CA1124868A - Charge redistribution circuits - Google Patents
Charge redistribution circuitsInfo
- Publication number
- CA1124868A CA1124868A CA320,239A CA320239A CA1124868A CA 1124868 A CA1124868 A CA 1124868A CA 320239 A CA320239 A CA 320239A CA 1124868 A CA1124868 A CA 1124868A
- Authority
- CA
- Canada
- Prior art keywords
- switching means
- common line
- plates
- capacitor
- commonly coupled
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired
Links
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M1/00—Analogue/digital conversion; Digital/analogue conversion
- H03M1/12—Analogue/digital converters
- H03M1/50—Analogue/digital converters with intermediate conversion to time interval
- H03M1/504—Analogue/digital converters with intermediate conversion to time interval using pulse width modulation
- H03M1/506—Analogue/digital converters with intermediate conversion to time interval using pulse width modulation the pulse width modulator being of the charge-balancing type
Abstract
ABSTRACT OF THE DISCLOSURE
A charge redistribution circuit for pulse code modulation encoding suitable for use in apparatus for encoding voice in a companded code to maximize signal-to-noise performance. The companded code is the fifteen segment approximation of the µ255 compression law, wherein each segment contains 16 equal size steps.
The circuit includes a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded. Each capacitor in the array has a pair of opposed plates, one plate of each capacitor being commonly coupled together.
The circuit further includes a binary comparator having a first input adapted to be coupled to a reference bias potential and a second input coupled to the commonly coupled plates. First switching means connect and disconnect the commonly coupled plates to and from ground potential. Second switching means selectively connect a first common line to an input analog voice signal or to ground potential. Third switching means selectively connect a second common like to a predetermined reference voltage Vref or to a reference voltage Vref/16 (which is 1/16 the predetermined voltage), and selectively couple the first comparator input to ground potential or to the reference bias potential. Fourth switching means, including a switch for each capacitor, selectively couple the other plates of the capacitors to the first or to the second common line. An eight-bit code generator provides a digital eight-bit output upon receipt of a read instruction and control control logic for controlling the four switching means, such logic being responsive to the output of the comparator and providing control signals to the generator. Simultaneously, it causes the third switching means to connect the first comparator input to ground when the sampled input signal lies in the upper four segments and it causes the third switching means to connect the first comparator input to a reference bias potential when the sampled input signal lies in the lower four segments.
A charge redistribution circuit for pulse code modulation encoding suitable for use in apparatus for encoding voice in a companded code to maximize signal-to-noise performance. The companded code is the fifteen segment approximation of the µ255 compression law, wherein each segment contains 16 equal size steps.
The circuit includes a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded. Each capacitor in the array has a pair of opposed plates, one plate of each capacitor being commonly coupled together.
The circuit further includes a binary comparator having a first input adapted to be coupled to a reference bias potential and a second input coupled to the commonly coupled plates. First switching means connect and disconnect the commonly coupled plates to and from ground potential. Second switching means selectively connect a first common line to an input analog voice signal or to ground potential. Third switching means selectively connect a second common like to a predetermined reference voltage Vref or to a reference voltage Vref/16 (which is 1/16 the predetermined voltage), and selectively couple the first comparator input to ground potential or to the reference bias potential. Fourth switching means, including a switch for each capacitor, selectively couple the other plates of the capacitors to the first or to the second common line. An eight-bit code generator provides a digital eight-bit output upon receipt of a read instruction and control control logic for controlling the four switching means, such logic being responsive to the output of the comparator and providing control signals to the generator. Simultaneously, it causes the third switching means to connect the first comparator input to ground when the sampled input signal lies in the upper four segments and it causes the third switching means to connect the first comparator input to a reference bias potential when the sampled input signal lies in the lower four segments.
Description
1124~368 .
CH~RGE REDISTRIBUTION CIRCUITS
This invention relates to new and improved charge redistribution circuitry for pulse code modulation encoding and, in particular, to such circuitry for use 5 in apparatus for encoding voice in a companded code to maximize signal-to-noise performance. Accordingly, it is a general object of this invention to provide a new and improved circuit of such character.
Voice signals can be coded in a nonlinear digital code for digital telephone transmission in order to maintain a high ratio of signal-to-noise over a wide range without requiring excessive digital transmission bandwidth. One widely used code in the United States is the 15 segment approximation of the ~=255 compression law. An a~alog signal, such as speech, can be con-verted into a digital coded output using an 8-bit pulse code modulation format in which~ in each of the 1~
segments, there are 16 equal steps, with the ex~eption of the segment passing through the origin which contains 31 equal steps. In such case, a standard sampling rate for voice signals is 8 KHz, resulting in a 64 kilobits PCM signal.
.
., ~1~24868 Coders implementing such code have been relatively comple~. For economic reasons, the coder/decoder (codec) is often shared over 24 or more channels through the use of analog multiplexers at the input and output. However, in many cases, a low cost codec per channel offers advantages. Recognition thereof has led to extensive studies of alternative codes, particularly linear and adaptive delta modulation which allow particularly simple codec hardware imple-10 mentations. Drawbacks to delta modulation are itsincompatability with standard PCM, and poor performance for data transmission.
A ~255 PCM voice encoder which is amenable to on-chip monolithic realization in standard MOS technology 15 has been proposed in the prior art. Such a coder includes some elements found in linear A/D converters suggested by others. The prior art encoder includes two binary-weighted capacitor arrays as the key precision elements. Other elements include a binary 20 comparator, a near unity gain buffer amplifier with very high input impedance, a reference source VR, and numerous analog switches. Binary sequencing and control logic totalling about 60 gates and flip-flops and a digital clock are required. A negative reference can 25 be developed by a sub-circuit, so that only a single positive reference is required in the main circuitry.
Further description of such a proposed coder is discussed hereinafter following the "Brief Description of the Drawingsl' under description of a "Prior Art 30 PCM Coder'l.
.
~124868 A disadvantage of prior art coding of companded pulse code modulation is that the implementation of the step voltage settings is achieved by means of a lower or step capacitor array feeding a unity gain 5 buffer amplifier. One of the main constraints permits the time required for the full encoding cycle to include the charging of the top and bottom capacitor arrays and the settling time of the buffer amplifier and comparator circuits. Thus, it is another object 10 of this invention to provide a new and improved charge redistribution circuit ~hich is faster than sim~lar circuits of the prior art.
Yet another object of this invention is to provide a new and improved charge redistribution circuit for 15 use as a shared multi-channel analog-to-digital encoder for pulse code modulation use in which the encoding cycle can be completed within l/m X 125 microseconds for multichannel encoding (i.e., for m channels).
Still yet another object of this invention is 20 provide a new and improved charge redistribution circuit in which faster operation than similar devices of the prior art obtained.
Another ob3ect of this invention is to provide a new and improved charge redistribution circu;t which need utilize reference voltage generation of but one polarity.
1~2~68 Yet still another object of this invention is to provide a new and improved charge redistribution circuit which can be produced on a monolithic large scale integrated circuit device using a minimal amount of die area.
In accordance with one embodiment of the invention, a charge redistribution circuit for pulse code modulation encoding is provided. Such circuit is suitable in apparatus for encoding voice in a companded code to 10 maximize signal-to-noise performance. The companded ~ode is the 15 segment approximation of the ~255 compression law, each segment containing 16 equal size steps. The circuit includes a single binary weighted capacitor array for the generation of both segment and 15 step boundaries of voice being encoded. Each capacitor of the array has a pair of opposed plates One plate of each capacitor is commonly coupled together. The circuit further includes a binary comparator having a first input adapted to be coupled to a reference bias 20 potential and a second input coupled to the commonly coupled plates of the capacitor array. A first switching means connects the commonly coupled plates to ground potential and disconnects the plates therefrom. The circuit further includes second switching means for 25 selectively connecting a first common line to an input analog voice signal or to ground potential. Further included is a third switching means for selectively connecting a second common line to a predet~rmined reference voltage Vref or to a reference voltage Vre~n 1~2~868 which is l/nth the predetermined voltage, and for selectively coupling the first comparator input to ground potential or to the reference bias potential.
The circuit also includes fourth switching means including a switch for eadh of the capacitors which selectively couples the other plates of eadh of the capacitors of the array to the first common line or to the second eommon line Code generating means provide a digital output (e.g., eight bits) upon 10 receipt of a read instruction and control signals applied thereto. Control logic means control the four switching means which are responsive to the output of the comparator and provide contro1 signals through the generating means. In accordance with certain features 15 of the invention, n = 16. In accordance with one embodiment of the invention in sequence: (1) The control logic means causes the first switching means to connect the commonly coupled plates to ground potential, causes the fourth switching means to couple the other plates 20 to the first common line, and causes the second switching means to connect the first common line to the receiving means, whereby the analog voice signal can be sampled and held as a charge stored on the capacitor array.
(2~ The control logic means causes the first switching 25 means to disconnect the commonly connected plates ~rom ground potential, and causes the second switching means ~248~8 to connect the first common line to ground potential so that the sign of the voltage on the second input o~ the comparator, as detected by the comparator, is the negative of the sign of the sampled analog voice signal, thereby determining the polarity and sign bit of the pulse code modul~tion. (3) The control logic means causes the third switching means to connect the second common line to the predetermined reference voltage Vref and causes the fourth switching means 10 to superimpose a reference voltage equivalent to the top of segment 4 to the commonly coupled plates of the capacitor array by connecting the four lower value ca~acitors to the second common line as an initial sequence of a st~ndard three sequence segment-selection 15 successive approxi.mation algorithm while the remainder of the array remains coupled to ground. (4) The control logic means causes the third switching means to continue to connect the second common line to the predetermined reference voltage Vref when the sampled 20 input signal lies in the upper four segments, and it causes the third switching means to connect the second common line to the reference voltage Vref/16 when the sampled input signal lies in the lower four segments.
112~868 Simultaneously, it causes the third switching means to continue to connect the first comparator input to g~ound when the sampled input signal lies in the upper four segments; and it causes the third switching means to connect the first comparator input to a reference bias potential when the sampled input signal lies in the lower four segments. It further causes the fourth switching means as part of the second sequence of the standard three sequence segment-selection successive 10 approximation algorithm to superimpose either a reference voltage equivalent to the top of segment 6 or segment
CH~RGE REDISTRIBUTION CIRCUITS
This invention relates to new and improved charge redistribution circuitry for pulse code modulation encoding and, in particular, to such circuitry for use 5 in apparatus for encoding voice in a companded code to maximize signal-to-noise performance. Accordingly, it is a general object of this invention to provide a new and improved circuit of such character.
Voice signals can be coded in a nonlinear digital code for digital telephone transmission in order to maintain a high ratio of signal-to-noise over a wide range without requiring excessive digital transmission bandwidth. One widely used code in the United States is the 15 segment approximation of the ~=255 compression law. An a~alog signal, such as speech, can be con-verted into a digital coded output using an 8-bit pulse code modulation format in which~ in each of the 1~
segments, there are 16 equal steps, with the ex~eption of the segment passing through the origin which contains 31 equal steps. In such case, a standard sampling rate for voice signals is 8 KHz, resulting in a 64 kilobits PCM signal.
.
., ~1~24868 Coders implementing such code have been relatively comple~. For economic reasons, the coder/decoder (codec) is often shared over 24 or more channels through the use of analog multiplexers at the input and output. However, in many cases, a low cost codec per channel offers advantages. Recognition thereof has led to extensive studies of alternative codes, particularly linear and adaptive delta modulation which allow particularly simple codec hardware imple-10 mentations. Drawbacks to delta modulation are itsincompatability with standard PCM, and poor performance for data transmission.
A ~255 PCM voice encoder which is amenable to on-chip monolithic realization in standard MOS technology 15 has been proposed in the prior art. Such a coder includes some elements found in linear A/D converters suggested by others. The prior art encoder includes two binary-weighted capacitor arrays as the key precision elements. Other elements include a binary 20 comparator, a near unity gain buffer amplifier with very high input impedance, a reference source VR, and numerous analog switches. Binary sequencing and control logic totalling about 60 gates and flip-flops and a digital clock are required. A negative reference can 25 be developed by a sub-circuit, so that only a single positive reference is required in the main circuitry.
Further description of such a proposed coder is discussed hereinafter following the "Brief Description of the Drawingsl' under description of a "Prior Art 30 PCM Coder'l.
.
~124868 A disadvantage of prior art coding of companded pulse code modulation is that the implementation of the step voltage settings is achieved by means of a lower or step capacitor array feeding a unity gain 5 buffer amplifier. One of the main constraints permits the time required for the full encoding cycle to include the charging of the top and bottom capacitor arrays and the settling time of the buffer amplifier and comparator circuits. Thus, it is another object 10 of this invention to provide a new and improved charge redistribution circuit ~hich is faster than sim~lar circuits of the prior art.
Yet another object of this invention is to provide a new and improved charge redistribution circuit for 15 use as a shared multi-channel analog-to-digital encoder for pulse code modulation use in which the encoding cycle can be completed within l/m X 125 microseconds for multichannel encoding (i.e., for m channels).
Still yet another object of this invention is 20 provide a new and improved charge redistribution circuit in which faster operation than similar devices of the prior art obtained.
Another ob3ect of this invention is to provide a new and improved charge redistribution circu;t which need utilize reference voltage generation of but one polarity.
1~2~68 Yet still another object of this invention is to provide a new and improved charge redistribution circuit which can be produced on a monolithic large scale integrated circuit device using a minimal amount of die area.
In accordance with one embodiment of the invention, a charge redistribution circuit for pulse code modulation encoding is provided. Such circuit is suitable in apparatus for encoding voice in a companded code to 10 maximize signal-to-noise performance. The companded ~ode is the 15 segment approximation of the ~255 compression law, each segment containing 16 equal size steps. The circuit includes a single binary weighted capacitor array for the generation of both segment and 15 step boundaries of voice being encoded. Each capacitor of the array has a pair of opposed plates One plate of each capacitor is commonly coupled together. The circuit further includes a binary comparator having a first input adapted to be coupled to a reference bias 20 potential and a second input coupled to the commonly coupled plates of the capacitor array. A first switching means connects the commonly coupled plates to ground potential and disconnects the plates therefrom. The circuit further includes second switching means for 25 selectively connecting a first common line to an input analog voice signal or to ground potential. Further included is a third switching means for selectively connecting a second common line to a predet~rmined reference voltage Vref or to a reference voltage Vre~n 1~2~868 which is l/nth the predetermined voltage, and for selectively coupling the first comparator input to ground potential or to the reference bias potential.
The circuit also includes fourth switching means including a switch for eadh of the capacitors which selectively couples the other plates of eadh of the capacitors of the array to the first common line or to the second eommon line Code generating means provide a digital output (e.g., eight bits) upon 10 receipt of a read instruction and control signals applied thereto. Control logic means control the four switching means which are responsive to the output of the comparator and provide contro1 signals through the generating means. In accordance with certain features 15 of the invention, n = 16. In accordance with one embodiment of the invention in sequence: (1) The control logic means causes the first switching means to connect the commonly coupled plates to ground potential, causes the fourth switching means to couple the other plates 20 to the first common line, and causes the second switching means to connect the first common line to the receiving means, whereby the analog voice signal can be sampled and held as a charge stored on the capacitor array.
(2~ The control logic means causes the first switching 25 means to disconnect the commonly connected plates ~rom ground potential, and causes the second switching means ~248~8 to connect the first common line to ground potential so that the sign of the voltage on the second input o~ the comparator, as detected by the comparator, is the negative of the sign of the sampled analog voice signal, thereby determining the polarity and sign bit of the pulse code modul~tion. (3) The control logic means causes the third switching means to connect the second common line to the predetermined reference voltage Vref and causes the fourth switching means 10 to superimpose a reference voltage equivalent to the top of segment 4 to the commonly coupled plates of the capacitor array by connecting the four lower value ca~acitors to the second common line as an initial sequence of a st~ndard three sequence segment-selection 15 successive approxi.mation algorithm while the remainder of the array remains coupled to ground. (4) The control logic means causes the third switching means to continue to connect the second common line to the predetermined reference voltage Vref when the sampled 20 input signal lies in the upper four segments, and it causes the third switching means to connect the second common line to the reference voltage Vref/16 when the sampled input signal lies in the lower four segments.
112~868 Simultaneously, it causes the third switching means to continue to connect the first comparator input to g~ound when the sampled input signal lies in the upper four segments; and it causes the third switching means to connect the first comparator input to a reference bias potential when the sampled input signal lies in the lower four segments. It further causes the fourth switching means as part of the second sequence of the standard three sequence segment-selection successive 10 approximation algorithm to superimpose either a reference voltage equivalent to the top of segment 6 or segment
2 to the commonly coupled plates of the capacitor array, dependent upon whether the reference voltage Vref or the reference voltage V~ f/16 respectively, is coupled ! 15 to the second common line of the capacitor array as a second sequence of a standard three sequence successive approximation algorithm for segment determination. (5) The control logic means causes the fourth switching means to superimpose a reference 20 uoltage equivalent to the top of segments 7 or 5, dependent on the polarity of the comparator output, to the com~only coupled plates ~ the capacitor array, for the predetermined reference voltage Vref, coupled to the second common line of the capacitor array, by 25 connecting the lower 7 or 5 value capacitors to the second common line as a third sequence of a standard three sequence successive approximation algorithm for segment determination. Alternatively, the control logic means causes the fourth switching means to super-~248~;8 impose a reference voltage equivalent to the top ofsegments 3 or 1, dependent on the polarity of the comparator output, to the commonly coupled plate~ o the capacitor array, for the predetermined reference voltage Vref~l6, coupled to the second common line of the capacitor array, by connecting the lower 7 or 5 value capacitors to the second common line as a t~d sequence of a standard three sequence successive approximation algorithm for segment determination.
10 (6) The control logic means causes the fourth switching means to superimpose a reference voltage to said commonly coupled plates of the capacitor array equivalent to the top of the eighth or middle step within the segment determined in the previous sequence, either by 15 retaining unchanged the highest capacitor switched to the second common line, while concurrently causing the capacitor lower in value than the highest valued capacitor connected to the second common line to be switched to the first common line, or, dependent on 20 the polarity of the comparator output at the end of said previous sequence, by causing the highest valued capacitor connected tD the second common line to be switched to the first common line while concurrently connecting the next highest capacitor to ~he second 25 common line, said swit~hing of the fourth switching means being an initial sequence of a stan~ard four sequence successive approximation algorithm for step determination. (7) Depending upon the polarity of the output of the comparator at the end of the previous 30 sequence5 the control logic means either causes the fourth switching means to supPrimpose a re~erence 112~868 voltage equivalent to the top of step 4 of the prede~er-mined segment to the commonly coupled plates ~ the capacitor array by switching the next two highest capacitors below the highest capacitor switched to 5 the second common line in the previous sequence to the first common line, or the control logic means causes the fourth switching means to superimpose a reference voltage equivalent to the top o~ step 12 of the pre-determined segment to the commonly coupled plates of lOthe capacitor array by switching the nex~ highest capacitor below the highest capacitor switched to the second common line in the previous sequence to the second common line~ concurrently switching the second h ~hest capacitor below the highest capacitor switched lSto the second common line in the previous sequence to ; the first common line as a second sequence of the standard four sequence successive approximation algorithm : for step determination, The algorithm for the ~our sequence step determination provides the relationship 20between the number of the step n wi~ in the predeter-mined segment whose reference voltage equivalent is switched to the commonly coupled plates o~ the capacitor array to the switching state of the four capacitors below the highest value capacitor switched to the second 25Common line in the final segment determination sequence and is defined as:
1124~68 n = 16( 1 ~ 1 ~ 2~ ~ 23 24 ) When ~a = kl, ~2~ ~3~ k4 = 1, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the first common line. When ka = kl, k2, k3, ~4 = 0, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the second common line.
(8) The control logic means causes the fourth switching means to superimpose a re~erence voltage equivalent to the top of steps 2, 6, 10 or 14 of the predetermined segment to the commonly coupled plates of the capacitor array as a third sequence of the standard four sequence successive approximation algorithm for step determination as defined in the previous section (7). (9) The control logic means causes the fourth switching means to superimpose a reference voltage equivalent to the top of steps 1, 3, 5, 7, 9, 13 or 15 of the pre-determined segment to the commonly coupled plates of the capacitor array as a fourth and final sequence of the standard four sequence successive approximation algorithm for st,ep determination as def~ned in the previous section (7). (10) The control logic means causes the circuit to present an eight-bit digitally encoded output word and to reset for subsequent pulse code modulation encoding.
In accordance with a preferred embodiment of this invention, an improved sample and hold circuit provides or the same polarity of sampled signal, irrespective of the polarity of the input signal, thus obviating the need for reference voltage supplies of both polarities and the need for close voltage matching between these supplies. In accordance with such embodiment, th~ circuit includes a fifth switching means for selectively connecting and dis-connecting the recei~-ing means with the commonly coupled plates, th~ control logic means controlling the fifth switching means.
ltZ4136S
BRIEF DESCRIPTION OF THE DRAWINGS
Other objects, advantages, and features of this invention, together with its construction and mode of operation, will become more apparent from the following description when read in conjunction with the accompanying drawings, in which:
FIG. 1 is a diagram illustrating a 15 segment piecewise linear approximation to the ~255 compression law;
FIG. 2 is a diagram of a conceptual achematic of a charge redistribution of ~255 encoder of the prior art;
FIG. 3 is a simplified diagram of a charge redistribution encoding system;
FIG. 4 is a diagram of a charge redistribution circuit for PCM encoding in accordance with one embodiment of the invention;
FIG. 5 is a diagram of a charge redistribution circuit for PCM encoding in accordance with a ~ preferred embodiment of the invention utilizing a modified sample and hold circuit; and FIG. 6 is a diagram illustrating the timing of some switches of the circuit shown in FIG. 5.
PRIOR ART PCM CODER
A proposed design for a standard 15 segment appro~imation to the ~255 compression law coder for pulse-code modulation telephony was presented by Paul R. Gray, David A. Hodges, John P. Tsividis and Jacob Chacko, Jr., in an article entltled "Companded Pulse-Code Modulation Voice Codec Using Monolithic Weighted Capacitor Arrays", in the IEEE 30urnal of Solid State Circuits, December, 1975, pp. 497-499.
The fifteen seg~ent approximation of the ~255 compression is shown in FIG. 1. Note that there are seven complete segments in the upper right quadrant and seven complete segments in the lower left q~adrant, with one segment going through the origin.
The proposed coder of Gray et al., diagrammed in FIG. 2, utilizes two binary-weighted capacitor arrays, C i and C i as the key precision elements.
Other element5 shown in FIG. 2 include a binary comparator 10, a near unity gain buffer amplifier 11 with very high i~put impedance, reference sources ~VR and -VR and numerous analog switches. Binary sequencing and control logic totalling about sixty gates and flip flops, and a digital clock (not shown), are required.
Referring to FI&. 2, the comparator 10 has a pair of input lines 12, 13 coupled thereto and an output line 14. The input line 13 is coupled to a point of reference potential, such as ground.
.. . . . . .. . . , . ~ .
112~868 The input line 12, which carries a signal Vx, is coupled to an arm of a single-pole double-throw switch SXO, one position of the arm being such as to couple the line 12 to a point o~ reference S potential such as ground, the other position thereof being such as to keep the line 12 from being coupled to such reerence potential.
The line 12 is commonly coupled to one plate of each of a plurality of capacitors Cxl, Cx2, Cx3 CX4~ CX5, Cx6~ CX7 and Cx8 The remaining plate~ of the capacitors Cxl, Cx2, X3 X4~ X5~ Cx6~ Cx7~ Cx8 are coupled, respectively to arms of three-position swit_hes Sxl, Sx2, S
X4~ X5 X6~ SX7~ Sx8~ respectively. A first contact position for each o~ the switches Sxl, Sx2, X3' SX4~ Sx5~ SX6' Sx7~ and Sx8 are coupled jointly to each other to a line 16 upon which a s ~nal Vz is impressed, which is coupled to the output of a near-unity gain buffer 11.
Intermediate contacts for each of the switches SXl' SX2' SX3~ SX4~ Sxs, Sx6~ SX7, Sx8 are jointly coupled, respectively, to a common line 17 which is connected to an arm 18 of SxR which can be selectively positioned between a line 19 which is coupled to a voltage source ~VR~ and to a line 21 which is coupled to voltage source -Vg.
.. . . . .
1124l~68 The third contacts for each of the switches Sxl~ Sx2~ SX3, SX4~ SX5~ SX6~ Sx7~ SX8 are jointly connected together, respectiv~, to a line 22 which is coupled to an arm of a single-pole double-throw S switch Sxx, which can be switched between a point of reference potential, such as ground and in input voltage VI~.
An arm of a single-pole double-throw switch Syo is coupled to a point of reference potential, such as ground. One position of the switch Syo is unconnected; the remaining position is connected to a line 23 which is coupled to the input of the near-unity gain buffer 11.
The line 23 is coupled jointly to one plate of each capacitor Cyl, Cy2, Cy3, Cy4 Cy5~ The other plates of the capacitors Cyl, Cy2, Cy3~ Cy4, Cy5 are coupled, respectively, to an arm of single-pole double-throw switches Syl Sy2, S~3, Sy4~Sy5~ One contact of each of the switches Syl, Sy2, Sy3~ Sy4 and Sy5 are jointly coupled together to a line 24 which is coupled to an arm of a single-pole double-throw switch SyR. The arm of the switch SyR can be selectively coupled between a voltage source +VR
and a voltage source -VR.
112~
The other contact of each of the switches Syl Sy2, Sy3~ Sy4~ Sy5 are coupled together to a common line 26 which is coupled commonly to a point of reference potential, such as ground.
The conversion proceeds as follows: Initially, the switch SxO is thrown to ground, and the switches S~l through Sx8 are positioned to receive the voltage source VIN Ithrough the switch Sxx). The voltage VIN
is thus stored an the capacitors Cxl thr~ugh Cx8.
10 ~ext, the switch S~O is opened, and the switches Sxl through Sx8 are thrown to ground (through the switch SxX). Ideally, the sign of the voltage Yx~ as detected by the comparator 10, is the negative of the sign of VIN. This determines the sign bit of the PCM code.
15 Also, the process just described, in effect, samples the signal, and the capacitor array is used as a holding capacitor. Next, the segment is determined by successively throwing Sxi in order, starting with Sx~, from grDund to the reference voltage until the sign 20 of Vx changes. (These switches are thrown to ~VR if the sign of VIN was previously determined to be positive, and to -V~ if it was determined to be negative). If, ~or example, the comparator output changes after the switch Sxn is thrown, this would indicate that the input 25 voltage lies within the nth segment (counting from the right) indicated in FIG. ~.
1~24~368 The final stage in the conversion is the determin-ation of the step within the segment. The capacitor array Cyl is used in this process. Initially, the switches Syl are all thrown to ground; that switch Sxi which was last thrown in the previous step is now switched to the voltage Vz.
Next, the switch Sy is opened and the switches Syl through Sy5 are throw~ sequentially from ground to VR. The successive approximation algorithm is used 10 or the in-chord coding, so that four clock periods are required to decide within which of the sixteen steps the sample lies. The setting of the switch S
at this point denotes the step within the segment and the conversion is complete.
15 DESCRIPTION OF ONE EMBODIMENT OF THE INV~TION
IN GENERAL
Applicant's new approach to encoding by charge redistribution techniques is set forth ~h ich uses only a single capacitor bank for both segment and step 20 selection and is capable of handling a number of channel inputs on a time-shared basis. The design of the single capacitor bank approach in a large scale integrated circuit implementation is consistent with multi-channel encoding speed requirements.
1~2~S68 INTRODUCTI~
Single channel charge redistribution encoder/
decoder techniques, using the storage and switching of charge on capacitors, have been described for companded PCM encoding use by P. ~. Gray, et al.
as set forth hereinabove. Such techniques use binary-weighted precision capacitor arrays which can readily be implemented in NM06 technology and realized in LSI
monolithic form. For single channel application, 10 encoding speed is generally not critical since a ull 125 microseconds is available. However, or multi-channel application (~hared codec), encoding speed is critical and attention should be given to the influence of each circuit element on the total encoang delay.
Thus, it is an object of this invention to provide new and improved circuitry which achieves faster encoding and provides for a new charge redistri-bution coniguration for encoding a number of channels on a time-shared basis.
1~2~68 MULTI-CHANNEL ENCODING
Encoding speed is a function of the number of switching sequences required to encode a sample. For PCM applications, voice is often coded in a nonlinear or companded code ~o maximize the overall signal-to-noise performance. The companded law most generally used in the U.S.A. is a fifteen segment approximation of the ~255 compression law. Each segment contains sixteen equal-size steps. For PCM encoding purposes, 10 the required switching sequences or time slots per sample may be considered as:
(a) Sample and Hold One time slot.
(b) Polarity determina- One time slot.
tion 1~ (c) Segment selection Three time slots.
(d) Step selection Four time slots.
(e) Output data strobe One time slot.
and general circuit re-set 20Total: ten time slots.
In the charge redistribution kechnique~ encoding circuit delay and therefore multi-channel capability is a function of the overall loop delay of the circuit shown in simplified form in FIG. 3.
11248~i8 As shown in FIGo 3, the input signal VIN to be sampled is coupled through a sample and hold circuit 41 to one input of a capacitor network 42. Reference voltage generation circuitry 43 is coupled to a second S input to the capacitor network 42. The output of the capacitor network 42 is coupled to one input of a comparator 44, the other input thereof being coupled to a point of reference potential, such as grDund.
The comparator 44 is clocked by an appropriate clock 10 signal 46. The output of the comparator 44 is coupled along a line 47 to control logic circuitry 48 which provides control signals to the reference voltage generation means 43, and also provides 20 gignals along output lines 49.
As shown in FIG. 3, the input signal VIN is sampled by the means 41 and is subtracted from successively deriv~d reference voltages. The difference is compared to a nominal ground reference potential.
25~he delay elements in the loop shown in FIGo 3 are:
(a) Charge redistribution (CR) encoding delay (this includes reference voltage generation and capacitor switch resîstance circuit delay)O
(b) Comparator circuit delay, ~cj Control logic circuit delay.
. . . , ..... ~ ~, ... .. ..
112413~;8 It is desirable in order to achieve the minimum encoding delay that the elements of the encoding loop be kept as simple as possible. Overall allowable delay for an eight channel analog to digital conversion is approximately 15.5 microseconds per sample.
Individual segment and step sequences must therefore be complete within 1.55 micorseconds, assumming the encoding sequences as discussed above.
CR ENCODING (REFERENCE VOLTAGE G~NERATION) Analog to digital converters, described above, use binary weighted capacitor arrays and a charge redistribution technique. The approach involves a sample and hold sequence followed by a succession of comparisons of the sampled signal with switched reference 15 voltages representing the boundary levels of the segments and the steps within the segment of the fiteen segment ~255 companded law. Much of the circuitry and associated delay of the encoder is in fact involved in generating these reference voltages. In the encoder ~O to be described in detail hereinafter, the comparison voltages are derived by capacitive division using a single binary-weighted capacitor array for the generation of both the segment and step boundaries.
Compared with earlier approaches, this technique avoids 25 the use of circuit~y which would otherwise add to the delay and canplexity of the encoder.
REFERENCE VOLTAGE REQUIREMENTS
Segment and step upper boundary level require-ments can be derived in terms of step units normalized to 8159 as suggested by H. H. Henning and J. W. Pan in an article entitled l'D2 Channel Bank: System Aspectsl', BSTJ~ pp- 1641-16~7 (October 1972). If Lk n is defined as the height of the [16-n]th step within the kth segment, then Lk,n [24 2~ -2 n Note that for 1 c k c 8 and 1 < n < 15~ the series function will start with the term 2k~4 for all values of k. Also, the step correction term can be expressed as: k k+3 + 2~2 q ~ 2k+l r ~ 2k s 15 where:
Pi' qi' ri' Si = , 1 This implies that step correction can be achieved for 5 c k c 8 by using combinations of the four terms of the series function below the highest term 2 20 Furthermore, step correction in this manner requires a minimum of five terms (i-e-, 5 < k < 8). For segments 1, 2, 3, 4, it can be readily shown that:
Lk O a,n + 31 = 2 [Lk = a ~ 4 nfl] where 1 < a < 4.
25 For ~xample, L + 31 = 2-4 [L ~1 1124fl68 This allows step correction to be obtained for segments 1, 2, 3, and 4 using series functions equivalent to segments 5, 6, 7, 8 and the step correction process described above. In this case, 5 however, a constant multiplier (2 4) and correctiD n term (+31) are required.
C II~CU IT IMPLEMENTAT ION
In a practical circuit implementation of this encoding approach, the segment and step boundary voltage levels Vk n can also be expressed in terms of a maximum reference vol~eage level VREE equivalent to the top of segment 8 such that:
Vk n Z k. n VRlEF
. k= o,n = o Or, approx imat ely, [(24 ~ 2 ) - 2 nl V = V E~ -- 1 J for k -L24~2i~ 5,6,70r8 V = V [(24 2i) 2k ~ 4~ ~ V (~) 16 [24 ~ 2 for k = 1~ 2, 3, or 4 The wors e cas e error introduced by the f oregoing approximation is 3% of the individual step heights within 30 the 5th and 1st segments and correspondingly less for other segments. These errors are well within the practica]
accuracy of the circuit~
1~2 ~8S8 FIG. 4 illustrates a practical circuit of an encoder using a single capacitor ladder array for both segment and step selection in this manner.
The capacitor array CAl, CA2, CA3~ CA4~ CA5~ CA6' CA7, CA8 includes eight capacitors which are binary weighted in the order 1, 2, 4, 8, 16, 32, 64, 128.
One plate of each capacitor CAl through CA8, respectively, is commonly coupled together to a line 51. The line 51 is coupled to one input of a comparator 52, the other input being coupled through a capacitor 53 to an arm of a single-pole double-throw switch S3. The line 51 is coupled to the arm of a single-pole double-throw switch S2.
The other plates of the capacitor array C
through CA8, respectively are c~ pled to arms, respectively, of single-pole double-throw switches ~1~ SA2~ SA3~ SA4~ SA5~ S.j,,6, SA7~ SA8-One contact, respectively, of all the switches SAl through SA8 are jointly coupled together to the arm of a single-pole double-throw switch Sl.
The other contacts of the switches SAl through SA8, respectively, are jointly coupled together to the arm of a single-pole double-throw switch S4.
: The switch Sl can be selectively coupled between ~5 a point of reference potential, such as ground, and a voltage source VIN (input analog signal).
The switch S2 can be selectively switched between a point o~ re~erence potentialJ such as ground, and - an open circuit.
1~24868 The switch S3 can be selectively switched between a point of reference potential and a voltage source Vl.
The switch S4 can selectively be switched between a reference voltage V~EF and a reference voltage VREF/16.
The output of the comparator 52 is coupled along a line 56 to control logic means 57.
A read signal, along the line 58, is coupled to means 59 which provides a digital output therefrom.
The means 59 and the control logic means 57 are inter-connected. The control logic means 57 also provides signals which controls the various switches Sl, S2, S3, S4 and the array switches SAl through SA8, respectively.
As stated hereinabove, the capacitors of the capacitor array CAl through CA8 are binary weighted as indicated. By switching the lower plates (see FIG. 4) of selected capacitors to VREF, for example, differing comparison voltages are derived at the input to the comparator 52 by capacitive division. Thus, if Cs is the total capacity switched to VREF and CT~T
is the total array capacity L~ ,~ n ~ CS
112~8~;8 -~6 -Test results of a circuit such as that shown in FIG. 4, using discrete devices, have demonstrated that the new encoding technique provides accurate segment and step selection for companded PCM encoding at higher encoding rates.
The sample and hold and polarity determination modes (modes 1 and 2) proceed as reported in the prior art, such as, for example, by Gray et al., supra.
Mode 3 now involves the derivation of the comparison voltages by connecting one or more of the capacitors to VREF (or VREFll6) while the remainder of the array remains connected to ground using a standard three sequence successive approximation algorithm.
In the first of this sequence, the capacitors CAl through CA4 are switched to VREF superimposing a reference voltage equivalent to the top of segment 4 to the top plates of the capacitor ban~. If the signal level lies in the upper four segments, VREF
is retained. If the signal level lies in segments 1, 2, 3j 4, VREF/16 is used for the next two segment selection sequences. In addition, a small correction b~as is applied via the switch S3.
In the fourth mode, the step is determined in a standard four sequence successive approximation algorithm similar to that used for segment selection, using the four capacitors lower in value than the highest values switched to VgEF (V~EF/16) ., ~i8 A PREFERRED EMBODrMENT
In the method of sampling the input signal as described in the preceeding sections, the capacitor array is connected between an input signal VIN and ground by switches Sl and S2 (FIG. 4). After allowing enough time for the capacitors to charge to VIN, switches Sl and S2 are reversed and a potential equal to -VIN remains on the commonly connected plates of the capacitors at the input to the comparator 52.
Since VIN can be either positive or negative, the resulting stored charge can also be either positive or negative necessitating the use of both positive and negative reference voltages for use in the successive approximation comparison proce~. Three positive and three negative reference voltages are in fact required in such encoder circuit as follows:
~ VREF + VREF/l6 Apart from generating these supplies, there is also a need to match accurately the positive and negative reference voltages to mLnimiZe distortion products.
112~68 The reference voltage generation is greatly simplified and mismatch between positive and negative - supplies obviated by using a modified sample and hold circuit as shown in FIG. 5. Three time periods are shown (FIG. 6). During time period 1, switch S5 is closed,connecting VIN to the input to the comparator 52 for polarity determination according to the principles used in the earlier embodiments. The comparator output then switches to a logic l'l" or a logic "O". This output is then used to determine the setting of the switches during period 2 depending on whether the input signal is positive or negative, as sh~wn in FIG. 6.
During period 2, the capacitor array is charged to VIN.
During period 3, the capacitor array is returned to ; 15 the state required for the standard successive approximation encoding process described in detail in above mentioned references.
CONCLUSIO~S
This invention utilizes a single capacitor bank (in lieu of multiple banks) and eliminates a buffer amplifier that had been used heretofore, in similar prior art devices, thus reducing overall circuit delay sufficiently to permit enhanced encoding speed and desired multi-channel operation.
~arious modifications can be performed by those s~illed in the art without departing from the spirit and scope of this invention as defined by the appended claims.
10 (6) The control logic means causes the fourth switching means to superimpose a reference voltage to said commonly coupled plates of the capacitor array equivalent to the top of the eighth or middle step within the segment determined in the previous sequence, either by 15 retaining unchanged the highest capacitor switched to the second common line, while concurrently causing the capacitor lower in value than the highest valued capacitor connected to the second common line to be switched to the first common line, or, dependent on 20 the polarity of the comparator output at the end of said previous sequence, by causing the highest valued capacitor connected tD the second common line to be switched to the first common line while concurrently connecting the next highest capacitor to ~he second 25 common line, said swit~hing of the fourth switching means being an initial sequence of a stan~ard four sequence successive approximation algorithm for step determination. (7) Depending upon the polarity of the output of the comparator at the end of the previous 30 sequence5 the control logic means either causes the fourth switching means to supPrimpose a re~erence 112~868 voltage equivalent to the top of step 4 of the prede~er-mined segment to the commonly coupled plates ~ the capacitor array by switching the next two highest capacitors below the highest capacitor switched to 5 the second common line in the previous sequence to the first common line, or the control logic means causes the fourth switching means to superimpose a reference voltage equivalent to the top o~ step 12 of the pre-determined segment to the commonly coupled plates of lOthe capacitor array by switching the nex~ highest capacitor below the highest capacitor switched to the second common line in the previous sequence to the second common line~ concurrently switching the second h ~hest capacitor below the highest capacitor switched lSto the second common line in the previous sequence to ; the first common line as a second sequence of the standard four sequence successive approximation algorithm : for step determination, The algorithm for the ~our sequence step determination provides the relationship 20between the number of the step n wi~ in the predeter-mined segment whose reference voltage equivalent is switched to the commonly coupled plates o~ the capacitor array to the switching state of the four capacitors below the highest value capacitor switched to the second 25Common line in the final segment determination sequence and is defined as:
1124~68 n = 16( 1 ~ 1 ~ 2~ ~ 23 24 ) When ~a = kl, ~2~ ~3~ k4 = 1, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the first common line. When ka = kl, k2, k3, ~4 = 0, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the second common line.
(8) The control logic means causes the fourth switching means to superimpose a re~erence voltage equivalent to the top of steps 2, 6, 10 or 14 of the predetermined segment to the commonly coupled plates of the capacitor array as a third sequence of the standard four sequence successive approximation algorithm for step determination as defined in the previous section (7). (9) The control logic means causes the fourth switching means to superimpose a reference voltage equivalent to the top of steps 1, 3, 5, 7, 9, 13 or 15 of the pre-determined segment to the commonly coupled plates of the capacitor array as a fourth and final sequence of the standard four sequence successive approximation algorithm for st,ep determination as def~ned in the previous section (7). (10) The control logic means causes the circuit to present an eight-bit digitally encoded output word and to reset for subsequent pulse code modulation encoding.
In accordance with a preferred embodiment of this invention, an improved sample and hold circuit provides or the same polarity of sampled signal, irrespective of the polarity of the input signal, thus obviating the need for reference voltage supplies of both polarities and the need for close voltage matching between these supplies. In accordance with such embodiment, th~ circuit includes a fifth switching means for selectively connecting and dis-connecting the recei~-ing means with the commonly coupled plates, th~ control logic means controlling the fifth switching means.
ltZ4136S
BRIEF DESCRIPTION OF THE DRAWINGS
Other objects, advantages, and features of this invention, together with its construction and mode of operation, will become more apparent from the following description when read in conjunction with the accompanying drawings, in which:
FIG. 1 is a diagram illustrating a 15 segment piecewise linear approximation to the ~255 compression law;
FIG. 2 is a diagram of a conceptual achematic of a charge redistribution of ~255 encoder of the prior art;
FIG. 3 is a simplified diagram of a charge redistribution encoding system;
FIG. 4 is a diagram of a charge redistribution circuit for PCM encoding in accordance with one embodiment of the invention;
FIG. 5 is a diagram of a charge redistribution circuit for PCM encoding in accordance with a ~ preferred embodiment of the invention utilizing a modified sample and hold circuit; and FIG. 6 is a diagram illustrating the timing of some switches of the circuit shown in FIG. 5.
PRIOR ART PCM CODER
A proposed design for a standard 15 segment appro~imation to the ~255 compression law coder for pulse-code modulation telephony was presented by Paul R. Gray, David A. Hodges, John P. Tsividis and Jacob Chacko, Jr., in an article entltled "Companded Pulse-Code Modulation Voice Codec Using Monolithic Weighted Capacitor Arrays", in the IEEE 30urnal of Solid State Circuits, December, 1975, pp. 497-499.
The fifteen seg~ent approximation of the ~255 compression is shown in FIG. 1. Note that there are seven complete segments in the upper right quadrant and seven complete segments in the lower left q~adrant, with one segment going through the origin.
The proposed coder of Gray et al., diagrammed in FIG. 2, utilizes two binary-weighted capacitor arrays, C i and C i as the key precision elements.
Other element5 shown in FIG. 2 include a binary comparator 10, a near unity gain buffer amplifier 11 with very high i~put impedance, reference sources ~VR and -VR and numerous analog switches. Binary sequencing and control logic totalling about sixty gates and flip flops, and a digital clock (not shown), are required.
Referring to FI&. 2, the comparator 10 has a pair of input lines 12, 13 coupled thereto and an output line 14. The input line 13 is coupled to a point of reference potential, such as ground.
.. . . . . .. . . , . ~ .
112~868 The input line 12, which carries a signal Vx, is coupled to an arm of a single-pole double-throw switch SXO, one position of the arm being such as to couple the line 12 to a point o~ reference S potential such as ground, the other position thereof being such as to keep the line 12 from being coupled to such reerence potential.
The line 12 is commonly coupled to one plate of each of a plurality of capacitors Cxl, Cx2, Cx3 CX4~ CX5, Cx6~ CX7 and Cx8 The remaining plate~ of the capacitors Cxl, Cx2, X3 X4~ X5~ Cx6~ Cx7~ Cx8 are coupled, respectively to arms of three-position swit_hes Sxl, Sx2, S
X4~ X5 X6~ SX7~ Sx8~ respectively. A first contact position for each o~ the switches Sxl, Sx2, X3' SX4~ Sx5~ SX6' Sx7~ and Sx8 are coupled jointly to each other to a line 16 upon which a s ~nal Vz is impressed, which is coupled to the output of a near-unity gain buffer 11.
Intermediate contacts for each of the switches SXl' SX2' SX3~ SX4~ Sxs, Sx6~ SX7, Sx8 are jointly coupled, respectively, to a common line 17 which is connected to an arm 18 of SxR which can be selectively positioned between a line 19 which is coupled to a voltage source ~VR~ and to a line 21 which is coupled to voltage source -Vg.
.. . . . .
1124l~68 The third contacts for each of the switches Sxl~ Sx2~ SX3, SX4~ SX5~ SX6~ Sx7~ SX8 are jointly connected together, respectiv~, to a line 22 which is coupled to an arm of a single-pole double-throw S switch Sxx, which can be switched between a point of reference potential, such as ground and in input voltage VI~.
An arm of a single-pole double-throw switch Syo is coupled to a point of reference potential, such as ground. One position of the switch Syo is unconnected; the remaining position is connected to a line 23 which is coupled to the input of the near-unity gain buffer 11.
The line 23 is coupled jointly to one plate of each capacitor Cyl, Cy2, Cy3, Cy4 Cy5~ The other plates of the capacitors Cyl, Cy2, Cy3~ Cy4, Cy5 are coupled, respectively, to an arm of single-pole double-throw switches Syl Sy2, S~3, Sy4~Sy5~ One contact of each of the switches Syl, Sy2, Sy3~ Sy4 and Sy5 are jointly coupled together to a line 24 which is coupled to an arm of a single-pole double-throw switch SyR. The arm of the switch SyR can be selectively coupled between a voltage source +VR
and a voltage source -VR.
112~
The other contact of each of the switches Syl Sy2, Sy3~ Sy4~ Sy5 are coupled together to a common line 26 which is coupled commonly to a point of reference potential, such as ground.
The conversion proceeds as follows: Initially, the switch SxO is thrown to ground, and the switches S~l through Sx8 are positioned to receive the voltage source VIN Ithrough the switch Sxx). The voltage VIN
is thus stored an the capacitors Cxl thr~ugh Cx8.
10 ~ext, the switch S~O is opened, and the switches Sxl through Sx8 are thrown to ground (through the switch SxX). Ideally, the sign of the voltage Yx~ as detected by the comparator 10, is the negative of the sign of VIN. This determines the sign bit of the PCM code.
15 Also, the process just described, in effect, samples the signal, and the capacitor array is used as a holding capacitor. Next, the segment is determined by successively throwing Sxi in order, starting with Sx~, from grDund to the reference voltage until the sign 20 of Vx changes. (These switches are thrown to ~VR if the sign of VIN was previously determined to be positive, and to -V~ if it was determined to be negative). If, ~or example, the comparator output changes after the switch Sxn is thrown, this would indicate that the input 25 voltage lies within the nth segment (counting from the right) indicated in FIG. ~.
1~24~368 The final stage in the conversion is the determin-ation of the step within the segment. The capacitor array Cyl is used in this process. Initially, the switches Syl are all thrown to ground; that switch Sxi which was last thrown in the previous step is now switched to the voltage Vz.
Next, the switch Sy is opened and the switches Syl through Sy5 are throw~ sequentially from ground to VR. The successive approximation algorithm is used 10 or the in-chord coding, so that four clock periods are required to decide within which of the sixteen steps the sample lies. The setting of the switch S
at this point denotes the step within the segment and the conversion is complete.
15 DESCRIPTION OF ONE EMBODIMENT OF THE INV~TION
IN GENERAL
Applicant's new approach to encoding by charge redistribution techniques is set forth ~h ich uses only a single capacitor bank for both segment and step 20 selection and is capable of handling a number of channel inputs on a time-shared basis. The design of the single capacitor bank approach in a large scale integrated circuit implementation is consistent with multi-channel encoding speed requirements.
1~2~S68 INTRODUCTI~
Single channel charge redistribution encoder/
decoder techniques, using the storage and switching of charge on capacitors, have been described for companded PCM encoding use by P. ~. Gray, et al.
as set forth hereinabove. Such techniques use binary-weighted precision capacitor arrays which can readily be implemented in NM06 technology and realized in LSI
monolithic form. For single channel application, 10 encoding speed is generally not critical since a ull 125 microseconds is available. However, or multi-channel application (~hared codec), encoding speed is critical and attention should be given to the influence of each circuit element on the total encoang delay.
Thus, it is an object of this invention to provide new and improved circuitry which achieves faster encoding and provides for a new charge redistri-bution coniguration for encoding a number of channels on a time-shared basis.
1~2~68 MULTI-CHANNEL ENCODING
Encoding speed is a function of the number of switching sequences required to encode a sample. For PCM applications, voice is often coded in a nonlinear or companded code ~o maximize the overall signal-to-noise performance. The companded law most generally used in the U.S.A. is a fifteen segment approximation of the ~255 compression law. Each segment contains sixteen equal-size steps. For PCM encoding purposes, 10 the required switching sequences or time slots per sample may be considered as:
(a) Sample and Hold One time slot.
(b) Polarity determina- One time slot.
tion 1~ (c) Segment selection Three time slots.
(d) Step selection Four time slots.
(e) Output data strobe One time slot.
and general circuit re-set 20Total: ten time slots.
In the charge redistribution kechnique~ encoding circuit delay and therefore multi-channel capability is a function of the overall loop delay of the circuit shown in simplified form in FIG. 3.
11248~i8 As shown in FIGo 3, the input signal VIN to be sampled is coupled through a sample and hold circuit 41 to one input of a capacitor network 42. Reference voltage generation circuitry 43 is coupled to a second S input to the capacitor network 42. The output of the capacitor network 42 is coupled to one input of a comparator 44, the other input thereof being coupled to a point of reference potential, such as grDund.
The comparator 44 is clocked by an appropriate clock 10 signal 46. The output of the comparator 44 is coupled along a line 47 to control logic circuitry 48 which provides control signals to the reference voltage generation means 43, and also provides 20 gignals along output lines 49.
As shown in FIG. 3, the input signal VIN is sampled by the means 41 and is subtracted from successively deriv~d reference voltages. The difference is compared to a nominal ground reference potential.
25~he delay elements in the loop shown in FIGo 3 are:
(a) Charge redistribution (CR) encoding delay (this includes reference voltage generation and capacitor switch resîstance circuit delay)O
(b) Comparator circuit delay, ~cj Control logic circuit delay.
. . . , ..... ~ ~, ... .. ..
112413~;8 It is desirable in order to achieve the minimum encoding delay that the elements of the encoding loop be kept as simple as possible. Overall allowable delay for an eight channel analog to digital conversion is approximately 15.5 microseconds per sample.
Individual segment and step sequences must therefore be complete within 1.55 micorseconds, assumming the encoding sequences as discussed above.
CR ENCODING (REFERENCE VOLTAGE G~NERATION) Analog to digital converters, described above, use binary weighted capacitor arrays and a charge redistribution technique. The approach involves a sample and hold sequence followed by a succession of comparisons of the sampled signal with switched reference 15 voltages representing the boundary levels of the segments and the steps within the segment of the fiteen segment ~255 companded law. Much of the circuitry and associated delay of the encoder is in fact involved in generating these reference voltages. In the encoder ~O to be described in detail hereinafter, the comparison voltages are derived by capacitive division using a single binary-weighted capacitor array for the generation of both the segment and step boundaries.
Compared with earlier approaches, this technique avoids 25 the use of circuit~y which would otherwise add to the delay and canplexity of the encoder.
REFERENCE VOLTAGE REQUIREMENTS
Segment and step upper boundary level require-ments can be derived in terms of step units normalized to 8159 as suggested by H. H. Henning and J. W. Pan in an article entitled l'D2 Channel Bank: System Aspectsl', BSTJ~ pp- 1641-16~7 (October 1972). If Lk n is defined as the height of the [16-n]th step within the kth segment, then Lk,n [24 2~ -2 n Note that for 1 c k c 8 and 1 < n < 15~ the series function will start with the term 2k~4 for all values of k. Also, the step correction term can be expressed as: k k+3 + 2~2 q ~ 2k+l r ~ 2k s 15 where:
Pi' qi' ri' Si = , 1 This implies that step correction can be achieved for 5 c k c 8 by using combinations of the four terms of the series function below the highest term 2 20 Furthermore, step correction in this manner requires a minimum of five terms (i-e-, 5 < k < 8). For segments 1, 2, 3, 4, it can be readily shown that:
Lk O a,n + 31 = 2 [Lk = a ~ 4 nfl] where 1 < a < 4.
25 For ~xample, L + 31 = 2-4 [L ~1 1124fl68 This allows step correction to be obtained for segments 1, 2, 3, and 4 using series functions equivalent to segments 5, 6, 7, 8 and the step correction process described above. In this case, 5 however, a constant multiplier (2 4) and correctiD n term (+31) are required.
C II~CU IT IMPLEMENTAT ION
In a practical circuit implementation of this encoding approach, the segment and step boundary voltage levels Vk n can also be expressed in terms of a maximum reference vol~eage level VREE equivalent to the top of segment 8 such that:
Vk n Z k. n VRlEF
. k= o,n = o Or, approx imat ely, [(24 ~ 2 ) - 2 nl V = V E~ -- 1 J for k -L24~2i~ 5,6,70r8 V = V [(24 2i) 2k ~ 4~ ~ V (~) 16 [24 ~ 2 for k = 1~ 2, 3, or 4 The wors e cas e error introduced by the f oregoing approximation is 3% of the individual step heights within 30 the 5th and 1st segments and correspondingly less for other segments. These errors are well within the practica]
accuracy of the circuit~
1~2 ~8S8 FIG. 4 illustrates a practical circuit of an encoder using a single capacitor ladder array for both segment and step selection in this manner.
The capacitor array CAl, CA2, CA3~ CA4~ CA5~ CA6' CA7, CA8 includes eight capacitors which are binary weighted in the order 1, 2, 4, 8, 16, 32, 64, 128.
One plate of each capacitor CAl through CA8, respectively, is commonly coupled together to a line 51. The line 51 is coupled to one input of a comparator 52, the other input being coupled through a capacitor 53 to an arm of a single-pole double-throw switch S3. The line 51 is coupled to the arm of a single-pole double-throw switch S2.
The other plates of the capacitor array C
through CA8, respectively are c~ pled to arms, respectively, of single-pole double-throw switches ~1~ SA2~ SA3~ SA4~ SA5~ S.j,,6, SA7~ SA8-One contact, respectively, of all the switches SAl through SA8 are jointly coupled together to the arm of a single-pole double-throw switch Sl.
The other contacts of the switches SAl through SA8, respectively, are jointly coupled together to the arm of a single-pole double-throw switch S4.
: The switch Sl can be selectively coupled between ~5 a point of reference potential, such as ground, and a voltage source VIN (input analog signal).
The switch S2 can be selectively switched between a point o~ re~erence potentialJ such as ground, and - an open circuit.
1~24868 The switch S3 can be selectively switched between a point of reference potential and a voltage source Vl.
The switch S4 can selectively be switched between a reference voltage V~EF and a reference voltage VREF/16.
The output of the comparator 52 is coupled along a line 56 to control logic means 57.
A read signal, along the line 58, is coupled to means 59 which provides a digital output therefrom.
The means 59 and the control logic means 57 are inter-connected. The control logic means 57 also provides signals which controls the various switches Sl, S2, S3, S4 and the array switches SAl through SA8, respectively.
As stated hereinabove, the capacitors of the capacitor array CAl through CA8 are binary weighted as indicated. By switching the lower plates (see FIG. 4) of selected capacitors to VREF, for example, differing comparison voltages are derived at the input to the comparator 52 by capacitive division. Thus, if Cs is the total capacity switched to VREF and CT~T
is the total array capacity L~ ,~ n ~ CS
112~8~;8 -~6 -Test results of a circuit such as that shown in FIG. 4, using discrete devices, have demonstrated that the new encoding technique provides accurate segment and step selection for companded PCM encoding at higher encoding rates.
The sample and hold and polarity determination modes (modes 1 and 2) proceed as reported in the prior art, such as, for example, by Gray et al., supra.
Mode 3 now involves the derivation of the comparison voltages by connecting one or more of the capacitors to VREF (or VREFll6) while the remainder of the array remains connected to ground using a standard three sequence successive approximation algorithm.
In the first of this sequence, the capacitors CAl through CA4 are switched to VREF superimposing a reference voltage equivalent to the top of segment 4 to the top plates of the capacitor ban~. If the signal level lies in the upper four segments, VREF
is retained. If the signal level lies in segments 1, 2, 3j 4, VREF/16 is used for the next two segment selection sequences. In addition, a small correction b~as is applied via the switch S3.
In the fourth mode, the step is determined in a standard four sequence successive approximation algorithm similar to that used for segment selection, using the four capacitors lower in value than the highest values switched to VgEF (V~EF/16) ., ~i8 A PREFERRED EMBODrMENT
In the method of sampling the input signal as described in the preceeding sections, the capacitor array is connected between an input signal VIN and ground by switches Sl and S2 (FIG. 4). After allowing enough time for the capacitors to charge to VIN, switches Sl and S2 are reversed and a potential equal to -VIN remains on the commonly connected plates of the capacitors at the input to the comparator 52.
Since VIN can be either positive or negative, the resulting stored charge can also be either positive or negative necessitating the use of both positive and negative reference voltages for use in the successive approximation comparison proce~. Three positive and three negative reference voltages are in fact required in such encoder circuit as follows:
~ VREF + VREF/l6 Apart from generating these supplies, there is also a need to match accurately the positive and negative reference voltages to mLnimiZe distortion products.
112~68 The reference voltage generation is greatly simplified and mismatch between positive and negative - supplies obviated by using a modified sample and hold circuit as shown in FIG. 5. Three time periods are shown (FIG. 6). During time period 1, switch S5 is closed,connecting VIN to the input to the comparator 52 for polarity determination according to the principles used in the earlier embodiments. The comparator output then switches to a logic l'l" or a logic "O". This output is then used to determine the setting of the switches during period 2 depending on whether the input signal is positive or negative, as sh~wn in FIG. 6.
During period 2, the capacitor array is charged to VIN.
During period 3, the capacitor array is returned to ; 15 the state required for the standard successive approximation encoding process described in detail in above mentioned references.
CONCLUSIO~S
This invention utilizes a single capacitor bank (in lieu of multiple banks) and eliminates a buffer amplifier that had been used heretofore, in similar prior art devices, thus reducing overall circuit delay sufficiently to permit enhanced encoding speed and desired multi-channel operation.
~arious modifications can be performed by those s~illed in the art without departing from the spirit and scope of this invention as defined by the appended claims.
Claims (9)
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:
1. A charge redistribution circuit for pulse-code modulation encoding comprising:
a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded, each capacitor of said array having a pair of opposed plates, one plate of each of said capacitors being commonly coupled together;
a binary comparator having a pair of inputs and an output, a first of said inputs being adapted to be coupled to a reference bias potential, and a second of said inputs being coupled to said commonly coupled plates of said capacitor array;
means for receiving an input analog voice signal;
first switching means for connecting said commonly coupled plates to ground potential and for disconnecting said plates from ground potential;
a first common line;
a second common line;
second switching means for selectively connecting said first common line to said receiving means or to ground potential;
third switching means for selectively connecting said second common line to a predetermined reference voltage VREF, or to a reference voltage VREF/n which is one nth said predetermined voltage, and for selectively coupling said first comparator input to ground potential or to said reference bias potential;
fourth switching means, including a switch for each said capacitor, for selectively coupling the other plates of each of said capacitors of said array to said first common line, or to said second common line;
code generating means for providing a digital output upon receipt of a read instruction and control signals applied thereto; and control logic means for controlling said four switching means, responsive to the output of said comparator, and for providing control signals to said generating means, wherein n is a positive integer.
a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded, each capacitor of said array having a pair of opposed plates, one plate of each of said capacitors being commonly coupled together;
a binary comparator having a pair of inputs and an output, a first of said inputs being adapted to be coupled to a reference bias potential, and a second of said inputs being coupled to said commonly coupled plates of said capacitor array;
means for receiving an input analog voice signal;
first switching means for connecting said commonly coupled plates to ground potential and for disconnecting said plates from ground potential;
a first common line;
a second common line;
second switching means for selectively connecting said first common line to said receiving means or to ground potential;
third switching means for selectively connecting said second common line to a predetermined reference voltage VREF, or to a reference voltage VREF/n which is one nth said predetermined voltage, and for selectively coupling said first comparator input to ground potential or to said reference bias potential;
fourth switching means, including a switch for each said capacitor, for selectively coupling the other plates of each of said capacitors of said array to said first common line, or to said second common line;
code generating means for providing a digital output upon receipt of a read instruction and control signals applied thereto; and control logic means for controlling said four switching means, responsive to the output of said comparator, and for providing control signals to said generating means, wherein n is a positive integer.
2. The. circuit as recited in claim 1 further comprising fifth switching means for selectively connecting and disconnecting said receiving means with said commonly coupled plates, and wherein said control logic means is adapted to control said fifth switching means.
3. The circuit as recited in claim 2 wherein said control logic mean, initially, in a sample and hold determination, in sequence (1) for a first time period, causes said first switching means to disconnect said commonly coupled plates from ground potential, causes said second switching means to couple said first common line to said receiving means, causes said fourth switching means to couple said other plates to said first common line, and causes said fifth switching means to connect said commonly coupled plates to said receiving means, thus providing for a comparator output in either of two different logic states;
(2) for a second time period, with a first logic state provided in (1) above, causes said first switching means to connect said commonly coupled plates to ground potential, causes said second switching means to continue to connect said first common line to said receiving means, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means; and with a second logic state provided in (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential causes said second switching means to couple said first common line to ground potential, and causes said fifth switching means to continue to connect said commonly coupled plates to said receiving means, thus charging said capacitor array;
and (3) for a third time period, with said first logic state provided in (1) above, causes said first switching means to disconnect said commonly coupled plates from ground potential, causes said second switching means to connect said first common line to ground potential, and causes said fifth switching means to continue to disconnect said commonly coupled plates from said receiving means, and, with said second logic state provided as (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential, causes said second switching means to continue to couple said first common line to ground potential, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means.
(2) for a second time period, with a first logic state provided in (1) above, causes said first switching means to connect said commonly coupled plates to ground potential, causes said second switching means to continue to connect said first common line to said receiving means, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means; and with a second logic state provided in (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential causes said second switching means to couple said first common line to ground potential, and causes said fifth switching means to continue to connect said commonly coupled plates to said receiving means, thus charging said capacitor array;
and (3) for a third time period, with said first logic state provided in (1) above, causes said first switching means to disconnect said commonly coupled plates from ground potential, causes said second switching means to connect said first common line to ground potential, and causes said fifth switching means to continue to disconnect said commonly coupled plates from said receiving means, and, with said second logic state provided as (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential, causes said second switching means to continue to couple said first common line to ground potential, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means.
4. The circuit as recited in claim 1 wherein n equals 2m, wherein m is a positive integer.
5. The circuit as recited in claim 1 wherein n equals 16.
6. In apparatus for encoding voice in a companded code to maximize signal-to-noise performance, the companded code being the 15 segment approximation of the µ255 compression law, each segment containing 16 equal size steps, a charge redistribution circuit for pulse-code modulation encoding comprising:
a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded, each capacitor of said array having a pair of opposed plates, one plate of each of said capacitors being commonly coupled together;
a binary comparator having a pair of inputs and an output, a first of said inputs being adapted to be coupled to a reference bias potential, and a second of said inputs being coupled to said commonly coupled plates of said capacitor array;
means for receiving an input analog voice signal;
first switching means for connecting said commonly coupled plates to ground potential and for disconnecting said plates from ground potential;
a first common line;
a second common line;
second switching means for selectively connecting said first common line to said receiving means or to ground potential;
third switching means for selectively connecting said second common line to a predetermined reference voltage VREF, or to a reference voltage VREF/16 which is one sixteenth said predetermined voltage, and for selectively coupling said first comparator input to ground potential or to said reference bias potential;
fourth switching means, including a switch for each said capacitor, for selectively coupling the other plates of each of said capacitors of said array to said first common line, or to said second common line;
eight bit code generating means for providing a digital 8 bit output upon receipt of a read instruction and control signals applied thereto; and control logic means for controlling said four switching means, responsive to the output of said comparator, and for providing control signals to said generating means.
a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded, each capacitor of said array having a pair of opposed plates, one plate of each of said capacitors being commonly coupled together;
a binary comparator having a pair of inputs and an output, a first of said inputs being adapted to be coupled to a reference bias potential, and a second of said inputs being coupled to said commonly coupled plates of said capacitor array;
means for receiving an input analog voice signal;
first switching means for connecting said commonly coupled plates to ground potential and for disconnecting said plates from ground potential;
a first common line;
a second common line;
second switching means for selectively connecting said first common line to said receiving means or to ground potential;
third switching means for selectively connecting said second common line to a predetermined reference voltage VREF, or to a reference voltage VREF/16 which is one sixteenth said predetermined voltage, and for selectively coupling said first comparator input to ground potential or to said reference bias potential;
fourth switching means, including a switch for each said capacitor, for selectively coupling the other plates of each of said capacitors of said array to said first common line, or to said second common line;
eight bit code generating means for providing a digital 8 bit output upon receipt of a read instruction and control signals applied thereto; and control logic means for controlling said four switching means, responsive to the output of said comparator, and for providing control signals to said generating means.
7. In apparatus for encoding voice in a companded code to maximize signal-to-noise performance, the companded code being the 15 segment approximation of the µ255 compression law, each segment containing 16 equal size steps, a charge redistribution circuit for pulse-code modulation encoding comprising:
a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded, each capacitor of said array having a pair of opposed plates, one plate of each of said capacitors being commonly coupled together;
a binary comparator having a pair of inputs and an output, a first of said inputs being adapted to be coupled to a reference bias potential, and a second of said inputs being coupled to said commonly coupled plates of said capacitor array;
means for receiving an input analog voice signal;
first switching means for connecting said commonly coupled plates to ground potential and for disconnecting said plates from ground potential;
a first common line;
a second common line;
second switching means for selectively connecting said first common line to said receiving means or to ground potential;
third switching means for selectively connecting said second common line to a predetermined reference voltage VREF, or to a reference voltage VREF/16 which is one-sixteenth said predetermined voltage, and for selectively coupling said first comparator input to ground potential or to said reference bias potential;
fourth switching means, including a switch for each said capacitor, for selectively coupling the other plates of each said capacitors of said array to said first common line, or to said second common line;
eight bit code generating means for providing a digital 8 bit output upon receipt of a read instruction and control signals applied thereto; and control logic means for controlling said four switching means, responsive to the output of said comparator, and for providing control signals to said generating means, said control logic means, in sequence, (1) causing said first switching means to connect said commonly coupled plates to ground potential, causing said fourth switching means to couple said other plates to said first common line, and causing said second switching means to connect said first common line to said receiving means, whereby said analog voice signal can be sampled and held as a charge stored on said capacitor array;
(2) causing said first switching means to dis-connect said commonly connected plates from ground potential, and causing said second switching means to connect said first common line to ground potential, whereby the sign of the voltage on said second input of said comparator, as detected by said comparator, is the negative of the sign of said sampled analog voice signal, thereby determining the polarity and sign bit of the pulse-code-modulation;
(3) causing said third switching means to connect said second common line to said predetermined reference voltage VREF, causing said fourth switching means to superimpose a reference voltage equivalent to the top of segment 4 to the commonly coupled plates of said capacitor array by connecting the four lower value capacitors to said second common line as an initial sequence of a standard three sequence segment selection successive approximation algorithm while the remainder of the array remains coupled to ground;
(4) causing said third switching means to continue to connect said second common line to said predetermined reference voltage VREF when the sampled input signal lies in the upper four segments, and for causing said third switching means to connect said second common line to said reference voltage VREF/16 when the sampled input signal lies in the lower four segments; and simultaneously, causing said third switching means to connect the first comparator input to ground when the sampled input signal lies in the upper four segments;
and causing said third switching means to connect said first comparator input to a reference bias potential when the sampled input signal lies in the lower four segments; and causing fourth switching means as part of the second sequence of the standard three sequence segment-selection successive approximation algorithm to super-impose either a reference voltage equivalent to the top of segment 6 or segment 2 to the commonly coupled plates of said capacitor array, dependent upon whether the reference voltage VREF or the reference voltage VREF/16, respectively, is coupled to said second common line of said capacitor array as a second sequence of a standard three sequence successive approximation algorithm for segment determination;
(5) causing said fourth switching means either (a) to superimpose a reference voltage equivalent to the top of segments 7 or 5, dependent on the polarity of the comparator output, to the commonly coupled plates of the capacitor array, for the predetermined reference voltage VREF, coupled to said second common line of said capacitor array, by connecting the lower 7 or 5 value capacitors to the second common line as a third sequence of a standard three sequence successive approximation algorithm for segment determination, or (b) to superimpose a reference voltage equivalent to the top of segments 3 or 1, dependent on the polarity of the comparator output, to the commonly coupled plates of the capacitor array, for the predetermined reference voltage VREF/16, coupled to said second common line of said capacitor array, by connecting the lower 7 or 5 , value capacitors to the second common line as a third sequence of a standard three sequence successive approximation algorithm for segment determination;
(6) causing said fourth switching means to super-impose a reference voltage to said commonly coupled plates of said capacitor array equivalent to the top of eight or middle step within the segment determined in the previous sequence, either by retaining unchanged the highest capacitor switched to said second common line, while concurrently causing the capacitor lower in value than the highest valued capacitor connected to said second common line to be switched to said first common line, or, dependent on the polarity of the comparator output at the end of said previous sequence, by causing the highest valued capacitor connected to said second common line to be switched to said first common line while concurrently connecting the next highest capacitor to said second common line, said switching of said fourth switching means being an initial sequence of a standard four sequence successive approximation algorithm for step determination;
(j) dependent upon the polarity of said output of said comparator at the end of the previous sequence, the control logic means either causes said fourth switching means to superimpose a reference voltage equivalent to the top of step 4 of the predetermined segment to said commonly coupled plates of said capacitor array by switching the next two highest capacitors below the highest capacitor switched to said second common line in the previous sequence to said first common line, or said control logic means causes said fourth switching means to superimpose a reference voltage equivalent to the top of step 12 of the predetermined segment to said commonly coupled plates of said capacitor array by switching the next highest capacitor below the highest capacitor switched to said second common line in the previous sequence to said second common line, concurrently switching the second highest capacitor below the highest capacitor switched to said second common line in the previous sequence to said first common line as a second sequence of the standard four sequence successive approximation algorithm for step determination, the algorithm for the four sequence step determination providing the relationship between the number of the step n within the predetermined segment whose reference voltage equivalent is switched to said commonly coupled plates of said capacitor array to the switching state of the four capacitors below the highest value capacitor switched to said second common line in the final segment determination sequence and is defined as:
wherein when ka = k1, k2, k3, k4 = 1, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the first common line, and wherein when ka = k1, k2, k3, k4 = O, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the second common line (8) causing said fourth switching means to superimpose a reference voltage equivalent to the top of steps 2, 6, 10 or 14 of the predetermined segment to said commonly coupled plates of said capacitor array as a third sequence of said standard four sequence successive approximation algorithm;
(9) causing said fourth switching means to superimpose reference voltage equivalent to the top of steps 1, 3, 5, Y, 9, 11, 13 or 15 of the predetermined segment to said commonly coupled plates of said capacitor array as a fourth sequence of said standard four sequence successive approximation algorithm; and (10) causing said circuit to present an eight-bit digitally encoded output word and to reset said circuit for a subsequent pulse-code-modulation encoding.
a single binary weighted capacitor array for the generation of both segment and step boundaries of voice being encoded, each capacitor of said array having a pair of opposed plates, one plate of each of said capacitors being commonly coupled together;
a binary comparator having a pair of inputs and an output, a first of said inputs being adapted to be coupled to a reference bias potential, and a second of said inputs being coupled to said commonly coupled plates of said capacitor array;
means for receiving an input analog voice signal;
first switching means for connecting said commonly coupled plates to ground potential and for disconnecting said plates from ground potential;
a first common line;
a second common line;
second switching means for selectively connecting said first common line to said receiving means or to ground potential;
third switching means for selectively connecting said second common line to a predetermined reference voltage VREF, or to a reference voltage VREF/16 which is one-sixteenth said predetermined voltage, and for selectively coupling said first comparator input to ground potential or to said reference bias potential;
fourth switching means, including a switch for each said capacitor, for selectively coupling the other plates of each said capacitors of said array to said first common line, or to said second common line;
eight bit code generating means for providing a digital 8 bit output upon receipt of a read instruction and control signals applied thereto; and control logic means for controlling said four switching means, responsive to the output of said comparator, and for providing control signals to said generating means, said control logic means, in sequence, (1) causing said first switching means to connect said commonly coupled plates to ground potential, causing said fourth switching means to couple said other plates to said first common line, and causing said second switching means to connect said first common line to said receiving means, whereby said analog voice signal can be sampled and held as a charge stored on said capacitor array;
(2) causing said first switching means to dis-connect said commonly connected plates from ground potential, and causing said second switching means to connect said first common line to ground potential, whereby the sign of the voltage on said second input of said comparator, as detected by said comparator, is the negative of the sign of said sampled analog voice signal, thereby determining the polarity and sign bit of the pulse-code-modulation;
(3) causing said third switching means to connect said second common line to said predetermined reference voltage VREF, causing said fourth switching means to superimpose a reference voltage equivalent to the top of segment 4 to the commonly coupled plates of said capacitor array by connecting the four lower value capacitors to said second common line as an initial sequence of a standard three sequence segment selection successive approximation algorithm while the remainder of the array remains coupled to ground;
(4) causing said third switching means to continue to connect said second common line to said predetermined reference voltage VREF when the sampled input signal lies in the upper four segments, and for causing said third switching means to connect said second common line to said reference voltage VREF/16 when the sampled input signal lies in the lower four segments; and simultaneously, causing said third switching means to connect the first comparator input to ground when the sampled input signal lies in the upper four segments;
and causing said third switching means to connect said first comparator input to a reference bias potential when the sampled input signal lies in the lower four segments; and causing fourth switching means as part of the second sequence of the standard three sequence segment-selection successive approximation algorithm to super-impose either a reference voltage equivalent to the top of segment 6 or segment 2 to the commonly coupled plates of said capacitor array, dependent upon whether the reference voltage VREF or the reference voltage VREF/16, respectively, is coupled to said second common line of said capacitor array as a second sequence of a standard three sequence successive approximation algorithm for segment determination;
(5) causing said fourth switching means either (a) to superimpose a reference voltage equivalent to the top of segments 7 or 5, dependent on the polarity of the comparator output, to the commonly coupled plates of the capacitor array, for the predetermined reference voltage VREF, coupled to said second common line of said capacitor array, by connecting the lower 7 or 5 value capacitors to the second common line as a third sequence of a standard three sequence successive approximation algorithm for segment determination, or (b) to superimpose a reference voltage equivalent to the top of segments 3 or 1, dependent on the polarity of the comparator output, to the commonly coupled plates of the capacitor array, for the predetermined reference voltage VREF/16, coupled to said second common line of said capacitor array, by connecting the lower 7 or 5 , value capacitors to the second common line as a third sequence of a standard three sequence successive approximation algorithm for segment determination;
(6) causing said fourth switching means to super-impose a reference voltage to said commonly coupled plates of said capacitor array equivalent to the top of eight or middle step within the segment determined in the previous sequence, either by retaining unchanged the highest capacitor switched to said second common line, while concurrently causing the capacitor lower in value than the highest valued capacitor connected to said second common line to be switched to said first common line, or, dependent on the polarity of the comparator output at the end of said previous sequence, by causing the highest valued capacitor connected to said second common line to be switched to said first common line while concurrently connecting the next highest capacitor to said second common line, said switching of said fourth switching means being an initial sequence of a standard four sequence successive approximation algorithm for step determination;
(j) dependent upon the polarity of said output of said comparator at the end of the previous sequence, the control logic means either causes said fourth switching means to superimpose a reference voltage equivalent to the top of step 4 of the predetermined segment to said commonly coupled plates of said capacitor array by switching the next two highest capacitors below the highest capacitor switched to said second common line in the previous sequence to said first common line, or said control logic means causes said fourth switching means to superimpose a reference voltage equivalent to the top of step 12 of the predetermined segment to said commonly coupled plates of said capacitor array by switching the next highest capacitor below the highest capacitor switched to said second common line in the previous sequence to said second common line, concurrently switching the second highest capacitor below the highest capacitor switched to said second common line in the previous sequence to said first common line as a second sequence of the standard four sequence successive approximation algorithm for step determination, the algorithm for the four sequence step determination providing the relationship between the number of the step n within the predetermined segment whose reference voltage equivalent is switched to said commonly coupled plates of said capacitor array to the switching state of the four capacitors below the highest value capacitor switched to said second common line in the final segment determination sequence and is defined as:
wherein when ka = k1, k2, k3, k4 = 1, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the first common line, and wherein when ka = k1, k2, k3, k4 = O, the ath capacitor below the highest value capacitor switched to the second common line in the previous sequence is switched to the second common line (8) causing said fourth switching means to superimpose a reference voltage equivalent to the top of steps 2, 6, 10 or 14 of the predetermined segment to said commonly coupled plates of said capacitor array as a third sequence of said standard four sequence successive approximation algorithm;
(9) causing said fourth switching means to superimpose reference voltage equivalent to the top of steps 1, 3, 5, Y, 9, 11, 13 or 15 of the predetermined segment to said commonly coupled plates of said capacitor array as a fourth sequence of said standard four sequence successive approximation algorithm; and (10) causing said circuit to present an eight-bit digitally encoded output word and to reset said circuit for a subsequent pulse-code-modulation encoding.
8 The circuit as recited in claim 7 further comprising fifth switching means for selectively connecting and disconnecting said receiving means with said commonly coupled plates, and wherein said control logic means is adapted to control said fifth switching means.
9. The circuit as recited in claim 8 wherein (1) for a first time period, causes said first switching means to disconnect said commonly coupled plates from ground potential, causes said second switching means to couple said first common line to said receiving means, causes said fourth switching means to couple said other plates to said first common line, and causes said fifth switching means to connect said commonly coupled plates to said receiving means, thus providing for a comparator output in either of two different logic states;
(2) for a second time period, with a first logic state provided in (1) above, causes said first switching means to connect said commonly coupled plates to ground potential, causes said second switching means to continue to connect said first common line to said receiving means, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means; and with a second logic state provided in (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential causes said second switching means to couple said first common line to ground potential, and causes said fifth switching means to continue to connect said commonly coupled plates to said receiving means, thus charging said capacitor array, and (3) for a third time period, with said first logic state provided in (1) above, causes said first switching means to disconnect said commonly coupled plates from ground potential, causes said second switching means to connect said first common line to ground potential, and causes said fifth switching means to continue to disconnect said commonly coupled plates from said receiving means, and, with said second logic state provided as (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential, causes said second switching means to continue to couple said first common line to ground potential, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means.
(2) for a second time period, with a first logic state provided in (1) above, causes said first switching means to connect said commonly coupled plates to ground potential, causes said second switching means to continue to connect said first common line to said receiving means, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means; and with a second logic state provided in (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential causes said second switching means to couple said first common line to ground potential, and causes said fifth switching means to continue to connect said commonly coupled plates to said receiving means, thus charging said capacitor array, and (3) for a third time period, with said first logic state provided in (1) above, causes said first switching means to disconnect said commonly coupled plates from ground potential, causes said second switching means to connect said first common line to ground potential, and causes said fifth switching means to continue to disconnect said commonly coupled plates from said receiving means, and, with said second logic state provided as (1) above, causes said first switching means to continue to disconnect said commonly coupled plates from ground potential, causes said second switching means to continue to couple said first common line to ground potential, and causes said fifth switching means to disconnect said commonly coupled plates from said receiving means.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US87423978A | 1978-02-01 | 1978-02-01 | |
| US874,239 | 1978-02-01 | ||
| US963,489 | 1978-11-24 | ||
| US05/963,489 US4195282A (en) | 1978-02-01 | 1978-11-24 | Charge redistribution circuits |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA1124868A true CA1124868A (en) | 1982-06-01 |
Family
ID=27128329
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA320,239A Expired CA1124868A (en) | 1978-02-01 | 1979-01-25 | Charge redistribution circuits |
Country Status (2)
| Country | Link |
|---|---|
| US (1) | US4195282A (en) |
| CA (1) | CA1124868A (en) |
Families Citing this family (38)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| USRE32313E (en) * | 1979-09-10 | 1986-12-23 | Hitachi, Ltd. | Digital-to-analog converter and PCM encoder using the converter |
| US4348658A (en) * | 1980-05-09 | 1982-09-07 | Motorola, Inc. | Analog-to-digital converter using half range technique |
| US4295089A (en) * | 1980-06-12 | 1981-10-13 | Gte Laboratories Incorporated | Methods of and apparatus for generating reference voltages |
| US4385286A (en) * | 1980-07-18 | 1983-05-24 | American Microsystems, Inc. | Use of single reference voltage for analog to digital or digital to analog conversion of bipolar signals |
| US4350975A (en) * | 1980-07-18 | 1982-09-21 | American Microsystems, Inc. | Dual bandwidth autozero loop for a voice frequency CODEC |
| CA1144653A (en) * | 1980-08-27 | 1983-04-12 | Adrian Worsman | Codec |
| JPS57123732A (en) * | 1981-01-26 | 1982-08-02 | Oki Electric Ind Co Ltd | Digital-to-analog converting circuit |
| US4384277A (en) * | 1981-05-08 | 1983-05-17 | Motorola, Inc. | Digital to analog converter |
| US4451821A (en) * | 1981-11-03 | 1984-05-29 | Texas Instruments Incorporated | Combination of analog to digital converter and error correction circuit |
| US4531113A (en) * | 1983-06-27 | 1985-07-23 | Gte Laboratories Incorporated | Capacitor array |
| EP0169535B1 (en) * | 1984-07-23 | 1992-06-10 | Nec Corporation | Analog to digital converter |
| JPH0813004B2 (en) | 1984-09-07 | 1996-02-07 | 株式会社日立製作所 | A / D converter |
| ATE58617T1 (en) * | 1985-09-23 | 1990-12-15 | Siemens Ag | CIRCUIT ARRANGEMENT FOR ECHO CANCELLATION. |
| US4668936A (en) * | 1985-10-15 | 1987-05-26 | Hughes Aircraft Company | Untrimmed 12 bit monotonic all capacitive A to D converter |
| SE452830B (en) * | 1985-12-02 | 1987-12-14 | Ellemtel Utvecklings Ab | DEVICE FOR TRANSFORMING AN ANALOGUE, BALANCED SIGNAL (V? 71I? 71N? 71 +, V? 71I? 71N? 71-) TO A DIGITAL SIGNAL THROUGH CHARGING DISTRIBUTION IN A NUMBER OF CAPACITIVE ELEMENTS |
| FR2599526A1 (en) * | 1986-05-29 | 1987-12-04 | Centre Nat Rech Scient | MOS adder and MOS binary multiplier comprising at least one such adder |
| JPH0761019B2 (en) * | 1986-06-19 | 1995-06-28 | 日本電気株式会社 | Analog to digital converter |
| EP0258842B1 (en) * | 1986-09-01 | 1994-11-30 | Siemens Aktiengesellschaft | Analog-to-digital converter with a capacitiv network |
| US5021787A (en) * | 1987-03-26 | 1991-06-04 | Texas Instruments Incorporated | Digital-analog converter for conversion of law A- encoded digital signals into analog signals |
| US4968989A (en) * | 1987-04-20 | 1990-11-06 | Olmstead John A | Switched capacitor filter for use with a digital-to-analog (D/A) converter |
| US5206650A (en) * | 1991-09-05 | 1993-04-27 | John Fluke Mfg. Co., Inc. | Charge-controlled integrating successive-approximation analog-to-digital converter |
| US5172019A (en) * | 1992-01-17 | 1992-12-15 | Burr-Brown Corporation | Bootstrapped FET sampling switch |
| JPH08125152A (en) * | 1994-10-28 | 1996-05-17 | Canon Inc | Semiconductor device, correlation operating unit empolying it, ad converter, da converter, and signal processing system |
| US5675340A (en) * | 1995-04-07 | 1997-10-07 | Iowa State University Research Foundation, Inc. | Charge-redistribution analog-to-digital converter with reduced comparator-hysteresis effects |
| JPH09129864A (en) | 1995-10-30 | 1997-05-16 | Canon Inc | Semiconductor device, and semiconductor circuit, correlation operating device and signal processing system which use the semiconductor device |
| US5920275A (en) * | 1996-09-09 | 1999-07-06 | Iowa State University Research Foundation, Inc. | Analog-to-digital converter using weighted capacitor array and interpolating comparator |
| US6259392B1 (en) * | 1997-10-08 | 2001-07-10 | Samsung Electronics Co., Ltd. | Multiplying digital-to-analog converters and methods that selectively connect unit and feedback capacitors to reference voltages and feedback voltages |
| US6963510B1 (en) | 1998-07-10 | 2005-11-08 | Xilinx, Inc. | Programmable capacitor and method of operating same |
| US6255675B1 (en) * | 1998-07-10 | 2001-07-03 | Xilinx, Inc. | Programmable capacitor for an integrated circuit |
| US7002506B1 (en) * | 2004-12-23 | 2006-02-21 | Texas Instruments Incorporated | Providing pipe line ADC with acceptable bit error and power efficiency combination |
| US7251121B2 (en) * | 2005-03-05 | 2007-07-31 | Innovation Engineering Llc | Electronically variable capacitor array |
| US7796079B2 (en) * | 2009-01-28 | 2010-09-14 | Freescale Semiconductor, Inc. | Charge redistribution successive approximation analog-to-digital converter and related operating method |
| US8125231B2 (en) * | 2009-01-28 | 2012-02-28 | Freescale Semiconductor, Inc. | Capacitance-to-voltage interface circuit, and related operating methods |
| US7969167B2 (en) | 2009-01-28 | 2011-06-28 | Freescale Semiconductor, Inc. | Capacitance-to-voltage interface circuit with shared capacitor bank for offsetting and analog-to-digital conversion |
| US8405407B2 (en) * | 2009-06-05 | 2013-03-26 | Chimei Innolux Corporation | Current measurement circuit and measuring method thereof including a binary weighted capacitor array |
| PL220486B1 (en) * | 2010-06-05 | 2015-10-30 | Akademia Górniczo Hutnicza Im Stanisława Staszica W Krakowie | Method and system for converting the electric charge into a digital word |
| DE102010063405A1 (en) * | 2010-12-17 | 2012-06-21 | Siemens Aktiengesellschaft | Analog-to-digital converter, method for operating an analog-to-digital converter and method for converting an analog input signal into a digital output signal |
| US9148603B2 (en) * | 2012-05-15 | 2015-09-29 | Semiconductor Components Industries, Llc | Offset injection in an analog-to-digital converter |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4129863A (en) * | 1977-10-03 | 1978-12-12 | Regents Of The University Of California | Weighted capacitor analog/digital converting apparatus and method |
-
1978
- 1978-11-24 US US05/963,489 patent/US4195282A/en not_active Expired - Lifetime
-
1979
- 1979-01-25 CA CA320,239A patent/CA1124868A/en not_active Expired
Also Published As
| Publication number | Publication date |
|---|---|
| US4195282A (en) | 1980-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA1124868A (en) | Charge redistribution circuits | |
| US5450085A (en) | Method and apparatus for high speed analog to digital conversion using multiplexed flash sections | |
| US4799042A (en) | Apparatus and method for offset voltage correction in an analog to digital converter | |
| GB2070364A (en) | Converter | |
| JPS63215223A (en) | Analog/digital converter | |
| US4404544A (en) | μ-Law/A-law PCM CODEC | |
| GB2080059A (en) | Digital-to-analogue converter | |
| CA2004317A1 (en) | Successive comparison type analog-to-digital converting apparatus | |
| US4384277A (en) | Digital to analog converter | |
| WO1982003955A1 (en) | A multiple function operational amplifier circuit | |
| US4384276A (en) | Capacitive DAC to filter interface circuit | |
| US5923275A (en) | Accurate charge-dividing digital-to-analog converter | |
| US4531113A (en) | Capacitor array | |
| EP0564143B1 (en) | Multi-mode analog to digital converter and method | |
| US3653035A (en) | Chord law companding pulse code modulation coders and decoders | |
| EP0251758A2 (en) | Digital-to-analog conversion system | |
| US4468654A (en) | Charge redistribution a-law PCM decoder | |
| US4983974A (en) | Analog-to-digital conversion by varying both inputs of a comparator utilizing successive approximation | |
| CA1237523A (en) | Linear codec with dual divider | |
| EP0247065B1 (en) | An untrimmed 12 bit monotonic all capacitive a to d converter | |
| JPH036126A (en) | High-speed analog-digital converter | |
| GB2142196A (en) | Improvements in or relating to a/d coders and decoders | |
| EP0164747A2 (en) | Successive approximation adc with variable frequency clock | |
| CN115441874A (en) | Fourteen-bit resolution two-stage cyclic analog-to-digital converter | |
| JPH04162828A (en) | Pcm encoder |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| MKEX | Expiry |