US3492424A

US3492424A - Time interval divider

Info

Publication number: US3492424A
Application number: US573285A
Authority: US
Inventors: Donald G C Hare; William L Rorden
Original assignee: Grass Valley Group Inc
Current assignee: Grass Valley Group Inc
Priority date: 1966-08-18
Filing date: 1966-08-18
Publication date: 1970-01-27
Anticipated expiration: 1987-01-27

Description

Jan. '27, 1970- D. G. c. HARE ET AL 3,492,424

TIME INTERVAL DIVIDER.

Filed Aug. 18, 1966 2 Sheets-Sheet 1 CONSTANT Z7 7 CURRENT i CHARGE saunas 0/005 CONDUCTS WHEN v V '1 16 Var I GLOSES a co/vsmv'r at MOMENTAR/LY T l4- cunnewr 540/1 PULSE l DISCHARGE 5L0 VI? P-2 P-3 P4 f I PULSE TRAIN r V7 V7 3 b ar :2 ..Z' g 2. 0

V 5 5 am et GROUND VOLTAGE T-t it] -4t I I PULSE TEA IN /N VENT 0/75 DONALD 6. 0. HA RE W/).L MM L RORDE' N United States Patent Office Patented Jan. 27, 1970 ABSTRACT OF THE DISCLOSURE Each of the pulses of a train of timing pulses which occur at intervals T but which may not .be evenly spaced, is anticipated at a time t prior to the occurrence of a pulse, where t in one time interval between pulses is deter-mined by dividing the duration of the time interval between two prior pulses by a constant radio t/ T.

This invention relates to a time-interval divider. More particularly, it relates to method and apparatus for defining portions of regularly or nearly regularly pulsed time intervals and the initiation thereof by reference to a constant ratio between each portion and the preceding time interval. In specific instances, it relates to method and apparatus for accurately synchronizing drive and blanking pulses and the like with each of a series of main actuating pulses, which may be evenly spaced or approximately but not quite evenly spaced.

Particular applications of the invention includ the provision of drive and blanking pulse systems and of noise gates from synchronizing pulses, but the invention is not limited to such uses.

For example, certain monochrome video synchronizing pulses used for horizontal timing may have a nominal width of 4.5 microseconds and may occur at a nominal interval of 63.5 microseconds, with a tolerance of one percent. The blanking or drive pulse has heretofore been derived by generating a fixed time delay from the preceding event. Thus, defining T=the actual interval between pulses, and

=a desired front porch interval, by which the blanking or pulse drive is to precede each pulse,

T-t=the time between each pulse and the initiation of succeeding front porch.

In actual practice the attempt has been to hold Tt to a constant time,

Suppose that t=1.5 microseconds. Then, a one-percent change in T would be 0.635 microsecond, and produces a forty percent change in the value of t. A decrease in T of a little more than two percent would cause complete missing of the synchronizing pulse. It has therefore been difiicult to assure accurate synchronization, for the tolerance has not always been sufiiciently small.

Even more critical tolerances have been required in the start of vertical blanking relative to the frame period, where accuracy of at least one part in five hundred is required-less than a thirty microsecond error in sixteen milliseconds.

Some prior-art methods of locking to a pulse train have employed phase-locked oscillators; such methods suffered from slow response, and they limited control range where high accuracy was to be achieved. In general, however, the prior art attempts to provide drive and blanking pulses have rested on deriving the gates (or delays and advances) by generating fixed time delays, using ramp generators, in which the instantaneous voltage value of the ramp was compared with a reference voltage V Each ramp was started and stopped at each pulse. So long as the reset time is constant and is short relativ to T, it may be neglected. Defining S=the slope of the ramp (a constant for the particular ramp generator),

V =the value of the voltage at the top of the ramp,

just before reset, and

V =the reference voltage for obtaining the blanking and drive pulse,

a value of V, was chosen such that V,=S(Tt) From this, it is apparent that t=TV,/S

Since both V, and S were constants, it is evident that (Tt) was a constant.

The ratio t/ T was, however, not a constant, but changed as T changed, for

(Equation 1) That the dependence of t/ T upon T was a serious matter can best be seen by considering the fact that necessarily (for video use) t/ T is much less than ,1 and then determining the ratio of a small change At to t and comparing it with the ratio of AT to T. From Equation 1 above,

At=AT--AV /S+ V /S-l-AS/S (Equation 4) Dividing this by t, as defined by Equation 1, and then rearranging the terms, We get A t 1 AT AV V I ASL t lV,/ST T V, ST ST ST (Equation 5) (Equation 2) (Equation 3) ST, and the term becomes a very large number. As a result, At/t varies greatly for a small change in any of V T, or S.

In contrast to the attempt to keep (Tt) constant, and thereby producing a wide variation in the value of t/ T, the present invention is directed to holding t/ T constant, or substantially so, thereby minimizing the effect of small changes in T or in S or in V,,. Thus At/t is substantially equal to AT/ T. As a result, so long as T varies slowly and AT/ T is small for successive intervals, the ratio t/ T is so nearly constant that it is independent of the small changes in T.

As a result, the invention enables accurate synchronization. It makes it possible to inhibit noise and other factors tending to cause misfire. Where there is no noise or where the noise is at a much lower level than the pulses, it is relatively easy to synchronize a system with original pulses, but Where the noise level is high, it produces misfires and complete loss of synchronization. Inhibiting enables the system to seek true pulses only during a short period of time preceding and following the expected event.

The present invention is not confined to use in noise gates, in horizontal scanning, or in vertical scanning, but is so useful in these applications that they are used herein as examples. Other applications will of course occur to those skilled in this art.

By working with the ratio t/ T as substantially constant instead of Widely varying, the front porch of each pulse is set at a desired value of t/T. For example, in horizontal scanning the value may be 2%, which is twice as large as the tolerance of T. In vertical scanning a different ratio may be chosen, if desired.

Our invention provides a simple, self-adjusting method and apparatus for holding t/T constant and, subject to certain restrictions, for eliminating all first-order error in the equality In the invention, the pickoff reference V, is defined in terms of V, or ST; fir example, by defining V =kST (Equation 6) where k is a constant, Equation 1 can be rewritten as Since k is small, this is substantially what is desired, namely;

At AT In our invention, we determine by measurement the value of one interval T and divide this measured interval as desired. The resulting ratio is applied to the next interval while simultaneously we measure that second interval, divide it, and apply the new ratio to the succeeding,

third, interval. The division is so performed that t/ T is held constant.

In other words, if we wish to anticipate the arrival of one of a succession of approximately evenly spaced pulses, we measure the interval between two successive pulses and then use this measurement to predict the arrival of the next pulse following the two which defined the measured interval. Further, we continue to measure the interval between each two successive pulses and to use it to predict the arrival of the next pulse thereafter.

In applying this new method to the preceding illustration and to Equation 6 we measure and store the value of V =ST for one period, divide the stored value appropriately to get a desired ratio, and store that ratio. Then the next ramp is compared to this stored ratio. The point of equality as indicated by the comparison is then the desired interval. The stored reference source is then reset and made ti store the next value of V -=ST.

Consider the example given above, where T=63.5 microseconds and t=1.5 microseconds and an error of 1% in T gave a 40 error in t in the prior-art system where T t was a constant. In that system, V, was constant and S was constant; in the present invention S remains constant, but Tt is defined in terms of the previous T, and V, also is so defined. Then, a one-percent change in T means AT At 0.01 t

and the change in t is also one percent and is 0.015 microsecond.

Similarly, in vertical synchronization of video systems, the stringent requirement that the accuracy of the advance t must be at least 0.2% is readily met in this invention. In contrast, in the prior art, careful and repeated adjustments were needed to maintain the required accuracy even with T held constant, and the prior-art system proved practically useless when T varied at all.

Other objects and advantages of the invention will appear from the following description of some preferred embodiments thereof.

In the drawings:

FIG. 1 is a diagram of a simplified circuit embodying the principles of the invention.

FIG. 2 is a waveform diagram obtained from the circuit of FIG. 1.

FIG. 3 is a waveform diagram superimposing two of the waveforms of FIG. 2.

FIG. 4 is a waveform diagram similar to FIG. 3 but with a delay in the start of the ascending ramp.

FIG. 5 is a circuit diagram of a specific interval divider embodying the principles of the invention.

FIG. 6 is a diagram of waveforms at various identified points in the circuit of FIG. 5.

Consider first the simplified form of circuit of FIG. 1. A switch 11 is closed momentarily by each pulse at, for example, P-l (see a in FIG. 2) to discharge the capacitor C (see b in FIG. 2) by placing ground potential on both plates thereof. A constant-current charge source 12 supplies a current I to the capacitor C after the switch 11 is opened, thereby charging it linearly at a slope S (see b in FIG. 2) to a voltage V by the time the next pulse P2 momentarily closes the switch 11 and discharges the capacitor C resetting the voltage there (V to zero. The waveform of the pulse train is shown at a in FIG. 2, and the waveform of V is shown at b. (Little error would result if the charge and discharge currents are not constant, so long as the resulting slopes do not depart too greatly from linearity.)

A diode 13 conducts current to a capacitor C, when V is greater than V The capacitance of C, is very much smaller than that of the capacitor C so that the charge flowing into C, via the diode 13 does not appreciably affect the charging rate of the capacitor C (However, this simplification is only for analysis of the FIG. 1 circuit, since by replacing the diode 13 with an emitter follower operating as a peak detector, the desired result can be achieved anyway. See FIG. 5.)

If, initially, V is less than V the capacitor C, will be charged during each cycle to V via the diode 13. When the switch 11 is closed and V falls to zero, the capacitor C, starts to discharge via a constant current 1 discharge or sink 14, with a slope S the waveform showing at c in FIG. 2. At some time V falls to a value equal to the now-increasing voltage V and then is once again charged to the value V The capacitor C is connected to a grounded resistor R-O, through which the charge and discharge current into and from the capacitor C necessarily flows. During discharge of the capacitor C the current produces a small negative voltage drop across the resistor R-O, as shown in the belowground portion of d in FIG. 2. At the moment when the descending ramp S intersects in value the ascending ramp S (See FIG. 3), the capacitor C begins charging, with a reversal of current through the resistor R-0. (This is the moment when V =V Thus, the voltage V across the resistor R-0 changes sign at this time, forming a pulse as shown in d in FIG. 2. The positive-going leading edge of this pulse determines the moment of intersection of the two ramps S and S (see c in FIG. 2), and this, in turn, defines the delay (Tt) and hence defines the advance t. The change of sign of the voltage V is not essential, but the occurrence of the positive-going pulse is. For example, V could rise from ground and yield the desired information.

FIG. 3 shows the two ramps b and c of FIG. 2 superimposed. V being defined as ST, in the ascending ramp V,=S(Tt) (Equation 9) In the descending ramp,

V =V S,(Tt) (Equation 10 -=STS (T-t) (Equation 11) Combining Equations 9 and 11 and solving for t gives t S+St (Equation 12) Therefore,

t S +8, fl l+S /S (Equation 13) Note how different Equation 13 is from Equation 3. It shows that t/ T is independent of T and depends only on the ratio of the two ramp slopes S and S Considering small changes, a little algebra applied to Equation 13 yields Ar 1 A] t T 1+S,/S S, S (Equation 14) To the degree that the slopes are constant, (AS-=0 and AS =0), or to the degree that they change in the same way (AS /S =AS/S), the equation is simplified into FIG. 4 is used to show that the system does not require an instantaneous or even relatively short reset of the ascending ramp. FIG. 4 differs from FIG. 3 in that the start of the ascending ramp has been delayed by an interval x. The ascending slope now becomes S, so that V =S'(Tx). On the ascending ramp,

(Equation 15) which, simplified gives S't-i-S t=S T (Equation 17) Solving for 1 gives S+S (Equation 18) Thus, Equation 18 is the same as Equation 1'2, except for S replacing S, and the interval x does not appear in it.

The simplified circuit of FIG. 1 can be realized by the practical embodiment shown in FIG. 5, used for horizontal video synchronization.

A voltage of about ten volts may be applied to the device by terminals 15 and 16 which are respectively connected to a positive bus 17 and a negative bus 18. A synchronizing pulse H sync is fed in through a noise gate 19, which may have a timing circuit identical to that shown in detail in the box 20 defined by broken lines. A pulse actuated drain switch 21 corresponds to the switch 11 of FIG. 1 and is connected to the noise gate 19 through a capacitor 0-1.

The capacitor C-l is connected to the base of a PNP transistor Ql and through a resistor R-1 to the positive bus 17 that leads to the emitter thereof. The collector of the transistor Ql is connected through a resistor R-2 to the base of an NPN transistor Q2. The base of Q2 is held at ground by a resistor R3, and its emitter is grounded. Its collector is connected to the capacitor C In FIG. 6, the waveforms are shown in their related patterns along the same time axis. The H sync pulse defines the time T at e. If the noise gate 19 is used, its inhibiting action is shown at f, preventing actuation by any pulses for the interval shown on that waveform. The sync pulses pass through and are differentiated by the capacitor C1 and the resistor R1, the negative going spike causing the PNP transistor Q-1 to conduct and raise the voltage at the base of the NPN transistor Q2. The patterns g and h respectively show the voltage waveforms at the base and collector of the transistor Ql. The transistor Q2 then conducts strongly and clamps the capacitor C to ground, thus discharging the run-up ramp, shown by waveform i in FIG. 6.

This ramp is generated by the current I flowing from a constant current charge source 22, corresponding to the element 12 of FIG. 1 and comprising a PNP transistor Q-3. The base of the transistor Q3 is connected to the positive bus 17 through a resistor R4 and bled toward ground by a potentiometer RS. Its emitter is connected to the bus 17 by a resistor R6. The capacitor C is charged by the constant current I via the collector of the transistor Q-3. The value of the constant current I is controlled by the resistors R4, R-5, and R6, together with the supply voltage. The slope S is determined by I and the value of the capacitor C With the value of S determined, the maximum height (voltage) of the ramp V is defined. The adjustable potentiometer RS enables adjustment of the slope S and thereby of the ramp height.

The diode 13 of the FIG. 1 circuit is here replaced by an emitter follower 23 acting as a peak detector, and comprising an NPN transistor Q4. The V voltage is applied to its base, and its emitter is connected to the capacitor C so that the voltage V is developed at the emitter of Q4.

A constant current discharge means 24 is provided by an NPN transistor QS with its collector connected to the emitter of the transistor Q4. The emitter of the transistor QS is connected to the negative bus 18 through the resistor R8 and its base is located between ground and the negative bus 18 by an adjustable potentiometer R7. Thus, I is controlled by the transistor Q-S, and for a fixed RS, the position of the slider of the potentiometer R7 varies I and thereby controls the timing of the output pulse relative to the input.

When the potential V of the base of the NPNtransistor Q-4 is above the potential V of its emitter, current flows from the collector to the emitter and charges the capacitor C The voltage V thus follows the voltage V up to approximately V at which time the capacitor C is discharged by the next pulse via the transistors Ql and Q-2 (see i in FIG. 6). The discharge brings V and hence the base of the transistor Q4 to zero, leaving the emitter of the transistor Q4 quite positive with respect to its base. Under these conditions, the transistor Q4 is cut off, and no current flows between its collector and its emitter (note i and k in FIG. 6). The condenser C discharges toward ground by means of the current I controlled by the transistor Q5, as stated above. Meanwhile, the capacitor C is being recharged toward the value V (as at i in FIG. 6). When the voltage V is once again more positive than the voltage V the base of the transistor Q4 is above its emitter, and the transistor Q4 conducts, recharging the capacitor C (see the waveforms i, j, and k in FIG. 6). This charging current flows into the transistor Q4 via its collector, which is connected to a network 25 that essentially replaces the resistor R-0 of FIG. 1. This network 25 gives an improved operation in forming the timing waveform 1 in FIG. 6, which is the equivalent of V in FIG. 2.

The network 25 comprises a PNP transistor Q6 with its base connected to the collector of the transistor Q4 and returned to the positive bus 17 through a resistor R9. The emitter of the transistor Q6 is connected to the positive bus 17, and the collector is connected to a grounded resistor RlO and to an output lead 26. When the capacitor C is discharging, the transistor Q4 is not conducting, and since there is no drop across the resistor R9, the base of the transistor Q6 is at the potential of its emitter; hence the transistor Q6 is also non-conducting, with its collector on ground. When the voltage V becomes more positive than the voltage V the transistor Q-4 conducts to recharge the capacitor C The charging current flows through the resistor R-9, making the base of the transistor Q6 more negative than its emitter, thereby causing the transistor Q-6 to conduct. The collector current of the transistor Q6 then abruptly causes the voltage of the ungrounded end of the resistor R-10 to rise toward that of the positive bus 17. This positive going voltage step across the resistor R-10 occurs when the ascending and descending ramps are equal (V =V and hence defines the desired interval (T t) and thus the desired advance t. The voltage across the resistor R-10 and hence the voltage of the output lead 26 is that indicated in the waveform 1 in FIG, 6. When the capacitor C is discharged and the capacitor C starts its rundown, the transistor Q-4 is not conducting, and there is no drop across the resistor R9; the transistor Q-6 is then not conducting, and the voltage across the resistor R-10 falls to zero. Unlike the resistor R in FIG. 1, there is no reversal in the direction of current in the resistor R-10, but the start of charging of the capacitor C is much better defined, because of the amplifying action of the transistor Q 6.

The remainder of the circuit of FIG. is not claimed as part of this invention; it is included for the sake of completeness in indicating how a blanking pulse of desired width and polarity is derived. The leading positivegoing edge of the voltage across the resistor R- defines the start of the pulse, but the width of this pulse is arbitrary, being a function of the advance t. In this example, it is desired to obtain a blanking pulse which 1) starts a time t before the leading edge of the horizontal synchronizing pulse, (2) has a desired, fixed duration, or pulse width, and (3) is negative going. Requirement (1) has already been attained, from the foregoing description.

Requirement (2), that of pulse width, may, as shown in FIG. 5, be obtained by use of the familiar monostable multivibrator (one-shot) circuit, made up of transistors Q7 and QS, a capacitor C and a resistor R and associated resistors. To trigger the one-shot, the pulse across R-10 is differentiated by a capacitor C4 and a base-return resistor R-11 of the transistor Q7. At the base of the transistor Q7 we then have the waveform m of FIG. 6, the positive spikes of which each trigger the one-shot, and these positive spikes coincide with the start of the advance interval t. The one-shot then generates a pulse such as the waveform n in FIG. 6, at the collector of the transistor Q-8. This pulse starts at the desired time, and its width is controlled by the capacitor C and the resistor R Requirement (3), the desired negative-going (from ground) of the pulse that is to be used, is obtained by using an inverter Q-9, whose emitter is on the negative bus 18 and whose collector is returned to ground by an appropriate resistor. This gives the waveform 0 at the collector of the inverter Q-9.

It may be of interest to recapitulate the comparison between the simplified functional circuit of FIG. 1 and its associated waveforms of FIGS. '2 and 3 with the specific example of a practical circuit shown in FIG. 5 with its waveforms of FIG. 6.

The positive-going pulse of FIG. 2 is generated from negative rectangular pulses by the transistor Q-1 of FIG. 5, while the switch 11 of FIG, 1 has its function performed by the transistor Q-2 of FIG. 5. The constant current network 12 of FIG. 1 is the functional equivalent of the network 22 in FIG. 5 The diode 13 corresponds to the transistor Q4, and the constant current discharge 14 is the network 24 of FIG. 5. The resistor R-0 is functionally replaced by the more sensitive network 25. The output signal V in FIG. 1 and FIG. 2 at waveform d is the signal 1 which appears on the lead 26 in FIG. 5.

Various values may be used in the components of FIG. 5. Assuming use in a horizontal blanking device, the following values may be assigned:

8 Terminal 15 +10 volts. Terminal 16 10 volts. Q-l, Q3, and Q-6 PNP silicon, e.g., 2N3638A. Q2, Q4, and QS NPN silicon, e.g., 2N3415. R1 and R-9 10,000 ohms each. R-2, R-3, R-4, and R10 1,000 ohms each. R5 and R-7 0 to 10,000 ohms (potentiometer). R-6 3,300 ohms. R-S 100,000 ohms. C-1 picofarads. C 0.001 microfarad. C 0.01 microfarad.

Equation 13 indicates that the general stability of the system is dependent to a degree on S /S. The slopes S and S are functions of their respective constant currents I and l and on the capacitors C and C Since the

generators

22 and 24 for the currents I and l are quite similar, they drift together, if they drift at all. The temperature coefiicients of C and C, can be chosen to nearly any desired degree of equality. Hence, the system is quite stable.

By a very small increase in circuit complexity, relative to the prior art, the invention achieves a great increase in usefulness. The old system required the maximum of criticality in its parameters just where the greatest accuracy was needed, i.e., at the time when the intervals are to be divided into ratios of t:T in the order of 1:100. In the old system the reference V and the slope S each had to have stability well in excess of the required accuracy, even when T was constant. In our system, we have no fixed and arbitrary reference and need require only that the two slopes S and S change in the same manner, as they do. Even if this were not achieved, the resulting error would in most places be trivial in comparison with the older method.

The horizontal blanking generator described is functionally identical to a vertical interval gate and to a noise gate; so these need not be described in further detail. It may be remarked that the required accuracies are readily achieved; noise gates, for example, can be set with an accuracy in the order of 0.25% to inhibit all ramp resets except those occurring within a fraction of a microsecond before or at the leading edge of the synchronizing pulse.

To those skilled in the art to which this invention relates, many changes in construction and widely differing embodiments and applications of the invention will suggest themselves without departing from the spirit and scope of the invention. The disclosures and the description herein are purely illustrative and are not intended to be in any sense limiting.

We claim:

1. A method for dividing each time interval T12, T T etc., of general duration T between each succeeding pair of regular pulses, P1, P2, P3, P4, etc., whether they come with absolute or only approximate regularity, into two sub-intervals (T -t) and t such that t/ T is substantially independent of T and therefore 12 12= 2s 2s= 34 a4 comprising:

storing a ratio t/ T at constant value,

measuring and storing the duration of time, T between P1 and P2,

dividing the stored duration T by the stored ratio of t/ T to give a time interval 13 initiating the time-interval at a time (T t following P2, during the interval T thereby anticipating the pulse P3 by an amount approximating 1 depending on the closeness of T to T in magnitude,

meanwhile measuring and storing the duration of time T between the pulses P2 and P3,

dividing the stored duration of time T by the stored constant ratio t/T to give the next anticipatory intervalt initiating the interval at the time (T t following P3, and

continuing the measuring, storing, dividing and initiating operations continuously for each time interval T.

2. A method for anticipating each of a series of approximately but not necessarily quite evenly spaced pulses by a short time interval bearing a fixed fractional relation to the time interval between the preceding two pulses comprising:

measuring the time interval between each two successive pulses,

storing the value of each such measured time interval during most of the next such time interval, dividing such stored value by a fixed divisor during said next such time interval, and

signaling the beginning of a short time interval equaling the resultant quotient at a time past the stored said time interval equal to the difference between said stored time interval and said quotient.

3. Given a series of events P1, P2, P3, etc., more or less regularly spaced in time, a method for dividing the intervals T12 T23, etc., between successive events, of general time T, into successive intervals (T t and t (T -4 and the intervals t 1? etc., being of general duration t, whereby the ratio t/ T is essentially independent of T so that if T varies, 1 will correspondingly change, comprising:

measuring the interval T between P1 and P2,

storing the resulting measurement, dividing it by the selected value of t/ T, generating an interval (t =T z/ T) commencing after P2 at time (T t and thereby anticipating P3 by approximately while measuring the interval T between P2 and P3 for identical storing, dividing, and generating of the next interval and performing such steps during each interval T.

4. A method for synchronizing an anticipatory pulse to regular pulses P1, P2, P3, etc., whether the regular pulses come with absolute or only approximate regularity, comprising:

measuring and storing the duration of time T between P1 and P2,

dividing T by a fixed divisor to give a short time-interval t shorter than T, and

initiating said interval 1 by a pulse coming at a time (T t) following P2.

5. The method of claim 4 wherein at each pulse the steps are repeated, for continuous synchronization.

6. A method for accurately synchronizing a controlled system with each of a series of approximately but not necessarily exactly evenly spaced pulses, comprising:

rendering said controlled system normally non-responsive to incoming signals,

rendering said controlled system responsive to incoming signals for a brief time interval following and spaced from each pulse, said time interval of responsiveness bearing a predetermined constant ratio to the between-pulse time actually elapsing between the two preceding pulses, and

initiating each said brief time interval at a time lapse from the pulse it is to follow, that, when said lapse is added to said time interval, equals said betweenpulse time between the two preceding pulses.

7. A method for accurately synchronizing a controlled system with each of a series of approximately but not quite evenly spaced pulses in the presence of noise tending to cause misfires between successive pulses, comprising:

rendering said controlled system normally non-responsive to both said pulses and said noise,

rendering said controlled system responsive after each pulse for a brief time interval constituting a predetermined fraction of the between-pulse time of the two preceding pulses, and

initiating said time interval at a set predetermined fraction of the said between-pulse time following the second of said preceding pulses.

8. A method for accurately synchronizing a controlled system with each of a series of pulses occurring at intervals of Tie, where T is a normal interval and e is the amount of tolerance from regularity for any particular pulse, in the presence of false pulses tending to produce misfires during the interval T, comprising:

rendering said controlled system normally insensitive to all pulses,

rendering said controlled system responsive to pulses for a time of T /K, where T /K 2e, and where T =the time actually between the two preceding pulses, and

initiating each said T /K interval at a time following the second of said two preceding pulses. 9. A method for synchronizing a blanking voltage pulse to a driving voltage pulse, comprising:

generating a constant-slope ramp voltage beginning at zero at each said driving pulse and cut oil at a value V at the succeeding said driving pulse,

measuring and storing the value of V between each two pulses,

dividing the stored V value by a fixed divisor and storing the resultant value V and initiating the next blanking pulse when the ramp voltage reaches V 10. Apparatus for dividing each time interval T T T etc., of general duration T between each succeeding pair of regular pulses P1, P2, P3, P4, etc., whether they come with absolute or only approximate regularity into two sub-intervals t and (T t) such that t/ T is substantially independent of T and therefore comprising:

means for storing a ratio t/ T at a constant value,

means for measuring and storing a duration of time T between P1 and P2,

means for dividing the stored duration T by the stored ratio t/ T to give the duration of the time interval 1 and means for initiating the time interval at a time (T -r following P2 and during the interval T thereby anticipating the pulse P3 by an amount approximating r depending on the closeness of T 23 to T in magnitude, while said means for measuring and storing is engaged in measuring and storing the duration of time T between the pulses P2 and P3, so that it can then be divided by said means for dividing to give the duration of the next anticipatory pulse r so that operation is continuous.

11. Apparatus for anticipating each of a series of approximately but not necessarily quite evenly spaced pulses by a short time interval having a fixed fractional relation to the time interval between the preceding two pulses, comprising:

means for measuring the time interval between each two successive pulses,

means for storing the value of each such measured time interval during most of the next such time interval, means for dividing such stored value by a fixed divisor during said next such time interval, and

means for signalling the beginning of a short time inter val equaling the resulting quotient at a time past the stored time interval equal to the difference between said stored time interval and said quotient.

12. A time interval divider which given a series of events P1, P2, P3, etc., more or less equally spaced in time, divides the intervals T T etc., between successive events, of general time T, into two successive intervals each (T 2f and I12, (T t and tag, etc., the intervals r etc., being of general duration t, whereby the ratio t/ T is essentially independent of T, so that if T varies, twill correspondingly change, comprising:

means for measuring the interval T between P1 and means for storing the resulting measurement,

means for dividing that measurement by the selected value of t/ T,

means for generating an interval T t/ T commencing after P2 at time (T -r and thereby anticipating P3 by approximately t while said means formeasuring and said means for storing are measuring the interval T between P2 and P3 for identical storing and dividing and the generating of the next interval and means for causing said means for measuring, means for storing, means for dividing, and means for generating to operate during each said interval T for continuous operation.

13. Apparatus for synchronizing an anticipatory pulse to regular pulses P1, P2, P3, etc., whether the regular pulses come with absolute or only approximate regularity, comprising:

means for measuring and storing the duration of time T between P1 and P2,

means for dividing T by a fixed divisor to give a short time-interval t shorter than T, and

means for initiating said time-interval t by a pulse coming at a time (T-t) following P2.

14. The apparatus of claim 13 having means for maintaining continuous synchronization by actuating each said means at each pulse, said means for setting said interval maintaining a fixed setting during operation.

15. Apparatus for accurately synchronizing a controlled system with each of a series of approximately but not necessarily quite evenly spaced pulses put out by a control system, comprising:

means for rendering said controlled system normally non-responsive to incoming signals,

means for rendering said controlled system responsive to incoming signals for a brief time interval following and spaced from each pulse,

means for adjusting said time interval of responsiveness to a predetermined constant ratio to the betweenpulse time actually elapsing between the two preceding pulses, and

means for initiating each said brief time interval at a time lapse from the pulse it is to follow that, when added to said time interval, equals said between-pulse time between the two preceding pulses. 16. Apparatus for accurately synchronizing a con trolled system with each of a series of approximately but not quite evenly spaced pulses in the presence of noise tending to cause misfires between successive pulses, comprising:

means for rendering said controlled system normally non-responsive to both said pulses and said noise,

means for rendering said controlled system responsive after each pulse for a brief time interval constituting a predetermined fraction of the between-pulse time of the two preceding pulses, and

means for initiating said time interval at a set predetermined fraction of the said between-pulse time following the second of said preceding pulses.

17. Apparatus for synchronizing a blanking voltage pulse to a driving voltage pulse, comprising:

means for generating a constant-slope ramp voltage beginning at zero at each said driving pulse,

means for cutting off said ramp at a value V at the succeeding said driving pulse,

means for measuring and storing the voltage of value of V between each two pulses,

means for dividing the stored V value by a fixed divisor,

means for storing the resultant voltage value V and means for initiating the next blanking pulse when the ramp voltage reaches V 18. The apparatus of claim 17 wherein said means for generating a constant-slope ramp voltage and said cut off means comprises a capacitor, constant current charge means for charging said capacitor at a constant rate, and pulse-actuated switch means for quickly discharging said capacitor at each said pulse.

19. The apparatus of claim 18 wherein said means for dividing comprises a second capacitor.

20. The apparatus of claim 19 wherein said means for measuring and storing comprises unidirectional means connecting said capacitors and passing current to said second capacitor only when it is at a lower potential than the first mentioned said capacitor, and constant discharge means for said second capacitor for discharging it at a rate slow relative to the pulse rate.

21. A time interval divider, for use with a series of approximately equal but not quite evenly spaced pulses, comprising:

a first capacitor,

a second capacitor,

charging means for applying a constant current to said first capacitor to charge it,

first discharging means at each pulse for discharging said first capacitor quickly to zero, so that said charging means charges said first capacitor at a constant slope between pulses to a value V gate means for conducting said constant current to said second capacitor only when the voltage of said first capacitor is higher than that of said second capacitor to a value approximately V second discharge means for discharging said second capacitor from approximately V at a constant current, slowly relative to said first discharge means, beginning at each pulse, and

means for obtaining a secondary anticipatory pulse when the voltage of said first capacitor equals that of said second capacitor.

22. The interval divider of claim 21 wherein said charging means comprises a power source connected to said first capacitor through a transistor having its collector connected to said first capacitor and a base connected to said power source through a resistor and to ground through a potentiometer, and an emitter connected through a resistor to said power source.

23. The interval divider of claim 21 wherein said first discharging means comprises a pair of transistors, one PNP and the other NPN connected to discharge said first capacitor through the collector of the second transistor when a said pulse is applied to the base of said first transistor, the collector of the first transistor being connected to the base of the second transistor.

24. The interval divider of claim 21 wherein said gate means comprises an NPN transistor with its base connected to said first capacitor and its emitter connected to said second capacitor.

25. The interval divider of claim 24 wherein said means for obtaining a secondary anticipatory pulse comprises a PNP transistor with its base connected to the collector of the aforesaid NPN transistor, and, through a resistor to a plus voltage, the emitter of the PNP transistor being connected to the plus voltage through a resistor and its collector being connected to an output lead and to a grounded resistor.

26. The interval divider of claim 21 wherein said second discharge means comprises an NPN transistor with its collector connected to said second capacitor, its emitter connected through a resistor to a negative potential, and its base connected to a voltage divider between ground and said negative potential.

27. Apparatus for providing and timing the initiation of a small time interval within a larger time interval as a set portion of said larger time interval regardless of minor variation of said larger interval, comprising:

a source of constant current,

a first condenser connected between said source and ground, so as to be charged thereby, during a time interval,

a switch connecting the source side of said first condenser to ground when said switch is closed, thereby reducing the first condenser voltage to zero,

means for closing said switch at the end of each said time interval,

a diode means connected to the source side of said first condenser,

a second condenser connected between said diode and ground, said diode passing current from said source to charge said second condenser when the voltage on said first condenser exceeds that on said second condenser,

a constant-discharge means connected to the diode side of said second condenser for discharging said second condenser slowly when said voltage on said first condenser is lower than that on said second condenser, and

means to generate an output pulse each time said second condenser begins charging, to initiate said small time interval, said small time interval enduring during the interval of charge of said second condenser.

28. A method of dividing each time interval T T T etc., of general duration T between each succeeding pair of regular pulses P1, P2, P3, P4 etc., whether they come with absolute or only approximate regularity, into two sub-intervals (Tt) and t such that t/ T is substantially independent of T and therefore t T =t T t /T comprising:

measuring the duration of time, T between P1 and P2 and storing a quantity representative thereof, dividing the stored quantity by a constant ratio of t/T to give a second quantity representing interval iz,

initiating the time-interval t in response to said second quantity at a time (T =t following P2, during the interval T thereby anticipating the pulse P3 by an amount approximating t depending on the closeness of T to T in magnitude,

meanwhile measuring the duration of time T between the pulses P2 and P3 and storing another quantity representative thereof,

dividing the stored constant ratio t/ T to give a further quantity representing the next anticipatory interval '23, initiating the interval t at the time (T =t following P3 in response to said further quantity and continuing the measuring, storing, dividing, and initiating operations continuously for each time interval T.

29. A method for anticipating each of a series of approximately but not necessarily quite evenly spaced pulses by a short time interval bearing a fixed fractional relation to the time interval between the preceding two pulses comprising:

measuring the" time interval between each two successive pulses,

storing a quantity representing the value of each such measured time interval during most of the next such time interval,

dividing such stored quantity by a fixed divisor during said next such time interval, and

signaling the beginning of a short time interval equaling the time represented by the resultant quotient at a time past the start of the next said time interval equal to the difference between the amount representing said stored time interval and said quotient.

30. In the occurrence of a series of events P1, P2, P3, etc., more or less regularly spaced in time, a method for dividing the intervals T T etc., between successive events, of general time T, into successive intervals (T in) and tn, (T =t2 and I23, the intervals [12, [23, etc., being of general duration t, whereby the ratio t/ T is essentially independent of T so that if T varies, I will correspondingly change, comprising:

' measuring the interval T between P1 and P2,

storing a quantity representing the resulting measurement,

dividing the quantity by the selected value of 1/ T,

generating an interval represented by the amount (t =T 2 t/ T commencing after P2 at time (T t and thereby anticipating P3 by approximately Z12,

while measuring the interval T between P2 and P3 for identical storing, dividing, and generating of the next interval t and performing such steps during each interval T.

31. A method for synchronizing an anticipatory pulse to regular pulses P1, P2, P3, etc., whether the regular pulses come with absolute or only approximately regularity, comprising:

measuring and storing a quantity representing the duration of time Tbetween P1 and P2, dividing the quantity representing T by a fixed divisor to give a quantity representing a short time-internal t shorter than T, and

initiating said interval t by a pulse coming at a time represented by the quantity representing (Tt) following P2. 32. The method of claim 31 wherein at each pulse the steps are repeated, for continuous synchronization.

33. A method for synchronizing a blanking voltage pulse to a driving voltage pulse, comprising:

generating a constant-slope ramp voltage beginning at zero at each said driving pulse and cutoff at a value V at the succeeding said driving pulse,

storing a voltage having the value V between each two pulses,

dividing the stored voltage V by a fixed divisor and storing the resultant voltage V and initiating the next blanking pulse when the ramp voltage reaches a value of V 34. A method anticipating the arrival of each of a plurality of pulses spaced by an interval T which may not be constant, at a time t prior to the occurrence of each pulse, comprising:

(1) varying a first voltage from a reference level at a preselected rate following each pulse, to establish a voltage value indicative of the time interval to the following pulse,

(2) varying a second voltage at a preselected rate following each pulse, from a value indicative of the prior interval,

(3) comparing the respective instantaneous values of such voltages, and

(4) providing an output when the ratio of the voltages reaches a predetermined value.

35. A system for providing an output at a predetermined time following the occurrence of each of a plurality of spaced timing pulses comprising:

(1) means responsive to each pulse :for deriving a voltage varying in amplitude from a reference level at a predetermined rate until the occurrence of the following pulse, and storing such voltage;

(2) means rendered operative after each pulse for reducing the amplitude of the stored voltage at a different rate,

(3) means for comparing the instantaneous values of such voltages,

(4) and means for producing an output when a predetermined ratio of said voltages is reached.

36. In a synchronizing system controlled by spaced 15 16 timing pulses which may occur at uneven intervals, pulse when the voltage of the second storing means means for producing an output at a time following a reaches a predetermined fraction of the voltage timing pulse which is a predetermined fraction of the prestored in the first storing means.

vious interval comprising:

(1) means for storing a voltage having an amplitude References Cited indicative of the time interval between each twO 5 UNITED STATES PATENTS timing Pulses 2 976 487 3/1961 Cohen 324-68 (2.) a second voltage storing means for momentarily 3395293 7/1968 Perlofi 228 storing at the start of one interval, a voltage equal to that representing the previous interval, 10 ROBERT L GRIFFIN Primary Examiner (3) means rendered operative by the occurrence of each timing pulse for initiating discharge of the sec- RICHARDSON, Asslstant EXaInlner ond storing means,

(4) and means for rendering the output producing means operative during the interval following the 15 178-7.l; 307-269; 32468; 328129