CYCLOCONVERTER DRIVES (Motors And Drives)

We conclude this topic with a discussion of the cycloconverter variable-frequency drive, which has never become very widespread but is sometimes used in very large low-speed induction motor or synchronous motor drives. Cycloconverters are only capable of producing acceptable output waveforms at frequencies well below the mains frequency, but this, coupled with the fact that it is feasible to make large induction or synchronous motors with high-pole numbers (e.g. 20) means that a very low-speed direct (gearless) drive becomes practicable. A 20-pole motor, for example, will have a synchronous speed of only 30 rev/min at 5 Hz, making it suitable for mine winders, kilns, crushers, etc.
Most of the variable-frequency sources discussed in this topic have been described as inverters because they convert power from d.c. to a.c. The power usually comes from a fixed-frequency mains supply, which is first rectified to give an intermediate stage – the ‘d.c. link’ – which is then chopped up to form a variable-frequency output. In contrast, the cyclo-converter is a ‘direct’ converter, i.e. it does not have a d.c. stage (see also Section 2.4.6). Instead, the output voltage is synthesised by switching the load directly across whichever phase of the mains gives the best approximation to the desired load voltage at each instant of time. The principal advantage of the cycloconverter is that naturally commutated devices (thyristors) can be used instead of self-commutating devices, which means that the cost of each device is lower and higher powers can be achieved. In principle cycloconverters can have any combination of input and output phase numbers, but in practice the 3-phase input, 3-phase output version is used for drives, mainly in small numbers at the highest powers (e.g. 1 MW and above).
Text topics often include bewilderingly complex circuit diagrams of the cycloconverter, which are not much help to the user who is seeking to understand how such systems work. Our understanding can be eased by recognising that the power conversion circuit for each of the three output phases is the same, so we can consider the simpler question of how to obtain a variable frequency, variable voltage supply, suitable for one phase of an induction motor, from a 3-phase supply of fixed frequency and constant voltage. It will also assist us to bear in mind that cycloconverters are only used to synthesise output frequencies that
are low in comparison with the mains frequency. With a 50 Hz supply for example, we can expect to be able to achieve reasonably satisfactory approximations to a sinusoidal output voltage for frequencies from zero (d.c.) up to about 15 Hz; but at higher frequencies the harmonic distortion of the waveform will be so awful that even the normally tolerant induction motor will balk at the prospect.
To understand why we use any particular power electronic circuit configuration, we first need to address the question of what combination(s) of voltages and currents will be required in the load. Here the load is an induction motor, and we know that under sinusoidal supply conditions the power factor varies with load but never reaches unity. In other words, the stator current is never in phase with the stator voltage. So during the positive half-cycle of the voltage waveform the current will be positive for some of the time, but negative for the remainder; while during the negative voltage half-cycle the current will be negative for some of the time and positive for the rest of the time. This means that the supply to the motor has to be able to handle any combination of both positive and negative voltage and current.
We have already explored how to achieve a variable-voltage d.c. supply, which can handle both positive and negative currents, in topic 4. We saw that what was needed was two fully controlled 3-phase converters (as shown in Figure 2.11), connected back to back, as shown in Figure 4.8. We also saw in topic 4 that by varying the firing delay angle of the thyristors, the positive-current bridge could produce a range of mean (d.c.) output voltages from a positive maximum, through zero, to a negative maximum; and likewise the negative-current bridge could give a similar range of mean output voltages from negative maximum to positive maximum. Typical ‘d.c.’ output voltages over the range of firing angles from a = 0° (maximum d.c. voltage) to a = 90° (zero d.c. voltage) are shown in Figure 2.12.
In topic 4, the discussion focused on the mean or d.c. level of the output voltages, because we were concerned with the d.c. motor drive. But here we want to provide a low-frequency (preferably sinusoidal) output voltage for an induction motor, and the means for doing this should now be becoming clear. Once we have a double thyristor converter, and assuming for the moment that the load is resistive, we can generate a low-frequency sinusoidal output voltage simply by varying the firing angle of the positive-current bridge so that its output voltage increases from zero in a sinusoidal manner with respect to time. Then, when we have completed the positive half-cycle and arrived back at zero voltage, we bring the negative bridge into play and use it to generate the negative half-cycle, and so on.
In practice, as we have seen above, the load (induction motor) is not purely resistive, so matters are rather more complicated, because as we saw earlier for some part of the positive half-cycle of the output voltage wave the motor current will be negative. This negative current can only be supplied by the negative-current bridge (see topic 4) so as soon as the current reverses the negative bridge will have to be brought into action, initially to provide positive voltage (i.e. in the inverting mode) but later – when the current goes negative – providing negative voltage (i.e. rectifying). As long as the current remains continuous (see topic 2) the synthesised output voltage waveform will be typically as shown in Figure 8.9.
We see that the output voltage wave consists of chunks of the incoming mains voltage, and that it offers a reasonable approximation to the fundamental frequency sine wave shown by the dotted line in Figure 8.9. The output voltage waveform is no worse than the voltage waveform from a d.c. link inverter (see Figure 8.2), and as we saw in that context, the current waveform in the motor will be a good deal smoother than the

Figure 8.9 Typical output voltage waveform for one phase of six-pulse cycloconverter supplying an inductive (motor) load. (The output frequency shown in the figure is one third of the mains frequency, and the amplitude of the fundamental component of the output voltage (shown by the dotted line) is about 75% of the maximum that which could be obtained. The fundamental component of the load current is shown in order to define the modes of operation of the converters.)
voltage, because of the filtering action of the stator leakage inductance. The motor performance will therefore be acceptable, despite the extra losses that arise from the unwanted harmonic components. We should note that because each phase is supplied from a double converter, the motor can regenerate when required (e.g. to restrain an overhauling load, or to return kinetic energy to the supply when the frequency is lowered to reduce speed). This is one of the major advantages of the cycloconverter.
It is not necessary to go into the detail of how the firing angle scheme is implemented, but it should be clear that by varying the amplitude and frequency of,the reference signal to the firing angle control, we can expect the output voltage to vary in sympathy. We then have the ability to keep the voltage-frequency ratio constant, so that the flux in the induction motor remains constant and we obtain a constant torque characteristic. It should also be evident from Figure 8.9 that as the output frequency is raised it becomes increasingly difficult to achieve a reasonable approximation to a sine wave, because there are too few ‘samples’ available in each half-cycle of the output. Cycloconverters are therefore seldom operated at more than one third of the mains frequency.
With the coniguration described above, each phase of the motor requires its own double-bridge converter, consisting of 12 thyristors, so the complete cycloconverter requires 36 thyristors. To avoid short-circuits between the incoming mains lines, the three motor phase-windings must be isolated from each other (i.e. the motor cannot be connected in the conventional star or delta fashion, but must have both ends of each winding brought out), or each double converter can be supplied from separate transformer secondaries.
In practice there are several power-circuit configurations that can be used with star-connected motors, and which differ in detail from the setup described above, but all require the same number of thyristors to achieve the waveform shown in Figure 8.9. This waveform is referred to as 6-pulse (see topic 2), because the output has six pulses per cycle of the mains. A worse (3-pulse) waveform is obtained with 18 thyristors, while a much better (12-pulse) waveform can be obtained by using 72 thyristors.