Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS3651414 A
Publication typeGrant
Publication dateMar 21, 1972
Filing dateApr 30, 1970
Priority dateApr 30, 1970
Also published asCA942387A, CA942387A1
Publication numberUS 3651414 A, US 3651414A, US-A-3651414, US3651414 A, US3651414A
InventorsJamieson Robert S
Original AssigneeLorain Prod Corp
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Variable frequency system
US 3651414 A
Abstract
A digital timing system of high precision and stability is provided with a selectively variable frequency and phase by employing a variable frequency dividing circuit that is responsive to a highly accurate precision crystal clock generator. The variable divider includes a fixed modulus counter which in turn triggers a variable modulus counter. High frequency precision clock pulses are fed first to the fixed counter which upon completion of its count, switches the clock pulses to the variable counter. The latter, when it completes its count, provides an output pulse that (a) constitutes one element of the system output, (b) resets the fixed counter, and (c) switches the clock pulses to the fixed counter whereby the above cycle is repeated. To vary the frequency or repetition rate of the described cycle of operation, the magnitude of the count provided by the variable counter is changed to thereby change the total interval or period of a single cycle. To control the phase of the output signal, the count of the fixed counter is changed during but one of its counts. The variable counter is formed of a pair of reversible, up-down counters each of which provides an output upon reaching the count of one when counting down. The frequency of the output is changed by controllably varying the number to which these variable reversible counters count.
Images(11)
Previous page
Next page
Claims  available in
Description  (OCR text may contain errors)

United States Patent J amieson [451 Mar. 21, 1972 [52] U.S. Cl ..328/44, 235/92 EV, 328/130 [51] Int. Cl. ..l-l03k21/30 [58] Field of Search..... ....328/39,41,42,44,129,186,

[561 References Cited UNITED STATES PATENTS Rosenberg et a1 ..235/92 EV ABSTRACT.

vided with a selectively variable frequency and phase by employing a variable frequency dividing circuit that is responsive to a highly accurate precision crystal clock generator. The variable divider includes a fixed modulus counter which in turn triggers a variable modulus counter. High frequency precision clock pulses are fed first to the fixed counter which upon completion of its count, switches the clock pulses to the variable counter. The latter, when it completes its count, provides an output pulse that (a) constitutes one element'of the system output, (b) resets the fixed counter, and (c) switches the clock pulses to the fixed counter whereby the above cycle is repeated. To vary the frequency or repetition rate of the described cycle of operation, the magnitude of the count provided by the variable counter is changed to thereby change the total interval or period of a single cycle. To control the phase 3:221? 253 of the output signal, the count of the fixed counter is changed 3378776 4/1968 g g /42 x during but one of its counts. The variable counter is formed of a pair of reversible, up-down counters each of which provides Primary Examiner.lhn S. Heyman an output upon reaching the count of one when counting Attorneyngausewitz Ca" & Rothenberg down. The frequency of the output is changed by controllably varying the number to which these variable reversible counters count.

..C i. 2 19 Q'QYEQEQ'T? .7 r

Pars; KNEW/70E l 65557 46 574677/1/6 PZ/ASE UP J a C/ECU/T Cal/N775? JUL UP-DOWA/ UT co/vravr COUNT/51? fgff L64 CONT/Q01. 5'W/7'0/ If;

F- I/VOQAS'E 2 JUL 0 FQWMWO :66 [4 70 4 p/ezcg/ a /v 5 6 0N T T 056667256 --1 6 44) FFZM/fA/C/ I P4035 57 155 K F W455 JUL uP-wW/v J5! 57 (O/WEN T caNTFOL (OZ/N756 T pow/v F5527 T L! I PATENTEUHARZI I972 SHEET UlUF 11 PATENTEDHARZI I972 SHEET O3UF 11 PATENTEDMAR21 m2 SHEET OSUF 11 W MGVKM N 0 5 N/ M mm 5 H y 0 P PAIENTEBMARm I972 SHEET D'IUF 11 INVENTOR. 9055/97 53 J4M/E50/V BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to timing systems and more particularly concerns a source of stable high precision timing signal that may be selectively varied in frequency and/or phase. In implementation of the timing aspect of the invention improved dividing circuits are employed.

2. Description of Prior Art Arrangements for controlling phase and frequency of oscillating circuits have long been employed in a variety of forms and involving many different techniques. Most commonly, an oscillator has its frequency varied by changing certain reactive elements in the circuit. Thus, a variable capacitor or inductor may be employed for frequency control. Many types of oscillators lend themselves to ready control by application of external synchronizing signals. Such oscillators, because of the variable element required for frequency control, are inherently subject to drift and lack suitably high tolerance, stability and precision of frequency output. For optimum precision of oscillator frequency, circuits are generally controlled by crystals which have a natural frequency that remains relatively stable and unaffected by environmental changes or aging. However, unless one goes to use of harmonic modes, the frequency of the crystal cannot be changed whereby the only way to change the frequency of a crystal controlled oscillator is by alternatively switching into the circuit one of a number of oscillators or crystals. This arrangement, of course, can provide but a limited number of frequency variations and does not readily lend itself to external synchronization or phase control.

Precision timing systems required in various power supplies, such as, for example, the uninterruptible power supply system employed in large digital computing machines, must be capable of being synchronized both in frequency and phase to an outside power source and yet, in the absence of such outside power source, must have an extremely stable frequency. Further, since a computer such as, for example, the IBM 360 computer, requires a power source that stays within a frequency band of i'r Hz., there is a maximum frequency variation that can be tolerated. Accordingly, it is desirable to provide a timing system of controllably variable frequency which, nevertheless, has set frequency limits so that it cannot be inadvertently varied in frequency to an extent that would cause damage to the system being powered thereby. No system meeting all of the above requirements has heretofore been available.

SUMMARY OF THE INVENTION In carrying out the principles of this invention in accordance with a preferred embodiment thereof, a fixed frequency clock providing basic frequency stability and accuracy is arranged to drive a variable divider that provides a system output that is selectively variable in phase and/or frequency. The variable divider may include means repetitively generating first and second chronologically successive time intervals, frequency control means for changing the duration of one of the intervals, phase control means for changing the duration of one of the intervals, and output means responsive to a predetermined point of one of the intervals. More specifically, the output of the fixed frequency clock is counted by a substantially fixed interval counter which upon termination of its count, switches the fixed frequency clock pulses to be counted by a variable counter. The latter upon reaching a predetermined point in its count provides an element of the frequency variable system output and initiates a repetition of the cycle of counting pulses of the train by the fixed counter and then by the variable counter. A significant aspect of the invention is the mutually opposite phase operation of a pair of oscillatory circuits, each operable in positive and negative half cycles, provided with means responsive to termination of one of the half cycles of the first circuit for terminating a concurrent half cycle of the second circuit, and wherein said means comprises means for initiating the next half cycle of the second circuit. In'a preferred embodiment the oscillatory circuits are reversible, up-down, counters.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a timing system according to principles of the present invention,

FIGS. 2a a-2d, collectively, comprise a synchrograph illustrative of the operation of FIG. 1,

FIG. 3 is a block diagram illustrating the fixed and variable modulus counters and their control,

FIGS. 4a, 4b, 4c and 4d illustrate details of the fixed modulus counter of FIG. 3,

FIG. 5 shows details of one of the reversible counters of FIG. 3,

FIG. 6 illustrates a starting circuit for the system,

FIG. 7 shows the details of the pulse train switch,

FIG. 8 shows the circuit of the direction control of the reversible counters,

FIG. 9 illustrates content control circuitry for the reversible counters,

FIG. 10 comprises an analog representation of the counter operation including increase and decrease frequency operation,

FIG. 11 is asynchrograph illustrating operation of phase advance and phase retard functions of the fixed modulus counter,

FIG. 12 is a synchrograph illustrating phasing of the frequency control signal, and

FIG. 13 is a block diagram of an arrangement for generating frequency and phase control command signals.

DESCRIPTION OF THE PREFERRED EMBODIMENT General System Advantages of digital integrated circuits, whether in timing devices, power systems, or other arrangements, are many and well known. Digital systems are not seriously affected by DC gains, drifts, temperature effects or the like. In most cases, noise is considerably less of a problem in a digital system. One can obtain greater precision of control in digital modes than in analog modes. Feedback is less of a problem. High feedback analog systems are subject to undesired oscillation problems requiring complex and sophisticated compensatory circuits. In a system such as an uninterruptible power supply requiring a crystal oscillator having a stability of 100 parts per million it is highly desirable to be able to synchronize the system in both frequency and phase to an outside power source. Notwithstanding such capability for synchronization, the system must exhibit extremely stable frequency in the absence of or upon failure of the outside power source. Such requirements of stability and accuracy can be met by employing a crystal clock as the basic frequency source. Suitable crystal controlled frequency sources are described in a copending application for Frequency Comparator, Ser. No. 33,208, filed on Apr. 30, 1970 by L.- C. Butler, Jr. and R. S. Jamieson, and in copending application for Multi-Channel Control Circuit, Ser. No. 8,877, filed on Feb. 5, 1970, by L. C. Butler, Jr., T. W. Grasmehr and R. S. Jamieson, both applications being assigned to the assignee of the present application and incorpOrated herein by reference. Such crystal clocks, although exhibiting high precision, are themselves subject to but a small tuning range. Accordingly, a fundamental advantage derived from use of the crystal clock, namely extreme stability, is antithetical to the presently identified requirement of a variable frequency and phase system that can be synchronized with another source. The two contradictory requirements, the first being for stable frequency, and the second being for external frequency and phase synchronization, are satisfied in the present invention by employing a fixed frequency clock to establish the basic frequency stability and accuracy and, in addition, a variable divider to provide the desired output signal of variable frequency and phase. The described embodiment of the present invention is particularly arranged for use in an uninterruptible power supply of the type referred to in the aforementioned application of Butler and Jamieson and of the type also referred to in copending application of L. C. Butler, Jr., Ser. No. 8,875, filed on Feb. 5, 1970 for Majority Logic System, assigned to the assignee of the present application and incorporated herein by reference. For reasons discussed in the above-identified applications, it is desirable to start with a precision clock signal of, for example, 230.4 kHz. and divide this down into a 360 l-Iz., six phase signal for driving inverters of an uninterruptible power system. Although these frequencies are employed in the illustrated embodiment by reason of the designed application for a specified system, it will be readily appreciated that the specific values of the various frequencies employed, the various moduli of the several counters, and particular details of magnitude of timing parameters all may be varied to suit particular needs and applications and are illustrated herein as being of the identified values solely for the purposes of exposition.

In order to provide a fixed frequency clock and a variable divider, one could simply employ a variable modulus counter, a counter whose division ratio could be changed at will and make such counter responsible to a fixed frequency clock. However, in accordance with a preferred embodiment of the present invention, a pair of interval generators are used. A fixed interval generator is followed by a variable interval generator as will be described more particularly hereinafter. The fixed interval generator is caused to actuate a variable interval generator which, in a preferred mechanization, is formed of a pair of up-down or reversible counters which exchange counts between them. As illustrated in FIG. 1, a clock pulse generator 16 provides a train of high stability and high precision pulses at a fundamental frequency, at a frequency in this example of 230.4 kHz. The pulses from the clock pulse generator are fed via a switch 18, when in the illustrated position, to a fixed time interval generator 20 which initiates generation of a fixed interval when the switch 18 is moved to the illustrated position. Upon termination of the interval provided by the generator 20, an output signal W is produced to cause the switch to move to the other position thereof which initiates operation of a variable interval generator 22. In the described embodiment both generators 20 and 22 are digital counters. When generator 22 times out, it produces an output signal C, which comprises one element of a train of pulses that forms the output of the described system. The output C, performs two other functions. First, it is fed back to the fixed interval generator 20 to reset the latter and place it in condition to initiate another of its time intervals. Second, it is fed back to operate the switch 18 and return it to the illustrated position thereof whereby operation of the fixed interval generator is once again initiated. Completion of this interval again operates the switch to initiate the timing of the second interval generator 22 and so on in continued cyclical steady-state operation.

The steady-state operation is illustrated in FIG. 2a wherein the first line illustrates the train of clock pulses 24 which comprise the units that are counted by the generators 20 and 22 for creation of their timed intervals. Upon first operation of the switch 18 into the position illustrated in FIG. 1, the fixed timed interval of generator 20 is initiated at time t, as indicated at 26. The interval of 26 terminates at time. 1 and thereupon operates switch 18 to effect initiation of the interval of the variable timer 22 as indicated in FIG. 2a at 28a. Interval 28a terminates at a time t, to thereupon operate switch 18 and initiate a subsequent fixed interval 26. Upon termination of the interval 28a there is produced the output pulse C, and upon the termination of the subsequent interval 26, a second interval 28a is initiated. Thus a steady-state cyclic operation continues with the two interval generators timing out alternately, one after the other, and an output pulse C, being generated at a fixed point in the cycle of one of the interval generators, specifically in this instance, upon the termination of the interval 28a. It will be readily appreciated that the train of pulses C, comprises a fixed frequency timing signal that may be taken from any readily identifiable point in the timing interval of either of the interval generators, but is suitably obtained at the fall of the interval designated as 28a in FIG. 2a.

In order to vary the frequency of the output of pulse train C these pulses are fed as illustrated in FIG. 1 to a frequency comparator 30 which receives as a second input thereto a reference frequency pulse train from a reference source 32. The difference in frequencies between the pulse train C, and the output of the reference frequency source 32 is fed via line 34 to a frequency adjust or command circuit 36 which produces a signal that is fed to the variable generator 22 to change the timing thereof in accordance with the sense of the frequency difference as determined by frequency comparator '30. A variety of methods and mechanisms may be used to change such timing as will be described hereinafter. Where the described system is to be employed as a standby power supply for an uninterruptible system, the reference system 32 may comprise the line frequency to which it is desired to synchronize the described system. It is not necessary to distinguish the magnitude of frequency difference, since it is necessary only to determine the sense of such difference, that is, to determine which of the two frequencies compared by comparator 30 is the higher. One such circuit that may be employed for this purpose is shown in the aforementioned application for Frequency Comparator of Butler and Jamieson and is described as a frequency difference sense detector. Other arrangements are readily available including the conventional frequency modulation servo that provides the sense of the frequency difference by.means of a shaft output rotation to thereby operate one of two switches depending upon which frequency is the higher and upon the direction of such a shaft rotation.

Functionally, operation of the frequency adjustment is illustrated in FIG. 2b wherein the situation for decrease of frequency is shown. The fixed interval pulse 26 remains the same, but for frequency decrease it is preferred to increase the variable interval to provide a longer period for the complete cycle whereby the repetition rate of the pulses C, is decreased. The relative timing of the various components of the system remains the same. That is, upon termination of interval 26, interval 28b is initiated and upon termination of the latter, the interval 26 is again initiated and the output pulse C, is generated. The difference resides in the fact that one of the intervals, interval 28b, has been lengthened.

As shown in FIG. 2c for increase of output frequency, the fixed interval 26 again remains the same. The output pulse C, still is produced upon the termination of the shorter variable interval, but the latter in this situation for increased frequency has been decreased as illustrated at 28c. Again the relative sequences remain the same, but the total period has been decreased to provide the commanded frequency increase.

For phase adjustment, output pulses C, are fed as one input to a phase comparator 37 (FIG. 1) which receives its reference input from the reference 32. Comparator 37 produces a signal that represents a phase error and is fed via line 39 to a phase adjust circuit 38. The latter produces signals that temporarily change the interval of the fixed interval generator 20 although temporary change of the variable interval generator may also control output phase. Illustrated in FIG. 2d are representations of the fixed and variable intervals and output pulse train C, for the condition of retarding of the phase of the output signals with respect to the phase of the pulse train illustrated in FIG. 20. The first of the intervals 26d and the shorter interval 28d, again remain the same as compared with the arrangement of FIG. 20 so that the interval between the first two output pulses C, shown in FIG. 2d remain unchanged. However in this situation, after the occurrence of the second output pulse C, a phase retard signal from the phase adjustment circuit 38, FIG. I, is fed to the fixed interval generator 20 to temporarily, or for one cycle of the output signal, increase its time period. Thus, the relatively fixed interval 26e has been lengthened for this one cycle. For subsequent cycles indicated at 26f and 263 this relatively fixed time interval goes back to its original length where it remains for subsequent steady-state operation. In each case, as before, the variable interval 28d is initiated upon termination of the fixed interval 26d or 262, etc., and upon termination of interval 284, the output pulse C, is generated and the fixed interval 26e or 26f is initiated. Thus, the fixed interval is changed but once in order to change the phase, and thereafter, the system returns to an unvarying steady-state operation. The magnitude of the phase change can be varied'by controlling the amount of lengthening of interval 26 or by repeating a selected lengthening. It will be readily appreciated that for a phase advance the fixed interval 26 would be shortened rather than lengthened during one (or several) of its occurrences to thereby advance the train of output pulses C,.

In the illustrations of FIGS. 2a through 2d, the fixed interval period 26 is shown to be of considerably greater duration than the nominal length of the variable interval 28. A significant advantage of employing both a fixed and variable interval generator resides in the protection afforded to the timed circuit. It was pointed out above that certain sensitive systems such as computing devices cannot tolerate a frequency variation beyond a certain amount such as plus or minus three percent, for example. For such a tolerance, it is possible with the illustrated arrangement to provide the fixed interval generator with a period equal to 97 percent of the desired period of the output frequency and to provide the variable interval generator with a nominal period equal to 3 percent of the desired output period. Accordingly, for a plus or minus 3 percent variation, the period of the variable interval may be varied from zero to an amount equal to a total of 6 percent of the desired output frequency. No greater variation can be achieved since even in the total absence of the variable interval generator, the fixed interval generator will still provide 97 percent of the output interval. Thus, the use of the fixed interval generator provides, in effect, a set of stops or limits for the frequency variation selection. It is not possible to inadvertently cause or command a frequency variation greater than the frequency tolerance of the system.

A number of different types of interval generators or timers, both analog and digital, are known and readily available for use in the practice of the invention as described in the system of FIG. 1. For example, relaxation oscillators, one-shot or monostable multi-vibrators may be employed with one triggering the other, and with various arrangements for effecting a temporary (in the case of the fixed interval generator or relatively a steady-state (in the case of variable interval generator 22) variation in the time interval thereof. Where conventional circuits such as one-shot multi-vibrators are employed, the clock pulse generator 16 need perform no function other than to initially trigger the interval generator. However, for high precision systems, the fixed interval generators are preferably made in the form of the more accurate counters which also perform the division of the high frequency clock pulse train of 230.4 kHz. down to a desired nominal 360 Hz. output timing signal.

A mechanization of the system of FIG. 1 employing a substantially fixed modulus counter and a pair of variable modulus reversible counters is illustrated in FIG. 3.

DIGITAL SYSTEM As indicated in FIG. 3, the clock pulse generator 16 which may be a source of high precision, high frequency train of clock pulses produced by equipment such as that more particularly described in the aforesaid copending applications, feeds the pulses thereof to a starting circuit 40 and thence through a pulse train switch 42 from which the pulses C,,, are fed alternately to a fixed interval counter 44, or to both of a pair of reversible counters 46, 48 through counter content control circuits 50, 52, respectively. The reversible or updown counters 46, 48 are substantially identical to each other and are arranged to operate under the control of a direction control circuit 54 to count in mutually opposite directions. Stated in another fashion, each of the counters is a cyclical counter or an oscillatory circuit that operates in a positive half cycle, say the up direction, for example, and thence in a negative half cycle, or down direction. When the counter 46 is counting up, the direction control for both counters is such as to cause counter 48 to be counting down and vice versa. When either counter reaches a predetermined point in its cycle or count, it feeds a signal to the direction control circuit 54 which thereupon performs three separate functions. First, the direction control circuit 54 produces the system output pulse indicated as C on output line 56. Second, via line 58, the output pulse C, is fed back to the fixed modulus counter 44 to reset this counter to 0. Third, via line 60 the output pulse C, is fed back to the pulse train switch 42 to cause this switch to feed the pulses from the pulse generator 16 through starting circuit to the fixed modulus counter 44 once again. The output pulse C, comprises one element or one pulse of the train of output pulses of the system. Initially, upon starting of the system, the pulse train switch 42 is in a position to feed pulses C,,, from the pulse generator to the input of the fixed modulus counter 44, which, accordingly, begins to count up from its zero count. When this counter attains its full count, it provides an output signal W on line 62 which is fed to the pulse train switch 42 to operate this switch into a position wherein the pulses from pulse generator 16 are now fed via line 64 to both of the content control circuits 50 and 52 of the two reversible counters. These counters then proceed to count the pulses of pulse train C one counter counting up and the other counting down under direction control of a signal provided on lines 66 and 68 from the direction control circuit 54. As previously indicated, when one of these counters reaches a preselected count, the counting direction of the two are switched, the fixed modulus counter is reset to zero, the pulse train switch is operated to switch the train C,,, to the fixed modulus counter and another output pulse C, is provided as a second element of the system output pulse train. Thus, the system will continuously recycle, counting a fixed interval by means of counter 44, and then a second interval by means of one half cycle of the reversible counters 46, 48.

The several periods and counts available from the fixed counter 44 and up-down counters 46, 48 are readily selected to meet a specific application. For example, for use in an uninterruptible power supply wherein a train of 360 Hz. pulses C, is required to drive the power supply inverters, the pulse generator may provide a pulse train of 230.4 kI-Iz. Thus, it is convenient to employ a fixed modulus counter 44 that will achieve a divide by 608 operation. In such an operation, the reversible counters are set to provide a nominal count of 32 and a maximum count of 64 so that the sum of the divisors, 608 and 32 total 640, which is the proper divisor of 230.4 kHz. to obtain the desired 360 Hz. For the exemplary counters in dicated, the tolerable frequency variation is from about whereby the nominal 360 Hz. pulse train C, may vary by :32 clock pulses C,,, for each repetition thereof. In most cases, however, the range of frequency variation is less important than the value of the nominal frequency, although either may be readily selected to suit a given application.

In order to vary the output frequency of the system, one or a group of increase or decrease frequency pulses f may be provided via line 70 to each of the counter content control circuits 50, 52. The addition of these pulses to the counters during the up or down counts thereof is but one of several ways to change the time required for either of these reversible counters to attain its selected count. Details of the counter and frequency control circuits thereof will be described 343( to about 378 hereinafter. Nevertheless, as previously described in connection with FIGS. 1 and 2, it will be understood that addition of pulses f, to the up-down counters will cause these to increase or decrease the total count thereof and thereby change the second of the two intervals that makeup the total period between output pulses C,. Furthermore, any change in the total count of these counters effects a steady-state change in the interval provided by these circuits whereby there is a relatively steady-state frequency change commanded. For phase control, on the other hand, it is convenient to provide a temporary change in the count of the fixed modulus counter 44 by adding or inhibiting a count of this counter. The addition of one or several pulses to this counter over and above the pulses received via the pulse train switch 42 will cause the fixed modulus counter to achieve its maximum count earlier whereby the entire cycle of the system is advanced. Conversely, inhibiting one or several pulses fed to the counting input of this counter causes it to reach its total count at a later time, thereby retarding the cycle of the system. This operation is temporary since it affects only a single cycle of the fixed modulus counter. On the next cycle of the system, the counter will receive the usual number of pulses and count to its predetermined value in the same time. Accordingly, there are two independent arrangements for controlling first, the phase by means of the fixed modulus counter, and second, the frequency by means of the reversible counters. It will be readily understood, however, that opposite sense variation of the frequency of the system on successive or different cycles thereof could also achieve a phase variation of the output. In other words, if input pulses f are fed in such a manner as to increase the frequency of the system at one instant and then a like number of such pulses are subsequently fed to decrease the frequency of the system, there will have been achieved no net change in frequency although the phase of the output will have been advanced. Similarly, although it is preferred to employ the fixed modulus counter 44 for phase control only, it may be repetitively employed to achieve relatively small amounts of frequency control. It is well known that phase modulation is, from one point of view, simply another form of frequency modulation.

Before a detailed description of the mode of operation of the digital system is presented, it will be convenient to describe the details of the fixed modulus and reversible counters and control circuitry therefor.

FIXED MODULUS COUNTER The fixed modulus counter 44 is a conventional arrangement of inter-connected flip-flops of .IK type which are operated to provide mutually exclusive outputs in any one condition. The flip-flops are conventional circuits which provide mutually exclusive outputs in any one condition as is well known. Each flip-flop as typically shown in FIG. 4d has direct set and reset terminals S and R which, when low, will shift and hold the flip-flop in its set or reset state respectively providing at the two flip-flop output terminals Q and Q, respectively, high and low outputs for the set condition of the flip-flop and respectively low and high outputs for the reset condition of the flip-flop. In addition to the direct set and reset terminals responsive to low signals, each flip-flop has a clock or toggle input I and a set and reset input gate indicated at s and r. Each of the set and reset input gates has two inputs which, when high, enable the input gates and allow the flip-flop to be toggled or to change its state when the clock or toggle input goes low. That is, the set or reset gate that is enabled by a high at its two inputs will provide a signal that allows the flip-flop to be set or reset upon the fall of the toggle if it is not already in such condition. Thus, the direct set and reset terminals are responsive to steady-state low signals, and the set and reset input gates are enabled by high signals to cause the flip-flop to be toggled on the fall of the toggle signal thereto. Typical flipflops of the type described are sold as microcircuit chips designated MC945FG, MC845F, P,G, MC948F,G, M C848F,P,G, and described in Integrated Circuit Data Book, First Edition, Aug. 1968, Motorola Semi-Conductor Products, Inc. All of the flip-flops described herein are of the type described so that a detailed description of only one is necessary. As shown in FIG. 4a, flip-flop 74, which comprises the first or input stage of the fixed modulus counter, has direct set and reset terminals 76 and 78 which, when low, will shift and hold the flip-flop to or in its set or reset state respectively which provide at the output terminals generally referred to as Q and 6, but more specifically for flip-flop 74, leads 80 and 82 thereof, high and low outputs for the set condition of the flipflop and low and high outputs respectively for the reset condition. In addition to the direct set and reset terminals, responsive to low signals, each flip-flop has a set and reset gate indicated at 84, 86 for flip-flop 74. Each of the set and reset gates has two inputs. These gates selectively enable the toggle input to the flip-flop. That is, upon the fall of the triggering input 84a, the set or reset gate that is enabled by a high on its two inputs will allow the flip-flop to be set or reset if it is not already in such condition. Thus, the direct set and reset terminals are responsive to steady-state low signals and force their respective output states independent of the condition of the other inputs whereas the set and reset gates of the flipflops are enabled by high signals, thereby permitting a specific state change upon the fall of the clock signal thereto. The first stage flip-flop 74 of the fixed modulus counter receives a triggering input via line 88 on which appears the pulse train C from the pulse train switch. This pulse train is also fed as the toggle input of flip-flop 94 of the second stage of the counter. All of the flip-flops employed herein and described in connection with other circuits of this system. are substantially identical to the flip-flop 74, its set and reset gates 84, 86, and its direct set and reset terminals 76, 78. Thus, it will be seen that when flip-flop 74 is set, its output terminal 80 provides a first enabling signal to the reset gate 92 of flip-flop 94, the latter when set provides a second enabling input to its reset gate so that upon the fall of a toggle or clock signal on lead 88, flipflop 94 will be reset. The 6 output of flip-flop 74 is fed via lead 82 and an OR gate 96 to one of the enabling inputs of the set input gate on flip-flop 94. Note that the OR gates described herein are structurally the same as the NAND gates, providing the logical OR function for low signals, with polarity inversion. Understanding of operation of the system is facilitated by referring to the various gates by their logical functions although the gate circuits are the same for OR and NAND. The Q output of flip-flop 94, the second stage of this counter, is fed as the toggle input of a third stage of the counter including a flip-flop 98. Additional counter stages including their set and reset gates as indicated at 100, 102, 104, 106, 108, and 112 (seealso FIG. 4b) provide this counter with a total count that produces the desired dividing ratio. Each of the counter stages is identical with all of the others and each is connected to the preceding stage in the same manner as counter flip-flops 94 and 98. The OR gate 96 between the zero or 6 output of flip-flop 74 and the set input gate 90 of flip-flop 94 is provided uniquely for the first two stages in order to provide for the advance retard feature functionally described above but yet to be described in detail. In order to provide the output signal W indicating attainment of the desired count by the fixed modulus counter, there is provided a pair of five input inverting AND or NAND gates 114, 116(FIG. 4c) having inputs (via leads not shown) as indicated in the drawing from E various flip-flops or counter stages 74, 94, 98, 100, 102, 104, 106, m, fill and l 12. It will be understood that the bar over the indicated counter stage indicates that the output is taken from the 6 output terminal of the specific stage, whereas the number without a bar indicates an output derived from the Q side. The outputs of the two NAND gates 116 are fed to a second NAND gate 118 to provide the positive going pulse W that is the output of this counter, and which is fed to operate the pulse train switch to be described in detail below. The NAND gate is a conventional circuit that provides a low output when and only when all of its inputs arehigh. When any one or more of its inputs is low, its output is high.

In order to cause the fixed modulus counter to attain its output count earlier or later in time, there is provided a pair of advance and retard flip-flops 120 and 122 (FIG. 4a) having set and reset input gates and direct set and reset input terminals all as described in connection with the other flip-flops employed herein. Each of the advance and retard flip-flops has a clock or toggle input via line 88 in the form of the clock pulses C,,,, and each is adapted to receive a negative retard or advance pulse applied to its direct set terminal from the phase advance or phase retard circuit as indicated in FIG. 2. The 2 outputs of both of the advance and retard flip-flops are fed as the inputs to NAND gate 124 which has its output inverted in a second NAND gate 126, that, in turn, provides an input to the set gate 84 of the first counter stage. The 6 output of the advance flip-flop is also fed via line 128 as the second of the two inputs of OR gate 96.

With a symmetrical square wave train of clock pulses C as indicated in FIG. 11, the first stage 74 of the fixed modulus counter of FIG. 4a sets and resets upon each fall of pulse train C,, to provide a train of square wave pulses, essentially a divide by two pulse train as indicated in the drawing. The second counter stage 94 sets and resets upon each fall of the Q output of the first counter stage as illustrated in the drawing. Upon occurrence of a negative going advance pulse 130 as indicated in the second row of output pulses of flip-flop 74, the advance flip-flop 120 has a low on its direct set terminal and is set to provide a low output at its 6 terminal, thereby disabling the NAND gate 124 and set gate 84. Concurrently, the low signal at the 6 output terminal of advance flip-flop 120 is fed through the OR gate 96 and inverted thereby to provide an enabling input to the set gate 90 of the second counter stage. Accordingly, the second stage flip-flop 94 will set on the next fall of a clock pulse as indicated at 132, whereas flip-flop 74, its set gate disabled, will remain low as indicated at 134. Flipflop 120 is reset or cleared by the next clock pulse whereby the relative phasing of flip-flops 74 and 94 thereafter continues exactly as it was prior to the occurrence of the advance pulse 130, except that both of these have now been advanced by one full cycle of the clock pulse C,, and, accordingly, the attainment of the desired total count of the fixed modulus counter is likewise advanced by one full clock pulse interval.

Upon occurrence of negative going retard pulse 136, also illustrated in FIG. 11, the direct set terminal of flip-flop 122 goes low whereby this flip-flop 122 is set and its 6 output is low, disabling NAND gate 124 and set gate 84 of the first flipflop stage 74. Accordingly, the Q output of this flip-flop remains low as indicated in the drawing at 138, upon occurrence of the next clock pulse. No additional input pulse is fed to input gates of the second stage flip-flop 94, although flipflop 74, its Q output remaining low and its 6 output high, provides a high output to the input of OR gate 96 which appears as a low input to the set gate 90 of flip-flop 94. Accordingly, the latter cannot be set upon occurrence of the next clock pulse and simply retains its low state for an additional clock pulse interval as indicated at 140. Retard flip-flop 122 is reset or clear upon the fall of the following clock pulse. Thereafter, in the absence of additional advance or retard pulses, the first two stages of the counter, stages 74 and 94, continue their cyclic operation as illustrated in the drawing and both have been retarded in phase by the period of one full clock pulse interval. Accordingly, the output pulse W has been delayed by such interval.

Illustrated in FIG. 4b is the starting gating for the fixed modulus counter. The output pulse C, which is provided by the reversible counter direction control circuit, is fed as the first input to an OR gate 142 which receives as a second input thereto a plurality of starting pulses identified as P. These pulses are fed via a pair of NAND gates 144 to a common reset line 146 connected to the direct reset terminal of all of the flip-flop stages except the advance and retard flip-flops, whereby each of the counter stages may be reset to zero upon start of the operation and upon each occurrence of an output pulse of C,,. Although the output pulse C, is not fed to the advance/retard flip-flops 120, 122, it is not necessary to place these in condition initially because they are immediately reset by clock pulses C, on line 88. The set gates of these flip-flops 120, 122 are disabled by the fixed ground or low input thereto.

REVERSIBLE COUNTERS Illustrated in FIG. 5 is an exemplary one of the two reversible counters, each of which is identical to the other, receiving only different control signals so that the two will always operate in mutually opposite phase relation during their reversible cyclical counting period. Direction and content control of the counters will be described below in connection with FIGS. 8 and 9. The various stages of the reversible counters each comprises a .ll( or a set and reset flip-flop having direct set and reset input terminals, and set and reset gates which permit the flip-flop to toggle upon fall of a clock input thereto just as described in connection with the corresponding flipflops and gating of the fixed modulus counter. Accordingly, each reversible counter, having in this expository arrangement a nominal counting half cycle of 32 and a total counting capacity of 64, comprises six stages of which the first two and last one are illustrated in detail with the three intermediate stages being illustrated in block. To cause these counters to reversibly count either up or down, the output terminals of each flip-flop are fed as toggling inputs to the input of the succeeding flip-flop stage via suitable gating which is identical for the interstage coupling of all of the counters as illustrated in FIG. 5. Thus, the first stage includes a flip-flop 148 that is followed by interstage gating comprising first and second NAND gates 150, 152 having the outputs thereof fed via an OR gate 154 to the clock or toggling input of the succeeding stage. Similarly, the Q and O outputs of the second stage flip-flop 156 are fed through similar gating to toggle the third stage 158. Similar connections are made to and between fourth and fifth stages 160 arid 163, with the outputs of the fifth stage 163 from the Q and Q terminals thereof being fed via NAND gates 164, 168, respectively, and OR gate to the toggle input of the final stage 172. Each of the NAND gates such as l50'and 164 that has a first input directly from the Q output terminal of the preceding stage flip-flop, also has a second input from a line 174 on which appears the up command signal. Whenever the signal on this line is high, the counter is counting in the up mode. Each of the NAND gates such as NAND gates 152 and 168 having a first input directly from the 6 side of the output terminal of the preceding stage flip-flop has a second input from a down command line 176. The signal on this line is high whenever the counter is commanded to count in the down direction. The third input to the interstage NAND gates, whether counting up or down, is provided for the first pair of NAND gates 150 and 152 from a counting input line 178 at which appears the pulse train C that is to be counted. The succeeding pairs of interstage NAND gates receive as their third inputs, the output of the OR gate of the preceding stage. Thus, NAND gates 150, 152 provide first and second inputs to OR gate 154. The latter provides its output as the clock input to the toggle terminal of the succeeding stage and also provides via line 180 the same output as a third input to both of the interstage NAND gates, (not shown) that follow the second stage flip-flop I56. Likewise, the third input on line 182 to the final stage NAND gates 164, 168 is provided by the output of the OR gate (not shown) that follows the fourth flipflop 160. Accordingly, with the arrangement of interstage gating illustrated in FIG. 5, each counter will count up or down in straight binary code depending upon the state of the signal on lines 174 and 176, respectively. Whenever the signal on line 174 is high, the signal on line 176 is low and vice versa, as will be described below with reference to FIG. 8.

As previously explained, the output of each of the reversible counters is provided when the counter attains a preselected count. In the embodiment illustrated herein, the preselected count for each of the counters is the count of one in the down counting mode. Accordingly, there is provided an eight input NAND gate 184 having a first pair of inputs from the down command line 176 and having six additional inputs from the various counter output stages as indicated. It is noted that identical input signals to plural terminals of certain of the gates are shown only because of the particular number of input terminals that are provided in the standard circuits that have been employed in mechanization of the described embodiment. All of the inputs to output gate 184 are provided from the 6 sides of the flip-flops except that output which is obtained from the first stage 148. Thus the input to gate 184 comprises the binary number 000001, the count of one. NAND gate 184 will provide a negative going output pulse when all of the inputs thereto are high. This negative going pulse is a pulse that is fed via line 232 for counter 46 and via line 234 for counter 48 through the direction control circuit of FIG. 8 to provide therefrom the output pulse C,, as will be described more particularly below.

The counter of FIG. will count input pulses provided on line 178 in straight binary notation, and will count in one direction until the control signals on the up-down command lines 174 and 176 are interchanged at which time the counter will start to count in the opposite direction. Accordingly, simply by changing the signal on lines 174 from high to low and concommitantly changing the signal on line 176 from low to high, the counter may be made to terminate a particular half cycle and initiate the succeeding half cycle, that is, to change the direction of its count. Further, by terminating the application of input pulses to the counter, the counter will stop its count, performing a memory function to store the particular count that had been reached when the input pulses are stopped.

STARTING CIRCUIT.

Illustrated in FIG. 6 is the starting circuit that has the function of initially providing on a line 186 the train of negative going preset pulses P which are fed to the various stages of the several counters and to other circuits to be described more particularly hereinafter to establish the several flip-flops in appropriate conditions for start of system operation. After a short start interval, the starting circuit of FIG. 6 operates to provide the clock train pulses C at an output terminal 188 thereof. The clock pulses C,,, are fed to the starting circuit from the clock pulse generator 16 (FIGS. 1 and 3) on a line a delay of several clock pulse intervals, the capacitor provides one input and one of the clock pulses C provides the other input to the NAND gate 204. This gate then provides a low output to gate 206 which enables the set input gate of flip-flop 202. The flip-flop is then set upon the fall of the next clock pulse. When the flip-flop 202 is set, its Q output terminal goes high to enable gate 192. Gate 194 is disabled to cut off the starting pulses P and the clock pulses C,,, are fed to the rest of the system from line 188.

PULSE TRAIN SWITCH The pulse train switch, as illustrated in FIG. 7, comprises a flip-flop 214 having direct set and reset input terminals and set and reset input gates connected and arranged as are all of the corresponding circuits of the other flip-flops employed herein. The clock train pulses C,, are fed from line 188 of the start circuit as the toggle input to flip-flop 214 and also fed as a first input to each of NAND gates 216 and 218. The reset input gate of pulse train switch flip-flop 214 is enabled by a positive going pulse W on line 215 received from the output of gate 118 of FIG. 4c. The set input gate of pulse train switch flipflop 214 is enabled by occurrence of the system output pulse C, appearing on line 217 from one of the outputs of the direction control circuit illustrated in detail in FIG. 8. The direct set input terminal of flip-flop 214 receives the preset starting pulses from line 186 and, accordingly, upon tum-on of the system flip-flop 214 is in the set condition. An output from its 0 terminal is fed to gate 216 which is, accordingly, enabled to receive and pass the clock train pulses C,,,. The output of gate 216 is inverted a second time in the NAND gate 220 which provides at an output terminal 222 the train of pulses 190 and also fed via a pair of NAND gates 192, 194, to a a second pair of NAND gates 196, 198. At the output of gate 196 appears the train of positive going clock pulses C,,,. These are fed to the pulse train switch which directs these pulses to the several counters. The output of NAND gate 198 is further inverted in a gate 200 at the output of which appears the starting pulses P on line 186.

When the circuit is first turned on, NAND gate 192 is disabled by the low input provided from the Q output of a starting flip-flop 202 which has the 6 output terminal thereof connected to enable the second starting gate 194. Initially, flipflop 202 is in reset condition and is arranged to receive a toggling input in the form of the clock pulses C on line 190. This flip-flop also receives enabling inputs to its reset and set gates from the output of a series of starting NAND gates 204 and 206, respectively. Initially, the 6 output side of flip-flop 202 is high, providing a high input to an OR gate 208 which provides a low input to gate 204 whereby both inputs to gate 206 (from the output of gate 204) are high. The output of gate 206 is low and the set input gate of the flip-flop 202 is disabled. When the clock pulse generator is turned on, pulses C,,, are fed through an input gate 210 through the OR gate 208 and to the NAND gate 204 to provide a first input thereto. A second input to the NAND gate 204 is provided by a capacitor 212 connected to an input diode (not shown) of NAND gate 204 substantially as described in connection with the timing flip-flop T described in FIG. 8 of the aforesaid copending application for Frequency Comparator. NAND gate 204 provides no output until a series of pulses are received via the OR gate 208 sufficient to charge the capacitor 212 to a value at which the capacitor provides the second input to the NAND gate 204. Thus, after C,, that is fed to the fixed modulus counter illustrated in detail in FIG. 4a, 4b and 4c. Accordingly, the fixed modulus counter proceeds to count to its selected value and when attaining such count, provides the output signal W on line 215 which enables the reset gate of flip-flop 214 whereby upon the fall of the next clock pulse C,,,, the flip-flop is reset. Since the fixed counter flip-flops are reset by starting pulses P, counting will commence only after start flip-flop 202 has been toggled. With flip-flop 214 reset gate 216 is no longer enabled and clock pulses no longer appear on output line 222. However, the 6 terminal of the pulse train switch flip-flop is now high whereby gate 218 is enabled and may pass to its output line 224 the clock pulses C,, that appear on line 188. Negative pulses on line 224 are fed to the content control circuits (FIG. 9) of the reversible counters which proceed to count. When that one of the reversible counters that is counting in the down mode reaches the count of one, an output pulse C, appears from the direction control circuit (FIG. 8) which enables the set input gate of flip-flop 214, whereby the latter will once again be set upon the fall of the following clock pulse. As previously described with the flip-flop 214 in set condition, the clock pulses are no longer fed to the up-down counters but are instead fed to the fixed modulus counter 44. Thus, the system recycles itself, alternately counting clock pulses in the fixed and reversible counters.

DIRECTION CONTROL CIRCUIT Illustrated in FIG. 8 are details of the direction control circuit which basically comprises a flip-flop 226 substantially identical to all of the other flip-flops described herein. The flip-flop receives a clock input from the clock train pulses C,, and is initially placed in set condition by receipt of starting pulses P which are applied to its direct set input terminal. With flip-flop 226 in set condition, its 0 output terminal is high and the circuit, as will be more particularly described below, is in condition to command counter 46 to count in the down mode and to command counter 48 to count in the up mode. The set and reset input gates of the direction control flip-flop 226 are enabled by the outputs of NAND gates 228 and 230, respectively, which receive outputs from the one count of the reversible counters 48 and 46, respectively, on lines 234 and 232. Upon start up, the starting pulses P are fed to the various counter input stages so that initially counter 46, which is to start counting in the down mode, is at its nominal high count of 32, and counter 48 which will start counting in the up mode, is at its lowest count. This initial condition is obtained by feeding the starting pulses to the appropriate direct set and reset terminals of the several counter flip-flops. For example, counter 46 is initially established with a count of 32 by feeding start pulses to the direct reset terminals of all but the last of its stages. The starting pulses are fed to the direct set input of flipflop 172 of this counter uniquely. Although, a one output from counter 48 would normally be fed to enable the set input gate of the direction flip-flop 226, the latter is initially in set condition. The selected count outputs from the reversible counters which occur when each of these reaches the count of one" when counting down, appear on input lines 232 and 234 of the direction control circuit and are also fed as inputs to an OR gate 236 which, accordingly, provides at its output line 217, the system output pulse C,,. It will be recalled that this is the pulse which (a) provides an individual element of the output pulse train of the system, (b) is fed to the pulse train switch to put the latter in condition to pass pulses C,, to the fixed modulus counter and (c) provides an input to the OR gate 142 (via a NAND gate 143) of FIG. 4a to reset the fixed modulus counter to its condition wherein it may initiate its fixed count. The Q and 6 outputs of the direction control flipflop 226 are fed to NAND gates 238 and 240, respectively, which invert the inputs thereto and provide on output lead 242 a high signal to command counter 46 to count down and counter 48 to count up, or on output lead 244 a high signal to command counter 48 to count down and counter 46 to count REVERSIBLE COUNTER CONTENT CONTROL Referring to FIG. 9, the inputs to the two reversible counters are provided via OR gates 246 and 248 as a train of positive going pulses on lines 250 and 252 which are fed to the inputs of the first stage of each of reversible counters 46, 48, respectively. These reversible counter input lines 250, 252 correspond to the input line 178 of the exemplary counter illustrated in FIG. 5. The OR gates are arranged to pass one of three possible input pulse trains as counting inputs to the reversible counters. The first of these pulse trains is that provided on output line 224 of the pulse train switch. This is the clock pulse train C As a second of the three possible inputs, each of the OR gates may receive an increase frequency command in the form of a series of positive going pulses fed from an increase frequency command line 254 via a down count NAND gate 256 for counter 48 (enabled only when counter 48 is in down mode) and via a down count NAND gate 258 for counter 46 (enabled only when counter 46 is in down mode). The third input to the OR gates 246 and 248 is provided by a series of decrease frequency command pulses on a line 260 which are fed to an up count NAND gate 262 for counter 46 and to an up count NAND gate 264 for counter 48. Up count gates 262 and 264 are enabled only when the respective countors are in up counting mode. Thus, the increase or decrease frequency command pulses are applied to the reversible counter input only when it is operating in an appropriate one of its two half cycles of operating.

As will be described in detail below, the count of the reversible counters may be varied to change frequency or to change phase in any one of a number of ways. There is selected for purposes of illustration an arrangement wherein frequency or repetition rate of the output signal C, is increased by feeding additional counting pulses to that counter which is counting down and decreased by feeding additional pulses to be counted by that counter which is counting up. Only two of the NAND gates 256, 258, 262, and 264 will be enabled at any one time and only one of the enabled pair receives a frequency change command signal. To command an increase of frequency, a group of increase frequency pulses is fed from line 254 to be passed through NAND gates 258 and 256 only when these NAND gates are enabled. For the method of frequency control illustrated, these are enabled only when the respective counters are counting down. A down direction command signal, appearing as a high on line 244, is fed as an enabling input to gate 256 when counter 48 is counting down and also fed as an enabling input to gate 262 when counter 46 is counting up. Similarly, for a down counting operation of counter 46, line 242 is high and enables the down counting gate 258 of the content control of counter 46 and simultaneously enables the up counting gate 264 of the content control of counter 48. Thus, the arrangement allows pulses to be added only to that counter counting down for a commanded frequency increase and only to that counter counting up for a commanded frequency decrease. The up-down count command signals on lines 242 and 244 are provided as outputs of the counter content control at output lines 242b and 244b which command counter 48 to count up and down respectively, and at output lines 244a, 242a, which command counter 46 to count up and down respectively. Up-down lines 174, 176 of FIG. 5 correspond to content control lines 244a, 242a, respectively, for counter 46 and to content control lines 242b, 244b, respectively, for counter 48.

FREQUENCY CONTROL OPERATION From the foregoing description and the following discussion it will be appreciated that the described reversible counters and control circuits therefore comprise an arrangement having exceeding flexibility of control that readily enables operation for a variety of applications in addition to that specifically described herein, including use as a variable divider, a time modulator or an externally controllable oscillator.

Illustrated in FIG. 10 is an heuristic representation of the operation of the digital frequency control system described above. The first line of this synchro-graph illustrates a group of decrease frequency command pulses 266. The second line indicates two groups 268, 270 of increase frequency command pulses. The third line of the graph schematically represents the operation of the fixed modulus counter 44. The next line shows the output pulses W of the fixed counter. The following two lines graphically represent the operation of the up-down X and Y counters 46 and 48. The final line illustrates the clock pulses C provided at the count of one by each of the reversible counters in the down counting mode. The analog representations of counter contents are employed in FIG. 10 solely to facilitate understanding of operation since the counter contents are digital, actually changing in discrete amounts. When the system is turned on and the preset pulses P have operated, the several counters are in the initial states ad indicated with the fixed counter being initially at the count of zero, the up-down counter 48 initially being at the count of zero and the up-down counter 46 initially being at its uppermost count for a nominal operation, namely, the count of 32 in the example previously discussed. As soon as the clock pulses begin to be applied to the counters with the pulse train switch in the initial position, the fixed modulus counter begins to count these and its count increases as indicated by the straight line portion 272 of the analog curve depicting the increasing count of this counter. When this counter reaches its full count at point 274, the first of its output pulses W is generated. It may be noted at this point that the pulses W are always synchronous with the variable frequency output C, since they are derived from the same basic source. Thus other pulse trains synchronous with the output pulses C, but relatively phase shifted are available from the output or from intermediate stages of the fixed counter. Where a variable frequency and phase control system such as described above is to be employed in an uninterruptible power system having a plurality of multiphase gated inverters, the availability of such a synchronous fixed frequency output pulse train such as the train of pulses of W, may be of great significance.

While the fixed modulus counter is counting its first full count, both of the reversible counters remain at their respective counts of zero and 32 as illustrated in FIG. 10. Upon occurrence of the first pulse W the pulse train switch is operated to feed the pulses C,,, to both of these counters which then proceed to count up and down respectively as indicated by the straight line portions 276 and 278 of the curves depicting contents of the reversible counters. It will be recalled that the direction control flip-flop of 226 is initially in set condition wherein it provides a high on line 242 and a low on line 244 to command counter 48 to count up and counter 46 to count down. As counter 46 counts down toward the count of one, counter 48 counts up. When counter 46 reaches its count of one, it provides an output signal from gate 184 (FIG. which appears on line 232 (FIG. 8). This low signal is produced at the output of gate 184 upon attainment of the count of one by the counter only when counting down because of input 176 to gate 184. The low on line 232 disables gate 230 whereby its output goes high to enable the reset input of the direction control flip-flop 226 and allows this circuit to be reset upon the fall of the next clock signal. Concomitantly the direction control flip-flop 226 is toggled, OR gate 236 provides a positive pulse C, which is fed via line 217 to enable the setting of pulse train switch flip-flop 214 and also to reset the fixed counter 44 by application of the reset pulse thereto as illustrated in FIG. 4b. Pulse C, is initiated via OR gate 236 (FIG. 8) when the down counter enters the one state. When pulse C, is high, the appropriate input gate of flip-flop 226 is also enabled. Thus, flip-flop 226 is toggled at the drop of the next clock pulse. The change of state of flip-flop 226 causes pulse C, to go low via the action of the down input on line 176 of gate 184 (FIG. 5). Thus pulse C, is one clock pulse period wide.

Similarly, the pulse train switch flip-flop 214 (FIG. 7) has its appropriate input gate enabled by pulse C, and, accordingly, toggled simultaneously with the change of state of direction control flip-flop 226. Thus the pulse C, is coincident with the one count of the down counting counter. At the end of this pulse C, (while 6 of pulse train switch flip-flop 214 is still high, just as this 6 is about to go low,) there is an additional toggle input to both up-down counters from the fall of the pulse C,,, on line 188 via gates 224 (FIG. 7) and 246, 248 (FIG. 9). Accordingly, the counters end their half cycle, not in counts of one" and thirty-two, but in counts ofzero and thirty-three and there remain when the clock pulse input train is cut off by operation of the pulse train switch. One full cycle has occurred and the first of the output pulses C, has been produced. The fixed modulus counter again begins its count while the two reversible counters remain at counts of thirty-three and zero", respectively, and the next cycle has started. The system will continue to recycle in .this manner, in steady state oscillatory operation, until it is disturbed. Each reversible counter operates intermittently in opposite half cycles, with a full count of counter 44 occurring between successive half cycles. Thus, one full cycle of the system includes (a) one full count of the fixed counter and (b) one half cycle (a full count in but one direction) of both of the reversible counters that are operating in mutually opposite phase. If deemed necessary or desirable for certain application the fixed interval between successive half cycles of reversible counter operation may be eliminated whereby each half cycle of one sense would immediately follow a half cycle of opposite sense for each counter. This operation is readily achieved by causing pulses C,,, to be fed to the reversible counters at all times, as by severing the line 217 between gate 236 and the pulse train switch, or eliminating, or by passing this switch entirely.

Consider now the situation where a burst of commanded increase frequency pulses 268 is provided as a counting input,

during the counting operation of the reversible counters. As discussed in connection with the counter content control circuit of FIG. 9, commanded increase frequency pulses are fed as inputs only to that one of the two counters that is counting down. Accordingly, counter 48 in this exemplary situation has begun its downward count indicated by the straight line portion 280 (FIG. 10) and upon receipt of the additional pulses 268, counts down more rapidly, or at an increased slope in this heuristic representation, as indicated by the increased slope potion 282. Upon cessation of the added input pulses 268, the counter resumes its normal rate of count as indicated at 284, until it reaches the count of one whereupon it provides at its output the negative going signal. This is fed to the direction control circuit to terminate thishalf cycle of the operation of both reversible counters and again provides an output pulse C,. It will be noted, however, that since the rate of down counting of the counter 48 was increased during this half cycle of operation, the total time required for the counter to count from its high count to the count of one is less in this instance than it would be if the counting rate had remained unchanged. Accordingly, if the interval from the beginning of the first cycle to the occurrence of the first pulse C, be designated as T and the following interval between pulses C, be designated as T it will be seen that the latter interval isless than the former, whereby at this point the second frequency will be greater than the first, presuming that no other changes are made in the normal counting operation of the several counters. This temporary change in down-counting rate of counter 48 is, in effect, remembered" by the system to result in a changed steady-state operation. This memory action may be understood by considering the effect of the shortened interval T upon the count attained by the counter 46 which was counting up during the half cycle of reversible counter operation in which the rate of counting of counter 48 was temporarily and momentarily increased. Since it is termination of the downward cycle of the down-counting counter that stops operation of the concurrent half cycles of both of the reversible counters and further, since the count of one is attained in a shorter time by the down-counting counter in this situation, the up-counting counter will have attained a lower value of count when this particular half cycle is terminated. Thus, for example, if seven counts are included in the burst of increased frequency pulses 268, the down-counting counter will reach the count of one when the up-counting counter reaches the count of 25 rather than its nominal high count of 32 (although both receive a single additional input, remaining at counts of zero and twenty-six respectively). Now the fixed modulus counter is reset, the pulse train switch feeds pulses to the fixed counter and the reversible counters temporarily stop operation. If no other changes were made in the system, the counter 46 which had reached a count of only 26 would thereupon count down from the count of 26 and both counters would, accordingly, continue to count to and from the counts of O and 26, thereby providing a steady-state oscillatory operation having a decreased period.

A second burst of commanded increased frequency input pulses 270 is shown in the next half cycle of reversible counter operation to illustrate the fact that these pulses may act upon either of the two counters. Thus, with regard to the frequency increase command pulses 270, counter 46 is counting down and temporarily, as indicated at 286, will increase the rate of its count upon receipt of pulses 270 and thereby reach the count of one sooner than it would have in the absence of the commanded increased frequency pulses 270. Accordingly, the next output pulse C, occurs at a time interval T following the end of interval T and the frequency is again increased to decrease the period so that interval T is still smaller than the preceding interval T Again, absent any further commanded increase or decrease frequency input pulses, the system has attained a steady-state operation at this now increased frequency and will continue to so operate until additional commands are received. 1

Now consider the case of commanded decrease frequency input as indicated by the burst of pulses 266. These decrease frequency pulses in the illustrated expository arrangement are provided only to the reversible counter that is in up-counting mode, namely, counter 48. Given the additional counting input, the up-counting counter has an increased rate of count as

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US2998190 *May 9, 1958Aug 29, 1961Gen Dynamics CorpAccumulator
US3044065 *Aug 5, 1957Jul 10, 1962Sperry Rand CorpElectronic programming means for synchronizing a plurality of remotely located similar means
US3096483 *Apr 6, 1961Jul 2, 1963Bendix CorpFrequency divider system with preset means to select countdown cycle
US3378776 *Oct 9, 1964Apr 16, 1968Stanford Research InstData shifting logic system with data and control signals advancing through the system in opposite directions
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US3786488 *Dec 30, 1971Jan 15, 1974Woodward Governor CoAlgebraic summing digital-to-analog converter
US3829785 *Apr 25, 1973Aug 13, 1974Philips CorpCircuit arrangement for digital frequency measurement
US3909791 *Jun 21, 1974Sep 30, 1975IbmSelectively settable frequency divider
US3997764 *Aug 23, 1974Dec 14, 1976Societe Generale De Constructions Electriques Et Mecaniques (Alsthom)Method for the conversion of a frequency into a number
US4473885 *May 28, 1981Sep 25, 1984Tokyo Shibaura Denki Kabushiki KaishaFrequency dividing ratio setting device for programmable counters
US4573175 *Sep 12, 1983Feb 25, 1986Case Communications Inc.Variable digital frequency generator with value storage
US4862096 *May 13, 1987Aug 29, 1989Tektronix, Inc.Programmable multiphase sequence controller
US5493322 *Apr 1, 1993Feb 20, 1996Eastman Kodak CompanyElectrophotographic image forming apparatus with non-uniformity correction of recording elements
US5657069 *Dec 13, 1991Aug 12, 1997Eastman Kodak CompanyMethod and apparatus for grey level printing
US6587528 *Oct 12, 1998Jul 1, 2003Ericsson Inc.Systems and methods for extracting and digitizing phase and frequency information from an analog signal
US8154321 *Dec 21, 2007Apr 10, 2012Power Integrations, Inc.Method and apparatus for time-differential comparison of an analog signal
US8324941Mar 1, 2012Dec 4, 2012Power Integrations, Inc.Method and apparatus for time-differential comparison of an analog signal
US8497729Oct 30, 2012Jul 30, 2013Power Integrations, Inc.Method and apparatus for time-differential comparison of an analog signal
US20090160489 *Dec 21, 2007Jun 25, 2009Power Integrations, Inc.Method and apparatus for time-differential comparison of an analog signal
CN101464658BDec 22, 2008Jan 23, 2013电力集成公司Method and apparatus for time-differential comparison of an analog signal
Classifications
U.S. Classification377/44, 377/45, 377/43, 327/49, 968/846
International ClassificationH03K3/72, H03K3/00, G04F10/04, G04F10/00
Cooperative ClassificationH03K3/72, G04F10/04
European ClassificationH03K3/72, G04F10/04