US 4,032,913 AGrant
Coding Equipment Providing Compressed Code
Issue Date:1977-06-28
•4 Claims
•10 Drawing Sheets
Abstract
In a parallel digital signal delivered from an AD converter, "1" appearing in a bit position nearest to the most significant digit (MSD) is detected, and the parallel digital signal is converted into a serial digital signal consisting of a weight character indicating this specific bit position and a data character indicating the numerical value of the data. Further, a weight bit is provided for changing the weight of the output of the AD converter depending on the analog quantity subjected to AD conversion so as to reduce the number of bits required for coding the parallel signal.
Metadata
Assignees
- Hitachi, Ltd.
- The Tokyo Electric Power Co., Inc.
Inventors
- Masami Okamura
- Fumio Aoki
- Yoshihiro Sano
- Jun-ichi Makino
- Yoshiteru Miki
- Hiroshi Sasaki
- Kazuo Seo
Application Information
Application Number:US 4593863
Filing Date:1974-04-09
Priority Date:1973-04-09
Art Unit:236
Classifications
IPC:
H03K 1302
Field of Search:
340179347 AD;347 DD15 AV
Patent Drawings (10 sheets)
Description
Background of the Invention
1. Field of the Invention
This invention relates to a coding equipment for converting a parallel digital signal into a serial digital signal.
2. Description of the Prior Art
It is commonly known that, in a centralized control of physical quantities such as voltage, current and temperature measured at remote points, these physical quantities are sampled with a suitable period at the individual measuring points, and after converting the respective sampled physical quantities into parallel digital signals by an analog-digital converter, these parallel digital signals are converted into a succession of serial digital signals to be transmitted to a central control station. In this case, in order that the information processing, storing and other operations in the central control station can be efficiently carried out, it is required that the information transmitted one after another in the form of the serial digital signal have the same number of bits regardless of the magnitude of the absolute value of the physical quantities sampled periodically. On the other hand, however, the fact that the information transmitted one after another in the form of the serial digital signal have the same number of bits means that the greatest value of each information is also limited to a predetermined number of bits even when great variations occur in the physical quantities. Thus, the magnitude of the physical quantity corresponding to the least significant digit (LSD) is inevitably increased resulting frequently in a great error when the physical quantity is small.
In an effort to obviate the above defect, a proposal is made in, for example, U.S. Pat. No. 3,298,016 in which the magnitude of a physical quantity is coded as a combination of a numerical value of a predetermined number of digits making possible to express the numerical value by a predetermined number of bits and an index representing the largeness of the numerical value. This proposal is effective from the standpoint that physical quantities are converted into signals of the same number of bits and can be coded with high precision. However, the means disclosed in this proposal cannot be utilized when a fully digital unit is desired due to the fact that means for handling an analog quantity is included therein.
Summary of the Invention
It is therefore a primary object of the present invention to provide a coding equipment in which a parallel digital signal can be converted into a serial digital signal of predetermined number of bits without handling any analog quantity.
According to the present invention, the bit position of the most significant digit (MSD) in a parallel digital signal corresponding to a physical quantity is detected and coding is controlled depending on this bit position so that the physical quantity can be coded with high precision in a fully digital fashion.
Brief Description of the Drawing
FIGS. 1a and 1b illustrate errors occurring during AD conversion.
FIG. 2 shows a manner of information transmission according to the present invention.
FIG. 3 is a block diagram of a typical embodiment of the present invention.
FIG. 4 shows the detailed structure of the block 5 in FIG. 3.
FIG. 5 shows the detailed structure of the block 6 in FIG. 3.
FIGS. 6a and 6b show the detailed structure of the block 7 in FIG. 3.
FIGS. 7a and 7b are time charts illustrating the operation of the embodiment of the present invention.
FIG. 8 shows a manner of AD conversion conveniently used in the practice of the present invention.
FIG. 9 is a block diagram of an AD converter based on the principle shown in FIG. 8.
FIGS. 10 and 11 are views similar to FIGS. 3 and 2 respectively but showing another embodiment which employs the AD converter shown in FIG. 9.
Description of the Preferred Embodiments
The present invention will be described in detail with reference to an application in which a physical quantity sampled with a predetermined period is subjected to analog-digital conversion for measurement.
When physical quantities are measured and the measured quantities are transmitted in large amounts, a digital signal transmission system is generally employed in view of the advantages from the standpoints of the required frequency band and information processing, storing and other operations in a receiving station. In this system, therefore, an analog-digital converter is disposed in the transmitting station for converting various analog values measured at measuring points into digital quantities, and then the digital quantities are arranged in the form of a series signal to be transmitted to the receiving station. In this case, the maximum value of the error due to the AD conversion is given by .+-.1/2 of the least significant digit (LSD). Therefore, when the full scale of the AD converter is N, the maximum value of the error is given by .+-.1/2.multidot. (1/2.sup.N), and the error .epsilon. due to the AD conversion is generally expressed as follows: ##EQU1##
This relation is shown in FIG. 1a in which the horizontal axis represents the scale N and the vertical axis represents the error .epsilon.. It will be seen that the error .epsilon. is .epsilon..sub.1 when N= N.sub.1, and this error .epsilon. is .epsilon..sub.2 (.epsilon..sub.2 < .epsilon..sub.1) when N= N.sub.2 (N.sub.2 > N.sub.1). Thus, the error .epsilon. is reduced with the increase in the scale N.
On the other hand, the maximum value I.sub.MAX and minimum value I.sub.MIN of the analog input I are determined depending on the transmission system. In order that the conversion error .epsilon. of the AD converter can be limited to lower than an allowable maximum error .epsilon..sub.MAX allowable for the specific system, these values must be selected to satisfy the following relation (2) which is a severest condition: ##EQU2##
FIG. 1b shows the relation between the error .epsilon. and the analog input I selected to satisfy the inequality (2). It is apparent from FIG. 1b that, when the inequality (2) is satisfied, the minimum error .epsilon..sub.MIN corresponding to I.sub.MAX is included with a sufficient margin within the range of the allowable maximum error .epsilon..sub.MAX.
A digital transmission system for measured quantities is designed so that each individual measured quantity is expressed in the form of N bits determined by the full scale N of the AD converter regardless of the margin of the input I relative to the allowable maximum error .epsilon..sub.MAX as shown in FIG. 1b. This system has been significant in that a compact and standardized structure can be provided for the hardware processing the transmitted and received measured quantity, but this system has not necessarily been most suitable from the viewpoint of the efficiency of transmission.
The present invention provides a practically useful coding equipment which retains the merit of capability of standardization of the transmitted and received information processing hardware in the digital transmission system of the kind above described and yet which provides a great improvement in the transmission efficiency which has been the demerit of the prior art system.
More precisely, in the transmission of the parallel digital signal which are the output appearing after the AD conversion, the present invention utilizes positively the fact that the margin of the conversion error .epsilon. in the output value relative to the allowable maximum error .epsilon..sub.MAX allowable for the specific transmission system varies greatly depending on the magnitude of the analog input I as shown by the inequality (2) and in FIG. 1b. According to the present invention, the margin of the error is uniformly distributed as far as possible for all the AD converter outputs so that, when the input has a large value, the number of bits transmitted can be reduced by the amount corresponding to the margin of the error.
The manner of reducing the number of bits according to the present invention will now be described. In the transmission of a binary information output of an AD converter having a scale of N bits, the digit corresponding to the maximum level of the value obtained by sampling of the information, (that is, "1" nearest to the MSD in the AD converted output when the output is expressed in the form of binary information) is employed as the MSD of the information to be transmitted, and this digit is called hereinafter a TMSD for distinguishing same from the term MSD commonly employed. A data character DC of X bits following the TMSD and a weight character WC of Y bits representing the weight of the TMSD are solely transmitted. In order that the error .epsilon. in all the sampled and transmitted information may not exceed the allowable limit .epsilon..sub.MAX, X and Y may be selected as follows: ##EQU3##
When the TMSD is situated in the position spaced by K bits (K< N) from the LSD of the AD converter output, X in the transmitted information must satisfy the allowable maximum error .epsilon..sub.MAX. From the above condition, the following relation holds: ##EQU4## By solving the inequality (5), X satisfying the above condition can be sought from the inequality (3). Further, the number of bits Y required for indicating the weight of the TMSD is given by the following inequality:
since ##EQU5## in the numerator and denominator of the inequality (5) is a finite geometric series, this can be expressed as ##EQU6## Practical numerical values will now be considered for ease of understanding. When, for example, N and K are supposed to be n= 12 and K= 6 and .epsilon..sub.MAX is taken as a parameter, X and Y are sought from the inequalities (3) and (4) as shown in Table 1.
From Table 1, it is apparent that the number of bits X+ Y required for information transmission is 10 in the case No. 1, 9 in the cases Nos. 2 to 4, and 8 in the case No. 5. Thus, it is apparent that the number of bits is less by 2 bits in the case No. 1, by 3 bits in the cases Nos. 2 to 4, and by 4 bits in the case No. 5 than when the full scale N= 12 bits of the AD converter is transmitted within the range of the allowable maximum error .epsilon..sub.MAX. It will thus be understood that the method above described is very effective in improving the efficiency of transmission.
The present invention has a great significance in that a practical coding equipment is proposed in which conversion based on the inequalities (3) and (4) is carried out by digital means described later so that the number of bits transmitted can be reduced to a minimum within the range of the allowable maximum error .epsilon..sub.MAX allowable for a specific transmission system. In the present invention too, it is apparent that standardization of the transmitted and received information processing hardware can be attained as heretofore due to the fact that individual measured values are naturally transmitted in the form of binary information of equal bit length.
As a typical example, the present invention will be described with reference to an application to a transmission system in which AD converted data are arranged in series in the order of D1, D2 . . . . Dm as shown in FIG. 2 and such data signals are cyclically transmitted. The scale N and allowable maximum error .epsilon..sub.MAX may be those determined depending on the transmission system. It is assumed herein that the scale N= 12 and the allowable maximum error .epsilon..sub.MAX = 0.02. Since this corresponds to the case No. 2 shown in Table 1, the numbers of bits X and Y required for transmission are X= 6 and Y= 3. In actual transmission, one parity bit P is added to these bits X and Y to give a total of 10 bits. As shown in FIG. 2, the weight character WC of Y= 3 bits starts from the MSD and is followed by the data character DC of X= 6 bits which starts from the TMSD, and the data character DC is followed by the parity portion of P= 1 bit.
FIG. 3 is a block diagram showing a typical structure of a measured data transmitting equipment in an embodiment of the present invention. Referring to FIG. 3, the block 2 designates a timing control circuit which is connected to a clock pulse generator 1 and applies timing pulses to various parts of the equipment for regulating the operation of these parts. The block 3 designates an AD converter having a scale of N= 12 bits. The block 4 designates a 12-bit buffer register for temporarily storing therein the AD converted measured data before transmission. The block 5 designates a transfer control instruction circuit which detects the position of the TMSD in the data information stored in the buffer register 4 and applies instruction signals to a transfer gate 7 and weight character coding circuit 6 (described below) so as to instruct the portion of X= 6 bits to be selected from the data information for transfer and to instruct the construction of the weight character WC. The block 6 designates a circuit for coding the weight character WC in response to the instruction signal applied from the transfer control instruction circuit 5. The block 7 designates a data transfer gate which, in response to the application of the instruction signal from the transfer control instruction circuit 5, transfers the portion of X= 6 bits starting from the TMSD in the data information stored in the buffer register 4 to a PS converter (described below) as the data bits to be transmitted. The block 8 designates a PS converter which, in response to the application of the timing pulses from the timing control circuit 2, receives the signals applied in parallel from the weight character coding circuit 6 and data transfer gate 7 and arranges such signals in a serial signal as shown in FIG. 2. The blocks 9 and 10 designate a parity generation circuit and a synchronizing signal generation circuit respectively which are connected to an OR gate 11.
The data information is applied to the AD converter 3 from an input terminal A, and the serial signal appears from an output terminal B of the OR gate 11, this output terminal B being the output terminal of the transmitting equipment. An AD conversion instruction pulse signal is applied from the timing control circuit 2 to the AD converter 3 by a line TAD. A write instruction pulse signal is applied from the timing control circuit 2 to the buffer register 4 by a line TBR for instructing the writing of the output of the AD converter 3 in the buffer register 4. A timing regulation instruction pulse signal is applied from the timing control circuit 2 to the PS converter 8 by nine lines TM1 to TM9 provided for individual digits. A parity generation timing regulation instruction signal is applied from the timing control circuit 2 to the parity generation circuit 9 by a line TM10. Twelve lines 300 to 311, one for each digit, are provided for applying the output of the AD converter 3 to the buffer register 4. Twelve lines 400 to 411 are provided for transferring the output of the buffer register 4 to the transfer control instruction circuit 5 and the data transfer gate 7. Seven lines 505 to 511 are provided for applying the transfer control instruction pulse signal to the weight character coding circuit 6 and data transfer gate 7. Three lines 600 to 602 are provided for applying the coded weight character WC of Y= 3 bits to the PS converter 8. Six lines 700 to 705 are provided for applying the data character DC of X= 6 bits starting from the TMSD to the PS converter 8. A synchronizing signal addition timing regulation instruction pulse signal is applied from the timing control circuit 2 to the synchronizing signal generation circuit 10 by a line TSYC.
The clock pulse generator 1, timing control circuit 2, AD converter 3, buffer register 4, PS converter 8, parity generation circuit 9 and synchronizing signal generation circuit 10 in the embodiment of the present invention shown in FIG. 3 may be conventional ones, and therefore, any detailed description of the structure of these parts is unnecessary.
The present invention is featured by the fact that the transfer control instruction circuit 5, weight character coding circuit 6 and data transfer gate 7 are disposed between the buffer register 4 and the PS converter 8 so as to reduce the number of bits transmitted by the system. Therefore, the embodiment of the present invention will be described while placing emphasis on the structure and operation of these parts.
The structure of the transfer control instruction circuit 5 will be described at first. This circuit 5 detects the position of the TMSD in the data information written in the buffer register 4 and informs the weight character coding circuit 6 and data transfer gate 7 of the detected position of the TMSD. To this end, the transfer control instruction circuit 5 must have a structure which satisfies the condition described below. That is, whether the position of the TMSD in the buffer register 4 lies in the bit position 2.sup.11, bit position 2.sup.10, bit position 2.sup.9, bit position 2.sup.8, bit position 2.sup.7, bit position 2.sup.6 or bit positions ranging from 2.sup.5 to 2.sup.0 must be detected. Therefore, the transfer control instruction circuit 5 satisfying the above condition may have a structure as shown in FIG. 4.
Referring to FIG. 4, the transfer control instruction circuit 5 includes a plurality of INHIBIT gates 51 to 55, an OR gate 56 and a phase inverting gate 57. A signal of "1" level is applied to this transfer control instruction circuit 5 by the lines 411, 410, 409, 408, 407 and 406 when the TMSD in the data in the buffer register 4 lies in the bit positions 2.sup.11, 2.sup.10, 2.sup.9, 2.sup.8, 2.sup.7 and 2.sup.6 respectively. It can be seen therefore that the TMSD lies in one of the bit positions ranging from 2.sup.5 to 2.sup.0 when the signal of "1" level does not appear in anyone of these lines 411 to 406. An output signal of "1" level is applied to the weight character coding circuit 6 and data transfer gate 7 by the lines 511, 510, 509, 508, 507, 506 and 505 when the TMSD is detected to lie in the bit position 2.sup.11, bit position 2.sup.10, bit position 2.sup.9, bit position 2.sup.8, bit position 2.sup. 7, bit position 2.sup.6 and bit positions 2.sup.5 to 2.sup.0 respectively, so as to instruct the transfer control of signals applied from the weight character coding circuit 6 and data transfer gate 7 to the PS converter 8.
The weight character coding circuit 6 acts to code the weight character WC of Y= 3 bits which is added to the data character DC to indicate the weight of the TMSD in the data character DC to be transmitted, that is, the position of the TMSD in the AD converted data of N= 12 bits. Thus, the weight character coding circuit 6 may be constructed to satisfy the condition described below. That is, the weight character coding circuit 6 may deliver "110" when the TMSD detected by the transfer control instruction circuit 5 lies in the bit position 2.sup.11, hence when the output signal of "1" level appears on the line 511. In other words, the weight character coding circuit 6 may apply a signal of "1" level, a signal of "1" level and a signal of "0" level to the PS converter 8 by the lines 602, 601 and 600 respectively. Similarly, the weight character coding circuit 6 may deliver "101" when the detected TMSD lies in the bit position 2.sup.10, hence when the signal of "1" level appears on the line 510, "100 " when the detected TMSD lies in the bit position 2.sup.9, hence when the signal of "1" level appears on the line 509, "011" when the detected TMSD lies in the bit position 2.sup.8, hence when the signal of "1" level appears on the line 508, "010" when the detected TMSD lies in the bit position 2.sup.7, hence when the signal of "1" level appears on the line 507, "001" when the detected TMSD lies in the bit position 2.sup.6, hence when the signal of "1" level appears on the line 506, and "000" when the detected TMSD lies in one of the bit positions 2.sup.5 to 2.sup.0, hence when the signal of "1" level appears on the line 505. The respective output signals are applied by the lines 602 to 600 to the PS converter 8. Therefore, the weight character coding circuit 6 may have a structure as shown in FIG. 5.
Referring to FIG. 5, the weight character coding circuit 6 includes a plurality of OR gates 61 to 66, and a plurality of INHIBIT gates 67 to 69. The input is applied by the lines 505 to 511 and the output is delivered by the lines 600 to 602. The OR gates 62, 64, 66 and INHIBIT gates 67, 68, 69 are not necessarily required due to the fact that the input signal of "1" level appears generally on only one of the lines 505 to 511.
The function of the data transfer gate 7 is such that, in response to the instruction signal applied from the transfer control instruction circuit 5, the data character DC of X= 6 bits starting from the TMSD in the sampled and AD converted data information written in the buffer register 4 is applied to the PS converter 8. Thus, the data transfer gate 7 must be constructed to respond to the instruction signal applied from the transfer control instruction circuit 5. That is, when the TMSD detected by the transfer control instruction circuit 5 lies in the bit position 2.sup.11, hence when the signal of "1" level appears on the line 511, the 6-bit data information ranging from 2.sup.11 to 2.sup.6 in the buffer register 4 may be applied to the PS converter 8 by the lines 705 to 700 in the sequential order starting from the TMSD. Similarly, the 6-bit data information ranging from 2.sup.10 to 2.sup.5 may be applied to the PS converter 8 by the lines 705 to 700 when the detected TMSD lies in the bit position 2.sup. 10, hence when the signal of "1" level appears on the line 510, the 6-bit data information ranging from 2.sup.9 to 2.sup.4 when the detected TMSD lies in the bit position 2.sup.9, hence when the signal of "1" level appears on the line 509, the 6-bit data information ranging from 2.sup.8 to 2.sup.3 when the detected TMSD lies in the bit position 2.sup.8, hence when the signal of "1" level appears on the line 508, the 6-bit data information ranging from 2.sup.7 to 2.sup.2 when the detected TMSD lies in the bit position 2.sup.7, hence when the signal of "1" level appears on the line 507, the 6-bit data information ranging from 2.sup.6 to 2.sup.1 when the detected TMSD lies in the bit position 2.sup.6, hence when the signal of "1" appears on the line 506, and the 6-bit data information ranging from 2.sup.5 to 2.sup.0 when the detected TMSD lies in one of the bit positions 2.sup.5 to 2.sup.0 , hence when the signal of "1" level appears on the line 505. Therefore, the data transfer gate 7 may have a structure as shown in FIGS. 6a and 6b.
Referring to FIG. 6a, the data transfer gate 7 is composed of six blocks 7A to 7F each of which has a structure 70 as shown in detail in FIG. 6b. Referring to FIG. 6b, the block 70 includes a plurality of AND gates 71 to 77 and an OR gate 78 to which the outputs of these AND gates 71 to 77 are applied. Input lines a, b, c, d, e, f and g of each of the blocks 7A to 7F shown in FIG. 6a must be connected to the corresponding seven lines among the lines 400 to 411. Further, an output line h of the block 70 shown in FIG. 6b must be connected to the corresponding line among the lines 700 to 705 as shown in FIG. 6a.
It is apparent that the reduction in the number of bits described previously can be realized by the provision of the parts having structures as above described. The PS converter 8 which provides a series signal on the basis of input signals may have a structure commonly known in the art.
FIGS. 7a and 7b are time charts showing the waveform of timing regulating pulses applied from the timing control circuit 2 during this PS conversion and showing the manner of providing a serial signal. As shown in FIG. 7a, an AD conversion instruction pulse TAD appears when the timing signal TM10 is in the "1" level, and a write instruction pulse TBR for writing the output of the AD converter 3 in the buffer register 4 appears also when the timing signal TM10 is in the "1" level, but the pulse TBR appears after the disappearance of the pulse TAD. Thus, by use of timing regulating pulses TM1 to TM10 for regulating the timing of the individual digits, the lth data D(l) starting from time I(l) and the (l+ 1)th data D(l++ 1) starting from time T(l+ 1) can be arranged in a serial signal.
The generation of the parity bit P is carried out in the block 9 in FIG. 3 in response to the application of the timing signal TM10. Further, the generation of the synchronizing signal SYC is carried out in the block 10 in FIG. 3 in response to the application of a timing signal TSYC as shown in FIG. 7b. In this manner, a data signal as shown in FIG. 7b can be sent out from the output terminal B in FIG. 3 toward a receiving station. A modulator (not shown) may be connected to the output terminal B so that the output signal may be transmitted to the receiving station after being subjected to high frequency modulation.
It is apparent from the foregoing description of a typical embodiment of the present invention that the objects of the present invention can be completely attained and the desired practical effects can also be exhibited.
While the embodiment of the present invention has been described with reference to cyclic transmission, it is readily apparent from the foregoing description that the present invention is applicable also to other forms of transmission. Further, the present invention provides a great flexibility due to the fact that the number of required bits of data can be sought from the inequalities (3) and (4) on the basis of the condition given by a specific transmission system. Further, as a matter of course, the desired reduction in the number of bits can also be attained when the blocks 5, 6 and 7 are used as input converting means in a display circuit and a conventional data processing unit in addition to the use as means for data transmission.
In the embodiment above described, AD converted information including a data character DC of X= 6 bits starting from the TMSD, a weight character WC of Y= 3 bits indicating the weight of the TMDS and a parity P of 1 bit is transmitted in a form as shown in FIG. 2 and all the outputs of the AD converter have a margin distributed as equally as possible for the allowable maximum error .epsilon..sub.MAX so as to realize the desired reduction in the number of bits to be transmitted. In an application of the present invention, another bit can be reduced under a certain limited condition.
This is limited to the case in which the number of coded forms of the weight character WC (corresponding to the number of TMSD positions to be detected) has a margin for the value of Y sought from the inequality (4). Under such a condition, information as to whether the TMSD is "1" or "0" 0 can be coded by utilizing the margin in the weight character WC.
For example, in the case of the embodiment above described, Y= 3 bits and the weight character WC can be coded in eight forms, while the number of TMSD positions to be detected is seven since the number of bits of X sought from the inequality (3) is X= 6 and the TMSD in the bit position 2.sup.11, bit position 2.sup.10, bit position 2.sup.9, bit position 2.sup.8, bit position 2.sup.7, bit position 2.sup.6 and bit positions 2.sup.5 to 2.sup.0 is detected. Thus, the former is more than the latter by one. This may be utilized so as to detect the TMSD in the bit position 2.sup.5 and bit positions 2.sup.4 to 2.sup.0 and to code this TMSD position by a separate weight character. In such a case, it is apparent that all the data can be correctly decoded in the receiving station even when the bit portion 2.sup.X.sup.-1 in FIG. 2 corresponding to the TMSD in the data character DC may be omitted. Thus, the data information transmitted is composed of a total of 9 bits consisting of Y= 3 bits, X= 5 bits and P= 1 bit, and the number of bits transmitted is less than in the above embodiment by one.
It can be known from the above description that the coding equipment according to the present invention can reduce the number of bits in a serial digital signal when a parallel digital signal representing a physical quantity are produced from the AD converter. The AD converter 3 shown in FIG. 3 may be suitably modified for move effective applications as will be described below.
An error is naturally involved in the quantization of a measured quantity as described with reference to FIGS. 1a and 1b. A suitable value I.sub.D intermediate between a maximum I.sub.MAX and a minimum I.sub.MIN of a measured quantity I.sub.i is selected as shown in FIG. 8, and a value between I.sub.MAX and I.sub.D and a value between I.sub.D and I.sub.MIN are converted by an AD converter to give the same number of bits. The output of the AD converter includes also a weight bit which indicates the fact that the measured value lies in one of the two ranges above described. This is more effective for the reduction of the number of bits transmitted. This idea will be described by taking an AD converter of successive approximation type as an example.
According to this idea, the level of an analog input to be AD converted is compared with a reference value I.sub.D calculated on the basis of an allowable error .epsilon..sub.AL of the entire system and the input is quantized into a digital signal of q bits, but the weight is different depending on whether the input level is larger or smaller than the reference value I.sub.D. More concretely, in this case, the AD converter delivers a digital output as described below depending on the level of the input I.sub.i.
weight bit WB-- "1"
Data bits-- 2.sup.q-1, 2.sup.q-2, . . . 2.sup.1, 2.sup.0
The input falling within the range of I.sub.D to I.sub.MAX is AD converted so that the allowable error of I.sub.D is less than .epsilon..sub.AL.
weight bit WB-- "0"
Data bits-- 2.sup.q.sup.- 1, 2.sup.q.sup.-2, . . . 2.sup.1, 2.sup.0
The input falling within the range of I.sub.MIN to I.sub.D is AD converted so that the allowable error of I.sub.MIN is less than .epsilon..sub.MAX.
In this case, therefore, it is important to select the value of I.sub.D so as to obtain a satisfactorily balanced AD converter. When the value of I.sub.D is properly selected, the input I.sub.i and the AD conversion error .epsilon. have a relation as shown in FIG. 8 so that the AD conversion error included in the allowable error of the entire system can be suitably distributed and the number of bits converted can be reduced by an amount corresponding to the distribution of the error.
Now, I.sub.D and q will be computed supposing that I.sub.MAX = 14100 A, I.sub.MIN = 100 A, .epsilon..sub.MAX-- .ltoreq. .+-. 0.3 and .epsilon..sub.AL-- .ltoreq. .+-. 0.04. In the AD conversion in which the weight bit WB is "0" , the severest condition for the inequality (2) must be satisfied in order that the conversion error is less than the allowable error .epsilon..sub.MAX shown in FIG. 8. That is, the following relation must hold since I.sub.MAX in this case corresponds to I.sub.D : ##EQU7## The numerator in the left-hand member in the inequality (8) can be expressed as .+-.1/2 .multidot. I.sub.LSD1 because ##EQU8## where I.sub.LSD1 is the weight of the least significant digit (LSD) of the quantized value. Thus, the inequality (8) can be expressed as ##EQU9## From this inequality (9), the weight I.sub.LSD1 of the least significant digit is given by
it is known from the inequality (10) that I.sub.LSD1 may be selected to satisfy the value I.sub.LSD1 < 60 when the values of I.sub.MIN = 100 A and .epsilon..sub.MAX = 0.3 are put in the inequality (10). In view of the arrangement of the quantized bits in the AD converter, it is desirable to select a suitable value of I.sub.LSD1 which satisfies the above condition in which I.sub.MIN is selected to be I.sub.MIN = 100 A. That is, it is desirable to select a value of I.sub.LSD1 to be
therefore, the arrangement of bits when the weight bit WB= "0" is determined as shown in Table 2.
However, the bit position of (I.sub.D /2) cannot yet be determined from the above condition.
In order that the AD conversion error be less than the allowable error .epsilon..sub.AL shown in FIG. 8, when the weight bit WB is "1" , the relation ##EQU10## must be satisfied since I.sub.D in this case corresponds to I.sub.MIN in the inequality (2). The numerator in the left-hand member in the inequality (11) can be expressed as .+-.1/2.multidot. I.sub.LSD2 where I.sub.LSD2 is the weight of the least significant digit of the quantized value. Thus, the inequality (11) can be expressed as ##EQU11## From the inequality (12), I.sub.LSD2 is given by
the value of I.sub.D relative to I.sub.MIN = 100 is selected to be I.sub.D = 12800, and this value and the value of .epsilon..sub.AL = 0.04 are put in the inequality (13). It is known that I.sub.LSD2 is selected to satisfy the relation I.sub.LSD2 .ltoreq. 1024. This I.sub.LSD2 is also desirably selected to be similar to one in the bit arrangement obtained when the weight bit WB is "0" or the value of I.sub.LSD2 is desirably selected from the numbers which are obtained by multiplying the numbers in the right-hand column of Table 2 above described. Suppose that the value of I.sub.LSD2 satisfying these conditions is I.sub.LSD2 = 800, then the bit arrangement when the weight bit WB is "1" is as shown in Table 3.
It will be seen that these values are well balanced with those in the case in which the weight bit WB is "0" , and I.sub.D = 12800 is an appropriate value. Therefore, the bit arrangement in this case is finally determined from Table 4.
It will be seen from Table 4 that I.sub.D is preferably selected to be I.sub.D = 12800 A, and the number of bits required for the AD converter in this case is q+ 1= 9 bits including the weight bit WB. It is thus apparent that the number of required bits is less by three bits than that in the successive approximation type in which twelve bits are required. This verifies the effectiveness of the AD converter employed in the equipment according to the present invention.
Referring to FIG. 9, an analog input is applied to an input terminal A of the AD converter, and after the period of time required for conversion, a digital output of (q+ 1) bits corresponding to the level of the analog input appears at a plurality of terminals at the output end DO of the AD converter. The structure and operation of this AD converter will be described in more detail with reference to FIG. 9.
Flip-flops FF.sub.7 to FF.sub.0 are arranged to correspond to the weights of individual bits and are set successively in the above order in response to the application of timing clock pulses from a pulse distributing circuit PD. The flip-flops FF.sub.7 to FF.sub.0 control associated switching means SW.sub.7 to SW.sub.0 and weight generating means W.sub.7 to W.sub.0 respectively so that the voltage of a reference voltage source (V.sub.1 or V.sub.2) can be applied or not applied to the weight generating means W.sub.7 to W.sub.0 depending on the state of the respective flip-flops FF.sub.7 to FF.sub.0. These voltage values are then added together in an analog adder means .SIGMA. to obtain an output voltage E.sub..SIGMA. which is compared in a comparator COM with an input voltage E.sub.i applied to the input terminal A of the AD converter. When E.sub.i .gtoreq. E.sub..SIGMA. , the output of the comparator COM is "1" and the flip-flops FF.sub.7 to FF.sub.0 remain in the set position, while when E.sub.i < E.sub..SIGMA. , the output of the comparator COM is "0" and the flip-flops FF.sub.7 to FF.sub.0 are reset in response to the application of the next timing clock pulses. The above operation is repeated by the number of bits according to the order of from the largest weight to the smallest weight. (In the case of conversion into a binary number, precision of (1/2.sup.q) is obtained with repetition of q times). When the above operation is repeated q times, that is, when one cycle is completed, the output values of the individual flip-flops FF.sub.7 to FF.sub.0 represent the AD converted output DO (2.sup.7 to 2.sup.0). INHIBIT gates G.sub.7 to G.sub.0 in FIG. 9 control the application of reset pulses to the respective flip-flops FF.sub.7 to FF.sub.0. For example, these gates G.sub.7 to G.sub.0 inhibit passage of the reset pulses when the output of the comparator COM is "1" and allow for passage of the reset pulses when the output of the comparator COM is "0" .
The operation of the AD converter shown in FIG. 9 will be described in more detail. The pulse distributing circuit PD generate successive timing clock pulses CP.sub.0 to CP.sub.9. An output appears at the terminal Q or Q of each of the flip-flops FF.sub.7 to FF.sub.0 when a set pulse or reset pulse is applied to the terminal S or R thereof. The INHIBIT gates G.sub.0 to G.sub.7 control the application of the reset pulses to the associated flip-flops FF.sub.0 to FF.sub.7. The reference voltage source V.sub.1 is used when the relation between the input E.sub.i and an intermediate reference voltage E.sub.D is given by E.sub.i .gtoreq. E.sub.D. The reference voltage source V.sub.2 is used when the relation between E.sub.i and E.sub.D is given by E.sub.i < E.sub.D. E.sub.i corresponds to I.sub.i described previously. The switching means SW.sub.0 to SW.sub.7 control the application of the reference voltage of the reference voltage source V.sub.1 or V.sub.2 to the weight generating means W.sub.0 to W.sub.7 depending on the output of the flip-flops FF.sub.0 to FF.sub.7. The weight generating means W.sub.0 to W.sub.7 generate corresponding weights in response to the application of the reference voltage from the reference voltage source V.sub.1 or V.sub.2. The analog adder means .SIGMA. add the outputs of the weight generating means W.sub.0 to W.sub.7 to provide the output E.sub..SIGMA. . The comparator COM compares the analog input E.sub.i applied to the input terminal A of the AD converter with the output E.sub..SIGMA. of the analog adder means .SIGMA. to deliver "1" when E.sub.1 .gtoreq. E.sub..SIGMA. and "0" when E.sub.i < E.sub..SIGMA. . An AND gate GW.sub.1 allows for passage of the timing clock pulse CP.sub.1 only when the relation E.sub.i .gtoreq. E.sub.D holds as a result of comparison between E.sub.i and E.sub.D immediately after the AD conversion is started. An INHIBIT gate GW.sub.2 allows for passage of the timing clock pulse CP.sub.1 only when the relation E.sub.i < E.sub.D holds. Two OR gates OG.sub.1 and OG.sub.2 are connected to the set and reset terminals S and R respectively of the flip-flop FF.sub.4 as shown. A flip-flop FF.sub.W acts as a weight bit memory. The digital output of the AD converter appears at the output end DO. Especially notable parts of the AD converter shown in FIG. 9 are the OR gates OG.sub.1 and OG.sub.2 for controlling the determination of the weight bit WB, the flip-flop FF.sub.W for memorizing the determined weight bit WB, and the two reference voltage sources V.sub.1 and V.sub.2 selectively used depending on the output of the flip-flop FF.sub.W. The remaining parts of the AD converter may be substantially similar to those commonly employed in the art and the function thereof is as described above.
The AD converter shown in FIG. 9 is featured by the fact that the level of the analog input E.sub.i is compared with the intermediate reference voltage E.sub.D computed on the basis of the allowable error of the system in response to the application of the timing clock pulse CP.sub.0 to the flip-flop FF.sub.4 from the pulse distributing circuit PD at the beginning of the AD conversion, one of the reference voltage sources V.sub.1 and V.sub.2 is selected by the flip-flop FF.sub.W depending on whether E.sub.i is larger or smaller than E.sub.D and the individual bits are quantized by different weights depending on the relation between E.sub.i and E.sub.D. Therefore, the operation of the AD converter will be described while placing emphasis on these parts.
As described previously, the AD converter shown in FIG. 9 is designed to suit the bit arrangement shown in Table 4. Thus, the intermediate reference voltage E.sub.D is selected to correspond to the bit position 2.sup.4. Therefore, the timing clock pulse CP.sub.0 appearing at the beginning of the AD conversion passes through the OR gate OG.sub.1 to be applied to the flip-flop FF.sub.4 corresponding to the specific weight above described. The switching means SW.sub.4 is closed and the output voltage E.sub.D of the reference voltage source V.sub.1 is applied through the switching means SW.sub.4 to the weight generating means W.sub.4 so that the output E.sub..SIGMA. of the analog adder means .SIGMA. at this time is the output voltage E.sub.D of the weight generating means W.sub.4. The reference voltage source V.sub.1 can be used initially due to the fact that the flip-flop FF.sub.W is placed in the set position by an instruction signal applied from a circuit (not shown) each time the AD conversion is started. Further, all the flip-flops FF.sub.0 to FF.sub.7 are placed in the reset position by the circuit (not shown) each time the AD conversion is started.
The comparator COM compares the output E.sub..SIGMA. = E.sub.D of the analog adder means .SIGMA. with the input E.sub.i applied to the input terminal A and delivers an output of "1" level when, for example, E.sub.i .gtoreq. E.sub..SIGMA. . As a result, the AND gate GW.sub.1 is opened to allow for passage of the next timing clock pulse CP.sub.1 so that the flip-flop FF.sub.W continues to be maintained in the set position. Thus, the reference voltage source V.sub.1 can be used for the subsequent AD converting operation.
On the other hand, when E.sub.i > E.sub..SIGMA. , the comparator COM delivers an output of "0" level and the INHIBIT gate GW.sub.2 is opened to allow for passage of the timing clock pulse CP.sub.1 thereby resetting the flip-flop FF.sub.W. In this case, the reference voltage source V.sub.2 can be used for the subsequent AD converting operation.
It is thus apparent that the weight for quantizing the input E.sub.i can be determined at the beginning of the AD conversion by the AD converter which is one of the features of the present invention. After the determination of the weight for quantizing the input E.sub.i, the timing clock pulse CP.sub.1 is applied through the OR gate OG.sub.2 to the reset terminal R of the flip-flop FF.sub.4, and at the same time, to the set terminal S of the flip-flop FF.sub.7 for generating the instruction signal for the greatest weight. It will be easily understood that this can be attained in the entirely same manner as that described briefly with reference to the AD conversion of successive approximation type. Upon completion of the conversion of all the bits, the outputs of the individual flip-flops FF.sub.W, FF.sub.7, FF.sub.6, . . . FF.sub.1, FF.sub.0 appear as the output of the AD converter.
Therefore, when the AD converter having a structure and function as shown in FIG. 9 is used in the coding equipment according to the present invention, the present invention may be applied to the bit positions 2.sup.0 to 2.sup.7 of the output DO in FIG. 9, and at the same time, the weight bit WB may form an element of the weight character WC. In this case, it is preferable that the weight bit WB is handled as an entirely independent element as shown in FIG. 10 and the serial digital signal may have a form as shown in FIG. 11. FIGS. 10 and 11 correspond to FIGS. 3 and 2 respectively and illustrate the case in which the AD converter of the type shown in FIG. 9 is employed in the coding equipment of the present invention. As mentioned above, it is possible to reduce the necessary number of bits of each output code for keeping the possible error of the output within a predetermined allowable value by use of the A-D converter as shown in FIG. 9 compared with the number of bits as being necessary for assuring the same error when the inputs are applied to the circuit as shown in FIG. 3. FIG. 10 shows a block diagram of an embodiment in which the inputs undergo discrimination by a circuit as shown in FIG. 9 as to whether the respective values of the inputs are greater than the predetermined value I.sub.D or not.
Claims
We claim:
1. A coding equipment for encoding a physical quantity comprising analog-digital converting means for converting the physical quantity to a digital code signal, said analog-digital converting means including discriminator means having comparator means for discriminating whether the physical quantity is larger than a predetermined value or not, selecting means for selecting one of a plurality of predetermined different reference values in accordance with the output of said discriminating means, means for determining a weight for the digits of the digital code signal, said comparator means comparing the physical quantity with a value which is a function of the selected reference value and a weight which is predetermined depending on the position of each of the digits constituting the digital code signal, and compressing means for determining the position of the most significant digit in said digital code signal and for compressing said digital code signal in dependence on the position of the most significant digit in said digital code signal.
2. A coding equipment according to claim 1, wherein said compressing means produces a train of codes including a code indicative of the position of the most significant digit and a predetermined number of digits succeeding the most significant digit.
3. A coding equipment according to claim 2, further comprising means for generating a parity bit and a synchronizing signal for the train of codes of said compressing means.
4. A coding equipment comprising: analog-digital converting means including a plurality of first flip-flop means arranged in a predetermined order and having first and second input terminals, each of said flip-flop means being initially at a first state and changeable to a second state from the first state in response to a pulse signal applied to its first input terminal, said first flip-flop means being changeable to the first state from the second state in response to a pulse signal applied to its second input terminal, a plurality of switch means, each of said switch means being coupled to a respective one of said first flip-flop means and being responsive to the second state of the respective first flip-flop means for being rendered operative to permit an input signal applied thereto to pass therethrough, a plurality of weighting means, each of said weighting means being coupled to a respective one of said switching means for receiving the signal passed through the respective switching means and producing an output which is a function of the signal and a weight predetermined for said weighting means, adder means for producing an output corresponding to the sum of the outputs of said weighting means, comparator means for comparing the output of said adder means with a physical quantity to be coded and for producing a signal of a first level when said physical quantity is not less than the output of said adder means and a signal of second level when said physical quantity is less than the output of said adder means, second flip-flop means having first and second input terminals, said second flip-flop means being initially at a first state and changeable from the first to a second state in response to a pulse signal applied to its second input terminal, said second flip-flop means being changeable from the second state to the first state in response to a pulse signal applied to its first input terminal, means for applying to each of said switching means, as the input signal thereto, a first reference signal indicative to a first reference value when said second flip-flop means is at the first state and a second reference signal indicative of a second reference value when said second flip-flop means is at the second state, pulse distributing means for producing a plurality of signals in a predetermined order and in a predetermined timing, said pulse distributing means applying the first pulse signal to the first input terminal of a predetermined one of said plurality of first flip-flop means and applying the second pulse signal to the first input terminal of the first of the predetermined arrangement of said first flip-flop means and the second input terminal of said predetermined one of said first flip-flop means, said pulse distributing means applying the third and succeeding pulse signals to the first input terminals of the second and succeeding first flip-flop means, respectively, means for applying a pulse signal to the first input terminal of said second flip-flop means in response to said comparator means producing the first level signal and the second pulse signal of said pulse distributing means, and for applying a pulse signal to the second input terminal of said second flip-flop means in response to said comparator means producing the second level signal and said second pulse signal of said pulse distributing means, means for applying the third and successive pulse signals produced by said pulse distributing means to the second input terminal of the first flip-flop means in accordance with the second level signal being produced by said comparator means, and output means for producing a plurality of output signals indicative of whether respective ones of the plurality of first flip-flop means are at the first state or the second state and an output signal indicative of whether said second flip-flop means is at the first state or the second state; means for determining the position of the first flip-flop means in the first state corresponding to an output in the most significant digit position; encoding means for producing a code in dependence on the position of said determined first flip-flop means; gate means for permitting selected ones of said plurality of output signals from said output means, respectively, corresponding to a predetermined number of the first flip-flop means succeeding said determined first flip-flop means to pass therethrough; and means responsive to at least the code produced by said encoding means and said selected output signals of said output means for producing a train of codes.
Patent Citations (11)
| Patent | Date | Inventor | Cited By |
|---|---|---|---|
| US2970309 | 1961-01-01 | Towles | |
| US3069504 | 1962-12-01 | Kaneko | |
| US3460132 | 1969-08-01 | Juliusburger et al. | |
| US3500247 | 1970-03-01 | Sekimoto et al. | |
| US3594560 | 1971-06-01 | Stanley | |
| US3618047 | 1971-11-01 | Hertz | |
| US3631464 | 1971-12-01 | Dahlberg | |
| US3681574 | 1972-08-01 | Caulkin | |
| US3688221 | 1972-08-01 | Fruhalf | |
| US3789392 | 1974-01-01 | Candiani | |
| US3855589 | 1974-12-01 | Solender |