DE1056395B - Electric row calculator - Google Patents

Description

The invention relates to an electrical serial computer for performing additions in decimal binary representation. The entered in series Digits are represented in decimal-binary format, and the result is also represented in decimal-binary format won. Since, however, when performing the calculation, the value 10 is in decimal-binary representation can be exceeded, in which case a ten-carry must take place, is in the invention carried out a correction process in the same way as in the known devices of a similar type. The principle of serial arithmetic assumes that the numbers or digits are entered in the arithmetic unit takes place in electrical pulse trains.

According to the invention, an arithmetic unit common to the digits or numbers is used for the series calculation provided, which is designed in a special way. An essential feature in this training consists in the fact that calculations are carried out in binary for each decimal place, even up to the ones above 10 resulting values; but that it is then determined by a comparison whether the result is a Decimal place exceeds the value 9, and if so, the binary represented value 10 is subtracted, so that the result is only a one-digit decimal number. Of course, a ten must then be carried over to the next higher decimal place, which is however again processed as a binary pulse.

However, the basic idea of the invention can just as well be used for subtraction. In this Case, minuend and subtrahend are compared in a comparator by decade, and then the minuend in a binary adder becomes dependent on the comparison in those decades around the Decimal value 10, which contain a lower value than · the corresponding subtrahend decade, increased. Then the subtrahend is deducted from the minuend in a binary subtraction unit, in which a pulse triggered by the comparator is stored as a binary carry before a decade is offset which reduces the minute decade to be offset by the value 1, if the previous one lower minute-end decade was increased by the decimal value 10, whereupon that of the subtraction unit The difference found is noted as the result in binary representation within the decades.

The multiplication can also be carried out with the device according to the invention if the known method of repeated addition is used for this purpose. In the same way you can Divisions can be carried out subtractively by repeated subtraction.

The device according to the invention enables an extraordinary-

Applicant:

Michael Maul,

Sdiwabach near Nuremberg,

Hindenburgstrasse 60

Max Pfeiffer, Nuremberg,
has been named as the inventor.

A very simple calculating device, especially if electronic means are used and the results and the numbers to be introduced are saved by marking combinations on magnetizable drums. -

A number of electronic devices have already become known, 'those in decimal-binary Calculate wisely. A correction process is also provided for the purpose of the tens transfer, by z. B. the filling number 3 · or the filling number 6 "can be used to trigger the tens carry on to initiate the next higher decimal place.

However, these known devices cannot be used in the same way for addition and subtraction, whereas in the present invention the same means for both addition and for Subtraction can be used. This is due to the fact that the application of the known Fill numbers do not allow the direct subtraction option without further ado, as is the case with the registration of the Case is. :

If the known computing devices are to be used for subtraction, this would have to be based on the known Ways of complement addition take place, with a considerable effort for the complement value conversion is required. By identifying the non-marks, the binary 1-2-4-8 key namely, the respective complement cannot be obtained, since the complement to 9 is formed must be, while not marking this key would result in the complement to 15. 'Also It must be taken into account that with the 1-2-4-8 key mentioned, a difference is not made simply through formation The complement of this difference can be queried because the complement is 15 and not a Complement to 9 would result, as would be necessary for the decimal binary representation.

509 508/211

An exemplary embodiment of the invention is shown in the drawings:

Fig. 1 shows schematically the calculation process of a numerical example for addition j

Fig. 2 shows schematically the arithmetic process of two S numbers for subtraction;

3 shows the binary addition possibilities of three binary values of the same binary digit;

4 shows the binary subtraction possibilities of three values of the same binary digit;

5 shows schematically the computing device in the setting for addition;

Fig. 6 shows the gearbox arrangement of the arithmetic unit;

Fig. 7 shows the one-turn clutch for switching on a transport mechanism for punch cards as well for the removal and insertion device;

8 schematically shows the computing device in a subtraction circuit;

Fig. 9 shows part of the development of the pulse generator and the arrangement of the sensing brushes;

10 shows the pulse generator drum and the cam contacts actuated as they rotate;

Fig. 11 shows the arrangement of the magnetic heads for reading or entering the magnetic marking points on the storage drum;

Fig. 12 shows the circuit diagram for a removal head;

Fig. 13 shows the circuit diagram for an insertion head;

14 shows the circuit diagram for the binary adder;

Fig. 15 shows a contact arrangement for reversing the computing device to addition or Subtraction;

Fig. 16 shows the circuit diagram of a binary subtracter;

Fig. 17 shows the circuit diagram of a comparison device to determine which of two encrypted values is the larger;

Fig. 18 shows the circuit diagram for sensing the markings from a punch card and entering the Markings in the map;

19 shows that part of the circuit diagram which controls the operation of the arithmetic unit;

Fig. 20 shows the key according to which the individual numerical values in the punch card and on the magnetic drum being represented;

Fig. 21 shows the cam diagram.

The machine operates according to an iso-called decimal-binary key, as shown in FIG. Each digit is represented by different markings within four positions, each of these positions being assigned a "specific valence 1, 2, 4 and 8. These valences are entered in column E. The individual digits are made up of the values of the The number 5 ζ »Β. is marked in the» 1 «position and in the» 4 «position, so that the sum of both valencies results in the number 5 A number is thus formed from several digits, each of which is represented in the manner explained, and each digit corresponds to a decimal place or "decade" of the number.

If the calculation work is only carried out within a decade, the calculation can be carried out using the numbers encrypted according to FIG. 20 in a purely binary manner, with each lower position corresponding to a higher binary digit. The value 2 ° (= 1) applies for the "1" position, the value 2 * (= 2) for the "2" position, the value 2 ^ (= 4) for the "4" position and for the "8" position the value 2 ³ (= 8). In the further course of the description (counting from above) the "1" position is always the first binary position, the "2" position as the second binary position, the "4" position as the third and the "8" position as fourth binary position of a digit.

The purely binary addition of two digits of the same decimal place value takes place in such a way that in the first Position of the two digits is used to determine whether a marking is provided or not. Only has one of the two digits in the first position, the result is also given in the first Position a mark. If, on the other hand, every first position of the two digits has a marking, then gets the result in the first position is not marked, however, for the second position of the A marking is offset against the result. One unit is therefore transferred to the next higher binary digit. In calculating machine technology, one speaks of »tens carry« if a »decimal place« reaches the value 10, although in reality only the value 1 is transferred to the next higher position. In an analogous application of this expression, one can therefore in binary calculation within the partial values 1, 2, 4, 8 of a marking of a "two-way carry" speak, although in this case of course only the value 1 to the next higher binary Computing station is transferred.

Together with the two-way carry, the second position of the two becomes in the example explained Numbers checked for markings. When a two-way carry is detected from the first position as well as a marker in the second position of each digit will be in the second position of the result a Mark entered and also forwarded a two-way carry to the third position. Is missing on the other hand, a marking in one of the two positions and is the two-way carry present, or two markings are present and absent at the same time the two-way carry, the second position of the result is not marked, but the Carry-two is still forwarded to the third position. Is only one position of the two digits is marked or there is only a two-way carry from the lower position, the Entry of a marker in the second position of the result, but no two-way carryover in the third position. This way of working continues up to the fourth position.

The possibilities of binary calculation for a position are shown in Fig. 3 in several examples I to VII, each example being shown in a horizontal line. Column α should represent one addend, column b the other addend and column c the two-way carry from a preceding binary digit. Column J> T represents the result and column U represents the two-way carry for the next higher binary digit. In the upper three examples I, II, III, "tur -eine" is marked in column α, column b .and column c . Since only a single marking appears with these three possibilities, W ¹ ITd in. In Examples IV, V and VI, on the other hand, there are two markings in each case, which lead to a two-way carryover, but no marking in the results column.

In Example VII, on the other hand, "all three columns a, b and c each have a mark, which in this case leads to an entry in the 4-result column and a two-way carryover.

When adding two digits according to the key f Fig. 20, it is possible that the result represents a value greater than 9 within a decade. In the selected binary representation, these larger values are made up of the individual valencies in the same way as the digits according to w> Fig. 20, If the result is greater than 15 Position shown with a value of 16. Since when adding a »16« position, the sum of all position values becomes 1 + 2 + 4 + 8 + 16—31, i $ can therefore represent the digit sums from 10 to 19. There are no higher digit sums, since the highest digit sum can only result from the addition of 9 + 9 in connection with a carry. zo

The binary subtraction unit works in a similar way to the binary adding unit. The various examples of binary subtraction are shown in FIG. The α column represents in this case the minuend, -the column b the subtrahend and column c to z $ twos carry from a previous BinäT site from. The D column is the result and the U column the two carry-in for a subsequent Bill again.

addition

The operation of the arithmetic unit for the entire number, ie not just for a decade, will be described for addition on the basis of FIGS. 1 and 5. The computing device has three storage drums B, D, E and a pulse generator drum C. On the storage drum E , the eye is entered in the form of a row of markings, while the addend is in the form of a row of marks on the storage drum B. The storage drum D is used for the * o temporary storage of auxiliary markings, to give corresponding work commands on the computing device at a later point in time in connection with the tens transfer between the decades. The pulse generator drum C has a row of markings which reproduce the number 10 for each decade according to the key in Fig. 20. The rows of marks on the two spreader drums E and B are applied in such a way that positions of a decade are one behind the other and the decades themselves are in turn, one behind the other and that the lowest position of the lowest decade (one) reaches the calculating device first. The distance between the individual decades is twice as large as that between the individual markings within a decade. This is used to take up another position with the value 16, which has already been mentioned.

With the start of the set-off, the first marking position of the memory E passes under the removal head Et , which sends a pulse to the input A \ of the adder A if a 6p marking is present. Simultaneously therewith, the first .Markierungsposition the storage drum B running under the extraction head Bl over, the one pulse to the A input of the adder A forwards upon detection of a mark. 3 These two pulses are added together in the adder according to the rule shown in FIG. 3. If only one pulse is present (line I or II in FIG. 3), a pulse is passed on to the entry head EZ via the output A 2 of the adder A , which enters a marking on the memory E. If, on the other hand, two pulses arrive in the adder A from the first position, a storage device is set which, when the pulses arrive from the second position, adds a two-way carry to the latter. This storage device consists of the capacitor Cc and the thyratron 12, which will be discussed in detail later. No pulse is passed on to the insertion head EZ , so that it erases any existing marking on the corresponding position of the storage drum, as will be explained later.

When sensing the second position under the sensing heads E 1 and B 1, the same process is repeated as in the first position, only with the difference that 'if a two-way carry is present, this is carried out in accordance with the scheme according to FIG is taken into account.

The new row of markings, which is formed by the entry header E 2 on the memory E , thus represents the sum of your two memories E and B. However , the representation is such that the sections corresponding to the decades can contain combinations of marks, whose value sum goes beyond -9 up to a total of 19. The latter value would thus be represented by markings in the positions with the values 1, 2 and 16 according to what has been said.

In order to determine those decade sums which are greater than 9, the series of pulses which runs to the entry head B2 is simultaneously passed to the input V1 of the comparison device V. Simultaneously with the arrival of the pulse series of a decade in the comparison device V , a pulse combination corresponding to the number 10 is passed from the pulse generator C via the sensing head C 1 to the input VZ of the comparison device V. The comparison device compares the two pulse series and determines whether the pulse EeJh.e coming to the input Vl represents a smaller value per decade than the pulse series coming to the input V2 with the value 10. If it is not smaller, the value is corresponding decade sum greater than -9, and a pulse is passed from the output V3 of the comparison device to the input S4 of the subtraction unit which causes the subtraction unit to transform the decade in the manner to be described. Simultaneously therewith runs from the output V 3 of the comparison before direction V also a pulse to input A 5 of the adder A, which prepares the two-way carry in the adder ^ before the following decade gets into the adder for calculation. When this two-way carry is triggered, the value 1 is added when the first position is offset, and with it the value 1, which means that the tens carry-over into the next higher decade was also carried out at the same time. The Markierungsr-eihedes intermediate result, which was entered through the head El, then runs under the dispense head 3 JS over, the * is derived according to the Abfühliung of markings pulses to the input S 1 of the 'Subtraktionswerks. 6 Simultaneously with this, a series of pulses is passed from the pulse generator C via the extraction head CZ to the input S3 of the subtraction unit S which reproduces the combination of the value 10 for each decade. The pulse series at the ■ input JT3 can only become effective in the subtraction unit if, before a decade from the comparison device V, a pulse to the

Input 6 * 4 of the subtraction unit, S was passed. In this case, the value 10 is subtracted from the corresponding decade. This deduction game only occurs through the selection by means of the comparison device if the intermediate result of this decade is greater than 9, so that a negative result can never appear at the output 6 "2. In all cases in which no pulse arrives at 6 * 4 , the pulse train emitted by the extraction head C 2 remains ineffective in the subtraction unit, so that the same pulse series appears again at the output S2 of the subtraction unit that was passed into the input S1 of the subtraction unit Entries the pulse series on the storage drum B and at the same time deletes the pulse series of the intermediate result, as will be described below. By subtracting the number 10 in the individual decades of the intermediate result that contain a value greater than 9, the intermediate result is returned to the final result.

The operation of the adding device and the formation of the individual pulse series is shown in FIG. 1 for an example in which the number 297 is added to the number 688. The course of the pulse series is to be viewed in such a way that the scanning or entry heads are at the position of the letters ASU etc. and the markings move from bottom to top. Column A gives the marking row for the end of the pulse, column 5 "for the addend, column U for the carry, column Z for the intermediate result, column 10 for the ten-pulse series, column H for the auxiliary pulse and N. for the final result 985. Field A + V shows the simultaneous operation of the binary adder and the comparison device, while field 5 ″ reproduces the binary subtraction that takes place at a correspondingly later point in time, as described above with reference to FIG became.

When the first decade is run through the adder, the two pulse combinations for the two digits 8 and 7 of the units are added to 15, which is entered as the intermediate result Z in the result memory E. Simultaneously with the entry, the pulse combination of the number 15 is compared with that of the number 10, whereby it is determined that the number 15 is greater than the number 10, and following the comparison an auxiliary pulse H appears. During the subsequent passage of the intermediate result 15 through the subtraction unit, the auxiliary pulse H is previously passed to the subtraction unit, which indicates that the value 10 'is to be subtracted from the intermediate result 15, resulting in the final result 5.

In the second decade (tens) the value 9 is to be added to the value 8. In addition, the determination that the intermediate result of the previous decade is greater than 9 triggered a carry 1 (column U) so that the values 8, 9 and 1 are added to the intermediate result 18. This is in turn compared with the value 10 and found to be greater than 9, so that an auxiliary pulse H is triggered in this decade as well. This auxiliary pulse enables a subtraction of the value 10 from the value 18 for the second decade, so that the value 8 appears as the final result.

In the third decade, the value 2 should be added to the value 6. In addition, the comparison device adds the value 1 as a carry of tens, so that the value 9 results as an intermediate result. A comparison with the value 10 shows that the intermediate result 9 is smaller than the value 10, and therefore no auxiliary pulse is triggered. The subtraction of the number 10 from the intermediate result is therefore also prevented and the intermediate result is entered on the storage drum B as the final result.

subtraction

In the event that one number is to be subtracted from another, the calculating device will do so switched that with the same computing elements with which the addition was carried out, also the Subtraction can be done. The functional diagram for this is shown in FIG. 8 and the pulse trains for a specific calculation example are illustrated in FIG. 2.

Between the works A, V, - (with the connections VlIVIlVh) - S and the storage drums B, C D, E there are different connections after the switchover (the switches are omitted for the sake of simplicity) in FIG. 8 than in FIG. 5, as a comparison of the two figures reveals. However, details of the switching are described in connection with FIG.

The row of marks for the minuend is on the storage drum B and the row of marks for the subtrahend is on the storage drum B. As is well known, the subtrahend is the number that is subtracted from another - namely the minuend. The arrangement of the mark row takes place here so that the first position passes the lowest decade of minuend and subtrahend simultaneously under the extraction heads Bl and Bl, as was true in a similar manner for augend and addend.

If the extraction head E1 detects a marking, it sends a pulse to the input V2 of the comparison device V, while the extraction head B1 sends a pulse to the input Vl of the comparison device V when a marking is detected. In the comparison device V it is determined whether the subtrahend is greater than the minuend within a decade. In the latter case, after the end of the comparison of a decade, a pulse appears at the output V 3 of the comparison device, which is forwarded on the one hand to the entry header Dl of the auxiliary memory D and on the other hand to the input A4: of the adder A. When a pulse arrives at input A 4 of the adder, the pulses arriving at input A 3 are released for accounting. These pulses represent the number 10 and are sent by a pulse generator C in the manner to be described below. If, on the other hand, no impulse arrives at input A4 at the beginning of a decade, the impulses arriving at input ^ 3 of the adder remain ineffective for the calculation.

The comparison device V is set so that the comparison result is not deleted after one decade, but is used again when comparing the following decade. This is necessary if z. B. after a decade in which the subtrahend was greater than the minuend, a decade follows in which the subtrahend is equal to the minuend. In this case, in the second decade, the subtrahend must also be displayed as being greater than the minuend, since the minuend is reduced by the value 1 due to the tens carryover

got to. If the comparison device is not deleted after a decade, the comparison result remains "Subtrahend greater than minuend" exist for the next decade, and nothing about that changes State of the comparison device if, when comparing the following decade, both values are considered to be equal to be established.

Then, after the marking has passed the extraction head E 1, the markings run past the extraction head E 2, which sends a pulse to the input A 1 of the adder A when a marking is detected. Simultaneously with this, pulses corresponding to the combination of the value 10 per decade are passed from the pulse generator C to the input A 3 of the adder A via the extraction head C 1. If input A 4 of the adder received a pulse before the decade was offset, the value 10 is added to the corresponding minute-end decade when the following decade is run through. If, on the other hand, no pulse arrives at input A 4 before a decade has been offset, the pulse series of this decade runs from the removal head ET. from unchanged by the adder.

The appearing at the output A2 of the adder A pulse series (in each decade of 10 of increased minuend) is the Subtraktionswerks 5 passed to the input 6 "1 *. Simultaneously therewith, the subtrahend is sensed on the storage drum B by the removal head B 2 and the pulse train to the input S3 of the subtraction unit 6 "passed. The sensing of the subtrahend by the sensing head 52 takes place in such a way that the pulses of the same positions of mutually corresponding decades arrive at the input S1 and the input ^ 3 at the same time. In the subtraction unit, the subtrahend is deducted from the minuend for each decade. The minuend cannot be smaller than the subtrahend in any decade, since a smaller decade of the minuend was increased by the value 10 in the previous work steps. Furthermore, before a decade is offset, a pulse is sent from the auxiliary memory D via the extraction head D 2 to the input S5 of the subtraction unit if it was found in the previous decade that this minute-end decade was smaller than the associated subtrah-end decade. When a pulse arrives at the input 6 * 5 of the subtraction unit, the two-way carry is triggered before the calculation of a decade begins, so that when the first position is calculated, in addition to the subtrahend, the value 1 is also deducted. The result obtained in the subtraction unit S runs through the output 6 "2 to the entry head £ 4, which at the same time erases the markings of the end of the minute with the entry of the result pulse series, so that the difference between the two numbers results on the drum behind the entry head £ 4 .

2 shows the pulse sequences of a subtraction for the following example: the number 788 (subtrahend) is to be subtracted from the number 497 (minuend); ie a higher number must be subtracted from a lower number, so that there is a negative difference. This example was chosen to illustrate how the negative difference that appears first with its complement value is converted from the complement value to the true negative value. Field V shows the pulse train in the comparison device, field A the pulse train in the adder and field S the pulse train in the subtracter. The same applies here to the passage of the pulses through the individual computing devices, as was stated above in the description of the adder. Column 5 ″ represents the minute end, column M the subtrahend, column U the carry, column if the auxiliary pulse, column 10 the combination of tens, column Z the intermediate result (the minute ends partially increased by 10) and D the difference.

In the one-decade, the comparison device determines that the minuend with the number 7 is smaller than the subtrahend with the number 8, and triggers an auxiliary pulse H. This auxiliary pulse H reaches the adder A before this decade is offset and causes the number 10 to be added to the minute end 7, so that the number 17 results: This intermediate result 17 reaches the subtraction unit, in which the subtrahend 8 is derived from the intermediate result is deducted and the difference 9 results.

In the second decade (tens place) the subtrahend 8 is to be deducted from the minute end 9. The comparison device determines that the minuend is greater than the subtrahend, and the comparison device is adjusted accordingly. This result is not influenced by the previous setting of the comparison device to "smaller". Since the minuend was determined to be greater than or equal to, no auxiliary pulse H is triggered and thus no value 10 is added to the minuend 9 in the adder A , so that the value 9 appears again in the intermediate result. Before the intermediate result 9 runs through the subtraction unit, the two-way carry is triggered by an auxiliary pulse from the previous decade, so that the minute end is reduced by the value 1. Subtracting the subtrahend results in the value 0 as the difference for the second decade.

In the third decade (hundred digit) the subtrahend 7 is to be subtracted from the minuend 4. The comparison device determines that the subtrahend is greater than the minuend and thus triggers an auxiliary pulse H at the end of the comparison. This auxiliary pulse adds the value 10 to the end of the minute 4, so that the intermediate result is the value 14, from which the value 7 is deducted in the subtraction unit, so that the value 7 appears as the difference. A two-way carry is triggered for the following decade, as already described for the first decade.

If it is assumed that the third decade is the highest digit of the values to be subtracted, then remains for all following decades, the markings neither for the minuend nor for the subtrahend have, the ■ two-way carry exist. This also means that the comparison device remains in the position indicating that the minuend is less than the subtrahend. These two characteristics indicate that the The difference obtained between these two values is not the real value but the complement value represents. To determine the true difference between these two values, it is possible to after determining the difference this from the value 999. . . deduct or re-invoice to be carried out, whereby the minuend and subtrahend are swapped.

Another possibility is to subtract one value from the other value, and vice versa, at the same time if the storage capacity of the two memories E and B is dimensioned so large that the two values are based on them

£ 09 508/211

Il

Stores can be applied one after the other, namely in the memory B first the number 5 "(Fig. 2) and then the number M, while on the storage drum B, conversely, first the number M and then the number JT is entered. Between the two numbers on a storage drum is provided by at least one to two decades a gap. Furthermore, the comparison device when switching is from the deleted a number to the other. in the subsequent passage of the so arranged pulse trains, the number S of the number M is first and then the number M of subtracted from the number 5 *, so that two differences arise, one difference being the real value and the other difference being the complementary value has already been described above. Furthermore, the real difference is also indicated by the auxiliary marking gen Ή is displayed.

Similar to how several differences can be formed in one run, the addition in the corresponding multiple sums can also be obtained from a run of several rows of items. It is only to be noted here that between the individual summands for the respective sums a A distance of several decades is maintained, so that one sum is not included in the following sum can run.

Mechanical construction

The removal and re-entry in the results memory E cannot take place at the same time in one and the same place. For this reason, viewed in the direction of rotation of the results memory, the respective entry head is located behind the corresponding removal head. Thus, only markings can reach the entry head which were read from the removal head at a previous point in time and passed on to the arithmetic unit; these markings are therefore no longer required for further billing and can therefore be deleted. By entering the result markings in a different place than the one at which the markings were taken, the result mark row leads to a lead over the mark rows of the other stores. Through the two-time removal and re-entry of the markings is produced while adding a two-time advance of the mark row in the result memory E. The arrangement of the Entnehme- or insertion heads at the result memory E is embodied such that the mark row by two or more decades leads. This lead is compensated by the fact that the drum of the results memory is driven at a reduced speed compared to the other memories so that after one revolution the lead has equalized again and the marking positions with the corresponding marking positions on the other memories arrive at the first removal head at the same time .

The mechanical structure as well as the drive of the storage drum is shown in Fig. 6. It is in the following Description assumed that the control is carried out by punch cards and the result is punched in cards again.

The two memories B and D, the pulse generator C and a pulse generator /, which will be returned to below, are combined to form a drum and are driven by the motor 50 via the shaft 51 at a constant speed. On the shaft 51 there is also the gear 52, which drives the storage drum E for the result via the gear 53. The translation between the gearwheels 52 and 53 is set in such a way that this compensates for the advance mentioned above.

At the right end (Fig. 6) of the shaft 51 is the clutch disc 54 of the single-turn clutch. The single-turn clutch is controlled by means of the magnet K, whose armature 55 (FIG. 7) is pivoted upwards when it is excited. As a result, the pawl 56 mounted on the clutch disk 58 is released and pivoted against the clutch disk 54 by a spring (not shown). The clutch disc 54 has a notch 57 into which the pawl 56 can fall. The single-turn clutch can only be effective in a very specific position of the shaft 51, namely when the notch 57 is located under the pawl 56. This notch 57 ensures that the output is always coupled in a very specific position to the storage drums. The clutch disk 58 is located on the worm shaft 59, which drives a shaft 61 via the worm wheel 60. A known card transport mechanism, not described in detail here, is driven by the shaft 61 in such a way that one revolution of the shaft 61 corresponds to the transport of the cards by an entire card.

On the shaft 61 there are also cams 62 which actuate the cam contacts η 1 to 8. These lie one behind the other, which is why only the foremost contact can be seen in FIG. 6. These cam contacts have the task of controlling the transfer of the values to be calculated from the cards to the storage drums and the punching of the result in the cards. Their operation will be described later.

When adding up several items, each taken from a card, the clutch magnet K is constantly energized and the one-turn clutch is constantly engaged, so that the cards pass under the sensing station at a constant speed. If, on the other hand, the sum train is triggered by a card, the coupling magnet K is briefly de-energized to punch the card, its armature drops and stops the card transport so that the individual positions can be punched.

The marking is stored on drums B, D and E by magnetizing individual points of a magnetizable layer. The arrangement for applying or removing these magnetic markings from a storage drum is shown in FIG. 11. On the storage drum 73 is the magnetizable layer 72, which passes under magnetic heads. These magnetic heads consist of a winding 70 through which a magnetic core 71 is passed. The magnetic cores 71 are designed in such a way that they form a narrowly localized magnetic field just above the magnetizable layer 72 which, when the winding is excited, leaves a corresponding marking in the magnetizable layer. Conversely, when a magnetic marking passes under the core 71, a magnetic flux can be induced in the core 71, which generates a voltage in the winding 70.

Fig. 10 shows the arrangement of the pulse generators C and J. The two pulse generators are located on a drum 64 made of insulating material, on the circumference

small conductive contact pieces 65 are embedded. The distribution of the contact pieces on the circumference of the drum 64 is shown in FIG. 9, which represents part of the development of this drum 64. The hatched areas indicate the conductive contact pieces. These contact pieces run under the sensing brushes il to Ϊ14 as well as Cl and C 2. The sensing brushes are designed as double brushes, with one brush serving to supply power and the other brush to discharge power. As FIG. 9 shows, the brushes are not located in a line, but are distributed over several points on the circumference in order to achieve a different time sequence of the pulses. On the left edge of FIG. 9 there are consecutive numbers 1 to 6. One step from one number to the next corresponds to a marking position or a binary digit, and six marking positions are each assigned to a decade (cf. FIGS. 1 and 2). The two sensing brushes C 1 and C 2 represent the two removal heads C1 and C2 according to FIGS. 5 and 8. The sensing brushes i 1 to il4 are used to control the arithmetic logic units, which will be returned to below in the description of the individual circuit diagrams. For the further description it should be pointed out that all of the cam contacts marked with η are actuated by cams on the shaft 61, and all of the contacts marked with i are actuated by cams on the shaft 51 or the pulse generator tronimel 64.

The circuit diagram for removing the magnetic markings from the storage drums is shown in FIG. 12 shown. The circuit is mainly used for regeneration and precise limitation of the pulses generated by the pick-up heads.

35 Circuit diagram of the sampling heads

The circuit for the pick heads is shown generally in FIG. The voltage generated by the extraction head designated here by Z is conducted to the base of the transistor 10 via the capacitor Ca and the rectifier Gl. The capacitor is charged from minus via the transistor 10 with the aid of the external voltage 50 V. Charging increases the voltage at the emitter. However, it collapses again as soon as it has reached a level corresponding to the data of the transistor, whereupon the capacitor Cb (which is connected to the right side of the capacitor) discharges via the emitter, while at the same time an amplified pulse is formed on the collector circuit of the transistor. The size and duration of this current pulse is independent of the input pulse and is only determined by the dimensions of the capacitor Cb and the base resistor R 1. A uniformly large pulse therefore always appears at the collector of the transistor, irrespective of the pulse size and pulse length that was passed from the removal head Z to the transistor.

60

Circuit diagram of the entry heads

The circuit for the insertion heads is shown in FIG. The circuit is designed in such a way that a current surge flows through the winding of the insertion head at every marking position that passes under the insertion head designated here as Y. The current surge is directed by a control pulse in such a way that it flows in one or the opposite direction, depending on whether a control pulse or no control pulse is received. Corresponding to the passage of the current through the winding of the recording head in two different directions, the magnetic head records markings with different polarities on the storage drum. The induction force of the magnetic head is so large that markings which are located under the magnetic head at the time of entry and have opposite polarity are erased and the new marking is entered with the corresponding polarity. For the removal, however, only markings with a certain polarity can be effective, which the rectifier E1 (FIG. 12) allows the current to pass through only in one direction.

The flow of current through the magnetic head is controlled by means of the tube 11 in connection with the pulse contact i 1, which is controlled by the pulse generator /, as already described above (cf. FIGS. 9 and 10). The pulse contact i 1 is shown in FIG. 9 and leads to an output arrow X 2. All inputs X 2 shown in the other circuit diagrams are connected to this output X 2. During the offsetting, the cam contact w3 has closed, so that a voltage of + 50 V is constantly applied to the pulse contact i 1 during this time.

The pulse contact i 1 is shown in simplified form in the circuit diagram (FIG. 19) as a contact, while in the practical version it is designed as a double brush (FIG. 9). The pulse contact i 1 closes each time a marking position is passed under the entry head. Appears at the same time so that no pulse at the grid of the tube 11, so it remains suspended because the ear via the resistor R 3 to the lattice of the ^R a negative bias voltage is set. Thus, only a current from X2 with + 50V can flow through the winding of the magnetic head and the resistor R2 to + 200V. If, on the other hand, a positive pulse arrives at the grid of the tube 11 at the same time as the closing of the pulse contact il , the tube becomes conductive and current can flow from zero through the tube 11, the resistor R 2 to + 220V and on the other hand from zero the tube 11, through the winding of the insertion head Y to X2 with + 50 V flow. When the pipes 11 are live, an opposite current flows through the insertion head Y than when the pipe 11 is blocked and marks with different polarity in accordance with the different current flows.

Adder

A purely binary working adder is to be shown with reference to FIG. B. can be used in the machine according to the invention. The designations A1 to A 5 correspond to the five inputs and outputs of the adder A in FIGS. 5 and 8. The inputs A 6 and A 7 are used together to block the input A3. The task of this blocking is described in more detail below. The two values to be offset are introduced at the two inputs A1 and A3. The pulse arriving at input A1 runs through the primary winding of transformer TrI and the pulse arriving at input through the primary winding of transformer Tr 2.

The two transformers TrI and Tr2 have the same structure and each have two secondary windings. One end of one of the secondary windings of TrI is connected to a line via the resistor R 5.

closed, which is designated by X 2 and connects to the line X 2 in FIG. The line X2 has voltage from the pulse generator contact il only as long as a marking position is sensed, as has already been described above. The opposite end of the other secondary winding of Tr 1 is connected to negative voltage via the resistor R 4t. If no pulse runs through the primary winding of the transformer Tr 1, there is a negative voltage at the output of one secondary winding with the resistor i? 4 and a positive voltage when the pulse contact il is closed at the output of the other secondary winding with the resistor R 5. The direction of winding of the two secondary windings is determined so that when a positive pulse arrives in the primary winding, the positive potential in the winding with the resistor R 5 is reduced to such an extent that a negative potential and the negative potential in the winding with the resistor i? 4 is increased so far that a positive potential appears at the output of this winding. In summary, it can be said that when no pulse arrives, one winding has a positive potential and the other winding has a negative potential, while when a pulse arrives, the potentials differ. reverse the ratios of the two windings. The same applies to the transformer Tr 2 with the input A 3 and the transformer Tr 3 for the two-way carry.

The representation is assumed in the circuit diagram (FIG. 14) in such a way that the upper output of a transformer has positive potential when a pulse arrives, while the lower output of the transformer has positive potential when no impulse arrives. As always in the further description, those lines of the secondary windings of a transformer which lead to one or more rectifiers are referred to here as outputs. The upper output of the transformer Tr 1 is led by the rectifiers G 1, G2 and G 3 to a voltage divider, which receives positive voltage via the line Z 2 via the resistors R7, R8, R9 and RIO when the pulse contact i 1 is closed. The rectifier Gl is connected to the resistor R 7, the rectifier G 2 to the resistor R 9 and the rectifier G 3 to the resistor R 10. The lower output of the transformer Tr 1 is connected to the resistor R 8 via the rectifier G 4. The current flow through the resistors R7, R8 and R9 takes place via rectifiers G14, G15 and G16, which are connected together to the resistor R 12, which in turn is connected to minus.

At a point between the resistors, a very specific voltage is established which can be taken from output A 2 for further control purposes. The resistances are dimensioned in such a way that when the pulse contact i 1 closes, a positive voltage appears at the output A 2. If, on the other hand, one of the transformer outputs has a negative voltage, this voltage is passed on to the voltage divider via the corresponding rectifier and the voltage at the associated resistor is thus reduced. The reduction in the voltage at output A 2 can only occur if there is a negative potential across all three resistors R7, RS and R9. In all other cases a positive potential remains at A 2 , since the negative potential is blocked by the rectifiers G14, G15 or G16. The voltage conditions are in this case so determined that at the appearance of a negative voltage to the resistors Rl, R 8 and R 9 is converted the positive potential at the output of A 2 to a negative potential.

The resistor R 10 and the resistor i? 13 via the rectifier G17 form a special voltage divider, which is connected to the previously described voltage divider via the rectifier G18. The voltage ratios of the second voltage divider behave similarly to those of the voltage divider described above. The potential existing between the resistors R 10 and R 13 is conducted to the grid of the thyratron 12, which is used to store the two-way carry, which will be returned to below.

If it is assumed that a pulse only arrives at the transmitter TrI , then, as already described above, a positive potential appears at the upper output of TrI , while the lower output has a negative potential. In this case, the voltage across the rectifier G 4 across the resistor R 8 is reduced. In the case of the transformers Tr 2 and Tr 3, the upper outputs carry negative voltage and the lower outputs carry positive voltage, since no pulse has arrived at the primary windings of this transformer. Due to the negative voltage at the upper outputs, resistors R 7 and R 8 receive negative voltage via rectifiers G 5 and G10. A negative voltage is applied to each of the resistors R7, R8 and R9 , so that a negative potential appears at the output A 2. The positive potential at the rectifiers Gl G 2, G 3, G 8 and G12, on the other hand, has no influence on the voltage divider, since the effect of the positive potential is blocked by the rectifier.

By connecting the various rectifiers to the two voltage dividers, the following voltage ratios result: If a pulse appears on only one of the three transformers TrI, Tr2 and Tr 3, the potential at A 2 becomes negative, but a pulse appears on each of two transformers, so the potential at A2 becomes positive. In the latter case, the positive potential at A2 is conducted to the thyratron 12 via the rectifier G18, so that it is ignited by the positive potential on its grid and thus prepares the two-way carry. If all three transformers each receive a pulse, the potential at A2 becomes negative, while, on the other hand, the potential at the grid of the thyratron 12 via the voltage divider with the resistors R 10 and R 13 becomes positive.

The following results for the further control of the computing device: If only two pulses are present, see FIG. 3), the entry in the results memory must be omitted. The positive potential at A 2 can thus be used to block or delete the markings in the results memory. If two pulses are present, the two-way carry (cf. FIG. 3) must be prepared, which is achieved via the rectifier G18 and the thyratron 12. If there are three pulses, the entry in the results memory must not be blocked (A 2 negative), but the two-way carry must be triggered, which is done via the voltage divider with resistors R 12 and R 13.

By igniting the thyatron 12, as already mentioned above, the two-way transfer is prepared. The pulse contact i 5 has closed for this purpose during the entire time of the pulse generation that when the thyratron is ignited, a current pulse from zero flows through the thyratron via the resistor R 11 and the contact ^ 5 to XX .

The line Xl leads to the output Xl in FIG. 19. During the offsetting process, there is positive voltage on the line X1 due to the closing of the cam contact w3. After the transmission has ended, the pulse contact i 6 briefly closes, which means that the capacitor Ce can be charged in accordance with the voltage drop across the resistor R 11. Before the impulses for the next position arrive, the impulse contact i 5 opens briefly and the thyratron goes out. Simultaneously with the arrival of these impulses, the impulse contact i 2 closes so that the charged capacitor Ce can discharge via the primary winding of the transformer Tr 3. A pulse thus only appears in the transformer Tr 3 if the thyratron 12 had ignited when the previous pulses were offset and the capacitor Ce was charged. If this is not the case, the capacitor Ce has no charge and no pulse can arrive at the transformer Tr 3.

As already mentioned above, the tens carry between the decades or decimal places is achieved by triggering the two carry before the offsetting of a decade begins, whereby "one" is carried over to the next decade. This transfer takes place when a positive pulse arrives at the input A 5, from where the pulse is passed on via the rectifier G19 to the grid of the thyratron 12, which ignites.

As mentioned above, the input A 3 must be blocked for the pulses so that the pulses of Al run without change by the computing device at each of arithmetic operations. This blocking takes place by sending a positive pulse to the input ^ 4. The means for applying and measuring the pulses at the inputs A 6, A 7, A 4 will be described in more detail later.

If a positive pulse is sent to input A 4 at the beginning of the offsetting of a decade, the cathode of the thyratron 13 receives positive voltage and cannot be ignited or extinguished. 4P The thyratron 13 is normally blocked because its grid is connected to minus via the resistor R 16. Once the thyratron has fired, a current flow from neutral through resistor R 15 represented by the thyratron 13 via the resistor R 17 and the closed at this time impulse contact i 10. The pulse contact i 10 is closed during the decade and only opens briefly at the beginning of the calculation of a decade in order to extinguish the ignited thyratron if necessary. A voltage divider with resistors R 20 and i? 21 is connected to the cathode of tube 13 and a voltage divider with resistors R 18 and i? 19 is connected to the anode. The two voltage dividers lead to the line with negative voltage. The resistors are dimensioned such that when the tube 13 is not energized, a positive voltage is established between the resistors R 18 and R 19 and a negative voltage is established between the resistors R 20 and R21. As soon as the tube 13 has ignited, a voltage drop occurs across the resistor R15 as well as across the resistor 17. As a result of these two voltage drops, the voltage on the cathode increases while the voltage on the anode decreases. This change in potential affects the voltage dividers, so that the voltage between resistors i-18 and R19 becomes negative and between resistors R 20 and i-21 becomes positive. The potentials are thus reversed when the tube 13 conducts current. The voltage between the resistors R 19 and i? 18 is passed on via the rectifier G20 to the upper output of the transformer Tr 2 , the rectifier only allowing a negative voltage to pass through for blocking. The voltage between the resistors i? 20 and R21 is passed on via the rectifier G21 to the lower output of the transformer Tr 2, the rectifier only allowing a positive voltage to pass through for blocking.

The tube voltage divider circuit just described can also be used in conjunction with the transistor 10 instead of the transformer Tr. The impulse from the extraction head Z (Fig. 12) is in this case directed to the grid of the tube 13. The decrease from the voltage dividers leads to the rectifiers in a similar way as the outputs of the transformers Tr.

Subtraction mechanism

The purely binary subtraction unit (FIG. 16) is constructed in a similar way to the adder unit described above. The pulse series of the minuend arrives at the input vS ¹ 1 of the primary winding of the transformer Tr 4, and the pulse series of the subtrahend arrives at the input 6 "3 of the primary winding of the transformer Tr 5. The transformer Tr 6 is used in a similar way to the adder for forwarding the two-carry impulses. the two secondary windings of the transformer Tr 4 are each with one of its ends through the resistor. ?? 30 on the line Z2 and the resistor R37 to the negative line. the same applies to the transformer Tr 5 with the resistors R 31 and R 38 and the transformer Tr 6 with the resistors R 39 and R410. The six outputs of the three transformers are in turn connected by means of a rectifier to a voltage divider with the resistors J-32 to R36 and R 42. The three transformers of the Subtraction units are constructed in the same way as the three transformers of the adder unit and, when a pulse arrives at the primary winding, have a po at the upper output positive and a negative voltage at the lower output, while if a pulse does not arrive, the lower output has a positive and the upper output has a negative voltage. The line Z 2 is the same as at the adder to the pulse contact il (Fig. 19) is connected and thus only receives voltage as long as a mark position is offset. The resistors R 32 to R36 are connected at one end to the line Z 2 and are combined at their other end by the rectifiers G 48 to G 51 on the one hand and by the rectifiers G 52 to G 55 on the other. The resistances of the voltage divider are dimensioned in such a way that a positive voltage is formed on the connecting lines between the resistors 32 to R 36 and the resistor R 42. If, on the other hand, a negative voltage is supplied via one of the rectifiers which are connected to an output of the transformer, the potential on the corresponding strand becomes negative as a result.

The lines leading down from the resistors R 32 to R 34 are combined via the rectifiers G 48 to G51 and lead to the output S 2, from which the pulses for marking the result memory are taken. The lines leading down from the resistors i? 33 to i? 36 are also combined by the rectifiers G 52 to G 55, the outputs of which lead to the grid of the thyratron 15 for the two-way carry, ".." -. '. .

909 508/211

The output "S" 2 has a positive voltage as long as positive voltage can be fed to the output via one of the four rectifiers G 48 to G 51. If, however, none of these four rectifiers has a positive voltage, the output S 2 has a negative voltage The same applies to the grid of the thyratron 15 in connection with the rectifiers G 52 to G 55.

The upper output of the transformer Tr 4 is connected with the rectifier G 57 to the resistor R 35 and with the rectifier G 58 to the resistor R 36. The lower output of the transformer Tr 4, however, is connected via the rectifier G 59 to the resistor R 32, the rectifier G 60 to the resistor R 33 and via the rectifier G 61 to the resistor R 34. If a pulse thus appears at input 6 "1, the lines leading down from resistors R 32 to R 34 become negative, while the lines leading down from resistors i? 35 and i? 36 remain positive. In a similar manner like the outputs of the transformer Tr 4, the outputs of the transformer Tr 5 with the rectifiers G 62, G 63, G 37, G 38 and G 42 are connected to the lines leading down from the resistors. The same applies to the transformer Tr 6 with rectifiers G 43 to G 47.

This rectifier circuit results in a positive pulse at output 5 "2 if only one of the three transformers receives a pulse. Furthermore , a positive pulse also appears at output S2 if all three transformers each receive a pulse. In all other cases output 6 "2 remains negative. On the other hand, a positive pulse appears on the grid of the thyratron 15 if only the transformer Tr 5 or only the transformer Tr 6 receives a pulse or both transformers each receive a pulse. In addition, a positive pulse appears on the grid when all three transducers each receive a pulse. This pulse distribution on the output S 2 or the grid of the two-carry thyratron 15 corresponds to the calculation scheme according to FIG. 4, which was set up for the binary subtraction.

The output 6 * 2 can be used to control the marking entries in the result memory in such a way that a marking is entered in the memory when a positive pulse appears, while existing markings are deleted if there is no positive pulse. This deletion takes place in the procedure described in the section »Entry head« by the fact that if the positive pulse is missing, the direction of the current in the entry head is reversed, thereby deleting an existing marking. A positive pulse on the two-carry thyratron 15 causes it to fire and, when the next marking position is transmitted, sends a two-carry pulse to the transformer Tr 6.

The operation of this thyratron together with the resistor R 41, the capacitor Cd and the pulse contacts i 4, iS and i8 is the same as with the adder described above. Similar to the adder, the blocking device for the pulse series works at input 6 * 3. The blocking action is achieved by applying a positive voltage to the input or before the beginning of a decade by a positive pulse at input .S "4. The pulse contact HO also corresponds to the same contact as in FIG. 14, but in the practical version for both circuits is provided jointly.

To take into account a ten-carry, the two-carry thyratron 15 is ignited before offsetting a decade via the input 5 "5 5- and the rectifier G 56 by a positive pulse, which means that the value 1 is subtracted from the end of the minute when offsetting the first Position is reached. As described above, output A 2

ίο with positive voltage a blocking or deleting effect on the marking of the result memory and the output 6 * 2 cause a marking with positive voltage. Both can be achieved in a simple manner in that the insertion head, which is connected to the output A 2 , has a winding with opposite winding directions than the insertion head, which is connected to S2 . Due to the different winding directions of the two entry heads, it is achieved that when carrying current

ao the Triodell (Fig. 13) the entry head connected to A2 enters a marking in the opposite polarity compared to the entry head, which is connected to the output S2 . Since only markings with a certain polarity are converted into pulses by the rectifier G1 of the removal head (FIG. 12), only the markings entered with this polarity can be effective for the readings.

Comparison device (Fig. 17)

The pulse series to be compared are sent to the inputs Vl and V2 of the two transformers Tr7 and Tr 8. The two transformers Tr 7 and Tr 8 each have two secondary windings and are constructed in a similar way to the transformers described above. The four outputs of the two transformers are connected via rectifiers G 70 to G 73 to two voltage dividers with resistors .ft "52 and R 60 as well as i? 53 and R 61.

The resistances of the voltage divider are dimensioned such that when the pulse contact is closed

, £ 1 (Fig. 19) and thus, when a positive voltage arrives on the line X2, there is normally a positive voltage on the two grids of the tubes 16 and 17. The upper output of the transformer Tr 7 and the lower output of the transformer Tr 8 are connected to the resistors 2? 52 and i? 60 existing voltage divider are connected, while the lower output of the transformer Tr 7 and the upper output of the transformer Tr 8 are connected via rectifiers G 71 and G 72 to the voltage divider formed by the resistors i? 53 and J261. When a pulse arrives at the transformer 7V7, a negative voltage appears at the rectifier G 71 and thus at the voltage divider with the resistors R 53 and 61, whereby the grid of the tube 17 becomes negative and blocks it. If, on the other hand, a pulse arrives at the transformer TrS, then a negative voltage is applied via the rectifier G 73 to the voltage divider consisting of the resistors i-52 and .R 60, so that the grid of the tube 16 becomes negative. In the absence of a pulse at the transformer Tr 7 , negative voltage is passed through the rectifier G70 and in the absence of a pulse in the transformer Tr 8, negative voltage is passed through the rectifier G 72 to the voltage dividers. This results in the following conditions: If no impulse appears on both transformers or if an im-

21 22

pulse, the brush is provided when the contact i 1 closes. The conversion of the sensing grid of the two tubes 16 and 17 is negative. As a result, if only the transformer Tr 7 receives a pulse at the brush b in a pulse series, then the commutator 21; he touches all his is the grid of the tube 17 is positive, during a time while the shaft 8, the 5 rotates once at fins 51 (Fig. 6) arrival of a pulse only at the transformer Tr. For each spei-grid on the drum the tube 16 becomes positive. A lamella is provided for the last decade. The tube 16 is designed as a thyratron and the tube 17 lamellae of the commutator 21 are formed via plug relay triode. If the grid of the tube 16 cords and rectifier G 74 the corresponding briefly positive, the thyratron ignites and brushes b, which sense the value to be calculated, are connected to a current flow across the two resistors ip. The rotor arm of the commutator 21 R 56 and 2? 57. This flow of current also remains open, one after the other, all lamellas on the contacts are maintained when the grid of the tube 16 is again chain with the cam contacts 5 to η 8 and the negative. On the other hand, it receives ill to il4 at a later pulse contact. The cam contacts point in time the grid of the tube 17 briefly a η 5 to n8 are actuated in such a way that when a positive voltage is sensed, the tube 17 is current- 15 leading to the first hole position of the card of the cam contact and thus closes the tube 16, so that »5, when the second hole position of the card is sensed, the thyratron goes out. With the decrease of the clamping the cam contact η 6 etc. is closed. The voltage on the grid of the tube 17, the current flow arrangement of the pulse contacts ill to il4 and also interrupted by these again, their time actuation can be seen from FIG. As already described above, the contacts 1 1 to 3 4 pass under the sensing heads practically as double pulse rows in such a way that brushes are formed which, via a contact piece track, touch the first position of the lowest decade of the pulse generator /. The four under the sensing heads and the highest position of the double brushes ill to il4 are arranged as the last decade under the sensing heads in such a way that the contact piece passes by when it passes through. Corresponding to the arrival of the first binary position a decade below the Marpulse at the inputs Vl and V 2 , the tube 'kierungskopf the first pair of brushes ill, ignited repeatedly at through-16 and then the second binary position again runs through the Tube 17 deleted. Depending on which pair i 12, in the third binary position the third transmitter alone receives a pulse, the fourth pair of brushes il4 is briefly ignited or not at the brush pair i 13 and when the tube 16 30 passes through the fourth or the end of the pulse series . If the tube 16 is ignited, it closes. The speed of rotation of the rotor arm of the last pulse arrived at the transformer TrS, commutator 21 is dimensioned so that, while the tube 16 is extinguished, it is the last pulse within the time in which a single hole has arrived at the transformer Tr 7 alone. Thus, when the tube 16 is ignited and the tube 16 is brushing at the input VX 35, all slats are swept over, the number that has arrived is displayed as the larger one, while the cam contact η is 5, as was the case with the tube 16 that was cleared above, the one at V2 incoming number as mentioned, only when sensing the first hole position the larger one is determined. If the two cards are identical, the tube 16 has not ignited during the pulse contact with the numbers. i 11 only if every first binary position is passed. After each decade has been compared, the cam contact i9 briefly closes a decade at the entry point. Depending on whether the head closes for a short time. This means that only one tube 16 has ignited or not, there is then a positive or negative pulse on the magnetic drum in the first output V 3. binary position of the decades if in the first the positive or negative pulse is given by the loc-hposition of the corresponding column of the map resistors 2? 56 to R 59 in connection with the 45 a perforation is sensed. Similarly, tube 16 is formed. If the tube 16 has not ignited, the calculations for each additional hole position are carried out so that the current flows from minus to plus only via that of the card.

Resistors 2? 56 to 2259, whereby the resistances After all the markings of the first hole are dimensioned in such a way that they are transferred between the counter positions of the card, the card stands R 58 and 22 59 results in a positive voltage. 50 with the second hole position under the row of brushes. However, has the tube 16 ignited, then the same time, opens so that the cam contact η 5, and voltage between the two resistors 2? 57 of the cam contact η 6 closes, the thus power and 2? 58 and thus also between the resistances via the pulse contact i 12 during the second pass R 58 and 2? 59 reduced, so that a negative voltage appears at the output of the rotor via the contact lamellae of the commu- V 3. 55 tators 21 conducts. In the following pass, _Λ7 -. ,, T-Xt. j τ- · ^ the marks of the second position of all columns device for removing and entering one after the other to the second position of the corresponding values in punch cards ^ _{transferred decades} . The same is repeated for The values to be calculated can, for example, be punched in the third and fourth positions with the cam contacts in punched cards, which are marked by a known % 7, nS and the pulse contacts il3 and il4. tes card transport, not described in detail here, to transfer the markings into the magnetic mechanism while running from a row of brushes with memory, four rotations of the memory of a single sensing brush per column sensed drum are required. In the subsequent fifth will be. The sensing by the brush row ER revolution setting off of the thus introduced follows in all columns simultaneously, but the one with the 65 value are befmdzelnen already in the result memory hole positions a column one after the loan value. After this rotation, another is sensed. It is assumed here; that-the further sixth rotation of the amount spei rather B cards according to the key in Fig. 20 are punched. erased and then a value from a new one. The cards run, as mentioned above, under card for offsetting with the result already in the brushes b , of which a value stored in memory is transferred for each column.

ΐ

As already mentioned above, Her result memory E does not run at the same speed as the "remaining memory and the pulse generator, in order to compensate for the lead required in the calculation. The lead or lag is only canceled out for one revolution, in which the calculation takes place, so that the result only reaches the first removal head E 1 correctly during this revolution. During the remaining five revolutions, however, the computing device is blocked so that the rows of marks on the result memory remain in the position entered in order that. after five revolutions the mark row on the result memory e returns to the correct position to the extraction head, the advance is determined by the insertion heads so that it takes up one fifth of the circumference of the storage drum e. it is thereby achieved that, after five revolutions the storage drum E the row of markings without changing the length on the drum mel has again reached its correct position in relation to the removal head E 1.

If the total train is triggered by sensing a special marking in a sum card lying between the post cards or by a known group control device (not described in detail here), the relay H (Fig. 19) is energized, the mode of operation of which will be returned below. By energizing the relay H , the contacts assigned to it h ^m to h ^{Yl are} actuated, which switch the removal device in such a way that from it the marking of the result drum is passed on to the punching mechanism, which punches the sum card according to the result.

As long as the relay H has not responded, the contact h ^{m is} closed and thus the contact plate of the brushes is connected to voltage. Furthermore, the contact h ^{lv is} closed and thus conducts the current via the pulse _{contact ill to il4 and the cam contact M 1} ^111, which is switched over during the transmission, to the magnetic head B 1, which in this case is used as a recording head. The magnetic head B 1 also serves to erase the previously entered and removed markings, as a direct current is fed through its winding during a total revolution of the storage drum after the calculation, which creates a permanent magnetic field on the magnetic head that erases the individual markings on the storage drum .

As soon as the relay H responds, it separates the contact plate of the brushes from the voltage ^{with the contact A 111.} Furthermore, the contact h ^1Y opens the circuit to the magnetic head B1, so that sensing and entry in the value memory is prevented. By switching over the contact h ^x the extraction head El is connected to the pulse contacts ill to i'14 and by closing the contact h ^yi positive voltage is applied to the second grid of the tube 18. The tube 18 is a thyratron with two grids on the other hand formed and lights only when both gratings positive voltage hold. if only one of the two gratings negative voltage, the tube remains blocked. as long as the Kentakt / i ^YI has opened, the top grating receives the tube 18 via the resistor i? 70 negative voltage, so that the tube 18. cannot ignite.-A tube 18 is provided for each column to be punched.

gnettrommel be "read, so to reinforce and to store that the perforated magnets L address and the associated punch voters can be adjusted. The magnets L serve here not directly for punching the card, but in a known manner just as voters magnets the punch for the Select perforation while the actual perforation process is carried out in a known manner by means of a punching yoke driven by the alotor.

After the relay H has responded and its contacts A ¹¹¹ to h ^Y1 has actuated accordingly, the first position of the lowest decade is assigned to the sensing button £ 1. Furthermore, at this point in time the first position of the card to be punched is under the punching mechanism, and at the same time the cam contact η 5 closes with the passage of all the first positions of the various decades, the pulse contact ill briefly closes and conducts the one triggered by sensing a marking Impulse from El to commutator 21, the rotor of which is located on one of the lamellae assigned to the decade being sensed. A tube 18 is connected to each of the lamellae of the commutator 21. Corresponding to the sensing result of the first position of all decades of the results memory, the associated tube 18 on the lower grid receives a positive pulse and ignites, since the upper grid also has a positive potential when the contact h ^ ^{1 closes.}

The tubes 18 are designed as thyratrons. The flow of current is maintained after ignition, even if one or both grids show negative potential again. Only when the circuit is interrupted by the cam contact w9 are the circuits to the tubes 18 interrupted, so that all tubes go out again. The cam contact η9 is controlled by the punching mechanism and opens briefly after each actuation of the punching yoke. The perforation magnets L are located in the circuits of the tubes 18 and are excited when their associated tube 18 becomes live. As already mentioned above, the brushes b are also connected to the lamellae of the commutator 21. In this case, however, they remain ineffective because the contact h ^{m has} opened and the rectifiers G 74 prevent a reverse current through the brushes b.

Switching of the computing device
on addition or subtraction or division

The switching device for the selection of the arithmetic operation is shown in FIG. Fig. 15 shows only the contacts that are used for switching, as well as the outputs and inputs to the individual computing devices. The line A 1 in this case corresponds to the input A of the adder 1, the "Line A 2 the output of the adder A2 etc. Similarly, the inputs and outputs for the comparator V, for Subtraktionswerk Ji as well as for the removal or Entry heads B, C ₁ D and E. The contacts f ⁿ to i ^XI serve to switch the computing device from addition to subtraction.

Calculation process control

With both types of calculation, both with addition and also with subtraction, takes place after each sensing

25 26

offsetting a card. Several cards are drumming tension. During this rotation, they are offset one after the other until the group is followed by the calculation of the transmitted numerical value, control device or by means of a special hole in Result in which is 5 with the special hole that the value to be offset is to be punched as a negative card. For both accounting items features, a current may voltage types of minus So a cyclic working through the closed at this point contact always gives way, since bt for the removal and Calculation of M ₁ ¹¹ and which entered through the special hole brush value the same time is needed. For that through the relay. T flow so that it responds. However, in many cases a corresponding and longer time is required by means of its contact t ¹ via the cam end, since the card does not make a holding _{circuit in such a way as to contact W 4} ^11. It can be quickly punched through as the sensing will trigger the relay T , as is done above. described the computing device on subtraction

The circuit diagram for controlling the arithmetic operations switched. After the calculation is finished,

shows Fig. 19. The relay K corresponds to the coupling 15 opens briefly the cam contact M ₄ ¹¹ and under-

Lungsmagneten K for the one-turn clutch in breaks the holding circuit to the relay T,

Fig. 6. The relay H is used in conjunction with the "~ so that it drops out and the computing device again

Brush bh to determine a cumulative draw hole in which switches to addition.

a card and triggers the addition and subtraction when a card is sensed, the brush bh by the sum move. The relay T controls the contact 20 a special hole, so it is indicated by this that the clock t and is in connection with a brush ibi is to be drawn in total. During the sensing special hole in the card firmly, indicating whether it has the cam contact M ₁ ¹¹ closed, so that a is a positive or negative value. With current from zero through the cam _{contact M 1} ¹¹ , the response of the relay T , the arithmetic device brush bh, through the relay H can flow. The rescue through the contacts i ¹¹ to t ^XI (FIG. 15) on sub- 25 lais H responds and is switched via its contraction. clock h ¹ a circuit. Assume that

Let us first assume that the special hole in the machine for the total move is in the last

Items are offset, with the negative or th hole position and the cam contact η 2,

positive values can occur. The with η denoted during the sensing of the last position

Neten contacts are cam contacts that opens cyclically at 30, closes before relay H drops out. It

the circulation of the card transport system is operated, the relay H remains on the holding circuit

as already described above. The closing with the contact h ^l and the cam contact η 2 so

ßungszeiten these contacts compared to the long excited until the end of the following sensing game

Transport of the cards or to the circulation of the memory in turn the cam contact η 2 opens and the

drum shows the cam diagram in Fig. 21. A 35 holding circuit is interrupted. By addressing them

The division of the cam diagram corresponds to the relay H , the contact h ^{u is also} reversed.

one revolution of the shaft 51 or the accumulator.

drums B, D and the impulse generator C and I or at the beginning of the next card game he has

the transport of the card by one position row. Cam contact M ₁ ¹ switched. At this time

Six of these revolutions correspond to one revolution. The previously sensed card is located with the

rotation of the shaft 61 or the transport of the first hole position shortly before the punching mechanism. ' Wäfa-

Card by one card division, with which the distance from rend to the following revolution of the storage drum

a certain point of a card up to the same, as described above, by the punching

Point of the next card is meant. The cam yoke of the cam contacts η 9 is closed and thus

Contact "9 becomes a difference from the 45 via the contacts M ₁ ^1, h ^n, and" 9 of the circuit to

Cam contacts n2 to w8 are not maintained by a cam on the clutch magnet K , where-

on the shaft 61, but remains effective with every downward movement of the one-turn clutch. - Of the

Punching yoke operated, assuming that the cam contact η 9 opens shortly before the second change

Stanzjoch during three revolutions of the shaft rotation of the storage drum, so that after one revolution

51 moved up and down once. 50 rotation, the clutch magnet K drops out and the em-

When a card is started, the touring clutch is uncoupled. During this

Cam contact nl ¹ toggled and thus a rotation is the first hole position of the sensed

Circuit through the contact h ^u to the clutch card pushed completely under the perforation and

magnets K made so that it remains energized. at the same time as the first loah position of that

At the same time, the cam contact η5 55 results memory to the perforation magnets, conducted,

(Fig. 18) and thus selects the first row of positions for, as already described above, since during

the transmission from, as already described above, this rotation closes the cam contact m5 and

was ben. Relay H closes for the next row of positions.

the cam contact η 6 and then η 7 and «8. The cam contacts η 5 to μ 8 remain uncoupled for 60 card transport units each time, so that the card stands still for one full revolution of the storage drum and the perforation can be carried out. After closed. After the transfer of the values in which the perforation was carried out, the memory closes again./? the cam contact η 3 (Fig. 19) briefly closes the cam contact η 9 and thus releases this and thus releases voltage to the line X 1 and, in the case of the transport mechanism, for the transport of the card to close the pulse contact i \ voltage to the 65 hole position. During this transportation trol cable X 2. The two lines Xl and X2 are following the transfer of the second Loehposition followed to supply the computing and Vergleichsvorrichtun- punching. This continues up to the gene, as already described above. The holes in the last hole position.

Calculating devices are thus only obtained via the After the last hole position has been punched,

Cam contact η 3 for one revolution of the memory 70, the cam contact W ₁ ¹ goes back to its starting point

position back and thus establishes a direct circuit to the Kuppktngsmagneten K. This excites and couples the card transport, which remains switched on until the relay H: - $ detects a total draw hole again. "

Bibliography: For an easier understanding of the operation of the binary: adder and the binary subtraction unit -.'- refer to the chapter "Rectifiers" of the book "Synthesis of electronic computing and control circuits". By the staff of the Computation Laboratory, Combridge, Massachusetts, Harward University-, Press 1951.

Claims (1)

PATENT.ANSP.RU C HE: 1. Electrical serial calculator for performing additions in ■ decimal-binary representation using a .. correction process and supplying the numbers to be calculated to the arithmetic unit in pulse series, characterized in that a binary adder (A) the decades of the addend one after the other and the binary digits within the decades also added one after the other to the decades of the eye end corresponding to their place values and both a voice below as well as one with or above the decimal value 10 is represented in binary form, that following the addition the recorded result is compared in a comparator (V) and when the decimal value 9 is exceeded in the result, a pulse from the comparator (V) is stored in the adder as a binary carry before adding up the next higher decade, and in a binary subtraction unit (6 *) depending on the result exceeding the value 9 automatically from the comparison the value 10 is subtracted and the difference is given in binary representation within the successive decades of the result. 2. Electrical serial calculator for performing subtractions in decimal-binary representation, characterized in that the minuend and subtrahend are compared in a comparator (V) for a decade and then the minuend in a binary adder (A) depending on the comparison in those decades around the decimal value 10, which contain a lower value than the corresponding subtrahend decade, and that the subtrahend is then deducted from the minute end in a binary subtraction unit (S) in which a pulse triggered by the comparator (V) is stored as a binary carry before a decade is calculated , which reduces the minute-end decade to be calculated by the value 1, if the previous lower minute-end decade has been increased by the decimal value 10, whereupon the difference found by the subtraction unit (S) is noted in binary representation within the decades as the result. 3. Apparatus according to claim 1 and 2, characterized in that the numbers to be calculated are shown in decimal-binary representation in the form of markings, the value of each decade is represented in binary and the marker combinations are introduced into an arithmetic unit a purely binary working adding unit (A), a comparison device (V) and a purely binary subtracting unit (S) work together as arithmetic groups in such a way that "when adding the adding unit (A) and when subtracting the subtracting unit - (S) the: actual - Carry out offsetting while the comparison device (V) together with the subtraction unit (S) for addition and together with the adder (A) for subtraction for the tens carry and for the return of the sum · ■ a decade to a value below 10 or the Difference of a decade to a ■ value above 0 serve. so · that while maintaining the xlekadenweise Offsetting the result of the values introduced in decimal binary form back into decimal-binary form is obtained. .- 4.- Device according to claim 3, characterized in that that synchronously with the passage of the marking rows through the individual -. Arithmetic groups (A, V and S) one pulse generator (C) the number corresponding to the number 10 when running through each decade Provides pulse combination. ■ 5.- device .according to claim 3 ,. .. characterized in that with '- the passage of the marking rows through the individual arithmetic groups (A, V and S) an auxiliary memory (D) circulates, to which an auxiliary marking is transferred for each decade, which reproduces the comparison result,' which occurs when the corresponding Decade is sensed again by the next arithmetic group and controls the effectiveness of the latter. 6. Apparatus according to claim 3 to 5, characterized in that the time interval between the last pulse position of a decade and the first pulse position of the following decade is at least twice as large as between the individual pulse positions. 7. Device according to one of claims 3 to 6, characterized in that the storage the pulses are carried out on rotating drums (73), preferably by means of magnetic markings. 8. Device according to claims 3 to 7, characterized in that a / "device (Fig. 18) is provided for transferring the values from a marking carrier to storage drums which, depending on the course of the calculation work, are carried out by means of a coupling (Fig. 7 ) can be coupled to the continuously rotating storage drums (B and D) in such a way that the coupling is only effective in a certain position of the storage drum. 9. ^ Apparatus according to claim 8, characterized in that the transmission from a marker carrier, whose rows of positions are sequentially and whose individual positions within each row are also sequentially sensed, in as many revolutions of the storage drum (I?), Are provided as rows of positions The positions of all columns are transmitted one after the other in each revolution and the entries of the individual positions of all columns are offset to one another in such a way that all positions of a column appear one after the other as a decade in the row of markings on the storage drum (B) . 10 '. Device according to claim 3 to 9, with a device which all columns of a mar- marking carrier at the same time, but the individual positions of a column marked one after the other, characterized in that the transfer from the results memory (B) to the marking mechanism (L) takes place in several revolutions of the storage drum (E) , and only the same positions of all in one revolution Transferring decades for marking in all columns of the marking carrier and then transferring the following positions of all decades in a further rotation, only taking the respective position of the individual decades. 11. The device according to claim 9 and 10, characterized in that the same device (Fig. 18) is used to transfer the information from the marker carrier to the memory (B) and vice versa from the memory (E) to the marker carrier. 12. The device according to claim 3 to 5, characterized in that in the adder (A) three inputs (TrI, Tr 2 and Tr 3) are provided, of which the first input its pulses from the results memory (E), the second its pulses from the value memory (B) and the third receives its impulses from a thyratron for the two-way transfer, and that at the outputs of the network at one output only one impulse appears if one impulse arrives at two inputs, and that at another Output a pulse appears when a pulse arrives at two or all three inputs, of which the pulses of the first output (A2) to control the entry in the results memory (E) and the pulses of the second output to ignite the two-way thyratron ( 12) serve. 13. The device according to claim 3 to 5, characterized in that in the subtraction unit (S) three inputs (Tr 4, Tr 5 and Tr 6) are arranged such that the first input (TV 4) to the result memory (E), the second input (Tr 5) is connected to the pulse generator (C) or the value memory (B) and the third input (Tr 6) to the thyratron (15) for the two-way transfer, and that at the outputs of the network on the an output (6 * 2) a pulse appears if a pulse is received in only one of the three inputs or in all three inputs, and a pulse appears at the other output if the second (TrS) or third input (Tr 6 ) alone one impulse or in the second and third together one impulse or at all three inputs one impulse each, of which the impulses of the first output (6 * 2) to control the entry in the results memory and the impulses of the second output for Ignition of the thyratron (15) is used for the two-way transfer. Considered publications: Book by C. W. Tompkins, J. H. Wakelin and W. W. Stifler, "High-Speed Computing Devices", McGraw Hill Book Comp. Inc., New York - Toronto * - London, 1950, pp. 410, 411; German Patent No. 930714; British Patent Nos. 687 427, 738 294; U.S. Patent Nos. 2,624,507, 2,668,661, 705,108. In addition 3 sheets of drawings ■ © 909 508/211 4.59