POWER SUPPLIES

HALF WAVE RECTIFICATION

The simplest PSU consists of half wave rectification followed by a smoothing capacitor. 
Shown in the figure 1, below.

Since current can flow through a diode in one direction only, a diode can be used to change alternating current (AC) into direct current (DC). It does this by permitting current flow when the anode is positive with respect to the cathode and shutting of current flow when the anode is negative.
An alternating voltage is from the secondary of the transformer is applied to the diode in series with the load resistor.
In the above circuit current only flows when the anode is positive with respect to the cathode. Therefore only half cycles flow through the load resistor. 
Looking at the graphic display, figure 1b, the shape of the output wave is shown by the solid lines of the sine wave. The dotted lines show the portion of the AC cycle when the diode is not conducting.
With no capacitor filtering the output voltage, that is the voltage read by a DC voltmeter across the load resistor is 0,45 times the Erms value of the AC voltage delivered by the transformer. With a capacitor connected across the load resistor the voltage will be 1,4 times Erms.
Because the frequency of the pulses in the output wave is low, one pulse per cycle, considerable smoothing is required to provide an adequately smooth DC output. For this reason this circuit is usually limited to applications where the current to be drawn is small. Another disadvantage of this type of rectification is that the transformer must have a considerably higher primary volt-amp rating, approximately 40% greater that for other types of rectification.

FULL WAVE RECTIFICATION

The most universally used rectification system is the Full Wave Rectifier as shown in figure 2, below.


Because the circuit has two diode rectifiers it makes use of both halves of the AC cycle. A transformer with a centre-tapped secondary is required. The voltage on either side of the centre tap being equal to E_rms in figure 1 above.
When the anode of DA is positive, current flows through the load resistor to the centre tap of the transformer. At this instant current cannot flow through DB because its cathode is positive with respect to its anode. When the polarity of the AC cycle reverses, DB will conduct and current again flows through the load resistor back to the centre tap of the transformer.
Looking at the graphic display, figure 2b, the shape of the output wave is shown by the solid lines of the sine wave when DA and DB are conducting. The dotted lines show the portion of the AC cycle when DA and DB are not conducting. You will notice that when DA is conducting DB is not. The graph of the output wave shows that the output pulse is twice that of the half wave rectifier. Much less capacitive filtering would be required. Since the rectifiers work alternately, each one handles half the average load current. The current rating of each diode need only be half the total load current. 
With no capacitor filtering the average output voltage, that is the voltage read by a DC voltmeter across the load resistor is 0,9 times the E_rms value of the AC voltage across half the transformer. With a capacitor connected across the load resistor the voltage will be 1,4 times E_rms.


FULL WAVE BRIDGE RECTIFIER

Another type of Full Wave rectifier circuit is shown in figure 3, below.


In this arrangement, two diodes operate in series on each half of the cycle, one diode being in the lead to the load and the other being in the return lead.
Because the circuit is using a bridge rectifier, which has four diodes encapsulated in a single unit it also makes use of both halves of the AC cycle. In this case the transformer does not have to have a centre tapped secondary. The voltage across the winding is equal to E_rms in figure 3 above.
When the anode of DA is positive, current flows through the load resistor to the anode of DC, through DC back to the transformer winding. At this instant current cannot flow through DB and DD because their cathodes are positive with respect to their anodes. When the polarity of the AC cycle reverses, DB will conduct and current again flows through the load resistor to the anode of DD, through DD back to the transformer winding.
Looking at the graphic display, figure 3b, the shape of the output wave is shown by the solid lines of the sine wave when DA/DC and DB/DD are conducting. The dotted lines show the portion of the AC cycle when DA/DC and DB/DD are not conducting. You will notice that when DA/DC is conducting DB/DD are not. Also note that the graphic display is the same as the one for two-diode Full Wave rectification. The only advantage of using a full wave bridge, is that you do not have to have a centre tapped transformer. If your transformer is centre tapped, just leave that tap unconnected. I will show you later how you can use this tap to, your advantage. The graph of the output wave shows that the output pulse is twice that of the half wave rectifier. Much less capacitive filtering would be required. The current rating of the bridge needs to be the same as the total load current, plus some.
With no capacitor filtering the average output voltage, that is the voltage read by a DC voltmeter across the load resistor is 0,9 times the E_rms value of the AC voltage across half the transformer. With a capacitor connected across the load resistor the voltage will be 1,4 times E_rms. 

SMOOTHING or FILTERING

In the three power supplies above no smoothing or filtering was mentioned.
When a electrolytic capacitor is added across the load resistor the peaks of the positive part of the AC cycle is smoothed out as shown in figure 4 below.

The illustrations above show smoothing for a half wave rectifier PSU.
Figure 4a shows no filtering. Figure 4b shows filtering with a small value capacitor. Notice that the filtered output voltage is not a straight line, normal for DC when viewed on an oscilloscope. Figure 4c shows filtering with a higher value capacitor. Notice that the filtered output voltage is now more of a straight line. Figure 5 shows filtering of a full wave rectified signal.
When Full Wave rectification is used, remember that the positive peaks are closer together, so adequate filtering can be accomplished with a lower value of capacitance.

VOLTAGE REGULATION

ZENER DIODES

Zener diodes are the simplest of the voltage regulators. They are also used in "Crow Bar" over-voltage circuits. These will be dealt with later.
A Zener or Zener Diode is a special device used primarily for holding a voltage within a specified limit. It must be noted that as the reverse voltage across a zener increases, a certain point will be reached where the zener breaks down and current flows through the device. This voltage is know as the zener voltage V_z. Zener diodes are available having zener voltages in the range 2 to 200 volts. At the instant the zener breaks down, the current at that point is termed the knee current and is represented by I_zk. This is the minimum current necessary for the operation of the zener diode. Below this value the zener will stop conducting and regulation will no longer be carried out. As the supply voltage increases so does the current through the zener. This current represented by I_zk. This maximum current is limited to the dissipation of the zener. For example a ½ Watt device at 12 volts would have a maximum current of 84mA, whereas a 50-Watt device would have a maximum current of 4 Amps.

The minimum current is obtained from diode tables or by measuring it in a test circuit.
The maximum current can be calculated as follows: -
The zener is a 12 volt 1 watt device.

I = P / E : I = 1 / 12 therefor maximum current is 0,084 Amps = 84 milli-Amps.

Figure 6 below shows how a zener diode is used as voltage regulator.

Figure 6	Figure 7

Figure 7 on the previous page shows the zener curve. The voltage across the zener is 0. As the voltage is slowly increased, the current through the zener starts to rise, slowly at first until the zener voltage Vz is reached. At this point the current Izk, also known as the Knee current rises sharply. It is now that we require a resistor in series, to limit the maximum current, Izm, through the zener so as not to exceed the power dissipation of the zener diode.
The value of R is calculated using the following formulae: -

                                  

Zener Diodes are only useful as regulators when the current required is reasonably constant.
Let us consider a case of an amateur transceiver. On receive this radio draws on average 0,7 Amps on receive. When this radio transmits it draws 2,8 Amps.
For a Zener diode regulator to work correctly the zener must be capable of running at 3 Amps. This would mean that it would have to be a 50-Watt or higher device. 

Dropping resistor = (18 - 12) ÷ (1,1 x 2,8)
                               = 6 ÷ 3,08
                               = 1,95 ohms.

Wattage of the Zener = (((18 - 12) ÷ 1,95)) - 0,7) x 12
                                     = 28,5 Watts.

LM780XX THREE TERMINAL FIXED VOLTAGE REGULATORS

Three terminal voltage regulators have been developed for a variety of output voltages, currents and for both positive and negative outputs.
The circuit of such a device is shown in figure 8 below.

Figure 8	Figure 9

Figure 9 above shows how the output voltage of a fixed three terminal voltage regulator can be increased by adding a zener diode.

LM317 THREE TERMINAL VARIABLE VOLTAGE REGULATORS


In addition to the above fixed voltage regulators, there are available three terminal adjustable voltage regulators. These can be used as shown in figure 10 below.

Figure 10

In the circuit on the LHS a potentiometer has been added. The output voltage can be adjusted from 5 to 24 volts output with 30 Volts into the device. The circuit on the RHS has fixed resistors for a fixed output voltage. The value of R2 can be obtained from the table below.

V_OUT                  R2                                          V_OUT                  R2

  5                    750                                         12                   2K0

  6                    910                                         15                   2K7

  8                    1K2                                         18                   3K3

  9                    1K5                                         20                   3K8

 10                   1K8                                         24                   4K3

The output current of these three terminal devices ranges from a few milliamps to ± 5 Amps. Ideal for a small PSU for a work bench. However for transceivers with higher power requirements, different voltage regulators have to be used. These adjustable voltage regulators have thermal and over current built into them. 



LM723 THREE TERMINAL VARIABLE VOLTAGE REGULATORS


There are a number of different circuits and IC's, which can be used in the high power adjustable voltage regulators. In this instance we will be dealing with the 723 Integrated circuit voltage regulator.
The voltage controller IC 723 enables high highly constant and stabilised power supplies to be constructed.
The IC contains a buffered reference voltage source, a correction amplifier to regulate the output voltage, an output stage and current limiting circuitry.
The final output voltage and current limiting capabilities can be selected from a wide range with the aid of a minimum of external components.
Operating conditions are as follows: -

Input voltage, Pins 11 and 12 : 10 to 37 Volts. Do not exceed 37 Volts.
Output current from pin 10 : 200mA to 900mA, depending on the package.

As can be seen from the above table this device can supply ± 900mA. This is too low a current for most applications. We therefore have to connect it to a series of pass transistors, depending on the current requirements.

Figure 11

Referring to figure 11 above. By using a variable feedback path potentiometer VR1, a variable regulated output voltage can be generated. The voltage reference is connected to the non-inverting input of the error amplifier and the output voltage via the pot, to the inverting input. The error amplifier drives the output transistor and hence the feedback voltage from VR1 controls the output voltage. 

A 100pfd capacitor is used to stabilise the device. R1 is used as a current limit control. When the current through R1, which is the load current, exceeds 100mA a voltage of 650mV is developed across it. 
This is sufficient to turn on the current limiting transistor within the 723, which in turn switched off the output regulating transistor, causing the output to voltage to drop to 0V.
To increase the current supplying capabilities of this regulator it is used with pass transistors as shown in figure 12 on the next page. This circuit can supply 5 Amps of current. Where more current is required two 2N3055's can be used in parallel, driven by TR1. 
This is shown in figure 13 on page 12. When even more output current is required TR1 can drive a 2N3054, which can in turn drive four or more 2N3055's. The Load Sharing Resistors in the collector leads of the 2N3055's are there to share the load between the two output pass transistors. This is necessary because these 2N3055's are mass-produced and are not matched. One may have different characteristics than the other. The load sharing resistors ensure that each pass transistor carries an equal share of the load. Should more pass transistors be used, each one must have a load-sharing resistor in its emitter lead. These resistors must be equal in value. In figure 12, current limiting is achieved across the 0,13 resistor in the emitter lead of TR2.
Voltage across the 0,13? resistor with 5 Amps flowing through it is equal to
5 x 0.13 which = 650mVolts. 
As was stated on the previous page the 723 requires 650mVolts to shut down the output to the driver transistor TR1.

Figure 12


Figure 13

Some electrical specifications for different transistors that can be used as pass transistors.

In all cases the dissipation across each 2N3055 should not exceed 50 Watts. This gives a margin of safety and you should never have trouble with the 2N3055's.
The 2N3055 transistors can be replaced by a number of other transistors, including a TIP142, which is a high efficiency Darlington pair in one case. 

MJ15004, 2N3772 and 2N3773's can also be used.
A typical drive chain would be a 2N3053, driving a 2N3054, driving 2 or more 2N3055's.
The power dissipated across the pass transistors is there worst enemy. Let us consider the case of a power supply, which has to deliver 12 Volts at 10 Amps. The raw DC power supply before the regulator is 20 Volts. The pass transistor has to dissipate a drop of 8 volts at 10 Amps. 
This power equates to E x I = 8 x 10 = 80 Watts. In this case you would use two pass transistors, each would then dissipate 40 Watts.

The designers say that the raw DC supply should be capable of 1,4 times the current required at the output. In the above case the raw PSU should be capable of 14 Amps. The more capable the raw supply is of being able to supply the required current, the less the voltage difference needs to be and therefore the power dissipated by the pass transistors also drops.

CROW BAR PROTECTION CIRCUIT

A crowbar protection circuit is used to protect the equipment being supplied by the PSU from an over voltage situation. Crowbar circuits come in a variety of different configurations. Figure 14 below shows a variation of the crowbar circuit, which has been tried and tested. The 500? pot and R1 form a voltage divider where the voltage feeding the zener diode may be set to the desired value to cause a trip when the output voltage of the PSU exceeds a specific value. The table below the figure shows different values obtained with different zener diodes.
The circuit operates as follows. When the voltage set by the voltage divider reaches the zener voltage, the zener conducts and a voltage is developed across R2. This voltage triggers the SCR. The SCR, connected across the input to the voltage regulator stage conducts heavily, shorting the supply to earth. Fuse F1 blows and no further output is present at the output of the PSU.

Figure 14

SWITCH MODE POWER SUPPLIES

Switch mode PSU's come in various configurations. The principle is however still the same. Figure 16 on the next page shows a block diagram of a typical Switch Mode Power Supply. The mains being fed into the power supply is first fed through a high frequency filter. This is necessary to prevent high frequency noise from being fed back into the mains supply lines.

Figure 15

A bridge rectifier and smoothing capacitors convert the mains voltage to DC. After smoothing the DC is fed to a pair of switching transistors which are driven by a transistor transformer combination from the Pulse Width Modulator. A schematic of the Pulse Width Modulator, TL494 is shown in figure 15 above.
The switching frequency is in the region of 33 kilohertz or higher. Remember that a transformer cannot operate with DC being fed into it. The DC voltage is therefor switched or pulsed into the Output transformer as an alternating voltage. This transformer then gives out the required output voltage depending on the turns ratio of the transformer. Because the switching frequency is high conventional iron core transformers cannot be used. The cores of these transformers are a magnetic ferrite material made up of combinations of ferric oxide and one or more oxides of bivalent metals. Manganese zinc ferrites are mostly use up to 1MHz. These transformer cores can supply more output current than the conventional iron core transformers. The higher the frequency a transformer operates at, the smaller it's physical size will be for a given amount of power output.

The output of the transformer is rectified and smoothed before being fed out. 

A sample of this voltage is fed back to the Pulse width modulator IC for control purposes. This output voltage is compared with a reference voltage and, if a difference exists, an error signal is generated and fed to the control circuitry. The control circuitry adjusts the Mark-Space ratio of the switching pulses to the switching transistors. Another method would be to vary the frequency of the switching.

It will be noted that two high frequency transformers are used (T1 and T2, figure 17 on page 17), one in the main output path and one in the feedback loop.

This ensures that the mains is well insulated from the equipment being fed by the PSU.
Also incorporated in this PSU is protection circuitry, which protects the equipment, being fed from a high output voltage. Other protection circuitry protects the PSU unit from excessive output current. When either of these protection circuits operate they shut down the pulse width modulator, so that little or no voltage is outputted.

Figure 16

Because the switching transistors are either ON or OFF, they dissipate nearly no power. The small size of the high frequency transformers makes it possible to build a high current power supply into a very small space.

The controller is typically a TL494 Pulse Width Modulator, which operates off the output voltage of the supply. This means that in order to start, there must already be an output voltage present! How they do this is really clever, and also extremely confusing. The output power transformer is itself, self-oscillating. This generates a rudimentary output voltage on the +12 volt output to allow the PWM to start. As soon as the auxiliary voltage UAUX is present at the input to the PWM the pulse width modulator TL494 swamps out the self oscillation with its own oscillations and normal operation commences. 

The error amplifier in the TL494 compares the voltage at the +5 volt output with a reference voltage, which is generated inside the TL494. It then calculates the analogue control variable and adjusts the pulse width modulator. The modulator sends alternate pulses to the driver transistors Q5 and Q6, figure 17. 

The pulse duration is proportional to the control variable rating. Increased loading on the + 5 volt output makes the pulses wider. Lighter loading causes narrower pulses. There is a finite minimum pulse width so a minimum load of about 200mAmps is required. Without this load the power supply may destroy itself.

Figure 17 on the next page shows the actual circuitry around the mains rectification and the switching of the DC voltage onto the output transformer T1. 

The switching pulses are fed in at points E and F from the pulse width modulator circuitry shown in figure 18 on the next page.

Figure 17


Figure 18

Figure 18 also shows the protection used by the PSU used in this article.
In these examples several protection circuits are included. Excessive primary current due to a very high secondary current leads to the 4,3 volts out of T3 increasing above a set threshold. Via one of the Op amps in IC3 and a transistor the pulse width modulator is shut down via its pin 4. 
The switching circuit and load are also protected against over-voltage at the +5 volt output and short circuits on the -12 volt and - 5 volt outputs. There are other forms of protection used instead of IC3. Some power supplies use discrete transistor and SCR monitoring circuits, offering the same protection.

Figure 19 below shows the original output circuitry. Full wave rectification is used for all outputs. The + 12 volt and + 5 volt outputs have higher current diodes to cope with the higher currents supplied by these outputs. The - 12 volt and - 5 volt outputs come from the same windings. 

Lower current diodes are used in these outputs. In figure 18 the - 5 volt output is obtained from a negative three terminal regulator fed by the - 12 volt supply line. L4a, b and c are windings on the same toroidal core. These are part of the smoothing circuits. 

In the PC PSU environment all three main outputs are wound on the same core, L4 in our case. This improves cross regulation between the windings.

Please remember that the diodes used for rectifying on the low voltage side must be Ultra fast recovery diodes or Schottky Barrier Diodes for high frequency operation.

Figure 19

Because these PC PSU's differ from model to model no specific reconstruction steps will be given, except to say that all components on the secondary side of the output transformer must be removed. You however need to identify the protection circuitry for reconstruction and modification. Strip all the windings from L4. Keep a note of the number of turns on the + 12 volt winding. L4 will need to be re-wound with thicker wire using the same number of turns. 

Figure 20 on the next page shows a modified output circuit. D5a&b, the original rectifier for the + 5 volt output has been retained because it has a high current carrying capability. It has to be put onto a larger heatsink. Extra filtering in the form of 2 x 2000Mfd, 2 x 0,47Mfd and a 100uH choke have been added for extra smoothing after rectification.

The 100Ω resistor provides a permanent load for the PSU. 

The UAUX voltage is taken from its own rectifier and smoothing circuit. 

The UA supply to IC 1 is taken from the + 12 volt output line via a suitable resistor R24. In this circuit the voltage at pin 1 of IC1 must be equal to 2,5 volts after loop stabilisation. That is half the 5 volt reference voltage when the output of the PSU is 13,8 volts.

In my prototype circuit I have put a 5KΩ potentiometer in place of R24 (figure 18). This gives me the option of varying the output voltage.

Figure 20

The END