Summary of the Invention
The present invention provides apparatus which can be used on-board a satellite to correct long-term crystal oscillator frequency drift. Typical high quality crystal oscillators drift, the drift being on the order of 1 part per 10.sup.-.sup.10 per day or less. The present invention corrects for this drift by establishing a drift rate based on the drift history. The invention also functions as a frequency translator to offset the frequency of oscillation of the source signal by a predetermined amount. The invention also removes any frequency bias in source signal. In the current application of this invention the offset amount is settable in the band of - 84.48 parts per million (PPM) to - 145.51 PPM when the source signal has a 5.0 MHz .+-. 5 Hz oscillation frequency.
The subject time and frequency control unit employs two counter/holding-register pairs which function as coarse and vernier frequency controls. Since the probability of crystal drift is known, pulses are clocked into the vernier counter at a specified rate and when an end point is reached the coarse counter will be incremented. The increments of the coarse control are approximately equal to 0.04 Hz with each increment having about 7000 divisions in the vernier control. The vernier end points are called roll-over numbers and the coarse control will be counted up or down depending on which end point the vernier control has traversed. Roll-over gates are supplied to automatically insert new roll-over numbers into the vernier control and also to gate the proper update pulse to the coarse control. By utilizing roll-over numbers and roll-over gates the invention provides a continuity of time and frequency control over the entire life of the satellite.
The time and frequency control unit may be controlled from the ground or from an on-board computer. A slope correction rate generator is provided by the invention to automatically update the coarse/vernier control for slope rate correction once a drift rate has been established. The rate generator provides increments for drift rate correction in a manner similar to that of the on-board computer.
Brief Description of the Drawings
FIG. 1 is a block diagram of the time and frequency control unit incorporating the principles of the invention;
FIG. 2 is a graph illustrating the comparative effects on frequency produced by the coarse control and vernier control;
FIG. 3 is a block diagram of an analog single sideband generator used as a mixer in the present invention;
FIG. 4 is a block diagram of a digital voltage controlled oscillator (VCO), used as a mixer in the present invention;
FIG. 5 is a graph showing frequency vs time, particularly showing the drift of input frequency f.sub.o, and the stabilized value of output frequency, f.sub.f ;
FIG. 6 is a graph of resolution versus register numbers.
Description of the Preferred Embodiment
Referring to FIG. 1, an input 2 is shown carrying a signal of frequency f.sub.o. In the preferred embodiment f.sub.o should be 5 MHz. The signal splits, with one path of it entering a conventional A counter 4 which divides the frequency f.sub.o by a modulus number a. The output from the A counter 4 meets the other path of the f.sub.o signal at a first mixer 6. The mixer 6 produces a signal of frequency ##EQU1## The frequency of this output is referred to as f.sub.1. The f.sub.1 signal enters a convention B counter 10 where it is divided by a second modulus number b before it is mixed in a second mixer 12 with the original f.sub.o signal. The output of the second mixer 12 is equal to ##EQU2## In expanded form this mixed output is equivalent to ##EQU3## It is apparent from this expression that a change in a is of much less significance than a change in b. The A counter 4 which divides by a is a vernier (or fine) control element; the B counter 10 is a coarse, translational control element. The distinction between the two control means is shown in FIG. 2. Referring again to FIG. 1, the mixed output from the second mixer 12 is passed through a narrow band output filter 14. The output from output filter 14 is the output of the time and frequency control unit. The system is designed to have a very carefully balanced analog single sideband mixer with the capability of suppressing undesired sidebands well over 40 db before filtering.
FIG. 3 is a block diagram of the second mixer 12 which is characterized in the preferred embodiment as a single sideband generator. FIG. 3 shows two outputs from B counter 10 going to an integrator controller 16 and a first integrator 18 respectively. The output from first integrator 18 is in the form of a triangular wave, i.e., the integration of the pulse coming from the B counter 10. The output from the first integrator 18 enters a zero crossing detector 20 which pulses whenever the leading edge of the integrated output from the first integrator 18 crosses the zero amplitude level. The pulse from the zero crossing detector 20 is integrated by a second integrator 22 which puts out a triangular wave similar to that of the first integrator 18 except that the output from the second integrator 22 is 90.degree. out of phase with the output from the first integrator 18. The outputs from the first integrator 18 and the second integrator 22 are truncated by truncator 24 and truncator 26, respectively, and then enter modulator 28 and modulator 30, respectively. The 5 MHz (f.sub.o) signal is shown entering a quadrature generator 32 which produces two signals which are out-of-phase by 90.degree.; one lead-signal 34 is shifted to lead in phase by 45.degree., the other a lag signal 36 to lag in phase 45.degree.. Lead signal 34 enters modulator 28 together with the output from truncator 24. Lag signal 36 enters modulator 30, together with the output from truncator 26, to yield a modulated signal similar to, but of different phase from, the signal from modulator 28. The modulated signals from modulator 28 and modulator 30 are summed in summer 38 to produce the single sideband (SSB) output which is then fed into the narrow band filter 14 as shown in FIG. 1. The SSB output mixer 12 is independent of frequency and amplitude. The raw output has no more than -35 dB unwanted sidebands harmonically related to the audio frequencies; the 5 MHz (f.sub.o) carrier is down by 50 dB.
In the present embodiment, a digital phase-lock loop is used as the first mixer 6. As seen in FIG. 4, the frequency f.sub.o enters a digital mixer 40 which transmits an output of frequency f.sub.o /a into a digital phase detector 42. The digital phase detector 42 pulses a switching circuit 44 which inputs to a crystal voltage controlled oscillator (VCO) 46. The output from the crystal VCO 46 is in the form of pulses carrying a signal indicating a frequency of ##EQU4## The "loop" of the phase-lock loop is completed by returning the output from the crystal VCO 46 back into the digital mixer 40 which mixes incoming f.sub.o with ##EQU5## to produce ##EQU6## or f.sub.o /a which is fed into digital phase detector 42 and so on.
Although the present embodiment includes a digital mixer stage for first mixer 6 it is to be noted that the analog single sideband generator stage of second mixer 12 could be used as the first mixer 6 as well as the second mixer 12. The analog mixer is considerably more sensitive than the digital VCO mixer stage described; thus, although the analog mixer could be substituted for the digital VCO mixer, the converse is not true.
FIG. 5 shows the gradual drift of f.sub.o over a time period such as 100 days. By altering the value of a and/or b it can been seen from the foregoing that final output frequency f.sub.f, can be held constant. Referring again to FIG. 1, in order to alter the values of a and b in the A counter 4 and B counter 10, respectively, an A holding register 50 which holds the value of a for A counter 4 and a B holding register 52 for holding the value of b to be used by B counter 10 are provided.
The values held in the A register 50 and B register 52 are controlled by two alternative sources. Which source is selected at any given time is determined by remote command source 57 which switches between sources as required. The first is an on-board computer 54 which clocks on clock line 70 at about 400 kHz and informs the A register 50 and B register 52 to increment or decrement on up/down line 72. The on-board computer 54 can control frequency to a resolution of <3.5 parts in 10.sup.-.sup.12 for each computer clock to compensate for a change in f.sub.o. The other alternative controller source is the slope correction rate generator 64 with granularity of 38 sec. minimum. Upon receipt of transmitted time markings from the satellite via conventional telemetry means, remote command source 57 communicates with a command data word holding register 60 via a command data line 56 and a command clock line 58 which operate in a conventional communication fashion. The data word communicated into the command data word holding register 60 includes data to be transferred into the A register 50, the B register 52, or a drift rate holding register 62. The first two bits of the data word indicate which register is to receive data from the command data word holding register 60. A transfer selection circuit 64 actually implements this decision by turning on the appropriate register to receive data when the proper two bit code is sensed on bit 1 line and bit 2 line.
The slope rate holding register 62 is an updated register which represents the drift slope of frequency over time as shown in FIG. 5. Again referring to FIG. 1, associated with the slope rate holding register 62 is a slope correction rate generator 65 which receives updated data from the slope rate holding register 62 and is turned on by the transfer selection circuit 64 and output from B counter 10. The output from the slope correction rate generator 65 is gated into OR gate 66 with the clock line 70 of the on-board computer 54 and into OR gate 68 with an up/down line 72 from the on-board computer 54. Outputs from OR gate 66 and OR gate 68 go to both the A holding register 50 and B holding register 52 to alter the values of a and b contained therein, respectively.
In order to preserve the continuity of the changing of a and b, the following scheme is employed. The value for a in A holding register 50 is fixed between a low-point and high-point determined by a gate-overflow circuit 70. For example, the range for a may be set as 6300 to 13,469. When a in the A holding register 50 is about to traverse 6300 (in a decrementing fashion) or 13,469 (in an incrementing fashion) gate-overflow circuit 70 inputs a signal to the B holding register 52 to change the value of b by one count and rolls a to the other end point. If b were 11,837 and a reached +13,469 (by incrementing), a signal would inform B holding register 52 to increment b by one to 11,838 while a simultaneous signal to roll-over gate 72 would be transmitted to reset the A holding register 50 value of a to 6300; a changes a maximum of about 7000 counts (in the above example) before b is changed by one count. Thus, when b is changed by a count, a is reset to a new value which offsets the large change effected by the change of b. The roll-over gating of the A register 50 acts with the changing of the value of b to provide continuity in the frequency control.
FIG. 6 shows the performance of the system in graphical form. The lower band showing the on-board computer control means (represented in FIG. 1 by numeral 54) indicates how an increment of a affects the frequency resolution on the order of 10.sup.-.sup.15 . The larger granularity of the slope rate controlled network is shown in the upper band. The graph of FIG. 6 is derived from the frequency relationships previously discussed where F = f.sub.f and .DELTA.F is the offset term ##EQU7##
It is believed apparent that the invention may be practiced other than as specifically detailed herein without departing from the scope of the invention as defined by the following claims.