|
Arabic
Bengali
Chinese
English
French
German
Hebrew
Hindi
Italian
Japanese
Korean
Malay
Polish
Portuguese
Spanish
Turkish
Ukrainian
Vietnamese|
HISTORY OF TECHNOLOGY, TECHNOLOGY, OBJECTS AROUND US
Calculating machine. History of invention and production
Directory / The history of technology, technology, objects around us The mechanization and mechanization of computing operations is one of the fundamental technical achievements of the second third of the XNUMXth century. Just as the appearance of the first spinning machines was the beginning of the great industrial revolution of the XNUMXth-XNUMXth centuries, the creation of an electronic computer became a harbinger of the grandiose scientific, technical and information revolution of the second half of the XNUMXth century. This important event was preceded by a long prehistory. The first attempts to assemble a calculating machine were made as early as the XNUMXth century, and the simplest computing devices, such as an abacus and an account, appeared even earlier - in antiquity and the Middle Ages.
Although an automatic computing device belongs to the genus of machines, it cannot be put on a par with industrial machines, say, with a lathe or weaving machine, because, unlike them, it operates not with physical material (threads or wooden blanks), but with ideal ones that do not exist in nature by numbers. Therefore, the creator of any computing machine (whether it be the simplest adding machine or the latest supercomputer) faces specific problems that do not arise for inventors in other areas of technology. They can be formulated as follows: 1. How to physically (objectively) represent numbers in a machine? 2. How to enter the initial numerical data? 3. How to simulate the performance of arithmetic operations? 4. How to present the input data and calculation results to the calculator? One of the first to overcome these problems was the famous French scientist and thinker Blaise Pascal. He was 18 years old when he began to work on the creation of a special machine with which a person who was not even familiar with the rules of arithmetic could perform four basic actions. Pascal's sister, who witnessed his work, wrote later: "This work tired my brother, but not because of the strain of mental activity and not because of the mechanisms, the invention of which did not cause him much effort, but because the workers with hard to understand him." And this is not surprising. Precise mechanics was just being born, and the quality that Pascal demanded exceeded the capabilities of his masters. Therefore, the inventor himself often had to take up a file and a hammer or puzzle over how to change an interesting but complex design in accordance with the qualifications of the master.
The first working model of the machine was completed in 1642. She did not satisfy Pascal, and he immediately began to design a new one. “I didn’t save,” he later wrote about his car, “neither time, nor labor, nor money to bring it to the state of being useful ... I had the patience to make up to 50 different models ..." Finally, in 1645, efforts he was crowned with complete success - Pascal assembled a car that satisfied him in every way. What was this first computer in history and how were the tasks listed above solved? The mechanism of the machine was enclosed in a light brass box. On its top cover there were 8 round holes, around each of which a circular scale was applied. The scale of the rightmost hole was divided into 12 equal parts, the scale of the hole next to it was divided into 20 parts, the remaining six holes had a decimal division. Such a graduation corresponded to the division of the livre, the main French monetary unit of that time: 1 sous = 1/20 livre and 1 denier = 1/12 sous. In the holes, gear setting wheels were visible, located below the plane of the top cover. The number of teeth of each wheel was equal to the number of scale divisions of the corresponding hole.
Numbers were entered in the following way. Each wheel rotated independently of the other on its own axle. The rotation was carried out with the help of a driving pin, which was inserted between two adjacent teeth. The pin turned the wheel until it hit a fixed stop fixed at the bottom of the cover and protruding into the hole to the left of the number "1" of the dial. If, for example, a pin was placed between teeth 3 and 4 and the wheel was rotated all the way, then it turned 3/10 of its full circle. The rotation of each wheel was transmitted through an internal mechanism to cylindrical drums, the axes of which were located horizontally. Rows of numbers were applied to the side surface of the drums. The addition of numbers, if their sum did not exceed 9, was very simple and corresponded to the addition of angles proportional to them. When adding large numbers, an operation called the transfer of ten to the highest digit had to be performed. People who count in a column or on an abacus should do it in their minds. Pascal's machine carried out the transfer automatically, and this was its most important distinguishing feature. The elements of the machine belonging to the same category were the adjusting wheel N, the digital drum I and the counter, consisting of four crown wheels B, one gear wheel K and a mechanism for transmitting tens.
Note that the wheels B1, B2 and K are of no fundamental importance for the operation of the machine and were used only to transfer the movement of the setting wheel N to the digital drum I. But the wheels B3 and B4 were integral elements of the counter and therefore were called "counting wheels". The counting wheels of two adjacent categories A1 and A2 were rigidly mounted on the axles. The mechanism for transmitting tens, which Pascal called the "sling", had the following device. On the counting wheel B1 of the junior grade in the Pascal machine, there were rods C1, which, when the axis A1 rotated, engaged with the teeth of the fork M located at the end of the two-knee lever D1. This lever freely rotated on the A2 axis of the senior category, while the fork carried a spring-loaded pawl. When, during the rotation of the axis A1, the wheel B1 reached the position corresponding to the number 6, the rods C1 engaged with the teeth of the fork, and at the moment when it passed from 9 to 0, the fork slipped out of engagement and, under the action of its own weight, fell down, dragging a dog. The latter at the same time pushed the counting wheel B2 of the highest order one step forward (that is, turning it along with the axis A2 by 36 degrees). Lever H, which ended with a tooth in the form of an ax, played the role of a hook that prevented the wheel B1 from rotating in the opposite direction when the fork was raised. The transfer mechanism operated only with one direction of rotation of the counting wheels and did not allow the subtraction operation to be performed by rotating the wheels in the opposite direction. Therefore, Pascal replaced subtraction with addition with a decimal complement. Let, for example, it is necessary to subtract 532 from 87. The addition method leads to the following actions: 532-87=532-(100-13)=(532+13)-100=445. You just need to remember to subtract 100. On a machine that had a certain number of digits, however, one could not worry about this. Indeed, let's subtract 532-87 on a six-bit machine. Then 000532+999913=1000445. But the very first unit will be lost by itself, since the transfer from the sixth category has nowhere to go. Multiplication was also reduced to addition. So, for example, if you wanted to multiply 365 by 132, you had to perform the addition operation five times: 365 But since Pascal's machine introduced the term anew each time, it was extremely difficult to use it to perform this arithmetic operation. The next stage in the development of computer technology is associated with the name of the famous German mathematician Leibniz. In 1672, Leibniz visited the Dutch physicist and inventor Huygens and witnessed how much time and effort was taken from him by various mathematical calculations. Then Leibniz came up with the idea of creating an adding machine. “It is unworthy of such wonderful people,” he wrote, “like slaves, to waste time on computational work that could be entrusted to anyone using machines.” However, the creation of such a machine required all his ingenuity from Leibniz. His famous 12-digit adding machine appeared only in 1694 and cost a round sum - 24000 thalers. The mechanism of the machine was based on the stepped roller invented by Leibniz, which was a cylinder with teeth of various lengths applied to it. In a 12-bit adding machine, there were 12 such rollers - one for each digit of the number.
The arithmometer consisted of two parts - fixed and movable. The main 12-bit counter and the stepped roller of the input device were placed in the fixed one. The installation part of this device, which consisted of eight small digital circles, was located in the moving part of the machine. In the center of each circle was an axle, on which a gear wheel E was mounted under the cover of the machine, and an arrow was installed on top of the cover, which rotated with the axle. The end of the arrow could be set against any number of the circle.
Data entry into the machine was carried out using a special mechanism. The stepped roller S was mounted on a four-sided axis with a toothed rack-type thread. This rail engaged with a ten-tooth wheel E, on the circumference of which the numbers 0, 1 ... 9 were applied. Turning this wheel so that one or another figure appears in the slot of the cover, the stepped roller is moved parallel to the axis of the gear wheel F of the main counter. If after that the roller was turned 360 degrees, then one, two, etc. were engaged with the wheel F. the longest steps, depending on the magnitude of the shift. Accordingly, wheel F turned 0, 1...9 parts of a full turn; the disk or roller R was also rotated. With the next revolution of the roller, the same number was again transferred to the counter. The computing machines of Pascal and Leibniz, as well as some others that appeared in the XNUMXth century, were not widely used. They were complex, expensive, and the public need for such machines was still not very acute. However, as production and society developed, such a need began to be felt more and more, especially when compiling various mathematical tables. Arithmetic, trigonometric and logarithmic tables became widespread in Europe at the end of the XNUMXth - beginning of the XNUMXth centuries; banks and loan offices used interest tables, and insurance companies used mortality tables. But astronomical and navigational tables were of absolutely exceptional importance (especially for England - the "great maritime power"). The predictions of astronomers regarding the position of celestial bodies were at that time the only means by which sailors could locate their ships on the high seas. These tables were included in the "Marine Calendar", which was published annually. Each edition required the enormous labor of tens and hundreds of counters. Needless to say, how important it was to avoid mistakes in compiling these tables. But there were still mistakes. Hundreds and even thousands of incorrect data also contained the most common tables - logarithmic ones. The publishers of these tables were forced to maintain a special staff of proofreaders who checked the calculations received. But this did not save from mistakes. The situation was so serious that the British government - the first in the world - took care of creating a special computer for compiling such tables. The development of the machine (it is called a difference machine) was entrusted to the famous English mathematician and inventor Charles Babbage. In 1822, a working model was made. Since the significance of Babbage's invention, as well as the significance of the method of machine calculations developed by him, are very great, we should dwell in more detail on the structure of the difference engine. Consider first, with a simple example, the method proposed by Babbage for compiling tables. Let's say you want to calculate the table of fourth powers of the members of the natural series 1, 2, 3...
Suppose that such a table has already been calculated for some members of the series in column 1 - and the resulting values are entered in column 2. Subtract the previous value from each subsequent value. You will get a sequential value of the first differences (column 3). Having done the same operation with the first differences, we obtain the second differences (column 4), the third ones (column 5) and, finally, the fourth ones (column 6). In this case, the fourth differences turn out to be constant: column 6 consists of the same number 24. And this is not an accident, but a consequence of an important theorem: if a function (in this case, it is a function y(x)=x4, where x belongs to the set of natural numbers) is a polynomial of nth degree, then in a table with a constant step its nth differences will be constant. Now it is easy to guess that you can get the required table based on the first row using addition. For example, to continue the started table by one more line, you need to perform additions: 156+24=180 590+180=770 1695+770=2465 4096+2465=6561 Babbage's Difference Engine used the same decimal counting wheels as Pascal's. Registers consisting of a set of such wheels were used to represent the number. Each column of the table, except for 1, containing a number of natural numbers, had its own case; in total there were seven of them in the machine, since it was supposed to calculate functions with constant sixth differences. Each register consisted of 18 digital wheels according to the number of digits of the displayed number and several additional ones used as a revolution counter for other auxiliary purposes. If all machine registers stored values corresponding to the last row of our table, then to obtain the next value of the function in column 2, it was necessary to sequentially perform a number of additions equal to the number of additions of the available differences. The addition in the difference engine took place in two stages. The registers containing the terms were shifted so that the teeth of the counting wheels meshed. After that, the wheels of one of the registers rotated in the opposite direction until each of them reached zero. This stage was called the addition phase. At the end of this stage, in each digit of the second register, the sum of the digits of this digit was obtained, but so far without taking into account possible transfers from digit to digit. The transfer took place in the next stage, which was called the transfer phase, and was carried out like this. During the transition of each wheel in the addition phase from 9 to 0, a special latch was released in this discharge. In the transfer phase, all latches were returned to their place by special levers, which simultaneously turned the wheel of the next highest rank by one step. Each such rotation could, in turn, cause a transition from 9 to 0 in one of the digits and, therefore, the release of the latch, which again returned to its place, making a transfer to the next digit. Thus, the return of the latches into place occurred sequentially, starting from the least significant bit of the register. Such a system is called addition with successive transfer. All other arithmetic operations were performed by addition. When subtracting, the counting wheels rotated in the opposite direction (unlike Pascal's machine, Babbage's difference machine allowed this to be done). Multiplication was reduced to sequential addition, and division was reduced to sequential subtraction. The described method could be used not only to calculate polynomials, but also other functions, for example, logarithmic or trigonometric, although, unlike polynomials, they do not have strictly constant leading differences. However, all these functions can be represented (expanded) as an infinite series, that is, a simple polynomial, and the calculation of their values at any point can be reduced to the problem that we have already considered. For example, sin x and cos x can be represented as infinite polynomials:
These expansions are true for all function values from 0 to p/4 (p/4=3, 14/4=0) with very high accuracy. For values of x that are greater than p/785, the expansion has a different form, but on each of these sections, the trigonometric function can be represented as some kind of polynomial. The number of pairs of terms in the series that are taken into account in the calculations depends on the accuracy that you want to get. If, for example, the requirements for accuracy are small, you can limit yourself to the first two or four terms of the series, and discard the rest. But you can take more terms and calculate the value of the function at any point with any accuracy. (Note that 4!=2•1=2; 2!=3•1•2=3; 6!=4•1•2•3=4, etc.) So the calculation of the values of any function was reduced by Babbage to one simple arithmetic operation - addition. Moreover, when moving from one section of the function to another, when it was required to change the value of the difference, the difference engine itself gave a call (it called after a certain number of calculation steps had been completed). The mere creation of a difference engine would have given Babbage a place of honor in the history of computing. However, he did not stop there and began to develop a much more complex design - an analytical engine, which became the direct predecessor of all modern computers. What was her specialty? The fact is that the difference machine, in essence, still remained only a complex adding machine and required for its work the constant presence of a person who kept the entire scheme (program) of calculations in his head and directed the machine’s actions along one path or another. It is clear that this circumstance was a certain brake in the performance of calculations. Around 1834, Babbage came up with the idea: "Is it not possible to create a machine that would be a universal calculator, that is, would perform all actions without human intervention and, depending on the decision obtained at a certain stage, would itself choose the further path of calculation?" In essence, this meant the creation of a program-controlled machine. The program that had previously been in the head of the operator now had to be decomposed into a set of simple and clear commands that would be entered into the machine in advance and control its operation. No one had ever tried to create such a computer, although the very idea of software-controlled devices had already been realized at that time. In 1804, the French inventor Joseph Jacquard invented the computer-controlled loom. The principle of its work was as follows. The fabric, as you know, is an interweaving of mutually perpendicular threads. This weaving is carried out on a loom, in which the warp threads (longitudinal) are threaded through the eyes - holes in the wire loops, and the transverse threads are pulled through this warp in a certain order using a shuttle. With the simplest weave, the loops rise through one, and the warp threads threaded through them rise accordingly. Between the threads raised and remaining in place, a gap is formed into which the shuttle pulls the weft thread (transverse) behind it. After that, the raised loops are lowered, and the rest are raised. With a more complex weave pattern, the threads had to be lifted in various other combinations. The weaver manually lowered and raised the warp threads, which usually took a lot of time. After 30 years of persistent work, Jacquard invented a mechanism that made it possible to automate the movement of loops in accordance with a given law using a set of cardboard cards with holes punched into them - punched cards. In Jacquard's machine, the eyes were connected with long needles resting on a punched card. Encountering holes, the needles moved upward, as a result of which the eyes associated with them rose. If the needles rested on the cards in the place where there were no holes, they remained in place, holding the eyes connected to them in the same way. Thus, the gap for the shuttle, and thus the weave pattern of the threads, was determined by a set of holes on the corresponding control cards. Babbage intended to use the same principle of control punched cards in his analytical engine. He worked on its device for almost forty years: from 1834 until the end of his life in 1871, but he could not finish it. However, after him there were more than 200 drawings of the machine and its individual components, provided with many detailed notes explaining their work. All these materials are of great interest and are one of the most amazing examples of scientific foresight in the history of technology. According to Babbage, the Analytical Engine should have included four main blocks.
The first device, which Babbage called the "mill", was designed to perform four basic arithmetic operations. The second device - "warehouse" - was intended for storing numbers (initial, intermediate and final results). The initial numbers were sent to the arithmetic unit, and the intermediate and final results were obtained from it. The main element of these two blocks were registers of decimal counting wheels. Each of them could be set in one of ten positions and thus "remember" one decimal place. The memory of the machine had to include 1000 registers with 50 numerical wheels each, that is, it could store 1000 fifty-digit numbers. The speed of the calculations performed directly depended on the speed of rotation of the digital wheels. Babbage assumed that adding two 50-bit numbers would take 1 second. To transfer numbers from memory to an arithmetic device and vice versa, it was supposed to use gear racks, which were supposed to mesh with the teeth on the wheels. Each rail moved until the wheel was in the zero position. The movement was transmitted by rods and links to an arithmetic device, where it was used by means of another rail to move one of the register wheels to the desired position. The basic operation of the analytical engine, like the difference one, was addition, and the rest were reduced to it. In order to turn many gears, a significant external force was required, which Babbage hoped to obtain through the use of a steam engine. The third device, which controlled the sequence of operations, the transfer of numbers on which operations were performed, and the output of results, was structurally two jacquard punched card mechanisms. Babbage's punched cards differed from Jacquard's punched cards, which controlled only one operation - lifting the thread to obtain the desired pattern in the fabric manufacturing process. The management of the Analytical Engine included various types of operations, each of which required a special type of punched cards. Babbage identified three main types of punched cards: operational (or operation cards), variables (or variable cards) and numerical. Operational punched cards controlled the machine. According to the commands knocked out on them, addition, subtraction, multiplication and division of numbers that were in the arithmetic device took place. One of Babbage's most far-sighted ideas was the introduction of a conditional branch command into the set of commands given by a sequence of operational punched cards. By itself, program control (without the use of a conditional branch) would not be enough to efficiently implement complex computational work. The linear sequence of operations is strictly defined at all points. This road is known in every detail to the very end. The concept of "conditional jump" means the transition of the computer to another section of the program, if a certain condition is previously met. Having the opportunity to use the conditional branch instruction, the compiler of the computer program was not required to know at what stage of the calculation the sign that influences the choice of the calculation course will change. The use of a conditional transition made it possible to analyze the current situation at each fork in the road and, on the basis of this, choose one or another path. Conditional commands could have a very different form: comparing numbers, selecting the required numerical values, determining the sign of a number, etc. The machine performed arithmetic operations, compared the received numbers with each other and, in accordance with this, carried out further operations. Thus, the machine could go to another part of the program, skip some of the commands, or return to the execution of some part of the program again, that is, organize a cycle. The introduction of the conditional branch instruction marked the beginning of the use of logical, and not just computational, operations in the machine.
With the help of the second type of punched cards - variables (or, in Babbage's terminology, "cards of variables"), numbers were transferred between memory and an arithmetic device. These cards did not indicate the numbers themselves, but only the numbers of memory registers, that is, cells for storing one number. Babbage called memory registers "variables", indicating that the content of the register changes depending on the number stored in it. Babbage's Analytical Engine used three types of variable maps: for transferring a number to an arithmetic unit and storing it further in memory, for a similar operation, but without storing it in memory, and for entering a number into memory. They are called: 1) "zero map" (the number is called from the memory register, after which the zero value is set in the register); 2) "saving card" (the number is called from memory without changing the content of the register); 3) "receiving card" (the number is transferred from the arithmetic unit to the memory and written into one of the registers). When the machine was running, there were an average of three variable cards per operational punched card. They indicated the numbers of memory cells (addresses, in modern terminology) in which the two original numbers were stored, and the number of the cell where the result was written.
Numerical punched cards represented the main type of punched cards of the analytical machine. With their help, initial numbers were entered for solving a certain problem and new data that might be required in the course of calculations. After performing the proposed calculations, the machine knocked out the answer to a separate punched card. The operator added these punched cards in the order of their numbers and later used them in his work (they were, as it were, her external memory). For example, when in the course of calculations the machine needed the value of the logarithm 2303, it showed it in a special window and gave a call. The operator found the required punch card with the value of this logarithm and entered it into the machine. “All cards,” Babbage wrote, “once used and made for one task, can be used for solving the same problems with other data, so there is no need to prepare them a second time - they can be carefully preserved for future use; over time, the machine will have your own library. The fourth block was intended for receiving initial numbers and issuing final results and consisted of several devices that provide I / O operations. The initial numbers were entered into the machine by the operator and entered into its storage device, from which the final results were extracted and output. The machine could output the answer on a punched card or print on paper. In conclusion, it should be noted that if the development of the analytical engine hardware is associated exclusively with the name of Babbage, then the programming of solving problems on this machine is with the name of his good friend - Lady Ada Lovelace, the daughter of the great English poet Byron, who was passionately fond of mathematics and perfectly understood in complex scientific and technical problems. In 1842, an article by the young mathematician Menabrea was published in Italy describing Babbage's analytical engine. In 1843, Lady Lovelace translated this article into English with extensive and profound commentary. To illustrate the operation of the machine, Lady Lovelace attached to the article a program she had compiled for calculating Bernoulli numbers. Her commentary is essentially the first ever work on programming. The Analytical Engine turned out to be a very expensive and complex device. The British government, which initially financed Babbage's work, soon refused to help him, so he was never able to complete his work. Was the complexity of this machine justified? Not in everything. Many operations (especially the input-output of numbers and their transmission from one device to another) would be greatly simplified if Babbage used electrical signals. However, his machine was conceived as a purely mechanical device without any electrical elements, which often put its inventor in a very difficult position. Meanwhile, the electromechanical relay, which later became the main element of computers, had already been invented at that time: it was invented in 1831 simultaneously by Henry and Salvatore dal Negro. The use of electromechanical relays in computer technology dates back to the invention of the American Herman Gollerith, who created a set of devices designed to process large amounts of data (for example, census results). The need for such a machine was very great. For example, the results of the 1880 census were processed in the USA for 7 years. Such a significant time was explained by the fact that it was necessary to sort out a huge number of cards (one for each of the 5 million inhabitants) with a very large - 50 headings - set of answers to the questions asked in the card. Gollerith knew firsthand about these problems - he himself was an employee of the US Census Bureau - a statistical agency that was in charge of conducting population censuses and processing their results. Working a lot on sorting cards, Gollerith came up with the idea of mechanizing this process. First, he replaced the cards with punched cards, that is, instead of pencil marking the answer option, he came up with punching a hole. To this end, he developed a special 80-column punched card, on which all the information about one person recorded during the census was applied in the form of punches. (The shape of this punched card has not changed significantly since then.) Usually, one strip of a punched card was used to answer one question, which made it possible to record ten answers (for example, to a question about religion). In some cases (for example, the question about age) two columns could be used, which gave one hundred answers. Gollerith's second idea was a consequence of the first - he created the world's first counting and punching complex, which included an input puncher (for punching holes) and a tabulator with a device for sorting punched cards. Perforation was carried out manually on a punch, which consisted of a cast-iron body with a receiver for a card and the punch itself. A plate with several rows of holes was placed above the receiver; when the punch handle was pressed over one of them, the card under the plate was punched in the required way. A complex punch punched through a group of cards with common data with one touch of the hand. The sorting machine consisted of several boxes with lids. The cards were pushed by hand between a set of spring pins and tanks filled with mercury. When the pin fell into the hole, it touched the mercury and completed the electrical circuit. At the same time, the lid of a certain box was lifted, and the operator put a card there. The tabulator (or adding machine) felt holes on punched cards, taking them as the corresponding numbers and counting them. The principle of its operation was similar to a sorting machine and was based on the use of an electromechanical relay (spring pins and cups with mercury were also used as them). When the rods, during the movement of punched cards, fell through the holes into cups with mercury, the electrical circuit was closed, and an electrical signal was transmitted to the counter, which added a new unit to the number in it. Each counter had a dial with an arrow that moved one scale unit when a hole was detected. If the tabulator had 80 counters, it could simultaneously calculate the results for 8 questions (with ten possible answers for each of them). To calculate the results for the next 8 questions, the same punch card was again passed through the tabulator by its other section. Up to 1000 cards per hour were sorted in one run. The first patent (for an idea) Gollerith received in 1884. In 1887, his machine was tested in Baltimore when compiling population death tables. In 1889, the decisive test of the system took place - a trial census was conducted in four districts of the city of San Louis. Gollerith's machine was far ahead of the two competing manual systems (it worked 10 times faster). After that, the US government entered into an agreement with Gollerith for the supply of equipment for the 1890 census. The results of this census, thanks to the tabulator, were processed in just two years. As a result, the machine very quickly gained international recognition and was used in many countries in the processing of population census data. In 1902, Gollerith created an automatic tabulator, in which cards were fed not manually, but automatically, and modernized his sorting machine. In 1908, he created a fundamentally new model of the adding machine. Instead of cups with mercury, contact brushes were used here, with the help of which the electrical circuits of electromagnets were closed. The latter ensured the connection and disconnection of the continuously rotating shaft with the digital wheels of the totalizer counter. The digital wheels turned through gears from a continuously rotating shaft that carried sliding dog clutches controlled by electromagnets. When a hole was found under the contact brush, the electrical circuit of the corresponding electromagnet was closed, and it turned on the clutch, which connected the digital wheel to the rotating shaft, after which the contents of the counter in this category increased by a number proportional to one turn of the wheel. The transfer of tens was carried out in much the same way as in Babbage's difference engine. The work begun by Gollerith continues to this day. Back in 1896, he founded the Tabulayting Machine Company, a company specializing in the production of perforated machines and punched cards. In 1911, after Gollerith left entrepreneurial activity, his firm merged with three others and was transformed into the now widely known worldwide corporation IBM, the largest developer in the field of computer technology. The Gollerith tabulator was the first to use electromechanical elements. The further development of computer technology was associated with a wide and multifaceted application of electricity. In 1938, the German engineer Konrad Zuse created the first-ever relay electronic computer Z1 on telephone relays (the recording device in it remained mechanical). In 1939, a more advanced Z2 model appeared, and in 1941, Zuse assembled the world's first working computer with program control, which used a binary system. All these machines died during the war and therefore did not have much impact on the subsequent history of computing. Regardless of Zuse, Howard Aiken was engaged in the construction of relay computers in the USA. As a graduate student at Harvard University, Aiken was forced to do a lot of complex calculations while working on his dissertation. To reduce the time for computational work, he began to invent simple machines for the automatic solution of particular problems. In the end, he came up with the idea of an automatic universal computer capable of solving a wide range of scientific problems. In 1937, IBM became interested in his project. A team of engineers was assigned to help Aiken. Soon work began on the construction of the Mark-1 machine. Relays, counters, contact and punch card input and output devices were standard parts of tabulators manufactured by IBM. In 1944, the car was assembled and donated to Harvard University. "Mark-1" remained a transitional type machine. It made extensive use of mechanical elements to represent numbers and electromechanical elements to control the operation of the machine. As in Babbage's Analytical Engine, the numbers were stored in registers consisting of ten-tooth counting wheels. In total, "Mark-1" had 72 registers and, in addition, an additional memory of 60 registers formed by mechanical switches. Constants were manually entered into this additional memory - numbers that did not change during the calculation. Each register contained 24 wheels, with 23 of them used to represent the number itself and one to represent its sign. Registers had a mechanism for transferring tens and therefore were used not only to store numbers, but also to perform operations on them: a number located in one register could be transferred to another and added to (or subtracted from) the number located there. These operations were carried out as follows. Through the counting wheels forming the register, a continuously rotating shaft passed, and any wheel could be connected to this shaft with the help of electromechanical switches for a time constituting a certain part of its revolution. A brush (reading contact) was attached to each number, which, when the wheel rotated, ran along a fixed ten-segment contact. This made it possible to obtain the electrical equivalent of the digit stored in a given bit of the register. To perform the summation operation, such connections were established between the brushes of the first register and the switching mechanism of the second register that the wheels of the latter were connected to the shaft for a part of the rotation period proportional to the numbers in the corresponding digits of the first register. All switches were automatically turned off at the end of the addition phase, which occupied no more than half of the turnover period. The summation mechanism itself did not essentially differ from the adder of the Gollerite tabulators. Multiplication and division were performed in a separate device. In addition, the machine had built-in blocks for calculating the functions sin x, log x, and some others. The speed of performing arithmetic operations averaged: addition and subtraction - 0 seconds, multiplication - 3 seconds, division - 5 seconds. That is, "Mark-7" was equivalent to about 15 operators working with manual calculating machines. The work of the "Mark-1" was controlled by commands entered using a perforated tape. Each command was encoded by punching holes in 24 columns running along the tape and read using contact brushes. Punching on punched cards was converted into a set of pulses. The set of electrical signals obtained as a result of "probing" the positions of a given row determined the actions of the machine at a given step of calculations. Based on these commands, the control device ensured the automatic execution of all calculations in this program: it fetched numbers from memory cells, gave the command for the required arithmetic operation, sent the results of calculations to a memory device, etc. Aiken used typewriters and perforators as an output device. Following the launch of the Mark 1, Aiken and his staff began work on the Mark 2, ending in 1947. This machine no longer had mechanical digital wheels, and electrical relays were used to memorize numbers, perform arithmetic operations and control operations - there were 13 thousand of them in total. Numbers in "Mark-2" were represented in binary form. The binary system was proposed by Leibniz, who considered it the most convenient for use in computers. (A treatise on this subject was written in 1703.) He also developed the arithmetic of binary numbers. In the binary system, just like in the decimal system we are used to, the value of each digit is determined by its position, only instead of the usual set of ten digits, only two are used: 0 and 1. In order to understand the binary notation of a number, let's first look at the meaning of well-known decimal notation. For example, the number 2901 can be represented as follows:
That is, the numbers: 2, 9, 0, 1 indicate how many units are in each of the decimal places of the number. If the binary system is taken instead of the decimal system, each digit will indicate how many units are contained in each of the binary digits. For example, the number 13 is written in binary like this:
The binary system is quite cumbersome (say, the number 9000 will be 14 digits in it), but it is very convenient when performing arithmetic operations. The entire multiplication table in it is reduced to a single equality 1 * 1 \u1d 1, and addition has only three rules: 0) 0 + 0 gives 2; 0) 1+1 gives 3; 1) 1+0 gives 1 and transfer XNUMX to the most significant bit. For example: 01010+
The approval of the binary system in computer technology was due to the existence of simple technical analogues of a binary digit - electrical relays that could be in one of two stable states, the first of which was put in line with 0, the other with 1. Transmission of a binary number by electrical impulses from one machine device to another is also very convenient. To do this, just two pulses of different shapes are sufficient (or even one, if the absence of a signal is considered zero). It should be noted that relay machines, created at the dawn of the history of computers, were not used for long in computer technology, since they were relatively slow-acting. Just as in a mechanical machine the speed of calculations was determined by the speed at which the digital wheels were turned, so the operating time of a circuit composed of a relay was equal to the time it took for the relay to operate and release. Meanwhile, even the fastest relays could not make more than 50 operations per second. For example, in Mark-2, addition and subtraction operations took an average of 0,125 seconds, and multiplication took 0,25 seconds. Electronic analogues of electromechanical relays - vacuum lamp triggers - had much greater speed. They became the basic elements in the first generation of computers.
The trigger was invented back in 1919 by the Russian engineer Bonch-Bruevich and independently by the Americans Eccles and Jordan. This electronic element contained two lamps, and at any moment could be in one of two stable states. It was an electronic relay, that is, in the presence of a control pulse signal, it turned on the desired line or electric current circuit. Like an electromechanical relay, it could be used to represent a single binary digit.
Let's consider the principle of operation of an electronic relay, consisting of two vacuum tubes-triodes L1 and L2, which can be located in one cylinder. The voltage from the anode L1 through the resistance R1 is supplied to the grid L2, and the voltage from the anode L2 is supplied to the grid L1 through the resistance R2. Depending on the position in which the trigger is located, it gives a low or high voltage level at the output. Let us first assume that lamp L1 is open and L2 is closed. Then the voltage at the anode of an open lamp is small compared to the voltage at the anode of a closed lamp. Indeed, since the open lamp L1 conducts current, then most of the anode voltage drops (according to Ohm's law u = i • R) at high anode resistance Ra, and on the lamp itself (connected in series with it) only a small part of the voltage drops. Conversely, in a closed lamp, the anode current is zero, and the entire voltage of the anode voltage source drops across the lamp. Therefore, much less voltage drops from the anode of an open lamp L1 to the grid of a closed lamp than from the anode of a closed lamp L2 to grid L1. The negative voltage Ec applied to the grids of both lamps is chosen such that at first lamp L2 is closed, despite the presence of a small positive voltage applied from the anode of the open lamp L1 to grid L2. Lamp L1 is initially open, since the positive voltage applied to the grid from the anode L2 is much greater than Ec. Thus, due to the connection between the lamps through the resistances R1 and R2, the initial state is stable and will persist for as long as you like. Let us now consider what will happen in the circuit if a negative voltage is applied from the outside to the grid of an open lamp L1 in the form of a short current pulse of such a magnitude as to close it. With a decrease in the anode current i1, the voltage at the anode of the lamp L1 will increase sharply and, consequently, the positive voltage on the grid L2 will increase. This will cause the anode current i2 to appear through the lamp L2, due to which the anode voltage on the lamp L2 will decrease. Lowering the positive voltage on the L1 grid will lead to an even greater decrease in the current in L1, etc. As a result of such an avalanche-like growing process of decreasing current in L1 and increasing current in L2, lamp L1 will close, and lamp L2 will be open. Thus, the circuit will move to a new stable equilibrium position, which will be maintained for an arbitrarily long time: the pulse applied to the input 1 is “remembered”. The return of the electronic relay back to its original state can be done by applying a negative voltage pulse to the input. The trigger has, therefore, two stable equilibrium positions: the initial one, in which L1 is open and L2 is closed, and the so-called "excited" state, in which L1 is closed and L2 is open. The time to transfer a trigger from one state to another is very short. Capacitors C1 and C2 serve to speed up the operation of the lamp. The idea of a computer in which vacuum tubes would be used as a storage device belongs to the American scientist John Mauchly. Back in the 30s, he made several simple computing devices on triggers. However, for the first time, another American mathematician, John Atanasov, used electronic tubes to create a computer. His car was already almost complete in 1942. But due to the war, funding for the work was cut off. The following year, 1943, while working at the Moore School of Electrical Engineering at the University of Pennsylvania, Mauchly, together with Presper Eckert, developed his own project for an electronic computer. The US Ordnance Department became interested in this work and ordered the construction of the machine from the University of Pennsylvania. Mauchli was appointed the head of the work. To help him, 11 more engineers (including Eckert), 200 technicians and a large number of workers were given. For two and a half years, until 1946, this team worked on the creation of an "electronic digital integrator and calculator" - ENIAC. It was a huge structure, covering an area of 135 square meters, having a mass of 30 tons and an energy consumption of 150 kilowatts. The machine consisted of forty panels containing 18000 vacuum tubes and 1500 relays. However, the use of vacuum tubes instead of mechanical and electromechanical elements allowed a sharp increase in speed. ENIAC spent only 0 seconds for multiplication, and 0028 seconds for addition, that is, it worked a thousand times faster than the most advanced relay machines. The ENIAC device in general terms was as follows. Every ten triggers were connected in it into a ring, forming a decimal counter, which acted as a counting wheel of a mechanical machine. Ten such rings plus two triggers for representing the sign of a number formed a storage register. In total, ENIAC had twenty such registers. Each register was equipped with circuitry for transmitting tens and could be used to perform summation and subtraction. Other arithmetic operations were performed in special blocks. Numbers were transmitted from one part of the machine to another through groups of 11 conductors - one for each decimal place and sign of the number. The value of the transmitted figure was equal to the number of pulses flowing through this conductor. The operation of individual blocks of the machine was controlled by a master oscillator that generated a sequence of certain signals that "opened" and "closed" the corresponding blocks of the electronic machine. Entering numbers into the machine was done using punched cards. Software control was carried out by means of plugs and typesetting fields (switching board) - in this way the individual blocks of the machine were connected to each other. This was one of the significant shortcomings of the described design. It took up to several days to prepare the machine for work - connecting the blocks on the switching board, while the task was sometimes solved in just a few minutes. In general, ENIAC was still a rather unreliable and imperfect computer. It often failed, and the search for a malfunction was sometimes delayed for several days. In addition, this machine could not store information. To eliminate the last drawback, Eckert in 1944 put forward the idea of a stored program in memory. It was one of the most important technical discoveries in the history of computing. Its essence was that the program commands had to be presented in the form of a numerical code, that is, encoded in the binary system (like numbers) and entered into the machine, where they would be stored along with the original numbers. To memorize these commands and operations with them, it was supposed to use the same devices - triggers, as for actions with numbers. From memory, individual commands were to be extracted to the control device, where their content was decoded and used to transfer numbers from memory to an arithmetic device to perform operations on them and send the results back to memory. Meanwhile, after the end of the Second World War, new electronic computers began to appear one after another. In 1948, the British Kilburn and Williams from the University of Manchester created the MARK-1 machine, in which the idea of a stored program was first implemented. In 1947, Eckert and Mouchli founded their own company, and in 1951 they launched the serial production of their UNIVAC-1 machines. In 1951, the first Soviet computer MESM by Academician Lebedev appeared. Finally, in 1952, IBM released its first industrial computer, the IBM 701. All these machines had a lot in common in their design. We will now talk about these general principles of operation of all computers of the first generation. Electronic computers, as you know, have made a real revolution in the field of application of mathematics to solve the most important problems of physics, mechanics, astronomy, chemistry and other exact sciences. Those processes that previously were completely uncalculable began to be quite successfully modeled on computers. The solution of any problem was reduced to the following successive steps: 1) based on the value of the physical, chemical and other essence of any process under study, the problem was formulated in the form of algebraic formulas, differential or integral equations, or other mathematical relationships; 2) using numerical methods, the problem was reduced to a sequence of simple arithmetic operations; 3) a program was compiled that determined the strict order of performing actions in the established sequence. (The computer carried out, in principle, the same procedure as a person working on an adding machine, but thousands or tens of thousands of times faster.) The instructions of the compiled program were written using a special code. Each of these commands determined some specific action on the part of the machine. Any command, except for the code of the operation being performed, contained addresses. Usually there were three of them - the numbers of memory cells, from where the two initial numbers were taken (1st and 2nd address), and then the number of the cell where the result was sent (3rd address). Thus, for example, the command +/17/25/32 indicated that the numbers in the 17th and 25th cells should be added and the result sent to the 32nd cell. A unicast command could also be used. In this case, to perform an arithmetic operation on two numbers and send the result, three commands were required: the first command called one of the numbers from memory to the arithmetic unit, the next command called the second number and performed the specified operation on numbers, the third command sent the result to memory. So the work of the computer was carried out at the program level. The computational processes proceeded as follows. The operation of the computer was controlled using electronic keys and switches, called logic circuits, and each electronic key, upon receiving a control voltage pulse signal, turned on the desired line or electric current circuit. The simplest electronic key could already be a three-electrode electron lamp, which is locked when a large negative voltage is applied to its grid, and opens if a positive voltage is applied to the grid. In this case, its operation can be represented as a control valve that passes pulse A through itself when a control pulse B is applied to its second input. When there is only one current pulse A or B, the valve is closed and the pulse does not pass to its output. Thus, only when both pulses A and B coincide in time, a pulse will appear at the output. Such a circuit is called a coincidence circuit, or a logical "and" circuit. Along with it, a whole set of other logical circuits is used in the computer. For example, the "or" circuit, which gives an output pulse when it appears on line A or B, or simultaneously on both lines. Another logical scheme is the "no" scheme. It, on the contrary, prohibits the passage of the pulse through the valve, if another inhibiting pulse is simultaneously applied, blocking the lamp. Using these two circuits, you can assemble a one-bit adder. Suppose that pulses A and B are simultaneously transmitted to the "no" and "and" circuits, and the "sum" bus (wire) is connected to the "no" circuit, and the "carry" bus to the "and" circuit. Suppose that a pulse (that is, one) is received at input A, but no input is received at input B. Then "no" will miss the pulse to the "sum" bus, and the "and" circuit will not miss it, that is, the bit will read "1", which corresponds to the binary addition rule. Assume that inputs A and B receive pulses at the same time. This means that the code of number A is "1" and the code of B is also "1". The "no" circuit will not miss two signals and the "sum" output will be "0". But the "and" circuit will skip them, and there will be a pulse on the "transfer" bus, that is, "1" will be transferred to the adder of the adjacent bit. In the first computers, triggers served as the main element of memory and an arithmetic adder. The trigger circuit, as we remember, had two stable equilibrium states. By assigning a code value of "0" to one state and a code value of "1" to another, it was possible to use trigger cells to temporarily store codes. In summing circuits, when a pulse is applied to the counting input of the trigger, it passes from one equilibrium state to another, which fully complied with the addition rules for one binary digit (0+0=0; 0+1=1; 1+0=1; 1+1 =0 and transfer of one to the most significant bit). In this case, the initial position of the trigger was considered as the code of the first number, and the applied pulse was considered as the code of the second number. The result was formed on the trigger cell. In order to implement a summing circuit for several binary digits, it was necessary to ensure the transfer of a unit from one digit to another, which was carried out by a special circuit. The adder was the main part of the machine's arithmetic unit. The adder for parallel addition of number codes at once for all digits had as many single-digit adders as the number code contained binary digits. The added numbers A and B entered the adder from memory devices and were stored there with the help of flip-flops. The registers also consisted of a series of interconnected flip-flops T1, T2, T3, T'1, T'2, etc., into which the number code was supplied from the recording device in parallel for all digits. Each flip-flop stored a code of one digit, so n electronic relays were required to store a number with n binary digits. The codes of numbers stored in the registers were added simultaneously for each digit using adders S1, S2, S3, etc., the number of which was equal to the number of digits. Each one-bit adder had three inputs. The codes of the numbers A and B of the same digit were fed to the first and second inputs. The third input served to transmit the transfer code from the previous digit.
As a result of adding the codes of a given bit, the sum code was obtained on the output bus of the adder, and the code "1" or "0" for transfer to the next bit was obtained on the "transfer" bus. Let, for example, it was required to add two numbers A=5 (in binary code 0101) and B=3 (in binary code 0011). When these numbers were added in parallel, the codes A1=2, A3=1, A1=2, A0=3 and B1=4, B0=1, B1=2, B1=3 were respectively applied to the inputs A0, A4 and A0 of the adder. As a result of summing the codes of the first digit in the adder S1, we get 1+1=0 and the transfer code "1" to the next digit. The adder S2 added three codes: the codes A2, B2 and the carry code from the previous adder S1. As a result, we get 0+1+1=0 and the transfer code "1" to the next third digit. The adder S3 adds the codes of the third digit of the numbers A and B and the transfer code "1" from the second digit, that is, we will have 1+0+1=0 and again the transfer to the next fourth digit. As a result of addition on the "sum" tires, we get the code 1000, which corresponds to the number 8. In 1951, Joy Forrester made an important improvement in the design of the computer, patenting the memory on magnetic cores, which could memorize and store pulses applied to them for an arbitrarily long time.
The cores were made from ferrite, which was obtained by mixing iron oxide with other impurities. There were three windings on the core. Windings 1 and 2 served to magnetize the core in one direction or another by applying pulses of different polarity to them. Winding 3 was the output winding of the cell, in which the current was induced when the core was remagnetized. In each core, by means of its magnetization, a record of one pulse was stored, corresponding to one digit of some number. From the cores connected in a certain order, it was always possible to select the desired number with great speed. So, if a positive signal was applied through the core winding, then the core is magnetized positively, with a negative signal, the magnetization was negative. Thus, the state of the core was characterized by the recorded signal. When reading through the winding, a signal of a certain polarity was applied, for example, positive. If before that the core was negatively magnetized, then it was remagnetized - and an electric current arose in the output winding (according to the law of electromagnetic induction), which was amplified by the amplifier. If the core was positively magnetized, then there was no change in its state - and no electric signal appeared in the output winding. After selecting the code, it was necessary to restore the original state of the core, which was carried out by a special circuit. This type of storage device allowed sampling of numbers in a few microseconds. Large amounts of information were stored on external media, such as magnetic tape. The recording of electrical impulses here was similar to recording sound on a tape recorder: current pulses were passed through the magnetic heads, which magnetized the corresponding places of the passing tape. When reading, the residual magnetization field, passing under the heads, induces electrical signals in them, which are amplified and fed into the machine. In the same way, information was recorded on a magnetic drum covered with a ferromagnetic material. In this case, information could be found faster. Author: Ryzhov K.V.
▪ Fire ▪ Karaoke
Artificial leather for touch emulation
15.04.2024 Petgugu Global cat litter
15.04.2024 The attractiveness of caring men
14.04.2024
▪ The Gulf Stream slowed down due to global warming ▪ ADC and DAC in an audio codec can operate simultaneously at different frequencies ▪ New Garmin fenix 5 multisport smart watch series
▪ section of the site for those who like to travel - tips for tourists. Article selection ▪ article by Chanakya Pandit. Famous aphorisms ▪ article What is bester fish and why is it called that? Detailed answer ▪ article by Chilibukh. Legends, cultivation, methods of application ▪ article Daylight from the battery. Encyclopedia of radio electronics and electrical engineering
Home page | Library | Articles | Website map | Site Reviews
www.diagram.com.ua |
We recommend interesting articles Section 
See other articles Section 
Latest news of science and technology, new electronics:
All languages of this page