MEASURING DEVICE FOR EXPERIMENTAL STUDY OF PULSE-WIDTH MODULATORS

Yu. M. Ageev, A. I. Gering, V. N. Kovalev

(Submitted by the scientific seminar of the Department of Automation and Telemechanics)

The main purpose of the experimental study of certain types of pulse-width (PW) modulators, in particular the magnetic-semiconductor PW modulator [1], is as follows: 1) determination of the properties of the modulator in the steady-state regime (conversion error, stability, etc.); 2) determination of the dynamic characteristics of the modulator.

The magnetic-semiconductor pulse-width modulator [1] is a highly stable device with a small static conversion error; therefore, for its study, equipment of a high accuracy class is required. It is expedient to represent the dynamic properties of PW modulators in the form of transient characteristics under a step input. It is therefore necessary that the measuring equipment make it possible to determine the transient characteristics of the modulators; moreover, recording of these characteristics is highly desirable.

In determining transient characteristics, particularly high measurement accuracy is usually not required; however, since high-accuracy equipment is necessary for taking static characteristics, it proves expedient to create a single measuring setup using digital measuring instruments for a comprehensive experimental study of pulse-width modulators. The use of digital instruments makes it quite simple to record the measurement results, and also to make the device semiautomatic, which will speed up the experiment and improve its quality.

In studying magnetic-semiconductor PW modulators, it is necessary to measure direct current or voltage and the duration of pulses. To measure direct current, either pointer instruments of the magnetoelectric system or digital instruments having an accuracy class not lower than 0.05 may be used. Measurement of pulse duration with high accuracy is best carried out by the method of filling the measured pulse with other pulses of stable frequency, which are then fed to an electronic pulse counter. The measurement error will be determined mainly by the discreteness error and will be the smaller, the higher the frequency of the filling pulses.

Method for Measuring Transient Characteristics and Functional Diagram of the Device

When static characteristics of a PWM are taken, no special difficulties arise, since the quantities being measured are steady-state ones. When transient characteristics are taken, the main difficulty consists in measuring the pulse durations. Since the transient-process time in some types of PWM is considerably greater than the pulse duration, the method of direct observation—and still more of measurement, for example by means of an oscilloscope—is practically inapplicable. In this connection, for taking the transient characteristics of a PWM, a method has been proposed whose essence is as follows.

A step action is applied to the input of the PWM. At the modulator output, the pulse duration begins to change. First the first pulse is measured, after which the circuit is returned to its initial state. Then the same step action is again applied to the PWM input, but now the second pulse is measured, and the circuit is again returned to its initial state. The step action is applied again and the next pulse is measured, and so on, until the transient process is completed.

At each measurement the pulse duration and its ordinal number from the moment the transient process begins are recorded; therefore, knowing the frequency of the PWM supply voltage, it is easy to determine the duration of the transient process.

The complete functional diagram of the device is given in Fig. 1. It contains the following main functional elements and units.

I. A generator unit intended for generating the necessary electrical signals.

At its output there are:

1. An AC voltage of rectangular waveform at various frequencies for supplying the PWM under investigation.

This function is performed by:

a) a quartz generator producing a pulse voltage of stable frequency 100 kc/s;

b) a frequency divider, which divides the frequency of the pulses arriving at its input and has division factors: 2, 10, 50, 250, 500, 1000, 2000;

c) a rectangular-voltage generator producing an AC rectangular waveform of the following frequencies: 50, 100, 200, and 400 c/s; 2, 10, and 50 kc/s.

1. A voltage in the form of unipolar rectangular pulses of stable frequency 500 kc/s, obtained with the aid of a frequency multiplier.

2. A DC voltage (or current), which is the input signal for the PWM under investigation and which is applied to the input of the latter through Kl1.

II. A measuring unit, consisting of:

1) a measuring circuit for measuring pulse duration (a circuit for selecting a single pulse, a filling circuit, and counter 1);

2) an instrument for measuring the input signal of the PWM;

3) a circuit for measuring the duration of the transient process.

III. A recording unit, which delivers the measurement results in digital form on paper tape and on a light display board.

IV. Start and reset circuits (the latter is not shown in Fig. 1).

Let us consider the operation of the measuring device when measuring static characteristics. In this case, a voltage or current from the input-signal shaper is continuously applied to the input of the PWM under investigation.

Fig. 1. Functional diagram of the measuring device.

Text visible in the diagram:

• Generator unit
• Input-signal shaper
• Sw1
• Frequency-amplitude
• Modulating-voltage generator
• Quartz generator, 100 kHz
• Frequency multiplier
• \(f = 500\) kHz
• Single-pulse selector circuit
• Filling circuit

• Instrument

• Single-pulse selector circuit

• PWM
• \(\Phi\)
• Sw2
• Sw3
• Sw4
• Sw9
• “M”
• Conversion circuit
• Reset
• Reset
• Tr1
• Tr2
• Tr3
• Switch
• Awaiting multivibrator
• Counter I
• \(\times 1\)
• \(\times 10\)
• \(\times 100\)
• \(\times 1000\)
• Reset
• Counter II
• \(\times 1\)
• \(\times 10\)
• \(\times 100\)
• B-I-1
• 1
• 10
• 100
• B-I-2
• 10
• 100
• B-I-3
• 1
• 10
• 100
• Registration unit
• PWM start
• End of operation
• Reset
• Filling
• Counter I
• Instrument-control unit
• Tr1
• Tr2
• \(U_1\)
• \(U_2\)

(\(K_l 1\) is permanently closed). The magnitude of this voltage (current) is measured by the measuring instrument. From the PWM output, a series of pulses, whose duration is constant and corresponds to the magnitude of the input signal, is fed to the single-pulse selection circuit. In this case, a signal is applied to the control input of this circuit directly from the start circuit, after which the very first pulse from the PWM passes to the filling circuit. In this circuit the measured pulse is filled with additional pulses of a stable frequency of 500 kHz, arriving from the frequency multiplier. Counter 1 measures the number of pulses arriving at its input and outputs information in a binary-decimal code of the type 1—2—4—2.

Let us consider the process of measuring transient characteristics. The device makes it possible to measure the transient characteristics of a PWM under a step input. In this case, depending on the duration of the transient process, it is possible to measure every pulse, every tenth pulse, or every hundredth pulse. Let us consider the case in which every pulse is measured. This case corresponds to the position of switch \(B\)-1 in Fig. 1. Before the measurement begins, the reset device brings the entire circuit to its initial state. The start pulse from the start circuit acts on \(Tr\ 1 \div Tr\ 3\) in such a way that \(K_l 1\) and \(K_l 2\) close, while \(K_l 3\) opens. Through the closed \(K_l 1\), the input voltage (current) is applied to the PWM input, and through \(K_l 2\) a pulse from the shaper \(\Phi\), coinciding in time with one of the edges of the PWM supply voltage, passes. This pulse is applied to the counting input of counter \(II\) and, through the closed \(K_l 4\), starts the single-pulse selection circuit. Thus the first pulse from the PWM will be measured, and a one will be recorded in counter \(II\).

The readings of counters \(I\) and \(II\) must be registered. This is done as follows. From the single-pulse selection circuit, a signal corresponding to the trailing edge of the measured pulse is applied to \(Tr\ 1\) and \(Tr\ 2\), returning them to their initial state. \(K_l 1\) and \(K_l 2\) open, and from the output of \(Tr\ 2\) the signal is applied to \(Tr\ 3\), closing \(K_l 3\) for resetting the recalculation circuit, and starts the digital printing machine. Upon completion of printing, the “end of printing” signal resets the readings of counters \(I\) and \(II\), while at the same time carrying out parallel transfer of information from the first decade of counter \(II\), through the closed \(K_l 3\), into the recalculation circuit.

When any number except zero is recorded in the recalculation circuit, \(K_l 4\) opens. After the one has been transferred from counter \(II\), the process of measuring the first pulse is completed. To measure the second pulse it is necessary to start the circuit again. The pulse from the start circuit again closes \(K_l 1\) and \(K_l 2\) and opens \(K_l 3\). However, now the first pulse from the shaper \(\Phi\) will not start the single-pulse selection circuit, since \(K_l 4\) is open. The first pulse from the shaper will be applied to the counting input of counter \(II\), will set \(Tr\ 4\), and, after a delay time equal to 10 μsec, will return \(Tr\ 4\) to its initial state, thereby recording minus one in the recalculation circuit.

In this device the recalculation circuit performs the function of a subtracting device. After minus one has been recorded in the recalculation circuit, zero will remain in it, and \(K_l 4\) will close, thereby preparing the start of the single-pulse selection circuit by the second pulse from the shaper \(\Phi\). The second pulse from the PWM will now be measured, the magnitude and serial number of which will also be registered. At the end of the cycle a two will already be recorded in the recalculation circuit, and the circuit is prepared for measuring the third pulse. The entire process is repeated until 10 pulses have been measured. If the transient process in the PWM lasts more than 10 periods of the supply voltage, then it is necessary to switch

Fig. 2

Visible labels in the diagram:

• Counters I
• Circuit
• Input
• Counters II
• Circuit
• Input
• Scaling circuit
• Reset
• Reset
• To the AND circuits
• To the counting input of the next decade
• Decade circuit
• PhЭ 3 [[unclear: abbreviation]]
• KP U-1 [[unclear: abbreviation]]
• KP U-2 [[unclear: abbreviation]]
• KP U-3 [[unclear: abbreviation]]
• KP U-4 [[unclear: abbreviation]]
• KP U-5 [[unclear: abbreviation]]
• KP U-6 [[unclear: abbreviation]]
• KP U-7 [[unclear: abbreviation]]

switch \(B\)-1 is set to position “10.” In this case the first pulse, the eleventh, the twenty-first, etc., will be measured. If switch \(B\)-1 is set to position “100,” then the first pulse, the one hundred first, the two hundred first, etc., will be measured.

Let us consider in more detail the operation of the recalculating circuit when carrying out carry and subtraction operations. Figure 2 shows structural diagrams of one decade of counter \(II\), of the recalculating circuit, and of \(K\ell\) 3. Before the carry operation, the recalculating circuit is reset to the zero state by a pulse on the “reset” bus, while \(Tr\) 3 is in such a state that \(K\ell\) 3-1–\(K\ell\) 3-4 are closed, and \(K\ell\) 3-5–\(K\ell\) 3-7 are open.

After registration of the measurement result is completed, the “end of printing” pulse arrives along the “reset” bus at counter \(II\), resetting this counter to the zero state. The information recorded in each trigger of this counter is applied to the inputs of the corresponding triggers of the recalculating circuit, thereby recording the number that had previously been in one decade of counter \(II\). The starting pulse switches \(Tr\) 3 into the opposite state, closing \(K\ell\) 3-5–\(K\ell\) 3-7 and opening \(K\ell\) 3-1–\(K\ell\) 3-4.

Now counter \(II\) operates as an adder, and the recalculating circuit as a subtracting device. Both circuits perform recalculation by 10 and use a code of the form 1–2–4–2.

Table 1 shows the states of the triggers of the recalculating circuit in the subtraction mode. Before pulses began arriving at its input, the number 9 had been entered into it.

Table 1

There is no need to examine the circuit elements in greater detail, since they may be chosen as standard types. We shall only point out that, when using as the registering device a digital printing machine of the TsPM-1 type, all triggers of counters \(I\) and \(II\), with the exception of the first decade of counter \(I\), should expediently be implemented with transistors of types P 25–P 26, choosing the supply-source voltage to be on the order of \((35 \div 45)\) V.

As the measuring instrument it is expedient to use a digital-printing voltmeter of the ETsPV-3 type, whose registering device can be used to record all measurement results.

LITERATURE

1. Yu. M. Ageev. Magnetic pulse-width converter of direct current, collection Proceedings of the Tomsk Polytechnic Institute named after S. M. Kirov, vol. 141. Publishing House of TSU, 1966.
2. G. Ya. Mirskii. Measurements of Time Intervals, Energiya Publishing House, 1964.