DE8902911U1 - Error detection device for detecting errors in moving material parts - Google Patents

Description

-6-

Error detection device for detecting errors in moving material parts

The invention relates to an error detection device for detecting defects in moving material parts, such as material taps or bottles, according to the preamble of claim 1.

Such devices are required, for example, for quality control in the manufacture of textiles. In fabrics manufactured in flat webs, a suitably designed device of this type can be used to identify both longitudinal defects, e.g. the breakage of a warp thread, and transverse defects, e.g. the breakage of a weft thread.

The problem of detecting defects in moving material webs is not only necessary in the field of textile production, but also, for example, in the production of metal strips, foils (plastic, metal), plates (plastic _t metal), fleece and glass fabrics (especially those used in the electronics industry), paper and tufting (carpets). While the quality control of textiles is preferably carried out using transmitted light, i.e. with an arrangement in which there is a light source on one side of the material web and an optics with an evaluation device on the other side,

-7-

in the quality control of e.g. metal strips reflected light. The present invention relates in particular to a device that works with transmitted light.

In the field of textile production, circular knitting machines have recently become more and more popular. These are automatic knitting machines equipped with several thousand needles that produce tubular knitted material. The tubular material moves from the circular knitting machine in the lengthwise direction of the tube while simultaneously rotating around the longitudinal axis. Since the machines are extremely expensive and downtimes should therefore be kept as short as possible, it is not only necessary to identify faults as quickly as possible, but also to differentiate between faults; not all faults require the machine to be switched off and repaired. In addition, it should be avoided that too much faulty, unusable knitted material is produced before a fault is identified, e.g. caused by a broken knitting needle.

Defects that typically occur in circular knitwear are so-called x^angular defects, which are defects that run in the direction of rotation of the material (ring or spiral), as well as transverse defects that result from a defective knitting needle, for example, and also point-shaped defects, i.e. either bumps (thickenings) or short holes.

A device for detecting various errors of the type mentioned above is described, for example, in DE-PS 29 25 734. In order to be able to distinguish between different types of errors, it is proposed

-8th-

to equip the rotated mirror wheel with two alternately effective types of mirror surfaces, whereby each type of mirror surface produces a specific light flake shape in the scanning plane (i.e. in the plane of the material web).

Apart from the fact that the manufacture of special mirrors suitable for the respective requirements is complex, the processing circuit must also be specially designed for each type of mirror.

US-PS 3,797,943 discloses a device for detecting defects in a moving flat material web, in which an aperture plate with three parallel aperture slots of different lengths is used and the images of the various aperture slots are projected from different photodiodes. The material web is scanned across its entire width with all three aperture slots.

The detection of defects in the moving material web only occurs in the longitudinal direction and inhomogeneously across the material width. This is because the sensitivity of the defect detection changes from the middle to the edge of the material web, and detection in the middle is sharper than at the edge.

The material to be tested must be guided very precisely, as there is no depth of field. Even the slightest fluttering of the material parts to be tested immediately distorts the measurement. The lack of depth of field is caused by the lack of parabolic

-9-

mirrors or Fresnellinae.

The detection of transverse faults is not possible here, since transverse faults only cause a level change around one reference point in '-or the total signal width of the detection.

The darkening or brightening over the entire width of the material (transverse error) causes a DC voltage change over the entire scanning width. Such DC voltage changes are eliminated by the downstream differentiator (Fig. 5(a), 5(b) of US Patent 3,797,943).

The invention is based on the task of creating a device of the type mentioned at the beginning, which requires the simplest means and yet can be used in a variety of ways with the most varied types of material parts to detect the most varied types of defects.

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

The device according to the invention can be used with a conventional, simple mirror wheel. Differentiation between different types of errors is achieved by splitting the optical signal into two partial beams. While a slit diaphragm arranged in the transverse direction can be arranged in one partial beam path, a slit diaphragm oriented in the longitudinal direction can be used in the other partial beam path.

-�IΩ-images of the two slit apertures then lead to a cross-shaped scanning spot in the plane of the material web. This makes the device according to the invention particularly suitable for use in circular knitting machines. However, the invention is in no way limited to such an application. The invention can also be used in conjunction with looms.

Apart from the relatively inexpensive polygon mirror wheel, the other elements of the device according to the invention consist of commercially available, inexpensive components. For example, the beam splitter can be a commercially available partially transparent mirror, as is generally known for various optical applications.

In order to take account of the specific types of errors to be detected, an amplifier/filter circuit is assigned to each partial beam path, the parameters of which are selected in such a way that the most favorable signal processing is achieved in each case. The amplifier/filter circuit can, for example, contain a bandpass filter to filter out certain interference frequency ranges. In addition, one or two lowpass filters can be provided which optically only let through that range of the spectrum that can be evaluated for error detection.

In particular, the invention provides that the output signals of the amplifier/filter circuit are fed to a multiplier. The product signal is characteristic of long transverse faults. Such transverse faults result from defective knitting needles in circular knitting machines and from defective weft dispensers in fabrics.

-11-

By appropriately differentiating and subtracting the signals obtained through the two partial beam paths, it is possible to ensure that the interference signals cancel each other out and only the useful signals are added together.

A particular advantage of the present invention is that a commercially available

r~Ohr6 Uses WGTdSH &Kgr;&agr;&Pgr;&Ggr;&idigr;, dis Slit SliiST

AC voltage with a frequency of about 50 kHz. All known fault detection devices of the type in question today work with a laser light source. The light emitted by the laser light source is monochrome light, which is not suitable for colored and patterned materials. In contrast, the fluorescent tube used in the invention is characterized by its broad usable spectrum and its considerably lower price. It was recognized that if the frequency of the AC supply voltage is high enough, the crosstalk _— the chromatic aberrations—that are otherwise disruptive in fluorescent tubes do not occur. In other words: if the operating frequency is high enough, the fluorescent tube behaves like a light source fed with direct current.

The detection of transverse defects (pinstripes, weft thread breakage or double weft) is carried out according to the principle of comparing two scans (longitudinal and transverse aperture), which are absolutely synchronous in terms of location and time and, however, require an absolutely exact parallel course of the transverse aperture under the scanning device. This parallelism of the transverse aperture cannot be guaranteed in practice for web widths of up to 2000 mm.

-12-

because stresses in the web material always cause a sickle-shaped course of the transverse error.

A preferred embodiment of the invention to solve this problem consists of a combination of individual defect detection systems according to the invention with a short scanning area over the entire width of the material. The scanning width of a system is preferably about 150 mm x 15 mm, so that a transverse defect with an inclination of 15 mm from edge to edge of the scan is detected over a width of 150 mm.

The invention can also be used in sorting systems to sort different objects by color. The solution specified in claim 14 provides for the detection of different colors of the objects using different color filters in the partial beam paths in order to obtain a criterion, for example, for separating out objects of a certain color. The error detection device according to the invention can be used economically, particularly when the bottles are being recycled. By scanning the individual bottles in their lengthwise direction, a meaningful signal is obtained which is characteristic, for example, of green or brown or white bottles. Bottles of a certain color can be collected using these signals.

A device for automatically distinguishing between brown and green glass bottles is already known from US-PS 4 076 979, whereby light sent through the bottle to be tested is reflected by a semi-transparent mirror on the one hand by a

-13-

green-permeable filter on a first photocell and on the other hand through a brown-permeable filter on a second photocell. However, in this case the scanning of the bottles is static and point-like. Therefore, an exact adjustment must be made according to the bottle size and bottle shape. This means that the bottles must be sorted beforehand.

The scanning according to the invention, however, preferably takes place at a total height of 150 mm and a width of 15 mm. This means that no pre-sorting is required and there is no dependency on bottle size and shape.

Furthermore, three-color recognition can be carried out by a downstream processing circuit.

While in the device according to the US patent, isolated external influences such as labels, burned-in writing or characters distort the recognition or make it impossible, these external influences are completely eliminated in the error detection device according to the invention by the scanning and the processing circuit.

• I · I , I ♦ · · *

-14-

In the following, embodiments of the invention are explained in more detail with reference to the drawing. They show:

Fig. 1 is a partially schematic view of a device for detecting defects in a moving material web,

\ Fig. 2 a detailed block diagram of the in Fig. 1

' schematically shown amplifier/filter

circuit and the processing circuit,

Fig. 3 shows a modified embodiment of a processing circuit compared to Fig. 2,

Fig. 4 is a schematic diagram illustrating different types of defects in a material web, and

Fig. 5 shows an example of a video error signal generated by the device.

First, reference is made to Fig. 4. Fig. 4 shows a section of a tubular material web. This is a knitted product produced using a circular knitting machine. Due to the special type of circular knitting machine, the tube moves in a spiral shape, as indicated by the arrow. Depending on the cause of the error, e.g. defect in one or more knitting needles, defect in the yarn or similar, the errors take on a variety of forms. There are transverse errors Q, which run transversely to the direction of movement of the material web 1, long longitudinal errors running parallel to the direction of movement.

-15-

defects L, thickenings (bumps) D and holes appearing as light spots H.

The tubular material web 1 is shown in Fig. 1 top left. It moves in its longitudinal direction according to the arrow PL and in the circumferential direction according to the arrow PR, so a given point on the material web 1 moves on a spiral path.

Inside the material web 1 coming from a circular knitting machine (not shown here) there is a commercially available fluorescent tube 2 as a light source, which is fed with a high-frequency alternating current with a frequency of 50 kHz.

The (white) light emitted by the fluorescent tube 2 falls on the inner surface of the material web 1 and passes partially through the material to a Fresnel lens 3 arranged outside the material web 1, which focuses the light on a specific point on a facet F of a polygonal mirror wheel 4 rotated in the direction of the arrow P-.

While the facet F of the mirror wheel 4 struck by the bundled light rotates a distance corresponding to the length of a facet, a line in the plane of the material web 1 is scanned in accordance with the rotation. The light reflected by the facet F of the mirror wheel 4 in the direction of the arrow P. onto a collecting lens 5 therefore corresponds at any time to the point of light which has passed through the material 1 at a specific point along the scanning line.

-16-

From the collecting lens 5, the light reaches a beam splitter 6, which lets through or deflects one half or part of the light along a partial beam path A, and the other half or the remaining part of the light along a partial beam path B.

The partial beam path A comprises an optical filter 2.5, a green filter ?A _; 5ovi« and a slit diaphragm 8A. The light emerging from the slit diaphragm A falls on a photodiode 9A which serves as a photoelectric converter device and which converts the light into an electrical signal. The output signal of the photodiode 9A reaches an amplifier/filter circuit 10A.

The partial beam path B is similar to the partial beam path A and contains a color filter 7B, for example also a green filter, a slit diaphragm 8B and a photodiode 9B, to the output of which an amplifier/filter circuit 1OB is connected.

The slits of the slit apertures 8A and 8B are oriented such that their image on the plane of the material web 1 defines a cross-shaped scanning spot 11.

The output signals of the amplifier/filter circuits 10A and 10B reach a processing circuit 100, which forms error detection signals from the signals.

Fig. 2 shows the structure of the amplifier-Zfilter circuit 10B in detail. The circuit 10A is constructed in the same way as the circuit 10B.

-17-

The amplifier-Zfilter circuit IOB contains a pre-amplifier 12, a bandpass filter 13, an amplifier j _: \

14, a low-pass filter 15, a peak detector 16 connected to the output of the low-pass filter 15 and an automatic gain control circuit (AVR) 17 connected to the output thereof, the output of which is connected to the amplifier 12. A further low-pass filter 18 is connected to the peak detector 16.

The output signal of the low-pass filter 15 is output via a line 19 as a video signal for cross errors at the output of the processing circuit 100. The output signal of the low-pass filter 18, which has a different characteristic curve than the low-pass filter 15, is fed as the first input signal to a multiplier 23 within the processing circuit 100.

The output signals of the amplifier/filter circuit 10A are passed via a line 22 as a video signal for longitudinal errors to the output 28 of the processing circuit or via a line 21 as a second input signal to the multiplier 23. The two multiplied input signals are output by the multiplier 27 at an output 24 as a transverse error signal.

The signal on line 22 is inverted in a negator 29 J and subtracted by means of an adder and differentiator 26. This creates a sum signal from the video signal for longitudinal errors V. and the video signal for transverse errors V ..

-18-

Through the inversion and addition, which results in a subtraction, the signal received at the output 27 is largely freed of higher frequency noise components. According to Fig. 2, the signal available at the output 27 has a front and a rear signal peak, which correspond to the beginning and end of a sampling period corresponding to a facet F of the mirror wheel 4. n<\s pulse-shaped signal lying between these two peaks corresponds to an error.

If one looks at the scanning spot 11 in Fig. 1 and the possible types of errors, it can be seen that, depending on the specific type of error, the error signal obtained occurs either in a large number of consecutive sampling periods or only in a very small number of consecutive sampling periods. In the case of a transverse error, a relatively long signal is generated in one partial beam path and the circuit connected to it, and a relatively short signal in the other partial beam path, because the slit diaphragm in this partial beam path runs transversely to the direction of the transverse error.

In the case of longitudinal defects, only very short signal pulses are generated. By appropriately selecting the time constants of the low-pass filters 15 and 18 in the two amplifier filter circuits 10A and 10B, it is possible to achieve a sufficient differentiation between transverse defects and longitudinal defects as well as between short and long defects in the material web 1.

«· λ Q ·μ

Fig. 3 shows a modified embodiment of a processing circuit 100'. The multiplier 23 corresponds to the multiplier 23 shown in Fig. 2. It is used here to determine a transverse error.

In addition to the differential element 101, which corresponds to the differentiating element 26 in Fig. 2, a second differentiating element 103 is provided in the circuit 100' according to Fig. 3, by means of which the second derivative dv2/dt2 of the sum signal (V1 + _VQ ) is formed. An integrator 107 is connected between the two differentiating elements 101 and 103. The first differentiation ensures that only the steep edges of an error signal are allowed to pass, but not the interference signals. However, the polarity of the error signal is thereby distorted. The integrator 107 attenuates high-frequency components in the output signal of the differentiating element 101 forming the first derivative. The second differentiating element 103 connected downstream of the integrator 107 cancels out the polarity distortion caused by the first differentiation.

From the output signal of the differentiator 103, an analysis device 104, which contains 5 processors CPU1-CPU5, determines error signals for transverse errors, longitudinal errors and point errors. A computer 105 is connected to the analysis device 104, which evaluates the error signals, if necessary, activates a display device or an alarm device,

&bull; » 9 · ft ·

-20-

All optical, mechanical and electrical components of the error detection device according to the invention are commercially available, inexpensive components.

The optical filters 7A and 7B shown in Fig. 1 can be _selected according to the color of the material of the material web 1 in order to obtain color signals A and B in the channels corresponding to the partial beam paths.

With reference to Fig. 5, the function of the scanning device according to the invention is explained again.

Fig. 5 shows video error signals provided by the scanning device in both the transverse and longitudinal error channels.

T - Inspection time

t - dead time (mirror surface change of the polygon mirror wheel)

A =· Amplitude level of the signal

T + t - effective time of a mirror surface of the polygon mirror wheel

When the mirror surface changes, the receiver (photodiode) is illuminated with twice the amount of light compared to the inspection time. The scanning produces the amplitude "A ¹¹ " in the video error signal at this change point, regardless of the aperture. The DC voltage value of this amplitude "A" is determined by a downstream peak detector 16 (Fig. 2).

-21-

The peak value determined is passed on to two circuits; firstly to the AVR circuit 17 (Fig. 2) for automatic amplitude control and secondly the signal passes through a low-pass filter 18 (Fig. 2) to a multiplier 23 (Fig. 2) with differential input.

The low-pass filter 18 only allows peak value changes to pass which have been caused by truer mirror surface changes.

In the case of a transverse error, the peak value change of the amplitude ¹¹ A" of the photodiode with transverse aperture is considerably higher than the peak value change of the photodiode with longitudinal aperture. The signal is switched from the respective low-pass filter 18 of the corresponding circuits 1OB or 1OA (Fig. 2) to the differential inputs of the multiplier 23.

In the multiplier 23 with differential inputs, equal parts of the two signals 1 on the lines 20 and 21 are eliminated and only the actual amplitude change of both peak values is multiplied by a constant factor.

This creates a signal at the output of the multiplier that clearly indicates a transverse error.

Claims (17)

Expectations 1. Error detection device for detecting errors in moving Material parts, with a light source (2) on one side and an optics (3-8) containing a rotated mirror wheel (4) on the other or the same side of the material piece, a photoelectric converter device (9A, 9B) which is arranged in the beam path of the light rays reflected by the mirror wheel, and a processing circuit (10A, 10B, 100; 100') which is coupled to the output of the converter device (9A, 9B). t · -2- characterized in that the optics have a beam splitter (6) which divides the light beam coming from the mirror wheel (4) into at least two partial beams, and that in each partial beam path (A? B) there is arranged a diaphragm (8A, 8S) and/or an optical filter (7A, 7B) which differs from the diaphragm or the optical filter in the other partial beam path with regard to diaphragm opening shape or filter characteristic. 2. Device according to claim 1, characterized in that the beam splitter is a partially transparent mirror (6). 3. Device according to claim 1 or 2, characterized in that the diaphragms (8A, 8B) in the two partial beam paths (A, B) are slit diaphragms, the images of which on the material web (1) are essentially perpendicular to one another and define a scanning spot (11). 4. Device according to at least one of claims 1 to 3, characterized in that an amplifier/filter circuit (ioA, 10B) is assigned to each partial beam path. 5. Device according to claim 4, characterized in that a multiplier (23) is coupled to the output of the amplifier-Zfilter circuit. 6. Device according to claim 4 or 5, characterized in that a subtraction and differentiation circuit is coupled to the output of the amplifier/ ^feeder circuit (10A, 10B) subtracted and differentiated. -3- 7. Device according to at least one of claims 1 to 6, characterized in that the light source is a fluorescent tube (2). 8. Device according to claim 7, characterized in that the fluorescent tube is connected to an alternating voltage source with a frequency of or more than 50 kHz. 9. Device according to at least one of claims tgr; to 8, characterized in that the game wheel is polygonal. 10. Device according to at least one of claims 1 to 9, characterized in that a Fresnel lens (3) is arranged between the material web and the mirror wheel (4). 11. Device according to at least one of claims 1 to 10, characterized in that the VexTäifbeiuünCjääChäluüriy (IGO; IGG) a longitudinal error, a transverse error and a point shape error detector circuit. 12. Device according to at least one of claims 1 to 8, for sorting material parts, for example bottles intended for reuse, according to color, characterized in that an optical filter (7A, 7B) is arranged in each partial beam path, which differs from the optical filter in the other partial beam path with regard to color permeability. 13. Device according to at least one of claims 4 to 12, characterized in that that each of the amplifier/filter circuits (10A, 10B) has a peak value detector (16) and a low-pass filter (18) connected downstream thereof. 14. Device according to claim 13, characterized in SGlCiinSi-, &ugr;&Xgr;&ugr; VjBITi opi L.ZSnVier"«.<uetäjCtuir ^1S) a StöüöiT&mdash; adjustable amplifier (12) is connected upstream, and that an AVR circuit (17) is connected between an output of the peak detector (16) and a control input of the amplifier (12). 15. Device according to at least one of claims 4 to 14, characterized in that the multiplier (23) has a differentiation input. 16. Device according to at least one of claims 4 to 17, characterized in that the outputs of the two amplifier/filter circuits (10A, 10B) are connected to an adder Nogrier element (29) are connected. 17. Device according to claim 18, characterized in that the differentiating element (101) is followed by an integrator (107) and a further differentiating element (103).