Discrimination System Using Delay Lines
Abstract Generated abstract
A discrimination method is described for extracting specified intensity or equipotential levels from time-varying detector signals, with particular relevance to flaw-detection signal processing. Instead of using separate diode limiters or Schmitt triggers with stabilized reference voltages for each level, the system employs a single zero discriminator comparing the input signal with a linearly increasing sawtooth voltage, while multiple delay lines define the desired levels through calibrated time delays. Coincidence circuitry then selects the instants at which the signal matches the prescribed levels, allowing synchronous probing of the process and display on an electron-beam tube. The authors report that level error is mainly limited by the stability of the zero device and can be reduced to about 0.1 percent, with example measurements giving a dead zone of 10 ± 2 mV and 56 mV resolution for 1 microsecond pulses.
Full Text
A Discrimination System Using Delay Lines
V. A. VOROB'EV, V. E. GOLOVANOV, A. N. LAFEROV
(Presented by the joint scientific seminar of the DCM and MRD sectors)
In the process of processing the signal from the detector of a flaw detector, in order to isolate characteristic intensity levels that make it possible to determine the desired parameters of a defect (filler material, thickness, dimensions, etc.), a device is needed that permits the time process under study to be “sliced” at the level of a given density. Such problems arise when representing the radiation-field pattern in equipotential lines.
This problem can be solved by means of diode limiters or by using Schmitt triggers as discriminators. In both cases, each level requires a separate discriminator and the creation of stabilized reference voltages. The hysteresis phenomenon inherent in the Schmitt trigger limits the range of application of this discrimination method.
Below a method is proposed for isolating the equipotential levels of a current process, containing a single discrimination element and, in this connection, possessing increased stability. The isolation of the required number of discrimination levels is carried out by means of delay lines.
The discriminator in such a device is a null organ, to one of whose inputs the process under study is applied, and to the other a linearly increasing sawtooth voltage. The instants at which the values of the input signal \(U(t)\) and the “sawtooth” voltage \(V(t)\) coincide are isolated by the null organ in the form of pulses \(\delta_u\). The position of the pulse \(\delta_u\) relative to the beginning of the sweep \(\delta_c\) is related by a linear dependence to the measured voltage \(U_t\)
\[ U(t)=\kappa \Delta t, \tag{1} \]
where
\[ \Delta t = |\delta_u-\delta_c|, \]
\[ \kappa=\frac{U_{\max}}{T} \]
for the sweep period \(T\).
Therefore, if the sweep-start pulse \(\delta_c\) is delayed by an amount \(\Delta t\), this will correspond to the value of the initial process determined by expression (1). And since the pulse \(\delta_c\) can be delayed simultaneously by several delay lines for different times \(\Delta t^i\), we obtain a series of pulses (after the mixer), the position of which uniquely determines the specified discrimination levels (Fig. 1). \(U_i=\kappa \Delta t^i\). Thus, the role of each reference voltage-
... of the discriminator in this method is played by the parameter \(\tau\) of the delay line; in this case \(\tau_{\text{dl}}=\dfrac{\Delta U_i}{K}\). In order that there be no superposition of levels during indication, it is necessary that \(\tau_i [[unclear: condition following \tau_i]]\).
The final operation of the method consists in isolating, by means of a coincidence circuit, the pulses \(\delta_u\) after the null element (Fig. 1).
The pulses of the coincidence circuit, when the signal is recorded on an electron-beam tube, are fed to the modulator and select (illuminate) those instants of the time process of the input signal at which the voltage values correspond to the prescribed reference levels. Although the measurement process takes place at a definite step determined by the sweep period of the “sawtooth,” the measurement can be made at any point of the process curve by time-shifting the “sawtooth” along the process by delaying the strobe pulse \(\delta_c\), which starts the “sawtooth.” The complete block diagram of the device is shown in Fig. 1.
Fig. 1
The period of the sawtooth scanning voltage may be determined from the condition of the required number of measurement points \((n)\) during a period equal to the maximum spectral frequency, i.e.,
\[ T_0=\frac{T_{\max}}{n}. \]
The maximum value of the duration of the operating pulses is limited by the flyback time of the “sawtooth,” which is usually
\(\Delta \tau < 0.1 T\); thus,
\[ \Delta \tau < 0.1 \frac{T_{\max}}{n}. \]
The error in the deviation of the levels from the specified value is determined mainly by the stability of operation of the zero device and can be brought down to \(0.1\%\).
Calibration of the level-setting unit was carried out by the counting method of measuring the time intervals \(\Delta t = (\delta_u - \delta_c)\) with the aid of an impulse-excitation generator filling this interval, and a pulse counter.
The advantage of this method consists, first of all, in the fact that a single discrimination element is used—the zero device—for all levels; secondly, in the simplicity of specifying the discrimination levels (\(\tau\)-delay lines); and thirdly, in the possibility of continuously “probing” the entire process with a series of specified levels moving synchronously (with a delay of the start of the “sawtooth” sweep relative to the process).
No less important is the fact that standard one-shots, sufficiently simple to adjust, can be used for the mixer and the delay lines.
Since the main units of the system determining the characteristics of the device are the zero device, the linear-voltage generator of period \(T\), and the duration of the operating pulses \(\delta\), an estimate of the capabilities can be made for specified values.
Thus, for \(\delta = 1\ \mu\text{s}\), \(f_{\text{in}} = 10\ \text{kHz}\), and with the simplest zero device, the following results were obtained:
- Dead zone — \(10 \pm 2\ \text{mV}\).
- Resolution (by coincidence pulses of duration \(1\ \mu\text{s}\)) — \(56\ \text{mV}\).