DEVELOPMENT OF MULTICHANNEL EDDY-CURRENT FLAW DETECTORS

V. K. Zhukov, B. A. Dobner

(Presented by the scientific and technical seminar of the Department of Information and Measuring Technology)

The noise immunity of an eddy-current flaw detector is an important operational characteristic that determines the reliability and dependability of inspection. It, in turn, is determined by the noise immunity of the primary transducer, which is a function of its design parameters, and by the selected method of information processing.

A number of works [1–3] have been devoted to the proper choice of the design parameters of the primary transducer. As for methods of signal processing, at present they by no means fully meet the increasingly complex methods of inspection, the mutually contradictory requirements for the speed of obtaining results and their accuracy, the capabilities of new instruments and methods, and also the knowledge that could be applied for this purpose [3]. This statement is especially true with respect to dynamic inspection by the eddy-current method. Classical methods for extracting the useful signal in this case do not always make it possible to obtain the desired result.

The principal disturbances affecting the operation of an eddy-current flaw detector are inhomogeneities of the electromagnetic properties along the axis of the inspected article, industrial interference caused by the operation of powerful technological equipment, and also vibrations and impacts of the inspected article in the inspection zone. Combating such disturbances is especially difficult when a through-type eddy-current transducer is used, when the quality of the article is judged from the results of a single scan. The probability of error in this case is considerably higher than when a surface scanning transducer is used, when the accumulation method, widely known in communication theory [4], can be used for processing the output signal.

The present work is devoted to increasing the noise immunity of flaw detectors with a through-type eddy-current transducer. Figure 1 presents a block diagram of one such flaw detector [6]. The principal difference between this scheme and the known ones consists in the presence of a second transducer and a second measuring channel \(IK_{II}\), as well as in the use of the logic circuit NOT for selecting the signal from a defect.

The operation of the flaw detector is easily understood from the plots in Fig. 2. Owing to the fact that the eddy-current transducer of the second channel \(D_2\) is displaced rela—

relative to the first, \(D_1\), the same message about the presence of a defect in the inspected article will appear in both measurement channels with a time shift \(t = d/v\), where \(d\) is the distance between the eddy-current transducers of the first and second channels, and \(v\) is the speed of motion of the inspected article. Signals from the action of the disturbances listed above appear in both channels simultaneously. This principle is the basis for extracting the useful signal against a background of random disturbances.

Fig. 1. Block diagram of a flaw detector with the NOT logic circuit

Signals from the outputs of the measurement channels, by means of the amplitude discriminators \(PU_1—PU_2\), are converted into rectangular pulses, which are normalized in duration by means of the shapers \(\Phi_I\) and \(\Phi_{II}\) and are fed to the inputs of the NOT coincidence logic circuit. The NOT circuit operates only in the case when the signal is absent at one of its inputs, which occurs when a defective section of the inspected article passes through the transducers. The output signal is recorded by the recording circuit \(SR\) (defect counter, paint marker, etc.). When a disturbance acts, the signals at the outputs of the measurement channels appear simultaneously and the NOT circuit does not operate.

Fig. 2. Time diagrams of operation of the flaw detector with the NOT logic circuit

Theoretical [2] and experimental studies of the flaw detector have shown that its noise immunity is significantly higher than that of an ordinary single-channel circuit, which makes it possible to use it for high-speed inspection of cylindrical articles.

The main disadvantage of the described circuit is that it records disturbances shifted in time. A shift of disturbance signals in time may occur, as studies have shown, because of nonidentity of the transducers, and also because of an incorrect choice of the distance between them.

A flaw detector whose block diagram is shown in Fig. 3 [5] is free of this drawback.

This flaw detector differs from the preceding one by the presence of an additional speed-measurement channel, consisting of a speed sensor \(DS\) and a speed-measurement circuit \(SIS\), at whose output a voltage proportional to the speed of the inspected arti-

tion. The circuit for selecting the defect signal consists of three monostable multivibrators and an AND cell.

If there is a defect in the inspected product, a signal first appears at the output of the first measuring channel (Fig. 4), and after a time \(t_1 = d/v\) it appears at the output of the second.

Fig. 3. Block diagram of a flaw detector with correction for the speed of motion of the inspected product

In the first channel, the defect signal will trigger the monostable multivibrator \(MM_1\), which generates a rectangular pulse of duration \(t_3 = t_1\). The monostable multivibrator \(MM_2\) is started by the trailing edge of this pulse, and a rectangular pulse delayed by time \(t_3\) will appear at its output. This pulse is fed to one of the inputs of the AND cell; to the second input is fed a pulse from the output of the monostable multivibrator \(MM_3\) of the second measuring channel. Since the pulses coincide in time, a signal indicating the presence of a defect will appear at the output of the AND cell; this signal is recorded by the registration circuit.

If, for some reason, the speed of motion of the inspected product changes, this change is recorded by the speed sensor and, after conversion by the speed-measuring circuit, changes the duration

Fig. 4. Timing diagrams of operation of a flaw detector with speed correction

of the pulse of the multivibrator \(MM_1\). If, in this case, disturbances shifted by a time \(t \ne t_3\) act on the measuring channels, then the pulses from the monostable multivibrators \(MM_2\) and \(MM_3\) will not coincide at the outputs of the AND cell, and there will be no signal at the output of the flaw detector.

The described circuits can be successfully used for inspection at high speeds, when the principal disturbances are impacts and vibration of the product.

References

1. V. K. Zhukov, B. A. Dobner. Analysis of a bridge circuit for connecting an eddy-current transducer. Defektoskopiya, 1973, No. 1.

2. V. K. Zhukov, B. A. Dobner. Two-channel flaw detector. Defektoskopiya, 1971, No. 6.

3. V. V. Klyuev, L. M. Fedotov. Application of through-type transducers and a fixed rotating electromagnetic field for flaw detection. Introscopy, Proceedings of NIIIN, issue 2, 1969.

4. A. A. Kharkevich. Essays on the General Theory of Communication. Moscow, “Energiya,” 1955.

5. V. K. Zhukov, B. A. Dobner. Electromagnetic flaw detector. Author’s certificate No. 386330, Bulletin, 1973, No. 26.

6. V. K. Zhukov, B. A. Dobner. Electromagnetic flaw detector. Author’s certificate No. 336588, Bulletin, 1972, No. 14.