DE2133037A1 - PROCEDURE AND CIRCUIT ARRANGEMENT FOR THE SELF-ADJUSTMENT OF AN ADAPTIVE EQUALIZER - Google Patents

Description

Method and circuit arrangement for the automatic setting of a adaptive equalizer.

The invention relates to a method and a circuit arrangement for the automatic setting of an adaptive, the undoing of linear distortions Equalizer serving on a transmission path, in particular tranaversal filter equalizer, in a time division multiplex system for pulse amplitude modulation, in which in addition to the message signals a periodically recurring additional signal is transmitted at the receiving end Samples of the amplitudes of the regenerated additional signal from samples of the The amplitudes of all incoming signals are subtracted from the difference signal and the regenerated additional signal setting criteria for setting elements of the equalizer The advantages of the time division multiplex transmission of pulse amplitude modulated Signals, which in the economic efficiency of the effort and the bandwidth requirement can only be fully exploited if they are on the transmission paths resulting linear distortions of the signals to be transmitted in each individual Multiple channel can be undone.

Linear distortion leading to noticeable crosstalk between the individual channels can be carried out over longer periods with fixed equalizers Only inadequate control over time, for example because of temperature fluctuations Can cause changes in the transmission properties of the transmission paths.

Automatic equalization with the help of special test signals at the beginning of the transmission with subsequent definition of the equalizer setting does not make sense because the transmission properties change during operation can no longer be regulated and since it is not possible on a regular basis Time intervals, for example, to shut down thirty transmission channels at the same time, to transmit a test signal sequence.

For pulse code modulated signals which only have the states 110.11 "+ 1 "and 11111 is from the publication" An automatic equalizer for Digital Repeatered Lines "in the Proceedings of the IEEE", March 1969, p. 363 and 364, a method for automatic equalization during. known of the transfer. There it is shown that under certain conditions the maximum amplitude of the equalize Signal is a measure of the cable attenuation at the "Nyquist frequency". Since the typical Cable attenuation curve is known, karin this level for setting an automatic Serve equalizer. However, this method cannot be applied to pulse amplitude modulation Transmit signals with more than two possible amplitude values.

This method also results in insufficient equalization, since the course of the cable attenuation is only roughly approximated.

Next is from the publication "An Adaptive Receiver for Digital Signaling Through Channels with Intersymbol Interference "by J.G. Proakis and J.H. Miller in "IEEE Transactions on Information Thecry", Vol. IT-5, No.4, July 1969, 5. 484-497 in particular p. 488, a method for the automatic setting of a Equalizer known that is used to form the setting criterion from a known test signal sequence Makes use of which at regular intervals in the flow of information to be transmitted is inserted. These Test signal sequence can be on the receiving side can be reproduced without errors and is used to minimize what is contained in the signal mean square error. However, this setting criterion cannot be used, when high crosstalk attenuation is required between adjacent channels.

This requires zero crossings of the impulse response of each channel to the Times at which the information in the other channels is sampled.

In "Computer Simulation of an Automatic Adaptive Equalizer for Real Telephone Channels and Free Data Format "by F. Eggimann," International Seminar on Digital Processing of Analog Signals ", Zurich, Switzerland 11.-13.3.1970, that will Method of integrating a known test pulse train into the one to be transmitted Information extended to the effect that a criterion is also taken into account, which zero crossings of the impulse response of each channel to the left and right of the Maximalvert of this impulse response forces points in time in which the information is queried in the adjacent channels. However, a procedure is used here at which-to set the variable parameters not the deviation of the distorted from the ideal impulse response, but rather the distorted impulse response at the sampling times is determined. As in the publication by R.W. Lucky "Techniques for Adaptive Equalization of Digital Communication Systems ", in" The Bell System Technical Journal, February 1966, pp. 255-286, this method is unfavorable because it is slow to converge. In addition, the term denoted in the following by the main word can be used in this way Maximum value of the impulse response cannot be regulated to the setpoint. For the correlation of the delayed equalizer output signals with the reference signal at the output of the reference signal generator or the decision circuit, analog multipliers are also required, which makes instruction difficult. A is used to equalize the post-oscillation riickgckoI, pelter transversal filter. This structure can, in particular at Occurrence of disturbances that lead to instabilities and self-excitement. As a test pulse train a binary bipolar pseudo-random signal is used, from which at regular intervals a pulse is inserted into the sequence of the information-carrying PAM signals. on the receiving side gets this reference signal through a received signal generated error-free synchronized pseudo-random generator in a suitable manner. This requires a not inconsiderable effort, especially because of the proportionate large period of the bear pseudo-Sufalls-Polge.

The invention is based on the object of the above-described Automatic adjustment procedure that avoids difficulties as completely as possible an adaptive equalizer in a time division multiplex system for pulse amplitude modulation to specify.

Based on a method for the automatic adjustment of a adaptive, the cancellation of linear distortions on a transmission link Serving equalizer, in particular transversal filter equalizer, in a time division multiplex system for pulse amplitude modulation, in which, in addition to the message signals, a periodic recurring additional signal is transmitted, with the receiving side samples the amplitudes of the regenerated additional signal of samples of the amplitudes of all incoming signals are subtracted and in the case of the difference signal and the regenerated additional signal setting criteria for adjustment glleder of the equalizer derived are, this object is achieved according to the invention in that as an additional signal the frame synchronizing signal is used.

A particularly simple variant of the method results when alternately positive and negative as a frame synchronization signal. in their amplitude same'and, measured on the transmitted signal, pulses with maximum amplitude are used will.

It is advantageous if, in order to form the setting criteria, the regenerated Additional signal in relation to the difference signal step by step by integer multiples the time interval between two consecutive samples is offset, if products from the samples of the incrementally shifted regenerated additional signal and the difference signal if formed from the products of each step a sum is formed and if these sums serve as setting criteria.

It is also advantageous if to form the setting criteria the regenerated additional signal compared to the difference signal step by step by whole numbers Multiples of the time interval between two consecutive samples offset when products are regenerated from the samples of the incrementally shifted Additional signal and the sign of the differential signal are formed if'aus the products of each step are summed up and when these sums are called Adjustment criteria are used.

Finally, it is advantageous if the setting of the adjusting members is slowed down or blocked when interfering signals occur.

In a preferred embodiment according to the invention as setting elements, step-wise adjustable circuits provided with the help of digital-to-analog converters or of magnetic memories, in particular ransfluxors, will be realized.

Bridge circuits can also advantageously serve as setting elements, the one by voltage negative feedback linearized Pelde effect transistor as controllable Item included.

It is also advantageous as controllable elements of the setting members Photoresistors; to be provided.

For the practical implementation of the subject matter of the invention, it is advantageous fixed pre-equalizers to be provided.

In the following, the invention is explained in more detail with the aid of exemplary embodiments explained.

FIG. 1 shows a block diagram and FIG. 2 shows pulse diagrams for explanation of the principle of the known transversal filter, Fig. 3 is a block diagram of the adaptive Equalizer according to the invention and FIG. 4 shows a pulse plan of the inputs and -am Output of the comparator signals present.

A transversal filter (echo equalizer) is expediently used as the equalization filter. used. This filter can zero crossings of the impulse response in sampling times force left and right of the maximum value of the impulse response and therefore minimized the mutual influence of the neighboring signals and thus the crosstalk. FIG. 1 shows the block diagram of a transversal filter and FIG. 2 shows pulse plans for more detailed explanation of its function.

The transversal filter in Fig. 1 consists of a number of in chain switched delay elements 1, 2, 3 and 4.

Each delay element 1, 2, bung 4 has a delay time corresponding to the distance between two consecutive characters sent. As well as the input 5 as well as each tap of the delay elements 1, 2, 3 and 4 formed delay line is via a setting member 6 to 10 and over Lines 11 to 15 are connected to the inputs of an adder 16, the output of which 17 at the same time the output of the transversal filter. represents.

For the sake of simplicity, the illustration of the pulse plans in FIG. 2 is used assumed that the signal at input 5 of the transversal filter with the aid of a sample-and-hold circuit is preformed. The signal is at the respective sampling times in the intervals T is sampled and the sampled value is stored with the aid of a storage element, for example a capacitance, held constant until the next sample is taken over.

The setting members 6 to 10 allow the signals to be multiplied at the respective taps with a real positive or negative numerical value. Such an adjusting element can in principle consist of a potentiometer and a switchable one Inverting amplifiers exist. In practice, of course, it is also implemented electrical means. Under the assumption = -me that the adjusting member 8 is superimposed on the cO = +1 is set, appears at point 18 in the transversal filter and likewise the line 13 the unchanged input signal. It points towards the signal on Input 5 only has a time delay.

The function of the transversal filter is based on the pulse plans in Fig. 2 explains in more detail. There is a single distorted one on line 13 Impulse response w shown.

This impulse response contains two pre-vibrations and two post-vibrations. To eliminate the first pre-oscillation, the distorted signal is reduced by the time 2 * X tapped earlier, i.e. at input 5 of the delay line, and the setting element 6 is set to such a value c 2 that the main value of this impulse response opposite to the pre-oscillator of the impulse response Xr on the line 13 is. If the impulse response u appearing on the bedding 11 in FIG. 1 becomes the impulse response w is added, the first pre-oscillation disappears.

The remaining pre- and post-oscillations can be treated accordingly will and the sum signal shown in FIG. 2 appears as the output signal at output t7 z. The distortions in the sum signal z have not yet been completely eliminated. This is coming partly because the delay line is not long enough in the example chosen is, secondly, from the non-optimal setting of the adjusting elements. With everybody Correction signal, new small overshoots are added to the signal so that the maximum Although the amplitude of the overshoot is getting smaller, the time dispersion of the resulting impulse response z is getting bigger and bigger. It's easy to show that the overshoots can practically be completely eliminated if the in the transversal filter included delay line is selected only sufficiently long, so if sufficient many taps are present, and if the distortion is present in the input 5 Signal are not extremely large.

Of course, the translte described is not the salfilter only possible equalizer structure. As already mentioned at the beginning, also use feedback filter structures and appropriate combinations and branching networks, such as those shown in FIG. 1, cost-effective implementation an adaptive equalizer for fast data transmission 'by K. Möhrmann, "NTZ", 1970, Issue 1, pages 36 to 42, are described. For the sake of simplicity, the Invention will be described below only with reference to the transversal filter. It however, it should be expressly pointed out that the following considerations can easily be transferred to the above-mentioned equalization structures.

The following is an adaptive equalizer for equalizing the transmission link described in a time division multiplex system, which both signals in the form of pure Pulse amplitude modulation as well as multi-level (quantized) pulse amplitude modulation allowed to rectify.

The transversal filter can, since on the receiving side, only to individual Sampling times individual, discrete amplitude values are queried, which the transferred information, expanded as an analog value shift register be.

An analog value shift register represents a chain connection of key-hold circuits The individual samples are from sample-and-hold circuit to sample-and-hold circuit shifted and the arrangement therefore replaces a delay line. On a thirty-two channel system must, for example, the required sample-and-hold circles per Sekujide with Pulsampli.tudenmodulation Transferring 32 x 8000 samples, i.e. 256 000 values.

The delay time of a delay element is therefore 3.91 / us FIG. 3 shows the block diagram of an adaptive equalizer according to the invention. This equalizer consists of several assemblies. A first assembly is in turn a transversal filter. This can also be done by another of the aforementioned Branch networks be replaced. Elements similar to those contained in FIG Elements that are effectively the same are denoted identically in FIG. 3, the transversal filter again consists of a chain of delay elements 1 ', 2', 3 ', 4' and 19 and a number of -N to + M setting members 6 ', 7', 8 ', 9' and 20. The outputs these adjusting members are via the lines 11 t g 12 ', 13', 14 'and 21 with the Inputs of the output-side Sunimierers 16 connected. The input of the transversal filter is denoted by 5, the output in turn by 17.

The dashed lines 22 and 23 indicate that the transversal filter also contain further delay and adjustment elements not shown here can.

The input 5 of the transversal filter is at the same time as the input a circuit 24 connected which the clock recovery from the time division multiplexed signal contained frame synchronization signal is used. For example, appears at the exit 25th the synchronization clock with the frequency 8 kHz, at output 26 the sampling rate for the analog shift register with the frequency 256 kHz and am Output 27 the regenerated frame synchronization signal, which in a 30/32 channel system was determined from the signal in the thirty-second channel which was used for transmission of the frame synchronization signal is used. This signal is referred to below as ak. The signal on the line 27 is obtained from the output signal yk at the output 17 with the aid. a differential amplifier 28 subtracted. A signal therefore appears at output 29, which is referred to below as the error signal ek. A decision is made in dashed lines 30. drawn, which at its output the sign of the applied input signal supplies.

The line 31 connects the output of the circuit 30 to the input a delay element 32 of the delay time N t.

Its output 33 is connected to the first inputs of multipliers 34, 35, 36, 37 and 38 connected. The line 39 connects the output 27 with the Input of a wide delay line, which consists of delay elements 40, 41, 42, 43 and 44 is formed with the delay a time. The taps of this delay line are connected to the second inputs of the multipliers 34, 5, 36, 77 and 38. Their outputs 45, 46, 47, 48 and 49 lead to those to be described in more detail below Circuits 50, 51, 52, 53 and 54. Their output lines 55, 56, 57, 58 and 59 lead to the adjustment members 6 ', 7', 8 ', 9' and 20 and provide the necessary Information about their automatic setting.

In the following, before going closer to the apparatus structure-the in Fig. 3 shown block diagram of an adaptive equalizer, is entered that Setting criterion derived from some theoretical explanations.

Prerequisite for the derivation of a setting or failure criterion it is, / on the receiving side, estimated values for the correct signals sent or "Ideal signals" can be derived. With pure pulse amplitude dinodulation, in general only known that the signals to be transmitted are the Have the shape of certain impulses. Since there are any number of amplitude levels can, the derivation of an ideal signal is difficult. But also the construction of one Quantizer for multistage pulse amplitude modulation becomes very expensive once more than four different amplitude levels are transmitted. The transmitted message flow therefore cannot be used to derive a criterion, especially often only some of the channels are occupied with calls. That is why it is used here as an ideal signal uses the regularly recurring rocker synchronization signal that is generated by a 30/32 channel system is transmitted in the thirty-second channel and that for example from an alternating positive and negative impulse, measured on the transmitted Signal, maximum amplitude exists. This synchronizing signal sequence then has the Fundamental oscillation 4 kHz. Other frame synchronization signals are of course also possible possible. For the sake of simplicity, however, the description is intended to be used in the following limit the exact example.

The use of this signal for derivation is under the following conditions one setting criterion possible: The dispersion of the distorted liapulse response on The output of the transmission channel must be so small that the individual successive Inpulanbvorten, which as a result of the thirty-second channel transmitted pulses occur, do not overlap and that the equalization of these impulse responses with the help of a transversal filter is possible without the temporal dispersion of the equalized impulse responses assumes too large values. When using a 30/32 channel system should in each case the post-oscillation acr impulse response in the thirty-second Channel will have subsided before significant pre-oscillation or the main value of the temporal next following impulse response occur in this channel.

If these assumptions, possibly enforced by fixed pre-equalization must apply, the regenerated frame synchronization signal can be used as the ideal signal This also applies if, as in our example, the frame synchronization signal is a periodic signal. The above-mentioned restrictions on dispersion the impulse response need not be adhered to when used as a frame synchronization signal binary bipolar pseudo-random text with a sufficiently long period is used will.

The frame synchronization signal can be derived from the incoming signal sequence with the aid of a peak value detector in the assembly designated by 24 in FIG. In our example, the ideal signal sequence only has the value "+ 11t or" - 1 "in the time intervals in which the thirty-second channel is sampled, otherwise the value" O "and is called a '(t) if the signal yk at the sampling time t0 + k # is at the output deu if the equalizer filter is not yet correctly set, the following applies: in this equation are the c. the coefficients given by the setting elements, the xk-j j are the delayed consumed th signals at the j-th tap of the delay line in the tranaversal filter. At the same time applies Here h (t) is the impulse response of the transmission channel and equalizer and hk = h (to + kt). The an contain the information, i.e. indicate the height of the amplitude levels of the transmitted signal pulses in the individual channels. Equation (2) states that the signal at the equalizer output is composed of a sum of superimposed, chronologically consecutive impulse responses weighted with factors, and n, as well as J are running, integer number variables here.

The deviation of the distorted impulse response h (t) from the ideal impulse response normalized to the value 1 at the sampling times is Here ho is assigned to the main value, ie the maximum value of the impulse response, the setpoint of which is normalized to the amplitude +1.

If one now compares the actually arriving signals, which are denoted by y (t) in FIG. 4, at time t =. t0 + j7 with the ideal signal a '(t), which is shown in FIG. 4, which thus consists alternately of "+1" and "-1", followed by thirty-one zeros in each case, the result is an "error signal" of size This signal is shown in FIG. 4 and denoted by e (t).

Accordingly, only the signal in the thirty-second channel is compared with the ideal signal sequence a '(t), while the signals in the other channels which are not correlated with the synchronization symbols are regarded as "interference" 1. As can easily be shown, can the sizes can be used to set the equalizer. The factor atk; In the majority of all cases, N is equal to zero.

The sum terms that contain this factor usually provide no contribution. Only every thirty-second character gives a sum term, because here ak'-j-N J N equals 41 "or" -1 ". in. The signals in the first through thirty-first Channels, which are considered here as interference, are represented by the in, equation (1,5) made summation integrated out. K is a sufficiently large one to be chosen integer. The delay of both quantities on the right-hand side of Eq. (5) by N.1rist required in order to be able to equalize pre-oscillation, i.e. to determine #hj for j <o to be able to.

According to equation (5), the adjusting members 6 ', 7', 8 ', 9' and 20 in Fig. 3 are continuously controlled so that the distortions, i.e. the dz for the different j, go to zero. This method has the advantage that the a ' k-j-N can only assume the states "G", "+1", "-1", which is the instrumentation simplified. However, a quick multiplication is an analog by one Sign size necessary, because e (t) can assume all possible amplitude values.

Using equation (5), the remaining part of the Describe the circuit in FIG. It is initially assumed that the dashed Circuit 30 does not exist, so that line 29 connects to line 31 directly connected. The quantity ek-N can be derived from ek with the aid of the delay line 32 win. The individual quantities a1 k-j-N are atk with the help of the from the elements 40, 41, 42, 43 and 44 generated delay line formed. This can be a suitable one modified digital shift register. The multipliers 34, 35, 36, 37 and 38 form the products of ek-N and a'k-j-N required according to equation (5) The necessary summation over K products of sampled values is carried out with the aid of the Circuits 50, 51, 52, 53 and 54. If # hi is greater than zero, 4ci in the Changed direction resulting in a decrease in #hj. Istli hj smaller than Zero, there is a change in c in the other direction. This setting can simultaneously and in parallel for all adjustment elements contained in the transversal filter take place.

A further simplification is possible by using the Relationship K 4c ae k = 1 kjN. kN (6, for automatic adjustment of cj.

Sgn ek-n here means the sign of ek-N '#cj ai the size the change of C3.

The adaptive equalizer shown as a block diagram in FIG. 3 can also be constructed according to the method described by equation 6). The instrumentation can largely be done digitally, since there is only a pure sign correlation occurs. The controlled variable according to equation (6) is used as the controlled variable for c. The size ak is formed anew at the point in time when the thirty-second Channel takes place.

Circuit 30 is used to determine the sign of ek.

The sign sgn ek appears on line 31 and can now be delayed with the aid of a digital shift register 32. This will make the Instrumentation greatly simplified.

As already mentioned, this delay is necessary if, according to Equation (6) also pre-vibrations of the impulse response should be removed. Corresponding becomes a'k N in turn from atk by delaying in a suitably modified digital Shift register from members 40, 41, 42, 43 and 44 obtained. These two sizes are multiplied in multipliers 34, 35, 7, 37 and 38. This multiplication. corresponds essentially to a modulo-two addition, since only the quantities "+1", "-1" and '0 "can occur. This sign multiplication can occur with purely digital Funds are executed. For example, the circuits 34, 35, 36, 37 and 38 can be implemented directly as a modulo-two adder. The setting of c is done in the following way: the result of the product is given at the moment of arrival a new value a'k-j-N N at the corresponding modulo-two adder 37 via a Gate circuit by means of a suitable auxiliary clock in one contained in circuit 53 Flip-flop stored until the next usable one arrives Value that appears after 125 / us, because in between a'K-j-N is the same Zero. In this Pall can consist of the links 40, 41, 42, 43 and 44 Delay line can be replaced by a simple digital shift register.

The product stored in this way is in turn an integrator or Totalizer suggests -which forms the sum according to equation (6). The outcome of the Circuit 53 controls the corresponding setting member via line 58 in turn 9 '. As a result of the intermediate storage of the signed products, disc lxi tegration or summation take place slowly and precisely with known circuits, since only all 125 µs a new product occurs.

The circuit of the control logic is not described in detail here, since it does not contribute to the elucidation of the method according to the invention. the Control impulses can be from the in J? Ig. 3 with 24 designated circuit supplied will.

In practice, it is often convenient not to use equation (5) and (6) to use the given relationships, since here the summation or integration after K products must again be started and since the adjusting elements 6 ', 7', 8 ', 9' and 20 must be able to retain the attitudes attained in each case, i.e. to be saved until a new value is available to improve the setting stands.

The relationships can take place instead instead of equation (5) or -Instead of equation (6), serve to set the cj. With better-valued equalization, the 1 hj tends to approach zero, and the controlled variables described by equations (7) and (8) tend to constant values. The starting time of the summation or integration can be chosen arbitrarily. q is a sufficiently small constant proportionality factor to be chosen, which determines the speed of the running in of the.

By the influence of the signals in the first through thirty-first channels in the formation of the controlled variables given by equations (5), (6), (7) and (8) In each of these equations there is an additional "perturbation element" that leads to a temporary "emigration" of the equalizer, dvh. to a deviation from the desired Behavior can lead. ' In order to minimize the influence of disturbances according to this sum To keep it low, the integration or summation time constant is sufficient selected large, i.e. the equalizer readjusts slowly. In equations (5) and (6) If K is chosen to be sufficiently large, in equations (7) and (8), however, q is sufficient small.

The influence of disturbances that occur on the transmission paths and which can also lead to a less than optimal setting of the equalizer, as well as the influence of the signals in the other channels can thereby be reduced or eliminated that the controlled variables have to be repeatedly checked for effects are subjected to interference signals contained in the quantities to be correlated. Based on the results obtained from the pen review, over the period the occurrence of the interfering signals the setting of the adjustment elements slows down or blocked. For example, in equations (5) and (6), K can be increased, in Equation (7) or (8) can be reduced to q.

This can be done, for example, by sequential Values of the controlled variables formed in adjacent time periods and related to their Signs are examined.

The technical realization of the adjusting members 6, 7, 8, 9, 6 ', 7', 8 ', 9' and 20 in FIGS. 1 and 3 can be adjusted by switching circuits be given that with the help of digital-analog-VJandlcrn or with magnetic Stores, for example transfluxors, are implemented. Next are as adjusting members Bridge circuits are possible that linearize one through negative voltage feedback Field effect transistor included as a controllable element. In addition, as ste: ilable Elements also photoresistors can be used.

The circuit according to the invention described has the advantage that no special test signal sequence is inserted into the information flow to be transmitted must be, since a frame synchronization signal in Zeiteul-tiplex-Pulsanitutudenmodulationssystemen already exists. A pseudo random text does not have to be used as the frame synchronization signal can be used, but a periodic signal is sufficient in most -Pallen to die, which can be recovered in a particularly simple manner on the receiving side leaves. The procedures and criteria described for the adaptive setting of the Entserrars allow a particularly simple instrumentation and an accurate Converge the circuit.

9 claims 4 figures

Claims (9)

P a t e n t a n s p r ü c h e ~~~ ~~~~~~~~~~~~~~~ ~~~~~~~~~~ 5) Procedure for the automatic setting of an adaptive, the cancellation of linear distortions equalizer serving on a transmission path, in particular transversal filter equalizer, in a Zeitmul1Aplexsystem for Pu1sam1i.t, -uaenmodulation, in which in addition to the message signals a periodically recurring additional signal is transmitted at the receiving end Samples of the amplitudes of the regenerated additional signal from samples of the The amplitudes of all incoming signals are subtracted from the difference signal and the regenerated additional signal setting criteria for the setting elements es Equalizer can be derived, that is, that as Additional signal the frame synchronization signal is used. 2. Verfaliren according to claim 1, d a d u r c h g e k e n n -z e i c n e t that as Ra} imensynchronisiersigii.l alternately positive and negative, equal in amplitude and, measured on the transmitted signal, impulses with maximum Amplitude can be used. 3. The method according to claim 1 or 2, d a d u r c h g e -k e n n z e i c h n e t that the regenerated additional signal is used to form the setting criteria compared to the difference signal step by step by integer multiples of the time Distance between two successive samples is offset that products from the samples of the incrementally shifted regenerated additional signal and the Difference signals are formed that from the products of each step a sum is formed and that these sums serve as setting criteria. 4. The method according to claim 1 or 2, d a d u r c h g e -k e n n z e i c h n e t that the regenerated additional signal is used to form the setting criteria compared to the difference signal step by step by integral multiples of the time Distance between two successive samples is offset that products from the samples of the incrementally shifted regenerated additional signal and the Sign of the difference signal are formed that from the products of each step a sum is formed and that these sums serve as setting criteria. 5. The method according to any one of the preceding claims, d a -d u r c It is noted that the setting of the setting elements when they occur is slowed down or blocked by interfering signals. 6. Circuit arrangement for Bu.rchführung the method according to a of the preceding claims, d a d u r c h g e -k e n n n z e i c h n e t that as Adjustment elements are provided with step-wise adjustable circuits of digital-to-analog converters or of magnetic memories, especially xransfluxors, are realized. 7. Circuit arrangement for carrying out the method according to a claims 1 to 5, d a d u r c Ii g e k. e n n -z e i c h n e t that as adjusting elements Bridge circuits are provided, which one through voltage negative feedback. linearized field effect transistor included as a controllable element. 8. Circuit arrangement for performing the method according to a of claims 1 to 5, d a d u r c h g e k e n n -z e i. c h n e t that as controllable Elements of the adjustment photoresistors are provided. 9. Circuit arrangement for performing the method according to a of claims 1 to 5, d u r c h g e n n n z e i c h n e t that fixed equalizers are provided. Blank page