UNIVERSAL TESTING SETUP

M. S. Roitman, A. I. Kramnyuk, V. I. Sedov

(Presented by the scientific seminar of the Department of Radio Engineering)

The continuous quantitative growth of electrical measuring instruments and the expansion of the functions they perform leads to a sharp increase in the stock of testing devices. And this is already turning out to be very disadvantageous from an economic point of view. Therefore, the creation of universal testing setups that provide comprehensive testing is a timely task. The present work is devoted precisely to a brief description of such a setup, developed at the Department of Radio Engineering of TPI. The setup is intended for testing active and passive multi-terminal networks, electronic voltmeters, phase meters, etc. It consists of a combination of a precision four-phase generator with automatic stabilization of amplitude and of a 90° phase shift of type GK-9, a system of broadband inductive dividers of type DI-2, and a vector-meter device with high resolving power and small capacitance with respect to “ground.”

For testing electronic voltmeters, any of the four outputs of the generator may be used (Fig. 1). The use of negative feedback with respect to the envelope of the output voltages of the generator, together with an rms-value comparator in the circuit of this feedback, ensures the setting of the output voltage of 5 V with an error of \(< 0.05\%\) and instability of not more than \(0.01\%\). In the generator, all possible measures have been taken to reduce nonlinear distortions, which do not exceed \(0.1\%\) at frequencies of \(20 \div 100\) kHz and \(0.05\%\) at frequencies of \(20\ \text{Hz} \div 20\ \text{kHz}\). All the above makes it possible to test voltmeters up to class 0.5 inclusive. The use of precision broadband inductive dividers makes it possible to perform this testing at different points of the scale and on different ranges.

Fig. 1. Block diagram for testing electronic voltmeters: GK-9—four-phase generator-calibrator, DI-2—inductive divider, V—voltmeter under test.

Small nonlinear distortions of the generator output voltage make it possible not only to realize more fully the potential capabilities of the compensation method for measuring nonlinear distortions of measuring amplifiers [1], but also to extend the range of measured values toward their reduction.

Thanks to the presence of the second channel, which forms the quadrature component of the voltage and consists of a quadrature phase shifter and a power amplifier, the setup can be used for verifying phasemeters.

Usually, in two-phase generators without frequency conversion, the accuracy of setting the voltage quadrature is ensured by a phase shifter through the use, in its circuit, of high-precision and highly stable elements. For these purposes we have used a method of automatic stabilization of the \(90^\circ\) phase shift. The stabilization circuit (Fig. 2) contains a phase comparator and a DC amplifier. The phase comparator produces a control signal when there is a deviation of the phase angle from quadrature, regardless of the causes of the latter. After appropriate amplification, the control signal acts on the controlled resistance of the phase shifter, changing its value in such a way that the deviation from quadrature is reduced.

Fig. 2. Block diagram of stabilization of a \(90^\circ\) phase shift: QPS — quadrature phase shifter, DCA — DC amplifier, PC — phase comparator.

This method makes it possible to sharply reduce the error in setting the quadrature. Its potential capabilities are limited by the errors of the phase comparator and of the sum-and-difference circuits used as output units of the generator or in the phase comparator. Using the inductive dividers included in the setup, one can obtain practically any phase shifts in the range of angles \(0 \div 360^\circ\) and thus verify phasemeters in the frequency range \(20\ \text{Hz} \div 100\ \text{kHz}\). The presence of four phases of the generator output voltage and of accurate dividers makes it possible, in addition to phasemeters, to verify various phase-measuring units, such as phase detectors, correlators, quadrature indicators, etc.

Each of the indicated instruments has a certain error that does not depend on the quadrant in which the voltages being compared are located. Therefore, if voltages with phase angles \(0\) and \(90^\circ\), then \(90\) and \(180^\circ\), \(180\) and \(270^\circ\), and finally \(270 \div 360^\circ\), are first applied to their inputs, then in each measurement an error will be recorded that is the sum of the error of the quadrature-setting setup and the error of the device being verified. The sum of all these resulting errors is four times the error of the device being verified, since the sum of the errors of the quadrature-setting setup for each quadrant is zero. As we see, when this verification method is used, the error in setting the quadrature of two voltages does not affect the verification result. It is interesting to note that, by the same scheme, one can also verify the error in setting the quadrature of each quadrant of a four-phase generator. After determining the intrinsic error of the quadrature indicator, the error in setting the phase shift of the generator is determined as the difference between the readings of the quadrature indicator and its intrinsic error. This method has the great advantage that it does not require precision quadrature indicators.

The inclusion in the setup of a vector-meter device with high sensitivity and selectivity makes it possible to considerably expand the range of application of the setup, verifying with it various four-terminal networks, both active and passive. The vector-meter device consists of a scale amplifier for increasing sensitivity, two phase detectors, and an rms voltmeter (Fig. 3).

The high selectivity of the vector-meter device is ensured by choosing the appropriate phase-detector circuit. The studies we have carried out make it possible to state that similar results can be obtained both when using sum-and-difference phase detectors based on comparison of voltages by rms value, and when using phase detectors employing integrated dc differential amplifiers. The latter are even preferable, since they have considerably lower current consumption. When compensation methods of measurement are used, the vector-meter device is “suspended” with respect to “ground.” Therefore, in order to reduce measurement error, especially at the upper operating frequencies, it must have a small capacitance with respect to “ground.”

Fig. 3. Block diagram of the vector-meter device: FD1, FD2—phase detectors; VAV3—rms voltmeter

Technical characteristics of the setup

1. The setup provides four voltages of \(5\ \mathrm{V} \pm 0.05\%\) at frequencies of 20, 45, 100, 400, 1,000 Hz, 5, 10, 50, 80, and 100 kHz, with phase angles of 0, 90, 180, and 270°.

2. The short-term instability of the output voltages at all frequencies does not exceed \(\pm 0.01\%\).

3. The error in setting the voltage quadrature in one quadrant is not more than \(\pm 0.01^\circ\) in the frequency range \(20\ \mathrm{Hz} \div 20\ \mathrm{kHz}\), and not more than \(\pm 0.03^\circ\) in the frequency range \(20\ \mathrm{kHz} \div 100\ \mathrm{kHz}\).

4. Nonlinear distortion on an internal load of \(100\ \Omega\) does not exceed \(0.1\%\) at frequencies of 20 Hz and \(20\ \mathrm{kHz} \div 100\ \mathrm{kHz}\), and does not exceed \(0.05\%\) in the frequency range \(45\ \mathrm{Hz} \div 20\ \mathrm{kHz}\).

5. The inductive divider included with the setup makes it possible to set any ratio of two voltages from 1 to \(10^{-4}\), in steps of \(1 \cdot 10^{-4}\), over the frequency range \(20\ \mathrm{Hz} \div 100\ \mathrm{kHz}\).

6. The basic error of the division coefficient of the inductive divider at a frequency of \(1\ \mathrm{kHz}\) is not more than

\[ 2 \cdot 10^{-6} \pm \frac{5\ \mu\mathrm{V}}{U_{\mathrm{in}}}. \]

1. The value of the frequency error of the inductive divider at a frequency of \(100\ \mathrm{kHz}\) is not more than \(3 \cdot 10^{-4}\) with a load of 50 pF.

2. The input resistance of the inductive divider is \(1 \div 30\ \mathrm{k}\Omega\) (depending on the operating frequency); the output resistance is less than \(5\ \Omega\).

3. The setup provides:

a) measurement of the transfer coefficient of active and passive four-terminal networks with an error of not more than \(\pm 0.05\%\) over the entire frequency range;

b) measurement of the phase angles of the transfer coefficients of active and passive four-terminal networks with an error of not more than \(0.01^\circ\) in the frequency range \(20\ \mathrm{Hz} \div 20\ \mathrm{kHz}\), and not more than \(0.03^\circ\) in the frequency range \(20\ \mathrm{kHz} \div 100\ \mathrm{kHz}\), with a phase-checking resolution of \(0.001^\circ\);

c) measurement of the nonlinear-distortion coefficient with a resolution of \(0.002\%\).

References

1. Investigation of ways and methods for creating a universal calibration setup. Scientific and technical report. State registration No. 71003862, 1970.