DE1299697B - Device for converting analog values into digital values - Google Patents

Description

in the binary intermediate code with a main line movement, rotary movement or other code corresponding digit to represent this signal as outward movements of a movable body, in

Digital signals were already

output signal is emitted, while if the signal in the binary intermediate code belongs to one of the likewise m different subcodes of η bits that differ from each of the main codes and represent the bitwise OR operation of two adjacent main codes, for the delivery of the output signal using Various devices have already been proposed for a carry signal of the next lower point of one of the two corresponding digital signals. A typical example of such known designs makes use of the combination of a coding plate with brushes. With such an arrangement, information of some bits is output in response to analog signals so that the value of the rectilinear movement or the rotational movement of the

adjacent main codes is selected, and that 30 movable bodies at a given time

the structure of the main codes is made so that each bit-wise OR operation of a main code with a man resulting from changing part of the bits from an adjacent main code the continuous movement can be transformed into a special code.

F i g. 1 shows an embodiment of a coding plate for converting such analog signals into

similar code to the main code itself or to the 35 a 4-bit binary code. The coding plate 1 according to

sub-code lying between the adjacent main codes.

With the method according to the main patent application it is possible to obtain an unambiguous and very precise F i g. 1 includes a number of conductive segments illustrated by hatched areas and correspond to the "l" of the binary code. The coding plate also contains a number of electrically iso

Conversion of analog values into digital values to 40 lated segments, which are indicated by non-dashed areas

parts are marked and correspond to the "0" of the binary code. In the individual sectors, the Numbers 0 to 15, which correspond to the angles of rotation of the coding plate 1, are coded by a 4-bit binary

achieve that even with the transition from one Main code to another main code on a coding plate by the scanning elements regardless of whose finite expansion does not lead to errors

can. The aim of the present invention is now to show a 45 decimal notation. There are still Bür devices indicate, with the help of this most 3 provided, each with a track of the coding process for the extraction of digital signals in plate 1 are in sliding contact. With this well-known

Can make use of the decimal system. Execution is a solid relationship between

According to the invention, this object is achieved by one of a track and a bit; is the number of tracks Structure of the device for converting analog 50 equals the number of bits. Trying to get the number values are resolved in digital values in such a way that each of the bits in a coding plate has a specific outer decimal place Increasing the diameter of the digital output signal therefore inevitably leads Coding plate is provided and that the main codes lead to a reduction in the width of the individual tracks, ten different, representing decimal digits, which in turn either increases the accuracy 2-out-of-5 codes and the associated sub-codes 3-out-of-55 when attaching the brushes or a narrower one 5 codes are. Training of the brushes is conditional. Trying to get wider

The details of the invention are now intended to preserve traces, on the other hand inevitably leads to to Fland the description of some designs with an enlargement of the external dimensions of the Ana

games are illustrated, reference is made to the drawings. In these show

F i g. 1 to 3 schematic representations for known converters and coding plates and associated Scanning devices,

F i g. 4 is a circuit diagram for a discrimination logic circuit for use in connection with a coding plate designed according to the invention,

F i g. 5 is a schematic perspective view of a known analog-digital converter,

log-digital signal converter. This magnification is Above all, it is undesirable because this also makes it effective when the coding plate is rotated Moment of inertia is increased. Even with coding disks with straight tracks, as in FIG. 2 illustrated is a large width of the tracks because of the resulting large width of the Brush holder undesirable.

In the case of code converters of the type described, the so-called »V-brush-

Method «and the Gray code used to prevent the generation of error signals when the To prevent brushes 3 from one contact position to the next. With this V-brush method, there are two Brushes 4 and 4 'are provided, which are shown in FIG. 3 obvious ways to interact with each bit. In the case of the Gray code, code patterns of a special shape are used in order to avoid uncertainties can occur at the transition between the contact positions; the coded signals are subsequently converted into the desired signals. About the complicated structure and the cumbersome Code pattern arrangement of the known analog-digital converters to avoid is the design according to the invention with a number of coding plates, with a coding plate for each decimal place, provided with the ten different code patterns of 2-out-of-5 codes that represent the decimal digits, leaving traces of the same, however appropriately offset in the circumferential direction patterns are formed and the bit-wise "or" of two each adjacent 2-out-of-5 codes form a series of 3-out-of-5 codes. The device according to the invention acts in such a way that when an encoded digital signal which is determined by a digit, the not the lowest digit corresponding to a mechanical quantity to be converted into digital form is one of the 2-out-of-5 codes, this coded signal is supplied unchanged as a true digital output signal. If, on the other hand, this coded signal is one of the 3-out-of-5 codes, one of the 2-out-of-5 codes changes to the opposite Pages of the 3-out-of-5 code delivered as a true digital output signal. What the most meaningless As far as digit is concerned, if an encoded digital signal is either a 2-out-of-5 code or a 3-out-of-5-code is' one of the 2-out-of-5 codes on opposite sides of it as true digital Output signal emitted.

To explain the principle according to the invention with reference to the drawing, the function of the V-brush analyzed from a novel point of view. You can say that a V-brush is on a Responses discriminating signal supplied by the neighboring lower digit to one of the two adjacent codes in a given point. It goes without saying that such a If desired, signal also from the point under consideration to the neighboring higher point can be delivered. This function of the V-brush is in fact very similar to rounding up and down in the numerical calculation in decimal notation. Using the decimal notation as an example be the structure of one in FIG. 5 illustrated analog-to-digital converter. The one in the The digit in the next higher position increases or decreases if in the relevant position a change from 9 to 0 or from 0 to 9 occurs. If, on the other hand, there is a digit changes from 4 to 5, no change occurs in the next higher position. A brush 3 then remains in the existing position in which it is in contact with a segment representing a specific digit. Whenever a digit changes from 9 to 0 or from 0 to 9 in a certain position, so has this has an influence on the next higher position. This point in time is referred to as the transfer point in time.

It should also be noted that 0 and 9 are two adjacent digits in the same place and that 0 - as far as these two digits are concerned - is greater than 9. In the case of decimal notation, there is also no transition between digits in the same place if a digit in the next lower digit changes from 4 to 5 or from 5 to 4. A V-brush that is assigned to a specific position can then select one of the two neighboring digits in this position in a suitable manner, namely the larger or smaller, depending on whether the V-brush in the next lower position is in contact with 0 , i, 2, 3 and 4 or in contact with "5, 6, 7, 8 or 9. In other words, the larger digit in the next higher digit is selected if the V-brush is in the lower digit in a digit range, which contains the digits 0, 1, 2, 3 and 4 and if signals from two digits in the higher digit are supplied; the smaller digit in the next higher digit is selected, however, if the V-brush in the lower digit is in a digit range which contains the digits 5, 6, 7, 8 and 9 and when signals of two digits are supplied in the higher digit As mentioned earlier, there is no possibility that signals of two digits are supplied when there is a digit in the neighboring th lower digit changes from 4 to 5 or from 5 to 4. Switching over the function of selecting one of two digits is only required if a digit in the next lower digit changes from 0 to 9 or from 9 to 0. The result is that the type of distinction explained is exactly the opposite of rounding up and down.

Based on this theory that the function of the V-brush is analyzed from a novel point of view, a general explanation will now be given of a case where n-bit signals are used to produce codes of m-coded notation. In this case both Ot and η are positive integers (not less than 2). There is of course the relationship 2 "> m between them. It is now assumed that an n-bit code has the form 00 .. .100 and corresponds to a digit i (0 < i < m) of the m-coded notation. Es let us then further assume that a code pattern a " which emits the above signal is expressed as follows:

α, - = 00 ... 100.

Since Ot codes are different from each other, the following relationship must exist:

α ,, φ Uj (i φ j, 0 <i < / η, 0 < j <m) . (1)

These m codes are hereinafter referred to as main codes and are provided as main code patterns on a coding plate.

In this context, it should be emphasized that a brush cannot cross the border between the codes α and a _{i ¥ 1} within an instant with regard to its width, however precisely these code patterns may be shaped. In the above description, the index i + 1 cannot become greater than m ; in such a case, m is subtracted therefrom so that the balance is between 0 and m-1 slit.

On the basis of this interpretation of the function of the V-brush, the following principle can be applied to the Determine the mode of operation of the V-brush. If a certain point is at the time of transmission, can send signals from the next higher point

two digits are supplied accordingly; a discrimination circle can distinguish these signals If, on the other hand, the next lower position is the furthest away from the transfer time, then A signal corresponding to a digit can be supplied by the point under consideration.

According to this principle, another signal must exist between the signals representing the main codes a- and a _{i + i.} Let this code be denoted by a _{i + ll2.} In this connection it should be remembered that a code α _π , _! Lying between the codes a _m _i and a " _+1/2 is. F i g. 6 shows, on an enlarged scale, the boundary between two code patterns on a coding disk. As can be seen from this, the code ", + ι 2" is the bit-wise "or" of the two codes a-, and fl _{f + l} . If the symbol © denotes the bit-wise »or«, the code « _{i +} , ₂ can be expressed as follows:

(2)

For example, in the case of 5-bit codes, these codes can be as follows:

«I =

1 1 0 0 0 0 1 1 1 0 1 1 1 1 0

flf + 1-2 =

Since the code a ^ _{ll2 is also} a / i-bit code and since accordingly all these codes can be distinguished from one another by logical distinction, the number of these codes must also be equal to m . These codes, referred to below as sub-codes, must be different from each other and from each of the main codes, ie

CI _{1 + 1} 2 * = «, + 12 ('- />

Since it is known from inequality (3) that 2m muslers of »bits are used, the previous relationship 2"> m is now corrected to 2 "> 2 m. From the above description it follows that the signal that comes from a specific digit of the m-coded notation, is either one of the main codes or one of the sub-codes. If one of the main codes comes from this position, the true output signal of this position is a digit that corresponds to the main code. If, on the other hand, one of the sub-codes comes from this position, the true output signal from this position is one of the two digits which correspond to the two main codes on opposite sides of the sub-code. In the latter case, the true output signal is determined according to whether the true output signal is the next lower digit

Range O, 1 ... - y- or to range - ^ - ... “I heard one.

Expressed differently:

The true output signal is / if the output signal of the coding plate is provided in the form of the main code a _t :

the true output signal is i +1 if the output signal of the coding plate is supplied in the form of the subcode « _{l + 1 2} and the true output signal of the next lower digit is 0.1 ...

ff. J

or τ is;

the true output signal is / if the output signal of the coding plate is supplied in the form of the subcode a _{i + l 2} and the true output signal of the next lower digit is a +1

from

.m-l is.

Since under the most meaningless digit there is no further «, = t Oj ₊₁₂ (/./: any whole number). (3) 40 digit is available, there is no over-

However, this is the condition when a brush is just in the middle of the transition from one code to the next. It is now assumed that a brush is on the code «, the number / and approaches the code _{i + 1} for / + 1. It is also assumed that the code <i _{I + 1} for i + 1 has two or more "1" for bits in which the code α, for ί has "0". There is then the possibility that a "1" touches the brush too early during the transition. If a code occurs at this point in time, a "1" is changed from a _{i + 1} to "0". This defective code of a _{i +} is denoted as 5 _{i +} i. Every code that occurs at this moment is therefore contained in α, © α _{ί + 1} . It should be remembered that this code appears at the moment when the information door ί or the information for the transition from ■ / ■ to i +1 should appear at the exit. The permitted output code is therefore either ti, - or «, _{+ 1 2} . So it applies

carry signal. In view of this, one can assume a signal from an imaginary location below. For this particular position, one of the following two conditions can be used instead of the above condition (6) to get voted:

The true output signal is / +1. if the output signal obtained from the combination of the coding plate and the brushes is «, - or« _{i + 1 2} .

or

(6'j

the true output signal is 1 if the output signal obtained from the combination of the coding plate and the brushes is α, - or α, _ι ₂ .

= a, or «, + in-

(4)

If the brush is on the code pattern α, _{+ (} for / +1 and if it approaches the code Ci ₁ for /. The following condition must be fulfilled in the same way:

^N ä, -® « _{i +} i =« _{1 + 12} or <i,.,. (5)

The true output signal can thus be easily determined using an electronic circuit which is the logical distinction between the conditions given in equations (6), (6 ') and (6 ") meets.

Various facts can be derived from the equations and inequalities (1) to (5). the The most important of these are the following:

a) The main codes invariably have the same number of "!" bits, while the number of "1" bits in the sub-codes by one greater than the number of "!" bits in the main codes isf.

b) If two adjacent main codes α, - or a _{i + l} and an intermediate sub-code a _{i + ll2} differ from one another only in their first two bits, then these codes have the form:

a _{t + l} = OV

IO

Here *** ... * mean the part in which these three codes have the same form. It surrenders from this that a number of these main codes and sub-codes form so-called Gray codes. The Gray condition is not fulfilled, however, since only the main codes are dealt with. Another big difference between the code according to the invention and the Gray code is that the width of the Brush in the transition from code to code according to the invention does not pose a problem. '

With the Gray code, a change in digits actually occurs when the rear edge of the brush 3 touches or leaves the boundary between code patterns when the brush 3 moves in one direction, in which there is a reduction by a "1" bit (cf. FIG. 7 a). The digit change takes place on the other hand instead of when the front edge of the brush 3 touches or leaves the boundary between the code patterns, if the brush 3 moves in a direction in which an increase by a "1" bit occurs (cf. Fig. 7b). When the code patterns are formed on the code plate so that they are for each code have the same width (or the same arc angle in the case of a rotating code plate), the Phase difference between the case of an increase by a "1" bit and the case of a decrease by a "1" bit twice the value of the brush width; one must therefore use the code pattern when producing it the coding plate take into account the brush width, which is both in terms of brush abrasion and is also disadvantageous in view of a possible brush replacement. The Gray code therefore does not represent any is a satisfactory solution in terms of brush width.

In the system according to the invention, one of the conditions (6 ') and (6 ") can suitably be so be chosen so that the actual digit change occurs whenever in a predetermined manner one of the brush edges touches or leaves the code boundary: have changes in the brush width no influence on the process of changing digits.

In the following, the conditions explained are used to search for codes of a decimal notation. which are preferably used according to the invention. From the relation 2 "> 2» i it is known that the The smallest number of bits that can be used with the current decimal codes is 5. You can do a combination from 5-bit decimal codes as shown in the following table.

60

a β V fr ^a k + 4l 1 0 0 0 1 0 0 1 <* k +61 0 0 1 1 O-k + ΊΙ 0 1 0 1 ^a k + 8l 1 1 0 0 ^a k + 9l 0 1 0 0

In the table above it is not indicated which code corresponds to the digits 0, 1, 2 etc. These codes rather have the following characteristics:

i) Interchangeability of the columns: The order of the columns can, as assumed in the table, be α βγδ f. Instead, however, other sequences are also possible, such as β Ö α e γ or γ t δ β a. ii) Interchangeability of the assignment: The number k used in the index can be any positive or negative integer or 0. In other words, the assignment of the digits 0 ... 9 to the codes can be started from any line.

iii) Reversibility of the assignment: / in the index can be either +1 or - /. In other words, the direction of association between the codes and the digits can be reversed.

iv) Rotation of the same pattern: All five columns have the same pattern 1110000000 and point successively have a phase difference of two bits.

Various forms of 2-out-of-5 codes are already known which are used for decimal codes and contain two "1" bits. However, mechanical V-brushes are needed in conjunction with a code plate if these codes are to be used to form code patterns in their original form. However, it has been proved by using a calculator that 5-bit decimal codes satisfying the formula 2 "> 2m and satisfying the conditions contained in equations and inequalities (1) to (5) are all included in the code group of the above table no mechanical V-brushes need to be provided for this. Although feature iv) has no relation to formula 2 "> 2m and the conditions in equations and inequalities (1) to (5), this feature can be used to create a to achieve a single-track coding plate that provides a 5-bit decimal code for each digit.

Various modifications of the code explained can be obtained by suitable determination of k and / in the index. However, since these modifications do not change the characteristics possessed by the above code, the association shown in the table below can be selected for the following description.

"k

"l · .1 ί

O O O 1 1 O O 1 O 1 O 1 1 O O 1 O 1 O O

fts A _x A ₂ A ₃

1 0 0 0 1 1 0 1 0 0 1 1 0 0 0 0 1 0 1 0 0 1 1 0 0 909 530/347

a β V δ 1 A ₅ O O 1 O O A, O O 1 1 O A ₁ 1 O O 1 1 A _n O O O 1 1 A _g O 1 O O

An essential feature of this assignment is that the signal γ β a ? δ (where γ is a “not” of ä ) supplies a sub-code that lies between the 10's and 9's complement of the original digit and can thus be used to calculate the complement. Of course, there are numerous other options for assignment apart from those assumed in the table above.

A coding plate, which has code patterns according to the table above, has the shape shown in FIG. 8, if it is produced in the usual way, one track each being required for the columns α, β, γ, δ and e.

However, all five columns and accordingly all five tracks have - as shown in feature iv) became - the same pattern. If this feature iv) is used, brushes that separate from each other Appropriate spacing to be used in conjunction with a single track to the Coding plate structure according to FIG. 9 to get. Their advantage is that the width is the only one Track larger or the diameter of the entire coding plate can be kept smaller than before (with a rectilinear encoder disk can accordingly reduce the width of the disk). The production the code pattern and the spatial arrangement of the brushes also need (with the exception of the most insignificant digit) not to be done so precisely, because of any ambiguity that occurs when passing the brushes can occur from one code pattern to the next, completely eliminated by an electronic circuit can be.

To fix the digit determined, in the form of a signal from a certain point accordingly one of a coding plate of the structure explained transmitted signal is to be transmitted, it is necessary that the Umschaltsägnal in the next lower Steadily exists. This means that any transition between two digits in a specific place depends on the transition of digits in the place of least importance. The coding plate, which represents the lowest digit of a multi-digit number must therefore be designed so that Signals from this are switched at equal intervals.

It can be economically expedient if logic circuitry of the same type can be used for all digits of a number, including the lowest digit, to correctly distinguish between two digits. Since in this case there is no more digit below the lowest digit, condition (6) is modified to one of the following two conditions: The output signal is α, · if the input signal is α, - or α, · + 1/2 is; or the output signal is a _h if the input signal is a _t or ß _; _i / ₂ is. The above condition indicates that a digit change occurs whenever an edge of a brush touches or leaves the boundary between two codes. It is accordingly known that the brush alignment is of considerable importance for the accuracy of the digit change. The Gray code, on the other hand, lacks any specific relationship between a given brush edge and the digit change (as is the case with the code according to the invention), since with the Gray code the digit change occurs at one brush edge when the brush moves in one direction which is increased by a "1" bit, while in the reverse case (movement of the brush in a direction in which a decrease by a "1" bit occurs) the digit change occurs at the other edge of the brush.

As a practical embodiment of the above code, in the following with reference to FIGS. 10 to 12 an analog-digital converter for three-part decimal numbers explained. In this converter, instead of the previous combination of code plates and Brushes the combination of a light source, code plates, slotted masks and photocells (Solar cells) provided. The code plate has the structure shown in Fig. 10 and forms one turn at a time in digits from decimal notation. The coding plate has openings 2 ', which the conductive surface parts according to F i g. 9 correspond. One shown in FIG. 11 shown mask with Slots 4 'is arranged under the coding plate, so that the through the openings of the coding plate and Light falling through the slits of the mask hits devices (not shown) with photocells and there generates an encoded signal. An embodiment of such an analog-to-digital converter is shown in Fig. 12 shown. According to this, three such coding plates 1, Γ and 1 ″ are connected to one another via reduction gears 5, 5 'connected. Each coding plate and mask is provided with a photoelectric device assigned to five cells, whereby the desired signal conversion takes place.

F i g. FIG. 4 shows part of a logic circuit which is preferably used with the analog-digital converter according to FIG. In Fig. 4, symbols a \ b \ c \ d ' and e' denote signals from the menu digit, and a general symbol a ' denotes a «, output signal from the i-th digit and represents a signal which a Digit j corresponds to. The signals a ', h' etc. indicate the switch-on state (state »1«), which is present when light hits the photocells, and also the switch-off state (state »0«), which is present when there is no light on the photocells hits. These signals are fed to the circuit shown in FIG. 4 fed from the left. An output "1" is developed at one of the a \ on the right side of the circuit, which clarifies what the true output of each digit is. In the present embodiment, the logic circle of the least digit is illustrated with a simpler structure; however, the same circuit as for the second digit can be used for the number of the least significant number; a carry signal m ^l can be specified with “1” or “0”. The logic circle of the third digit is not illustrated because it corresponds to that of the second digit. The output signal of one of the photocells of the analog-digital converter according to FIG. 12 has the waveform Λ according to FIG. 13 and can be without limitation and / or amplification or by passing through a Schmitt trigger

Bring difficulty into waveform B. There is a 2% phase difference between the output signal according to FIG. 13B and the input signals of the five photo lines a, b, c, d and e. A practical embodiment will now be explained to show the practical application of this code. A practical advantage of using the code according to the invention is that the switch-on and switch-off periods (with the exception of the lowest digit) do not have to be as precise.

To facilitate understanding, it is assumed that one revolution is represented numerically by K) O ⁰ ',,. According to the invention, the output signal of a photocell, after amplification and / or limitation, does not have to be exactly analogous to the code pattern (1110010000), or the ratio of switch-on to switch-off period does not need to be exactly 30%: 20%: 10%: 40%, as in Fig 13C shown. There is also no absolute requirement that there must be a 20% phase difference between the signals from the five photocells a, b, c, d and e . In fact, this on and off ratio can be about 32 to 36%: 18 to 14%: 12 to 16%: 38 to 34%, as indicated in Figure 13D; the phase difference between the signals of the five photocells a, b, c, d and e can accordingly deviate in a range of 1 to 2% compared to the ideal phase difference of 20%. The only requirement in this context is that two or three of the five photocells of a point in their switched-on state ("1" state) and less than one or more than four cells not in their switched-on state ("!" State) in each Moment and that all five photocells are switched on before the arrival of the carry signal from the next lower point. With regard to this requirement, which is easy to meet, the production and assembly of this converter results in a considerable simplification even if the gear backlash, which leads to a less precise conversion from waveform A to waveform B according to FIG. 13, is also taken into account .

As far as the smallest point is concerned, the conditions must be stricter. The code pattern must accordingly be shaped somewhat more precisely. With regard to this particular point, there must be an exact ratio of 30% ^: 70% between the switch-on and switch-off periods if the logic circle of this point fulfills 10; the switch-off-switch-on period must have an exact ratio of 50%: 50% if the logic circle of this position 11 satisfies. In addition, there must be an exact 20% phase difference between the signals from the five photocells in this particular location. It is clear that these requirements are also relatively easy to meet from a mechanical and electrical point of view.

It is also possible to provide a track for a 10 digit and a track for a 100 digit on a single coding plate in order to effect an analog-digital conversion in these locations. In this case, the brushes can easily be assigned to such a coding plate, since the brushes can be set at a predetermined phase distance for each point, so that there is considerable freedom with regard to the size of the brush holder. 14 shows a coding plate for use with such an analog-digital converter, the coding plate generating a signal after every '/,' and ¹ Z _{100 of} a revolution. 15 shows a mask which is arranged on the light-sensitive cells opposite the coding plate according to FIG. 14 and which bears the illustrated slit pattern. The pattern on the coding plate can also be modified in such a way that a signal is generated _{after every V 20, V40 ° or any other division.}

A rotary switch is designed in such a way that it only rotates at predetermined particular angular positions interrupts. This rotary switch can be coupled with an analog-digital converter, so that coded signals are emitted depending on its angular position. An embodiment with a such a rotary switch is shown schematically in FIG. 16 illustrates. The brushes are stationary in shape Contact formed up to 9; a guiding pattern in Form of a contact strip 10 is provided with a number of radial projections which are in sliding contact come with contacts 5 to 9. The power supply to the sliding contact 10 takes place via a Connection 11.

The above description mainly related to analog-digital converters with a combination of coding plates and brushes. It is to be understood, however, that the invention is in no way based thereon is limited, but also, for example, for converters with electromagnetic or photoelectric Elements is suitable.

The use of sub-codes results even when using code patterns for four-part numbers a relatively large tolerance in code pattern production. The attachment of the elements to determine the transformed signals (for example from brushes) must be observed because of the low Tolerance convenient to carry out. In an arrangement with a plurality of coding plates, the are connected to each other via gears, a considerable gear backlash can also be allowed will. It is also advantageous that there is no logical discrimination circuit for V-brushes in the converter is needed. The smallest possible number of transmission lines is sufficient for the desired process for the transformed signal. A simple logical discrimination circuit can also do any desired code conversion carry out.

Claims (1)

Claim: Device for converting analog values into digital values with a plurality of coding plates connected to one another via reduction gears, each of which corresponds to a digit of the digital output signal in a specific number system and carries m different main codes each consisting of η bits, each corresponding to a specific digit in this number system bit-by-bit OR operation with a defective code resulting from a change in part of the bits from an adjacent main code leads either to the main code itself or to the sub-code lying between the adjacent main codes and generated by their bit-by-bit OR operation and from which, after track-by-track scanning, via a binary intermediate code the digital output signals corresponding to an analog input signal are obtained by means of logic circuits. Mood of one of the signals in the binary intermediate code with a digit corresponding to a main code, this signal is output as an output signal, while if the signal in the binary intermediate code belongs to one of the likewise m different sub-codes, which also differ from each of the main codes, each of ii bits, for the delivery of the output signal using a carry signal of the next lower position of one of the two adjacent main codes is selected for implementation of the method according to patent application P 1290961.4-31, characterized in that that a coding plate is provided for each decimal place of the digital output signal and that the main codes ten different 2-out-of-5 codes representing decimal digits and the associated ones Subcodes are 3 out of 5 codes. For this purpose 2 sheets of drawings