EP1520323A1

EP1520323A1 - Positioning of flat conductors

Info

Publication number: EP1520323A1
Application number: EP03739844A
Authority: EP
Inventors: Nikola Dragov
Original assignee: I&T Innovation Technology Entwicklungs und Holding AG
Current assignee: I&T Innovation Technology Entwicklungs und Holding AG
Priority date: 2002-07-02
Filing date: 2003-07-02
Publication date: 2005-04-06
Anticipated expiration: 2023-07-02
Also published as: JP2005532772A; AU2003281375A1; DE50306756D1; WO2004006390A1; EP1520323B1; ATA9872002A; US20040200068A1; ATE356448T1; AT413164B

Abstract

The invention concerns a method for positioning of the stripped sites of two flat flexible cables (FFC) to be mechanically and electrically connected to each other. The invention is characterized by the fact that the stripped sites of the two flat flexible cables being joined are optically detected, both in terms of their location, on the corresponding flat flexible cable, and also according to their shape and size, that they are positioned one on the other, so that, in a first variant, the smallest of the overlapping stripped surfaces will be a maximum. In a second variant, positioning occurs so that the largest current density occurring in one of the overlapping stripped surfaces will be a minimum. The invention also concerns an apparatus for execution of the method.

Description

Positioning flat conductors

The invention relates to the positioning of flat conductors (FFC), more precisely the positioning of stripped points of two flat conductors to be electrically connected and mechanically connected to one another.

If stripped areas, so-called windows, of flat conductors are to be conductively connected to one another, this is done manually in the prior art and comprises almost exclusively only one, in rare cases a maximum of two conductor paths per FFC, since the windows are not produced with sufficient accuracy in order to enable the currently existing machines to achieve sufficient coverage of the conductor tracks exposed in the windows for the respective current load.

According to the invention, the inclusion of an image processing device in the positioning and connecting process of flat conductor cables improves the accuracy of the connection and thus improves the conductivity. In the following, the advantage of the invention is shown by a worst-case estimate for both methods, for the classic method previously used, which, however, does not represent a previously published state of the art but internal knowledge of the applicant, and the image processing method.

The invention is explained in more detail below with reference to the drawings. 1 shows the effects of the stroke error, FIG. 2 the additional effects of the robot error, FIGS. 3 and 4 the effects of the output data errors, FIGS. 5 to 8 the formation of the coverage error, FIG. 9 the overlap error, 10 the stroke error, FIG. 11 the robot error, FIGS. 12 to 14 the effects of twists, FIG. 15 the effect of the combinations of the individual errors, FIG. 16 the situation of a single window in the image processing method, FIG 17 the combinations of the individual errors in the image processing method, FIGS. 18 and 19 the basic situation in the case of several windows, 20 to 24 the effects on the individual windows, FIGS. 25 to 28 representations analogous to FIGS. 20-24, but with the image processing method, FIGS. 29 to 32 and 33 on the one hand and FIGS. 34 to 37 representations on the other the overlap of the windows in the classic or in the image processing method.

First, the less favorable case for image processing is dealt with, in which no deviations occur during production. In order to be better than the classic method with image processing, the camera must not deviate by more than 0.08475 mm in accuracy.

The influence of the rotation of the positioning robot is estimated as well as the theoretical possibility of causing a short circuit by connecting two conductor tracks. The worst possible values for the cable production are assumed, but they are still within the tolerance range. These deviations in production are called output data errors.

In the following, both methods are analyzed in a single window view and in a much more practical matrix view: Single window:

Matrix consideration:

Based on the representation in the coordinate system, the center crosses and the effects of the individual errors on the displacement of these crosses are easiest to display graphically.

If the camera error is below 0.08475 mm, the image processing method is superior in all cases to the classic system in terms of accuracy. The increase in the common contact area, which is a measure of the improvement in accuracy, can be summarized as follows:

Ultimately, a major advantage of image processing is the detection of defective products and the possibility of specifying a threshold value for the common contact surface, below which the workpiece (or workpieces) are rejected.

Matrix Positioning Considerations

1. Description of the work processes

1.1 Classic procedure (existing, but unpublished, system)

Production of the workpieces - errors that occur are due to both the tolerances of the conductor tracks and the offset of the matrix, which is due to the inaccuracy of the laser during production. The data sheets for the applicant's FFC were used to calculate the maximum error, which is still in the tolerance range. Stop - The process of connecting starts when the lower flat conductor cable is attached to a stop. This deviation from the ideal stop (stop error) is also taken from the data sheet and represents the lateral tolerance outside the copper tracks of ± 0.12 mm.

Robot positioning - A robot grips the flat cable and positions it on the carrier using a vacuum suction device. The inaccuracies in the positioning of the robot gripper are included in the calculation. An estimation of the repetition accuracy and the rotational error must be carried out.

Stop - The second (upper) flat conductor cable is placed against the stop. The same tolerance estimates apply as for the first stroke error.

Robot positioning - The second flat conductor cable is also positioned on the carrier by the robot gripper. This cable is rotated by 90 ° relative to the first (lower) flat conductor cable.

Connect - The flat conductor cables are conductively connected to the carrier at the defined and exposed copper windows using a material or non-positive connection process. Welding, soldering, crimping or similar processes can be used as connection processes, provided that they produce an electrically conductive connection. Because the same robotic arm is used for positioning, the same error estimate can be used.

1.2 Image processing methods

Production of the workpieces - The same assumptions are made for the production and positioning of the matrix itself as for the classic manufacturing process.

Stop - The first flat conductor cable is also placed against a stop in the image processing process. Errors that occur here are interpreted and compensated for by the camera as translation or rotation of the matrix on the flat conductor cable. The stroke error is not included in the accuracy in the image processing process. Robot positioning - The robot gripper positions the first workpiece on the carrier. Errors that occur here are interpreted and compensated for by the camera as translation or rotation of the matrix on the flat conductor cable. The robot error in the image processing method does not affect the accuracy of the first flat conductor cable.

Image processing - The flat conductor cable lying on the carrier is picked up by the camera in reflected and / or backlit mode. The window size, the matrix structure and the position of the windows with respect to the conductor strips are determined. The image processing calculates a center cross of the matrix and center crosses of the individual copper windows detected. The camera error is included in the calculation here.

Stop - The second (upper) flat conductor cable is placed against the stop. Image processing also compensates for any errors here and the accuracy is not affected by this step.

Robot positioning - The robot gripper holds the second flat conductor cable with the contact side in the camera. Any positioning errors are compensated for by the camera.

Image processing - The flat cable held in the camera is picked up by the camera in reflected and / or backlit mode. The window size, the matrix structure and the position of the windows with respect to the conductor strips are determined. The image processing calculates a center cross of the matrix and center crosses of the individual copper windows detected. The camera error is included in the calculation here.

Calculation - The trajectory for transferring the center cross of the matrix of the upper flat conductor cable into that of the lower flat conductor cable is calculated. No error is assumed in the calculation.

Robot positioning - The robot gripper positions the second workpiece according to the calculated trajectory after a 90 ° rotation over the first cable. This step can no longer be checked by the camera and errors that occur are included in the accuracy in the second robot positioning as well as in the classic process. Connect - On the carrier, the flat conductor cables are conductively connected to each other at the defined and exposed copper windows using a connection process. Because the same robotic arm is used for positioning, the same error estimate can be used.

1.3 Fixed values

The stroke error is assumed to be ± 0.12 mm from the data sheet.

For robot positioning (robot error), a repeat accuracy of 0.07 mm based on the TCP (tool center point) applies. When rotating, the robot gripper does not deviate by more than 0.05 mm with a cable length of 600 mm.

The I&T FFC data sheet was used as the basis for the tolerances in cable production.

Although based on algorithms, sub-pixel-precise edge detection can be achieved using image processing, the worst case for the camera error is the deviation from a full pixel (currently 0.025 mm).

2. Estimation without tolerances in the manufacture of the flat cable

The image processing method recognizes several errors in the course of the entire production process and can compensate for them. The greater the errors that occur, the greater the improvement in the image processing method compared to the classic (but unpublished) method described. The behavior and function of both manufacturing processes under optimal starting conditions, i.e. no tolerances in the production of the cables can be estimated.

2.1 Classic (but unpublished) procedure

Assuming that both the conductor strips and the matrix are correctly positioned, the stroke error and the robot error remain in the classic manufacturing process only twice to limit the accuracy.

The observation is made for the smallest possible single window, a 1.5 mm x 1.5 mm window. Since the stop error can occur on both flat conductor cables, one loses 0.12 mm in height and one side from the common top surface of the copper squares.

Fig. 1 stop error

Fig. 2 additional robot error

The window is reduced by the stop error to a 1.38 mm x 1.38 mm window, which still represents 84.64% of the original contact area.

In the worst case, the robot error in the positioning of the workpieces causes the squares to drift further apart by 0.07 mm ^• sin 45 ° = 0.0495 mm in terms of side and height.

The coverage window is thus reduced to 1.281 mm x 1.281 mm or 72.93% of the original area.

2.2 Image processing methods

The impact error is compensated for by the image processing on both flat conductor cables.

The robot error only occurs once (when the second cable is removed after image processing). The window is reduced to a 1.4505 mm x 1.4505 mm square, which has 93.51% of the original area. In the processing method with image processing, the camera error limits the accuracy for both workpieces.

In order to achieve better results with the image processing method than with the previously used method even with this estimation without manufacturing tolerances of the cables, the camera error must be less than 0.08475 mm. The same lower limit for the camera error is also obtained for other single window sizes (up to 19 mm x 19 mm possible according to the data sheet). This value is also undercut under the worst case assumption of a deviation of a full pixel of 0.025 mm and leads to a contact area square of 1.4005 mm x 1.4005 mm or 87.17% of the original area.

The improvement of the image processing method compared to the classic method is thus 16.34%.

3. Errors in the classic (but unpublished) process

3.1 Output data error

For a worst-case assessment, the deviations that occur during the production of the flat conductor cables must also be added to the errors in the processing and connection process.

The sum of the errors that occur in cable production are referred to here as output data errors, since both the classic method and the image processing method are subject to these specifications.

The following tolerances now affect production:

Conductor stripe offset (± 0.15 mm) Conductor thickness change (± 0.05 mm) Matrix offset (± 0.12 mm) Window size change (± 0.05 mm)

For a worst-case assessment, a 1.5 mm x 1.5 mm window assumes that the conductor strip offset and the matrix offset are maximally offset from one another. In addition, both window size and stripline thickness are reduced to a minimum. Such a worst-case window has dimensions of 1.45 mm x 1.2 mm instead of 1.5 mm x 1.5 mm and the visible copper area has dropped to 77.33%.

3 and 4 output data errors 3.2 Coverage errors

For the calculation of the coverage error, the contact area reduction that arises when one assumes a worst case situation with the output data errors from point 3.1, the individual steps leading to the maximum deviation are detailed. It can be seen that the common contact area has a minimum if the copper window of both

Cables have a size of 1.45 mm x 1.20 mm and are shifted to the left and up.

This conclusion is obtained from the following assumptions:

Fig. 5, 6, 7, 8 misregistration

To connect the two flat conductor cables, the second workpiece is turned and positioned on the first workpiece rotated by 90 °. If the same position dimensions apply to both flat conductor cables, in the worst case the overlap looks as follows:

Fig. 9 overlap error

The full overlap on the narrow side is also given in the worst case (previously only manufacturing tolerances and coverage errors were taken into account). The contact area is a 1.2 mm x 1.2 mm window, which still has an area of 64% of the original window size.

3.3 Stop errors

The stop error has the greatest influence on the accuracy if both the lower and the upper flat conductor cable at the stop are too narrow by the full tolerance of 0.12 mm. When considering the worst case in the classic method, it must then be assumed that this deviation can be found in both workpieces.

Fig. 10 stop error The 90 ° rotation of the upper flat conductor cable reduces the copper contact surface by 0.12 mm in height and in the side due to the stop error. The resulting contact area is a 1.09 mm x 1.09 mm square when considered worst case, which is only 52.80% of the original size.

3.4 Robot errors (positioning inaccuracy)

The repetition accuracy of the positioning robot is 0.07 mm, based on the TCP. In conjunction with the other errors, the worst case occurs when the deviation vectors point left-up or right-down in the coordinates.

Fig. 11 robot error

The greatest reduction in contact area is thus characterized by a decrease in the side lengths of the contact area square by 0.07 mm ^* sin 45 ° = 0.0495 mm in each case.

Finally, after considering all (maximum possible) errors, a square with 0.991 mm x 0.991 mm is the result. The copper-copper contact area is only 43.65% of the original contact area!

4. Further estimates

4.1 Short circuit

It is investigated whether the maximum can be exploited to the maximum that the case may arise that a window can be placed over two different conductor tracks and thus leads to a short circuit after the connection.

If you take the value x as the ideal distance between two strips, in the worst case the strip spacing is x - 0.20 mm, since the strip spacing is assumed to be too small by 0.15 mm and the strip thicknesses by 0.05 mm are assumed to be too wide ,

The laser-applied window can be 0.05 mm too large and 0.12 mm too deep (or too high) on the flat cable. Together, these deviations result in a streak stood from x - 0.345 mm. Since the spacing of the strips is in practice more than 1 mm (i.e. significantly more than 0.345 mm), there can be no short-circuit, even in the worst case, if the tolerances are observed.

4.2 Effects of rotation on the robot gripper

4.2.1 Single window viewing

For the maximum rotation error, a deviation of 0.05 mm is specified for a 600 mm long workpiece. The error angle is consequently

_, 0.05mm a = tan '- = 0.00477 °

600mm

In the worst case, there is an opposite rotation by this error angle and thus a relative rotation by 0.00955 ° between the squares.

Fig. 12 and Fig. 13 effect of a twist

The result is 8 loss _{triangles of} the total area A _verl = 4. - .tan 0.00955 ° (with x as

Side length of the square).

The relative error is 0.016% in each case.

This relative error of less than 0.02% is not taken into account for the error due to rotation when viewing single windows, but would speak in favor of the image processing method if it were taken into account, since only there can compensation by the camera be made.

4.2.2 Matrix consideration

Assuming the maximum matrix size of 128 mm x 128 mm and a positioning of a (real) 1.45 mm x 1.2 mm window in a corner of this matrix, the error that arises can be estimated by the following consideration:

Fig. 14 Effect of twisting The window is now no longer approximated 1.45 mm x 1.20 mm but only 1.45 mm x 1.1787 mm, which corresponds to a surface coverage of less than 1.8%. Since the common contact area only drops by less than 1.8% under worst-case conditions, the rotation errors of the robot gripper can also be neglected when considering the matrix. With the positioning method with image processing, it could also be reduced.

5. Crosses for the classic (but unpublished) process

When looking at the coordinate system, one can use the center crosses of the individual windows or their displacement due to the individual errors.

Fig. 15 combination of the individual errors

6. Summary of the classic, unpublished. Single window viewing procedure

Due to the output data errors, a resulting window size of 1.45 mm x 1.20 mm is assumed. The misregistration limits the minimum contact area to a square with a side length of 1.2 mm x 1.2 mm. It must be expected that the stop error occurs in each of the two workpieces and is therefore included in the calculation twice. It is also assumed that the robot error for each workpiece is included in the accuracy calculation as badly as possible.

7. Summary of the image processing method for single window viewing

As with the classic, unpublished process, the image processing process is also dependent on the specifications from flat conductor cable production and the output data errors must be used as the starting value. For the single window view, the misregistration can be compensated in such a way that the center crosses of the upper and lower workpiece overlap. The robot error is only included in the accuracy calculation once, since the image processing cannot compensate only for the second workpiece. An overlapping surface is obtained which can be sketched as follows:

Fig. 16 Situation of the single window

For this procedure, the camera error KF (difference between reality and digital image) goes into the accuracy twice, since each of the two workpieces is "measured". The resulting total contact area can now be specified using the following formula:

for KF <0.03775 mm: relative cover area = 64%

for KF> 0.03775 mm: (.100%)

It also shows for a worst case camera error of 0.025 mm that the lateral "overlap reserves" are large enough and the common contact area is not restricted by the camera error.

A worst-case improvement of 31.80% for the image processing method compared to the classic method can be specified for single window viewing.

The following examination of the center crosses shows how small the deviations due to errors are:

Fig. 17 Combination of the individual errors

8. Multi-window matrix consideration for the classic procedure: The worst case for multi-window matrices (practical case) should be sketched:

Fig. 18 and Fig. 19 Basic situation with several windows

In this case, all windows drift apart to the maximum tolerances on one cable and the windows on the other flat conductor cable move together to a minimum. For an originally 1.5 mm x 1.5 mm window, which was reduced to 1.45 mm x 1.20 mm due to the manufacturing tolerances, the following overlap occurs:

20, 21, 22, 23 and 24 effects on the individual windows

When considering the matrix, an additional coverage error of 10.54% can occur, which ultimately reduces the contact area to a 1.135 mm x 1.135 mm square. The area of the contact square corresponds to only 57.25% of the original area.

Here too, the stop error can occur for both cables and thus reduce the contact area to 45.79% of the original area.

The robot error occurs again twice and has the contact area reduced to a 0.916 mm x 0.916 mm square, the size of which corresponds to 37.29% of the original size.

If a possible connection error, which has the same deviation as the robot error, were included, the surface of the joined copper-copper connection is 0.8665 mm x 0.8665 mm and thus 33.37% of the original coverage area.

9. Multi-window matrix viewing for the image processing method

If you again take the case of the maximum apart windows on one workpiece and the maximum apart windows on the other workpiece for viewing, you can use the image processing to reduce the misregistration. The procedure is to reduce the rather large error in the lower right window to the detriment of the other windows and thus to correct errors in all windows. In the worst case, this error compensation corresponds to a translation of the upper flat conductor cable by 0.0375 mm to the left and up. With this translation, the overlaps listed under point 8 become overlaps of the following structure:

25, 26, 27 and 28 affect the individual windows

The worst case for the contact area square is 1.1725 mm x 1.1725 mm side length and thus with a contact area of 61.10% of the original area.

With the image processing method, the stroke error is not included in the accuracy calculation - this error is compensated for by the image processing.

The robot error only has to be taken into account once, namely when the second flat conductor cable is positioned on the first cable after analysis by the camera. If you deduct the effects of this one robot error from the contact area, you get a copper-copper coverage area of 56.05% of the original area (1.123 mm x 1.123 mm instead of 1.5 mm x 1.5 mm).

Now the camera error has to be taken into account twice. In order to be better than the worst case in the classic case with 0.916 mm x 0.916 mm, the camera error has to be less as 0.1035 mm based on the TCP. The camera test in practice confirms a deviation of 0.025 mm as a worst case error.

For the matrix analysis that occurs in practice, the increase in the common copper-copper contact area and thus the improvement in accuracy is 27.12%.

Classic (but not prepublished) procedure

Fig. 29, 30, 31 and 32 overlap of the windows in the classic method

From the lower right window follows a detailed drawing for a clearer representation:

Dashed square: ideal square with 1.5 mm x 1.5 mm

Dotted lines: diagonals of the ideal square and the resulting contact area square for determining the center crosses

The center point of the remaining contact area is shifted right-up compared to the ideal center point

Fig. 33 Overlap of the windows in the classic method

Matrix view image processing

34, 35, 36 and 37 overlap of the windows in the image processing method

Classic (unpublished) procedure:

Image processing methods:

Summary of the comparison:

As can be clearly seen from the above explanations, the invention achieves through its measures:

Optical detection of the stripped points of two flat conductor cables to be connected both according to their location on the respective flat conductor cable as well as according to their shape and size, the stacking in such a way that the (relatively) smallest overlapping stripped surface is one Maximum and the placement of the tool of the machine that makes the electrical connection in the middle of the overlapping stripped surface that a reduction in the size of the actual contact surfaces reduced and thus a significant reduction in electrical resistance and a substantial increase in mechanical strength is achieved ,

The "relatively smallest area" is not necessarily the one with the smallest area, but usually the one that is to be regarded as the most critical due to the width of the copper conductor to be connected and / or the intended specific current load If, for example, there are two overlapping areas of the same geometry with different current loads, the area with the higher expected current load is the "relatively smaller area".

The invention is not limited to the examples described, but can be modified in various ways. By appropriate adaptation of the formulas, which can easily be carried out for the person skilled in the art with knowledge of the invention, a corresponding application can be made for flat conductor cables to be connected to one another at an angle (and not at right angles, at 90 °). The invention can be used for all types of robots and flat cables (laminated and extruded, with mutually identical copper conductors or with mutually different, etc.), the stripping can be done in many different ways, even if special attention was paid in the application to laser ablation of the insulation , The shape of the stripped surfaces does not have to be rectangular, circular or oval shapes are also possible.

It should also be taken into account that the implementation of the method according to the invention and the comparison with the internal state of the art was carried out using a special exemplary embodiment, tolerances being used as they occur in practice. Of course, other possible improvements are obtained when comparing examples with different tolerances, but the example clearly shows how an automatic connection of FFCs is possible through optical detection in combination with the method according to the invention. The question of the "center" of the overlapping stripped surface can be solved either by choosing the center of gravity or by another choice that is adapted to the respective connection method. For example, it is possible to carry out a center of gravity with a weighting, linear or square, for special needs Particularly in the case of complexly shaped overlapping surfaces, the intersection of the shortest with the longest chord or the like can be used for simplification, and the use of specially shaped tool tips, which in turn make a choice of one's own possible and necessary, is also conceivable.

All devices that are controlled by an electronic control and that are able to manipulate an FFC with repeatable accuracy (gripping, moving, depositing) can be considered as robots. In principle, it does not matter whether it is gripped by suction cups or clamps.

The invention also relates to a device for performing the method, the at least one robot that can handle the FFC, a fixing device for the positioned FFC on a carrier, a camera, a device that provides the electrical connection, and a control device for the robot, which also processes the signals coming from the camera. The control device does not have to be a physical unit; the description and the claims include the sum of the (mostly electronic) components and devices together with their sensors, which in total allow the inventive method described above to be carried out.

Since the FFC is positioned in its end position after the optical detection, without necessarily using a stop, the FFCs only have to be fixed in their end position, which is possible, for example, using suction devices, pressure devices, etc., which are preferably mounted on the carrier is.

The carrier can be a flat plate (for example a work table), or it can have the shape of an elongated storage surface, for example the surface of an extruded profile. The fixing devices are preferably mounted on the carrier, since this is the easiest way to prevent or minimize any undetected and thus undesirable movement of the FFC during fixing. For this purpose, the FFC is placed and held by the robot on the carrier, then the fixing devices (here in the description always addressed in the majority, without that this would be a technical necessity) and only then does the robot let go of the now fixed FFC.

The second FFC is also deposited by the robot, preferably the same as for the first FFC, after the calculation of the desired end position and fixed by the fixing devices assigned to it before the robot releases it.

The tool of the connecting device is then brought into the desired position and activated so that it produces the electrically conductive and mechanical connection between the two FFCs by connecting the exposed and brought into contact with one another at the first window. The tool is then brought into the appropriate position for the connection in the region of the second window and activated, etc., etc., until the conductor tracks of all windows of the matrix are connected to one another. In this case, preference is given to the order from the smallest overlapping area to the largest, so as not to provoke any changes in the relative position of the FFC due to thermal expansions etc. in the most critical connection (s).

After all connections have been made, the finished part is gripped by the robot (or another robot), released by the fixing devices and transported away for further use.

Claims

claims:

1. Method for positioning the stripped points of two flat conductor cables (FFC) to be electrically connected and mechanically connected to one another, characterized in that the stripped points of the two flat conductor cables to be connected optically both according to their location on the respective flat conductor cable and according to their shape and size are detected so that they are placed on top of one another in such a way that the smallest of the overlapping stripped areas becomes a maximum.

2. A method for positioning the stripped points of two flat conductor cables (FFC) to be electrically conductively and mechanically connected to one another, characterized in that the stripped points of the two flat conductor cables to be connected optically both according to their location on the respective flat conductor cable and according to their shape and size detected that they are placed on top of one another in such a way that the greatest current density occurring in one of the overlapping stripped areas becomes a minimum.

3. The method according to claim 1 or 2, characterized in that the tool of the machine that produces the electrically conductive connection is placed in the middle of each overlapping stripped surface.

4. The method according to claim 3, characterized in that the center of the overlapping stripped surfaces is the center of gravity.

5. The method according to claim 3, characterized in that the center of the overlapping stripped surfaces is the center of gravity with a weighting, for example linear or square.

6. The method according to claim 3, characterized in that the center of the overlapping stripped surfaces is the intersection between the longest chord of the surface and the shortest chord of the surface.

7. The method according to any one of the preceding claims, characterized in that the connection method is a welding method.

8. The method according to any one of the preceding claims, characterized in that the first connection is made at the smallest of the overlapping stripped surfaces.

9. The method according to claim 8, characterized in that the connection is subsequently made at the next larger overlapping stripped surface.

10. The device for performing the method according to one of claims 1 to 9, characterized in that it is a carrier for the storage of the FFC, a robot for the

Handling the FFC, at least one fixation device for each of those on the carrier positioned FFC, a camera, a connecting device providing the electrical connection between the conductor tracks of the FFC and a control device for the robot and the fixing devices, which also processes the signals coming from the camera.

11. The device according to claim 10, characterized in that the Fixiervorrichrung is mounted on the carrier.