JPS5942802B2 - Method for detecting deviations that occur in electrical signals at repetitive periods - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for detecting periodically occurring deviations in a signal representing instantaneous variations in the transverse dimension of a moving endless element, such as a woven yarn sensing signal.

British Patent No. 795,808 discloses a method and a device for detecting periodic phenomena of variable size, particularly in the material cross-section of woven yarns.

This method is based on frequency analysis of electrical signals representing successive values of the above magnitude. Typically, such electrical signals have, in addition to frequencies characterizing periodic phenomena, a wide continuous frequency band due to aperiodic events and irregular transient phenomena. Therefore, it is almost impossible to reliably separate the periodic signal from the original electrical signal in the filtering process, even with the use of fairly large and expensive equipment. The main object of the invention is therefore to provide a method for the examination or recognition of periodically occurring deviations in the sensing signal, with which such deviations and periodicities can be reliably evaluated. Another object of the invention is to convert the original sensed signal into a signal pulse sequence or train of signal pulses and convert it into a reference pulse having a variable pulse period and having a variable phase relationship with respect to said signal pulse sequence. The object of the present invention is to provide a method for sequentially scanning signals.

A more particular object of the invention is to detect periodic thickness variations of very small amplitude which cannot be reliably detected by the methods and apparatus commonly used for thread guiding. The object of the present invention is to provide a new and improved method for detecting so-called moire patterns or moire effects in rotor-spun yarns, such as those present in rotor-spun yarns.

According to the novel method of the invention, the detection device is simplified based on the fact that a pulse sequence is derived from the original sensing signal and that this pulse sequence, rather than the original sensing signal, is processed. Ru.

The term "pulse sequence" is used herein to refer to a non-periodic series of single pulses or individual pulses, as opposed to a periodic pulse train. Another advantage of the invention is that the periodic phenomena of the original signal appear in the pulse sequence in a simplified form and can therefore be detected by their periodicity alone. The detection of periodic phenomena is thus improved to such an extent that even such phenomena, which have a small amplitude compared to the original signal and therefore cannot be detected visually, can be recognized. The above-mentioned and other objects of the invention will become apparent from the following detailed description taken in conjunction with the accompanying drawings. The new and improved solution of the present invention is used to detect only periodic deviations from a series of all cross-sectional or thickness deviations occurring in a moving thread.

For this purpose, the yarn movement signal F' (see FIG. 1a), which is obtained by sensing the moving yarn or the like and represents the varying instantaneous value of the cross section or diameter of the yarn, has a certain threshold value S. Pulse sequence S representing the peak value of the thread movement signal F'
' (see Figure 1b). The pulse sequence S' preferably consists of rectangular pulses of constant amplitude and different pulse widths and is stored in a memory device. In a first step, the pulse sequence S'
is a needle pulse Zn(n
21, 2, 3, . . . ) pulse train Z' and compared with the pulse train Z'.

In a subsequent step, the phase of the limited pulse train Z' is shifted by a small step Dn. In each of these limited pulse trains obtained from z', the number of coincidences of individual pulses within a predetermined length section of S' is determined, while also the Z' train is determined. From this number of matches or "coincidences" it is possible, using certain criteria described later, to recognize whether there is a periodicity in the pulse sequence S'. If the first test, which includes a large number of such steps, does not pass, one or more other tests, each having a pulse period varied by an increment Dp with respect to the period of the preceding pulse train, are performed. The process described above can be repeated by selecting the pulse train Z'. However, the variation in the duration of the pulse period p should be limited to a suitable range based on experience, for example when testing rotor-shaped yarns. In this case, the length of the cycle is approximately the circumference of the rotor, and the cycle of the first series of tests can be determined based on the size of this circumference. One can determine approximately the correct value of the pulse period p from one trial, i.e. one production trial carried out previously.One series of five trials can be compared to the time required to generate the signal pulse sequence S'. can be carried out within a very short time, so that even in the case of textile machines in operation, e.g. The same signal pulse sequence S' can be scanned by the same test series.

FIG. 1d shows the phase shift of the limited pulse train Z' by one step Dn, and FIG. 1e shows the variation or increment DP of the pulse period p.

In the following description, the pulse sequence S' will be referred to as the signal pulse sequence, while the limited pulse train Z
′ will be referred to as a reference pulse string. A detailed explanation will be given below with reference to FIG. In FIG. 1a, the line M represents the time-averaged value of the cross-section or diameter of the thread, while the dashed line defines the threshold value S on which the formation of the signal pulse sequence S' is based. As shown in FIG. 1b, one individual pulse Sm (m=1, 2, 3...) of the signal pulse sequence S' has the time coordinates Am and the leading and trailing edges of the pulse. Em
determined by. These time coordinates are measured from an arbitrarily chosen time zero TO. The needle pulse Zn of the reference pulse train Z' is a multiple (integral multiple) of the time coordinate d and pulse period p of the first needle pulse Z1.
determined by. FIG. 1d shows a phase shift of the reference pulse train Z' by one step Dn. As the scanning progresses, the phase shift amount changes by steps Dn, 2dn, 3dn, etc. sequentially. The step Dn should be chosen as a fractional value of the pulse period p. For example, it should be selected within the range of 0.1-0.01p. In some cases, the pulse period p
is an integer multiple of the increment Dp, 2dp for each value of phase shift.
, 3dp, and so on. The value S (FIG. 1a) is automatically set such that in the signal pulse sequence S' the relationship between the average pulse intensity and the average pulse pause period, ie the pulse impulse factor, is approximately 10%, for example.

In order to detect periodic pulse trains within the signal pulse sequence S'', it is advantageous to select a section or part of S' that includes approximately 10 periods.

One of the pulse trains S' or Z' that are phased relative to each other during the scanning process should contain at least one extra period of pulses. This means that if a fixed part or section of S' having, for example, 10 periods is to be scanned with a pulse train z', the pulse train Z' should have at least 11 periods; The reverse is also possible. The electronic device shown in FIG. 2 is an embodiment for implementing the method of the invention described above.

The electronic device is equipped with evaluation channels 1-6, in which a signal pulse sequence S' is derived from the yarn movement signal F'. Furthermore, a minicomputer 7 is provided, in which the signal pulse sequence S' is
Scanned to detect periodically occurring pulses.
Furthermore, a display device 14 is provided. A sensing device or sensor 1 is provided to sense the thread F during movement.

The sensor 1 can be a known transducer, such as a capacitive or photoelectric transducer, with a measuring field through which the thread F can pass. The sensor 1 generates a sensing signal representing successive values of the diameter or cross-section of the thread F, and the amplifier 2, for example an AC amplifier, generates a sensing signal representing the fluctuations of the thread movement signal F shown in FIG. 1a. ’. Thread movement signal F
' is supplied as a signal input of a comparator 4 arranged in the evaluation circuit 3. A bias voltage is applied to the second input of the comparator 4, which defines a threshold value S (see FIG. 1a). The magnitude of the bias voltage is adjusted by a feedback loop consisting of a series connection of an integrator 5 and a DC amplifier 6 connected to the output of the comparator 4. When the thread F is moving, the output terminal of the comparator 4 receives a signal pulse train S as shown in FIG. 1b.
By suitably setting the gain of the DC amplifier 6, the threshold S of the comparator 4 is determined by a predetermined relationship between the average pulse width and the average pulse pause period for the signal pulse sequence, e.g. 1: 10 or 10%.The signal pulse train S' is controlled by the minicomputer 7.
A signal storage device 8 is provided within the signal storage device 8.

The minicomputer 7 further comprises a reference storage device 9, a numerical comparator 10, a counting and addition device 11, a timing circuit 12 and a control device 13. The storage device 8 stores time coordinates Am and Em (see FIG. 1b). The reference storage device 9 stores the time coordinates d+Kp (k=0, 1,
2, 3...) are stored. In a numerical comparator 10, the signal pulse sequence S' is compared with a reference pulse train Z'. This means testing whether the condition is true for individual values of m and k within a predetermined portion of the signal pulse train S' that includes approximately 10 periods.

If there is a match, it is considered a coincidence. In the counting and summing device 11, the number of coincidences T is evaluated for each test sequence or signal pulse sequence S'(7) for each scan, and for this purpose the number of coincidences Tm
compared to

As a first criterion for a periodic yarn error pattern, for example, the occurrence of a moiré effect, it is determined whether the following relationship is realized.

Tm should be set larger than the average value of the values T occurring in all comparison stages. If a moiré effect occurs within a portion of the yarn movement signal F' that includes 10 periods, approximately 8
and 11, and in the absence of moiré effects, the value of T is in the range 1 to 5.

In this case, the minimum coincidence Tm can be fixed at 7. The second criterion is for increasing the accuracy of moiré detection.

This second criterion is as follows. That is, in the scanning process, as soon as the first criterion is satisfied, consecutive coincidences T>Tm are added, resulting in a sum of coincidences S(1). If the sum of the coincidences is larger than a predetermined minimum value Sm(T), it is assumed that the second criterion is satisfied. In this case, the measuring pulses are applied to a display device, which can be, for example, a counter. If the first criterion or two criteria are not realized, the reference pulse train Z/
The value of the period p is varied in the reference storage device 9 by an amount DP and the scanning process described above is repeated using this new value.

This process is repeated until either the second criterion is satisfied and a moiré effect is displayed, or the period p has been varied over the range in which periodic errors can be expected from experience. The control device 13 controls the entire program described above.

The control device 13 includes a command storage device in which program steps and numerical constants Tm, Sm(1), D, dp, Hd and Hp are stored. Hd is the deviation of magnitude d in increments of steps Dn and thus also the deviation of the phase of pulse Zn of the reference pulse train Z' with respect to the signal pulse sequence S'(7) pulse Sm;
Hp is the deviation of the period p varied by successive increments Dp. Timing circuit 12 is used to time the various program steps of controller 13. The following table shows exemplary numerical examples of the method of the invention.

Yarn sensing signal F' is generated from rotor cotton spinning exhibiting moderate moiré effects. The length of the thread portion was 11 times the peripheral length of the groove of the spinning rotor. That is, it corresponds to 11 periods p. The reference pulse train Z' contains 10 periods p and therefore 11 pulses Zn. This means that pulse train Z' is scanned by pulse sequence S'. The deviation Hd was equal to one period and occurred at 60 steps Dn. The first criterion is T>
It was 7. In the table below, n means the number of scanning steps Dn.

The second column of the table shows T belonging to each step, and the third column shows the sum S(T) taken from consecutive steps when Tm=7 is realized for the first criterion T>Tm. The sum is 74 for 8 consecutive steps, thus clearly indicating the presence of a moiré effect. Although the novel method of the invention chooses rectangular pulses as the pulses of the pulse sequence or pulse trains S' and Z', it is by no means limited to the signal and pulse shapes as shown in FIG. . For example, the pulse sequence S' may include a needle pulse that coincides with the midline of the peak value above the threshold S, as shown in FIG. 1a, for example.
The pulse train ZI consists of pulses or needle-like pulses, while the pulse train ZI consists of rectangular pulses of equal constant width such that the relationship between the pulse width and the pulse pause period is 1:10 as described above for the pulse sequence S'. It can be constructed from. Although preferred embodiments of the present invention have been described above with reference to the drawings, the present invention is not limited to these embodiments, and various embodiments and specific examples can be made without departing from the scope of the present invention. It goes without saying that there is.

[Brief explanation of drawings]

1 is a signal waveform diagram illustrating the steps of the detection method of the present invention, and FIG. 2 is a block diagram of the electronic equipment used to carry out the method of the present invention. F...Thread, 1...Sensor, 2...AC amplifier, 3
...Evaluation circuit, 4...Comparator, V...Bias voltage, 5...Integrator, 6...DC amplifier, 1-6...
・Evaluation channel, 7...Mini computer, 8...
- Signal storage device, 9...Reference storage device, 10
... Numerical comparator, 11... Counting/adding device, 12.
...Timing circuit, 13...Control device, 14...
Display device.

Claims (1)

[Scope of Claims] 1. A method for detecting a deviation occurring at a certain repetition period in an electric signal representing an instantaneous value of a diameter variation of a thread-like element during movement of a textile machine, comprising: (a) setting the electric signal to a certain threshold value; converting it into a series of signal pulses representing the electrical signal peak value exceeding
(b) defining a reference pulse train having a limited number of said certain repetition periods; and (c) defining said reference pulse train as a portion of said signal pulse train having a fixed phase relationship with respect to said reference pulse train in a first comparison step. (d) further comparing said reference pulse train with a portion of said signal pulse sequence in a plurality of comparison steps, where each comparison step includes a comparison preceding said one portion of said certain repetition period; (e) estimating the number of pulse coincidences that occur in each comparison step, using a phase relationship that is different from the phase relationship of the steps; A method comprising the step of detecting as representative a periodic deviation in the electrical signal. 2. The predetermined large number is determined as a number T that is equal to or larger than the minimum value Tm exceeding the average value of all the numbers T occurring in the comparison step. Method. 3. A predetermined large number is determined as a number T that is equal to or greater than the minimum value Tm that exceeds the average value of all the numbers T occurring in said comparison step, and said large number occurring in successive comparison steps. The method according to claim 1, which calculates the sum of . 4. The method of claim 1, wherein the threshold value is selected such that the duty cycle in the signal pulse sequence takes on a predetermined cycle. 5. The method of claim 4, wherein the predetermined value of the duty cycle is approximately 0.1. 6. The method of claim 1, wherein the pulses of the signal pulse sequence have essentially equal amplitudes. 7. The method of claim 1, wherein the reference pulse train includes at least 10 repetition periods and the duration of the repetition periods is determined from previous runs of the textile machine.