurements. In indirect measurements, it is not the desired parameter that is measured, but other parameters with which the desired one is related by a functional dependence. For example, power can be determined by measuring voltage and resistance and calculated using the formula P=U 2 /R. When using the method of cumulative measurements, several physical quantities of the same name are simultaneously measured, at which the desired values of the quantities are found by solving a system of equations. In this case, the MT is programmed to implement the necessary analytical dependencies.
9.2 Examples of using MP in measuring equipment
9.2.1 Microprocessor digital frequency counter
To measure high frequencies, a direct method is used, in which a certain time interval is selected and the number of periods of the signal under study is counted. The measurement accuracy increases with the increase in the number of periods N. At low frequencies, this would require too much time. Therefore, at low frequencies, an indirect method is used. The width of the temporary gates is selected as a multiple of the period of the signal under study qT x , the gates are filled with pulses of the generator of a known frequency F sch, and the number of pulses n is counted. Both methods are illustrated in Figure 9-1
Fig.9-1 Timing diagrams of the frequency measurement process.
Here:
a - measured signal;
b - signal converted into a sequence of pulses;
c - time interval for indirect measurement;
d - filling pulses during indirect measurement;
e - time interval for direct measurement;
e - burst of pulses during direct measurement.
Figure 9-2 shows a block diagram of a device for measuring the frequency of a signal by direct and indirect methods under the control of the MP, in which the points corresponding to the timing diagrams are marked.
Fig.9-2
direct method
When A 0 =1, a direct measurement method is implemented. Multiplexers select inputs x 1 . The MP creates a temporary gate with a duration of T. If the counter counted N pulses in this interval, then T=nT x, or T=n/F x, hence F x =n/T.
indirect method
When A 0 =0, x 0 inputs of the multiplexers are selected, and an indirect measurement method is implemented. The temporary gate shaper contains a frequency divider with a conversion factor q=2 k, where k is chosen so as to obtain the number of pulses (plot d) that provides the required measurement accuracy F x . In the interval qT x fit n pulses qT x =nT mid or q/F x =n/F mid, so F x =qF mid /n.
9.2.2 Wide range counter
It uses a heterodyne method for lowering the frequency of the measured signal. If you mix the measured signal F meas with the local oscillator signal (auxiliary generator) F 1 , the result is signals with frequencies F meas +nF 1 and F meas -nF 1 . To lower the frequency, the variant F meas -nF 1 =F pr is used, where F pr is the intermediate frequency allocated by the next block.
Fig.9-3
PSCH - programmable frequency synthesizer (local oscillator).
UPCH - intermediate frequency amplifier.
TsCH - digital frequency meter type fig.9-2
During operation, the MP changes F synth to the value F "synth, at which
F meas -F "synth \u003d F ave. Then F meas \u003d F pr + nF "synth.
9.2.3 Measuring generator with MP control
The most commonly used are functional generators that generate signals of various shapes (triangular, rectangular, sinusoidal, and others) with normalized metrological characteristics. The frequency range of such generators is 10 -6 Hz - 50 * 10 6 Hz. Figure 9-4 shows a block diagram of such a generator.
Fig.9-4
Here, BS is a programmable block of counters, GTI is a programmable clock generator.
After the operator enters the function f(t) to generate a signal of the same form, the MP calculates the samples f(t i) on the interval of one period with a given sampling rate. Readings are written to RAM. The GTI output signal goes to the BS, where the RAM address is formed.
9.2.4 Digital filters
A digital filter is a device that converts one discrete signal x n into another discrete signal y n , and the signals x n and y n themselves are binary digital codes.
The analog filter is a frequency selective circuit that performs some linear transformation from a continuous input signal U 1 (t) to a continuous output signal U 2 (t). In contrast, the digital filter converts the input digital sequence x(nT) into the output digital sequence y(nT). Consider the conversion of an analog filter into a digital one using the simplest filters as examples.
The simplest analog high-pass filter is an RC circuit (Figure 9-5).
Fig.9-5
Let's define the ratio between input and output voltage.
U 2 (t) \u003d i (t) * R \u003d RC * d (U 1 -U 2) / dt (1)
Let's represent U 1 (t) and U 2 (t) by the corresponding digital sequences U 1 =x(nT) and U 2 =y(nT), then:
Substituting (2) into (1), we get:
Denote
.
The resulting expression determines the algorithm for calculating the output signal of the filter Y n at the n-th quantization step, depending on its value at the previous n-1 -th step, the values of the input signal X n , X n -1 and the sampling step τ. Let's define the transient response of the high-pass filter.
If we choose the sampling step τ=1, then we get
X(nT)=1 for n>=0,X(nT)=0 for n<0.
With a smaller step τ=0.125 we have
When using an analog filter, solving its differential equation gives
Figure 90-6 shows the values of the output signal calculated by formulas (3), (4) and (5) and the corresponding graphs.
Fig.9-6
It can be seen that as the sampling interval τ decreases, the transient response of the digital filter approaches that of the analog filter.
The simplest analog low-pass filter is shown in Figure 9-7.
Figure 9-7
It is described by the equation:
Let's move on to increments:
and finally:
It can be shown that in this case, as τ decreases, the transient response of the digital filter indefinitely approaches the transient response of the analog filter.
In digital filters, it all comes down to operations of multiplication by some coefficients and addition. The above filters are first order filters. The best results are obtained by filters of higher orders, in which the values of x and y delayed by several steps are used to calculate the output value Y n.
The calculation of such an expression is very easy to program and perform on the MP. Delayed signals are placed on the stack.
10 Testing of microprocessor systems
10.1 Testing with static signals
In microprocessor systems, data flows are aperiodic, signal durations change, which causes great difficulties in testing and diagnostics - determining the cause of errors. One way to overcome these difficulties is to statically test the system. For MP K580VM80, this is done as follows. MP is not soldered into the board, but is installed in the panel. When testing, the MP is removed and an adapter block for simulating and indicating signals is inserted. Toggle switches are connected to the address bus pins, toggle switches are connected to the data bus through tri-state circuits and LEDs through open-collector logic elements. By dialing the necessary addresses and output signals of the MP with toggle switches, you can test the system.
10.2 Autodiagnostics of microprocessor systems
Autodiagnostics is a built-in diagnostics based on the use of internal diagnostic programs. These programs may be self-executing or invoked by the user of the system. They are laid down when designing a microprocessor system.
10.3 Logic analyzers
Testing with static signals is a slow and not always applicable process. More universal is the use of special devices - logic analyzers.
10.3.1 Logic state analyzers (synchronous mode)
They are available in 8-, 12-, 16-, and 32-bit versions. Output information is given in the form of tables of ones and zeros, octal or hexadecimal codes. The analyzer is connected to the bus under test, and a table or display shows a table of n bus states, starting from the specified state, or n previous states. Similar analyzers are built according to the block diagram of Fig. 10-1.
Fig.10-1
K0-K15 - input signal comparators;
R - potentiometer for setting the level of comparison;
KC - word comparator;
Kl - word input keyboard;
FUS - control signal generator;
Rg0-Rg15 - shift registers (module 2 chapter 7.2) for recording 16 values of the i-th input;
f:n - frequency divider; BPR - conversion block.
At the beginning of the work of the logic analyzer, a word is typed on the keyboard, starting from which the analysis is performed. If the code at the outputs K0-K15 and the dialed code match, the COP generates a pulse, under the influence of which the FUS generates control signals US1 and US2. With the arrival of each clock pulse TI, a counting pulse US1 * TI appears at the output of the counter - divider. After n clock pulses arrive, the conjunctor &2 closes and writing to the registers stops. The conversion block from n output values of registers Rg0-Rg15 forms a table on the display screen containing n lines.
10.3.2 Logic timing analyzers (asynchronous mode)
Such analyzers scan the input signals at a frequency much higher than the frequency of the signals. This allows not only to determine the presence or absence of a signal in each clock period, but also to investigate the dynamics of change, detect front distortions, short-term peaks, dips, etc. Asynchronous mode analyzers are clocked at a much higher internal frequency. Devices are produced with f=20, 50, 100, 200 MHz. They use additional trigger circuits for fixing false pulses up to 5 ns, which makes it much easier to detect such pulses.
10.4 In-circuit emulators
Emulation is a process in which one system is used to reproduce the properties of another system. In-circuit emulators are used to organize the emulation of various components of the developed microprocessor device. They are intended for the organization of complex debugging of development. The industry produces emulators as stand-alone devices. They emulate the behavior of the microprocessor, storage devices, peripheral devices.
The in-circuit emulator can operate in the modes of polling the status of various MPS nodes, step-by-step execution of the user program. With its help, the MPS core, trunks are checked, ROM and RAM tests are performed. The best testing option is a combination of in-circuit emulation and signature analysis methods.
10.5 Signature analysis
A signature is a number consisting of 4 characters of a hexadecimal code and conditionally, but unambiguously characterizing a certain node of the controlled device. The signature is determined at the factory - the manufacturer of the device and is indicated at individual points of the circuit (Fig. 10-2) or in the instructions for the device.
Fig.10-2 Signatures indicated on the device diagram
The signature is formed from the test signal (test sequence) generated by the MP. A test sequence consisting of at least 16 zeros and ones is fed to the input of any node. From the output of the node (controlled point), the already converted sequence is removed and fed to the input of the signature analyzer. The signature analyzer contains a BFS signature generation block (Fig. 10-3), consisting of 16 triggers interconnected through adders modulo 2. When the analyzer is running, the operation of dividing polynomials is performed. The input sequence forms a dividend, the FFS circuit is a divisor, and the result recorded in triggers after the end of the test sequence is the remainder of the division. If the test sequences at the manufacturer and at the consumer conducting the test are the same, as well as the same BFS, then when checking a working unit, the resulting signature matches the signature specified in the documentation.
Fig.10-3
The probability of obtaining the same signatures for two binary sequences that differ from each other by one bit is equal to zero, and that differs by several erroneous bits is 0.00001526. In other words, the error detection confidence >=99.998%. Checking individual nodes of the device is reduced to determining the signature at the output of the node. If it matches the factory one, the unit is working.
11 Ensuring noise immunity of microprocessor systems
11.1 Primary line suppression
When developing microprocessor systems, it is necessary to pay attention Special attention to protect against interference that leads to malfunctions. A significant part of the interference comes from the mains. An MPS that is well-established in the laboratory may be completely inoperable in a production environment due to interference. Interference occurs when abrupt changes network load, for example, when turning on a powerful electric motor, furnace, welding machine. Therefore, if possible, isolation from such interference sources should be carried out through the network. Figure 11-1 shows various options for connecting devices that include a microprocessor. The best option is to power the MPS and consumers that create powerful current pulses (motors).
Fig.11-1
To suppress short-term interference, a network filter is installed Fig. 11-2.
Fig.11-2
In some cases, it is necessary to introduce an electrostatic shield (for example, an ordinary water pipe connected to a grounded power board housing) for laying network wires inside it.
11.2 Mains interference suppression in the power supply
Despite the correct connection, the electrostatic screen and the presence of a surge protector, interference still partially penetrates the mains input of the device. Due to the capacitive coupling between the mains and secondary windings, the impulse noise passes through the power transformer and enters the rectifier and beyond.
Suppression methods:
1. The primary and secondary windings of a power transformer are located on different coils. This significantly reduces interwinding capacitive coupling, but reduces the efficiency of the transformer.
2. The windings are located on one coil, but are separated by a copper foil screen with a thickness of at least 0.2 mm, which is connected to the body ground. The shield must never be short-circuited!
3. The primary winding is completely enclosed in a screen (not short-circuited), which is grounded.
4. The primary and secondary windings are enclosed in separate screens, and a separating screen is placed between them. All screens are grounded. In parallel with the primary winding, a chain of series-connected C \u003d 0.1 μF and R \u003d 100 Ohm is connected to quench energy at the time of shutdown. 11.3 Grounding rules
In structurally finished units, there are always two types of "ground" buses - body and circuit.
According to the safety regulations, the body bus is necessarily connected to the ground bus laid in the room. The circuit bus (“ground” of the device circuit) should not be connected to the body bus, but there must be a separate clamp for it, isolated from the body. If the system includes several devices connected by information lines, then it is far from indifferent how their body and circuit ground buses are connected to the room ground bus.
If connected incorrectly, the surge voltages generated by the equalizing currents on the ground bus will actually be applied to the inputs of the devices, which can cause them to falsely operate.
The least mutual interference is obtained when the circuit ground buses are combined at one point, and the case buses at another point (Fig. 11-3). The distance between the points is selected experimentally. In some cases point A may not be connected to the room ground bus.
Fig.11-3
11.4 Interference suppression on the secondary supply circuits
At the moments of switching of integrated circuits and in push-pull output circuits, large current surges occur. Due to the finite inductance of the power rails on the boards, they cause voltage pulses. If the tires are thin, and there are no decoupling capacitances, then pulses with an amplitude of up to 2V appear at the “far” end of the bus! The level of such pulses corresponds to a logical unit, which causes failures. To eliminate this effect, follow these recommendations:
1. The power and ground rails on the boards must have a minimum inductance. To do this, they are given a lattice structure that covers the entire free surface of the board.
2. External power and ground buses are connected to the board through several contacts evenly distributed on the connector.
3. Interference is suppressed near the places of their occurrence. To do this, a capacitor C = 0.02 μF is installed near each TTL circuit to eliminate high-frequency interference, and an electrolytic capacitor C = 100 μF is additionally installed on a group of 10-15 circuits.
