In the current article, we will turn our attention to the basic concepts of Linear Power Supplies. All circuits require some source of power to operate and the most convenient source of such power is an AC wall outlet. Unfortunately, most of electronic circuits cannot make use of AC directly. Instead, some way to convert the AC to DC is required.
The first step for converting AC to DC is Rectification. Rectification is achieved using a Rectifier such as a simple junction diode. A diode only conducts when its anode is positive with respect to its cathode. That property is important when we are dealing with AC. If the diode were connected in a series circuit along with an AC supply and a load, its presence would mean that current could only flow through the load during the half of the AC cycle when the anode was positive with respect to the cathode. During the other half cycle the diode would not conduct and no current could flow. Such an arrangement is referred to as a half-wave rectifier because only half the waveform (i.e. alternate half-cycles) is allowed to pass freely. The other half of the waveform is cut off. The presence of those half-cycles of current causes pulsating DC to be generated across the load. The amount of voltage variation in that pulsating DC can be reduced by wiring a "filter" capacitor across the load. The amount of ripple in the output is then determined by the values of the capacitor and the load.
Fig.1 A half-wave rectifier
When dealing with electronic circuits, the power source should be as stable (i.e . free of ripple) as possible. The ideal power source then is a battery, as all DC voltages that are derived from an AC supply have some ripple. Using a battery is not always possible or practical, but fortunately most circuits can tolerate the presence of ripple if it is sufficiently attenuated. One way to minimize ripple is to use a full-wave rectifier. Such a circuit is shown in Fig. 2. Note that the circuit consists of a center tapped transformer, with the tap grounded, and two diodes.
Fig. 2 Full wave rectifier based on silicon diodes and a center tapped transformer
Let's see how this circuit works. We'll start by looking at what happens during the positive half cycle. During that half-cycle the polarity of the applied voltage is such that the upper terminal of the transformer's secondary is positive with respect to the center tap and the lower terminal. Also, during that half-cycle the polarity across D1 is such that the anode of the diode is positive with respect to its cathode and the device conducts. Thus, current flows from the upper transformer terminal, through D1 and RL, and back to the center tap through the ground. Note that the voltage during this half-cycle varies in phase from 0 to 180 degrees and that the current varies from zero, to some peak value, and then back to zero. Because of that varying current, the voltage developed across RL varies identically with the input waveform. Finally, during the positive half-cycle the cathode of D2 is more positive than its anode, so the diode does not conduct and no current flows through it.
The polarity of the voltage across the transformer is reversed during the negative half-cycle. Now, the bottom terminal of the transformer is positive with respect to ground and with respect to the top terminal. Diode D1 ceases to conduct because its cathode is more positive than its anode. But as for D2, its anode is now positive with respect to its cathode and the device conducts. Thus, current flows from the lower terminal of the transformer, through D2 and RL, and back to ground and the center tap, and a positive half-cycle of voltage is developed across RL. Note that here, once again, the voltage across RL varies identically with the input waveform, but the polarity of the voltage across the resistor is reversed (it is positive).
That sequence repeats during the succeeding positive and negative half-cycles. Note that current always flows through RL in the same direction so that only a positive voltage with respect to ground is across the load. That is true regardless of the instantaneous polarities of the AC voltage applied to the circuit.
The advantage of the full-wave rectifier over the half-wave rectifier lies in the fact that in the half-wave circuit no voltage is developed across the load during negative half-cycles. Because of that, the ripple in the output of the half-wave rectifier is higher.
Fig. 3 The full-wave diodes bridge
For ripple to be minimized in either type of circuit, some type of filtering must be used. To do so, a large capacitor is usually placed across RL. That capacitor is charged to the peak voltage, Vp, during the first half-cycle. Between peaks, it discharges slowly through RL. But it does not have enough time to discharge substantially before the next half-cycle appears and recharges it.
Fig. 4 A large capacitor is usually placed across RL for ripple filtering
Without the capacitor, the ripple voltage across RL varies from +Vp to 0 volts. But with the capacitor present, it varies from + Vp to whatever its voltage dropped to before the next half-cycle appeared to recharge it. From that, you should be able to see why the ripple is easier to filter in a full-wave rectifier. The reason is that the filter capacitor is recharged once during each half-cycle in a full-wave circuit, while in the half-wave arrangement it is recharged only once during each full cycle. Because of this longer recharge cycle, the voltage across the capacitor drops to a lower level. The ripple voltage, the voltage variation from +Vp to that discharge voltage level, is therefore larger for the half-wave than the full-wave circuit.
When the rectifying diode is not conducting, twice the peak supply or transformer secondary voltage may be across the device. This is true for the full-wave and the half-wave circuit with the exception of the full-wave bridge. So when designing a power supply circuit, be certain that the diodes have a sufficient voltage rating. The average current flowing through the diode is equal to the voltage across the load resistor divided by its resistance. The diode once again must be capable of accommodating that amount of current. Power dissipation capabilities of the diode are limited. Information as to just what these limits are is supplied by the manufacturer and can be found on data sheets. The power the diode must be able to dissipate is about the average current it passes in the forward direction multiplied by 1 volt. At times, it may be necessary to mount the diode on a heat sink so that its operating temperature will not exceed its specified limit.
If there is a filter capacitor, when the circuit is first turned on, the filter capacitor being charged by the DC behaves as a short circuit. Because of that, a large initial current surges through the diode. That surge current is equal to the supply-voltage peaks divided by all resistance in the circuit other than the resistances wired across the shorting capacitor. If the surge current is more than the diodes being used can accommodate, the device will be damaged. The best way to avoid damage is to use diodes that can safely handle that initial current surge. Alternately, you can connect a small resistor in series with each diode to limit current surges to safe levels. As for the transformer, it, too, can overheat if it conducts excessive quantities of current. Be sure to use a transformer which has sufficient current capabilities.
Throughout the discussion, it was assumed that the AC line voltage is fixed and that the load does not change in resistance but remains a constant RL. If anyone assumes that to be a realistic condition, then he is living in a dream world. Line voltage fluctuates from minute to minute. Over time it can vary ± 10% or more. During periods of extremely heavy usage, power companies have been known to greatly reduce voltage levels.
As for the load, it is not always a fixed resistor. If the supply is feeding an audio, RF, or electronic-switching circuit, the load impedance varies, sometimes from instant-to-instant, with the signal or switch current fed to it.
A fixed, stable voltage is frequently required when powering an electronic circuit. That constant voltage is not present when there are either supply-voltage or load variations. Most often, a filled voltage developed across a Zener diode can be used to stabilize the voltage across a load if the Zener is placed across that component or circuit. That is fine where low currents are involved. But when large quantities of current must flow through the load, the Zener diode can seldom be used economically as the sole regulating device for the circuit. Series, parallel, and feedback circuits using Zener diodes along with one or more transistors have been developed as practical regulators.
A Zener diode is a type of diode in which breakdown occurs at a well-defined reverse voltage. For instance, you can buy a Zener diode with a rated voltage of 6.8 V if you want to stabilize a supply voltage at this value. Figure 5.1 shows the corresponding basic circuit. The operating principle of this circuit can be seen from the characteristic curve of a typical Zener diode (Figure 5.2).
Fig. 5.1 Voltage stabilization with a Zener diode
First breakdown occurs when the reverse voltage rises above a certain value (Uz), leading to a sharp increase in the reverse current. The voltage across the diode remains stable at the breakdown voltage, as long as you don’t overdo it with the reverse current. Second breakdown is a frequently observed fault with Zener diodes. If the Zener diode becomes too hot, the junction shorts out, and after this the diode ‘stabilises’ the voltage at something close to zero volts.
Fig. 5.2 Characteristic curve of a Zener diode
Strictly speaking, the designation ‘Zener diode’ is not always correct, because two different phenomena are responsible for the breakdown effect with voltages over the range of 3 V to 200 V. The true Zener effect predominates at voltages below 5.6 V. It has a negative temperature coefficient, causing the Zener voltage to drop by up to 0.1% per degree. The avalanche effect, which predominates above 5.6 V, has a positive temperature coefficient. Zener diodes with a rated voltage of 5.1 V have the lowest temperature coefficient, while Zener diodes rated at 7.5 V or so have the steepest characteristic curves and therefore the lowest differential internal resistance. This means that they provide the best voltage stabilization with variable Zener current.
Although voltage stabilization with a Zener diode is easy, it has some drawbacks. One of the major drawbacks is power dissipation. This results from the fact that the series resistor must be dimensioned for the lowest input voltage and the highest output current. For example, if the circuit shown in Figure 5.1 has to supply a maximum current of 2 mA, the maximum output power is just 18 mW. The voltage over the series resistor is 3 V at the lowest input voltage of 12 V. This means that 1 mA flows through the Zener diode and 2 mA flows through the load. A current of less than 1 mA through the Zener diode is undesirable because it places the operating point on the knee of the characteristic curve, resulting in higher internal resistance and poorer voltage stabilization. However, even at this current level one-third of the input current is ‘wasted’ in the Zener diode. With even higher load requirements, the recommenced minimum Zener current is 5 mA. Things are even worse when the input voltage rises to 24 V. In this case the voltage drop over the series resistor is 15 V and the current is 15 mA. The resulting total input power is 360 mW. Compared with the useful power of 18 mW, this yields an efficiency of just 5%, which is terrible and is hardly tolerable in times of energy crisis. Fortunately, there is a solution to this problem.
Efficiency can be improved significantly if the Zener diode is followed by a transistor operating in common-collector mode, with the collector of the transistor connected directly to the positive terminal of the supply voltage (Figure 6a). This type of circuit is also called an emitter follower because the voltage on the emitter always follows the voltage on the base, with an offset of 0.7 V. Here the Zener circuit only has to supply the base current for the transistor. As a result, the input current is only slightly higher than the output current of the circuit over a wide range of operating conditions. Most of the power dissipation occurs in the series-pass transistor, and it depends only on the output current and the difference between the input voltage and the output voltage.
The circuits of Figure 6a and 6b are also known as series-regulators because DC current flows from the unregulated portion of the DC power supply through a transistor to the load. In both of those circuits, current flows through R1 and Zener diode D1 which causes a fixed voltage to be developed across D1. In Fig. 6a, current flowing through R1 also flows through the base-emitter junction of Q1. A fixed voltage, about 0.6 or 0.7 volt, is developed across this junction, turning on Q1. The voltage between the emitter of Q1 and ground, or across RL, is about 0.7 volt plus the voltage across D1. That fixed voltage is across RL regardless of supply voltage or load variations.
There are two types of parallel regulator circuits; one supplying a voltage that is only slightly lower than the breakdown voltage of the Zener diode used in the circuit, and one supplying a voltage that is considerably higher than that of the diode. Both are shown in Fig. 8.
Fig. 8 Parallel regulator circuits. The output from (a) is 0.7 volt above the Zener breakdown voltage; the output from b Is considerably higher.
In Fig. 8-a. current flows through R1, D1, and the base-emitter junction of Q1. Fixed voltages are developed across D1 and the base-emitter junction of Q1. The sum of those two voltages is the regulated voltage applied across RL.
In Fig. 8-b. current flows through R1, R2, the base-emitter junction of transistor Q1, and Zener diode D1. A fixed voltage is developed between the emitter and collector of Q1. The circuit's regulated output, VR which is across RL, is equal to the sum of the Zener voltage, Vz, and the voltage developed across Q1. It can be shown that that voltage is equal to VZ ( R2 + R3 ) / R3.
Resistor R4 is critical in and must be selected by trial and error. That resistor should be selected for the minimum variation of voltage across RL as the unregulated input voltage is varied from its minimum to its maximum.
Performance can be improved by using Darlington pairs rather than individual transistors and by replacing R1 with a constant-current source.
A commonly used series regulator-circuit using feedback is shown in Fig. 9. Current from R2 flows into both the collector of Q3 and the base of Q2. Because of D1, the emitter of Q3 is at a fixed voltage with respect to ground. Note that the regulated voltage is across RL as well as across R3 so that R3 can be used to adjust the voltage across RL.
Fig. 9 An example circuit of a using feedback regulator
When the voltage VR, across RL, increases above the desired level, the voltage at the base of Q3 rises. That transistor conducts more heavily than when VR is at its proper level. The base of Q3 is then more positive with respect to its emitter than it is when the level of VR is correct. That causes the transistor to draw more current than it does normally, reducing the available amount of current at the base of Q2. Because current through Q2, and consequently the current through Q1, are reduced, less current remains for RL.
In the opposite condition, when the voltage across R3 and RL is below the desired fixed level, less current flows through Q3. More current is now available to flow through Q2 and Q1, rebuilding the output voltage to its desired level.
Figure 10 shows a typical IC regulator and some of its surrounding circuitry; the part of the circuit enclosed by the dashed box is usually part of the IC.
A fixed voltage is developed across D1. A portion of that voltage, as set by R2, is applied as a reference voltage to the non-inverting input of the op-amp. The output from the op-amp is passed on to Q1. The voltage at the emitter of Q1, which is close to the voltage at the output of the op-amp, is fed back through RF to the inverting input terminal of the op amp. That inverting input is connected to ground through RIN. The voltage at the inverting input, and at the emitter of Ql, is equal to the voltage at the non-inverting input multiplied by 1 + (RF/RIN). The output voltage is therefore fixed by the voltage across D1, the setting of R2, and the ratio of resistors RF and RIN at the inverting input.
Fig. 10 A typical linear IC regulator circuit. The circuitry within the dashed box is usually contained in the IC
In Fig. 10 we've added a circuit to protect against damage in the event there is a demand for excessive current from the regulator. Excess current flow can not only damage a transistor, but can destroy an op-amp, and consequently an IC. Transistor Q2 is in the IC to protect it from being damaged. When excess current flows, sufficient voltage is developed acrossR3 to turn on Q2. When turned on, the base-collector circuit of Q2 is across the base-emitter circuit of Q1 , preventing it from conducting excess current.
It’s good that low-cost integrated voltage regulators are available for all common output voltages. A 7805 can deliver up to 1 A at 5 V, although a heat sink is necessary at such high current levels. In many situations the current is much lower, and in such cases the 78L05, with a maximum current rating of 100 mA, is sufficient. However, you should note that the 78L05 has a different pinout than its larger cousin. These voltage regulators require two capacitors - one at the input and the other at the output - to prevent oscillation at frequencies of several hundred kilohertz (Figure 11). These voltage regulator ICs contain everything already described in this instalment of our basics course using discrete semiconductor devices.
Fig. 11 7805 voltage regulator circuit
If you examine the internal circuit shown in Figure 12, you will see a lot of familiar things, such as the Zener diode with its series resistor. The actual control circuit is somewhat more complicated and includes a differential amplifier as well as a current mirror (see below). The series pass transistor is implemented as a Darlington pair consisting of Q11 and Q12, with most of the power dissipation occurring in Q12. Current limiting is handled by transistor Q10, which chokes off the base current to the Darlington pair Q11/Q12 if and when necessary. As might be expected from the 3 Ω value of the current sense resistor, the cutoff current is 200 mA. However, the IC is already very hot at this point, since the voltage on the base of Q10 is less than 0.6 V. The IC is protected against over-current and over-temperature. The over-temperature protection circuitry is built around Q7, Q8 and Q9.
Fig. 12 Internal structure of the 78Lxx
A current mirror, as illustrated by this circuit, is a distant cousin of a constant-current source. A current mirror is primarily used in ICs. The (constant) current through the 1 kΩ resistor is mirrored by the two transistors, and the collector current of the right-hand transistor is nearly the same as the current of the left-hand transistor. The base and collector of the left-hand transistor are connected together, which causes the base-emitter voltage to automatically assume a value that results in the specified collector current.
Fig. 13 A simple current mirror
In theory, if the second transistor has the same characteristics it should have the same collector current at the same base-emitter voltage. In practice, the current is usually slightly different because it’s difficult to obtain identical transistor characteristics. A current mirror is primarily used in ICs, where a large number of transistors on the same chip have the same characteristics. It’s also important that both transistors have the same temperature, since the transfer characteristics are temperature dependent. A current mirror of this sort can therefore be used as a temperature sensor.
Try touching one of the transistors with your finger. The resulting heating changes the output current, which can be seen from the change in the brightness of the LED. Depending on which of the two transistors you touch, you can make the LED a bit brighter or a bit darker. The temperature dependence of the current mirror is actually a drawback of this circuit. This sort of thing is often seen in electronics, where something that is an undesirable ‘degrading’ effect in one situation is a desirable ‘useful’ effect in another situation.