DE2637977A1 - PROCEDURE FOR CHECKING PATTERNS OF MATERIAL PROVIDED WITH PERIODICALLY RECURRING PATTERNS FOR ERRORS - Google Patents

Description

G.A.O. GESELLSCHAFT FÜR AUTOMATION UND ORGANIZATION MBH

Method for testing material webs provided with periodically recurring patterns for mistakes

The invention relates to a method for testing web materials provided with periodically recurring patterns for errors in which the web area to be tested is scanned in parallel tracks.

From DT-PS 2 322 803 it is known to paper webs, which consist of two or more with respect to the sequence of areas different Competing railways are transparent about checking for errors. In this process, especially with the watermark provided banknote tracks are to be checked, there are always two feature tracks running parallel in the transport direction each with a photocell row extending transversely to the direction of travel of the paper web is scanned and the output signals those two photocells that are assigned to homologous locations in the two feature tracks, by means of a differential circuit compared to each other. In the case of error-free feature trajectories, the result is that the compared with one another The competing feature sequences should have the same transparency, a difference signal that has a certain threshold value does not exceed. If the difference signal rises above the specified threshold value limits, this is a sign that that there is an error in one of the two compared feature paths.

809809/0084 ./.

- Ir-

It turns out to be disadvantageous that the testing of a periodic pattern present only once in the paper web is not possible, since two parallel trajectories are compared with each other.

The object of the invention is therefore to provide a device that is economically adapted to various quality requirements To propose test methods for testing material webs with periodically recurring samples, with which also individual Feature tracks are testable.

This object is achieved by the means specified in claim 1.

In a further development of the invention, the predetermined number is derived from the threshold values of homologous subsections a mean value of samples. To compensate for permissible long-term tolerances that are not to be assessed, the the averaging obtained and used for comparison by newly determined threshold values at certain intervals Threshold values overwritten.

It has proven to be particularly advantageous in this method that the desired accuracy is achieved with the present Track width and number of tracks can be varied almost at will by varying the length of the subsection.

If, for example, only those outside the pattern, the optical impression particularly disturbing, errors as well as the correct pattern spacing to each other were found which can be the case especially in pattern areas with a small percentage of area, so one forms in the vellum part, i.e. in the non-patterned areas of the web, a threshold value that is slightly above that of the cloudiness of the paper, for example returning noise level. The track levels in the pattern area, on the other hand, are raised so much for the whole pattern, that the greatest amplitude value of a pattern found to be in order is still below the threshold values. In this way

809809/0084

Only those defects can be recognized in the patterns that stand out above the general printing or embossing level. the however, errors in the unpatterned area, which are usually much more disturbing, are almost completely recognized.

If, on the other hand, you want an inspection accuracy, even the slightest If deviations in the pattern can be recognized, the pattern area in particular is divided into so many sections as possible, so that in the ideal case at least one threshold value is assigned to each signal amplitude. On. this way lays one can use the output signals of the pattern areas, depending on the size of the subsections, to create a more or less precise envelope with the the congruence of the periodically recurring lane areas can be monitored. To the permissible, partly above the good output signals lying on the arithmetic mean are not to be assessed as error signals, the envelope curve is lifted Addition of a constant voltage value at a certain distance over the output signal. The vellum part is checked against it also here with a uniform threshold.

Since the errors in the amplitude spectrum of the output signal can assume any value, e.g. in the positive amplitude range an otherwise positive partial section value will not only assume a smaller, but even a more negative value due to an error. Such opposing errors are recognized by the opposing envelope curve.

In order to limit the effect of such errors when determining the mean values used for comparison, it is useful to for this purpose, only allow the amplitude values up to zero potential.

309809/0034

- Y-

Depending on the need, the number of subsections in the area of the pattern can be determined according to various aspects will. If, for example, sample areas with low position and sample tolerances are to be checked, such as print samples, then the test accuracy can be increased significantly by as many sections as possible, so that even the smallest deviations in the pattern can be recognized. In this case, the setting is optimal if the segment duration is the half-wavelength of the smallest recognizable error, i.e. the setting is made depending on the highest frequency.

A different definition is recommended for the watermark check. Due to the manufacturing process, watermarks have larger tolerances than e.g. print samples. Short-term signal amplitudes from homologous Sub-sections would not result in a continuous envelope curve if the permissible position tolerance was greater. Considering of the permissible positional deviations, the subsections should therefore be chosen so large that homologous points are consecutive Watermark can barely be detected from a partial section with a maximum positional deviation. The larger the allowable If a positional deviation is selected, the greater the length of the subsections or the coarser it is while the envelope.

The output signals required for the evaluation can be transmitted in the form supplied by photocells, in which contains the average paper thickness as a direct voltage component, or as a differentiated photocell signal to be provided. Since the average paper thickness is usually checked much earlier during paper production as the differentiated signal also characterizes changes in contrast due to significantly steeper signal changes and since such a signal is automatically assigned a fixed zero level, one becomes the differentiated signal, however give preference.

809809/0084 · /.

* 2837977

A further improvement in error detection is achieved through special signal processing. Looking at the frequency spectrum a transparent, pattern-free paper web, this spectrum can basically be divided into three frequency ranges subdivide. The first area (I) mainly contains the long-wave changes in thickness of the paper and in the third area (III) the high-frequency amplitudes occur that result from sharp-edged differences in brightness, e.g. from small errors. The middle area (II) mainly contains the background noise or the cloudiness of the paper. In the second frequency range errors can only be detected if their amplitude exceeds the general noise level. Through a suitable Filter technology can thus make errors according to the criteria

a) Edge steepness of the error margins (III)

b) amplitude of the errors (II)

c) thickness error (I)

be separated. Depending on the application, the number of Sub-sections as well as the frequency range for the evaluation of the output signals can be freely selected. Likewise are different Combinations possible for the different railway areas. Should e.g. general changes in the thickness of paper knots or The same can be determined, the test in the first frequency range (I) will be carried out with only a section in the feature-free Perform rail area. Watermark ranges, on the other hand, make more sense with the help of the second and third frequency ranges to be checked taking into account the positional tolerance for the section length.

Further features and advantages of the invention emerge from the claims and the following description of an exemplary embodiment, which describes the testing of a watermarked paper web.

809809/0084

Show it:

Fig. 1 Schematic representation of the expedient for watermark paper Transmitted light scanning according to the principle of the invention;

2 shows a paper web with an indicated pattern area;

3 shows an output signal of a scanning track with threshold value increase ;

4 shows an output signal of a scanning track with an envelope curve;

5 shows a block diagram of the scanning device according to the invention and

6 shows a basic circuit of the averaging.

Fig. 1 shows the basic structure of a test device according to the invention and useful for testing paper webs, the Function when checking a watermarked banknote web is to be explained. In the use case is this test station in front of the sheeter and sorter from which the paper web cut into individual sheets at the correct distance from the feature areas and incorrectly marked sheets be sorted out by means of a switch. It is clear to the person skilled in the art that this is only one of many application examples and that in the same way the examination of print samples and the like is possible.

A paper web 1, which is moved in the direction of an arrow 4 via transport rollers 2, 3, points periodically in the transport direction recurring watermarks, their local position and structure together with the watermark-free areas (vellum part) should be checked for deviations and errors. For this purpose, a rod-shaped lamp 5 is below the paper web 1 as Lichtsöhder provided, which the paper over the entire web

809809/0084

wide homogeneously illuminated. On the opposite side of the paper web, a row of photocells 6 is attached as a receiver. For the light penetrating the paper web 1, given the continuously running web, the different Transparency fluctuations in brightness caused by the individual photocells are converted into amplitude fluctuations in their electrical output signals. The individual analog output signals the photocells are fed to a signal conditioning circuit 7, in which they are amplified and converted into one for the further processing necessary form can be brought. A clock generator 8, a cross-mark sensor 10 and a marker generator 9 are also provided.

Since the lamp 5 and the photocell row 6 transverse to the transport direction are arranged, as shown in FIG. 2, the paper web 1 continuously transported in the direction of arrow 4 and thus also the watermarks 11 made in the paper in the presence of 1 photocells in al - al -parallel, tracks 12 running along the direction of transport. Since these watermark areas 14 are on the paper web 1 in generally repeat at regular intervals, the same output signals of the photocells will also be periodic Repeat the order, whereby they exceed the noise level of the vellum areas several times in amplitude. In order to be able to check these repeating watermark signals of the individual tracks 12 according to the method according to the invention, the al - al individual tracks 12 are subdivided into bl - bn subsections 13. This makes the watermark areas 14 of superimposed on an imaginary two-dimensional Piüfraster, its Fineness depends on the length of the sections 13 and the number of tracks 12 (Fig. 2).

For this purpose, a clock signal of the clock generator 8 that is dependent on the paper transport and one that is dependent on the position of the watermark is used Synchronous pulse of cross mark sensor 10 required. In the use case is in the paper web 1 to limit the sheets

809809/0084 ./.

a so-called transverse mark (watermark mark) is provided for each sheet, which is also detected by a photo cell for synchronization of the subsections 13 is used. The geometric dimensions of the watermark, i.e. length and width, are manually entered in a programming unit.

By separate determination of those present in the individual tracks 12 highest and lowest amplitude values in the output signal of the photo cells and through the assignment of threshold values for this purpose, an upper and a lower envelope curve is assigned to each track for comparison with subsequent samples, i.e. for the positive and negative amplitude range of the output signal one each. The envelopes therefore represent thresholds that are set at fault-free web material must not be exceeded or undercut at any point.

3 and 4 show two output signals 15 obtained during the scanning of one of the tracks together with the associated envelope curves. The relatively strongly fluctuating opacity ratios in the watermark 11 are a region 17 of these output signals 15, whereas the more uniform output signal sections caused by the cloudiness of the paper of the vellum part are assigned to an area 16.

In Fig. 3 is a division of the track for less high quality requirements shown, with which primarily only the particularly annoying errors in the vellum part are recognized and displayed should be. The subsections are chosen so large that the output signal 15 during a watermark in the range 17 from only a single upper positive and a lower negative threshold value and the output signal 15 during the Velin part im Area 16 is also only limited by a single positive and a negative threshold value. The by identifying the The threshold values formed for the maximum amplitudes are shown by lines 22. The hatched areas 20 identify the

809809/0084

-X-

2537977

voltage value added as a tolerance range. Due to the thresholds formed in this way and relatively well-fitting in the vellum part 21, changes in opacity in the pattern-free area are recognized with great accuracy. Errors in the watermark, the smaller amplitudes than generate the greatest amplitude in each case, however, due to this very coarse rasterization, they remain undetected. Since smaller Differences in brightness in the watermark are by far not as disturbing as in the vellum part, the quality is such tested paper web 1 is sufficient for many cases despite the relatively low testing effort.

For high demands and especially for subsequent machine authenticity checks To enable bank notes, however, in the case of bank note paper, it is advisable to place the thresholds 21 above and below the watermark 11 to be changed in such a way that even smaller differences in brightness in the watermark are recognized as errors and eliminated can be. This is done in the pattern area by clocking the tracks into much smaller sections. In FIG. 4 For example, the watermark area was divided into 15 equal subsections bl - bl5. Just like that shown in FIG Simplified procedures, the highest and lowest amplitudes are determined for each section, the threshold values are generated and each assigned a threshold 21 increased or decreased by a constant voltage value. Despite the still quite coarse grid, in contrast to the example according to FIG. 3, a relatively recognize good adaptation to the output signal.

It is easy to see that the comparison of output signals with such envelopes reveals far more errors and that the test accuracy can be increased with ever finer timing up to the exact proof of concreteness.

An optimization of the subsection length can in particular be achieved with. very small errors to be detected can be achieved if the length of the subsections is the half-wavelength of the smallest

809809/0084. _A.

identifying error. In order to obtain a coherent envelope curve with certainty when averaging, However, the length of the section should not fall below the maximum permissible position tolerance and / or distortion deviation. In the case of very small permissible tolerances, the length of the subsections is therefore equal to the half-wavelength of the smallest Adjust the error to be assessed. If larger tolerances are to be taken into account, the length of the sections is selected using the maximum tolerance deviation.

In Fig. 5, the block diagram is one according to the invention Method functioning test device shown for one lane and the arrangements for the other lanes indicated.

The test device essentially consists of a programming unit 25, in which the current watermark position or register position is stored and which from the clock generator 8 and from the cross mark sensor 10 is controlled; a preamplifier 24 in which the output signal of a photocell 23 of the photocell row 6 is amplified and, if necessary, is divided into the desired frequency ranges by means of suitable filter circuits; a memory and Arithmetic unit 26 in which the envelope curves are generated and stored; a control unit 27, which the timing for Measured value acquisition, storage and playback controls and one Comparison unit 28, which compares the respectively valid threshold values with the output signal of the photocell 23 and the exceeding a threshold as an error. The aforementioned preamplifier 24 and the units 25 to 28 and one later explained OR circuit 69 are contained in the signal conditioning circuit 7 (Fig. 1).

The programming unit 25 and the control unit 27, since they control the entire system, are only per test device to be provided once, whereas the storage and arithmetic unit 26,

809809/0084

a-

the comparison unit 28 and the preamplifier 24 for each track to be checked must be present once. With 1 lane 12 the test device consists of 1 photocell 23 in the photocell row 6, 1 comparison units as well as 1 storage and arithmetic units 26 and one programming unit each and control unit 25, 27 as well as the clock generator and the cross-mark sensor 10. Should a paper web 1 follow If several separate criteria are checked, then there is a separate one for each track 12 for each frequency range Provide storage and arithmetic and comparison units 26, 28, as shown in FIG. 5 with blocks 26a, 26b, 28a and 28b is indicated.

The paper web 1 to be tested, on which, as already mentioned, the individual watermarks 11 become so-called registers lined up and on which cross marks are provided for synchronization, the test device runs through the manner shown in Fig. 1 and is scanned by the consisting of 1 photocells 23 photocell row 6 for errors. With the help of a photocell 32 and a lamp 33 in the cross mark sensor 10, the synchronous

809809/0084

marks in the paper determined. In the photocells 23 arise the output signals corresponding to the output signals 15 shown in FIGS. 3 and 4, whereas the photocell 32 only emits a pulse intended for synchronization at the beginning of a register. In addition to these two signals supplies the one which is driven by the transport system and consists of a photocell 29, a slotted disc 30 and a lamp 31 Clock generator 8, a square-wave signal that is permanently assigned to the paper flow and is used to clock the tracks 12.

Depending on the desired length of the subsections 13 are in the programming unit 25 in a circuit 34, which is from the clock is controlled, for this purpose combined several clocks for a section with the length b. The designation (_) in The circuit diagram symbol indicates that 3 clocks for a length b of a section 13 are combined. The size of the Sub-sections and the point in time at which the formation of the sub-sections in the register areas begins can be set for each Paper type can be entered manually in a matrix circuit 35 as a test program. The matrix circuit 35 is from the cross mark sensor 10 controls and controls the circuit 34 in turn.

Via a clock line 36, the segment length adapted, Clock signal from circuit 34 to a memory pulse generator 38 and to a memory address generator 39 passed on. From the storage pulse generator 38, the signals formed in a rectifier circuit 46 are then converted into individual sections chopped up. The rectifier circuit is constructed from a peak-value measuring rectifier of a known type of circuit. In this rectifier circuit 46 can also a differentiating circuit may be included. The differentiated signals are preferably processed further. Also will a summer 49, analog memories 50,52 and a comparator 55 from the Speieherimpulsgenerator 38 via a line for the Threshold formation and the comparison necessary clock cycle imprinted. The generator generates with every partial pulse

809809/0084

thereby a pulse sequence which, as will be described in more detail later, essentially the write and read cycles the memories 50 and 52 control. The memory address generator is meanwhile, as will also be explained in more detail, to control the memory 50,52 switched by one memory address with each partial segment clock. The name () in Generator 39 should say in the application failure that n memory addresses are available, one of which is coded out one after the other.

A clock line 37 emanating from the matrix circuit 35 is supplied a pattern counter 41 in a similar manner with clock pulses which mark the beginning and the end of the individual patterns. In the pattern counter 41, the number of pattern areas used for averaging is preset. Go from him hence the signals which, after a preset number of patterns m have been reached, replace the ones stored in memory 52 effect old comparison threshold values against the newly determined values formed in memory 50. The pattern counter 41 thus counts the successively scanned watermarks and emits a coincidence signal if there is a tie between the counter and the preset, which transfers the memory content from memory 50 via an AND circuit to memory 52 and, on the other hand, the generators 38, 39 resets via a control line 42. The control unit 27 also has a logic circuit 43 that detects any errors registered during the averaging via a line 45 and influenced the pattern counter 41 and thus dependent on the The number of errors or error size prevents the new envelope curve from being overwritten, otherwise the threshold values for checking will be falsified the following watermark series would be used.

For the sake of simplicity, the storage and computing unit 26 is explained with reference to the formation of a positive envelope curve by means of the rectifier circuit 46, the adder 49 and the memory 50. The formation of the negative envelope takes place in the same way with the same circuit units indicated in the lower half of the unit 26. . ^ _ä "_.

809809/0084

2S37i? 7

The output signal of the photocells 23 passes through the preamplifier 24 once to the rectifier circuit 46, in which the largest amplitudes within the subsections are determined and threshold values dependent thereon and thus thresholds 21 are formed, as well as to an isolating amplifier 54, from which it is compared with the threshold values available in the memory 52. The output of the rectifier circuit 46, which has only one voltage value depending on the respective highest amplitude per section passes over a line 47 and a quotient circuit 40 to the adder In the quotient circuit 40 is controlled by a preselection switch 44 and, as will be described in greater detail later, each voltage level assigned to the sub-sections to the m-th value divided down, where m corresponds to the preset number of patterns. This voltage value is used for averaging in the summer 49 for the previously determined values of the same Sub-sections are added and stored in a memory cell of the memory 5O assigned by η to the respective sub-section. After checking m watermarks, the η memory cells of the memory 50 accordingly contain those from m Watermarked mean values of the individual subsections. Applied after completion of the mth test the pattern counter 41 has an input 53 of the AND circuit 51 with a voltage pulse to which the AND circuit 51 turns on. So that the content of the memory 50 is a Line 67 reloaded in sequence into memory 52 after deleting the previous content. So during the individual storage processes the individual »voltage levels always get into the correct memory cells of the memory 50 and 52 and thus at the When comparing the current level, the corresponding threshold values of the memory 52 are always available, opens and the memory address generator 39 closes the individual memory cells of the memories 50 and 52 one after the other in the rhythm of signals supplied by circuit 34 ..

809809/0084

The voltage levels stored in memory 52 are used for ongoing comparison with those output by isolating amplifier 54 current levels. They are each applied to an input 57, 58 of the comparator 55 and generate 59 at its output an error signal if the signal at the first comparator input 57 applied voltage value is exceeded by the voltage value applied to second input 58. This error signal is given to an output terminal 70 via OR circuits 68,69. By means of the error signals at the output terminal 70, the marker transmitter 9 indicated in FIG. 1 can then be actuated. A comparator 56 does the same Function and evaluates the signal for the negative amplitude range and also outputs its error signal via the OR circuit 68,69 to output terminal 70.

Since, during the restoring of the memory contents from memory 50 to memory 52, there is no clear comparison of the data from the isolation amplifier 54 delivered signals with the stored values of the memory 52 is possible, the comparators 55 and 56 for these periods of time are blocked by the generator 38 via the line 60. The output of an error signal that occurred during restoring based on undefined tension is therefore not possible. Since the Umspeieherzeit only fractions of the actual Cycle time is required, the resulting fade-outs for the test are negligible.

With the aid of a basic circuit diagram shown in FIG. 6, the mode of operation of the memory and arithmetic unit 26 carried out is intended Averaging will be explained again in more detail.

When examining subsection bl of the first of m watermarks 11 (FIG. 2), the voltage value already processed by the rectifier circuit 46 arrives via the supply line 47 to the quotient circuit 40. With the aid of the preselection switch 44, which is also mechanically coupled to the pattern counter 41, the division ratio of the quotient circuit 40, adjusted so that only the corresponding part of the applied voltage at the input of an operational amplifier 61 in the summer 49 becomes effective. The resistors 62 and 63 have a ratio of 1: 1

809809/0084 _{m / m}

to each other, so that the full voltage value of a buffer cell 64 becomes effective at the summing input. Since a switch 65 formed from a field effect transistor (FET) initially remains blocked during the first watermark, 0 potential is accordingly present at the summing input. Thus, at the output of the operational amplifier 61, the m-th part of the voltage value al previously opened, registered. The voltage values of the subsections b2-bll are transferred to the memory cells S ₂ S in the same way.

The threshold value formed for the vellum part is stored in the memory cell S _{+ -.} This threshold value is determined exactly like the individual threshold values of the envelope curve by determining the highest amplitude. Since the checking of the entire vellum part, as already mentioned at the beginning, is carried out only with a constant threshold, the entire vellum area located between two watermarks is, however, equated to a partial section by appropriate programming in the matrix circuit 35.

In the case of the second of the m watermarks, the buffer cell is activated via switch 65 at the beginning of each subsection 64 are charged in sequence with the voltage value present in the respective cell S of the memory 50 and closed the voltage levels provided by the quotient circuit 40 are added. At the output of the operational amplifier 61 are that is, the newly arriving voltage levels are added to the associated memory values of the same subsections and Stored again in the memory cells S via the switch 66. After passing m watermarks, they are the Memory cells S ^ ... S assigned to subsections are therefore loaded with the mean values formed over m watermarks.

809809/0084

These mean values are, as already explained, before the next, i.e. (m + 1) watermark via line 67 "reloaded" into the memory 52. The memories 50 and 52 are during the test and averaging and during the reloading of the memory address generator 39 controlled with the same addresses, i.e. it will be completely the same memory cells are opened or closed synchronously. While the next m. Watermarks with the levels of the Memory 52 are compared by the comparison unit 28, the memory 50 is filled in parallel with new mean values. In this way, the test apparatus is rendered uninterrupted in a relatively economical manner during the paper test of the test process, with the exception of the negligible Hiding during memory overwriting, supplied with new threshold values.

As already mentioned at the beginning, the individual mean values are increased by a constant amount. This is done by a simple summing circuit consisting of an operational amplifier and some resistors '. For the sake of simplicity, this unit is not shown separately. It is included in the memory 52, ie connected upstream of the memory output.

When testing the surface of samples with a uniform threshold (see Fig. 3), the effort can be substantial be reduced. The signal coming from the rectifier circuit 46 is here instead of via summing and storage circuits Passed through a less complex low-pass circuit and compared with the output signal of the isolating amplifier 54. In this modification of the invention, an occurring error causes an immediate change to the one used for comparison Threshold. In the case of a time constant corresponding to the passage of five watermarks, for example, this has the effect one-time errors but only a short-term change in the threshold value

around the actual value. The error detection will

therefore only very insignificantly affected, although the device-related Effort by omitting the quotient circuit 40,

809809/0084

the summer 49, the memories 50, 52 and the AND circuit can be reduced very substantially. A further simplification can be achieved if only during the examination the positive signal components are taken into account. By rectifying the signal coming from the preamplifier 24, in particular by means of a two-way or bridge rectifier, which causes the negative signal components to flip up the resulting loss of quality can be kept within limits.

Without departing from the principle according to the invention, the paper web 1 can of course also be checked transversely to the transport direction instead of with tracks running in the longitudinal direction. The scanning of the tracks is then no longer carried out in parallel, but in series, e.g. by means of a row of scanned photocells or by a light beam deflected by a mirror wheel or the like. The basic idea of the sub-section and threshold value assignment but can be taken over in full regardless.

809809/0084

Claims (18)

Claims Procedure for testing material webs provided with periodically recurring patterns for defects, in which the web area to be tested is scanned in parallel tracks, characterized in that the individual Tracks are divided into subsections by a clock synchronized with the recurring pattern, that the Sub-sections of the output signals of the sampling dependent threshold values are assigned, which over the individual output signals of the tracks form an envelope curve, that these envelopes are stored and that when the output signals subsequently determined are compared with the stored envelopes, when exceeded, indicate the presence of an error. 2. Test method according to claim 1, characterized in that the parallel tracks to be scanned in the longitudinal direction of the Material web run. 3. Test method according to claim 1, characterized in that the parallel tracks to be scanned transversely to the transport direction run along the web. 4. Test method according to claims 1-3, characterized in that the length of the subsections in the individual tracks correspond to the length of a pattern. 5. Test method according to claims 1-3, characterized in that the individual tracks in the area of each pattern in several equally long subsections are subdivided and the individual subsections separately determined threshold values be assigned. _. , 809809/0084 6. Test method according to claim 1-3, characterized in that the length of the subsections as a function of the Positional tolerance between the feature areas in the patterns can be determined. 7. Test method according to claim 1-3, characterized in that the length of the subsections as a function of the Pulse width of the smallest permissible error is determined. 8. Test method according to claim 1, characterized in that the individual threshold values by adding a constant Voltage value can be formed for the largest amplitude present in the respective subsection. 9. Test method according to claim 1 and 8, characterized in that that the formation of the threshold values in the individual subsections after the scanning of a preselectable Number of patterns done. 10. Test method according to claim 9 ^ characterized in that that the threshold values determined according to a preselectable number of samples for testing the following, same number of patterns can be used. 11. Test method according to claim 9 and 10, characterized in that the threshold values of the individual subsections replaced by new threshold values after the predetermined number of patterns have been checked. 12. Test method according to claim 1, characterized in that the new threshold values during or in parallel are determined for checking the predetermined number of samples. 809809/0084 13. Test method according to claim 9, characterized in that from the threshold values of homologous subsections the predetermined number of patterns is averaged. 14. Test method according to claim 1, characterized in that the stored threshold values of the individual subsections can be changed immediately by the homologous output signal components of the current pattern. 15. Test method according to claim 14, characterized in that the speed of the threshold value change by Provision of delay elements is adjustable. 16. Test method according to claim 1, characterized in that the formed by the individual threshold values and stored envelopes of the output signals of the tracks with the differentiated output signals of the current scanning are compared. 17. Test method according to claim 16, characterized in that the rectification of the output signals of the straight subsequent sampling is carried out with a full-wave or bridge rectifier method. 18. Test method according to claim 1, characterized in that the output signals of the sampling of the individual Tracks are divided into at least 2 frequency ranges and that the test method for one or more these frequency ranges are carried out separately. 809809/0084