Signal generators: Design and construction: DC power supplies (part 2)

Home | Articles | Forum | Glossary | Books

cont from part 1

-Designing a power-supply circuit-

So far, we have been looking at power-supply circuits in a very general way. In practical electronics work, you will obviously need to convert this theoretical information into a practical circuit with specific component values to suit the needs of your intended application. Designing a simple power-supply circuit is seldom difficult, Only a handful of components are required, and usually there will be consider able leeway in their values. You usually won’t need to work out any calculations too precisely.

In most simple power-supply circuits, the transformer is the most critical (and most expensive) component. There are four basic specifications for a power transformer:

  • • Primary voltage
  • • Secondary voltage
  • • Power rating
  • • Regulation factor

The primary voltage is simply the AC voltage the transformer expects to see at its primary winding, or its input. In the U.S.A., most commonly available transformers are designed for use on ordinary AC house current, which has a nominal value of about 120 Vac. Sometimes the primary transformer is marked 110 Vac or 117 Vac. Different manufacturers use slightly different standards. The actual voltage of AC house current tends to fluctuate, so these are all just nominal values, and they are all considered to be identical.

You might encounter a transformer with a primary voltage rating of 220 Vac or 240 Vac. Can you use such a transformer with ordinary house-current? Sure, but you will get a different secondary voltage than what is marked on the transformer’s housing. The primary voltage is basically a reference that determines the secondary voltage. Reducing the primary voltage will result in a lower secondary voltage, and vice versa.

Occasionally you might come across a power transformer with a low primary voltage rating, such as 32 Vac. You might be able to use such a transformer on ordinary house current, but I wouldn’t count on it. The excessive input voltage might cause the transformer to over-heat and burn itself out. If left unattended, a fire could result.

The secondary voltage of a power transformer is generally considered the most important specification. It’s given in Vac rms, for the stated primary voltage as the input. The secondary voltage rating always assumes the full rated power load is being used—that is, the maximum rated current drain is being used by the load circuit. With a smaller load, the actual secondary voltage will be higher than its rated value.

In selecting a power transformer for a specific application, you must use a some what higher secondary voltage than the nominal desired value. For example, if you are building a bridge rectifier power-supply circuit with an intended output voltage of +12 volts, the transformer’s secondary will need to put out somewhat more than 12 volts. Why? Because there is a voltage drop across the surge resistor, the filter resistor, and each of the diodes.

The power supply’s output voltage will always be somewhat lower than the transformer’s secondary voltage, because the power-supply circuit itself needs to use some of the power in order to operate. Manufacturers of transformers take this into account. This is why the standard secondary voltage ratings of off-the-shelf power transformers tend to look so awkward. A transformer rated for 6.3 Vac at its secondary would be used in a 5-volt or 6-volt power supply The other two most common transformer secondary voltages are 12.6 Vac (for 12-volt power supplies) and 25.2 Vac (for 24-volt power supplies).

In practical design work, the power rating specification is almost as important as the secondary voltage. This specification tells us the maximum load the transformer can reliably feed without the risk of burning itself out. The secondary voltage rating assumes the full rated load is being applied to the output of the transformer.

Usually the power rating is given as a wattage, or volts-amperes (VA) value. These two terms are interchangeable, since wattage is equal to the voltage times the current:


To find the maximum acceptable current drain, just divide the power rating by the secondary voltage:


For example, let’s assume we have a power transformer rated for a secondary voltage of 12.6 Vac, and a power rating of 15 watts, or 15 VA. The maximum current drain this transformer can safely handle is about:

I=15/12.6 = 1.2 amps

The transformer can handle brief load surges of a higher current, as long as this excessive power drain doesn’t continue too long.

Some manufacturers and parts dealers give the transformer’s power rating directly in current (amperes, or mA ( instead of in power (watts or VA). For most hobbyist work, this tends to be more convenient. If you need to determine the actual power, you can just multiply the current value by the secondary voltage rating, as suggested earlier:


The regulation factor is probably the least familiar transformer specification for most people working in electronics, whether professionally or on the hobbyist level. It’s generally less important than the other specifications we have already considered. In many practical applications, it can safely be ignored.

Earlier, we mentioned that the secondary voltage rating for a power transformer assumes that the full-power load (the power rating) is being applied to the transformer’s output (secondary). Reducing the current drain will cause the actual secondary voltage to rise. The regulation factor is a measurement of how much the secondary voltage will increase if the load current is reduced to zero. A typical regulation factor is about 10%. For a 12.6 Vac transformer, with a regulation factor of 10%, with little or no load current being drawn, the actual secondary output voltage could be as high as:


= 12.6 + (0.1 x 12.6)

= 12.6 + 1.26

= 13.86 Vac

(In this equation Vs stands for the rated secondary voltage of the transformer, and Vout is the actual output voltage, with a 0 ([or near 0]) load current.)

If the transformer is wired into the circuit backwards, it will act like a step-up transformer. For example, a standard 12.6 Vac power transformer wired backwards, so a 120 Vac source is applied across its nominal secondary winding (here being used as the primary), will produce an output voltage of 1,143 Vac. Only a small current can be drawn through the step-up transformer, since its power rating remains the same, and power is the product of voltage times current.

Accidentally hooking a power transformer up backwards in a power-supply circuit is liable to destroy every semiconductor and many of the other components in the power-supply circuit and in the load circuit.

For the sake of completeness, we will also mention the isolation transformer, which has an equal number of turns in its primary and secondary windings. The out put voltage is equal to the input voltage (ignoring some small losses within the transformer itself), but the output circuit will be electrically isolated from the input circuit by the transformer.

Selecting the diodes for use in a power-supply circuit is fairly easy. For the most part, we are just concerned with two key specifications. The first of these is relatively simple, although it could have several different names, which can lead to some confusion:

  • • PIV (Peak inverse voltage)
  • • PRV (Peak reverse voltage)
  • • Maximum reverse voltage
  • • Maximum reverse-bias voltage

You might encounter other names for the same specification. They all mean the same thing. What is the largest voltage the diode can safely and reliably block when it’s reverse-biased? If the applied voltage exceeds this maximum voltage, the diode’s pn junction will break down, and start to conduct heavily. The odds are very good that the diode will be permanently damaged or destroyed by this uncontrolled avalanche effect.

In power-supply circuits, the PIV just needs to be higher than the maximum expected input voltage on the primary winding of the transformer. Even the worst possible short circuit is unlikely to feed more voltage through the diode(s) than this. Still, it never hurts to allow a little extra headroom in the PIV specification.

For most power-supply circuits running off of ordinary line current, the diodes should have PIV ratings of at least 200 volts, but 300 volts or 400 volts (or even higher) would offer a little more protection at little or no added expense.

A more critical diode specification is the max (or peak) current-handling capability. All current drawn by the load circuit must pass through the diode(s). For safety and reliability, over-rate the diode on current as much as possible. For example, if the load circuit is expected to draw currents up to 1 ampere, don’t use a diode rated for just 1 ampere, or even 1.5 ampere. Use at least a 2-ampere diode, and preferably an even heftier unit.

In some cases, you might need to take the voltage drop across the diode(s) into consideration when designing a power-supply circuit. This is due to the small internal resistance of the pn junction when forward-biased. Usually the voltage drop across a forward-biased diode will be minimal, and can be considered negligible. There is some variation in the exact value, but for most silicon diodes (the type most commonly used in modern electronics), the nominal forward-biased voltage drop is about 0.7 volt-per-diode. If there are multiple diodes in series, the voltage drops will be cumulative. Older type germanium diodes have a lower voltage drop—typically only about 0.3 volt, but they are usually less reliable, and have lower maximum cur rent and PIV ratings than silicon diodes.

In a simple half-wave rectifier circuit, there is only one diode, of course. In a full- wave rectifier circuit there are two, and a bridge rectifier circuit has four. In these multiple diode circuits, all of the diodes should be matched—that is, they should all be of the same type number.

The lN400x series of diodes is very well suited to power-supply applications. The larger the last digit, the greater the maximum ratings of the device. The 1N4002 is good for low-power circuits, but I’d recommend using at least a 1N4003 or 1N4004 in most power-supply circuits, just to be on the safe side. In some high-power applications, you might need something like a 1N4007. I don’t recommend using the 1N4001 in power-supply circuits. The lN400x series diodes are all rated for 1 ampere maximum continuous current, but they can handle surges of up to 30 amperes. The difference is in the PIV ratings:

  • 1N4001 PIV = 50 volts
  • 1N4002 PIV = 100 volts
  • 1N4003 PIV = 200 volts
  • 1N4004 PIV = 400 volts
  • 1N4005 PIV = 600 volts
  • 1N4006 PIV = 800 volts
  • 1N4007 PIV = 1000 volts

If your power-supply circuit must supply more than about 0.75 ampere, I’d suggest using a heavier-duty diode, such as the 1N540x series, which is rated for 3 amperes continuous, and surges up to 200 amperes:

  • 1N5400 PIV = 50 volts
  • 1N5402 PIV = 200 volts
  • 1N5404 PIV = 400 volts

In multiple-diode circuits, especially bridge rectifiers, the full current load is shared by the diodes to some degree.

For the filter capacitor, as a rule of thumb, the larger the capacitance, the better the filtering. For most rectifier-based power-supply circuits, at least 100 uF should be used for the filter capacitor, and preferably a lot more. 500 uF and 1000 uF capacitors are commonly-used values. Since we are talking about such large capacitances, electrolytic capacitors are usually the only practical choice. Electrolytics are polarized, so they must be installed with care. If installed backwards, they won’t work, and will almost certainly be permanently damaged. In addition, there is a fairly good chance that an electrolytic capacitor connected to a reversed polarity for an extended period could explode.

The working voltage of the filter capacitor should be at least twice the nominal output voltage of the power-supply circuit. For example, if you are building a +12 volt power supply, use filter capacitors rated for at least 25 volts. However, don’t over-rate the working voltage too much. Some electrolytic capacitors can dry out and age prematurely if they are operated on too low a voltage for an extended period. They can even wear out just sitting on a shelf (applied voltage = 0). Newer de vices are less susceptible to such problems, but it’s still not a good idea to use a 1,000-volt electrolytic capacitor in a 12-volt circuit. Besides, it would be far more ex pensive and bulky than it needs to be to do the job properly.

The amount of ripple in a power supply’s output signal is directly proportional to the load current, and inversely proportional to the filter capacitor’s value. That is, in creasing the load current will tend to cause more ripple to appear in the output volt age. Using a larger capacitance value will reduce the ripple for a given amount of load current. This is illustrated in the graph of Fgr. 25.

_25 Using a larger capacitance value will reduce the ripple for a given amount of load current.

-Using standard voltage-regulator ICs-

Earlier in this section, we briefly mentioned voltage-regulator ICs. Now is the time to look at their use a little more closely.

The most popular hobbyist-level voltage-regulator ICs are the 78xx series. There are seven commonly available entries in this series. The “xx” part of the number identifies the regulated output voltage:

  • 7805 5 volts
  • 7806 6 volts
  • 7808 8 volts
  • 7812 12 volts
  • 7815 15 volts
  • 7818 18 volts
  • 7824 24 volts

The output of a 78xx voltage-regulator IC is always positive (with respect to ground or common). There is also a comparable 79xx series that puts out regulated voltages that are negative (with respect to ground, or common):

  • 7905 -5 volts
  • 7906 -6 volts
  • 7908 -8 volts
  • 7912 -12 volts
  • 7915 -15 volts
  • 7918 -18 volts
  • 7924 -24 volts

Each of these devices is available for a variety of current ratings. A voltage-regulator IC with a lower current rating will tend to be less expensive and bulky, but many applications will call for a larger current capability. Typical current ratings for voltage-regulator ICs include:

  • 100 mA (0.1 ampere)
  • 500 mA (0.5 ampere)
  • 1000 mA (1.0 ampere)
  • 3000 mA (3.0 amperes)

Inexpensive voltage-regulator ICs rated for more than 3 amperes are not easily found on the hobbyist market. Surplus houses that handle discontinued industrial- grade parts would probably be your best source for heftier voltage regulators.

The simplest and most direct way to use a voltage-regulator IC is in place of the R2 resistor in the output filter network of an ordinary rectifier-type power-supply circuit. A voltage-regulator chip can be used with a half-wave rectifier circuit, as shown in Fgr. 26, or a full-wave rectifier circuit. But for the best results, a bridge rectifier circuit, like the one shown in Fgr. 27, is recommended.

_26 A voltage-regulator chip can be used with a half-wave rectifier circuit.

Notice that filter capacitors are used on both the input and the output lines of the voltage regulator. The values of these filter capacitors can be much smaller than in an unregulated power-supply circuit. The input capacitor (C1) will usually have a value of about 0.1 uF to 0.5 uF, and should be mounted physically close to the body of the voltage-regulator IC itself. The value of the output capacitor is typically about 10 uF to 100 uF, with values in the lower end of this range being the norm.

To help the voltage regulator do its job most efficiently, a large filter capacitor of the usual type is also commonly used. It’s connected in parallel with the voltage regulator’s input, as illustrated in Fgr. 28. As usual, the larger the value of this capacitor, the better. It will usually be mounted at least a couple inches away from the voltage-regulator chip.

_28 To help the voltage regulator do its job most efficiently, a large filter capacitor is connected in parallel with the voltage regulator’s input.

Some designs omit the rectifiers altogether, and drive the voltage regulator directly from the secondary winding of the power transformer (through a surge resistor), as illustrated in Fgr. 29-—although this probably isn’t a good idea in most cases. The voltage-regulator IC will have to work a lot harder than usual in this circuit. At the very least, additional heatsinking should be used. Another disadvantage of this configuration is that the input signal seen by the voltage-regulator IC goes both positive and negative, which could damage the chip under some conditions. Usually it won’t be a problem, but it could become one rather unexpectedly.

_29 Some power-supply circuits omit the rectifiers altogether, and drive the voltage regulator directly from the secondary winding of the power transformer through a surge resistor.

_27 For the best results, a bridge-rectifier circuit is recommended for use with a voltage-regulator IC.

It’s generally best to use a rectifier circuit before the voltage regulator so at least the chip’s input is closer to the desired output voltage, and at least approximately resembles DC .

==Changing the output voltage==

In some applications, the power-supply circuits we’ve described so far might not be satisfactory because they don’t put out the desired level of voltage. For example, you might want to build a regulated power supply that will drive a 9-volt device, but there is no 7809 voltage-regulator IC— you have to use either a 7808 (8 volts) or 7812 (12 volts), and neither could be close enough. Some applications might require a supply voltage that can be fine-tuned over a specific range. And there are a few cases when a given power supply might need to be used to drive different loads at different times, each with differing supply voltage requirements. An example of this is a universal power supply on an electronics technician’s workbench, which is used for various testing and design functions.

An obvious solution would be to add a simple voltage-divider network across the power supply’s output, as shown in Fgr. 30. Often this will be good enough, but it can often defeat the purpose of a voltage regulator somewhat. Remember, the load circuit electrically acts like a variable-resistance element, changing its value as the current drawn by the load changes. This variable resistance load is in parallel with the lower half of the voltage divider network, inevitably affecting its value, causing the supplied voltage to fluctuate with changes in the current drawn by the load. Such a crude solution therefore usually won’t be too helpful in most practical applications.

_30 An obvious, though not always effective way to obtain a different output voltage from a fixed-voltage regulator chip is to add a simple voltage-divider network across the power sup ply’s output. Unregulated voltage in; +Vb; +Vc; Common

A simple 78xx series voltage-regulator IC has just three leads, and is so simple, it would seem you’d be pretty much stuck with its designed output voltage. After all, there is no “voltage adjust” pin or anything like that. These chips were designed for fixed voltages, but they can be tricked into operating at voltages somewhat different from their original design value.

The secret is in where you put the true ground in the circuit. The output voltage of a three-pin voltage-regulator IC is referenced to its common pin. Ordinarily, this pin is connected to true ground potential, so the output voltage from a 7805, for ex ample, is five volts above ground—that is, just plain +5 volts.

Most voltage-regulator ICs draw only a small quiescent current (typically just a few milliamperes) that flows to ground from the common pin. By applying a bias voltage to the common pin, we can “float” it above ground, and fool the voltage regulator into putting out a higher-than-normal output voltage.

The simplest way to do this is to add a variable resistance (normally a potentiometer of some sort) between the common pin and true circuit ground, as illustrated in Fgr. 31. If you don’t need a true variable output voltage, just an “oddball” value, you could substitute a fixed resistor with an appropriate value. The best and most re liable way to find the correct resistance is to temporarily hook up a potentiometer, as shown here, and adjust it carefully for the desired output voltage. Then, without moving the potentiometer’s shaft, carefully measure its resistance value. You will probably need to use a precision resistor to come close enough to the desired value.

_31 The easiest way to “fool” a voltage regulator into putting out a higher than normal output voltage is to add a variable resistance between the common pin and true circuit ground. Unregulated voltage in; Regulated voltage out

This common-to-ground resistance should not be very large. I would not recommend using much more than a 1-k-Ohm (1,000 ohms) potentiometer. You can “fool” a 78xx just so much before the trick ceases to work reliably, and the IC could possibly be damaged.

This circuit is simple and inexpensive, but the regulation of the output voltage can be adversely affected. The output voltage will change with any shifts in the quiescent current, which can occur for a variety of different reasons.

An improved variable output voltage-regulator circuit is shown in Fgr. 32. Here we have added a second, feedback resistor (R2) from the chip’s output to its common pin. A value of about 1 k-Ohm (1,000 ohms) should be used in most cases. You might want to decrease the maximum value for potentiometer R1 just to be safe. Try using a 500-Q potentiometer in this circuit. Yes, you will get a smaller range of output voltages with a smaller potentiometer, but this is a reasonable trade-off for significantly improved regulation of the output voltage.

_32 This is an improved variable-output voltage regulator circuit.

For even greater stability when adapting a standard voltage-regulator IC for a fixed “oddball” output voltage, you can connect a zener diode from the common pin to circuit ground as shown in Fgr. 33. An output to a common feedback resistor should always be used in this case, probably with a somewhat higher value, say, around 3.3 k (3,300 ohms) to 4.7 k-Ohm (4,700 ohms). The circuit’s output voltage will be equal to the sum of the voltage-regulator IC’s nominal output voltage, and the zener diode’s avalanche voltage. For example, let’s say we are using a 4.2-volt zener diode with a 7815 voltage regulator. The output of this circuit will then be a fairly well-regulated 19.2 volts (4.2 + 15 volts). The regulation won’t be quite as good as the voltage regulator by itself, but it will still be reasonably close to the original specifications, and should be adequate for most practical applications. It’s certainly better than nothing.

Semiconductor manufacturers quickly recognized the usefulness of such circuit design tricks, and they soon came out with special adjustable voltage-regulator ICs that were intentionally designed to accommodate such external changes to the out put voltage. One of the simplest devices of this type is the LM317, which has three terminals just like a 78xx voltage regulator, except instead of INPUT—COMMON—OUT PUT, the pin functions of the LM317 are INPUT—ADJUST—OUTPUT.

_33 Better variable voltage regulation can be achieved by adding a zener diode.

The basic LM317 variable-output voltage-regulator circuit is shown in Fgr. 34. It can be operated reliably over a much wider range and with better regulation than a 78xx chip or other fixed voltage-regulator device. The LM317 is designed to accept input voltages from 4 to 40 volts, and can put out regulated voltages ranging from 1.25 to 37 volts. The current rating for this IC is 1.5 amperes.

_34 This is the basic LM317 variable-output voltage regulator circuit.

A similar device is the LM338, which can handle currents up to 5 amperes. The input voltage range is the same as for the LM317, but the maximum regulated out put voltage is just 32 volts for the LM338. The lower end of the output voltage range is still 1.25 volts.

A somewhat more sophisticated and versatile variable-voltage-regulator IC is the 723. This chip is housed in a standard 14-pin DIP package, but three of the pins are not internally connected to anything. The pin-out diagram for the 723 is shown in Fgr. 35.

This chip will accept unregulated input voltages of up to 40 volts, and its regulated output voltage can be anything from 2 to 37 volts. Normally, the maximum out put current for the 723 is just 150 mA (0.15 ampere), but it’s a fairly simple matter to add some external power transistors as current amplifiers to supply currents of up to 10 amperes.

For a variety of technical reasons, slightly different circuitry should be used for low-output voltages (2 to 7 volts) than for higher-output voltages (7 to 37 volts). The low-voltage version is shown in Fgr. 36.

The reference voltage Vref is fed into pin 6, through resistors R1 and R2. This reference voltage will normally be between 6.8 volts and 7.5 volts. The formula for this circuit’s output voltage is:

The value of resistor R3 should be equal to the parallel combination of R1 and R2. That is:

_35 A somewhat more sophisticated and versatile variable voltage-regulator IC is the 723.

_36 This 723 circuit can be used to generate low output voltages (from 2 to 7 volts).

For best results, precision resistors are recommended, but for general purpose applications, standard 5% resistors should be close enough. Standard resistor values for some typical output voltages are summarized in TBL 3. Remember, these resistor values are rounded off to the nearest standard values, so the output voltages won’t be exact.

___ TBL 3. Standard resistor values for some typical output voltages for the low-voltage 723 power-supply circuit of Fgr. 35

The 7- to 37-volt version of the 723 variable voltage-regulator circuit is shown in Fgr. 37. Basically, we’ve just moved the resistors around some. The reference voltage Vref is still on pin 6, and resistor R3 is still equal to the parallel combination of R1 and R2:

_37 This 723 circuit can be used to generate higher output volt ages (from 7 to 37 volts).

In some applications, resistor R3 is optional, but including it gives the circuit better temperature stability. Resistors are so inexpensive, I see little reason to want to omit this component.

The output voltage formula for this version of the 723 is a little different:

Once again, for the best possible results, precision resistors are recommended, but for general-purpose applications, standard 5% resistors should be close enough. Standard resistor values for some typical output voltages in this circuit are summarized in TBL 4. Remember, these resistor values are rounded off to the nearest standard values, so the output voltages won’t be exact.

__ TBL 4. Standard resistor values for some typical output voltages for the high-voltage 723 power-supply circuit of Fgr. 36.

-Current regulation-

Ordinarily, the load determines how much current it will draw from the power supply in response to the voltages and resistances within the load. The current drawn by the load is determined by Ohm’s law.

In some specialized applications, however, we want the current level to be consistent and independent of the load. In such applications, we need a fixed cur rent source, which always puts out a specific, pre-determined amount of current. This type of circuit is sometimes called a constant current source, or a current limiter.

Roughly speaking, this type of circuit could be considered a current regulator, analogous to a voltage regulator. The output current in this case is regulated, or held to a specific value, and is not permitted to fluctuate during normal operation. A simple, but practical fixed current source circuit built around one section of an LM3900 quad Norton amplifier IC is illustrated in Fgr. 38. A typical parts list for this circuit is given in TBL 5.

Because the output current from this type of circuit is related to the voltage, the actual power supply should be well-regulated, with the voltage regulation circuitry coming before the current source circuit.

Throughout the following discussion, we will assume that a +15-volt voltage regulator is being used to drive the fixed current source circuit. Using the component values from the suggested parts list of TBL 5, the output current from this circuit will be fixed at 1 mA (0.001 ampere).

_38 A LM3900 Norton amplifier can be used as the heart of a simple, but practical fixed-current source circuit.

TBL 5. Suggested parts list for the fixed-current source circuit of Fgr. 37.

IC1 LM3900 quad Norton amplifier (one section only)

Q1 pnp transistor (2N3906, or similar)

The load (RL) being driven by the fixed current must be connected between the collector of output transistor Q1 and ground. The current source will hold its constant value as long as the load impedance is no greater than about 14 k (14,000 ohms). If the load impedance is higher than this, the actual current value will drop. But the current will never exceed its rated, constant value, regardless of the load resistance. The load impedance might drop all the way down to 0 ohms (a dead short) without affecting the current source.

The input to the Norton amplifier (IC1) is derived from a simple voltage divider network made up of resistors R2, R3, and R4. The values of these resistors are selected to present a 14-volt input to the Norton amplifier’s non-inverting input, through input resistor R1.

For high-precision applications, it’s a good idea to use high-grade, 1% tolerance resistors in this voltage-divider network. Any inaccuracy in these resistance values will affect the output current value. You could simplify the circuit slightly by replacing R3 and E4 with a single 14-kQ 1% tolerance resistor. Unfortunately, this is not a standard value for 5% tolerance resistors, which is why we had to make it up from a 10-kQ resistor and a 3.9-kQ resistor in series.

In this circuit, feedback resistor ES has the same value as input resistor El, so we have a non-inverting unity gain amplifier:


=1,000,000/ 1,000,000


The Norton amplifier automatically adjusts its output to provide an output volt age at the junction of resistors ES and E6, that is identical to its input voltage. Again, using 1% resistors for El and ES will improve the overall accuracy and precision of the circuit.

This voltage is also fed to the emitter of transistor Q1, while the direct output signal from the Norton amplifier feeds the transistor’s base. The collector is connected to the external load circuit or device.

Because there are +14 volts at the ES end of resistor R6, and the full supply volt age (+15 volts) at the other end of this resistor, it necessarily follows that the voltage drop across this component must always be 1 volt. Knowing the value of resistor R6 (from the parts list), we can now use Ohm’s law to find the current flowing through it. For greatest accuracy, a 1%-tolerance resistor is recommended for E6 too.

According to the parts list, resistor R6 has a value of 1 k-Ohm (1,000 ohms). The current flowing through this component therefore works out to:



= 0.001 ampere

= 1 mA

This 1-mA current is derived from the emitter of transistor Q1. You should recall from your basic electronics theory that a transistor’s emitter and collector currents are virtually identical, so the output current to the load (EL) is also about 1 mA, and is held at that constant value, regardless of any small-to-moderate fluctuations in the load impedance of EL.

The fixed-output current from this circuit can be increased by decreasing the value of resistor E6. Reducing this resistance by half doubles the output current. It’s a fairly simple matter to use Ohm’s law to calculate the necessary value of resistor R6. Assuming that none of the other values throughout the fixed current source circuit are changed, the formula is:



The voltage drop across R6 should always be 1 volt, assuming all other resistor values (and the supply voltage) remain the same. Id is the desired output current value in amperes (not mA).

Don’t try to make the output current value for this circuit too large, or the transistor, the IC, or both, could be damaged. Check the manufacturer’s specification sheets for the particular semiconductor components you are using in your circuit to determine the maximum safe output current value. Be sure to leave some headroom—don’t try to use a current value right at the component’s absolute maximum limit.

Transistor Q1 can be almost any standard pnp-type unit. The 2N3904 recommended in the parts list is a good, general-purpose choice. Just make sure that the transistor you select for use in this circuit can safely and reliably handle the desired output current. As a general rule of thumb, over-rate the transistor’s current-handling capability by at least 20% to 30%. That is, if the manufacturer says the absolute maximum current for your transistor is 2.5 amperes, treat it as if the practical maxi mum current value is somewhere between 1.75 amperes and 2.0 amperes.

==Project #1—Multiple-output, dual-polarity power supply==

I have already given all the significant technical details for the two power supply projects of this section. In both projects, we are simply putting together several circuits that have already been described.

In certain applications, multiple-supply voltages, perhaps both positive and negative, might be needed by different sections of the load circuitry. Our first project is designed to simultaneously put out three positive voltages and three negative volt ages, each individually regulated.

The schematic diagram for this multiple-output dual-polarity power supply project is shown in Fgr. 39. The suggested parts list for this project is given in TBL 6.

Notice that a center-tapped power transformer is required in this project to permit both positive-(above ground) and negative- (below ground) output voltages. The transformer’s center-tap is grounded. The full secondary voltage of the power transformer must be a little more than twice the largest desired output voltage, which we are assuming to be 15 volts. Fifteen volts above ground (+15 volts) plus fifteen volts below ground (—15 volts) equals a total of 30 volts peak-to-peak. The nearest standard transformer voltage is 36 volts, which is just about perfect for our purposes here.

For the time being, let’s just consider the positive side of the circuit. The unregulated positive voltage is tapped off the bridge at the junction between D2 and D4. C1 is a large filter capacitor. It’s exact value is not critical, but the larger it is, the better job of filtering it will do. The nominal DC voltage at this point is approximately 17 volts—one half the voltage of the secondary winding (referenced to ground) less the normal voltage drop across the active bridge diodes. Let’s mentally disconnect resistor R1 from the circuit for now. This means there is no current path to the circuitry surrounding 1C2 and 1C3—they are not part of the circuit until R1 is replaced.

_39 Project # 1—Multiple-output, dual-polarity power supply.

__ TBL 6. Suggest parts list for Project #1—Multiple-output, dual-polarity power-supply of Fgr. 39

7815 +15-V, 500-mA voltage-regulator IC

7812 +i2-V, 500-mA voltage-regulator IC

7805 +5-V, 500-mA voltage-regulator IC

7915 —15-V, 500-mA voltage-regulator IC

7912 —12-V, 500-mA voltage-regulator IC

7905 —5-V, 500-mA voltage-regulator IC

Power transformer—secondary 36 Vac, center-tapped

F1 4-A fuse and holder

R1, R4 2.7-k-Ohm, 5%, 1/4-W resistor

R2, R5 8.2-k-Ohm, 5%, 1/4-W resistor

R3, R6 6.8-k-Ohm, 5%, 1/4-W resistor

C1, C2 2,500-pF, 25-V electrolytic capacitor

C2, C4, C6, C9, C11, C13, 0.22-uF capacitor

C3, C5, C7, C10, C12, C14 10- 20-V electrolytic capacitor

This leaves us with just a simple, standard-voltage circuit, built around IC1. This IC is a 7815, so it’s regulated-output voltage is +15 volts. Capacitors C2 and C3 are just the standard filter capacitors, almost always used with 78xx series voltage regulators.

So far, we have nothing unusual. Now, let’s mentally reconnect resistor R1 to the circuit. This resistor, along with R2 and R3, forms a simple resistive voltage-divider network. Using the component values suggested in the parts list, 1C2 (a 7812) sees an unregulated-input voltage of a little less than + 14.5 volts, and puts out a regulated +12 volts. Similarly, 1C3 (a 7805) sees an unregulated input voltage of a little more than +6.5 volts, and puts out a regulated +5 volts.

Because the resistive-voltage divider network comes before the voltage regulators, there are no loading effects, and fluctuations in the load circuit make no difference.

Strictly speaking, these resistors aren’t absolutely necessary. Even a 7805 can take an unregulated input voltage up to about 30 volts, but there seems little reason to make the voltage-regulator chip work that hard. The three resistors are likely to be considerably less expensive and bulky than the additional heatsinking that might be required without them.

If you are using voltage regulators with larger output currents than what is recommended in the parts list, you might need to use resistors with higher wattage ratings. If the resistor can’t handle the current drawn through it, it could change value, which usually won’t be too much of a problem in this particular application. More seriously, the resistor could burn itself out totally, and act essentially like an open circuit. Any later voltage-regulator stages won’t receive any input voltage, so that output will be dead.

The negative-output voltage section of the circuit works in just the same way, except 79xx voltage-regulators are used in place of the positive 78xx devices of IC1, IC2, and IC3. The unregulated negative voltage is tapped off between bridge diodes D1 and D3.

The input fuse (F1) should be selected to handle a little more than the sum of the maximum output currents for each voltage regulator. Since the parts list recommends 500-mA voltage regulators, and there are six of them, this is a total acceptable current of 3 amperes. A 3.5-ampere to 4-ampere fuse will offer sufficient headroom, but still should blow before any damage is done in case of a short circuit in the power supply or in the load circuit. For greater protection, you might want to add additional fuses in each of the output lines. Using the recommended voltage-regulators, each output fuse should be rated for ampere (500 mA) Use automotive fuses here, because they are designed for lower DC voltages. A regular 120-volt fuse might not blow in time, even if the rated current value is exceeded.

==Project #2—Variable-output, current-limited power supply==

Our next power supply project permits manual adjustment of the output volt age. In addition, the output current is limited, and the user can manually adjust the maximum output current as well. The schematic diagram for this project is shown in Fgr. 40, with a suitable parts list appearing in TBL 7.

Once again, we are just putting together a couple of the basic circuits discussed earlier in this section. This project is designed around an LM317 adjustable voltage- regulator IC (IC 1). We’re using just the basic LM3 17 variable output voltage circuit here, except for the addition of diode D5 and capacitor C3, which improves the stability and reduces the output ripple. This circuit can offer a ripple rejection figure of up to 80 dB, which is excellent for most practical purposes.

Another addition to the basic LM317 circuit here is meter M1. This is just a small DC voltmeter, to permit the user to know what output voltage the power supply is currently set for. The output voltage is adjusted via potentiometer R3, which should be a front panel control, mounted as close as possible to Ml for convenience.

The output of the LM3 17 voltage-regulator circuit is then fed through the same current-limiter/fixed current-source circuit we discussed in an earlier section. A milliammeter (M2) is added in series with the current determining resistance, which in this circuit is comprised of the series combination of R8 and R9. R9 is another front panel mounted potentiometer, near M2. Resistor R8 is included to prevent the possibility of setting R9 too close to zero.

Because of the placement of the milliammeter (M2) in this circuit, the actual output current being drawn by the load circuit is not indicated. The actual load cur rent might be lower than the reading on the meter, but it won’t be permitted to exceed it. I believe this is more useful in practical applications. The user can adjust R9 for the desired current limit, without worrying about what the load impedance is at the moment. There is also less possibility of loading problems in this arrangement.

In use, adjust the desired output voltage first (via R3), then select the desired current limit (via R9). Don’t reverse this sequence. The current limit is dependent on this circuit’s input voltage. Changing the voltage setting, without moving the shaft of R9, will result in a different current-limit setting.

Because of the inherent inaccuracies of this voltage dependence, there wouldn’t be much point in using high-precision 1% tolerance resistors for R4 through R8. Any inaccuracies in these resistances can be compensated for by adjusting potentiometer R9 until the desired current value is read on the milliammeter (M2).

The three extra sections of the quad-Norton amplifier chip (IC2) can be left disconnected, or they can be used independently (except for the supply voltage) in other circuitry as part of a larger system.

_40 Project #2—Variable-output/current-limited power supply.

__ TBL 7. Suggested parts list for Project #2—Variable-output, current-limited power supply of Fgr. 39.

  • IC1 -- LM317K adjustable voltage-regulator IC
  • IC2 -- LM3900 quad Norton amplifier (one section only)
  • Q1 -- pnp transistor (2N3904, or similar) 1N4003 diode
  • D1, D2, D3, D4, D5
  • T1 Power transformer—secondary 40 Vac
  • F1 3-amp fuse and holder
  • R1, R2, R8 100-ohm 5%, ¼-W resistor
  • R3 5- potentiometer
  • R4 1-k-Ohm, 5%, ‘%-W resistor
  • R5 -- 10-ku, 5%, ‘A-W resistor
  • R6, R7 1-Me, 5%, Y resistor
  • R9 2.5-k-Ohm potentiometer
  • C1 -- 2,200-uF, 50 V electrolytic capacitor
  • C2 0.1uF- capacitor
  • C3, C4 -- 10-uF, 50 V electrolytic capacitor
  • M1 DC voltmeter 0—50 V
  • M2 DC milliammeter 0—1 amp

This power supply project is designed to provide positive-regulated output volt ages ranging from +2 to +37 volts. Major changes h the design would be needed to accommodate negative-output voltages.


Top of Page

PREV   NEXT   Guide Index HOME