Gate-commutated inverters (DC/AC converters)

Home | Articles | Forum | Glossary | Books

AMAZON multi-meters discounts AMAZON oscilloscope discounts

Most modern AC VSDs in the 1-500 kW range are based on Gate-commutated devices such as the GTO, MOSFET, BJT, and IGBT, which can be turned ON and OFF by low power control circuits connected to their control gates.

Operating principle:

An inverter works on a DC supply giving a variable frequency AC output. It can be operated either as a step wave inverter or a PWM inverter.



In a step wave inverter, the transistors are switched such that the phase difference is 60º and each transistor is kept on for 180º. To vary the output AC waveform frequency, the duration between the turn on of transistors is changed. The output of AC voltage is varied by changing the DC input voltage. This type of an inverter has problems of a pulsating torque due to harmonics in the output voltage. This gives a pulsating motion of the rotor at low speeds.

The pulsating torque can be eliminated by the use of pulse width modulation (PWM) type inverters since their output has low harmonic content. The details of this type of inverter are explained later in this section. With a DC supply and semiconductor power electronic switches, it’s not possible to obtain a pure sinusoidal voltage at the load. On the other hand, it may be possible to generate a near-sinusoidal current. Consequently, the objective is that the current through the inductive circuit should approximate a sinusoidal current as closely as possible.

Single-phase square wave inverter

To establish the principles of gate-controlled inverter circuits; four semiconductor power switches feeding an inductive load from a single-phase supply.

++++21 Single-phase DC to AC inverter.

This circuit can be considered as an electronic reversing switch, which allows the input DC voltage VD to be connected to the inductive load in any one of the following ways:

1. S1 = on, S4 = on ... giving + VD at the load

2. S2 = on, S3 = on ... giving - VD at the load

3. S1 = on, S2 = on ... giving zero volts at the load

4. S3 = on, S4 = on ... giving zero volts at the load

5. S1 = on, S3 = on ... giving a short-circuit fault

6. S2 = on, S4 = on ... giving a short-circuit fault

However, these four switches can be controlled to give a square waveform across the inductive load. This makes use of the switch configurations (1) and (2), but not the switch configuration (3) or (4). Clearly, for a continued safe operation, option (4) should always be avoided. In the case of a purely inductive load, the current waveform is a triangular waveform. In the first part of the cycle, the current is negative although only switches S1 and S4 are on. Since most power electronic devices cannot conduct negatively, to avoid damage to the switches, this negative current would have to be diverted around them. Consequently, diodes are usually provided, anti-parallel with the switches to allow the current flow to continue. These diodes are sometimes called reactive or freewheeling diodes. These diodes conduct when the voltage and current polarities are opposite. This occurs when there is a reverse power flow back to the DC supply.

The frequency of the periodic square wave output is called the fundamental frequency.

Using Fourier analysis, any repetitive waveform can be resolved into a number of sinusoidal waveforms. Each comprises one sinusoid at a fundamental frequency and a number of sinusoidal harmonics at higher frequencies, which are multiples of the fundamental frequency. The harmonic spectrum for a single-phase square wave output. With an increase in frequency, the amplitude of the higher-order harmonics voltages fall off rapidly.

++++22 Square wave modulation waveforms.

++++23 Square-wave harmonic spectrum.

The RMS value of the fundamental sinusoidal voltage component is:

The RMS value of the nth harmonic voltage:

This illustrates that the square wave output voltage, has many unwanted components of reasonably large magnitude at frequencies close to the fundamental. The current flow in the load is due to an output voltage distortion, as demonstrated by the non-sinusoidal current wave-shape. In this example, the current has a triangular shape.

If the square-wave voltage were presented to a single-phase induction motor, the motor would run at the frequency of the square-wave. Being a linear device (inductive/resistive load), however, it would draw non-sinusoidal currents and would suffer additional heating due to the harmonic currents. These currents may also produce pulsating torques.

To change the speed of the motor, the fundamental frequency of the inverter output can be changed by adjusting the switching speed. To increase frequency, the switching speed can be increased, and to decrease frequency, the switching speed can be decreased.

The output voltage magnitude can also be controlled. The average inverter output voltage can be reduced by inserting periods of zero voltage, using a switch configuration (3). Each half cycle then consists of a square pulse, which is only a portion of a half period.

++++24 Square wave modulation with reduced voltage pulse width.

The process of changing the width of the pulse, to reduce the average RMS value of a waveform is called PWM. In the single-phase, PWM makes it possible to control the RMS value of the output voltage. The fundamental sinusoidal voltage component is continuously variable in the following range:

The harmonic spectrum of this modified waveform depends on the fraction, that the pulse is, of the full square wave, but is broadly similar to the waveform shown earlier.

Single-phase pulse width modulation (PWM) inverter

The fact that the voltage supply to the stator, of an AC induction motor, is a square wave and is not distorted in itself is a problem to the motor. The main problem comes from the distortion of the current waveform, which results in extra copper losses and is due to shaft torque pulsations. The ideal inverter output is one, which results in a current waveform of low harmonic distortion.

An AC induction motor is predominantly inductive, with a reactance that depends on the frequency (XL = j2pfL). It is, therefore, beneficial if the voltage harmonic distortion can be pushed into high frequencies, where the motor impedance is high and not much distorted current will flow.

One technique for achieving this is the sine-coded pulse width modulation (sine-PWM). This requires the power devices to be switched, at frequencies much greater than that of the fundamental frequency, producing a number of pulses, for each portion of the desired output period. The frequency of the pulses is called modulation frequency. The width of the pulses is varied throughout the cycle in a sinusoidal manner, giving a voltage waveform. The illustration also shows the current waveform for an inductive load, by showing the improvement in a waveform.

The improvement in the current waveform can be explained by the harmonic spectrum shown in ++++26. It can be seen that, although the voltage waveform still has many distortion components, they now occur at higher harmonic frequencies, where the high- load impedance of the motor is effective in reducing these currents.

++++25 Sine-coded PWM voltage and current

++++26 Harmonic spectrum for a PWM inverter

Increasing the modulation frequency will improve the current waveform, but at the expense of increased losses in the switching devices of the inverter. The choice of modulation frequency depends on the type of switching device and its frequency. With the force-commutated thyristor inverter, a modulation frequency of up to 1 kHz was possible with the older technologies. With the introduction of GTOs and BJTs, this could be pushed up to around 5 kHz.

With IGBTs, the modulation frequency could be as high as 20 kHz.

In practice, a maximum modulation frequency of up to 12 kHz is common with IGBT inverters up to about the 22 kW motor size and 8 kHz for motors up to about 500 kW. The choice of modulation frequency is a trade-off, between the losses in the motor and in the inverter. At low-modulation frequencies, the losses in the inverter are low and those in the motor are high. At high-modulation frequencies, the losses in the inverter increase, while those in the motor decrease.

One of the most common techniques for achieving the sine-coded PWM in practical inverters is the sine-triangle intersection method. A triangular saw tooth waveform is produced in the control circuit at the desired inverter-switching frequency. This is compared in a comparator, with a sinusoidal reference signal, which is equal in frequency and proportional in magnitude to that of the desired sinusoidal output voltage. The voltage VAN is switched high whenever the reference waveform is greater than the triangle waveform. The voltage VBN is not controlled by the same triangle waveform but with a reference waveform shifted by 180º.

++++27 Principle of triangle intersection PWM.

The actual phase-to-phase output voltage is then VAB, which is the difference between VAN and VBN, which consists of a series of pulses, each of whose width is related to the value of the reference sine wave at that time. The number of pulses in the output voltage VAB is double that in the inverter leg voltage VAN. For example, an inverter switching at 5 kHz should produce a switching distortion at 10 kHz in the output phase-to-phase voltage. The polarity of the voltage is alternatively positive and negative at the desired output frequency.

It can also be seen that the reference sine wave is given a DC component so that the pulse produced by this technique has a positive width. This puts a DC bias on the voltage of each leg. However, each leg has the same DC offset which disappears from the load voltage.

The technique of using a sine-triangle intersection is particularly suited to the old analog control circuits, where the two reference waveforms are fed into a comparator and the output of the comparator is used to trigger the inverter switches. Modern digital techniques operate based on a switching algorithm. For example, by producing triggering pulses proportional to the area under a part of the sine wave. Recently, manufacturers have developed a number of different algorithms that optimize the performance of the output waveforms for AC induction motors. These techniques result in PWM output waveforms. The sine-coded PWM voltage waveform is a composite of a high-frequency square wave at the pulse frequency (the switching carrier) and the sinusoidal variation of its width (the modulating waveform). It has been found that, for the lowest harmonic distortion, the modulating waveform should be synchronized with the carrier frequency, so it contains an integral number of carrier periods.

This requirement becomes less important with high carrier frequencies of more than twenty times the modulating frequency. The voltage and frequency of a sinusoidal PWM waveform are varied by changing the reference waveform giving outputs.

++++28 Variation of frequency and voltage with sinusoidal PWM

++++28(a) a base case, with the rated V/f ratio. ++++28(b) the case where the voltage reference is halved, resulting in the halving of each pulse. --- the case where the reference frequency is halved, resulting in the extension of the modulation over twice as many pulses.

The largest voltage with the sine-coded PWM occurs when the pulses in the middle are the widest, giving an output with a peak voltage equal to the supply.

Modulation index:

This is defined as the ratio of the peak AC output to the DC supply. Thus, the largest output voltage occurs when the modulation index is 1. It’s possible to achieve a high value of modulation index by abandoning the strict sine PWM and by adding some distortion to the sinusoidal reference voltage. This results in the removal of some of the pulses near the center of the positive and negative parts of the waveform. This is a process called pulse dropping. In the limit, a square voltage waveform can be achieved with a modulation index of 1.

Three-phase inverter

A three-phase inverter could be constructed from three inverters of the type shown earlier. However, it’s more economical to use a six-pulse (three-leg) bridge inverter.

In its simple form, a square output voltage waveform can be obtained by switching each leg high for one half-period and low for the next half-period, at the same time ensuring that each phase is shifted one third of a period (120º).

++++29 Three-phase inverter using gate-controlled switches.

++++30 Quasi-Square wave modulation output waveforms.

The resulting phase-to-phase voltage waveform comprises a series of square pulses whose widths are two-thirds of the period of the switch, in each phase. The resulting voltage waveform is called a quasi-square wave (QSW) voltage. This simple technique was used in early voltage source inverters (VSI), which used the forced commutated thyristors in the inverter bridge. To maintain a constant V/f ratio, the rectifier bridge controlled the magnitude of the DC bus voltage, in order to keep a fixed ratio to the output frequency, which was controlled by the inverter bridge. This technique was also known as the pulse amplitude modulation (PAM). The output voltage of a three-phase converter has a harmonic spectrum, very similar to the single-phase square wave, except that the triplen harmonics (harmonics whose frequency is a multiple of three times the fundamental frequency) have been eliminated.

In an inverter with a three-phase output, so that the 3rd, 9th, 15th, 21st, etc. harmonics are eliminated. To develop a three-phase variable voltage AC output of a particular frequency, the voltages VAN, VBN, VCN on the three output terminals a, b, and c can be modulated on and off to control both the voltage and the frequency.

The pulse-width ratio over the period can be changed according to a sine-coded PWM algorithm. When the phase-phase voltage VAB is formed, the present modulation strategy gives only positive pulses for a half period followed by negative pulses for a half period, a condition known as consistent pulse polarity. It can be shown that the consistent pulse polarity guarantees the lowest harmonic distortion, with most of the distortion being at twice the inverter chopping frequency. Therefore, these are the types of inverters used in industrial applications. The same methods are also used in AC drives.

++++31 Output voltage waveform of a three-phase sine-coded PWM.

Top of Page

PREV: Power electronic rectifiers (AC/DC converters) | NEXT: Overall protection and diagnostics Guide Index | HOME