PROCEEDINGS
OF THE TOMSK ORDER OF THE RED BANNER OF LABOR POLYTECHNIC
INSTITUTE named after S. M. KIROV

Volume 141 1966

MAGNETIC PULSE-WIDTH DC CONVERTER

Yu. M. AGEEV

(Presented by the scientific seminar of the Faculty of Automation and Computer Engineering)

A magnetic pulse-width DC converter, intended for converting direct or slowly varying currents into the duration of voltage pulses, belongs to the measuring devices used in digital measuring and information technology, telemechanics, and industrial automation. Therefore, the specified device must satisfy a whole set of requirements imposed on measuring devices, the chief of which are sensitivity, linearity of characteristics, stability, etc. Known magnetic pulse-width converters are practically impossible to use for measuring weak direct currents and voltages because of their low sensitivity, while electronic and semiconductor converters are very complex and expensive devices.

The magnetic pulse-width converter described below makes it possible to obtain at the output bipolar voltage pulses of constant amplitude, the duration of which is proportional to the magnitude of the input signal, while the phase depends on the polarity of this signal. This device has a small zero drift, sufficiently high sensitivity, linear characteristics, and an adjustable conversion coefficient. The converter has small dimensions and weight and consumes little energy. The use of such converters will make it possible, in multichannel measuring and other equipment, to dispense with DC commutators, the creation of which remains a problem to this day, with DC amplifiers and a number of other units, which will make newly designed systems simpler and more reliable. When the proposed converters are used, it will become fundamentally possible to increase the scanning rate of monitored points to several thousand and even several tens of thousands of points per second.

The use of a pulse-width converter in DC amplifier circuits makes it possible to create practically drift-free amplifiers; and in industrial automation devices the converter can be used as a contactless sensitive relay reacting to the polarity of the input signal (an analogue of a polarized relay), in temperature-control circuits, and in a number of other cases.

Operating Principle of the Pulse-Width Converter

The magnetic pulse-width DC converter (Fig. 1) is made on four identical magnetic cores, whose magnetization characteristic is shown in Fig. 2. On each core there are three windings:

\[ W_p \]

—the operating windings of the converter,

\[ W_{\mathrm{cm}} \]

—the bias windings,

\[ W_y \]

—the control windings.

The operating windings of the converter are connected in a bridge circuit. A rectangular supply voltage is applied to one diagonal of the bridge, and from the other diagonal of the bridge the output voltage is taken, which is fed to the load resistance \(Z_n\). The waveform of the converter supply voltage is shown in Fig. 3. For normal operation of the converter it is necessary that the initial position of the operating points for each core be quite definite. This task is performed in the converter by the bias circuit. Direct currents \(i_{\mathrm{cm}1}\) and \(i_{\mathrm{cm}2}\) flow through the bias windings. In the bias circuit there are two resistances \(R_1\) and \(R_2\). The resistance \(R_1\) serves to limit the bias currents, and \(R_2\) to balance them (by changing the position of the slider of resistance \(R_2\), equality of the bias currents can be established, i.e., \(i_{\mathrm{cm}1}=i_{\mathrm{cm}2}\)).

Fig. 1. Schematic diagram of a magnetic pulse-width converter.

The bias windings are connected in such a way that the initial position of the operating points in the cores on their magnetization characteristics ...

…of magnetization was as shown in Fig. 4. For this it is necessary that \(H_{\mathrm{cm}} = H_s\), where \(H_s\) is the saturation-field strength for the cores. Then the point \(O_{1,2}\) is the initial operating point for cores I and II, and the point \(O_{3,4}\) is the initial operating point for cores III and IV.

Fig. 2.

Fig. 3.

The control windings of the converter are connected in such a way that, in the control circuit, the alternating voltages induced in each control winding are compensated. If the material of the converter cores has an ideal magnetization characteristic, i.e., has no hysteresis loop, then on this characteristic one may distinguish sections with different magnetic permeability (Fig. 2). On the section \(NS\) of this curve the magnetic permeability of the material is large and equal to \(\mu_2\), while on the sections \(MN\) and \(SQ\) the magnetic permeability is small and equal to \(\mu_1\). It is expedient to introduce such an idealization for analyzing the operation of the converter.

Fig. 4. Magnetic state of the converter cores at \(H_{y1} = 0\).

In Fig. 5 and Fig. 6 a graphical explanation is given of the principle of operation of the converter. In the absence of a control signal \((U_y = 0)\), the initial points are the points \(O_{1,2}\) and \(O_{3,4}\). If the control signal is not equal to zero, then the current \(i_y\) flowing in this case creates in the cores

Fig. 5.

a control magnetic field \(H_y\), under the action of which the initial operating points for each core are displaced along the magnetization characteristics. In this case

point \(a_1\) is the initial operating point for core I,
point \(a_2\) is the initial operating point for core II,
point \(a_3\) is the initial operating point for core III, and
point \(a_4\) is the initial operating point for core IV.

Since a rectangular voltage is applied to the operating windings of the converter, the changes in the total magnetic inductions \(B_{1,2}(t)\) and \(B_{3,4}(t)\) in cores I, II, III, and IV will be linear (Fig. 6). The change of magnetic induction in cores I and II may be regarded as the change of one total magnetic induction \(B_{1,2}\), on the grounds that the operating windings of these cores are connected in series, and their magnetic characteristics and geometrical dimensions are identical. This statement is also valid for cores III and IV. During operation of the converter each core will be remagnetized according to the following cycles (Figs. 5 and 6):

core I according to the cycle \(a_1,\, b_1,\, o_{1,2},\, c_1,\, o_{1,2}, \ldots\)

core II according to the cycle \(a_2,\, b_2,\, o_{1,2},\, c_2,\, o_{1,2},\, b_2, \ldots\)

core III according to the cycle \(a_3,\, o_{3,4},\, b_3,\, o_{3,4},\, c_3,\, o_{3,4},\, b_3, \ldots\)

core IV according to the cycle \(a_4,\, o_{3,4},\, b_4,\, o_{3,4},\, c_4,\, o_{3,4},\, b_4, \ldots\)

In the process of remagnetizing the cores according to the indicated operating cycles, their magnetic permeability changes periodically, which leads to a periodic redistribution of voltages on the operating windings for each pair of cores I, II and III, IV.

Fig. 6.

Let us consider this process over the individual intervals of the operating cycles. In the interval $a_1 b_1$ the magnetic permeability of core I is equal to $\mu_2$, while in the interval $a_2 b_2$ the magnetic permeability of core II is equal to $\mu_1$. Since $\mu_2 \gg \mu_1$, the inductive resistance of the operating winding of core I in the interval $a_1 b_1$ will be much greater than the inductive resistance of the operating winding of core II in the corresponding interval $a_2 b_2$. Therefore the voltage on the operating winding of core I ($U_{\mathrm{I}}$) will be much greater than the voltage on the operating winding of core II ($U_{\mathrm{II}}$), Fig. 6. In the intervals $b_1 o_{1,2}$ and $b_2 o_{1,2}$, the magnetic permeability of core I is equal to $\mu_2$, and the magnetic permeability of core II is equal to $\mu_1$. Therefore in these intervals $U_{\mathrm{I}} \gg U_{\mathrm{II}}$.

Considering in this way the remagnetization processes over all sections of the working cycles, one may conclude that:

\[ \begin{aligned} &\text{on sections } b_1 o_{1,2}\,(o_{1,2}b_1) \text{ and } b_2 o_{1,2}\,(o_{1,2}b_2), && U_{\mathrm{I}} \gg U_{\mathrm{II}},\\ &\text{on sections } o_{1,2}c_1\,(c_1o_{1,2}) \text{ and } o_{1,2}c_2\,(c_2o_{1,2}), && U_{\mathrm{I}} \ll U_{\mathrm{II}},\\ &\text{on sections } o_{3,4}b_3\,(b_3o_{3,4}) \text{ and } o_{3,4}b_4\,(b_4o_{3,4}), && U_{\mathrm{III}} \gg U_{\mathrm{IV}},\\ &\text{on sections } o_{3,4}c_3\,(c_3o_{3,4}) \text{ and } o_{3,4}c_4\,(c_4o_{3,4}), && U_{\mathrm{III}} \ll U_{\mathrm{IV}}. \end{aligned} \]

Thus, during operation of the converter, the voltage applied to the working windings of cores I, II and III, IV will be redistributed depending on the change in the magnetic permeability of the cores. The instant of redistribution of voltage on the working windings of the cores will be called the instant of their switching. If the control signal is equal to zero, then the switching instants for cores I, II and III, IV coincide in time, and the voltage at the converter output is equal to zero.

If, however, the control signal is not equal to zero (Figs. 5, 6), then the switching instants of the above-mentioned cores do not coincide in time, and the output voltage \((U_{\mathrm{n}})\) is not equal to zero. In the latter case, the output voltage of the converter has the form of bipolar rectangular pulses, whose repetition frequency is equal to the frequency of the converter supply voltage (Fig. 6). The amplitude of the output-voltage pulses \((U_m)\) is constant, and the duration of the pulses \((\tau_{\mathrm{u}})\) is proportional

Fig. 7. Experimental characteristics of the converter at different values of bias current.

to the magnitude of the input signal, while the phase of the output voltage relative to the converter supply voltage changes by \(180^\circ\) when the polarity of the input signal is changed.

Figure 7 shows the experimental characteristics of a converter built according to the circuit of Fig. 1. The characteristics are given for two values of the bias current. A pulse-width converter with the following data was tested:

core material—K260 ferrite \((7 \times 4 \times 2)\) mm,
winding data—\(w_p = 50,\ w_y = 100,\ w_{\mathrm{cm}} = 20\),
supply voltage—amplitude \(0.2\) V, frequency—15 kHz.

Conclusions

1. A magnetic pulse-width converter built according to a bridge circuit can be used as the principal unit of an analog-to-digital converter for converting weak dc signals without preliminary amplification.

2. By changing the magnitude of the bias current, the conversion factor of the converter can be regulated smoothly and over sufficiently wide limits.

References

1. M. A. Rozenblat. Magnetic Amplifiers. Sovetskoe Radio, 1960.

2. F. I. Kerbnikov, M. A. Rozenblat. Sensitive magnetic modulator with two-phase supply. Avtomatika i Telemekhanika, vol. XXII, No. 3, 1961.

3. K. M. Polivanov, A. A. Braude. Hysteresis-free magnetization and its application to the measurement of small currents and emf. Proceedings of MEI, issue 18, 1956.