Talk Page



Switching Power Supplies

Many methods can be used to convert DC to AC. One of the most common methods is to use a device called an inverter. An inverter takes a DC voltage and power switches it using solid-state switching devices, thereby converting it into an AC waveform.

In this article, the fundamental theory of inverter operation is discussed. The basics covered here provide a better understanding of the operation of a typical single-phase inverter.


Functional Blocks

Figure 1 is a typical block diagram of an inverter. The function of each block is as follows.

  • Battery, Battery Input, Charger, and DC Filter The purpose of this block is to provide the DC power from a bank of batteries, connect the batteries together, provide a charger for replenishing the battery charge, and filter any unwanted AC from the battery charger output.
  • Power Switching This block is where the inverter function is performed. This block contains the power switching devices (SCRs) used to convert the DC into AC.
  • Controls This block is the control panel that allows the user to operate the inverter. This block also contains the circuits that generate the gating pulses. The gating pulses are used to switch on and off the power switching devices. Also, this block contains circuits used to synchronize the inverter output with an alternate power source.
  • Isolation, Output, and Harmonic Filter This block provides for transformer isolation of the output for safety and reliability. The transformer used is a constant voltage transformer (CVT) that provides for output voltage regulation. Also, since the inverter produces square waves and they contain a large number of odd-ordered harmonic frequencies, a harmonic filter is used to remove these undesired signals from the output.
  • Static Switch This block contains the high-speed automatic electronic switch that is used to transfer a critical load from the output of an overloaded or failed inverter to an alternate power source or standby inverter without interruption of power.

Figure 1: Typical Inverter Block Diagram
Virtually all inverter systems will contain these basic blocks. The circuitry and electronic components used will vary from manufacturer to manufacturer. A more detailed discussion of the purpose of each of these blocks will follow, starting with the heart of the inverter: the power switching circuit.

Power Switching Circuit

The power switching circuit converts the filtered DC input into a square-wave output. To be able to deliver substantial power to the loads, the switching devices used must be able to interrupt very large current flows. In the early days of inverter technology, mechanical switches called relays were used to chop the DC into a square-wave AC. These relays were bulky and required constant replacement of their contacts due to shock and pitting. Because of this, early inverters were unreliable and costly. Also, because of the switch bounce associated with mechanical switches, the output waveforms of these inverters were electrically noisy. The development of reliable solid-state high-power switching devices came shortly after the invention of the transistor. These static electronic switches eventually replaced old-fashioned mechanical relays as the power switching devices in inverters.

During the years since the invention of the transistor, a wide variety of solid-state switching devices has been produced. Thyristors are a family of switching devices whose members include the silicon controlled rectifier (SCR), the triac, and the diac. The operation of these extremely useful devices was discussed in another article.

Other more exotic power switching devices include gate turn-on devices (GTOs), metal oxide semiconductor field effect transistors (MOSFETs), and insulated gate bipolar transistors (IGBJTs). These devices are also very useful as switching devices. The theory behind their operation is beyond the scope of this article.

No matter what power-switching device is used, it must be able to take the DC input and re-shape it into a square-wave AC output. For the purpose of explanation, we will examine the operation of a simple power switching circuit using mechanical switches. This circuit is for the purpose of understanding only and would not be practical to use for a real world inverter. This circuit is shown in Figure 2 below.

Figure 2: Simulated Static Inverter Using Mechanical Switches
Switches may simulate the operation of a static inverter. Switches 1 and 1 are operated together. Switches 2 and 2 are also operated in unison. When 1 and 1 are closed and 2 and 2 are open, load current flows in the direction shown by the arrow above the load. With 2 and 2 closed and 1 and 1 open, load current flows in the reverse direction. The important thing to note is that while the current from the DC battery input always flows in the same direction, the current through the load reverses direction with each different switch combination. This means that the voltage polarity being supplied to the load by the load current also reverses its polarity. An inversion of the polarity has been performed by this circuit, hence the name inverter. The mechanical switches used in this circuit would not be used in an actual inverter. A power-switching device such as the SCR would be used instead in the place of each switch. This will cause the simple mechanical switching circuit to evolve into the circuit shown in Figure 3.

Figure 3: Bridge SCR Static Inverter
The circuit shown in Figure 3 contains the SCRs and various other components used to support the SCRs operation. These include the addition of the inductors above and below the SCR bridge and a capacitor connected across the bridge. The inductors and capacitor are commutating components. These components are necessary to turn one pair of SCRs on as the other pair is being switched off. These reactive components ensure that the switching is accomplished smoothly.

The SCR gate circuits are the triggers that actually turn the SCRs on at the proper times. The gate circuit applying a positive voltage to the gate terminal of the SCR accomplishes SCR turn-on. The gate circuitry is complex, but it is fundamentally an oscillator circuit. This circuit also includes the rectifier diodes, D1, D1', D2, and D2'. These diodes are not part of the inverter process, but they are used to clamp the amplitude of the load voltage to a value approximately equal to the magnitude of the source voltage. Figure 3 shows the diagram of a four-SCR parallel-commutated inverter bridge. Figure 4 is a two-SCR bridge. The operation of the two-SCR bridge is similar to that of the four-SCR bridge. It differs in that half the number of controlled rectifiers is used, and each must hold off a voltage approximately equal to twice the supply voltage. The center-tap of the transformer provides a return path for the direct current.

Figure 4: Center-Tap SCR Inverter
The circuits used to power inverters will now be discussed. These circuits are the battery and battery charger (rectifier).

Battery Charger/Rectifier

The inverter is powered via a bank of batteries. Various types of batteries can be used. The output voltage of the batteries depends on the output voltage that the inverter is designed to produce.

The bank of batteries must be capable of providing a filtered DC source for the power switching circuit. This ensures that the output of the power switching circuit is transient-free. Usually, an AC source, a transformer, a rectifier, additional filtering, and a logic diode are used to maintain the batteries in a fully charged state. If a power failure occurs that results in a loss of AC power, the batteries will continue to power the inverter for a time. This feature is what makes the inverter so useful in uninterruptible power supplies (UPS).

Since the AC power that is used to charge the batteries may vary in voltage at different times of day and under different load conditions, the battery charger must be capable of producing a constant output voltage. This voltage regulation is accomplished by using a specially designed rectifier whose output waveform is controllable. Figure 5 shows a basic block diagram of the battery charger/rectifier. The circuit is supplied by an AC input. The function of the blocks is described below.

Figure 5: Block Diagram of Battery Charger/Rectifier

  • Input Transformer The input transformer is designed to step up or down the incoming AC supply voltage to the proper value for the rectifier. The transformer also provides isolation between the supply and rectifier circuits for safety and ground reliability.
  • Rectifier This block is where AC is converted into a pulsating DC output. Typically, the rectifier is made up of three diodes and three SCRs (or 6 thyristors) connected in a full-wave array. The conduction interval of the SCRs is adjusted by firing at the proper time during each half-cycle. This change in the conduction interval allows the rectifier to maintain a constant output with varying input conditions. A longer conduction period increases the DC voltage output, and conversely, a shorter conduction period lowers the DC output voltage. This method of regulation is called phase angle control.
  • Output Filter This block contains the circuits that smooth the pulsating DC output from the rectifier. The filter is made up of a choke (inductor) and a capacitor bank. These are connected together in an "L" configuration. The capacitors oppose changes in voltage, and the inductor opposes changes in current. After filtering, the DC output voltage will be relatively smooth and ripple-free. The capacitor part of the filter may consist of the capacitors located in the inverter.
  • Control Board The control board supplies the pulses that gate the SCRs at the proper times. The exact point of pulse generation is a function of the output voltage of the rectifier. The control board compares the output of the rectifier to an internal reference and then generates an error signal, which in turn adjusts the firing angle of the SCRs. If the output of the rectifier begins to drop, a signal is generated that increases the conduction interval of the SCRs and thus returns the output voltage to its normal value. In addition, the control board has built-in overvoltage protection. If the DC output should go abnormally high, the gate pulses to the SCRs are immediately reduced. This feature protects the load against the high DC voltages that may occur if a drastic increase in AC supply voltage were to occur.

To better understand the operation of the rectifier portion of the charger, a simple half-wave rectifier using an SCR for phase control will be examined. See Figure 6.

Figure 6: Half-Wave Rectifier
Figure 7 shows a half-wave rectifier using an SCR for phase control. The SCR provides the rectification. The gate circuit controls the firing of the SCR. As an example, suppose that at steady state the conduction angle of the SCR is 90. The input and output waveforms are shown below.

Figure 7: Rectifier Input and 90 Output Waveforms
The output is a pulsating DC half-wave output with half of the waveform eliminated. Although the filtering of the output filter is not shown, the final output would be a constant DC voltage of a certain level. The level would represent the average DC value of the output waveform shown above.

Suppose for example, that the AC input voltage increased. The gate circuit would sense this and adjust the conduction angle so that it would be less than the 90 angle shown in Figure 7 above. The conduction angle may decrease to 45; this output waveform is shown below (see Figure 8).

Figure 8: Rectifier Input and 45 Output Waveforms
After filtering, the average DC output voltage of this waveform would be less than the waveform shown in Figure 7. This would compensate for the increase from the AC input.

If, for example, the AC input voltage decreased, the gate circuit would detect this and generate a new firing angle that would cause the SCR conduction angle to increase from the 90 steady-state value to a value of 135. This is shown in Figure 9 below. Now the output is receiving power for a larger percentage of each sine-wave input, so the average output voltage of this waveform after filtering would be higher than the steady-state value shown when the conduction angle was 90. This increase in average output voltage would compensate for the decrease in the AC input voltage.

Figure 9: Rectifier Input and 135 Output Waveforms
The half-wave single-phase rectifier circuit shown above has some serious disadvantages. First, since this is a half-wave rectifier, half of the AC input waveform is wasted since it is blocked by the SCR. The negative half-cycle is never used. This makes the half-wave rectifier very inefficient. Second, since there are large time gaps between the DC output peaks, filtering this output is difficult and would require large capacitors and inductors to produce an acceptable ripple-free output.

Using a full-wave rectifier instead of a half-wave rectifier increases the efficiency of the circuit. This allows the entire sine-wave input to be used. Also, full-wave rectification causes the peaks to be moved closer together in time, which again makes the output waveform easier to filter.

The control board controls the firing of the SCRs in the proper sequence. This is important so that the average DC output of the rectifier can be controlled.

Isolation and Harmonic Filter

The output of the power switching circuit is a square wave. This square wave must be wave shaped into a relatively pure sine-wave output. This means that most of the harmonic content of the square wave must be filtered out. A constant voltage transformer circuit is used to accomplish this. The constant voltage transformer circuit contains as its key component a sine-wave ferroresonant transformer (see Figure 10).

Figure 10: Sine-Wave Ferroresonant Transformer
Referring to Figure 10, inductor L1 and capacitor C1 form a series resonant circuit, which is tuned to the fundamental frequency (ordinarily 60 Hz). This connection will allow heavy current to flow out of the source and through L1 and C1, which causes the voltage across the capacitor to rise quickly.

At some value of volt-seconds across saturating transformer ST1, its magnetic core will become saturated, which destroys the resonant condition and limits the voltage rise across the capacitor. In effect, this saturating element is clamping the voltage across the capacitor to a specified value that is dependent on the number of turns on the winding and the cross-sectional area of the core of ST1. Since the voltage across ST1 may not be exactly the level desired at the output, a separate transformer, labeled T2, is used to adjust the level of the output voltage. In addition, a small transformer labeled T1 has been included. The purpose of this transformer is to compensate for a slightly rising output voltage that is directly proportional to the input voltage. This voltage is developed in the secondary of T1 and is added to the output voltage out of phase so that, in effect, it adds to the output negatively and tends to flatten the regulation curve.

The filter components, which are labeled CH1 and CH2, are used to improve the waveform so that the load is presented with an approximate sine wave. Choke CH1 is a linear device and is tuned together with C1 to the third harmonic of the input frequency. These components, therefore, present very low impedance to the third harmonic and reduce its amplitude in the output to a very low value.

Choke CH2 is allowed to saturate and, therefore, generates harmonic currents that are transformer-coupled out of phase into choke CH1. The addition of these out-of-phase harmonics in CH2 tends to cancel the higher order harmonics that appear across the primary winding of choke CH1. The result of using this type of arrangement is the generation of a sine wave with a total harmonic content of approximately 3%. A short circuit on the output will have a current level determined by the magnitude of input voltage divided by the reactance of L1. This component is responsible for the current-limiting characteristic of the ferroresonant transformer in the overload region.

Automatic Static Transfer Switch

The automatic static transfer switch is a high-speed electronic transfer switch used in uninterruptible power systems to automatically transfer a critical load from the output of a failed or overloaded inverter to a standby or alternate source of power without interruption. The alternate source can be another inverter or any other source of reliable power compatible with the output of the inverter.

The static switch has been designed to detect the only two failure conditions possible with ferroresonant inverters. These are:

  • Loss of square wave resulting from a power or control circuit failure
  • Loss of inverter AC output voltage resulting from ferroresonant transformer failure

Provisions have also been made to detect and initiate transfers of load from the inverter to an alternate source upon load faults and overload conditions. This mode is very useful for clearing load circuit branch fuses in the shortest possible time by relying on the superior fault clearing capabilities of the alternate source. Figure 11 shows a block diagram of the typical output switching arrangement.

Figure 11: Alternate Power Source Switching
The heart of the static switch is a reed relay that provides superior isolation from externally generated transients. The reed relay provides bistable switch operation, resulting in fail-safe performance. The reed relay primarily is used to steer the gate signals to the power SCRs located inside the switch. These SCRs act like solid-state circuit breakers. Logic power for the control circuitry of the switch is derived from a reduced level, rectified square wave generated in the inverter power switching bridge circuit. Failure of the inverter will interrupt the supply power to the static switch SCR gate control circuits, turning off the SCRs and causing a change in the state of the reed relay and resulting in a transfer of the load.

Failure of the inverter square wave is detected virtually instantaneously, causing the static switch power SCRs to switch before depletion of the stored energy within the output filter of the ferroresonant transformer. This results in a truly uninterrupted, make before break, transfer of power to the load.

Transfers that result from failure of the inverter output transformer and load fault or from overcurrent are initiated by simulating a failure of the inverter. This is done by removing power to the reed relay coil when failure or abnormal conditions have been detected. In the case of a voltage loss resulting from transformer failure, an interruption of cycle maximum will be experienced because no energy is available to maintain the load during the switching interval.

A two-position manual switch is often used to bypass the inverter and static transfer switch for maintenance purposes. It can also be used as a manual transfer switch when a static switch has not been included as part of the uninterruptible power system. The switch is available with "make before break" contacts when used with a system employing either a static transfer switch or an inverter line synchronizing accessory. In Figure 11, a single pole switch is shown for simplicity. In three-phase units, a 3PDT or 4PDT switch would be used. One pole of the switch may be used to activate an alarm.