DE2911845A1 - ADJUSTMENT DEVICE CONNECTED TO A TRANSMISSION LINE FOR ELECTRICAL SIGNALS FOR COMPENSATING SIGNAL DISTORTION - Google Patents

Description

description

The invention is concerned with alignment devices for data transmission channels, in particular with automatically adapting matching devices that are used in modems for very fast data. The inventive Adjustment is particularly suitable for a microprocessor control and enables a method and a device for performing extremely accurate ultrafast matching by looking at the type of Responding to only a single transmitted test pulse is being worked on.

The balancing device according to the invention is special suitable for quick broadcasts. In such applications it is common to have a number of remote stations, for example, to address quickly one after the other in different locations from a central point. the Transmission to each remote station uses a new transmission line, so the alignment is reset must be when a remote station

is called. Therefore, it is extremely desirable to shorten the adjustment time as much as possible in order to achieve the Increase data throughput.

Many known balancing devices adjust relatively slowly when they are first used with a new transmission line connected and consume - hence valuable data throughput time. Many are time consuming incremental settings of the balancing device taps during an analysis of a relatively lanafm Period of time from random data or a sample or run-in signal sequence with numerous test pulses required.

909845/0684

■% ■

Bell System Technical Journal, Volume 50, No. 6, pages 1969-2041 suggests using a to adjust a single transmitted test pulse, but the system described there actually requires a Diversity of many test pulses.

In the US patent specification to Mr. David Motley et al, there is discussed a facility in which the trimming is achieved during the response time for two transmitted pulses using a technique in which the known, forcibly leading to zero scheme is approximated, and in which a correlation of signal samples is not is used. The adjustment in this device requires an additional pulse to adjust the phase of the Line signal so that it is properly sampled. Since this facility is a forced to zero scheme is used and only approximates it, it can be higher for a strongly distorted line or for applications Precision cannot be used. It does operate at 4800 bps in the Motley et al patent recommended, but you cannot work with a data speed of 9600 bps or more, a speed with which the adjustment device according to the invention can work without difficulty.

In IEEE Transactions On Communications, Volume Com-23, No. 6, June 1975, P. Butler et al Method that enables a direct solution of a matrix equation in real variables, where the constants of the Matcher setting in a one-sided tape system will be described. This procedure requires a much longer run-in sequence than for the subject of the invention, to achieve a reasonable accuracy estimate. In addition, the method does not solve a matrix equation with complex variables. Therefore, this concept cannot

909845/0684

can be used to achieve alignment in a two-side-band system, but this is what the invention does.

The invention is therefore based on the object of a balancing device for data transmission systems that work faster and more accurately. The invention is further intended create a practical matching device that allows the analysis of only a single transmitted pulse to the initial Adjustment of the taps of the balancing device for a large range of different and fluctuating Line interference and data speeds requires. Furthermore, the invention is intended to provide a balancing device, in which the constants for the adjustment device can be calculated exactly instead of just approximated. Next should With the invention, an exact calculation by means of an iterative method can be made possible during the Period during which a transmitted pulse is received, so that the constants for the Equalizer tap is calculated and this is calculated during the interval between the end of the pulse time and the time when the first received data reach the main tap of the matching device, are set. Further The aim of the invention is to provide a balancing device which allows an exact initial setting within a very rapid period of time Time, for example 30 milliseconds for a data rate of 9600 bps and 15 to 20 milliseconds for a data speed of 4800 bps.

The invention achieves these objects and further advantages in that one of a transmitter Via a transmission device transmitted signal is analyzed. The signal can, for example, a single comprise received impulse or a single received impulse response. Forms the from the received signal Setup in-phase and 90 ° out of phase

909845/0684

Pulse signals. The in-phase and ura 90 ° phase-shifted Pulse signals * can be in the passband or baseband frequency or in another translated frequency. Sampled values from these two impulse response signals are used to calculate a complex matrix equation define. A complex matrix of that equation has elements that result from the auto and transverse relations of the two impulse response signals sampled values are formed. The scanning device according to the invention solves the complex matrix equation thus formed and receives the exact values for the initial setting of all taps the scanning device. The initial scanner settings fully and fully compensate the distortion of the transmission medium over which the sample signal was transmitted. After the initial Adjustment occurs with the adjustment device according to the invention, a more adapted adjustment during data reception on.

The device according to the invention has the main advantage Possibility of an exact calculation of equation constants from the temporal samples in the order of magnitude the duration of a single received pulse despite the wide range Variations in the distortion caused by the medium over which the received pulse was transmitted was introduced. In carrying out the invention, an adjustment device was built with which an initial adjustment in a period of time on the order of 30 milliseconds at a data rate of 9600 bits per second. could be achieved. This is nearly five times faster than the match that is currently on the market Facilities can be reached. In the preferred embodiment of the invention, the adjustment is executed in response to a single received pulse,

* English: the in-phase and quadrature phase impulse response signals

909845/0684

and an important feature of the preferred, erfindungsgeraäßen device consists in the SCIU ₁ 1fung of devices that determines the optimal points prior to the reception, where the received pulse is sampled. While the adjustment is carried out with in-phase and 90 ° phase-shifted received pulse signals at baseband frequencies in the preferred embodiment, such signals can be derived at frequencies of the passband or at other corresponding translated frequencies.

The invention is described below using a preferred exemplary embodiment with reference to the accompanying drawings described in detail. Show it:

Fig. 1A shows the received signal from the preferred balancing device according to the invention is used;

Fig. 1B phase-shifted components resulting from the and 1C received signal according to the time scale can be generated according to FIG. 1A;

Figure 2 is a schematic diagram illustrating the structure and operation of the preferred balancing device according to the invention;

Fig. 3 is a schematic diagram for explaining that in the preferred embodiment the digital processor used in the invention;

Fig. 4 is a flow chart illustrating the overall structure and operation of the a correct timing of the sampling of the received signal in the ■ preferred embodiment of the invention allow;

Fig. 5 is a detailed flow chart for

Explanation of the procedure and the facility used to determine de, s Presence of a received signal can be used in the preferred embodiment of the invention;

909845/0684

BATH ORIGINAL

Fig. 6 is a detailed flow diagram for illustration of the method and apparatus used in the preferred embodiment of the invention used for the correct setting of the sampling points on the received signal will;

Figure 7 is a detailed flow chart illustrating the method and preferred ones device according to the invention for determining the constant settings on Matching device tap from samples obtained on the received signal;

Fig. 8 is a continuation of the flow chart of Fig. 7;

9 shows a method with which phase quadrature

signals can be generated for use in the device according to the invention;

10 shows an alternative method for generating the phase quadrature signals; and

11 shows a further embodiment of the method for generating the phase quadrature signals for use in the context of the invention Furnishings.

The automatically / adapting adjustment device according to the invention is best explained in conjunction with the sample signal sequence used to set the automatic Gain (automatic gain control), timing and adjustment according to FIG. 1 is used. The signal sample according to Fig. 1 is the analog demodulated signal sequence at the receiver after transmission. The preferred facility refers in particular to quadrature amplitude modulation (QAM), or tetravalent amplitude modulation.

The "run-in" signal sequence (sample signal sequence, control signal sequence) consecutively has a number of bauds only of the pure carrier 11; a number of bauds of the pure clock 13,

a rest or squelch period 15, a receive pulse and a further squelch period 19. The "run-in" signal sequence can also be a fine-tuning sequence if desired

909845/0684

include before the transmission of the useful data 20.

In the preferred embodiment, eight bauds of the pure Carrier 11 and 17 bauds of the pure clock 13 sent. The squelch periods 15, 19 are twenty-seven and twenty-seven, respectively. twenty-one bauds long and the pulse period is one baud long. Each baud period is 416.7 milliseconds. Of course, other baud periods could also be used. Also depends on the length of periods 11, 17, 15 and 19 depends on the maximum distortion. If the line distortion is less, there will be fewer bauds for each of them, especially for periods 17, 15 and 19, and the total time for the adjustment will be shorter. If the Line distortion is high, more bauds are required for each period and the total alignment time becomes longer.

During the period 13 when only the carrier is there (Fig. 1), the initial incidence of the carrier energy is on determined by the management (carrier determination) and the operation of the facility begins. When a carrier is noted a coarse time counter KSTMX is started, which finally announces the occurrence of the received pulse 17. KSTMX counts once per baud.

The automatic gain control is then carried out on the carrier signal 11. After a fixed number of Bauds of the carrier signal 11 has been detected, the system knows that it is actually an "in" signal sequence receives and can then expect the pure clock signal 13.

The device checks during the pure clock pulse sequence 13 the transmitted pulse train to determine the optimal points where the incoming received pulse 17 is sampled shall be. The facility then sets up the sampling procedure for the received pulse. The first step is to

909845/0684

/ lh

to use the previously calculated sample point and jump the sample clock to the optimal sample position leave or pre-set. So it can be seen that the preferred embodiment of the invention is actually the Reception scans for a received pulse, although other devices for generating signals also use the representative of such samples could be used.

During the squelch period 15, the coarse time counter KSTMX, which is triggered when the pure carrier signal is detected, is set was to continue its count in order to then indicate to the device the point in time at which the sampling of the received pulses (Fig. 1B, C) in phase and out of phase by 90 ° must be started. The scans are then taken, stored and correlated to form a matrix. After the matrix is formed, a special one is made Iteration technique to determine the exact initial tap settings used by the balancer as discussed below.

The device used to perform these operations is conceptually shown in the block diagram of FIG shown. The device has an analog-to-digital converter (A / D) and an automatic gain control (AGC) 21, a digital processor 23 and a transversal adjustment device 28. The matching device 28 is for explanation only shown in principle and is expediently carried out digitally, in which case it also as Part of the digital processor 23 could be represented.

The digital demodulation of the QAM signal into in-phase (X) etc. (Y) baseband components displaced by 90 ° is expediently executed in the digital processor 23. The analog form the demodulated in-phase and out-of-phase baseband components are shown in Figures 1B and 1C, respectively. the

909845/0684

2311845 • AS-

Demodulation can again be carried out by methods known per se, and can expediently by a particular the apparatus provided for this purpose can be carried out outside the digital processor 23. Neither the automatic Gain control (fading compensation) nor the type of demodulation used form part of the Invention.

The X and Y phase components generated by the demodulation represent digital samples of the baseband signal, wherein the Y component of the sample is demodulated by a carrier to which the X component sample is supplied Carrier is 90 ° out of phase. In order to calculate the correct sampling time during the pure clock period, the preferred embodiment two samples taken per baud. After the optimal clock phase is set, the device takes one sample per baud.

The samples X, Y are fed to separate channels 25, 27 of the transversal adjustment device 28. Each channel 25, 27 has equally spaced digital delay elements 29, 31 and digital multipliers 30, 32, 33 and 35. The multipliers 32, 33 multiply the delayed samples X _m by the constants Cp ₁ and Cq ₁ , and the multipliers 30, 35 multiply the delayed samples Y _m by the constants cp. and cq. The output of a summing network 46 is subtracted from that of another summing network 42 in a summer 37 and results in the output data signal EQX. The outputs of the other two summing networks 40, 44 are summed by a summing network, which results in the output data signal EQY.

909845/0684

As shown in further detail in FIG. 3, the preferred embodiment of the invention is programmed Microprocessor structure and an adjustment unit. The adjustment unit 34 includes the functions of Adjustment device 28 from FIG. 2 and carries out a continuously adapted adjustment according to US Pat. No. 4,035, for example the end. Contains in the preferred structure according to the invention the trimming unit 34 also has a circuit that performs the initial setting of the trimming device, followed by will be discussed below.

The microprocessor structure shown in Fig. 3 has a program memory 16, a program counter 18 for Addressing the program memory 16, an instruction decoder for decoding instructions from the program memory 16 and generating control signals and an arithmetic unit 22, which the commands as a function of the Control signals from the decoder 14 executes. The microprocessor structure furthermore has a data memory 26 and an address decoder 24 for addressing the memory. The program memory 16 is a read-only memory (ROM) of sufficient capacity and stores the instructions for execution the adjustment operations to be described below and can, for example, be built up from four AMD 9216 ROM chips be. The program counter 18 can, as required, depending on the control signals from the command decoder loaded or jump. The arithmetic unit 22 performs the necessary operations which will be explained below and of course has sufficient capacity for this. The data memory 26 has a memory for constants and 256 words of random access memory (RAM) and can consist of three AM91L12ADC RAM chips and one General Instrument's R03-5120 chip. The random access memory is used to store incoming Samples from the receive pulse 17 and subsequent data 20 while the calculations are carried out will·

909845/0684

'Al-

The device indicated in FIG. 3 carries out a quick initial adjustment by adding the multiplier constants (Factors) for the initial taps of the matching device can be calculated in a very rapid manner. The type of calculation and the function and structure of the device from FIG. 3 are explained in detail below.

The multiplication constants (factors) to be calculated are denoted by c .., c ~, c ₃ ··· c. as well as c .., c? / c 2 '"' ⁰ ^ ( ^{V <} ? 1 * Fig. 2).
In complex notation, the constants for the balancing device taps can be written down in the form:

c _i = c _pi + jc _qi i = 1, 2, 3, ... η

To calculate the adjustment constants c. demodulated from the pair Received pulses according to FIGS. 1B and 1C are is based on the following definitions:

Σ (χ ₊ γ)

Σ »" ¹ ₍ x _m x _{m + 1 +} Y _n Y _{n + 1} ) ₊ i Σ »(X _1n Y _{n + 1} - Y _n X _{n + 1} , ₍₂ ,

m = l _{m = 1}

909845/0684

The following elements are defined:

\ = ^X nqk ~ ^ ^Y nqk k = l, 2 ..n (5)

^rT 0

rT r. = —- 1 = 1,2 ..n-1 (6)

* ^rT 0

In the above equations (1) - (6), X and Y are the m th samples of the in-phase and quadrature phases, respectively of the received pulse, which is used to calculate the Autocorrelation and cross-correlation is used. Equations (2), (3), (4) represent the autocorrelation and represents the cross-correlation of samples X and Y.

m m

In the above equations (1) - (6), M represents the total number of samples used for calculating the autocorrelation and the cross-correlation can be used, and η denotes the number of taps (in the preferred embodiment of the Invention, M is chosen to be 20 and η to 16), and q means the appendix of the first sample actually used, which for Calculation of h, is used. The variable q takes into account the fact that in the preferred embodiment of the invention and not all samples taken are used, i.e. that η is less than M, as explained below will. If η = M, then q = 1.

With these definitions, the _{equations defining the optimal tap constants C 1} , c- ... c for a balancing device of η taps can be written down in the following matrix form:

909845/0684

- ψ - ./19.

^r 2 ^r l

r -. n-1

V ^r nl

V ^r n-2

1 r * ... r *

^{X r} l ^r n-3

r 1 r * τ- *

^r l ^{l r} l ^r n-4

"" Γ ^h 2 H
η ^C l ^C 2 C.
η

(7)

In equation (7), r means ..... r. as well as h .... h complex constants and C ₁ ... c are complex vari; indicates the complex conjugate form.

stants and C ₁ ... c are complex variables. The star (*)

According to the invention, a special solution to this equation (7) allows an exact iterative calculation of the tap constants c .... c- within the time interval required for the entry scheme from FIG. 1, plus the time delay required for the data to pass through from the input to the output of the balancing device is required. According to this solution, the following definitions are made:

^e 1 =

- H ²

/ r .. / means the absolute value of the complex quantity r ₁ ;

- -r

The exponent fD "indicates the first iteration i = 1;

(8th)

(9)

(10)

909845/0684

With these definitions, the exact iterative solution for the tap constants results as follows:

' ¹ Vi - Ji «."'"W'V - ^{(r +} Ji ⁸ J" W ⁽ V

(16)

Again, the exponent indicates the value of the variable for a particular iteration. Create equations (7) - (16) a simple way of quickly and precisely calculating the tap constants c in the complex matrix equation (7). This iterative method enables the device according to the invention to use the constants c. to calculate and to adjust the adjustment device so that the initial adjustment occurs during a total "run-in" time of about 30 milliseconds from the beginning of the pure carrier to the first bit of the user data in a 2400 baud machine can be reached. Variations of the matrix equation (7) can be written down and replaced by the above Technology can be solved without deviating from the inventive concept.

The construction and operation of the device, which is explained in Fig. 3, will now, as far as it relates to the ■ preferred embodiment of the invention, described in more detail in connection with Figs. 4-8.

909845/0684

. 84 ·

After the carrier has been determined and the clock KSTMS has started, the facility begins accordingly to operate in accordance with the flow chart of FIG. The flow chart of Figure 4 illustrates the automatic gain control function (the function of the fading compensation), the filtering, the demodulation and the detection of the presence of the Retraction pulse train (TPP = training pattern present) and the calculation of the in the following equations according to Fig. 7 optimal sampling point to be used. Two samples of the pure carrier signal are processed, each Baud for eight bauds of pure carrier signal 11 · A counter N is set to -8 and determines the operation.

As long as N is less than zero, the left branch 45 of the flow chart is traversed and each sample becomes one automatic gain control 47, a filtration and subjected to demodulation 49 and detection measures relating to the presence of the entry pulse train. The test pulse train detection checks six consecutive bauds of the pure carrier signal and then sets a mark indicating that a test signal sequence is actually being received. At each baud the counter N as well as the counter KSTMX advanced by one unit.

When N equals zero, the pure clock signal 13 begins. The AGC is frozen and it becomes now entered right branch 53 of the flow chart of FIG. In this right branch 53 a filter and Demodulatxonsoperatxon 55 carried out and a test 57 of the TPP marking carried out. When TPP is established has been and the TPP flag is set, the rapid retract stroke operation 59 is performed. During this Operation, which is denoted by FLCLK = fast learn clock, the device calculates the optimal sampling point for the incoming pulse, based on the demodulated pure clock information. After demodulating two samples per baud

909845/0684

and have been used in the FLCLK process, the counter KSTMS as well as the counter N are incremented by one unit. When FLCLK is performed, the flow chart shown in FIG. 7 is entered.

The manner in which the entry beacon determination is performed is shown in detail in FIG. There the X and Y samples of the demodulated baseband signal are sent to the respective input at one speed of two probes per baud.

The following work is then carried out on the samples fed to the X input. Each sample is first passed through the multiplier 63 squared and the output of the multiplier 63 is for a sampling time in a delay element 65 is stored. The present one

Exit/

of multiplier 63 becomes the negative value of the previous output of multiplier 63 in FIG an adder 67 is added. The output of the adder 67 is fed as an input to a second adder 69. The X input is also fed to a second delay element 71, which delays by a sample time. The output of the second delay element 71 is sent to a second multiplier 73 together with the X input so that the current input X sample is multiplied by the immediately preceding X sample. The output of the second multiplier 73 is fed to an input of a third summing network 75.

Work is carried out on the Y input in a similar way. A delay element 77 delays the first sample of the Y input and a multiplier 79 multiplies the first sample of the Y input by the delayed sample and outputs this to the third summing network 75. The Y input is also squared and the Y input value is squared a delay element 81 is supplied. The delayed

909845/0684

. as.

The squared sample is subtracted from a currently present squared sample by a summing network 83, whose Output is fed to the second summing network 69.

The output of the first summing network 73 is multiplied by two by a multiplier 85 and forms an output labeled ACK. The output of the second summing network 79 is labeled BCK. The arcus tangens of ACK / BCK is then formed to determine the sample angle θ _η . The current value of θ is stored by a delay element 87. The stored value of θ is in the clock advanceii
will be discussed below.

of θ is used in the measure preset to which

It is now determined whether θ is within the limits for any number of counts NTPP. If NTPP is greater than 10, five sample bauds have been checked. Thus, when each -9o / for more than 10 STPP counts less than 15 • η '

it is confirmed that the carrier has been received for 5 bauds and that the TPP flag has therefore been set to 1 is. This mode of operation is shown in FIG. 5 when the process continues through blocks 91, 92, 93 to block 94 becomes where TPP = 1. If / θ -9 Oj is greater than 15 and TPP = 1, then the clock preset routine occurred.

However, in the event that θ _η ~ 90 | is greater than 15 for one of the processed pure carrier samples, the test block 95, namely TPP = 1, is not fulfilled and NTPP is reset to zero. In this case, if KSTMX is greater than 19, indicating that 19 bauds have occurred without TPP being detected, then the absence of TPP is indicated. Failure to detect TPP usually indicates line failure.

909845/0684

When the run-in pulse sequence has been determined, it is necessary to align the timing of the sampling of the received pulse 17 correctly. The rehearsal points are calculated so that the balancer can best minimize the output error. The procedure and the structure of a preferred device for carrying out the presetting of the sample cycle (FLCLK) is detailed shown in FIG. 6.

Assuming that there is no distortion or noise, then the difference between consecutive Angles θ be 180 °. Therefore, if the size

K ~ ^θ η-1 l ¹⁸⁰ '

less than 12 ° for several consecutive samples then the clock is in a good range.

If θ extends over several sampling intervals of the pure η

If the clock pulse train is in a good range, then the loop consisting of blocks 101, 102, 103, 104, 105, 106 and 107 from FIG. 6 comes into play. At the beginning the three counters NCNT, NSAMP and NAV are set to zero. When the first sample is tested, the sample counter NSAMP is incremented by 1, as indicated in block 101. The angle φ is then checked, and if it is within the range, the counter NCNT is incremented by 1. After four consecutive good samples, the test to determine whether NCNT is greater than or equal to four (block 104) is positive and NAV = 0. In this case, the test indicated by block 106 is performed to determine whether the magnitude of θ is less than 90 °. If so, the counter NR is set to 1. The counter NAV, which represents the number to be averaged, is set to 1 and TAGL (the total angle) is defined to be equal to θ at this moment. The next time it is run through the loop, the test NAV = falls

909845/0684

2311845

. as

not positive and NR is advanced by one unit (block 108). NR + 1 is then two. NR is then not odd and another sample is taken. After this test, NR is odd (= 3) if φ is still within the bounds. Therefore the average number is NAV + 1 = 2 and TAGL is equal to the previous θ value plus the new θ value. Therefore, two angles must be averaged or balanced. Assuming that θ still remains within the bounds, then the number of averaged θ samples is increased to four, and then the angle P _fi at block is determined by calculating the quotient of TAGL and NAV. P _ft indicates the number of degrees by which the sample point used differs from the optimal sample point. Therefore, 10 angle differences are taken within the limits in order to achieve the block P_ = TAGL / NAV.

However, if distortion occurs, other measures are taken to calculate P_. For example, if the ratios are such that φ is greater than and thus block 102 is satisfied, a test 110 is carried out to determine whether the number of good samples NCNT counted is greater than or equal to 4, ie whether four angle test outputs are within the limits appeared. If so, a test 111 is made by the NAV counter to determine whether any TAGL samples have been stored for averaging. If any samples have been stored, the mean θ = TAGL / NAVL is calculated as indicated by the four av

Blocks 112, 113, 114, 115 is indicated. These four blocks indicate that P _{Q is taken} equal to θ,

ö av

if NR is even, whereas P _Q equals [ ^SGN < ^e av>] [ ¹⁸⁰ ° "j ^e av {]

909845/068 4

is taken when NR is odd. But when on test 111 NAV is found equal to zero, indicating that no samples θ have been collected, then the present sample θ is taken for P_.

If the NCNT > _ 4 test 110 fails, a test 117 of the number of samples, indicated by counter KSTMX, is carried out. If the number KSTMX is greater than (> 14 bauds), V is taken back to θ. If NCNT > A

ö η -

is not fulfilled and KSTMX _> 29 is also not fulfilled then NCNT is set to zero and another sample is tested. This procedure ensures that then, if the angle determination is at the beginning or occasionally outside the limits, the following angles can be checked in order to average out the clock according to the previously explained procedure.

Pq then determines the phase shift or the pulse sample rate, which is to be used in the matrix test according to FIG.

According to FIG. 7, the first run of the test 121 for the new program (NP) will result in a positive result and the left one Branch 123 of the flow chart is traversed. Here, a random access memory (RAM) 26 becomes the matching device deferred. Furthermore, the optimal sampling point determined by FLCLK is used to determine the sample clock to bring it to the optimal sample location within each baud. The clock speed is reduced to half 2400 Hz so that a sample per baud of the received pulse is taken from each of the X and Y channels.

During the second pass through the flow chart according to FIG. 1, a second branch 125 is passed through. Detected samples are demodulated (block 127) and then a test is performed

909845/0684

executed on KSTMX to see if it is greater than 45. If not, KSTMX is incremented according to block 131. As soon as KSTMX is greater than 45, matrix formation 133 begins from pulse 17 subjected to the sample.

The branch that is run through at KSTMX ^ _45 includes a demodulated output energy test, which ensures that the device receives the calibration test signal sequence and not user data. After KSTMX is greater than 38, the energy is determined according to blocks 134, 138. If the energy remains below a set amplitude E _ref , it is assumed that the squelch period has been detected and the device then knows that a run-in signal sequence is present. If the energy is higher than the amplitude E _f , then it is indicated that the lead-in pulse train TPP is not present. This results in a double check for the presence of the run-in pulse train.

The scanning of the pulse shape 17 is indicated by the vertical lines in FIGS. 1B and 1C. The number of samples is counted by a counter K which starts when KSTMX = 45. Each sample generates an X component x. and a Y component y .. If the samples are x., y. are taken successively, the formation of the matrix equation (7) begins in accordance with the definition equations (1), (2), (3), (4) given above. For example, X ₁ and V _{1 are} picked up and stored in RAM 25 during the first baud

2 2 and can then be used to compute X ₁ + y * , the first iteration of rT _{Q /} according to equation (1). During the second and subsequent samples, the iterations of rT _Q and the correlated equations TT ₁ , rT ₂ ... are calculated.

When the second sample x_, y_ is taken, it is in

9

stored in RAM 26 and the amplitude square X ₀ + y is compared with the first amplitude square χ ^ + γ ^ ,

909845/0684

2311845

to determine which is larger. The larger one is kept and compared with the size square of the next sample in order to determine the largest sample and thus the peak 180 of the received pulse 17 examined. The baud KP during which the peak 180 occurs is stored for use in subsequent operations. All samples x ^, y ^ are also saved.

The execution of the samples is terminated at one of two conditions indicated by test 137 (Fig. 7) will. In the preferred embodiment of the In the invention, it is advantageous to use eight samples before the tip and eleven samples after the tip. If 11 samples after the peak occurred, then K = KP + 11 and matrix formation is terminated. If K <8, then KP becomes set to eight (blocks 135, 136) so that at least nineteen samples must be taken before formation can be terminated at K - KP + 11. Otherwise the matrix formation is terminated when a total of 24 samples have been taken and a mark is then placed.

If a matrix flag is set, a branch 132 is made on the next pass through the flowchart Test 141, K> 20 (Fig. 8) occur. If more than 20 samples have been taken, test 141 is positive, and the processor 23 goes on to correct the influence of taking too many samples on the matrix.

Thus, because of the coarse alignment of the counter KSTMX, K> 20 indicates that too many samples were taken before the peak 180 appeared. These samples are likely to lead to inaccuracies and their effect is subtracted by an operation 143, labeled SUB 1. This subtraction is carried out by taking the first samples X ₁ , Y ₁ from memory, you

909845/0684

~ ίΓ / 2311845 .89-

Influence on the values for equations (1), (2), (3) for rT _n , rT ₁ , rT n, etc. are calculated and this influence is subtracted. After the effect of the first sample X ₁ , Y _{1 has been} subtracted, the sample counter K is decremented by one unit (block 145) and the K> 20 test 141 is carried out again. If the test 145 does not result in a positive result, the influence of the second pair of samples X ₂ , y _{2 is} calculated and subtracted, etc., until K <_20. If K is £ 20, those of the remaining samples will be χ. , y ^ certain values are used in the following matrix calculations.

When K is reduced to 20, branch 147 occurs which is used to calculate the tap constants, equations (5) - (6) and (8) - (16), first running a test 149 to determine if the computation is already has been executed. At the beginning of the calculation, a counter N is set to zero. A test 151 of the value N becomes then executed.

The first step 153 in the calculation with N = zero is a normalizing process (normalizing process). During this Step the r. and the h i of equations (4) and (5) are calculated by the microprocessor of FIG.

On the next pass through branch 147, where N = 1 (block 155), the microprocessor develops the trim constants c. calculated by taking successive iterations of equations (11), (12), (13), (14), (15), (16) as explained above.

If N = 2, the microprocessor of FIG. 3 begins to interact with the matching unit 34 in the following manner. The microprocessor calculates equations (11) and (12) and then transfers the "r" matrix (equation 7) and other interim calculation results for the reconciliation unit

909845/0684

2S11845. 30-

In this way, the processor shifts some of the computational responsibility to the reconciliation unit 34 to give the processor a free hand to process other operations on the incoming data. The matching unit 34 contains wired logic that performs the subsequent iterations of equations (11) - (14) or calculated. When N = 2, the matcher only calculates equations (13) and (14). At the end of every calculation an iteration of equations (11) through (14) in FIG Adjustment unit 34 (13 and 14 only for N = 2), the processor calculates the quantity z. + 1, equations (15) and (16) and feeds this value to the adjustment device 34 to execute the next iteration of equations (11) to (14) to. This assignment of the calculation between the microprocessor and the adjustment unit is only due to the effort of one efficient use of the facility. It can be seen that the assignment of the calculations of equations (11) to (14) one possibility for a circuit schematically assigned to the adjustment unit 34 is the instantaneous Calculate matching device settings. Other options, such as using a more powerful one Microprocessor, which can perform all calculations, are of course also within the scope of the invention.

When N = 15 the matrix for the tap constants is c, solved and a flag indicating the execution of the iteration is set. If the last tap constants are calculated, the h, and the last are saved Matcher constants c. Determined according to the method just described set.

The operation just explained is sufficient, 16 taps to put. If the line signal is of such poor quality that additional taps are needed, fine-tuning can be carried out in which

909845/0684

additional bauds of known two-phase data are sent and the error difference is determined and used for setting additional taps are used in accordance with standard practice.

As indicated several times in the preceding discussion, modifications and adaptations are numerous the preferred embodiment of the invention possible without thereby deviating from the inventive concept.

For example, those shown in FIGS. 9, 10 and 11 can be used in-phase and quadrature-phase signals used by the adjustment device according to the invention, at frequencies other than baseband and in systems using different demodulation schemes use.

9 shows a simple quadrature demodulation method, in which baseband signals x (t) and y (t) represent the in-phase and quadrature-phase signals. In Fig. the received signal is fed to first and second mixers 203 and 204 on an input line 201, in which the line signal is mixed with the respective signals cosu> t and - sinto t, where (O is the carrier frequency

C C

is. The components are then filtered by respective baseband filters 207, 209 and result in the baseband quadrature components x (t) and y (t).

According to Fig. 10, the received signal is a first band filter 211 supplied, which has an impulse response characteristic h (t), and supplied to a second bandpass filter 213, whose Impulse response characteristic h (t) is what the Hubert transformation is the impulse response characteristic h (t) of the first filter 211. The respective outputs h (t), h (t) filters 211 and 213 are at the passband frequency

909845/0684

and form the in-phase and quadrature-phase signals that are sent by the calibration device according to the invention for sampling can be used.

In FIG. 11, an output h (t) of the filter 211 is fed to a first mixer 215 and a third mixer 219. The output h (t) of the filter 213 is fed to a second mixer 217 and a fourth mixer 221. The four mixers 215, 217, 219, 221 each take the second inputs of cos «) t, sinct; t, sinu) t, cosu / t, where W

Angela C

again is the carrier frequency. The outputs of the first and second mixers 215, 217 are then provided by a summing network 218 are summed and result in the demodulated baseband signal x (t). The output of the fourth mixer 221 becomes is subtracted from the output of the third mixer 219 at a summing network 220 and results in the demodulated baseband signal y (t). Referring to Fig. 11, x (t) and y (t) are the in-phase and quadrature-phase signals that are also made in accordance with the invention can be used for sampling to obtain the initial setting of the balancer taps.

It can thus be seen that the invention can also be carried out under other conditions than those shown, without thereby deviating from the invention.

Overall, an automatically adapting adjustment device has been described, which with a test signal sequence works, to which a pure carrier signal period, a pure clock signal period and a single test pulse belong. The equalizer uses a transverse filter that is under the control of a microprocessor. Samples of the in-phase and quadrature phase components of the received pulse are cross-correlated and autocorrelated, so that one obtains elements of a complex matrix equation representing the optimal settings of the balancer tap describes. The microprocessor performs a special

909845/0684

iterative operation using the elements of this equation so that the optimal initial settings for the tap constants can be calculated quickly and precisely. The pure clock section of the test signal sequence is analyzed so that the sampling rate can be set exactly to the correct sampling at the receiving pulse, so that the tap constant calculation within the duration of the received pulse can be carried out. The initial alignment can take place in the Of the order of 30 milliseconds at a data rate of 9600 bits per second.

The signals or signal components referred to as "in phase" in the above description are in phase Signals, and the signals referred to as "quadrature phase signals" are those which are in phase angle are shifted by 90 °.

In addition, the copy of the prior application and the drawings are independent means of disclosure of the invention to be viewed in such a way that, for example, only features which can be inferred from these belong to the disclosed invention.

909845/0684

. -34-

Blank page

Claims (22)

RACAL-MILGO, INC., 8600 N.W. 41st Street, Miami, Florida 33166, (V. St. A.) Compensating device that can be connected to a transmission line for electrical signals to compensate for signal distortions 1. To a transmission line for electrical signals connectable device to compensate for distortions in the transmitted signal, in which the temporal The course of at least one received signal is scanned and one with the samples obtained from the scanning Correction device is applied, which the extent of the corrections to be made to the received signal certainly. 2. Device according to claim 1, characterized in that that a detector is provided for determining a pure carrier signal in an incoming signal sequence, the one signal transmitter during the presence of the HZ / il pure carrier signal controls in such a way that from the samples taken from the received signal a sequence of arcus tangens values are formed and fed to a test device, which the arcus tangens values for Checks the presence determination of the pure carrier signal. 3. Device according to claim 1 or 2, characterized in that it has several adjustable taps for compensation of interference in the received signal and a device for generating several signal samples that correspond to the first and second signals from the received pulse, which forms elements of a matrix equation from the samples, the exact values of the optimal Settings of the taps for the elements are calculated iteratively and the taps are set to the optimal values, where the samples are both autocorrelated and cross-correlated and the matrix equation is complex constants and contains complex variables. 4. Device according to one of the preceding claims, characterized in that the successive Samples are derived from a test pulse and the exact values are calculated over a period of time, which is shorter than the duration of the test pulse received. 5. Device according to claim 3 or 4, characterized in that the first and second signals are in phase and 90 ° phase shifted (quadrature phase) components of the received pulse. 6. Device according to one of the preceding claims, characterized in that a clock device for successive Sampling is provided from the received signal; that a phase correction circuit for the Clock device is provided, which a device 909845/0684 to generate successive arcus tangens values of the samples taken, a computing unit for calculating the phase error from the arcus tangens values and a Contains phase adjustment device for correcting the phase error. 7. Device according to claim 6, characterized in that the computing unit determines whether the respective arcus tangent values lie within a defined range for several sampling intervals and are dependent on the determination of the arcus-tangens values determined within the value range for the formation of the phase error averages. 8. Device according to claim 6 or 7, characterized in that the computing unit only during the reception of a pure clock signal works. 9. Device according to one of the preceding claims, characterized by a transmission channel connection, by several adjustable taps for equalization, by a detector device that the presence of a Detects the inflow pulse sequence and triggers a sample count; by a responsive to the run-in pulse train Sample count setting means; by a signal generator for the formation of in-phase and 90 ° phase-shifted Signals from the received pulse; by a scanning device for the in-phase and by 90 ° phase-shifted signals according to the set Sample counting, forming multiple samples; by a memory for the samples, by a correlation circuit for the samples, by an evaluation circuit for the correlated samples in a group iterative Calculations whereby the exact values of the optimal settings for the taps are obtained, as well as by a setting device for the taps to the optimal values. 909845/0684 10. Device according to one of the preceding claims, characterized by a receiving device for a first signal transmitted over the transmission line; by a signal generator to generate several second Signals representing samples of in-phase and 90 ° out of phase received pulse signals; as well as through an evaluation device which autocorrelates and cross-correlates the second signals and elements of a matrix equation forms from the auto-correlated and cross-correlated signals, the elements being complex constants and contain complex variables, the latter defining optimal settings for the taps, and which the iteratively calculate exact values of the optimal settings using the constants. 11. Device according to one of the preceding claims, characterized in that the second signals from the first received signal can be derived. 12. Device according to one of the preceding claims, characterized in that the device for formation of the in-phase and 90 ° phase-shifted receive pulse signals are in-phase and 90 ° phase-shifted Samples signals at successive locations to form the second signals, and that the successive Points at which the first signal is sampled are determined as a function of the incoming signal sequence will. 13. Device according to one of the preceding claims, characterized in that the part of the first received signal which is used to generate the second signals is a single receive pulse. 909845/0884 14. Device according to one of the preceding claims, characterized in that the received signal is amplitude-modulated is. 15. Device according to one of the preceding claims, characterized in that the received signal is a contains single transmitted pulse and that the second signals by scanning the in-phase and by 90 ° phase-shifted components of the received signal are generated. 16. Device according to one of the preceding claims, characterized in that the detector is the second signal as close as possible to the top of the received signal and the influence of other second signals on the Elements subtracted taking into account their location. 17. Device according to one of the preceding claims, characterized in that the detector indicates the presence a run-in signal sequence determines and as a function from the determination of the sampling points in the received pulse certainly. 18. Device according to one of the preceding claims, characterized in that a clock generator for determination the samples as well as a device for determining the phase angle errors of the clock from the samples, generated by the scanning during the transmission of the lead-in pulse train, as well as a phase adjustment device are provided for the clock, which compensates for the phase error. 19. Device according to claim 18, characterized in that the device for determining the phase error consecutive arcus tangens values resulting from consecutive Samples obtained are checked. 909845/0684 20. Device according to claim 19, characterized in that the testing device determines whether the arcus tangent value lies in a defined range for several intervals of the pure clock signal. 21. Device according to one of the preceding claims, characterized in that successive arcus
tangent values for determining the phase angle error
averaged when the arcus tangens value is within the predetermined range for several intervals. 22. Device according to one of the preceding claims, characterized in that the phase angle error is determined if the arcus tangens is not within the defined area for multiple intervals. 909845/0684