ISSUES IN THE VERIFICATION OF GENERATOR-CALIBRATORS

R. F. Aknaev, Yu. M. Fomichev, A. A. Kush

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

The rms voltmeters currently being produced, of accuracy class 0.1–0.2, require verification at least once every three months. Their verification in a centralized manner requires large expenditures of time and resources.

Therefore, it is desirable to have laboratory-type equipment that meets metrological requirements and would make it possible to verify voltmeters at the places where they are used.

In recent years, a number of calibrators of sinusoidal rms voltage (GK-3, GK-4, GK-6, GK-7), satisfying the requirements imposed on verification equipment, have been developed and manufactured at the Tomsk Polytechnic Institute.

Verification of generator-calibrators was carried out at VNIIM.

Fig. 1. Block diagram for verification of generator-calibrators

Functional verification diagram

Verification of generator-calibrators was carried out according to the diagram shown in Fig. 1.

1 — GK-6 generator-calibrator;
2 — R33 resistance box;
3 — auxiliary dc voltage source (storage battery);
4 — reference thermoelectric converter Trb No. 13;
5 — 1:10 divider, class 0.001;
6 — R345 semiautomatic potentiometer, class 0.001.

Reference thermoelectric converter

The converter has a coaxial design (Fig. 2). As the measuring element, a TVB-4 vacuum contactless converter with an asymmetry error \(\delta = 0.005\%\), connected in series with an additional resistance, is used.

A blank MLT-1 resistor without notching or paint coating was used as the additional resistance. The value of the additional resistance was \(R_g = 500\ \Omega\).

The input is implemented through a plug with a characteristic impedance \(R_w = 50\ \Omega\), and the TEMF output is through a connector mounted on the fixed wall of the transducer.

Fig. 2. Design of the standard thermoelectric transducer

Fig. 3. Design of the photoelectric transducer

A transducer of this type was used in comparisons with generator-calibrators in the frequency range \((0.2 \div 30)\ \text{MHz}\). The error of the standard transducer when measuring AC voltage is:

\[ (0.2 \div 1)\ \text{MHz}\quad \delta \le \pm 0.05\%, \]

\[ (1 \div 10)\ \text{MHz}\quad \delta \le \pm 0.2\%, \]

\[ (10 \div 30)\ \text{MHz}\quad \delta \le \pm 0.5\%. \]

For verification of generator-calibrators in the range \(5\ \text{Hz} \div 100\ \text{kHz}\), a standard transducer was used in which, as the measuring unit, a TEM with an additional MLT-1 resistance, \(R_g = 500\ \Omega\), was employed; structurally, the transducer was made analogously to that described above. The error of such a transducer is:

\[ 5\ \text{Hz} \div 15\ \text{Hz}\quad \delta \le \pm 0.03\%, \]

\[ 20\ \text{Hz} \div 100\ \text{kHz}\quad \delta \le \pm 0.02\%. \]

To eliminate the influence of temperature, both transducers were placed in a passive thermostat.

Measuring Unit GK-6

The measuring unit is a photoelectric transducer (PT) made on the basis of an incandescent lamp and a photoresistor. Structurally, the PT is made in the form of a section of coaxial line. The input is implemented through a coaxial plug. The photoresistor is fixed to one of the walls of the transducer housing. A type S2-10 resistor was used as the additional resistance. The incandescent lamp is of the soffit design, type NSM 6.3×20.

Design data of the incandescent lamp:

lamp filament: coiled;

filament diameter — \(0.005\ \text{mm}\);

coil diameter — \(0.05\ \text{mm}\);

total number of turns — 124;

lead-ins — kovar, diameter \(0.25\ \text{mm}\).

The SF2-5 photoresistor is one arm of the comparator bridge. The sensitivity of the balance indicator is \(0.005\%/\text{division}\). The calculated frequency error of the photoelectric transducer at \(30\ \text{MHz}\) is \(\le 0.05\%\). The PT design is shown in Fig. 3.

Table 1

Fig. 4. Frequency-error graph of generator-calibrator GK-6 No. 1

Fig. 5. Frequency-error graph of generator-calibrator GK-6 No. 2

Fig. 6. Frequency-error graph of generator-calibrator GK-6 No. 4

Table 2

\(f = 30\ \text{kHz}\)

The instability of the output voltage is 0.013% over 16 min.

Table 3

\(f = 1\ \text{MHz}\)

The instability of the output voltage is 0.03% over 32 min.

Table 4

\(f = 30\ \text{MHz}\)

The instability of the output voltage is 0.2% over 27 min.

Fig. 7. General view of the setup for verification of generator-calibrators

Comparison Procedure

The error in setting the reference voltage of the GK-6 is included in the error in setting the variable output voltage. To exclude instrument error when measuring the reference voltage, the latter is measured by means of the same divider and potentiometer as the auxiliary dc voltage.

The variable voltage of the GK-6, set by means of the built-in measuring unit, is measured with the Prb No. 13 standard thermoelectric converter. The thermal emf of the Prb obtained in this way is recorded on the R345. By connecting the Prb to the auxiliary dc voltage source and setting the value of the thermal emf, we obtain a dc voltage which, in rms value, is equal to the variable voltage of the GK-6. The instrument error on dc, introduced as a result of the measurements, is equal to

\[ \sqrt{\delta_{\mathrm{d}}+\delta_{\mathrm{p}}+\delta_{\mathrm{ne}}+\delta_{\mathrm{prb}}}=\delta_{\Sigma}, \]

where
\[ \begin{aligned} \delta_{\mathrm{d}} &= 0.001\% &&\text{— error of the divider;}\\ \delta_{\mathrm{p}} &= 0.001\% &&\text{— error of the potentiometer;}\\ \delta_{\mathrm{ne}} &= 0.001\% &&\text{— error of the standard cell;}\\ \delta_{\mathrm{prb}} &= 0.005\% &&\text{— error of the converter.} \end{aligned} \]

\[ \delta_{\Sigma}\approx 0.0053\%. \]

Comparison Results

The comparison results obtained for generator-calibrators GK-6 Nos. 1, 2, and 4 are summarized in Table 1; the graphs are given in Figs. 4, 5, and 6.

Instability Study

The instability study was carried out according to the block diagram presented in Fig. 1. The standard converter is connected to the output of the GK-6, and the value of the thermal emf is recorded at certain time intervals. To eliminate error due to drift of the converter, the value of the thermal emf on dc is measured at the beginning of the measurement. At the end of the measurement, by setting the same voltage as at the beginning of the measurement, a new value of the thermal emf is obtained. The difference between these thermal emfs gives the drift of the converter itself during the measurement. This difference is excluded from the results obtained on ac.

The results of the instability study are summarized in Tables 2, 3, and 4.