US3379960A

US3379960A - Automatic voltage regulator with buck-boost transformers

Info

Publication number: US3379960A
Application number: US521723A
Authority: US
Inventors: Joseph C May
Original assignee: Superior Electric Co
Current assignee: Superior Electric Co
Priority date: 1966-01-19
Filing date: 1966-01-19
Publication date: 1968-04-23
Anticipated expiration: 1985-04-23

Description

J. C. MAY

Aprili 23, R968 AUTOMATIC VOLTAGE REGULATOR WITH BUCK-BOOST TRANSFORMERS Filed Jan. 19, 1966 ,QTTOENEYS United States Patent Otiice 3,379,960 Patented Apr. 23, 1968 AUTOMATIC VOLTAG REGULATOR WITH BUCK-BOST TRANSFORMERS Joseph C. May, Cheshire, Conn., assigner to The Superior Electric Company, Bristol, Conn., a corporation of Connecticut Filed Jan. 19, 1966, Ser. No. 521,723

9 Claims. (Cl. 323-45) The present invention relates to an automatic voltage regulator for maintaining substantially constant an A.C. output voltage.

In some regulators which are connected to an A.C. source to provide an output voltage of substantially constant value, the value of the input voltage may be altered to the desired value of output voltage by algebraically adding a corrective voltage of the proper value to the input voltage. The corrective voltage may be added by the use of a single unit which is capable of adding the maximum value of corrective voltage but is controlled to provide only the required value. Alternatively, the corrective voltage may be applied in discrete values by having a plurality of different sized devices which are selectively energized. While both of the above ways of adding the corrective voltage have been suggested, the former has not been found completely satisfactory by reason of its relatively slow speed of response and increased introduction of distortion while in the latter method the addition of voltages of suiiciently small discrete values in order to provide a suticiently sensitive regulator has tended to introduce instability in operation. Additionally, the regulators heretofore suggested for adding the corrective voltage to the input voltage have been either/ and relatively complex, susceptible to malfunctioning, possessed :a relatively high internal impedance, slow in response which in many uses of regulators has been found undesirable.

It is accordingly an object of the present invention to provide an automatic voltage regulator which is rapid in response, yet is quite sensitive to correcting small changes without tending to be unstable and which inherently introduces very little distortion in the output voltage.

Another object of the present invention is to provide an automatic voltage regulator which has a low internal impedance, is eicient in operation and which is relatively simple in construction yet durable in use.

A further object of the present invention is to utilize at least one magnetic structure for the algebraic addition of a corrective voltage, which is simple in construction, relatively economical to manufacture and which inherently isolates the remainder of the regulator including the control circuit from abnormal conditions occurring outside the regulator.

ln carrying out the present invention, a feature resides in a plurality of separate units that are serially connected between the input and output terminals of the regulator. Each unit is capable of being either passive or activated to cause the algebraic addition of a definite value of corrective voltage to the input voltage to provide the desired value o-f output voltage. Each of the units includes a pair of magnetic devices with one device upon being actuated adding a known quantity of voltage while the other device upon being actuated subtracting the same value of the voltage. If neither is actuated then no voltage is added to the input voltage. While any desired number of units may be employed, depending for example on the maximum value of the corrective voltage, in the specific embodiment herein disclosed three units are provided. The units vary in the amount of corrective voltage which each may supply and the capability of each is made to be related to each other in proportion of the powers of two. Thus the smallest unit is capable of adding or substracting a corrective voltage of plus or minus one integer of value, another unit is capable of supplying a corrective voltage of plus or minus two integers of value, the third unit is capable of supplying a corrective voltage of plus or minus four integers. In addition to the above described units that are exponentially interrelated, there is a fractional unit which is capable of adding or subtracting a corrective voltage that is the same `as the smallest unit, namely one integer of value.

By the present invention, each of the exponential units except the fractional unit is actuated in a manner which individually renders each to produce its maximum capable corrective voltage. The fractional unit on the other hand is continuously adjusted to add a fraction of one integer of value of corrective voltage. The exponential units thus supply the large quantities of corrective voltage by their being added digitally while a small value of corrective voltage less than an integer of value is added by adjusting the fractional unit.

A further feature of the present invention resides in the structure of each of the units. While each may vary in size by reason of the quantity of the corrective voltage it is capable of supplying, in the present invention each of the devices is a saturable transformer having a primary winding, secondary winding and a control Winding.

Other features and advantages will hereinafter appear.

The sole figure is a block and schematic diagram of the components of the :automatic voltage regulator of the present invention.

Referring to the drawing, the automatic voltage regulator is generally indicated by the reference numeral 10 and includes a pair of input terminals 11 and 12 which are adapted to be connected to a source of alternating current. At a pair of output terminals 13 and 14, an A.C. voltage of substantially constant value appears after a corrective voltage has been algebraically added to the input voltage. The corrective voltage is sufficient to not only overcome the deviation in value of the input voltage from the desired value of output voltage but also may include a relatively constant amount to overcome the small losses in the regulator.

For adding the quantities of corrective voltage there is provided a plurality of

units

15, 16, 17 and 18. Each includes a magnetic device 15a, 16a, 17a and 18a respectively which when actuated provides a corrective voltage which is in phase with the input voltage and thus adds its value to the value of input voltage and a similiar magnetic device 15b, 16h, 1711 and 18h respectively which when actuated provides a corrective voltage which is out of phase with the input voltage and hence subtracts its value from the value of the input voltage.

Referring specifically to the device 15a, it has a pair 0f

secondary windings

15e and 15d which are serially connected to the input terminal 11. Each of the other devices also has a pair of secondary windings and the secondary windings are all connected in series between the input terminal 11 and the output terminal 13. All current ow between the two terminals accordingly passes through ea-ch ofthe secondary windings.

Further provided on the device 15 is a pair of primary windings 15e and 15f which are connected together in series and are connected to the input terminal 12 and the output terminal 13. Each of the other devices has a similar pair of serially connected primary windings. The primary windings of each unit are each connected across the output terminals 13 and 14 and thus are .connected in parallel to each other with the output voltage of the regulator being applied to each pair of primary windings.

The device 15a further includes a control winding 15g and each of the other devices similarly has a control winding. The control winding regulates the flux linkage between the primary and secondary windings of each device and when fully energized saturates the core of the device preventing a voltage from being induced in the secondary windings.

Referring to the unit 15 and assuming the condition when there is no current -owing in the control windings of either device 15a or 15b, it will Ibe understood that current flows through the primary windings of each between the termin-als 12 and 13. This will induce in the secondary windings 15C and 15d a voltage which is added to the input voltage while in the secondary windings of the device 15b a voltage is induced which opposes the input voltage and hence is subtracted therefrom. Both of the devices 15a and 15b are of substantially identical construction thereby having a voltage of identical magnitude induced in their respective secondary windings. Accordingly, as the secondary windings of both devices are connected in series, they nullify each others induced voltage and the total effect of the voltage induced in the secondary windings on the input voltage is substantially zero, causing no change in the value of the input voltage.

If however control current flows in the control winding 15g, the flux linkage between the primary and secondary windings of device 15a is decreased preferably to zero such that there is no induced voltage in the secondary winding of device 15a and thus the input voltage will only have algebraically added to its, the subtractive voltage induced in the secondary winding of device 15b. More over, the voltage subtracted by the device 15b will be about twice as large as that when the device 15a is not saturated by reason of the primary windings 15e and 15j being caused to be only substantially resistive yand thus effectively short-circuited to cause the substantially total voltage between terminals 12 and 13 to be applied across the primary windings of device 15b rather than just one half. If the control winding of device 15b has control current owing therein, then the device 15a would produce a value of voltage of substantially similar magnitude and instead of being subtractive would be Iadditive to the input voltage. Each of the

other units

16, 17 and 18 functions similarly to add or subtract a voltage to the input voltage depending upon the value of corrective voltage needed Ito produce the desired value of output voltage.

As will be hereinafter understood, only three conditions exist for the unit 15, either no current flows in both control windings; or saturating current ows in device 15a and none in 15b; or saturating current flows in 15b and none in 15a. While it is within the scope of the present invention to cause each unit to assume one of the three conditions, in the specific embodiment hereinafter set forth, the

units

15, 16 and 17 have at all times saturating current owing to either add or subtract their maximum set value of corrective volt-age.

While each of the

units

15, 16 and 17 are shown in the diagram as being identical, each varies from the other with respect to the magnitude of the corrective voltage which it may `add to the input voltage with the difference being in accordance with the denominational order of a binary numerical system. The smallest unit 17 may be capable of algebracially adding a value of 1, or for illustrative purposes only to more clearly describe the present invention a value of plus or minus l volt which will effect a total change in the input voltage of 2 volts. The unit 16 adds a value which is twice that of the unit 17, namely plus or minus 2 volts for a 4 volt total change and the unit 15 adds a value which is four times the unit 17, namely plus or minus 4 volts fora total change of 8 volts.

The fractional unit 18 is similar in construction to the unit 17 and is capable of algebraically adding the same value of corrective voltage to the input voltage as the unit 17, namely a maximum plus or minus l volt for 'a total change of 2 volts. As will be more hereinafter appreciated, the fractional unit 18 however rather than being essentially an On-oi device is operated as an amplifier to -usually add only a fraction of its maximum capable corrective voltage -to the input voltage. Also it differs from the operation of the

units

15, 16 and 17 in that it usually has control current flowing in both its adding and subtracting devices 18a 4and 18b respectively simultaneously rather than just in one device or the other.

In the operation of the regulator when the input voltage equals the desired value of output voltage, except for losses in the regulator, then control current flows in the

units

15, 16 and 17 to either add or subtract a voltage that is oppositely opposed by a voltage from the unit 18. Continuinfy the assumption that the common multiple is l volt, for deviations of the output voltage requiring :a Icorrective voltage after the addition of the voltage from the

units

15, 16 and 17 of not more than plus or minus 1 volt, the only change will be in the value of control current flowing in the fractional unit 18. If the corrective voltage needed is adding or boost, then the device 1811 has a value of current flowing in its control winding that partially saturates its core enough to decrease its subtraction voltage and increase its additive voltage to provide the needed corrective voltage value and it will add or subtract the necessary fraction of the common multiple voltage. When sa `greater value of corrective voltage is required, then initially the fractional unit 18 will add its maximum 1 volt and if it is insufficient then the

units

15, 16 and 17 change to algebraically add greater increments of voltage. If the needed voltage is greater than l volt, the

units

15, 16 and 17 change to add 3 volts which if still insufficient are subsequently changed to add 5 volts and then if still insufficient their maximum of 7 volts is added. The fractional unit 13 continues to supply the required quantity of fractional voltage between the integers of voltage added by the

units

17, 16 and 15. It will be appreciated that the maximum corrective voltage that can be added is the sum of l for the unit 1S, l for the unit 17, 2 for the unit 16 and 4 for the unit 15 totaling 8 volts which is also the maximum which can be subtracted and hence the regulator would be able to correct over a total range of 16 volts. For producing the integers of corrective voltages, some units are operated to provide a voltage that opposes the voltages produced by the other units. For example for a corrective voltage of plus 5 volts, units 15 land 16 will supply plus 4 and 2 volts while the units 17 will supply a minus l volt. The hereinafter set forth chart shows the manner of selective energization of the devices of the

units

15, 16 and 17 that will provide the necessary integral corrective voltages between values of plus 8 and minus 8 volts.

The flow of control currents in the units 15 through 13 is controlled by a control circuit 19 that generally includes the remaining parts of the regulator. An error detector 2t) is connected across the input terminals 13 and 14 to provide a DC. output signal that is related in magnitude to the deviation of the output voltage from the selected value with the polarity of the signal indicating the direction of the deviation. The detector may be a bridge circuit of the type that includes rectiers :and a Zener diode reference though other well-known detector circuits may be employed.

The signal from the error detector 2t) is introduced into the input of an integrating amplifier 21 which functions -to provide in a lead 22, a substantially pure D.C. signal With a magnitude which has a rate of change depending upon the magnitude of the signal from the error detector with a larger magnitude producing a larger rate of change. The direction of change is determined by the polarity of the error signal. Interconnected across the integrating amplier is an integrating amplifier output limiter 23 which limits the output voltage in the lead 22 to for example plus or minus 10 volts. The limiter 23 is set, as by feed back, to prevent the integrating amplifier from producing its maximum output in the lead 22 and thus cause it to operate in its intermediate range wherein it retains its rapid speed of response.

The D.C. signal in lead 22 is introduced into a driver amplifier 24 which preferably consists of a plurality of stages of D.C. amplifiers that substantially linearly amplify the signal 22. The signal from the driver amplifier 24, if positive in polarity, appears on a lead 25 and if negative in polarity, appears on a lead 26.

Negative voltages on the lead 26 are introduced to a buck linear amplifier 27 which linearly amplifies the voltage in the lead 26 to produce a direct current in the control winding of the device 1811. The lead 25 is connected to a boost linear amplifier 28 which linearly ampiifies the positive voltage on the lead 25 to produce a direct current in the control winding of the device 18a. The driver amplifier 24 additionally supplies a linearly amplified signal on one of a pair Of leads 29 and 30, with the former being connected to a down level detector 31 and the latter being connected to an up level detector 32. Each of the

level detectors

31 and 32 are monostable circuits, such :as Schmitt trigger circuits in which a -given voltage level on their

respective leads

29 and 30 will cause the detectors to change their state and maintain the changed state only While the given voltage level exists. It will be understood that only positive voltages on the lead 22 appear amplified on the lead 29 while a minus voltage on the lead 22 appears as an Vamplified voltage only on the lead 30 and thus only one or the other detector is caused to be operable .at a time.

Each of the

detectors

31 and 32 is designed to change its state when the magnitude of the signal on the lead 22 indicates that there is a need for a value of corrective voltage that is greater than that value 'which the fractional unit `18 is capable of supplying.

iAssuming a condition for a corrective voltage greater than or -1 volt, either the

lead

29 or 30 will be of suflicient value to operate its -associated detector. If it appears on the lead 29 indicating that there should be a decrease in the corrective vol-tage, then the down level indicator changes its state and supplies a signal on a lead `3i3 to a limited range up-down `binary counter 314. The signal on the lead 33 is also transmitted through a lead 33a to an OR gate 35 which in turn transmits a signal through a lead 36 to a pulse generator 37. The up level detector has an output lead 38 Which is also connected to the counter 34 and by a lead 38a to the OR gate '35. Thus irrespective of which detector 311 or 32 changes its state, a signal will appear on the lead 36 to the pulse gen- .rator 37. The pulse generator is an oscillator that produces pulses having Ia selected time interval upon it being actuated and delivers said pulses vby .a lead 39 to the updown binary counter 34.

The counter 34 is in the embodiment herein described at least a three-stage binary counter and with the occurrence of a signal on the lead 33 and a pulse on the lead 39 it changes its state to decrease its binary count one digit while with a signal on the lead 38 and a pulse appearing on the lead 3-9 it changes its state to increase its binary count by one digit. Whenever there simultaneously occurs a pulse on the lead 39 and a signal on either the

leads

37 or 38, the counter will change its binary count. However, once it has attained either of its maximum count states, it will retain said maximum count and will only decrease therefrom. Thus the counter can only count or assume a binary condition representative of a number from 0-7 inclusive and it is prevented from shifting with only one pulse from 7 directly to 0 0r 0 directly to 7, i.e., recycle.

The counter 34 has a plurality of output leads 40, 41, 42, 43, 44 and 45. The lead 40 is connected to a boost switching amplifier 40a which is connected to control current passing through the control winding g of the device 15a. The lead 41 is connected to a buck switching amplifier 41a that controls current to the control winding of the device 15b. The leads 42 and 43 are connected respectively to a buck switching amplifier 42a and a buck switching amplifier 43a, each of which is connected to control conduction of current through control windings of the devices 16a and 16b respectively. The leads 44 and 45 are connected to a boost switching amplifier 44a and a buck switching amplifier 45a which in turn respectively control the current conduction in the control windings of' the devices 17a and 17b.

Each of the switching amplifiers a through 45a upon receiving a signal on its respective lead from the counter 34 produces in its respective control winding a value of current that is sufficient to cause saturation of its associate-d device and thus each of the amplifiers is effectively either completely on to cause saturation or completely off introducing no saturating current.

The count which the counter 34 has determines which of the switching amplifiers is caused to be conducting. Thus with the counter capable of having 8 conditions, namely 0 through 7 inclusive and with the value of the corrective voltage capable of being varied (without now considering the unit 18) from a minus 7 to a plus 7 volts in integral volts and with a 2volt differential between each change, the following chart indicates for each value of corrective voltage the required switching amplifier which must be actuated to produce the proper value of corrective voltage. If a switching amplifier is actuated, saturating its associated device, then an X appears on the chart, While if a switching amplifier is not actuated a 0 appears.

Count State of Switching A mnlitiers Value of Counter 34 Corrective 40a 41a 42a 43a 44a 45a Voltage X O X 0 X O -7 X O X O O X -5 X O O X X O -3 X O O X O X -1 O X X O X O +1 O X X O O X +1 O X O X X O +5 O X O X O X +7 From the chart, for exam-ple when -7 or buck corrective voltage is needed, all boost switching amplifiers 40a, 42a and 44a are actuated to saturate their respective boost devices and enable the bucking device `15b to provide -4 volts, the bucking device 16h to provide -2 volts and the bucking device .1711 to provide -l volt of corrective voltage for a total corrective voltage of -7 volts from the

units

15, 16 and `17.

For a condition of +3 volts of corrective voltage being required, then from the chart it will be understood that the buck switching amplifier 41a is actuated to provide +4 volts of corrective voltage, the boost switching amplitfier `42a is actuated to provide +2 volts and the buck switching amplifier a is actuated to provide +1 volt of corrective voltage with the algebraic sum thereof being +3 volts. It will accordingly be appreciated that the -count of the counter 34 determines the total sum of corrective voltage added by the

units

15, 16 and 17 as shown on the chart.

For even integral values of corrective voltage, the counter 34 assumes a state which supplies either one volt more or one volt less of corrective voltage than the required value of corrective voltage and the fractional unit supplies the other one volt. Thus if +4 volts of corrective voltage is needed, the counter 34 may have a count of 5 producing +3 volts and the Ibuck linear amplifier 28 is energized to saturate the device 18b to provide a +1 volt corrective voltage from the unit 18 which provides a total of +4 volts of corrective voltage. Alternatively, the count may be 6, producing +5 volts and the device 18a of the fractional unit 18 is completely saturated to provide a -1 volt, which when algebraically added produces the total of +4 corrective volts.

`In both of the above conditions it will be understood that the value of the signal in the lead 2.2 is sufiicient to provide a signal in either lead 2'5 or 26 which will effect saturation but as it is not more than the value of the signal caused by 1 volt of deviation, a signal will not appear on either lead 29 or 30 to change the count.

For variations Within the range of the fractional unit 18, i.e., +1 through l volt, the unit i8 will consistently adjust to provide a corrective voltage that is a fractional part of its total range as determined by the value of the signal on either lead or 26. The use of .an integrating amplifier causes the value of the signal on the lead 2.2 and hence on the

leads

25 or 26 to be a constant value when a Zero voltage deviation exists and to change slowly for small deviations, thereby increasing the sensitivity of the regulator. Additionally, the maximum value of the signal on the lead 22 is limited to a value just slightly more than is required -to saturate the fractional unit devices and produce a signal on one of the

leads

29 and 30. Thus the signal will always be at its maximum value whenever the deviation is greater than 1 volt with the fractional unit operating at its maximum capability. This will cause the counter to be operated at a rate determined by the pulse generator 37 until the deviation becomes less than 1 volt from the desired value, at which time the counter will maintain its count.

It will be appreciated with the above structure that the analog or fractional unit 18 has a range of corrective voltage which it may add or subtract which is equal to the range of voltage that can be algebraically produced by a change in count of the counter 34. Thus at all times the analog unit 18 functions to supply the corrective voltage which has a value that is equal to the incremental change between the digital units.

It has been found that by the use of independently operated digital units which are either on or off and only a small fractional unit that these devices may -be quickly operated to provide a fast speed of response. In one embodiment, only one-tenth of a second was required `for maximum correction.

Each of the devices 15a, 15b, 16a, 1615, etc., is essentially a transformer having primary windings and secondary windings. Thecontrol windings of each determines the degree of saturation of the core interconnecting the two windings and hence the induced voltage in the secondary winding. The control winding is however made independent of the currents in the prima1y and secondary windings by the opposing flux of substantially equal magnitude passing through the portion of the core encircled by the control winding, thereby inducing substantially no voltage in the control winding. Specifically in the device 15a, the ux produced by the primary winding 15e traverses the path indicated by the arrow 15h while the flux produced by the primary winding 15f traverses the path indicated by the arrow 151'. Accordingly, the control circuit i9 is effectively decoupled from high voltage or currents which may occur in the regulatOI'.

The use of saturable transformers in place of other magnetic devices enables the regulator to have a low internal impedance which renders them particularly useful when combined to control polyphase A C. Additionally, they have lower power consumption rendering the regulator substantially eiiicient in use. It will of course be understood that other magnetic devices such as a saturable reactor may be employed if desired.

An extremely low distortion in the output voltage of the herein described regulator has been found to occur by reason of having the devices 15a-17h that supply the larger values of corrective voltage operate in either their completely off or saturated condition. At such conditions, they introduce very low distortion. However, though larger distortion is produced between these two maximum conditions, such distortion only occurs in the fractional unit i8 which controls the smallest quantity of corrective voltage and accordingly the production of larger distortion is limited to the smallest corrective voltage, thereby minimizing the total distortion of the regulator.

Variations and modifications may be made within the scope of the claims and portions of the improvements may be used without others.

l claim:

l. An automatic voltage regulator comprising input terminals connectible to a source of A.C.; outp-ut terminals at which a desired value of output voltage is maintained substantially constant; a digital unit for supplying a corrective voltage of a set value and having at least a rst condition wherein its corrective voltage is additive and a second condition wherein its corrective voltage is subtractive; an analog unit for providing an adjustable value of corrective voltage; means serially interconnecting the two units between the input and output terminals; means for providing a deviation signal related to the deviation of the output voltage from a desired value; switch means connected to said digital unit and responsive to a signal for causing said unit to change from one of its conditions to the other; adjusting means connected to the analog unit and connected to receive said deviation signal and to adjust its value of corrective voltage within its maximum corrective voltage range in response to the value of the ydeviation signal; means for supplying said deviation signal continuously to said adjusting means; and means for supplying a signal to the switch means whenever the deviation signal to said adjusting means is greater than the value required to provide substantially the maximum corrective voltage of the analog unit.

Z. The invention as defined in claim 1 in which the analog unit is capable of providing an additive or a subtractive corre-ctive voltage; the deviation signal contains information for deviation in one direction and for deviation in the other direction and the analog unit provides a corrective voltage in the opposite direction of the deviation.

3. The invention as defined in claim 1 in which the value of the set corrective voltage of the digital unit is substantially equal to the value of the maximum corrective voltage of the analog unit.

4. The invention as defined in claim l in which there is a plurality of digital units; a plurality of switch means, there being a switch means connected to each unit; each of said switch means being responsive to a signal for causing its associated digital unit to assume at least either a rst condition wherein it provides a set value of an additive corrective voltage or a second condition in which it provides substantially the same set value of a subtractive corrective voltage; the values of the set corrective voltages of the units being related to each other in accordance with the denominational orders of a binary numerical system.

5. The invention as defined in claim 4 in which the means for supplying a signal includes a limited range Counting means having a plurality of states, an output connected to all of said switch means to provide an individual signal to each, said output changing with each state, and means for changing the count whenever a signal is supplied by the supplying means.

6. The invention as defined in claim l in which the means for providing a deviation signal includes an integrating amplifier and means for limiting the output of the amplifier to a value only slightly greater than the value required to cause a maximum corrective voltage from the analog unit.

'7. The invention as defined in claim l in which each unit includes at least one saturating transformer having a core, a primary winding connected between an input terminal and an output terminal, a secondary winding inductively associated on the core with the primary winding and being serially interconnected between the input and output terminals and a control winding for controlling the saturation of the core.

8. The invention as -defined in claim 1 in which each of the units includes a first saturable transformer and a second saturable transformer; each saturable transformer having a core, a primary winding, a secondary winding inductively associated by the core with the primary winding and a control winding wound to control the saturation of the core that inductively connects the primary and secondary windings; means serially connecting the primary windings of both transformers; and means serially connecting the secondary windings, said transformers having their windings wound oppositely to provide opposite corrective Voltage.

9. The invention as defined in claim 1 in which each unit includes a first device and a second device; said first device being arranged to provide an additive corrective voltage; said second device being arranged to provide a subtractive corrective voltage; a switch means for each device; and in which there are means at all times for only having one of said switch means actuated.

References Cited UNITED STATES PATENTS JOHN F. COUCH, Primary Examiner.

WAR-REN E. RAY, Examiner.

Claims

1. AN AUTOMATIC VOLTAGE REGULATOR COMPRISING INPUT TERMINALS CONNECTIBLE TO A SOURCE OF A.C.; OUTPUT TERMINALS AT WHICH A DESIRED VALUE OF OUTPUT VOLTAGE IS MAINTAINED SUBSTANTIALLY CONSTANT A DIGITAL UNIT FOR SUPPLYING A CORRECTIVE VOLTAGE OF A SET VALUE AND HAVING AT LEAST A FIRST CONDITION WHEREIN ITS CORRECTIVE VOLTAGE IS ADDITIVE AND A SECOND CONDITION WHEREIN ITS CORRECTIVE VOLTAGE IS SUBTRACTIVE; AN ANALOG UNIT FOR PROVIDING AN ADJUSTABLE VALUE OF CORRECTIVE VOLTAGE; MEANS SERIALLY INTERCONNECTING THE TWO UNITS BETWEEN THE INPUT AND OUTPUT TERMINALS; MEANS FOR PROVIDING A DEVIATION SIGNAL RELATED TO THE DEVIATION OF THE OUTPUT VOLTAGE FROM A DESIRED VALUE; SWITCH MEANS CONNECTED TO SAID DIGITAL UNIT AND RESPONSIVE TO A SIGNAL FOR CAUSING SAID UNIT TO CHANGE FROM ONE OF ITS CONDITIONS TO THE OTHER; ADJUSTING MEANS CONNECTED TO THE ANALOG UNIT AND CONNECTED TO