A method for measuring the clock error of a frequency shift repeater

By constructing a dual-state measurement method and utilizing edge capture with different local oscillator paths and a shared counting reference, the problem of difficult-to-distinguish clock error coupling in frequency-shift repeaters was solved, achieving low-cost and high-precision clock error measurement and quantifying the error contributions of near-end and far-end clocks.

CN123113273APending Publication Date: 2026-11-17SHENZHEN KAIYIXIN TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611306532.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively distinguish and quantify the error contributions of near-end and far-end clocks in frequency-shift repeaters, leading to a contradiction between the high-cost, high-precision measurement requirements and low-cost mass deployment. Furthermore, traditional methods are difficult to accurately measure minute frequency differences.

Method used

The first and second measurement states are constructed. The continuous phase changes of the input-side reference signal and the frequency shift link output signal are obtained through different local oscillator paths. The error contribution coefficients of the near-end and far-end clocks are determined by using a dual-channel edge capture method with a shared counting reference. The error contribution of each clock is obtained by solving the system of equations simultaneously.

Benefits of technology

It achieves high-precision measurement of clock errors in frequency-shift repeaters with low-cost hardware, and can quantify the error contributions of near-end and far-end clocks separately, overcoming the problem of difficult-to-distinguish error coupling in traditional methods, and providing an engineering implementation path for high-precision measurement.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113273A_ABST
    Figure CN123113273A_ABST
Patent Text Reader

Abstract

This invention proposes a method for measuring clock error in a frequency-shift repeater. The method includes: constructing a first measurement state and a second measurement state. The first measurement state uses a near-end low-side local oscillator path, and the second measurement state uses a near-end high-side local oscillator path. The nominal frequency conversion results of the two states are consistent, but the clock error contribution relationships are different. The method further involves determining the near-end and far-end clock error contribution coefficients for each measurement state; acquiring the continuous phase change between the input-side reference signal and the frequency-shift link output signal in each measurement state; and determining the relative errors of the near-end and far-end base clocks based on the continuous phase changes and corresponding clock error contribution coefficients for the two states. This method achieves high-precision clock error measurement using low-cost hardware, can quantify the error contributions of the near-end and far-end clocks separately, and obtains the residual frequency offset of a normal frequency-shift link.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of communication technology, and in particular relates to a method for measuring the clock error of a frequency shift repeater. Background Technology

[0002] Frequency-shift repeaters typically consist of a near-end device and a far-end device. The near-end device uses its local base clock to generate a local oscillator signal via a frequency synthesizer, converting the received first operating frequency into a relay frequency suitable for near-end and far-end transmission. The far-end device then uses its own base clock to generate a recovered local oscillator, converting the relay frequency back to the first operating frequency, thereby realizing frequency-shift transmission between the near-end and the far-end.

[0003] In practical engineering, independent crystal oscillators are usually configured at the near and far ends, with different sources for their base clocks. Even if the crystal oscillators have the same nominal frequency and the same device model, a slight frequency difference will occur due to initial frequency deviation, device variability, temperature changes, power supply variations, and aging conditions. This frequency difference will first act on the local oscillator through frequency multiplication, division, or other frequency synthesis relationships at each end, and then be transmitted to the final recovered frequency through a two-stage mixing process. This causes the relay frequency generated at the near end to deviate from its theoretical value, and further, together with the error of the local oscillator at the far end, it forms the final residual frequency deviation.

[0004] Currently, the most direct solution is to use high-precision temperature-compensated crystal oscillators, oven-controlled crystal oscillators, or other high-stability frequency references at both the near and far ends, or to measure the clocks at both ends separately using high-precision frequency meters, spectrum analyzers, and external frequency references during production, debugging, and maintenance. While these solutions can improve absolute frequency accuracy, the cost of components, testing, and maintenance increases significantly with the target accuracy, contradicting the need for mass deployment of low-cost frequency-shift repeaters.

[0005] Another approach uses frequency division and digital counting to count the number of clock cycles under test within a specified observation period and determine the frequency error accordingly. This approach is relatively simple to implement in hardware, but when the frequency difference under test is less than the integer count change that a single observation window can reflect, different tiny clock errors may produce the same counting result. Simply extending the observation time can easily allow slow clock drift at different times and changes in the measurement environment to mix into the measurement results.

[0006] More importantly, traditional clock testing typically measures the near-end and far-end crystals as two independent devices. However, what frequency-shift repeaters truly need to address is the impact of two heterogeneous base clocks on the actual recovered frequency after passing through different frequency synthesis paths and participating in two stages of mixing. Therefore, obtaining the frequency deviation of a single crystal or only the final residual frequency deviation of the entire frequency-shift link is insufficient to directly distinguish the error contributions of the near-end and far-end clocks. Summary of the Invention

[0007] This invention discloses a method for measuring the clock error of a frequency shift repeater, in order to solve the above-mentioned problems in the prior art.

[0008] To achieve the above objectives, the present invention provides a method for measuring the clock error of a frequency shift repeater, the method comprising: A first measurement state and a second measurement state are constructed; the first measurement state adopts the near-end low-side local oscillator path, and the second measurement state adopts the near-end high-side local oscillator path; the nominal frequency conversion results of the first measurement state and the second measurement state are consistent, and the near-end clock error effect relationship in the first measurement state is different from that in the second measurement state; Determine the near-end clock error factor and the far-end clock error factor under the first measurement state and the second measurement state, respectively; In the first measurement state and the second measurement state, the continuous phase change between the input-side reference signal and the frequency shift link output signal is acquired respectively. Based on the continuous phase changes of the first measurement state and the second measurement state, and the corresponding near-end clock error action coefficient and far-end clock error action coefficient, the near-end base clock relative error and the far-end base clock relative error are determined respectively.

[0009] Furthermore, the construction of the first measurement state and the second measurement state includes: Verify that the candidate second measurement state meets the following conditions: the nominal relay frequency and nominal recovery frequency of the candidate second measurement state are consistent with those of the first measurement state, and the two sets of near-end clock error action coefficients and far-end clock error action coefficients corresponding to the candidate second measurement state are not proportional to the two sets of action coefficients corresponding to the first measurement state, thus forming a reversible binary linear observation relationship. Candidate paths that meet the conditions are determined as the second measurement state.

[0010] Further, determining the near-end clock error action coefficient and the far-end clock error action coefficient under the first measurement state and the second measurement state, respectively, includes: A known relative test deviation is applied to the near-end base clock. The offset corresponding to the near-end local oscillator is obtained according to the frequency division configuration of the near-end frequency synthesizer. The relative change of the final recovery frequency relative to the first operating frequency is obtained along the near-end mixing and far-end recovery path. The relative change is compared with the applied relative test deviation to obtain the near-end clock error action coefficient. A known relative test deviation is applied to the remote base clock, and the relative change of the final recovered frequency is obtained along the remote local oscillator and mixing path, thus obtaining the remote clock error effect coefficient.

[0011] Further, acquiring the continuous phase change between the input-side reference signal and the frequency-shift link output signal in the first measurement state and the second measurement state respectively includes: Phase acquisition for a dual-state observation group is completed in the order of first measurement state, second measurement state, second measurement state, and first measurement state, so that the two first measurement state windows in the same observation group are located at the beginning and end of the time interval, and the two second measurement state windows are located in the middle region.

[0012] Furthermore, the method also includes: The reference monitoring signal is obtained through the radio frequency coupling branch on the first operating frequency input side; The recovered monitoring signal is obtained through the remote monitoring coupling branch at the first working frequency after remote recovery, and then sent to the near-end measurement module through the backhaul branch after fixed frequency division. The reference monitoring signal and the recovery monitoring signal use the same frequency division configuration.

[0013] Further, acquiring the continuous phase change between the input-side reference signal and the frequency-shift link output signal in the first measurement state and the second measurement state respectively includes: A dual-channel edge capture method with a shared counting reference is used to acquire edge information of the reference monitoring signal and the recovered monitoring signal respectively; The continuous phase change is determined based on the edge information.

[0014] Further, determining the continuous phase change based on the edge information includes: Get the remaining difference count within the period at the start position of the window and the remaining difference count within the period at the end position of the window; During the window operation, when the remaining difference within a period crosses from one period boundary to another, the number of complete period crossings is counted in a uniform direction. The continuous phase change is determined based on the number of complete cycle spans, the remaining differential count within the cycle at the start position of the window, the remaining differential count within the cycle at the end position of the window, and the count value corresponding to a complete cycle of the reference monitoring signal.

[0015] Furthermore, the method also includes: The continuous phase changes obtained from the same measurement state in different windows are accumulated to a unified observation scale to form the cumulative equivalent phase change of the measurement state; The step of determining the relative error of the near-end base clock and the relative error of the far-end base clock based on the continuous phase changes of the first measurement state and the second measurement state, and the corresponding near-end clock error action coefficient and far-end clock error action coefficient, includes: The relative frequency offset of the first state is determined based on the cumulative equivalent phase change of the first measurement state and the corresponding cumulative equivalent reference period number, and the relative frequency offset of the second state is determined based on the cumulative equivalent phase change of the second measurement state and the corresponding cumulative equivalent reference period number. Based on the relative frequency offset of the first state, the relative frequency offset of the second state, and the corresponding near-end clock error action coefficient and far-end clock error action coefficient, the near-end base clock relative error and the far-end base clock relative error are obtained by solving simultaneously.

[0016] Furthermore, the step of accumulating continuous phase changes acquired in different windows under the same measurement state to a unified observation scale includes: The phase growth ratio of each window is obtained by comparing the continuous phase change of each window with the corresponding effective reference period. The phase growth ratio of the two windows under the same measurement state is averaged and converted to a common equivalent reference period scale to obtain the equivalent phase increment of the measurement state. The equivalent phase increments of each observation group are continuously accumulated to form the cumulative equivalent phase change.

[0017] Furthermore, the method also includes: The residual frequency offset of the normal frequency shift link is determined based on the relative error of the near-end base clock, the relative error of the far-end base clock, and the near-end clock error effect coefficient and the far-end clock error effect coefficient corresponding to the normal frequency shift operation state.

[0018] The beneficial technical effects of this invention are as follows: By constructing a first measurement state and a second measurement state, the nominal frequency conversion results of the two states are consistent, but the clock error effects are different. This transforms the originally coupled near-end and far-end clock errors into distinguishable observation results, thereby enabling the separate quantification of the error contributions of the near-end and far-end clocks. This overcomes the technical problem in traditional methods where the clock errors at both ends are coupled and difficult to quantify separately. This method acquires the continuous phase change between the input-side reference signal and the frequency shift link output signal in each measurement state, and accumulates the phase changes of the same measurement state in different windows to a unified observation scale. This converts minute clock frequency differences into quantifiable cumulative phase differences, enabling low-cost digital measurement circuits to identify minute frequency differences that are originally difficult to distinguish within a single observation window, overcoming the technical problem that minute clock frequency differences are difficult to directly distinguish with low-cost hardware. This method solves the cumulative observation results of two measurement states and the corresponding clock error action coefficients simultaneously to determine the relative error of the near-end base clock and the relative error of the far-end base clock, and obtains the residual frequency offset of the normal frequency shift link. It provides an engineering implementation path for high-precision clock error measurement using ordinary base crystal oscillators and general-purpose digital counter devices, and overcomes the technical problem that high-precision clock measurement depends on high-cost devices. Attached Figure Description

[0019] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0020] Figure 1 This is a flowchart of a method for measuring the clock error of a frequency shift repeater according to the present invention. Detailed Implementation

[0021] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0022] In one or more embodiments, such as Figure 1 As shown, a method for measuring the clock error of a frequency shift repeater is disclosed, the method comprising the following: S101. Construct a dual-state clock error observation relationship. Construct a first measurement state and a second measurement state; the first measurement state uses the near-end low-side local oscillator path, and the second measurement state uses the near-end high-side local oscillator path; the nominal frequency conversion results of the first measurement state and the second measurement state are consistent, and the near-end clock error effect relationship in the first measurement state is different from that in the second measurement state; determine the near-end clock error effect coefficient and the far-end clock error effect coefficient in the first measurement state and the second measurement state, respectively.

[0023] Specifically, this step builds upon the existing frequency synthesis link and two-stage mixing link of the frequency-shift repeater to construct the dual-state measurement parameters required for subsequent measurements. The near-end device receives the RF signal corresponding to the first operating frequency. The near-end base clock drives the frequency synthesizer to generate a near-end local oscillator. The near-end local oscillator and the first operating frequency signal enter the near-end mixer, and the subsequent filtering channel selects the mixing product corresponding to the relay frequency. The far-end device receives the relay frequency signal. The far-end base clock drives the far-end frequency synthesizer to generate a far-end local oscillator. The far-end local oscillator and the relay frequency signal are mixed again, and the first operating frequency is recovered through the corresponding filtering channel. During implementation, the device's main controller reads the reference clock configuration, multiplication configuration, division configuration, local oscillator output channel, and lockout status configuration of the near-end and far-end frequency synthesizers. Simultaneously, it reads the currently connected mixing channel and filtering channel configuration of the RF switch. The above data for the programmable frequency synthesizer comes from the configuration content written and stored in the control register during device initialization. The fixed-connection mixing and filtering channels are given by the channel configuration table corresponding to the hardware design in the device firmware. These configurations allow for the step-by-step determination of the actual frequency propagation paths between the near-end base clock and the near-end local oscillator, the near-end local oscillator and the relay frequency, the far-end base clock and the far-end local oscillator, and the far-end local oscillator and the first operating frequency after recovery. The mixing process employs the classic heterodyne mixing principle. The mixer output includes the sum-frequency component and the difference-frequency component formed by the RF signal and the local oscillator signal. The actual RF channel selects the target mixing product through the corresponding filter channel according to the equipment frequency plan. The fundamental frequency relationship is expressed as: ; in, This indicates the target frequency configuration value selected by the current mixer through the filter channel; This indicates the RF frequency configuration value for the current mixer; This represents the local oscillator frequency configuration value currently fed into the mixer by the frequency synthesizer. All the above frequency configuration values ​​use the same frequency unit, or the same nominal frequency scale already converted from register configuration, thus ensuring that the addition and subtraction relationships in the sum and difference frequency calculations remain consistent in dimensions. The near end selects the corresponding difference frequency component to form the relay frequency based on the first operating frequency and the local oscillator position. The far end selects either the sum or difference frequency component to recover the first operating frequency based on the actual configuration of the relay frequency and the far-end local oscillator. For example, with a first operating frequency configuration value of 1000 and a relay frequency configuration value of 900, when the near end uses a low-side local oscillator with a configuration value of 100, 900 is obtained through the difference frequency between 1000 and 100; when the near end uses a high-side local oscillator with a configuration value of 1900, 900 is also obtained through the difference frequency between 1900 and 1000; when the far end uses a local oscillator with a configuration value of 100, 1000 is recovered through the sum frequency between 900 and 100. Therefore, the same nominal relay frequency can be obtained from different near-end local oscillator injection relationships, while still being able to recover the same nominal first operating frequency from the far end.

[0024] The near-end low-side local oscillator path used during normal equipment operation is taken as the first measurement state. The main controller determines the role of the near-end fundamental clock error and the far-end fundamental clock error in this state according to the actual frequency multiplication, division, and mixing paths of the equipment. Specifically, in the configuration calculation, a known small relative test deviation is applied to the near-end fundamental clock. The offset corresponding to the near-end local oscillator is obtained based on the actual frequency multiplication and division configuration of the near-end frequency synthesizer. Then, the relative change of the final recovered frequency relative to the first operating frequency is obtained along the near-end mixing and far-end recovery paths. This relative change is compared with the applied near-end fundamental clock relative test deviation to obtain the near-end clock error effect coefficient in the first measurement state. Subsequently, the near-end nominal configuration is restored, and the known relative test deviation is applied to the far-end fundamental clock in the same way. The relative change of the final recovered frequency is obtained along the far-end local oscillator and mixing paths to obtain the far-end clock error effect coefficient in the first measurement state. The aforementioned relative test deviation is only used by the control program to determine the proportional relationship in the configuration calculation, and does not require actual adjustment of the crystal oscillator. This test deviation uses a dimensionless relative frequency deviation, and its value only needs to be a small non-zero quantity, maintaining the same sign convention in near-end and far-end calculations. Therefore, the clock error relationship is clearly expressed as the direction and proportion of the relative errors of the two basic clocks at the near and far ends on the relative offset of the final recovered frequency, and the obtained values ​​can be directly used in subsequent steps.

[0025] Taking the aforementioned frequency configuration as an example, when both the near-end low-side local oscillator configuration value and the far-end recovery local oscillator configuration value are 100, after the near-end base clock generates a given relative deviation, the near-end local oscillator changes proportionally, and the relay frequency changes in the opposite direction by subtracting the difference frequency relationship of the near-end local oscillator from 1000. Relative to the first operating frequency configuration value of 1000, the near-end clock error action coefficient corresponding to this change is -0.1. When the far-end base clock generates a relative deviation of the same nature, the far-end local oscillator configuration value of 100 changes proportionally, and the final recovery frequency is affected by the sum-frequency relationship between 900 and the far-end local oscillator. The far-end clock error action coefficient obtained relative to the first operating frequency configuration value of 1000 is positive 0.1. Therefore, the clock error action relationship in the first measurement state is jointly determined by the near-end action coefficient of -0.1 and the far-end action coefficient of positive 0.1.

[0026] The second measurement state uses the near-end high-side local oscillator path. The near-end frequency synthesizer switches the local oscillator configuration to 1900, and the mixing and filtering channel selects the difference frequency of 900 formed by 1900 and 1000. The far end continues to restore 900 to 1000 according to the predetermined recovery path. The main controller sets the known relative test deviation to the near-end and far-end base clocks respectively in the same way as in the first measurement state and calculates the relative change of the final recovered frequency along the true frequency propagation path. In this example, the near-end local oscillator with a configuration value of 1900 changes proportionally with the near-end base clock. This change acts on the relay frequency in the same direction through the difference frequency relationship between the high-side local oscillator and the first operating frequency, forming a near-end clock error factor of +1.9 relative to the first operating frequency configuration value of 1000. The far end still uses the recovered local oscillator with a configuration value of 100, so its far-end clock error factor continues to be +0.1. Therefore, the first measurement state establishes an error action relationship of "negative 0.1 and positive 0.1", and the second measurement state establishes an error action relationship of "positive 1.9 and positive 0.1". Both states have the same nominal relay frequency and nominal recovery frequency, and simultaneously form two different sets of deterministic observation relationships for the same pair of near-end and far-end fundamental clock errors. Since the near-end action coefficient is significantly different in the two states while the far-end action coefficient remains the same, the two sets of coefficients formed by the two states are disproportionate, allowing for a unique solution to the two unknown relative errors at the near and far ends.

[0027] The main controller further verifies the candidate second measurement state according to the aforementioned test deviation propagation method. When the nominal relay frequency and nominal recovery frequency of the two states are consistent, and the two sets of near-end and far-end clock error action coefficients can form two distinguishable observation relationships, the candidate path is determined as the second measurement state. Here, "distinguishable" means that the proportional relationship between the two sets of action coefficients on the two unknown errors at the near and far ends is different, forming a reversible binary linear observation relationship. If the proportions of the two sets of action coefficients obtained from the candidate path are the same, or the difference between them is lower than the calculation resolution threshold determined by the counter resolution and the numerical calculation accuracy of the control program, then it is not considered the second measurement state. After completing the hardware configuration for both states, the near-end frequency synthesizer register configuration, far-end frequency synthesizer register configuration, mixer and filter channel configuration, state identifier, conditions allowing the start of effective measurement, and the near-end and far-end clock error action coefficients corresponding to that state are saved respectively, forming the dual-state measurement parameters. The frequency synthesizer lock indicator and RF channel status are used to determine the effective acquisition starting point after each state switch; the two sets of clock error coefficients are continuously propagated with the corresponding state and are directly used as the numerical basis for the dual-state observation relationship when solving the final error.

[0028] S102. Obtain valid phase change records for both states. Under the first measurement state and the second measurement state, respectively, obtain the continuous phase change between the input-side reference signal and the frequency shift link output signal.

[0029] Specifically, step two receives the dual-state measurement parameters formed in step one and completes phase acquisition for a dual-state observation group in the order of first measurement state, second measurement state, second measurement state, and first measurement state. Each time the main controller enters a measurement state, it writes the corresponding near-end and far-end frequency synthesis configuration, switches the corresponding mixer and filter channels, and continuously reads the conditions for starting valid measurements specified in step one. When the local oscillator configuration is complete, the frequency synthesizer is locked, and the target RF channel is connected, the subsequent complete reference edge is used as the starting position of the current valid window. Each window saves the current state identifier and the near-end and far-end clock error action coefficients determined in step one, ensuring that the first-state data always corresponds to the error propagation relationship of the first state, and the second-state data always corresponds to the error propagation relationship of the second state. Using the above symmetrical state sequence, the two first-state windows in the same observation group are located on opposite sides of the time interval, and the two second-state windows are located in the middle, providing a continuous data basis for step three to form dual-state cumulative results with similar time backgrounds.

[0030] Phase changes are obtained using a dual-channel edge capture method with a shared counting reference. The reference monitoring signal is obtained from the first operating frequency input side via an RF coupling branch, and the recovered monitoring signal is obtained from the first operating frequency recovered at the far end via a far-end monitoring coupling branch. For devices with physically separated near and far ends, the recovered monitoring signal is first divided by a fixed frequency, and then the frequency-divided edge sequence is sent to the near-end measurement module via the device's existing monitoring feedback interface or a dedicated calibration feedback branch. The feedback branch maintains a fixed connection, and its propagation delay exists simultaneously at the start and end boundaries of the same window. The reference monitoring signal is processed using the same frequency division configuration at the near end. The two frequency-divided edge sequences enter two capture channels of the same main controller and share the same free-running counter. The main controller simultaneously records the number of complete cycles of the reference channel; therefore, the subsequent observation length is determined by the number of complete cycles experienced by the reference signal itself. The free-running counter is mainly used to determine the edge position within one reference cycle.

[0031] At the start of the current effective window, the main controller obtains the remaining differential count within the period of the recovery frequency edge relative to the reference frequency edge, and records it as... During window operation, the relative positions of the two edges are continuously monitored. When the remaining difference within a cycle crosses from one cycle boundary to another, the number of crossings in the complete cycle is recorded. The count is performed in a unified direction; at the end of the window, the remaining difference count within the period is obtained in the same manner. Here and All are limited to the remaining edge difference within a reference period after completing the full period expansion; the full period portion consists only of... Recording is done so that the complete cycle span and the remaining changes within the cycle occupy different data portions. The free-run count difference between two adjacent complete edges of the reference channel is denoted as... This is used to convert the remaining count within a period into the proportion of the reference period. The continuous phase change value of the current window is expressed as: ; in, This represents the continuous phase change value of the recovered frequency component relative to the reference frequency component within the current effective window; This indicates the relative number of directional complete cycles crossed by the two edges during the window period; This indicates the remaining difference count within the period after the period expansion is completed at the window's starting position; This indicates the remaining difference count within the period obtained using the same rule at the end position of the window; This represents the free-running count value corresponding to one complete cycle of the reference monitoring signal after frequency division. , and All count values ​​were obtained from the same free-running counter. For a reference period count greater than 0, therefore This indicates the proportion of the remaining phase change within the period. and All are expressed in terms of reference period number, which is a dimensionless phase period quantity. Taking a certain effective window as an example, when For 18, For 43, When the value is 100, and the recovery frequency edge completes two full cycle crossings along the specified positive direction during the window period, the remaining change within the cycle is 0.25, ultimately yielding... A value of 2.25 indicates that two complete cycles have been formed within the window, and a continuous phase shift of 0.25 cycles continues to accumulate.

[0032] The main controller also synchronously counts the number of complete cycles experienced by the reference monitoring signal within each valid window and uses this as the number of valid reference cycles. Since the reference monitoring signal and the recovered monitoring signal use the same frequency division configuration, the continuous phase change of the recovered signal relative to the reference signal, divided by the number of cycles experienced by the reference signal, directly characterizes the relative frequency change relationship between the two original operating frequencies. The frequency division process produces the same frequency scaling for both signals. After each window ends, the main controller displays the status flag and the continuous phase change value. The effective reference period number and the near-end and far-end clock error action coefficients corresponding to the current state form a dual-state effective phase record, which is continuously saved in the actual acquisition sequence of the first measurement state, the second measurement state, the second measurement state, and the first measurement state. Thus, the dual-state error action relationship formed in step one is converted into actual observation data with clear state attribution, continuous phase change, and a unified reference period scale, and is directly used as the input for step three.

[0033] S103, forming a dual-state cumulative observation result.

[0034] Specifically, step three receives the dual-state valid phase records formed in step two and, according to the state identifier, continuous phase change value, number of valid reference periods, and corresponding near-end and far-end clock error action coefficients in the records, divides multiple valid windows into cumulative observation results for the first and second measurement states respectively. The main controller, following the sequence of first measurement state, second measurement state, second measurement state, and first measurement state formed in step two, assembles a complete observation group of four valid records. The two first-state windows at the beginning and end of the observation group are located on the front and back sides of the time interval, respectively, while the two second-state windows are located in the middle region. Therefore, the representative observation times of both states are close to the time center of the same observation group. The clock error action coefficient of each record continues to be bound to the state identifier, ensuring that the accumulation process always maintains the two sets of error propagation relationships determined in step one.

[0035] The four windows within the same observation group may contain slightly different numbers of effective reference periods. Therefore, the main controller first uses the continuous phase change value of each window and the corresponding number of effective reference periods to obtain the phase growth ratio of that window. Then, it converts two windows in the same state to the common equivalent reference period scale of the current observation group. This processing is derived from the relationship between the relative phase change and the number of reference periods under stable frequency difference conditions. Combined with the dual-state symmetrical acquisition structure of this patent, it forms an equivalent phase increment representing the current observation group in that state from two windows with different time positions in the same state. ; in, Indicates the first Measurement status in a complete observation group The corresponding equivalent phase increment; Corresponding to the first measurement state or the second measurement state; Indicates the sequential numbering of the complete observation group; and The continuous phase change values ​​are taken from the two valid windows belonging to this state in step two, respectively; and These represent the number of valid reference periods corresponding to the two windows mentioned above; This represents the common equivalent reference period number obtained by averaging the effective reference periods of the four windows in the current observation group. and Expressed in terms of the number of cycles, , and All these are the number of reference periods, and the number of effective reference periods involved in the calculation is greater than 0. Dividing the continuous phase change value by the number of effective reference periods yields the dimensionless relative frequency offset ratio, which is then multiplied by the number of common equivalent reference periods to obtain the equivalent phase increment. Therefore... It is still expressed in terms of the number of periods. In the calculation process, the continuous phase change value is first divided by the corresponding effective reference period to obtain the phase growth ratio of each of the two windows. Then, the two ratios of the same state are averaged and finally converted to the same common equivalent reference period scale so that the first state and the second state have a consistent observation length benchmark when entering the accumulation process.

[0036] For example, in an observation group, the continuous phase changes for the two windows of the first measurement state are 0.18 and 0.20, corresponding to 10000 and 10200 effective reference periods, respectively; the continuous phase changes for the two windows of the second measurement state are -0.31 and -0.34, corresponding to 9900 and 10100 effective reference periods, respectively. The common equivalent reference period for the four windows is 10050. After calculation according to the above relationship, the equivalent phase increment for the first measurement state is approximately 0.189, and the equivalent phase increment for the second measurement state is approximately -0.327. The two results retain the error propagation direction for the two measurement states in step one, and have been converted to the same reference period scale.

[0037] For each complete observation group, the main controller adds the equivalent phase increment of the first measurement state to the first state accumulation register and the equivalent phase increment of the second measurement state to the second state accumulation register. Simultaneously, the common equivalent reference period number of the current observation group is added to the cumulative reference period records corresponding to both states. Subsequent observation groups continue to accumulate in the same way, so smaller phase changes within a single window will continuously increase along their respective states. After completing the predetermined observation process, the first measurement state generates the cumulative equivalent phase change, the cumulative equivalent reference period number, and the near-end and far-end clock error coefficients for the first state. The second measurement state generates a corresponding result with the same structure. Together, they constitute the dual-state cumulative observation result. Step four directly uses the cumulative equivalent phase change and the cumulative equivalent reference period number of each state to determine the relative frequency offset of the state, and uses the near-end and far-end fundamental clock errors continuously retained with that state to solve for them.

[0038] S104. Determine the quantization results of the near-end and far-end clock errors. Based on the continuous phase changes of the first measurement state and the second measurement state, and the corresponding near-end clock error action coefficient and far-end clock error action coefficient, determine the near-end base clock relative error and the far-end base clock relative error, respectively.

[0039] Specifically, step four receives the dual-state cumulative observation results formed in step three, which includes the cumulative equivalent phase change, cumulative equivalent reference period count, and near-end and far-end clock error coefficients corresponding to the first measurement state, as well as the corresponding data for the second measurement state. For any measurement state, the main controller divides the cumulative equivalent phase change of that state by the cumulative equivalent reference period count to obtain the state relative frequency offset of the recovered frequency relative to the first operating frequency. This calculation is directly derived from the basic relationship between phase accumulation and frequency difference: the effective reference period count represents the number of periods experienced by the reference signal during the observation period, and the cumulative equivalent phase change represents the number of periods experienced by the recovered signal more or less than the reference signal in the same period. The ratio of the two represents the relative frequency offset of the recovered signal relative to the reference signal. In step two, the reference monitoring signal and the recovered monitoring signal use the same frequency division configuration, so this ratio also corresponds to the relative frequency offset between the first operating frequency and the recovered frequency before frequency division.

[0040] Step one has already obtained the near-end and far-end clock error coefficients for the first and second measurement states, respectively. Steps two and three consistently correlate these coefficients with the actual observation data for the corresponding states. The main controller organizes the relative frequency offsets of the two states in a fixed order of first state and second state, and simultaneously organizes the two sets of clock error coefficients in a fixed order of near-end and far-end, thus forming two linear observation relationships regarding the relative error of the near-end base clock and the relative error of the far-end base clock. Its matrix form is expressed as follows: ; in, This represents the clock error vector, with two elements representing the relative error of the near-end base clock and the relative error of the far-end base clock, respectively, in a fixed order. This represents the dual-state relative frequency offset vector. The first element is obtained by dividing the cumulative equivalent phase change of the first measurement state by the cumulative equivalent reference period number, and the second element is obtained by the corresponding cumulative result of the second measurement state. This represents the dual-state error action matrix. The first row takes the clock error action coefficients of the near end and far end of the first measurement state determined in step one, and the second row takes the two action coefficients corresponding to the second measurement state. This represents the inverse solution relationship formed based on two sets of distinguishable clock error interaction relationships. Since clock error, relative frequency offset, and interaction coefficient are all relative quantities... and All elements are dimensionless relative frequency deviations. The action coefficients in the equation are also dimensionless proportionality coefficients; the reverse solution is only executed when step one confirms that the two sets of action coefficients are not proportional and can form a reversible relationship. In actual execution, the control program can directly use the binary linear equation elimination method, that is, first use the action coefficient of one end of the two states to eliminate the error of that end, obtain the error of the other end, and then substitute the result back into the observation relationship of one of the states to obtain the remaining clock error.

[0041] Using the same configuration as in Step 1—first operating frequency configuration value 1000, relay frequency configuration value 900, first state near-end local oscillator configuration value 100, second state near-end local oscillator configuration value 1900, and far-end recovery local oscillator configuration value 100—the clock error coefficients for the first measurement state are -0.1 and +0.1, respectively, and the clock error coefficients for the second measurement state are +1.9 and +0.1, respectively. Assuming the cumulative results from Step 3 show a relative frequency offset of 0.00000004 for the first state and 0.00000044 for the second state, then the first state represents 0.000000044 formed by a negative 0.1 times the near-end fundamental clock error and a positive 0.1 times the far-end fundamental clock error, and the second state represents 0.00000044 formed by a positive 1.9 times the near-end fundamental clock error and a positive 0.1 times the far-end fundamental clock error. When subtracting the first state from the second state, the contributions of the far-end clock error cancel each other out, and the difference in the relative frequency offset between the two states is 0.00000040. The difference in the near-end action coefficient is 2, thus the relative error of the near-end base clock is 0.00000020. Substituting this result back into the first state observation relationship, the relative error of the far-end base clock is 0.00000060. Substituting the two results back into the second state observation relationship, the near-end error contribution of +1.9 times and the far-end error contribution of +0.1 times together give 0.00000044, which is consistent with the actual observation result of the second state.

[0042] After determining the relative errors of the near-end and far-end base clocks, the main controller, based on the first measurement state interaction relationship as the normal operating path in step one, recombines the clock errors at both ends to obtain the residual frequency offset of the normal frequency shift link. In the example above, the relative error of the near-end base clock is 0.00000020, corresponding to a negative 0.1 times contribution in the first state; the relative error of the far-end base clock is 0.00000060, corresponding to a positive 0.1 times contribution in the first state. The combination of the two yields 0.00000004, consistent with the actual relative frequency offset in the first state. Finally, the relative error of the near-end base clock, the relative error of the far-end base clock, and the residual frequency offset of the normal frequency shift link are combined to form the clock error quantization result.

[0043] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for measuring the clock error of a frequency-shift repeater, characterized in that, The method includes: A first measurement state and a second measurement state are constructed; the first measurement state adopts the near-end low-side local oscillator path, and the second measurement state adopts the near-end high-side local oscillator path; the nominal frequency conversion results of the first measurement state and the second measurement state are consistent, and the near-end clock error effect relationship in the first measurement state is different from that in the second measurement state; Determine the near-end clock error factor and the far-end clock error factor under the first measurement state and the second measurement state, respectively; In the first measurement state and the second measurement state, the continuous phase change between the input-side reference signal and the frequency shift link output signal is acquired respectively. Based on the continuous phase changes of the first measurement state and the second measurement state, and the corresponding near-end clock error action coefficient and far-end clock error action coefficient, the near-end base clock relative error and the far-end base clock relative error are determined respectively.

2. The method for measuring the clock error of a frequency-shift repeater according to claim 1, characterized in that, The construction of the first measurement state and the second measurement state includes: Verify that the candidate second measurement state meets the following conditions: the nominal relay frequency and nominal recovery frequency of the candidate second measurement state are consistent with those of the first measurement state, and the two sets of near-end clock error action coefficients and far-end clock error action coefficients corresponding to the candidate second measurement state are not proportional to the two sets of action coefficients corresponding to the first measurement state, thus forming a reversible binary linear observation relationship. Candidate paths that meet the conditions are determined as the second measurement state.

3. The method for measuring the clock error of a frequency-shift repeater according to claim 1, characterized in that, The determination of the near-end clock error action coefficient and the far-end clock error action coefficient under the first measurement state and the second measurement state, respectively, includes: A known relative test deviation is applied to the near-end base clock. The offset corresponding to the near-end local oscillator is obtained according to the frequency division configuration of the near-end frequency synthesizer. The relative change of the final recovery frequency relative to the first operating frequency is obtained along the near-end mixing and far-end recovery path. The relative change is compared with the applied relative test deviation to obtain the near-end clock error action coefficient. By applying a known relative test deviation to the remote base clock, the relative change of the final recovered frequency is obtained along the remote local oscillator and mixing path, thus obtaining the remote clock error effect coefficient.

4. The method for measuring the clock error of a frequency-shift repeater according to claim 1, characterized in that, The step of acquiring the continuous phase change between the input-side reference signal and the frequency shift link output signal in the first measurement state and the second measurement state respectively includes: Phase acquisition for a dual-state observation group is completed in the order of first measurement state, second measurement state, second measurement state, and first measurement state, so that the two first measurement state windows in the same observation group are located at the beginning and end of the time interval, and the two second measurement state windows are located in the middle region.

5. The method for measuring the clock error of a frequency-shift repeater according to claim 4, characterized in that, The method further includes: The reference monitoring signal is obtained through the radio frequency coupling branch on the first operating frequency input side; The recovered monitoring signal is obtained through the remote monitoring coupling branch at the first working frequency after remote recovery, and then sent to the near-end measurement module through the backhaul branch after fixed frequency division. The reference monitoring signal and the recovery monitoring signal use the same frequency division configuration.

6. The method for measuring the clock error of a frequency-shift repeater according to claim 4, characterized in that, The step of acquiring the continuous phase change between the input-side reference signal and the frequency shift link output signal in the first measurement state and the second measurement state respectively includes: A dual-channel edge capture method with a shared counting reference is used to acquire edge information of the reference monitoring signal and the recovered monitoring signal respectively; The continuous phase change is determined based on the edge information.

7. The method for measuring the clock error of a frequency-shift repeater according to claim 6, characterized in that, Determining the continuous phase change based on the edge information includes: Get the remaining difference count within the period at the start position of the window and the remaining difference count within the period at the end position of the window; During the window operation, when the remaining difference within a period crosses from one period boundary to another, the number of complete period crossings is counted in a uniform direction. The continuous phase change is determined based on the number of complete cycle spans, the remaining differential count within the cycle at the start position of the window, the remaining differential count within the cycle at the end position of the window, and the count value corresponding to a complete cycle of the reference monitoring signal.

8. The method for measuring the clock error of a frequency-shift repeater according to claim 4, characterized in that, The method further includes: The continuous phase changes acquired under the same measurement state in different windows are accumulated to a unified observation scale to form the cumulative equivalent phase change of that measurement state; The step of determining the relative error of the near-end base clock and the relative error of the far-end base clock based on the continuous phase changes of the first measurement state and the second measurement state, and the corresponding near-end clock error action coefficient and far-end clock error action coefficient, includes: The relative frequency offset of the first state is determined based on the cumulative equivalent phase change of the first measurement state and the corresponding cumulative equivalent reference period number, and the relative frequency offset of the second state is determined based on the cumulative equivalent phase change of the second measurement state and the corresponding cumulative equivalent reference period number. Based on the relative frequency offset of the first state, the relative frequency offset of the second state, and the corresponding near-end clock error action coefficient and far-end clock error action coefficient, the near-end base clock relative error and the far-end base clock relative error are obtained by solving simultaneously.

9. The method for measuring the clock error of a frequency-shift repeater according to claim 8, characterized in that, The step of accumulating continuous phase changes acquired in different windows under the same measurement state to a unified observation scale includes: The phase growth ratio of each window is obtained by comparing the continuous phase change of each window with the corresponding effective reference period. The phase growth ratio of the two windows under the same measurement state is averaged and converted to a common equivalent reference period scale to obtain the equivalent phase increment of the measurement state. The equivalent phase increments of each observation group are continuously accumulated to form the cumulative equivalent phase change.

10. The method for measuring the clock error of a frequency-shift repeater according to claim 1, characterized in that, The method further includes: The residual frequency offset of the normal frequency shift link is determined based on the relative error of the near-end base clock, the relative error of the far-end base clock, and the near-end clock error effect coefficient and the far-end clock error effect coefficient corresponding to the normal frequency shift operation state.