KR100873008B1 - Measurement device of frequency offset of atomic clock and control method - Google Patents

Abstract

A device for measuring a frequency offset of an atomic clock and a control method thereof are provided to implement a stable synchronization system by using the method to add the weight to the measured current frequency offset value. A device for measuring a frequency offset of an atomic clock(100) includes a phase comparator(10), a defective data deleting unit, an average calculator(30), a filter(40), a phase change value setting unit, a frequency offset calculation unit(60), a weight value calculation and application unit, a control signal output unit(80), and a frequency controller(90). The phase comparator compares the phase difference of the second pulse outputted from the frequency controller and the reference second pulse. The defective data deleting unit deletes the defective data among the data outputted from the phase comparator. The average calculator takes an average of the data outputted from the defective data deleting unit. The filter performs filtering of the averaged data by the line fitting method. The phase change value setting unit determines whether the phase of the signal outputted from the filter is changed or not. The frequency offset calculator calculates the frequency offset based on the data outputted from the phase change value setting unit. The weight value calculation and application unit determines whether the weighted value is applied to the frequency offset data. The output unit outputs the data with the calculated weight. The frequency controller outputs the second pulse based on a control signal outputted from the control signal output unit.

Description

Measurement device and control method of atomic offset of atomic clock {Measurement device of frequency offset of atomic clock and control method}

1 is a conceptual diagram of frequency offset measurement and control for a conventional atomic clock.

Figure 2 is a block diagram of a frequency offset measuring apparatus for the atomic clock according to the present invention.

Figure 3 is a block diagram showing the configuration of the abnormal data remover according to the present invention.

4 is a block diagram showing a configuration of a phase change value setting unit according to the present invention;

Figure 5 is a block diagram showing the configuration of the weight calculation and application unit according to the present invention.

6 is a procedure showing a process of measuring and controlling a frequency offset according to the present invention.

<Explanation of Signs of Major Parts of Drawings>

10: Phase comparator,

20: abnormal data removal unit,

21: phase inversion detector,

22: abnormal data detector,

23: output data generator,

30: average calculator,

40: filter,

50: phase change value setting section,

51: reference phase value setter,

52: phase change value calculator,

53: threshold comparator,

60: frequency offset calculator,

70: weight calculation and application,

71: offset inversion detector,

72: weight generator,

73; step setter.

The present invention relates to a frequency offset measuring apparatus and a control method of an atomic clock, and more particularly, to compare a phase and a high pulse accurately synchronized to a reference super pulse coming into the input, and to process the phase comparison data to obtain an appropriate frequency offset value. The present invention relates to a frequency offset measuring apparatus and a control method for providing an output pulse with high accuracy synchronized with an input pulse.

An atomic clock is a clock that uses the resonance frequency of atoms to maintain time with extreme accuracy. Electronic components of atomic clocks are controlled by the frequency of microwaves, which are electromagnetic plates that are emitted or absorbed by the transition of atoms or molecules. In atomic clocks, these quantum transitions produce very regular electromagnetic waves, which, like other phenomena that occur periodically in clocks, count the waves and measure time.

Atomic clocks are widely used to reliably provide the frequencies that are the basis for building high-precision time syncronization systems. These atomic clocks provide high frequency stability and, when synchronized to an external reference time, provide synchronous accuracy of tens of nanoseconds per day. However, because atomic clocks are not ideal, there is a frequency offset, and a technique for measuring and compensating for the frequency offset of the atomic clock is required in order to continuously maintain synchronization within several nanoseconds for a long time.

1 shows a conceptual diagram of a conventional technique for measuring and controlling the frequency offset of an atomic clock. The phase comparator 10 is a device for comparing the phase of the reference superpulse with the output superpulse of the frequency controller 90 to calculate a difference value between the two values. The average calculator 30 outputs average data according to a predetermined number of measured data. However, when a predetermined number is not large when an incorrect data value is input, the average calculator 30 outputs an incorrect frequency offset. In other words, the output of this wrong frequency offset has errors caused by bit error and phase inversion. Since such bit errors occur very rarely and do not occur consecutively, the problem is neglected. However, the error that occurs when the inversion of the whistle occurs because the time interval counter (not shown) in the phase comparator 10 only shows a positive value and does not represent a negative value.

Various methods may be used for the filtering method. In the conventional filtering method, a line fitting technique is used. [Equation 1] shows the equation for calculating the frequency offset using the visual variation value obtained by the line fitting.

From here,

T: measurement time,

ΔT: amount of time change during measurement time,

f _norm : input reference frequency,

Δf: Frequency offset value to be obtained.

The measurement time is the total time used to generate the data to be used for line fitting. The frequency offset value obtained as described above is sent to the frequency controller 90 to add the value corresponding to the frequency controller 90 to change the value of the output frequency. The frequency offset value is measured continuously and the value input to the frequency controller 90 is added to the previous frequency offset value to exit. If the frequency offset value of the atomic clock 100 is not changed over time and is constant, high-precision synchronization is achieved by applying the same weight to the frequency controller and equally weighting the previously measured frequency offset value and the current frequency offset value. Can be.

However, in the actual measurement, the frequency offset measurement value of the atomic clock 100 is changed every measurement time, and the change amount is not constant but has a random value. Therefore, when the frequency offset value is added to the frequency controller 90 by simply adding the previously measured value and the currently measured value, the frequency offset value of the atomic clock 100 which is changed by the accumulated and calculated frequency offset value is immediately changed. It cannot be reflected and compensated quickly. If the cumulative previous frequency offset value is opposite to the currently measured frequency offset sign and is larger, there is a problem in that the synchronization accuracy is lowered by applying the frequency offset value in a direction in which the synchronization error increases.

Therefore, even if the current frequency offset value is measured within the error range, a problem arises in that the error of the frequency offset value input to the frequency controller 90 is increased by the process of adding the previous frequency offset.

The present invention has been made to solve such a problem, and an object of the present invention is to achieve high-precision visual synchronization by a method of removing such abnormal data when the data output using the phase comparator includes the abnormal data. .

Another object of the present invention is to determine the phase change state of the input super pulse by the phase change value setting unit and to control it, and to calculate the frequency offset value of the atomic clock by weighting the current frequency offset value stably high precision It is to achieve visual motivation.

Other objects, specific advantages and novel features of the present invention will become more apparent from the following detailed description and the preferred embodiments associated with the accompanying drawings.

The apparatus for measuring frequency offset of an atomic clock according to the present invention includes a phase comparator for comparing a phase difference between a reference pulse and a superpulse output from a frequency controller, an abnormal data remover for removing abnormal data from data output from the phase comparator, Set an average calculator for calculating the average of the data output from the abnormal data removal unit, a filter for filtering the averaged data by the line fitting method, and a phase change value for judging whether the signal from the filter is in phase change. And a frequency offset calculator for calculating a frequency offset based on the data from the phase change value setting unit, a weight calculation and application unit for determining whether or not to apply weights to the frequency offset data, and the weighted data. On the basis of the signal from the control signal output unit for outputting the control signal output unit And a frequency controller for outputting super pulses.

2 is a block diagram showing a frequency offset measuring apparatus for the atomic clock 100 according to the present invention. Referring to FIG. 1, the abnormal data removing unit 20 is added before the average calculator 30, the phase change value setting unit 50 is added after the filter 40, and the frequency offset calculator ( 60, it can be seen that the weight calculation and application unit 70 is added next.

In brief, the abnormal data removal unit 20 is a device for deleting data determined as abnormal data through the phase comparator 10. If the abnormal data removal is not performed, an error occurs in the calculation of the average as described above, which causes an error in the frequency offset calculation. The average (m) is calculated by the average calculator 30 with respect to the output data passed through the abnormal data removing unit 20.

The averaged output data is input to the filter 40 described above, and filtering is performed by the line fitting technique. Data passing through the filter 13 is input to the phase change value setting unit 50. The phase change value setting unit 50 determines whether the phase change is performed and accordingly, the phase change value or the value of 0. It will go through the process of outputting. The frequency offset Δf is calculated by the frequency offset calculator 60 using Equation 1 on the data output from the phase change value setting unit 50.

In addition, the weight calculation and application unit 70 generates different outputs according to whether or not frequency offset inversion occurs. Using this, when the frequency offset is inverted, it is possible to output the frequency offset value in response to the change of the current frequency offset value quickly by weighting the current frequency offset value.

3 shows a configuration of the abnormal data removal unit 20 according to the present invention. When the data of the phase comparator 10 enters the input, the phase inversion detector 21 first detects whether or not the phase is inverted. The reference point of phase inversion becomes the initial measured initial phase value and can be judged by comparing whether the subsequent data deviated from this value more than 0.5 second. Next, the abnormal data detection unit 22 detects whether the collected data contains data that is difficult to see with the data values normally measured. Anomaly data refers to data having extreme values that are very large or small compared to normal data. The most widely used method for detecting abnormal data includes a median absolute deviation (MAD) measurement, which is defined as shown in [Equation 2].

From here,

m: Median [y (i)],

Median [x]: median of x values,

0.6745 is the standard deviation of the data with a normal distribution.

Normally, a value that is five times smaller or larger than the MAD value is represented as the abnormal data. When the abnormal data detector 22 passes, the abnormal data can be determined. The type of abnormal data is the same as the above-described processing in the case where an error occurs in the bit and is output as a value different from the original value, and the following processing is performed for the second error. That is, since the time interval counter (not shown) which is generally used in the phase comparator 10 shows only a positive value and does not represent a negative value, when phase inversion occurs, the actual phase data value is -0.5 seconds to 0.5. It has changed in the range of seconds, and the measured value is measured to have changed by 0.5 seconds or more. For example, if the first measured value is 0.9 seconds and this value continues to increase while the frequency offset is corrected, and the phase difference between the two pulses increases by 0.15 seconds, the actual phase difference value is 1.05 seconds, whereas the measured phase difference value is Is 0.05 seconds. When the frequency offset is calculated based on the measured data, the frequency offset is actually increased but the frequency offset is decreased. Therefore, an incorrect control value is applied, and this phenomenon occurs repeatedly, so that no control is possible. A method of compensating for 0.9 seconds of data by the action of the phase inversion detector 21 is used. That is, when the synchronization boundary value after synchronization with reference input pulse is T _SyncBound , the phase difference value measured through phase comparison is T _SyncBound. Or 1-T _SyncBound . The synchronization accuracy for the reference superpulse varies with the performance of the frequency controller 90, but typically achieves initial synchronization within hundreds of nanoseconds. Therefore, the initial phase difference value output from the phase comparator 10 may be as small as 1 to several hundred nanoseconds or several hundred nanoseconds. The algorithm for removing abnormal data using this method is as follows.

Case 1: Phase inversion does not occur and abnormal data does not occur

y _out = y _in

Case 2: Phase inversion does not occur and abnormal data occurs

y _out = m

Case 3: Phase inversion occurs and no abnormal data occurs

Y _out if initial phase value is more than 0.5 = 1+ y _in

Y _out if initial phase value is less than 0.5 = y _in -1

Case 4: Phase Inversion Occurs and Abnormal Data Occurs

Y _out if initial phase value is more than 0.5 = 1+ m

Y _out if initial phase value is less than 0.5 = m-1

Y _out in the above algorithm And y _in are the output and input of the abnormal data remover 20, respectively, and m is the median. Through this process, the output data generator 23 determines whether the input data is inverted in phase, determines whether it is abnormal data, compensates for the output, and then outputs the data.

4 shows a phase change value setting unit 50 according to the present invention. As described above, when the data filtered by the least-squares line fitting technique is input, the reference phase value setter 51 sets a phase value to be a reference. The reference phase value at this time is programmed to be set to the middle value of the first line fitted data. Next, the phase change calculator 52 calculates a difference between the reference phase value and the last value of the line-fitted data using the output of the line-fitted data from the second time onwards and sets it as the phase-change value. The threshold comparator 53 is provided at the rear end of the phase change calculator 52. The threshold comparator 53 outputs a phase change value when the absolute value of the phase change value is larger than the threshold value using a predetermined threshold value, and outputs a value of 0 when the threshold value is smaller than the threshold value.

5 is a block diagram showing the configuration of the weight calculation and application unit 70 according to the present invention. As shown in FIG. 5, the data output from the frequency offset calculator 60 is a current frequency offset value. When this current frequency offset value is output, it is first input to the offset inversion detector 71. The offset inversion detector 71 compares the sign of the previous frequency offset value with the current frequency offset value and detects whether the frequency offset value is inverted. Next, the weight generator 72 applies different weight values to the following two cases.

Case 1: Frequency offset inversion does not occur.

Δf _out = Δf _Current + Δf _Previous

Case 2: Frequency offset inversion occurs.

Δf _out = [a × Δf _Current + Δf _Previous ] / a, a> 1

Here, Δf _{Current and} Δf _Previous are the output current and previous frequency offset values, respectively, and a is a real value greater than one. By using this, when the frequency offset is inverted, the current frequency offset value can be weighted to quickly reflect the changing trend of the current frequency offset value. Finally, the step setter 73 adjusts the frequency offset value finally provided as an input to the frequency controller 90 using Equation 3 below.

Δf _final = N × max [R _spec , R _theory ]

here,

이고,

ego,

max [x, y] means the larger of x and y, and min [x, y] means the smaller of x and y. Also, [x] means the largest integer smaller than x. R _spec represents the frequency offset resolution described in the specification of the frequency controller 90, and R _theory represents the frequency offset resolution required for the synchronization system. Δf _specMax is the maximum adjustable frequency offset shown in the specification of the frequency controller, and Δf _theoryMax is the maximum frequency offset possible when the atomic clock is operating normally.

,

,

b> 1

Where σ _spec is the frequency stability value given in the oscillator specification, and T _Meas Is the data measurement time used to find the frequency offset value, E _spec Is the error value specified in the oscillator specification, and b is a real number greater than 1.

As such, a method of weighting the current frequency offset value by adjusting the output of the frequency offset value according to whether or not the frequency offset is inverted is possible.

6 is a flowchart illustrating a process for measuring a frequency offset of the atomic clock 100 and controlling the output of the frequency offset according to the present invention. As shown in FIG. 6, first, the abnormal data is removed by comparing the phase of the reference pulse and the output pulse emitted from the frequency controller 90 (S10). During the process of removing such anomaly data, it is necessary to first determine whether the input data has been phase-inverted by placing the phase inversion detector 21, so that accurate frequency offset calculation is possible as described above.

As described above, compensation for the abnormal data is performed along with the determination of phase inversion and the frequency offset measurement can be precisely performed. As such, the data compensated through the abnormal data removal unit 20 passes through the average calculator 30 to calculate the median value m. As described above, the filtering is performed by the line fitting technique in the situation where the intermediate value m is calculated (S20).

The filtered data is first used to set the reference phase value. After the reference phase value is set in this way, the data input to the phase change value setting unit 50 thereafter calculates an output for the phase change in the threshold comparator 53 according to the phase change. That is, the threshold comparator 53 outputs data of 0 or outputs a phase change value with a reference for phase change (S30).

In the frequency offset calculator 60, the frequency offset value is calculated for the data in consideration of the phase change value. When the frequency offset value is input to the weight calculation and application unit 70, the current data is judged in advance whether the frequency offset value is inverted. Judging whether the frequency offset value is inverted is because the frequency offset value of the atomic clock changes with time. That is, in order to minimize the synchronization error that may occur when the current frequency offset value is opposite to the previous frequency offset value and larger, a method of weighting the current frequency offset value is used (S40).

Finally, when the control signal is output and the frequency offset value is adjusted according to the data in which the weight is reflected, it is possible to provide an output pulse pulse that is highly precisely synchronized with the reference pulse pulse coming into the input (S50).

The present invention effectively removes abnormal data generated during measurement in the implementation of a synchronization system requiring high-precision visual synchronization. In addition, a stable synchronization system can be constructed by effectively compensating for the change in the current frequency offset by using a method of weighting the measured current frequency offset value.

In particular, in the construction of a time synchronization system requiring ultra-precision time synchronization of several nanoseconds or less, it can be usefully used in an application for continuously maintaining stable synchronization with an external reference signal.

Claims (9)

In the frequency offset measuring device of the atomic clock 100, A phase comparator 10 for comparing the phase difference between the reference superpulse and the superpulse output from the frequency controller 90; An abnormal data removal unit 20 for removing abnormal data from data output from the phase comparator 10; An average calculator 30 for obtaining an average of data output from the abnormal data removing unit 20; A filter 40 for filtering the averaged data by a line fitting method; A phase change value setting unit (50) for determining whether the signal from the filter (40) is a phase change; A frequency offset calculator (60) for calculating a frequency offset based on the data from the phase change value setting unit (50); A weight calculation and applying unit (70) for determining whether to apply weights to the frequency offset data; A control signal output unit 80 for outputting the weighted data; and And a frequency controller (90) for outputting a superpulse based on the control signal output from the control signal output unit (80). The method of claim 1, The abnormal data removal unit 20, A phase inversion detector 21 for determining whether or not the phase is inverted; An abnormal data detector 22 for determining whether abnormal data is present after determining whether the phase reversal is performed; and And an output data generator (23) for outputting the data output from the abnormal data detector (22). The method of claim 1, The phase change value setting unit 50, A reference phase value setter 51 for setting a phase value serving as a reference to data output from the filter 40; A phase change value calculator 52 for calculating a difference between the reference phase value set by the reference phase value setter 51 and the data output from the filter 40; and And a threshold comparator (53) for determining the output data by comparing the phase change value with a predetermined threshold value. The method of claim 1, The weight calculation and application unit 70, An offset inversion detector 71 for comparing a sign of a current frequency offset value with a previous frequency offset value; A weight generator 72 that determines whether weights are generated according to the determination of the offset inversion detector; and And a step setter (73) for adjusting a frequency offset value by reflecting the output of the weight generator (72). In the frequency offset control method of the atomic clock, (a) comparing the phase of the output superpulse output from the reference superpulse with the frequency controller 90 and removing abnormal data (S10); (b) obtaining an average of the output data from which the abnormal data is removed and performing filtering by a line fitting method (S20); (c) performing a comparison process between the calculated phase change value and a predetermined threshold value on the filtered data (S30); (d) calculating weights based on the change of the current frequency offset value with respect to the output data on which the comparison process is performed (S40); and and (e) adjusting the frequency offset value by outputting a control signal according to the output data reflecting the weight (S50). The method of claim 5, The step of removing the abnormal data (S10) further comprises the step of determining the phase inversion before removing the abnormal data, characterized in that the frequency offset control method of the atomic clock. delete The method of claim 5, In the step S30 of comparing the calculated phase change value with a predetermined threshold value, a reference phase value is set using line-fitted data, and when the calculated phase change value is larger than the threshold value, the reference phase value is set. And using a phase change value and setting the phase change value to zero when the calculated phase change value is smaller than the threshold value. The method of claim 5, In the calculating of the weight by reflecting the change of the current frequency offset value (S40), the frequency offset of the atomic clock is set by giving a higher weight to the current frequency offset value than the previous frequency offset value. Control method.

Images