Automatic equalizing method and system

What is claimed is:

1. Apparatus for use at the receiving terminal of a suppressed carrier data communications system, receiving transmitted data signals over a transmission channel, for automatically compensating for static transmission channel characteristics, and for dynamic variations in the amplitude and phase characteristics of the locally-generated carrier frequency, comprising:

a transversal filter having an input for receiving distorted data signals from the transmission channel, said transversal filter having a controllable transfer function, and an output for data signals as modified by said transfer function;

dynamic compensating means operatively connected to said output of said filter for correcting distortion of the modified data signals caused by the dynamic variations in the amplitude and phase characteristics of the locally-generated carrier, said dynamic compensating means having controllable compensating characteristics and having an output for providing distortion corrected data signals;

decision means having an input and an output, with said input connected to the output of said dynamic compensating means, said decision means responding to the distortion corrected data signals from said dynamic compensating means for producing a digital output representation of the transmitted data signals at said output thereof;

error signal generating means connected to said input and said output of said decision means for producing an error signal proportional to the difference between the signals appearing at said input and said output of said decision means;

processing means connected to said decision means and to said error signal generating means, and responsive to said output of said decision means and to said error signal for processing the error signal to produce a first control signal representative of only the data signal distortion due to the static transmission channel characteristics, and a second control signal representative of only the data signal distortion due to the dynamic locally-generated carrier characteristics;

first control means operatively connected to said transversal filter, to said processing means, and to said error signal generating means, and responsive to the first control signal and to said error signal for varying said controllable transfer function of said filter to minimize the error signal, thereby substantially correcting the distortion of the data signals due to the static characteristics of the transmission channel, and

second control means operatively connected to said dynamic compensating means and to said processing means, and responsive to the second control signal for varying said controllable compensation characteristics of said dynamic compensating means to minimize the error signal, thereby substantially correcting the distortion of the data signals due to the dynamic variations in the characteristics of the locally-generated carrier.

2. The apparatus as recited in claim 1 in which said processing means includes:

second order integrator means for receiving the error signal from said error signal generating means and the digital output of said decision means and producing the second control signal; and

first multiplying means for receiving the error signal from said error signal generating means and the second control signal from said second order integrator means and multiplying same for producing the first control signal.

3. The apparatus as recited in claim 2 in which said dynamic compensating means includes second multiplying means for producing the product of the modified data signals and the second control signal.

4. The apparatus as recited in claim 2 in which said second order integrator means has a transfer function of the form (a + b/s)1/s.

5. The apparatus as recited in claim 1 in which said error signal generating means includes a subtraction circuit for producting the error signal.

6. The apparatus as recited in claim 1 in which said dynamic compensating means has a speed of response selected to allow tracking of the phase jitter of the locally-generated carrier independently of the accuracy of said transversal filter compensation characteristic.

7. The apparatus as recited in claim 1 in which said transversal filter includes:

a delay line tapped at intervals providing delays equal to the time duration of a data symbol of the data signals;

a tap gain circuit connected to each of the taps of said delay line;

a tap gain control circuit operatively connected to each of said tap gain circuits, and responsive to the first control signal; and

summing means connected to each of said tap gain circuits for summing the outputs of said tap gain circuits to produce the modified data signals.

8. The apparatus as recited in claim 7 in which said first control means comprises:

multiplying means operatively connected to each tap circuit of said delay line for multiplying the first control signal and the tap output signal; and

averaging means connected to output of said multiplying means for averaging the output signal of said multiplying means, the output of said averaging means being connected to said tap gain control circuit for control thereof.

9. The apparatus as recited in claim 7 having a number of taps of said delay line selected to provide a predetermined accuracy of compensation for distortion due to the static transmission channel characteristics independently of distortion due to dynamic variations of the locally-generated carrier.

10. In a transmission system comprising a receiver having a decision circuit for making a decision as to the identity of received digital signals, an apparatus for automatically correcting both a steady state portion of distortion in digital signals due to static characteristics of the transmission system and a dynamic portion of distortion in digital signals due to dynamic characteristics of the transmission system carrying such signals, comprising;

means for producing an error signal substantially proportional to the distortion of the digital signals;

error signal processing means for receiving the error signal from said producing means, and responsive thereto and to said decision from said decision circuit for resolving the error signal into a first component substantially proportional to said steady state portion of the distortion due to the static characteristics of the system, and a second component substantially proportional to said dynamic portion of the distortion due to the dynamic characteristics of the system;

first adjustable compensating means for modifying the distorted digital signals, said first adjustable compensating means receiving and being adjustable in response to the first component of the error signal and to said error signal for minimizing the first component of the error signal, and thereby correcting for the static distortion; and

second adjustable compensating means for modifying the distorted digital signals, said second adjustable compensating means receiving and being adjustable in response to the second component of the error signal and to said error signal for minimizing the second omponent of the error signal, and thereby correcting for the dynamic distortion.

11. The apparatus as recited in claim 10 further comprising:

digital processing means for processing the distorted digital signals; and

program storage means operatively connected with said digital processing means and having a predetermined digital program stored therein for controlling said digital processing means to operate in a predetermined sequence to process the distorted digital signals whereby the functions of said error signal producing means, said error signal processing means, said first adjustable compensating means, and said second adjustable compensating means are performed by said digital processing means in response to said digital program.

12. Apparatus for automatically correcting distortion including both static distortion due to static transmission channel characteristics and dynamic distortion of baseband digital data signals due to variations of amplitude and phase characteristics of the system carrying such signals comprising:

digital processing means for processing said baseband digital data signals; and

program storage means controllably connected with said digital processing means and having a preselected sequential digital program stored therein for controlling said digital processing means so as to cause said digital processing means to convert the baseband digital data signals to digital codes including said static and dynamic distortion, to generate a code output determined by amplitude and phase values of each successive data period represented by the digital codes, to measure the difference between the code output and the corresponding digital code to develop an error signal, to process said error signal to further develop a first error value indicating said static distortion due to static transmission channel characteristics and a second error value indicating said dynamic distortion due to variations of the amplitude and phase characteristics, and to use said first and second error values to modify the digital code to minimize said first and second error values, respectively, thereby correcting the static distortion and the dynamic distortion, respectively, of the data signals.

13. The apparatus as recited in claim 12 in which said sequential digital program causes said digital processing means to separate the error signal developed from said measured difference into said first error value indicating the static amplitude and phase characteristics of the system and a second error value indicating the dynamic amplitude and phase characteristics of the system, to cause the first component to modify the digital code to minimize distortion due to such static characteristics, and to cause the second component to modify the digital code to minimize said dynamic distortion due to such dynamic characteristics.

14. The apparatus as recited in claim 12 in which said digital processing means includes a digital decision circuit, said circuit generating the code output under control of said digital program with each such code output consisting of a 3 bit code in response to a quadrature amplitude modulation data signal represented by a complex vector, said digital decision circuit including:

a first digital register for storing a digital representation of the polarity of the X-axis component of the vector;

a second digital register for storing a digital representation of the polarity of the Y-axis component of the vector;

a third digital register for storing a digital representation of the polarity of the X'-axis component of the vector where said X'-axis represents the X-axis rotated 45.degree. counterclockwise;

a fourth digital register for storing a digital representation of the polarity of the Y'-axis component of the vector where said Y'-axis represents the Y-axis rotated 45.degree. counterclockwise;

a latch circuit for reading out each of the digital polarity representations of said first through said fourth registers sequentially; and

an output code generator for sequentially receiving during one data period the digital polarity representations of one of eight complex vectors from said latch circuit and for generating the 3-bit code output uniquely determined by the sequence of polarities of the digital polarity representations.

15. A method for automatically correcting static and dynamic distortion of digital data signals having a selected data symbol duration, comprising the steps of:

passing the signals through a transversal filter having a plurality of adjustable tap gain circuits;

passing the signals from the transversal filter through an adjustable network for correcting data signal distortion due to variations of the channel;

performing a decision operation on the signals from the adjustable network to generate an output code group for each data symbol duration period;

comparing each output code group with the signals from the adjustable network to generate an error signal;

processing the error signal so as to separate same into a first component representing needed adjustment of the tap gain circuits to compensate for static distortion of the digital data signals and a second component representing needed adjustment of said adjustable network to compensate for dynamic distortion of the digital data signals;

adjusting the adjustable tap gain circuits in accordance with the first component to minimize the value of the first component; and

adjusting the adjustable network in accordance with the second component to minimize the value of the second component.

16. A method for automatically correcting static and dynamic distortion of digital data signals having a selected data symbol duration, comprising the steps of:

converting a baseband data signal occurring in a data symbol duration period to a digital code value;

storing a selected number of such digital code values in a tap register;

modifying the value of each of the selected number of stored digital code values according to the value of corresponding digital coefficients stored in a tap gain register;

combining the modified stored digital code values to produce a first compensated coded value;

modifying the first compensated coded value in accordance with a digital coefficient stored in a CA register to produce a second compensated coded value;

storing the second compensated coded value in an A register;

judging the second compensated coded value from the A register to determine an output digital code group denoted by such second compensated coded value;

generating the digital code group so determined;

comparing the output digital code group with the second compensated coded value stored in the A register to generate a first error signal representing needed adjustment of the digital coefficients stored in the tap gain register to compensate for static distortion of the digital data signals and a second error signal representing needed adjustment of the digital coefficients stored in the CA register to compensate for dynamic distortion of the digital data signals;

adjusting the digital coefficients stored in the tap gain register in accordance with the first error signal in a manner to minimize succeeding values of the first error signal;

adjusting the digital coefficients stored in the CA register in accordance with the second error signal in a manner to minimize succeeding values of the second error signal; and

repeating each of such steps for each successive baseband data signal occurring in each data symbol duration period.

17. The method described in claim 16 which includes the further step of controlling each of such steps by means of a digital program stored in a digital program storage element.

18. The method described in claim 16 in which the baseband data signal is composed of two signal components in quadrature phase relationship and in which said judging step includes the further steps of:

resolving the second compensated coded value from the A register into four polarity indicating components; and

determining the polarity sequence of the polarity-indicating components to determine a 3-bit code group represented by such sequence.

19. Apparatus for automatically correcting static and dynamic distortion of digital data signals having a selected data symbol duration, comprising:

digital multiplier means for producing products of digital coded values;

digital adder means for producing sums and differences of digital coded values;

analog-to-digital converter means for receiving baseband digital data signals and generating a digital coded value for each signal occurring in a data symbol period;

tap register means for storing a selected number of the same digital coded value;

tap gain register means for storing a set of tap gain digital coefficients equal in number to said selected number, said multiplier means multiplying each of such stored coded values by a corresponding one of such tap gain coefficients, said digital adder means producing the sum of such multiplied digital coded values, such sum representing a first compensated coded value;

sum-signal register means for storing said first compensated coded value;

sum-signal coefficient register means for storing sum signal coefficients, said multiplier means multiplying said first compensated coded value by said sum-signal coefficient to produce a second compensated coded value;

digital decision means for receiving said second compensated coded value and generating an output digital code group representing a best estimate of the transmitted digital code group, said multiplier and said adding means performing computations on the code output and second compensated coded value to produce a first error signal representing needed adjustment of the tap gain register means to compensate for static distortion of the digital data signals and a second error signal representing needed adjustment of said tap gain register means to compensate for dynamic distortion of the digital data signals;

means for performing computations on the first error signal and the set of tap gain coefficients thereby to change the values of the coefficients in a manner to minimize the value of the first error signal, and performing computations between the second error signal and the sum-signal coefficient thereby to change the value of said coefficients to minimize the value of the second error signal; and

program storage means having a plurality of control buses and signal buses interconnecting said multiplier means, said adder means, said converter means, said tap register means, said tap gain register means, said sum-signal register means, said sum-signal coefficient register means, and said decision means, said program storage means having a digital program stored therein for sequentially controlling each of said interconnected means to minimize the first and second error signals whereby the distortion of the digital data signal is substantially corrected.

Description Submit all comments and votes

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to a transversal-type equalizing apparatus and method for use in a suppressed carrier data transmission system, and more specifically to apparatus and method that automatically compensates for phase pertubations in the reinserted carrier signal at the receiving terminal of such system as well as for variations in phase and amplitude of the received signal due to the transmission channel.

2. Description of the Prior Art

The demands for data transmission have greatly increased in recent years, resulting in development of techniques for transmission of high-rate pulse-type data signals over bandlimited transmission channels such as telephone lines and the like. Additionally, demands are also placed on greater accuracies in detection of such data signals at the receiving terminal.

Two significant factors influencing the capability of a data communication system to correctly receive the transmitted signals are: (1) the noise characteristics of the transmission channel, and (2) the intersymbol interference caused by the time dispersion in the channel. For low-noise environments, as typified by telephone channels, the most important consideration is minimization of the latter factor.

For a given telephone channel, the transfer function of the channel is characterized by its attenuation and delay as a function of frequency. The envelope delay, defined as the derivative of the phase characteristic, represents the relative time of arrival of the various frequency components of the input signal; thus, a sequence of pulses, or symbols, in passing through the channel are distorted and overlap in time between successive symbols. This distortion is referred to above as intersymbol interference.

While the above-mentioned channel characteristics may be relatively constant for a specific telephone line, in switched telephone networks a channel is generally chosen at random for a given transmission; therefore a wide variation of channel characteristics may be expected with respect to a given receiving terminal. To solve this problem, it is well-known to provide at the receiving terminal an automatic equalizer having the capability to compensate for the expected range of attenuation and phase variations of the incoming channels, thus minimizing intersymbol interference in the received data.

The most effective prior art equalizer is the transversal filter wherein a delay line tapped at intervals equal to the symbol period has each tap connected through a variable gain element to a summing bus. The gain elements are automatically controlled by means of an error signal to introduce echoes of the signal of such amplitudes as to compensate for the overall channel delay characteristics.

The requirements for higher data rates have led to development of multi-level modulation methods (e.g., multiphase and quadrature amplitude modulation methods) to increase the data rate in a given channel bandwidth. As is well known, it is common to translate the baseband data signal to a higher carrier frequency, and to frequency multiplex a number of such frequency-translated signals over a single wide-band channel. Generally, the carrier frequency is suppressed and only the upper and lower sidebands are transmitted, this method being known as double sideband, suppressed carrier (DSBSC) transmission. Therefore, the pilot signal is not required. This makes unnecessary the use of complicated carrier regeneration circuit which inevitably was required for regenerating the pilot signal. Moreover, since there is no need to allocate within the bandwidth available for transmission the pilot signal adjacent to the data signal, a comparatively simplified filter circuit can be used. This is a large merit of such a transmission system.

At the receiving terminal, however, it is necessary to regenerate the carrier frequency and to insert the carrier at the correct phase in order to demodulate the data signal and thereby recover the baseband signal. The requirement for locally generating the carrier for demodulating the DSBSC signal gives rise to an additional source of signal distortion. Unavoidable variations in generation of the carrier may result in frequency offset and phase jitter. Prior art automatic equalizers of the transversal type have attempted to solve this second distortion problem by two basic compensating, or correcting, techniques. The first is to use a local carrier generator or of fixed frequency, and to rely on the normal automatic adjustment of tap gains of the equalizer to correct for carrier frequency deviation and phase jitter. The second is to use a voltage controlled local carrier-generating oscillator and to control same as a function of the error signal of the automatic equilizer.

Such compensating methods have the disadvantage that high channel equalizing accuracy requires a relatively large number of delay line taps, and the response time of the filter increases with the number of taps; yet, the response time must be short if the system is to be able to follow the rapid carrier phase jitter. Therefore, a satisfactory compromise between equalizing accuracy and phase jitter compensation is difficult or not possible -- i.e., increasing equalizer accuracy increases jitter, and decreasing jitter results in decreased accuracy.

SUMMARY OF THE INVENTION

The present invention overcomes the disadvantages of the prior art by providing means, used in combination with an automatic equalizer, for automatically compensating for variations in phase of the locally-generated carrier in a DSBSC data transmission receiving terminal without affecting the equalizing accuracy of the automatic equalizer with respect to the data channel amplitude and phase characteristics. Hence, the present invention provides an automatic equalizing method and system which can effectively compensate, or correct the phase of the carrier without deterioration of the equalizing accuracy of the automatic equalizing circuit.

One embodiment of the invention involves modifying the signal appearing on the transversal filter summing bus in accordance with a characterization of the error signal caused only by the carrier phase pertubations. The modified sum signal is then applied to the decision circuitry and a composite error signal derived conventionally. The composite error signal is next processed to remove the carrier phase pertubation error signal and thus to provide a gain control signal that is relatively free from the effects of the phase variations. The resultant gain control signal is used to control the tap gain circuits of the transversal filter.

It is therefore a primary object of the present invention to provide apparatus and method for automatically compensating for distortion in received data symbols caused by phase pertubations in the re-inserted carrier of a DSBSC data transmisison receiving terminal.

It is a further object of the present invention to provide apparatus for and a method of compensating for carrier phase pertubations without affecting the accuracy of the automatic compensation for variations in the data transmission channel amplitude and phase characteristics.

It is another object of the present invention to provide an automatic equalizing apparatus and method that will closely track phase variations of the locally generated carrier signal.

It is yet another object of the present invention to provide an automatic equalizing apparatus in which variations in phase of the locally-generated carrier signal are compensated through the use of digital circuitry.

These and other objects and advantages of the present invention will become apparent to one skilled in the art to which the invention pertains from a perusal of the appended claims and the detailed description when read in conjunction with the attached drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a conventional prior art automatic equalizing circuit of the transversal-filter type presented to delineate the operating principles thereof;

FIG. 2 is a block diagram of one embodiment of an automatic equalizing circuit in accordance with the present invention;

FIG. 3 is a block diagram of another embodiment of the present invention, showing an automatic equalizing circuit realized with digital circuit elements;

FIG. 4 is a diagram of the coding technique used for transmission of information in a quadrature amplitude modulation (QAM) system that may utilize the present invention; and

FIG. 5 is a detailed block diagram of the decision circuit shown in FIG. 3 suitable for use in the QAM system referred to by FIG. 4.

DETAILED DESCRIPTION OF THE INVENTION

Before describing embodiments of the present invention, it is helpful to analyze the prior art automatic equalizer of the transversal-filter type as applied to a DSBSC data communication system. Accordingly, in FIG. 1, a typical prior art automatic equalizer is shown. Input terminal 1 of demodulation circuit 2 receives the DSBSC signal. Demodulator 2 includes a local oscillator 2a which generates the demodulation signal .theta.c of the required carrier frequency, applied to multiplier 2b. The DSBSC signal is then multiplied by multiplier 2b with signal .theta.c producing a baseband signal at the output of the demodulator 2. A low-pass filter, although not shown, is assumed to be (and typically is) included in demodulator 2 to pass only the fundamental baseband frequencies to the output.

As will be recognized, the baseband signal can include distortion from the following sources:

1. the amplitude vs. frequency characteristic of the transmission channel is not a constant;

2. the phase vs. frequency characteristic of the transmission channel is not linear with frequency;

3. pertubations in the phase of carrier .theta.c.

The baseband signal is passed through a series of delay circuits 3 in the delay interval of each of which is selected to be equal to the data symbol duration. A tap at each delay circuit 3 supplies the signal from each delay 3 through a corresponding variable tap gain circuit 4, the outputs of all circuits 4 being supplied to summing circuit 5. The summed signal at the output of summer 5 is thus the baseband signal modified by weighted and delayed versions, or sample portions of the signal. By properly adjusting each tap gain circuit 4, the distortions of the baseband signal can be minimized, within the inherent capabilities of the equalizer.

The summed signal is accordingly analyzed by decision circuit 6 which judges and regenerates the data signal, based on a prior knowledge of the total ensemble of possible signals, as to the correct signal (typically by a thresholding operation) which then appears as the digital output at output terminal 10. Difference circuit 7 substracts the output signal of summer 5 from the digital output signal of the decision circuit 6, with the difference thus representing the residual distortion, or error, signal. The error signal is utilized in a negative feedback fashion to modify the transfer characteristics of the transversal filter to cause the summed signal from summer 5 to more nearly approximate the digital output signal of decision circuit 6, thereby reducing the error signal. Accordingly, the error signal is multiplied by multipliers 8 respectively associated with the taps, the products therefrom being smoothed or averaged by associated integrators 9, and the outputs of the latter then are utilized as gain control signals for the corresponding tap gain circuits 4. It is to be noted that the signals present at each multiplication and summing point are considered to be complex, varying in both amplitude and phase.

Having briefly summarized the operation of the prior art circuit of FIG. 1, an analysis will be described to illustrate the operation of the circuit and the problems of such a circuit to which the present invention is directed.

Case (1) Static channel characteristics:

The baseband input signal to the delay circuit is expressed as x(t) and is assumed to have the form ##EQU1## where: a(K) is the code transmitted over the channel;

h.sub.1 (t) is the impulse response of the equalizer input; and

T is the symbol or pulse duration of the code.

The automatic equalizer output, y(t), can be expressed as: ##EQU2## where: h.sub.2 (t) is the impulse response of the equalizer, and is given by: ##EQU3## where: M is the number of taps before the center tap,

N is the number of taps after the center tap, and

C.sub.n is the tap gain coefficient.

The relation between x(t) and y(t) can be expressed as: ##EQU4## and the equalizing output after sampling becomes: ##EQU5##

Referring to the output of the decision circuit 6 as a(.tau.), the error .epsilon.(.tau.), comprising the difference between the equalizer output and the decision circuit output and hereinafter referred to as the "equalizing error," is seen to be

.epsilon.(.tau.) = y(.tau.,T) - a(.tau.). (6)

The automatic equalizing circuit is adjusted by an algorithm that determines the correlation of the input signal x(t) and the equalizing error .epsilon.(.tau.) to compensate the tap coefficient C.sub.K, as expressed by the following equation:

C.sub.K.sup.(n+1) = C.sub.K.sup.(n) - .alpha..sub.c .multidot..epsilon.(.tau.).multidot.x {(.tau.-K)T} (7)

where .alpha..sub.c is the control coefficient, and C.sub.K.sup.(n) is the gain value of the Kth tap adjusted by n times.

Assume that the optimum C.sub.K has been obtained by minimization of the mean-square (MS) distortion and is C.sub.OK ; then, the root mean-square error defined by the following equation becomes a minimum when C.sub.K = C.sub.OK : ##EQU6##

Case (2) Dynamic channel characteristics:

The general case for which the channel amplitude and phase characteristics vary with time, as will occur due to jitter in the locally-generated carrier, will be considered with the input signal to the automatic equalizer being designated as x'(t). x'(t) may be expressed as:

x'(t) = x(t)e.sup..gamma.(t) (9)

where:

.gamma.(t) = .alpha.(t)+i .beta.(t) and

.alpha.(t) = time variation of amplitude and

.beta.(t) = time variation of phase (including phase jitter and any frequency offset).

The output y'(t) of the transversal filter may be expressed as ##EQU7##

If the total delay time, (M+N)T, through the automatic equalizing circuit is much smaller than the dynamic channel variation rate, the value of e.sup.(t-nT) will appear to be a constant for n = -M through N, which may be expressed as e.sup..gamma.(t).

In this case, the expression for channel variation reduces to ##EQU8##

The sampled equalizing error .epsilon.'(.tau.) now may be expressed as:

.epsilon.'(.tau.) = y'(.tau.,T) - a(.tau.)

= y (.tau.,T)e.sup..gamma.(.tau.,T) -a(.tau.). (12)

In order to minimize the inter symbol interference, the following expression must be optimized: ##EQU9##

For this expression to become a minimum in accordance with equation (8), the following relation must obtain:

C.sub.n ' = C.sub.on .multidot.e.sup.-.gamma.(t) (14)

In other words, C.sub.n ' must follow or track the channel fluctuation e.sup.-.gamma.(t).

Thus, the fundamental problem in the prior art transversal-filter type automatic equalizer can be now seen. Increasing the speed of response of the filter by reducing the number of delay line taps to allow C.sub.n ' to track the channel fluctuation e.sup.-.gamma.(t) results in decreased accuracy of equalization of the static channel characteristics; and in general, the equalizer cannot be optimized for both types of compensation. Furthermore, when the carrier includes a frequency offset, e.sup.-.gamma.(t) exhibits a linear increase of phase, and prior art circuits do not have sufficient control to track such error.

Turning now to FIG. 2, the block diagram of a preferred embodiment of the present invention is shown. Elements in FIG. 2 that correspond to the identical elements in FIG. 1 are identified by the same numbers. As may be noted, the identical prior art transversal-filter type automatic equalizer section of FIG. 1 is included in the instant invention in combination with a phase compensating network 11 and an equalizing error compensation multiplier 12.

Compensating network 11 is connected at the output of the transversal-filter and serves to compensate for phase jitter of the channel characteristic independently of the static characteristic as will be explained in detail hereinbelow. The equalizing error compensation multiplier 12 is connected in the tap-gain control feedback line and is utilized to prevent the phase jitter error component in the equalizing error signal from affecting the static compensation of the equalizer.

The Signal y"(t) at the output of summer 5 may be conveniently expressed in terms of the sampled output of decision circuit 6 and the equalizing error signal as follows:

y"(t) = {a(.tau.,T) + .epsilon.(.tau.)} e.sup..gamma.(t) (15)

.epsilon.(t) as defined in equation (6) represents the equalizing error for the static channel characteristics and will hereinafter be referred to as the steady state error.

At point C in network 11, signal C.sub.APC appears as will be shown hereinbelow and approaches e.sup.-.gamma.(t) with an error of .DELTA.(t). This residual error, .DELTA.(t) will hereinafter be referred to as the dynamic error. With this notation, the value of C.sub.APC may be expressed as

C.sub.APC = e.sup.-.gamma.(t) + .DELTA.(t) (16)

The value C.sub.APC is applied to multiplication circuit 13 with the product of y"(t) and C.sub.APC designated as z(t), which has the form

z(t) = y"(t).multidot.C.sub.APC

= {a(.tau.,T) + .epsilon.(.tau.)} e.sup..DELTA.(t)

.apprxeq. a(.tau.,T) + .epsilon.(.tau.) + a(.tau.,T).DELTA.(t) + .epsilon.(-.tau.).DELTA.(t). (17)

As may now be seen from equation (17), z(t) is the decision circuit output term a(.tau.,T) plus a residual error involving the steady state error .epsilon.(t) and the dynamic error .DELTA.(t) and represents the input to decision circuit 6. As is well known, a digital decision circuit generates a discrete output for all input levels lying within a set of limits, effectively determining within which set of limits the analog signal lies. Therefore, the residual error of Equation (17) need be only sufficiently small so that the vector y"(t) lies within the limits which define its correct digital value. Since .DELTA.(t), the dynamic error, can be made to approach zero as shown hereinbelow, the circuit can thus be compensated for the source of this error.

The manner in which the instant invention extracts or compensates for the dynamic channel variations at the output of summer 5 will now be explained. Assuming that the sampled output of decision circuit 6 is a(.tau.,T), the overall equalized error signal E(t) at the output of difference circuit 7 may be expressed as follows:

E(t) = z(t) - a(.tau.,T)

= .epsilon.(.tau.) + a(.tau.,T).DELTA.(t) + .epsilon.(.tau.).DELTA.(t)

.apprxeq. .epsilon.(t) + a(.tau.,T).DELTA.(t), where .epsilon.(t)