Digitally-controlled AC voltage stabilizer

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to a voltage stabilizer apparatus for use with alternating current (AC) power sources. In particular, the present invention relates to an AC voltage stabilizer with digitally-controlled emulation of servotransformer operation using tap-switching technology.

2. Discussion of Background

AC power lines are subject to a number of different types of voltage disturbances. These include spikes (i.e., pulses of very high voltage and current), which are most commonly caused by lightning and usually last less than a second; surges and sags, which are periods of less severe, too-high or too-low voltage lasting from several seconds to a few minutes, usually caused by faulty power company switching or by other devices on the line; brownouts, which are longer sags lasting from several minutes to several hours; and blackouts, which are periods of near-zero voltage, which can be caused by blown fuses or tripped circuit breakers after a lightning strike or other line problem.

Users of computers and other electronic equipment are familiar with the problems caused by power-line spikes and surges. So-called “surge-suppressing” (more correctly, “spike-suppressing”) power strips or adapters are now widely available. However, these devices are of little value in protecting against sustained power-line surges and of none at all against sags or brownouts. In areas where severe sags, surges or brownouts occur frequently, voltage-sensitive devices such as motors, computers and other electronic equipment, and even ordinary light bulbs have a limited life expectancy. These problems are ubiquitous in the historically less-developed countries, and even in remote areas of the developed countries where generators or long stretches of low-tension cable may be needed to provide electrical power. In many countries, the infrastructure needed to support widespread use of these technologies is inadequate. In particular, the lack of steady, dependable electrical power has proved to be an important factor in limiting the potential market for electrically-powered consumer goods (refrigerators, air conditioners, television sets, personal computers, etc.).

Many methods and devices are available for stabilizing local line voltage, but typically represent difficult compromises, playing off smoothness and accuracy of control against cost. Most commercially-available voltage stabilizers rely on transformers whose step-up or step-down ratios can be changed to help compensate for input-voltage changes. These devices fall into two broad classes: discrete tap switchers or “relay boxes,” and servo-transformers.

Tap switchers are simple and inexpensive, but do not provide smooth voltage regulation. In such a device, one or more relays select various taps of a transformer (typically an autotransformer) so that the output voltage is raised or lowered in steps to help compensate for changes in the input voltage. This principle is illustrated in

FIG. 1A

for a nominal 230-volt line using a single relay, where an autotransformer

10

is provided with three taps

12

,

14

and

16

, arranged so that an AC voltage applied between taps

12

and

14

results in a higher (“stepped-up”) voltage appearing between taps

14

and

16

. (For clarity, transformer

10

, an iron-cored AC power transformer, is shown as a simple coil or series of windings in FIG.

1

A and the following Figures.) The selected step-up ratio depends upon the particular application, but is typically within the range of about 10-30%.

In an unbalanced circuit, the AC (alternating current) hot line is normally connected to tap

12

through a terminal

12

a,

and the neutral line to tap

14

through a terminal

14

a.

A relay

20

connects either the input voltage at tap

12

or the stepped-up voltage at tap

16

to a terminal

22

. Relay

20

is driven by a control circuit (represented schematically as

24

) which selects tap

12

when the input voltage is generally above a threshold

26

, or tap

16

when the input voltage is generally below this threshold. By “generally” it is meant that the control is not exact; hysteresis effects are usually present, and indeed are desirable to prevent relay chatter and prolong the life of relay

20

. The output is then taken between terminal

22

, which is the new AC hot line, and neutral output terminal

14

b.

Threshold

26

is chosen so that the output voltage remains in the vicinity of a selected voltage represented by a line

36

(FIG.

1

B), over a wider range of input voltages than if no compensation were made.

The result is shown graphically by line segments

30

,

32

and

34

(FIG.

1

B). When the input voltage is below threshold

26

, and neglecting the effects of loading, relay

20

selects tap

16

. The output voltage measured between taps

14

and

16

is higher than the input voltage by the step-up factor of transformer

10

, typically between about 10-30%, as shown by line segment

30

. As the input voltage rises above threshold

26

, control circuit

24

causes relay

20

to change position, selecting tap

12

instead of tap

16

, and the output voltage changes abruptly as shown by line

32

. At still higher input voltages, the output equals the input as shown by line segment

34

.

If the input voltage then drops below threshold

26

, relay

20

again changes position to select tap

16

and the output voltage is again stepped up to compensate for the input voltage drop. For comparison, line segment

38

represents the output voltage if no compensation is made.

When transformer

10

is supplying current to a load, the graph shown in

FIG. 1B

is not completely accurate since some voltage drop occurs within the transformer, and the output voltage is correspondingly lower. Hence, selecting the transformer tap according to the output voltage, rather than the input, would in theory provide superior control. In commercial tap switchers, however, this is not often done because it can cause oscillation between positions, so that the “stabilizer” actually makes the output voltage less rather than more stable.

The quality of voltage regulation available from a tap switcher can be improved by adding more steps, so that the intervals between the steps can be made smaller and/or so that a wider span of input voltages can be handled. More steps, however, require correspondingly more relays and transformer taps, resulting in greater circuit complexity and higher cost. These types of devices normally operate in simple “daisy-chain” fashion, so that a separate tap and relay are required for each step.

For example,

FIG. 2A

illustrates a tap switcher using an autotransformer

10

with four taps (not separately labeled) and three relays

20

a,

20

b,

and

20

c

acting successively at three different thresholds

26

a,

26

b,

and

26

c,

respectively, under the control of circuit

24

. Such a stabilizer might, for example, provide either straight-through operation as shown by line segment

34

(FIG.

2

B), boosts of 20% or 40% for line undervoltages as shown by line segments

30

a

and

30

b,

respectively, or 20% bucking (voltage decrease) for line overvoltages as shown by line segment

30

c.

Line segments

38

a

and

38

b

represent the output voltage if no compensation were made.

As in the circuit of

FIG. 1A

, abrupt voltage steps

32

a,

32

b

and

32

c

occur each time a relay switches. However, a larger number of steps permits either stabilization of the output voltage near a desired level

36

over a wider input voltage range, smaller individual steps so that the output voltage remains closer to level

36

over this range, or both.

Some products designed mostly for audiophile or recording-studio use perform a similar function to the circuits of

FIGS. 1A and 2A

using solid-state switching devices rather than electromechanical relays. This approach speeds response and minimizes switching noise, but tends to be quite costly because of the added complexity of the circuitry.

Servo-transformers provide much better and smoother, nearly stepless regulation than tap switchers, but at the cost of greater mechanical complexity, higher price, and relatively slow response. Such a device and its voltage response are shown in

FIGS. 3A and 3B

. Instead of multiple discrete taps, autotransformer

10

is typically equipped with only two taps

12

and

14

plus a sliding brush

50

which is able to contact any selected one of the transformer windings. This is usually accomplished by winding transformer

10

with heavy copper wire on a ring-shaped armature, then grinding one end surface flat (including all windings) so that the flat surfaces of successive windings lie in a common ring-shaped plane. The ground surfaces of the windings are usually plated with a precious metal to prevent tarnishing. The brush, which is usually made of graphite, is mounted on a pivoting arm so that, as the arm rotates, the brush contacts each winding in turn.

As brush

50

moves along the windings of transformer

10

, and with an AC voltage applied between taps

12

and

14

, the step-up or step-down ratio depends upon the position of brush

50

. This ratio changes as brush

50

is moved from one winding to another. Typically, one tap

14

is located at an end of the winding and serves as a common neutral, while the other tap

12

is located partway along the winding and serves as the hot input. Brush

50

is connected to output terminal

22

. As a result, the ratio of the output to input voltages can be changed from near zero to well above unity in a series of very small steps: one step for each turn of winding

10

which is exposed along the path of brush

50

.

A mechanically-variable autotransformer such as this, without other components, is sometimes called a “variac.” Because of its mechanical complexity and the grinding and plating operations required in its manufacture, a variac is typically more expensive, as well as significantly larger and heavier, than a multitapped autotransformer with the same overall power rating.

A variac can be made into a voltage stabilizer by adding a servomotor (represented schematically as

52

) which moves brush

50

back and forth along the windings of transformer

10

under the control of a circuit

24

. Linkage between servomotor

52

and brush

50

is usually through a gearbox or other suitable type of speed reducer

54

.

Since the voltage steps are small, feedback control can be used to minimize output-voltage changes as the load varies. Typically, two threshold voltages

60

a

and

60

b

are set up,

60

a

above and

60

b

below a desired output voltage

36

, and spaced further apart than the maximum voltage step which occurs on the motion of brush

50

from one winding of transformer

10

to the next adjacent winding. The output voltage is monitored, typically by being rectified, filtered and compared with DC (direct current) references corresponding to these thresholds. Should the output voltage move above level

60

a,

brush

50

is moved to contact a lower-voltage turn of winding

10

. Conversely, should the output voltage move above below

60

b,

brush

50

is moved to contact a higher-voltage turn. The output voltage is continually stabilized in this way.

On a rising input voltage, the output voltage repeatedly rises to threshold

60

a

and is corrected downward in small steps throughout the range over which stabilization is possible, as generally indicated by line

62

(FIG.

3

B). If the input voltage falls again, the output voltage will repeatedly decrease to threshold

60

b

and be corrected upward in small steps, as generally indicated by line

64

. (For purposes of illustration, thresholds

60

a

and

60

b

are shown relatively further apart, and the voltage correction steps in lines

62

and

64

correspondingly larger, than would be typical in a real stabilizer. As in

FIG. 1B

, line

34

represents the output voltage if no correction were being performed.)

Since a variac can provide voltage step-up or step-down ratios ranging down almost to zero but upward only to some set ratio above unity, the range over which it is possible to keep the output near desired output level

36

ranges from a minimum voltage

66

set by the maximum step-up ratio, upward to a limit

68

determined by the ability of winding

10

to withstand overvoltage. With an input voltage less than voltage

66

, the output voltage equals the input voltage multiplied by the maximum step-up ratio as indicated by a line

70

. Operation at an input voltage greater than limit

68

should be avoided, since equipment damage could occur.

Because of the need for mechanical motion, correction for input-voltage or load changes in the circuit of

FIG. 3A

is typically rather slow. If control circuit

24

is improperly designed, moreover, or if thresholds

60

a

and

60

b

are set too close together, brush

50

may move too far before circuit

24

can respond. In such a case, oscillation takes place, the output to the load becomes unstable, and brush

50

may quickly wear out.

The advantage of a servo-transformer voltage stabilizer such as that shown in

FIG. 3A

is that, since it provides a large number of very small steps, voltage regulation can either be very accurate, be performed over a very wide range of input voltages, or both. Because of its mechanical complexity, however, such a stabilizer is larger, heavier, and more costly than a tap switcher of equal power-handling capacity. Because of the large number of moving parts it contains, it is also more liable to failure and requires much more frequent maintenance.

Response approximating that of a servo-transformer can be achieved using conventional tap-switching technology if a very large number of taps and relays are used. However, because of the expense of so many relays and the complexity of the circuitry needed to drive them, and because the geometry of a conventional power transformer limits the number of taps which can be placed on its windings to a dozen or so, this is not achievable at reasonable cost.

Additional types of voltage stabilizers use devices such as ferroresonant transformers. These provide much smoother voltage regulation and have no moving parts, but are cumbersome, relatively costly, and often generate a continuing audible hum that is annoying to many listeners.

There is a need for a voltage stabilizer apparatus that provides smooth, noise-free regulation in a relatively small, lightweight package, at low cost and with high power efficiency.

SUMMARY OF THE INVENTION

According to its major aspects and broadly stated, the present invention is a voltage stabilizer apparatus that uses cost-effective, digitally-controlled emulation of servotransformer operation to provide smooth voltage control. The invention makes use of tap-switching on both the primary and the secondary sides of a transformer, thereby allowing a relatively small number of relays to provide a relatively large number of voltage step-up and step-down ratios.

An important feature of the present invention is the use of tap-switching technology to emulate servotransformer operation, that is, to “servo” a voltage up or down under control of a feedback loop. In a conventional tap switching circuit, the number of step-up and step-down ratios is equal to the number of relays used to perform the switching, that is, one less than the number of taps. By switching on both the primary and secondary sides of an autotransformer, the maximum number of achievable ratios is increased to N

max

=(m*n−p+1) where m is the number of primary taps, n the number of secondary taps, and p the number of taps which are common to both sets. For m=n=p=5, for example, there are 21 possible ratios, more than four times as many as for a conventional tap switcher with 5 relays. For m=n=p=10, the invention offers 91 possible ratios instead of the 9 afforded by a conventional device.

Another feature of the present invention is the spacing of the taps. For an approximately evenly spaced series of step-up and step-down ratios, the taps are placed at logarithmic (or approximately logarithmic) intervals along the transformer winding. Two overlapping sets of such taps can be used: one set placed at a basic narrow spacing ratio R

n

and the other at a wider spacing ratio Rw, so that combinations of the two sets form ratios at approximately equal logarithmic spacing. Using one set of taps as the inputs and the other set as the outputs then gives the desired succession of ratios.

Still another feature of the present invention is the ability to add additional components to the apparatus, including but not limited to devices for sensing under- or over-voltages, devices for preventing unwanted cycling under load, rectifiers, arc-suppression devices, and circuitry for interfacing with the user's electronic equipment. If desired, LEDs or other indicators may be included for providing feedback on the operational state of the voltage stabilizer apparatus.

Other features and advantages of the present invention will be apparent to those skilled in the art from a careful reading of the Detailed Description of Preferred Embodiments presented below and accompanied by the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

In the drawings,

FIG. 1A

is a circuit diagram of a tap switching circuit;

FIG. 1B

is a graph of the output voltage vs. the input voltage of the circuit of

FIG. 1A

;

FIG. 2A

is a diagram of another tap switching circuit;

FIG. 2B

shows the output voltage vs. the input voltage of the circuit of

FIG. 2A

;

FIG. 3A

shows a typical servo-transformer;

FIG. 3B

shows the output voltage vs. the input voltage of servo-transformer of

FIG. 3A

;

FIG. 4A

is a voltage stabilizer with a multitapped autotransformer according to a preferred embodiment of the present invention;

FIG. 4B

shows the output vs. the input voltage of the circuit of

FIG. 4A

;

FIG. 5

shows a control circuit according to the present invention;

FIG. 6

is a tap-selecting device according to the present invention;

FIG. 7A

is another tap-selecting device according to the present invention;

FIG. 7B

is a graph of the output vs. the input voltage of the circuit of

FIG. 7A

; and

FIGS. 8-13

illustrate the components of a voltage stabilizer circuit according to another preferred embodiment of the present invention, wherein

FIG. 8

shows a tap selecting circuit,

FIG. 9

shows a rectifying and averaging circuit,

FIG. 10

shows a voltage comparator circuit and subsequent control logic,

FIG. 11

shows an electromechanical relay coil with a driver,

FIG. 12

shows an arc suppression network, and

FIG. 13

shows a low-pass resonant filter.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

In the following detailed description of the invention, reference numerals are used to identify structural elements, portions of elements, surfaces or areas in the drawings, as such elements, portions, surfaces or areas may be further described or explained by the entire written specification. For consistency, whenever the same numeral is used in different drawings, it indicates the same element, portion, surface or area as when first used. Unless otherwise indicated, the drawings are intended to be read together with the specification, and are to be considered a portion of the entire written description of this invention as required by 35 U.S.C. §112. As used herein, the terms “horizontal,” “vertical,” “left,” “right,” “up,” “down,” as well as adjectival and adverbial derivatives thereof, refer to the relative orientation of the illustrated structure as the particular drawing figure faces the reader.

Conventional tap-switching voltage stabilizers perform switching only on the secondary side of the transformer: that is, the number of turns to which the input voltage is applied is not varied. Hence, the maximum number of steps achievable with this technology equals the number of relays used to perform the switching, and is one less than the number of taps. In essence, the available voltage step-up or step-down ratios are N

1

/N

x

, N

2

/N

x

, N

3

/N

x

, . . . , N

y

/N

x

where N

1

, N

2

, . . . , N

y

are the numbers of turns in each secondary winding (that is, between each of taps

1

through y and neutral) and N

x

is the number of turns in the primary winding between the tap used as the hot input and neutral. The maximum number of available ratios thus equals the number of taps.

If, however, switching could be performed on both the primary and the secondary sides of the transformer, the number of available step-up and step-down ratios would be much greater. These ratios would then be given by a rectangular matrix such as that of Table 1. Excluding all but one of the “unity” ratios (that is, ratios N

x

/N

y

where x=y), the maximum number of achievable ratios would now be given by N

max

=(m*n−p+1) where m is the number of primary taps, n the number of secondary taps, and p the number of taps which are common to both sets. For example, for m=n=p=5, there are 21 possible ratios. Sample multitap switching step-up/step-down ratios for this configuration are shown in Table 1.

TABLE 1

Sample multitap switching step-up/step-down ratios for

N

1

= 100, N

2

= 120, N

3

= 150, N

4

= 180, N

5

= 220.

# Turns

# Turns

(secondary winding)

(primary winding)

100

130

160

180

200

100

1.00

1.30

1.60

1.80

2.00

130

0.77

1.00

1.23

1.38

1.54

160

0.63

0.81

1.00

1.13

1.25

180

0.56

0.72

0.89

1.00

1.11

200

0.50

0.65

0.80

0.90

1.00

The resulting step-up and step-down ratios in ascending order are: 0.50, 0.56, 0.63, 0.65, 0.72, 0.77, 0.80, 0.81, 0.89, 0.90, 1.00 (five times), 1.11, 1.13, 1.23, 1.25, 1.30, 1.38, 1.54, 1.60, 1.80, 2.00. These ratios are very unevenly spaced, having successive ratios between them as low as 1.0125 (from 0.80 to 0.81) or as high as 1.12 (0.50 to 0.56). As a result, some steps would be smaller than others, so that the transformer's voltage-correcting ability would be inefficiently used.

In order to take maximum advantage of this scheme, therefore, it would be necessary to determine those values of m, n, and N

1

−N

y

which provide a series of step-up and step-down ratios which are themselves spaced with approximately the same ratio between successive ones. Surprisingly, this problem can be solved by placing taps at two overlapping sets of logarithmic intervals along the transformer winding: one set placed at a basic narrow spacing ratio R

n

and the other at a wider spacing ratio Rw, so that combinations of the two sets form ratios at equal logarithmic spacing, covering the region of interest with no gaps wider than R

n

and with a minimum number of ratios repeated. Any individual tap may thus belong to the “narrow” set, the “wide” set, or both sets at once. Using one set as the inputs and the other as the outputs then gives the desired succession of ratios.

For example, to provide output voltage stabilization with a nominal accuracy of ±2%, one set of taps is placed at a narrow spacing ratio R

n

of 1.04, while another is placed at a wider spacing R

w

=R

n

k

, where k is an integer. As will be explained further below, it is convenient to make this k=4, so that R

w

=R

n

4

.R

w

is thus ideally 1.16986 . . . , or very nearly 1.17. Since taps must be separated by integral (or sometimes half-integral) numbers of transformer turns, however, neither of these turns ratios is precisely achievable in reality, and the successive ratios merely approximate these ideal values. If the tap nearest neutral is separated from it by “K” turns, it is convenient to label this tap “R

n

0

” and the other taps “R

n

p

”, where “p” represents the distance from the first tap expressed in logarithmic intervals of R

n

and rounded to the nearest whole (or half) turn. An example of such a set of tap spacings (expressed as turns numbers) and the resulting step-up and step-down ratios is given in Table 2.

TABLE 2

Sample multitap switching step-up/step-down ratios:

narrow spacing R

n

= 1.04; wide spacing R

w

= R

n

4

= 1.17

# Turns

# Turns

(secondary winding)

(primary winding)

100

104

108

112

73

1.37

1.42

1.48

1.53

85

1.18

1.22

1.27

1.32

100

1.00

1.04

1.08

1.12

117

0.85

0.89

0.92

0.96

Tap labeling, in R

n

p

notation: K=73, R

n

=1.04, 73=R

n

0

, 85=R

n

4

, 100=R

n

8

, 104=R

n

9

, 108=R

n

10

, 112=R

n

11

, 117=R

n

12

. The resulting ratios in ascending order are: 0.85, 0.89, 0.92, 0.96, 1.00 (once), 1.04, 1.08, 1.12, 1.18, 1.22, 1.27, 1.32, 1.37, 1.42, 1.48, 1.53.

The ratios between successive pairs of these values all lie between 1.029 and 1.047, with the errors largely balancing out as wider intervals are considered. Larger numbers of turns in the windings would allow these errors to be reduced still further.

For purposes of AC line voltage correction, it makes little difference which set of taps is used on the primary side and which is used on the secondary side. However, since lower-voltage analog and digital electronics are used in relay control, there is some advantage in spacing the primary taps at R

w

and the secondary taps at R

n

so that less overall voltage variation occurs across the primary winding from one end of the stabilization range to the other. One or more auxiliary taps placed between R

n

0

and neutral, optionally supplemented by another one or more taps placed on an extension of the windings beyond neutral so as to have an opposite phase relationship, can then supply low-voltage AC for rectification and regulation to a constant DC voltage using simple, conventional means.

Referring now to

FIG. 4A

, there is shown a voltage stabilizer

90

according to the present invention. Voltage stabilizer 90 includes a premultitapped autotransformer

100

made according to the above-described approach, along with other components of the stabilizer. A tap

102

is neutral; primary taps

104

a,

104

b

and

104

c

are denoted by R

n

0

, R

n

4

and R

n

12

, respectively, in the nomenclature set forth above; secondary taps

106

a,

106

b

and

106

c

are denoted by R

n

9

, R

n

10

and R

n

11

; a tap

108

can serve either as a primary or as a secondary tap, and is denoted by R

n

8

; and taps

110

a

and

110

b

are low-voltage AC taps having opposite phase relationships and used to power the control circuits

24

through a rectifier and regulator

112

.

A primary tap-selecting device

120

a

is connected to AC hot input

12

a

and connects it in turn to any one of primary taps

104

a,

104

b,

108

or

104

c.

Similarly, a secondary tap-selecting device

120

b

is connected to output terminal

22

and connects it in turn to any one of secondary taps

108

,

106

a,

106

b

or

106

c.

The interaction of the two tap-selecting devices

120

a,

120

b

thus permits any step-up or step-down ratio between R

n

−4

(tap

104

c

primary, tap

108

secondary) and R

n

11

(tap

104

a

primary, tap

106

c

secondary), inclusive, to be selected. In other words, the circuit as shown offers the selection of eleven different step-up ratios for use during voltage sags or brownouts, straight-through operation for periods of near-nominal line voltage, and four step-down ratios for use during voltage surges, as shown in Table 3. The resulting regulation tolerance and range of acceptable input voltages are determined chiefly by the value chosen for R

n

in designing the transformer.

TABLE 3

Result of multitap switching in the circuit of FIG. 4:

step-up/step-down ratios with R

n

= 1.04.

Selected Secondary Tap

Selected Primary Tap

R

n

8

R

n

9

R

n

10

R

n

11

R

n

12

0.85

0.89

0.92

0.96

R

n

8

1.00

1.04

1.08

1.12

R

n

4

1.18

1.22

1.27

1.32

R

n

0

1.37

1.42

1.48

1.53

The output voltage from terminal

22

is also fed to control circuit

24

. A circuit

24

usable with the invention is shown in block-diagram form in

FIG. 5

(it should be understood that this is an example only; other circuits, both analog and digital, may also be suitable for use with the invention). First, a divider circuit

140

reduces the raw AC voltage at terminal

22

to a level more easily handled by analog integrated circuits such as operational amplifiers. This voltage is then processed by a rectifier

142

and a low-pass filter

144

. Either of blocks

142

and

144

may be of any convenient type, and may have either a linear or nonlinear response of any convenient form so long as the output from filter

144

has a unique DC value for each different input AC voltage and the relationship between the two is monotonic over the full range of interest. These functions are conveniently carried out using neutral line

14

as common or signal ground.

A change in the AC input voltage does not cause an instantaneous change in the DC output of filter

144

. Rather, at least one full power cycle must be averaged. Where non-periodic disturbances such as lightning spikes or switching transients are present, a longer averaging period may be desirable. Care should be taken in the design of rectifier

142

and filter

144

to minimize the resulting delay.

The output of filter

144

is then compared with a pair of DC reference voltages by comparators

146

a

and

146

b.

The reference voltages may be supplied by a voltage divider

148

or other suitable device. As illustrated in

FIG. 4B

, these voltages represent upper and lower limits

60

a

and

60

b

between which it is desired to stabilize the output voltage, close to optimal value

36

.

The results of the comparison cause a digital emulation of a motor's action, incrementing or decrementing the count in an up/down counter

150

. Counter

150

is preferably binary, for example, a CD4029-type, 4-bit CMOS up/down counter, which can hold counts ranging from 0000 to 1111 inclusive. However, other counters may also be suitable.

A supplementary “stop” logic

152

prevents the count from ascending beyond a set maximum value or descending beyond a set minimum value so that no abrupt rollover from maximum to minimum counts occurs. For a CD4029 counter, for example, these values may be set at 1111 and 0000, respectively. Logic

152

monitors binary outputs

154

of counter

150

and ensures that with a below-optimum input voltage, the count increases steadily until the input voltage becomes optimum or until the maximum permitted count has been reached. Similarly, with an above-optimum input voltage, the count decreases steadily until the input voltage becomes optimum or until the minimum permitted count has been reached. When the voltage becomes optimum—that is, between the two limits represented by voltages

60

a

and

60

b

—no counting occurs and the counter maintains whatever value it holds.

As noted above, the wider tap-ratio spacing R

w

is made equal to R

n

raised to some integral power, and this power is conveniently made equal to four so that R

w

=R

n

4

. The advantage of this is that the last two (least significant) bits of the binary count can now select one of four secondary-winding taps separated successively by R

n

, while the higher-order bits can select from a number of primary-winding taps separated successively by R

w

=R

n

4

. This number is also conveniently made four, so that a total of four bits, as provided by the CD4029 or a similar device, suffices for control with sixteen available voltage ratios according to the scheme shown in Table 3. For example, if primary taps are set at R

n

0

, R

n

4

, R

n

8

, and R

n

12

and secondary taps at R

n

8

, R

n

9

, R

n

10

and R

n

11

, the binary outputs from counter

150

may select among them for a total of sixteen step-up or step-down ratios as listed in Table 4.

An advantage of this arrangement is that it permits simple selection among four taps using only two relays, as shown in FIG.

6

: a single-pole, double-throw (SPDT or “form C”) relay

20

a

and a double-pole, double-throw (DPDT or “2 form C”) relay

20

b.

Relay

20

a

is controlled by the more-significant bit, and relay

20

b

by the less-significant bit, of two successive outputs from counter

150

.

Alternatively, another number of bits may be used either for the primary or the secondary tap selection, or a non-binary type counter (for example, a decimal or binary-coded-decimal counter) may be used as counter

150

, with some appropriate tap selection scheme.

The rate at which the count rises or falls is set by a free-running clock

156

, which may be of any convenient type and which, once during each clock cycle, permits counter

150

to count up or down if this is warranted by the other conditions just described. Clock

156

is set to the fastest speed which, combined with the response time of filter

144

, results in stable operation without oscillation or overshoot.

TABLE 4

Binary Control of Tap Selection and Step-Up/Step-Down Ratios.

Binary

Primary

Secondary

Ratio

Ratio with

count

tap sel.

tap sel.

in R

n

p

R

n

= 1.04

0000

R

n

12

R

n

8

R

n

−4

0.85

0001

R

n

12

R

n

8

R

n

−3

0.89

0010

R

n

12

R

n

8

R

n

−2

0.92

0011

R

n

12

R

n

8

R

n

−1

0.96

0100

R

n

8

R

n

9

R

n

0

1.00

0101

R

n

8

R

n

9

R

n

1

1.04

0110

R

n

8

R

n

9

R

n

2

1.08

0111

R

n

8

R

n

9

R

n

3

1.12

1000

R

n

4

R

n

10

R

n

4

1.18

1001

R

n

4

R

n

10

R

n

5

1.22

1010

R

n

4

R

n

10

R

n

6

1.27

1011

R

n

4

R

n

10

R

n

7

1.32

1100

R

n

0

R

n

11

R

n

8

1.37

1101

R

n

0

R

n

11

R

n

9

1.42

1110

R

n

0

R

n

11

R

n

10

1.48

1111

R

n

0

R

n

11

R

n

11

1.53

Outputs

154

then drive relays or other devices which make up tap-selecting devices

120

a

and

120

b.

For example, each of devices

120

a

and

120

b

may be made up of relays as illustrated in FIG.

6

. Often, such a device requires more voltage or current than is available directly from the output of a logic device such as counter

120

. Hence, these devices are usually driven through buffer amplifiers

158

, which may be of any convenient type.

As will be described below, an arc suppression or voltage limiting device is placed across devices

120

a

and

120

b

to minimize arcing when switching inductive loads. For example, metal-oxide varistors (MOVs) can be connected across the contacts of each relay

20

a,

20

b,

or between the input and each of the four output terminals of the relay cluster shown in FIG.

6

. The MOVs would not only prolong overall relay life, but would also substantially eliminate the possibility of slowed response due to localized welding of the contacts.

Alternatively, devices

120

a

and

120

b

may be replaced by suitable networks of solid-state switching devices, such as silicon controlled rectifiers or triacs. Some additional control circuitry may then be needed to ensure that switching takes place in a “break-before-make” fashion with only one of the four possible connections being made at any given instant. Otherwise, portions of the transformer windings may be shorted out, with potentially catastrophic effects.

The circuit shown in

FIGS. 4A and 5

does not incorporate any form of control beyond the transformer step-up or step-down range. If a too-high or too-low voltage is input to the circuit, for which adjustment to acceptable output voltages is not possible, the circuit merely goes to its lowest or highest ratio respectively. The result is shown in FIG.

4

B: a line segment

70

a

representing circuit performance at too-low voltages goes right down to zero, while a segment

70

b

representing circuit performance at too-high voltages goes up without limit. Operation at too-high or too-low voltages, as was previously noted, can damage or destroy equipment, in this case, the equipment being powered by the stabilizer.

As an added safety feature, as partly shown in

FIG. 7A

, an additional relay or other switching device

160

is preferably connected between output

22

of tap-selecting device

120

b

(where the feedback voltage is measured) and a final output terminal

162

. Device

160

is driven by an auxiliary set of comparators

164

a

and

164

b,

a voltage divider or other source of reference voltages

166

, a control circuit

168

and a buffer amplifier

170

, sensing under- or overvoltages at output

22

and disabling output

162

if an acceptable output voltage cannot be produced by tap switching. To prevent unwanted cycling under load, at least lower comparator

164

b

preferably includes hysteresis. This safety feature results in the plot of input versus output voltages shown in

FIG. 7B

; at excessively low or high input voltages, the output is disconnected and falls to zero. Hence, out-of-range performance as represented by line segments

70

a

and

70

b

is limited to safe output voltages.

These functions may be implemented by other suitable techniques. For example, rectifier

142

may be replaced with an RMS (root-mean-square) extracting integrated circuit such as the AD535 or AD536, or blocks

142

and

144

together may be replaced with a peak-detecting circuit having exponential reset to zero with a time constant which is about an order of magnitude longer than the power cycle. Alternatively, the processed voltage may be digitized at some point and some or all of the described functions performed by a microprocessor, digital signal processor (DSP) or similar device. Similarly, any of a wide variety of different analog and digital integrated circuits, discrete-component technologies and combinations of these may be used to implement the functions of blocks

146

a

and

146

b,

148

,

150

,

152

,

156

,

164

a

and

164

b,

166

,

168

and

170

.

Referring now to

FIGS. 8-12

, there is illustrated a preferred embodiment of the present invention using, as nearly as possible, the numbering scheme used in the previous Figures. Depending upon context, the same identifying number may apply to a signal or to the line, node or connected system of lines and nodes on which it appears.

A voltage stabilizer according to the invention includes a tap selecting circuit

180

with a transformer

10

, which is made and functions according to the principles described above and illustrated in

FIG. 4

, using R

n

=1.04 and Rw=R

n

4

. As shown in

FIG. 8

, tap selection is made through two single-pole, double-throw relays

20

a

and

20

d

and two double-pole, double-throw relays

20

b

and

20

c.

A fifth, single-pole, single-throw relay

160

provides the safety disconnect feature to output

162

. Low-voltage DC power for the control circuitry, referenced to neutral line

14

, is provided to a positive rail

190

through diodes

192

a

and

192

b,

a capacitor

194

, and a voltage regulator

196

. Diodes

192

a

and

192

b

may be 1N4001's, capacitor

194

an aluminum electrolytic capacitor of at least 3300 μf, and regulator

196

an LM7812 or equivalent, 12-volt positive monolithic regulator. However, other components may also be suitable for use with the invention.

Input AC voltage, after passage through a switch and circuit breaker (not shown) is applied to node

12

at the input of relay

20

a.

Voltage is sampled from node

22

at the output of relay

20

d,

and reduced, full-wave rectified and averaged by a circuit

182

shown schematically in

FIG. 9

, which corresponds to blocks

140

,

142

and

144

of FIG.

5

. For example, all resistors shown in

FIG. 9

may be 270 KΩ, 5% tolerance, ¼W types except that resistor

200

a

may be a 50 KΩ, 10-turn trimpot and resistor

200

b

chosen so that its value and that of resistor

200

a

total about 6 KΩ per volt at the intended line voltage, resulting in approximately 8 volts RMS across resistor

200

a.

Diodes in this and the following Figures, unless otherwise stated, may be 1N914's or similar devices. Capacitors

202

a

and

202

b

may be 0.1 μf and 0.22 μf, respectively. Operational amplifiers

204

a

and

204

b

may be LM324 or other single-supply types powered by rail

190

. The output

206

is a DC level which is approximately linearly related to the input AC voltage but may be adjusted up or down using resistor

200

a.

As for above-described device

180

, other circuit components may also be suitable.

As shown in

FIG. 10

, DC voltage

206

is next compared with a set of four DC reference levels established by a voltage divider

210

in a voltage comparator circuit

184

. From highest to lowest, these reference levels represent the maximum input voltage for a safe output, the upper and lower regulating levels, and the minimum input voltage for a safe output. A Zener diode

212

, preferably rated around approximately five or six volts, stabilizes the reference voltages even at low AC input voltages when regulator

196

cannot establish a stable 12 volts on rail

190

. For example, voltage divider

210

may be made up of resistors (from top to bottom) having values of 10 KΩ, 10 KΩ, 5.6 KΩ, 27 KΩ and 180 KΩ respectively, and diode

212

may be rated at 5.1 volts. The resulting reference voltages are then about 5.1, 4.9, 4.7 and 4.1 volts respectively.

Comparators

214

a

through

214

d

may be formed by LM324 or other single-supply operational amplifiers. To prevent cycling under load, resistors

216

a

and

216

b

are preferably used at the lowermost limit to provide hysteresis. Resistors

216

a

and

216

b

may, for example, have values of 470 KΩ and 10 MΩ respectively, giving thresholds of 4.4 volts as voltage

206

rises and 3.9 volts as it falls again. Care should be taken that the upper threshold resulting from hysteresis does not overlap the next higher threshold set by divider

210

.

Logic gates

218

a

through

218

d,

which may be CMOS NAND gates such as the CD4011B or other suitable types, interpret the resulting signals. Gates

218

a

and

218

b

generate a logic high output

220

whenever the voltage at node

22

is within a safe range; this drives safety relay

160

through a buffer amplifier

222

a

whose makeup will be described below.

A resistor

230

and a capacitor

232

prevent output

220

from responding to brief deviations outside the safe range, especially the periods of zero voltage during transitions of relays

20

a

through

20

d,

so that relay

160

remains stable during such transitions. The RC time constant of these two components together is preferably on the order of one second or thereabouts. Capacitor

232

is connected to positive rail

190

rather than common/neutral

14

, so that the output also remains low on initial power-up; if this were not so, relay

160

could close before the voltage at node

22

had been stabilized.

An output

240

of comparator

214

b

is applied to the U/D input of a CD4029B or equivalent, up/down counter

150

to determine the count direction. Output

240

is also combined with all four binary outputs Q

1

through Q

4

from counter

150

through AND and OR gates formed by resistors

242

a

and

242

b

and the diodes connected to lines

224

and

226

respectively. (Alternatively, conventional CMOS or other compatible type logic gates could perform the AND and OR functions.) Resistors

242

a

and

242

b

are preferably each in the range between 22 KΩ and 100 KΩ, inclusive; however, values outside this range may be useful depending on the selection of the other components of the apparatus. Resistors

242

a,

242

b

function to make line

224

logic high only when signal

240

and all of counter

150

's binary outputs are also at logic high, and logic low otherwise. Similarly, a line

226

will be logic low only when signal

240

and all the binary outputs are at logic low, and otherwise at logic high.

The outputs of comparators

214

b

and

214

c

are applied to logic gate

218

c

to generate an output signal

242

which is at logic high when the AC voltage at node

22

is outside the regulation range. Gate

218

d

performs a NAND operation between output signal

242

and a signal

226

so that when both inputs are high, an output

244

is at logic low. A diode

246

and a resistor

248

perform a further AND operation between signals

244

and

224

, so that the resulting signal

250

is high only when signals

224

and

244

are both high. (For reliability, resistor

248

is preferably about ten times resistor

242

a.

Again, the same function could also be performed by conventional CMOS or other compatible type logic gates.) Signal

250

is then applied to the “Count In” (C

in

, active low) input of counter

150

, so that counting either up or down is enabled under the conditions set forth in Table 5.

TABLE 5

Control of Counter 150 by Comparator Inputs and

Counter 150's Own Binary Outputs.

Kh

Kl

T1

T2

T3

C

in

Binary

Signal

214c

Sig.

Sig.

Sig.

Sig.

Count

Count

240

Out

224

226

242

250

Enabled?

0000

0 (down)

1

0

0

1

1

No

xxxx

0 (down)

1

0

1

1

0

Yes

1111

0 (down)

1

0

1

1

0

Yes

0000

1 (up)

1

0

1

0

1

No

xxxx

1 (up)

1

0

1

0

1

No

1111

1 (up)

1

1

1

0

1

No

0000

1 (up)

0

0

1

1

0

Yes

xxxx

1 (up)

0

0

1

1

0

Yes

1111

1 (up)

0

1

1

1

1

No

(Comparators

214

b

and

214

c

never give “zero” outputs at the same time. “xxxx” indicates any four-bit combination other than 0000 or 1111.)

Thus, counting is enabled except when either signal

240

and Q

0

through Q

3

are all zeros (preventing rollover past 0000 when counting downward), the same signals are all ones (preventing rollover past 1111 when counting upward), or the input voltage is between the regulation limits (so that an acceptable voltage is maintained). The circuit described above is only one of many possible ways of achieving this result, other circuits may also be suitable for use with the invention.

When enabled, counting is triggered by the rising edge (if counter

150

is a CD4029B or equivalent) of a periodic rectangular pulse or square wave generated by a clock circuit

156

. This may be formed from spare logic gates or a spare operational amplifier in any of several well-known ways. For example, the operational-amplifier circuit shown in block

156

may be used. The output frequency is not especially critical, but is made as high as possible without introducing instability or overshoot. For example, a frequency in the range between 15-20 Hz may be useful. To achieve this with the circuit shown, for example, resistor

260

may be about 100 KΩ, capacitor

262

may be 0.47 μf, and the other resistors in the block may all be equal with any convenient value in the range from 100 KΩ to 1 MΩ, such as 270 KΩ.

Alternatively, resistor

260

may be made adjustable—for example, in the circuit shown resistor

260

may be a 200 KΩ or 250 KΩ trimpot—so that the count rate may be changed to accommodate the characteristics of varying types of switching devices. This feature would be especially useful in case the control circuitry were being offered by itself with the user providing the transformer and switching devices.

Counter

150

, if a CD4029B or similar device, also has its preset inputs P

0

through P

3

so connected that, if input PE (“preset enable”) is brought to logic high, Q

0

through Q

3

will take on the values of P

0

through P

3

respectively. This is done upon power-up (or after loss of power) through a resistor

270

and a capacitor

272

, so that the step-up voltage ratio is brought up slowly from a minimum value rather than beginning at a higher and possibly dangerous level. Resistor

270

and capacitor

272

may, for example, have values of 820 KΩ and 0.1 μf, respectively.

Since false triggering of PE may occur upon relay transitions, another capacitor

274

, having two or more times the capacitance of capacitor

272

, is preferably added as close as possible to resistor

270

and capacitor

272

to bypass the power supply locally. Additional bypassing is desirably added by placing one or more other capacitors (not shown), totaling a few microfarads, between rails

14

and

190

wherever circuit layout makes this convenient.

Each of buffer amplifiers

222

a

through

222

e

needs to be designed for the specific relay type being used, as in a circuit

186

shown in FIG.

11

. The logic input signal is applied through a resistor

280

to the base of a transistor

282

, whose emitter is connected to common rail

14

. The collector is then connected to a relay coil

284

, which is bypassed by a free-wheeling diode

286

(preferably a diode that is able to handle the maximum expected voltage and current). Relays are not driven from positive rail

190

, and coil current does not pass through regulator

196

, as this could create disturbances in the DC power which might upset other circuit functions. Instead, coil

284

is connected directly to a positive terminal

288

of capacitor

194

.

For example, OMI-SS-212D relays may be used in circuit

186

. These are DPDT relays with contacts rated at 250 volts, 5 amperes, and with 12-volt, 200-ohm coils; thus, the driving current for each relay is about 60 milliamperes. To drive these relays, resistor 280 is 10 KΩ, transistor 282 a 2N2222 type transistor, and diode

286

a 1N4001 diode.

A light-emitting diode (not shown) may optionally be placed as an indicator, either in series with resistor

280

, or in series with a second resistor of lower value, the pair placed in parallel with coil

284

and diode

286

so that the LED glows when the coil is energized.

Where the control circuitry is meant to interface with a user's transformer and tap-selecting devices, the output may be of any type widely used in industry, such as a contact closure, current-loop output or open-collector output. Optionally, several such outputs may be provided for each relay, providing greater ease of use.

To prolong the lives of the relays or other switching devices, arc-suppression circuitry may be added between each pair of points involved in switching to absorb the energy stored in an inductive load, which would otherwise cause arcing between the contacts on opening. For example, a separate metal-oxide varistor (MOV) can be placed between the common and normally-open contacts of each relay used, and another between the common and normally-closed contacts. This approach, however, uses a great many MOVs, each of which is used only when its own particular set of contacts opens. Furthermore, since for any configuration of the contacts there are two relays connected in series on each side of the transformer, and since because of slight variations between the relays in such a pair one invariably opens slightly faster than the other, only the first-opening contact pair arcs. For arc suppression, therefore, the cluster containing the pair may be treated as a single, four-position switch or relay.

A second configuration of MOVs, therefore, might have one MOV placed between the common node of each tap-selecting device (as represented generally by multiposition switch

120

a

or

120

b

in

FIG. 4

) and each of the other nodes to which this MOV may be connected. For four-position tap-selecting devices, one located on the primary and one on the secondary side of the transformer as shown in the examples, this would reduce the number of MOVs needed to eight.

A further reduction is possible if a network of steering diodes is used so as to select the most positive and the most negative voltages present at any of these nodes. A single MOV may then be placed across the diode network, so as to prevent the maximum voltage between any pair of these nodes from exceeding a set limit. Such diode network

188

is shown in

FIG. 12

, where the transformer, switching arrangement and numbering system of

FIG. 4

are repeated for clarity, but the arc suppression circuit

188

is shown separately. Nodes

12

and

22

, representing the common nodes of the two tap-selecting devices, and nodes

104

a,

104

b,

104

c,

106

a,

106

b,

106

c,

and

108

, representing the possible nodes of connection, each have two diodes added, one diode with its anode facing the node and the other with its cathode facing it. The opposite ends of these diodes are connected together at two additional nodes

290

and

292

: node

290

receiving the cathode ends, and node

292

the anode ends.

Node

290

thus takes on the most positive voltage, and node

292

the most negative voltage, present at any of these nodes. By connecting a suitably-chosen MOV

294

between these nodes, therefore, the entire set of relay contacts may be protected simultaneously.

The diodes used may be taken from the 1N400X series, where “X” is a digit representing the peak inverse voltage of the diode; such diodes, while rated for one ampere of continuous current, can withstand brief forward spikes of 50 amperes or more. For the 230-volt stabilizer used in the examples, with R

n

=1.04 and R

w

=1.17, diodes with PIV=200 volts or more (e.g., 1N4003's or 1N4004's) and an MOV with a breakdown voltage of 150 volts or more provide adequate protection. Heat dissipation in the MOV is minimal since it receives inductive spikes only when the relay contacts change position, at most 15 to 20 times a second.

To minimize radio-frequency interference which might be generated during relay transitions and passed on to the load, a low-pass resonant filter is preferably added at the AC output. This may, for example, take the form of a resonant filter

192

shown in

FIG. 13

in which two identical capacitors

300

a

and

300

b

and an inductor

302

form a simple “pi” network. For example, capacitors

300

a

and

300

b

may be 0.02 μf and 145 microhenries, respectively, for an overall resonant cutoff frequency of about 130 KHz.

Spike suppression is also conveniently added here by placing a metal-oxide varistor (MOV)

304

between hot output line

162

and neutral line

14

. Optionally, one or more additional MOVs such as

308

may be placed between hot line

162

and either neutral line

14

, ground line

306

, or both.

The description above is given merely as an example of the invention, and should not be construed as limiting the scope of the invention. Many modifications to this basic circuit or to its constituent parts, performing the general functions described above, should now be apparent to any person of usual skill in the art of analog, digital and power electronic circuit design, and are therefore included within the scope of the invention.

With respect to the above description of the invention, it is to be realized that the optimum dimensional relationships for the parts of the invention, to include variations in size, materials, shape, form, function and manner of operation, assembly and use, are deemed readily apparent and obvious to one skilled in the art, and all equivalent relationships to those illustrated in the drawings and described in the specification are intended to be encompassed by the present invention.

Therefore, the foregoing description is considered as illustrative only of the principles of the invention. Further, since numerous modifications and changes will readily occur to those skilled in the art, it is not desired to limit the invention to the exact construction and operation shown and described, and accordingly, all suitable modifications and equivalents may be resorted to, falling within the scope of the invention. Thus, it will be apparent to those skilled in the art that many changes and substitutions can be made to the preferred embodiment herein described without departing from the spirit and scope of the present invention as defined by the appended claims.

Number	Name	Date	Kind
3818402	Golaski et al.	Jun 1974	A
5825164	Williams	Oct 1998	A
5900764	Imam et al.	May 1999	A

Digitally-controlled AC voltage stabilizer

Information

Patent Number

Date Filed

Date Issued

Inventors

Examiners

CPC

US Classifications

Field of Search

US

International Classifications

Abstract

Description

Claims

Parent Case Info

US Referenced Citations (3)

Provisional Applications (1)