DE1290961B - Process for converting analog values into digital values - Google Patents

Description

The invention relates to a method for converting analog values into digital values using a converter system 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 m different ones, each corresponding to a specific digit in this number system Main codes each consisting of η bits, from which the digital output signals corresponding to an analog input signal are obtained by means of logic circuits after the tracks on the coding plates have been scanned via a binary intermediate code.

In such a method, the invention consists in that when one of the signals in the binary intermediate code coincides with a digit corresponding to a main code, this signal is emitted as an output signal, while if the signal in the binary intermediate code is also different from m , from each of the Main codes differing and the bit-wise ORing of two adjacent main codes representing subcodes of η bits each belongs, one of the two adjacent main codes is selected for the delivery of the output signal using a carry signal of the next lower digit, and that the structure of the main codes is made in this way that every bit-wise ORing of a main code with a defective code resulting from a change in part of the bits from an adjacent main code leads to the main code itself or to the sub-code lying between the neighboring main codes.

In an expedient embodiment of the invention, the function of conventional V-brushes used to derive the main and sub-codes, creating a logical code system that essentially represents an extension of the known Gray code system.

Details of the invention will appear from the following description. Show in the drawing

F i g. 1 and 2 each a diagram of a known coding plate with two different brush arrangements,

F i g. 3 and 4 each a diagram of two known analog-digital converters,

F i g. 5 and 6 diagrams to explain the mode of operation of the coding plate mentioned, F i g. 7 schematic representations of two coding plates according to the invention,

F i g. 8 shows the principle diagram of a logical discrimination circuit for use in connection with the coding plate according to FIG. 7th

Various devices have already been proposed for converting analog signals, which represent linear displacements, angles of rotation or similar movement quantities of movable bodies, into corresponding digital signals. A known device of this type includes a combination of a coding plate and brushes. Among these devices, there is one in which information of a few bits is provided in response to analog signals so that the value of rectilinear displacement or rotation angle of the movable body at a certain time during the continuous movement can be converted into a particular code .
F i g. 1 shows an exemplary embodiment of a known coding plate for converting such analog signals into a 4-bit binary code. The coding plate 1 according to FIG. 1 contains a large number of conductive segments, which are represented by dashed areas and correspond to the "1" of the binary code. In addition, the coding plate 1 contains a large number of electrically isolated segments which are identified by non-dashed areas and which correspond to the “0” of the binary code. In the individual sectors 2, the numbers 0 to 15, which correspond to the angles of rotation of the coding plate 1, are represented by a 4-bit binary code. The coding plate 1 is assigned brushes 3 which slide on the four tracks of the coding plate 1.

One difficulty that arises with this combination of the coding plate and the brushes lies in the significant errors due to a slight displacement of the brushes. To avoid these mistakes, the so-called "V-brush method" was proposed, in which - as shown in FIG. 2 illustrates - each bit two brushes 3, 3 'are assigned to the coding in the usual way after a ordinary binary system coded coding plate. This V-brush method becomes one of the two output signals of the brushes of a certain number selected depending on whether the brushes deliver the output signal »1« or »0« at a less important point. In this way it is possible to eliminate the influence of a displacement of the brushes.

In this V-brush method, however, a signal 4 from a lower adjacent location in the area indicated by FIG. 3 in order to select the one of the two signals supplied by the two sliding contacts of a V-brush when a signal at a certain point transitions from one code to the next. In this case, eight transmission wires 6 must be provided per coding point in order to transmit the signal to a receiver if a selection circuit 5 is provided in the receiver for selecting one of the two signals. If you want to have only four transmission wires instead of eight, the selection circuit 5 must be arranged in the converter. In order to avoid the use of V-brushes for signal generation, a system is known which uses the so-called Gray code system. The coding is provided in such a way that only one bit is changed between adjacent codings. Since in this system the output signal is not in the form of a conventional binary system, logical discrimination must be provided in the receiver in order to obtain the desired code. However, if more than two coding plates are connected to one another via gears, then V-brushes must be used for the coding plates of the digits with the exception of the lowest digits even when using a Gray code system. The disadvantages of the known systems are completely avoided by the device according to the invention for converting an analog signal into a digital signal. An essential feature of the invention is the use of a coding plate for each memory location, these coding plates displaying m different main codes with η bits and also m sub-codes with η bits, which represent the bit-by-bit "or" of two main codes, the neighboring numbers of the m -coded notation

and are different from each of the main codes mentioned. According to the invention, one distinguishes Device which causes a discrimination of a coded signal that depends on the coding plate the mechanical variable to be converted into a digital signal is used to determine whether this Signal belongs to the main code or sub-code, so that the signal is accepted as a true output signal becomes when it belongs to the main code, while when this signal belongs to the sub code frozen, one of the neighboring numbers of the »! -coded notation as a true output signal depending on is chosen from the true output of the lower adjacent digit.

In the following, the function of the above-mentioned V-bristles is described from a different point of view analyzed from. One can say that a V-bristle responds to a discriminatory signal, that is supplied by the lower neighboring point to one of the two neighboring codes of the considered Position. It goes without saying that such a signal can also be of the higher neighboring Location from the location under consideration can be supplied if desired. This function of the V-brush is in fact the rounding up or down when decimaliating similar.

With reference to FIG. 4 is the structural analysis of an analog-digital converter using a decimal notation. The number in the next higher position is increased or decreased by one unit if the number in the considered position changes from 9 to 0 or from 0 to 9. If, on the other hand, a digit in a position changes from 4 to 5 or from 5 to 4, there is of course no change in the number of the next higher position; a brush 3 '' remains in the existing stiffening, in which it is a segment in contact mil, which is a certain number. Each time \\ hen a number from 9 to 0 or from 0 to 9 changes in any stiffness , this affects the next higher position; this moment is referred to in the following as the transfer time.

In each digit there are 0 and 9 adjacent digits; Furthermore, 0 represents the higher digit compared to 9. As mentioned, in the case of a decimal numbering in one place there is no change in the digit present there if the digit 4 changes to 5 or 5 to 4 in the adjacent lower position. This means that a V-brush in a certain position can select one of the two adjacent digits of the part (namely the larger or the smaller), depending on the position of the V-brush in the next lower SteUe, ie depending on whether this V-brush is in contact with 0, 1, 2, 3 or 4 or in contact with 5,6,7,8 or 9. In other words, the next higher digit can be selected if the V-brush is in the lower digit in the area with the digits 0, 1, 2, 3 and 4 and if signals from two digits are supplied in the higher digit while the lower digit is selected in the next higher digit if the V-brush is in the lower digit in the area with the digits 5, 6, 7, 8 and 9 and if signals of two digits are supplied in the higher digit . As mentioned earlier, there is no way that signals of two digits will be delivered when in the. next lower digit changes a digit from 4 to 5 or from 5 to 4; The toggle function for selecting one of two digits is only required if a digit changes from 0 to 9 or from 9 to 0 in the next lower digit. It is obvious that the nature of the discrimination is exactly the opposite of whether it is rounded up or down. On the basis of the above theory in which the function of the V-brush is analyzed from a new point of view, a general explanation will be given for the case where n-bit signals are used to produce codes of ro-coded notation. In this case, m and η are positive integers not less than 2, where m and η have the relationship 2 "2: m. It is now assumed that an« -bit code has the form 00 ... 100 and corresponds to a number / (0 5Ξ / < m) of the m-coded notation Further assume that a code α, for this signal is expressed with

α, - = 00 ... 100.

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

ο. ψ aj (i φ ./, 0 ^ i <1 », 0 £ j < m). (1)

These m codes are called main codes in the following and are arranged on a main coding disc.

In this connection it should be pointed out that even if the codes α, and a _{i + l} are shaped as precisely as possible, a brush cannot cross the border between the two codes within an instant with regard to its width. The mentioned index / + 1 cannot be larger than m ; in such a case, m is subtracted so that the result falls between 0 and m- 1.

On the basis of the above-mentioned interpretation of the function of the V-brush, the mode of action set up the following principle for the V-brush. If there is a signal at one point in the transmission period, signals that correspond to two neighboring digits can be sent from the next higher position, are supplied, a discrimination circuit distinguishing between these signals; if, on the other hand, the next lower position is very far away from the transfer time, then the considered can be Place a signal corresponding to a digit.

According to this principle, it is known that a further signal between the signals of the main codes α _; and « _{i + 1} must be present. Let this code be denoted _{as α ί +} ' _2. It should be recalled here that a code that is located between the codes' _i _m and a _0, ö _ra _, +> _2. F i g. 5 shows, on an enlarged scale, the transition between two codes on a coding disk. In this context, it is useful to consider the code a _{i +} i ₂ as the bit-wise "or" of the two codes α and a _{i + l} . Assuming that the symbol © denotes the bit-wise "or", the code _{i +} > ₂ can be expressed as follows:

öj 0 α _{/ +} ι = α _{ί +} . ₂ . (2)

In 5-bit codes, these codes can be expressed as follows:

a, 1 1 0 0 0 a _{i + i} = 0 1 1 1 0 "I ₊ V ₂ = 1 1 1 1 0

Since the code a _{i +} ' _{2 is} also a «-bit or 5-bit code, the number of these codes must be equal to m in order to distinguish all these codes from one another by means of logical discrimination. These codes, referred to below as sub-codes, must be different from each other and also from each of the main codes, ie

' ₂ Φ

<* i Φ

(3)

IO adjacent lower digit is either 0, 1 ... or - ^ - - or one of. '" ⁺ * ... m - 1. In other words:

The true output is f if the output of the encoder disk is in the form of the main code a _; is given; the true output signal is / + 1 if the output signal of the coding disk is given in the form of the subcode a _{i +} ' _z and the true output signal of the adjacent

where / and j are any integer.

However, this is the condition when a brush is just in the middle of the transition from one main code to the next. It is now assumed that a brush is on the main code a, - for /, which has at least two "O" bits,., And is approaching the main code a _{i + 1} for / + 1. and further that the main code a _{i + i} for / + 1 has two or more "1" bits, where the main code contains α, - for / "O" bits. Then there is the possibility that only one of the "I" bits will touch the brush during the transition. Therefore, a code appearing at this point in time is given by the expression α, · - ä _{i + 1} . In this expression, the part ä _{i + 1} denotes a code which is derived from the main code α, ₊₁ by changing some "1" bits to "O" bits and can be described as the defective code of the main code a _{i + 1} .

Such a code α, ■ ä _{i + l} can, as mentioned above, occur at the moment of the transition of a brush from a main code a _{ to a main code a _{i + 1} , but this does not matter if the main codes of the coding plate are constructed in such a way that that the resulting code a _; ~ 5 _{i + 1 is the} same as the main code a _t itself or is the same as the sub-code a _{i +} ' _{2 lying between the main codes α, and a I + 1} , since both the main code itself and this sub-code are in any case the above under (1) to (3) are met. Correspondingly, for the code a, ä _{i + l} the condition barten lower digit 0,1 ... or

m- \

is;

ö, ₊₁ = ti, - or ei, +.,

(4)

or in a similar manner for the case of the transition of a brush from the main code a _{i + 1} to the main code α, - and the code ä, a _{i + 1} occurring in this case, the condition

45

= α, + ι, or a _{i + 1}

(5)

are valid.

Since it follows from inequality (3) that 2 »! Codes of η bits are used, the previous relationship 2" ^ m is now corrected to 2 "^ 2m.

From the above description it follows that the signal of a digit of the »»! -Coded notation is either a signal of the main codes or one of the subcodes. If one of the main codes comes from the position, the true output signal of the position is a digit corresponding to the main code. On the other hand, if one of the sub-codes comes from the position, the true output signal of the position is one of the two digits corresponding to the two main codes on opposite sides of the sub-code. In this case, the true output signal is determined according to whether the true output signal is the true output signal z, if the output signal of the coding disk is given in the form of the subcode «, · +>, and the true output signal of the adjacent lower digit ^m + ¹ . .. or m - 1 is.

Since there is no further digit below the lowest digit, there is also no carry signal in this respect. In view of this, one can assume a signal from an underlying imaginary digit; For this particular location, one of the following two conditions can therefore be selected instead of the above condition (6): The true output signal is / + 1 if the output signal from the combination of the encoder disk and the brushes a _; or a _{i +} i _z .

or deviating therefrom, the true output signal is /, if

the output from the combination of the encoder disc and the brushes is α, or a _{i +} > ₂ .

The true output signal can therefore 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), of the most important of which are the following:

a) The main codes invariably have the same number of "!" bits, while the number of "!" Bits in the subcodes is one greater than the number of "1" bits in main codes.

b) If there are two adjacent main codes. a _; and a _{; +1} and an intermediate sub-code a _{i +} > ₂ only differ in their first two bits, then these codes have the form

55

_a . ₌

*** ... * denotes the part in which the three codes are identical. From this it follows that a group of such main and sub-codes forms a so-called "Gray code". However, the Gray condition is not met in the system according to the invention, since it is according to the invention is all about the main codes. An even more significant difference between the one according to the invention Code and the Gray code system is that, according to the invention, the width of the brush 3 at

transition from one code to another poses no problem.

In the Gray code system, the digits are actually changed when the rear end of the brush 3 touches or leaves the boundary between adjacent codes when the brush 3 is in one direction wanders in which a "!" bit changes (see Fig. 6a). This change occurs when the leading edge of the Brush 3 touches or leaves the boundary when the brush moves in a direction in which a "!" Bit is added (see FIG. 6b). Do the individual codes on the coding disc have the same width? or the same arc angles in the case of a rotating coding disk, then there is between the case of one Increase by a "!" Bit and in the case of a decrease by a "!" Bit, a phase difference that corresponds to twice the brush width; it is therefore necessary to place the codes on the coding plate when producing the codes the brush width take into account what is happening in view of the wear and tear on the brush as well is disadvantageous with regard to a brush replacement. The Gray code system is therefore with regard to the brush width no sensible solution is possible.

In the system according to the invention, one of the aforementioned conditions (6 ') and (6 ") can be selected in this way that the digit change actually always takes place when a brush edge crosses the code limit touches or leaves, so any changes in brush width will not affect the Perform digit change process. At the border between two adjacent main codes can also a code-free area can be provided, the width of which is smaller than the width of the brush. Also can a sub-code can be provided at the boundary between two main codes.

According to the invention, analog-digital conversion is thus carried out by using codes which have the relationship 2 ″ Ξ> 2m and the conditions of equations (1), (2), (3), (4) and (5) the code conversion is effected by suitably selecting the conditions of equations (6), (6 ') and (6 ").

A preferred embodiment of the coding according to the invention will be described in more detail below explained. A number of 5-bit codes can be expressed as follows:

O ₀ = 00011
ö, = 00110
α, = 01010

«3 = 11,000
«4 = 01100
α ₅ = 01001

7a and 7b show a rotating coding disk and a linearly movable coding plate, marked with these codes.

A logical discrimination circuit suitable in connection with these coding plates is shown in FIG. 8 illustrates. It contains OR circles OR, AND circles AND and NOT circles NOT S ₁ and S ₂ denote the carry signal from the next lower position and the carry signal to the next higher position. The symbols O ₀ to a ₅ designate the output signals as a function of the input signals to f.

Although in this embodiment fl, and a, + ', correspond to a single code, so can but the arrangement can also be made in this way.

that different codes correspond to the same digit i , thus adding the error correction function. This error correction function can be achieved in detail as follows. It is assumed that instead of the code α, - a group of codes A ₁ is used and that the codes belonging to this code group / 4, - are expressed by a _x , k , where 0 ^ k <k; and k _{x is} the total number contained in the code group A ₁

, ο means code variations. Then you get the following Equations:

«Ι., Φ «;., ('^ J. 0 ύ P <h 0 ^ q <kj) + ", -." (0 g P <h 0 g q <k _b Jf φ q) s äj + i, _g e (Ai -r

In the above relationships, e means that the part on the left is included in the part on the right. The conditions given by these expressions now replace the conditions in equations and inequalities (2), (3) and (4) explained earlier. Based on these new conditions, 6-bit ternary codes can be expressed as follows, for example:

A ₀ :

A ₂ :

000001,

001111,
001011,
000111,

001000,

111100,
101100,
011100,

100,000

110011,
110010,
110001,

000010,

001101,
001001,
000101,

000100,

110100,
100100,
010100,

010000,

010011,
010010,
010001,

000011

001110,
001010,
000110

001100

111000,
101000,
011000

110000

100011,
100010,
100001

A coding plate can expediently be equipped with codes that have the largest number of "1" bits included under the above codes. The advantage of this solution is that even then there is no mistake results when a "!" bit is missing in the code. A disorder that is common when combining one Encoding plate and brushing occurs because a "!" Bit is faulty due to poor contact or is changed to an "O" bit by some other cause. However, such a mistake will corrected automatically using the codes of the system explained. These codes can thus be provided with an error correction function.

In the next digit there is a group of main codes and sub-codes of a special kind for the lowest digit. It should be remembered that one can treat α, + ·., In the lowest digit in the same way as 0, or a _{i + 1.} In principle, the code group / J _{1 +} ^ can therefore be used in the same way

treat like A _x or / Ij _{+ 1} .

The above description was based on an analog-to-digital conversion using a combination of coding plates and brushes. It understands

909512/1421

however, that the invention is also suitable for other converters, for example those based on electromagnetic basis or work with photoelectric elements.

The use of subcodes results even when using four-part coding devices a large allowable tolerance in code production; the position of the detector elements (for example Brushing) to pick up the transformed signal can be easily adjusted; at a facility with a large number of coding plates connected to one another, for example by gears, is a considerable one Gear backlash permitted. Furthermore, there is no logical discrimination circuit for V-brushes in the converter necessary. The smallest possible number of lines is still sufficient to transmit the formed Signal. Finally, a simple logical discrimination circuit can do any desired code transformation carry out.

Claims (1)

Claim: 20th Method for converting analog values into digital values using a converter system 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 m different, respectively a specific digit in this number system carries corresponding main codes of η bits each, from which, after scanning tracks on the coding plates via a binary intermediate code, the digital output signals corresponding to an analog input signal are obtained by means of logic circuits, characterized in that if one of the signals in the binary intermediate code with a digit corresponding to a main code, this signal is emitted as an output signal, while if the signal in the binary intermediate code leads to one of likewise m different subcodes that differ from each of the main codes and represent the bitwise ORing of two adjacent main codes each η bits belongs, one of the two adjacent main codes is selected for the delivery of the output signal using a carry signal of the next lower digit, and that the structure of the main codes is made such that each bit-wise OR operation of a main code with a defective code resulting from a change in part of the bits of an adjacent main code leads to the main code itself or to the sub-code lying between the adjacent main codes. 1 sheet of drawings