A broadband phase measurement method

Abstract

The present invention discloses a wide-band phase measurement method, which includes digitizing a reference signal and a signal to be measured that have passed through anti-aliasing filtering respectively; mixing and filtering the reference signal with two first local oscillator quadrature signals generated by a first digital control oscillator respectively to solve the pseudo-phase of the reference signal; performing frequency estimation on the signal to be measured after frequency measurement to obtain the frequency of a second digital control oscillator for generating two second local oscillator quadrature signals; mixing and filtering the signal to be measured with the two local oscillator quadrature signals respectively to solve the pseudo-phase of the signal to be measured; solving the pseudo-phase of the signal to be measured relative to the reference signal according to the pseudo-phase of the reference signal and the pseudo-phase of the signal to be measured; correcting and compensating the pseudo-phase to obtain the true phase of the signal to be measured relative to the reference signal. The present invention effectively solves the problem of adaptive measurement of the phase of a signal to be measured in a large dynamic range by evaluating the frequency of the signal to be measured and adaptively tracking the frequency of the local oscillator signal.

Description

Technical Field

[0001] The present invention relates to a broadband phase measurement method, belonging to the technical field of metrological calibration. Background Art

[0002] With the development of human society and the progress of technology, the importance of time and frequency has become increasingly prominent. It plays an important supporting role in the development of fields such as national defense, metrology, major national projects, and national infrastructure. Independent time and frequency are related to national security and core interests. High-precision time and frequency have been widely applied in various fields, including space technology, national metrological applications, and national infrastructure construction. High-precision phase comparison measurement technology is the basis for time and frequency measurement, maintenance, and transmission, and its measurement level directly affects the application ability of time and frequency.

[0003] Currently, high-precision phase comparison measurement methods mainly include two categories: phase comparison measurement methods based on time interval counters and phase comparison measurement methods based on digitization. Among them, the phase comparison measurement methods based on time interval counters include: frequency deviation multiplication method, beat method, and dual-mixer time difference method. The frequency deviation multiplication method uses a frequency deviation multiplier to multiply the frequency deviation between the reference signal and the signal to be measured by several times, and then uses a counter to measure the amplified frequency deviation. This method can improve the measurement resolution. The beat method obtains the frequency deviation between the reference signal and the signal to be measured through mixing and filtering, and then uses a counter to measure the time of the frequency deviation signal. This method has less additional noise and high measurement accuracy. The dual-mixer time difference method uses two beat structures to achieve phase measurement between the reference signal and the signal to be measured. This method uses a double-balanced mixer and a common reference source, and the common reference source provides a reference source for the two beat structures. Since the two groups of symmetric beat structures are affected by the same noise, this method has the ability to cancel the phase noise of the common reference source and can improve the phase measurement resolution. These methods all transform the frequency of the signal to be measured into a low-frequency beat signal through frequency conversion, then use a high-speed counter to count the number of cycles of the beat signal, and finally calculate the phase between the reference signal and the signal to be measured according to the number of cycles. The commonality of these methods is to determine the frequency range of the signal to be measured, and then measure and calculate the phase between the reference signal and the signal to be measured by setting a specific functional circuit.

[0004] The digitization-based phase comparison measurement method digitizes the analog reference signal and the analog signal to be measured respectively by using an analog-to-digital converter, and then calculates the phase difference between the reference source and the source to be measured through an algorithm. Usually, the digitization-based phase comparison measurement methods include: the function correlation method, the spectral analysis method, and the I-Q demodulation method. The function correlation method refers to directly obtaining the phase difference between the reference source and the source to be measured by using the cross-correlation function and the auto-correlation function of the measurement reference source and the source to be measured. This method can effectively suppress Gaussian white noise, and the phase measurement result is theoretically independent of the signal frequency. The more sampling points, the more accurate the measurement. However, for interference with strong correlation such as harmonic signals, the phase measurement error is large. The spectral analysis method is a method that uses spectral analysis processing of the source to be measured and the reference source, extracts their phase information in turn, and then calculates the phase difference. This method can effectively suppress the influence of Gaussian white noise and harmonics. At present, the commonly used spectral analysis methods are the FFT phase measurement method and the all-phase FFT phase measurement method (APFFT). The FFT phase measurement method has the phenomenon of spectral leakage. Only when the signal is sampled at a positive period, the estimated value is equal to the actual value. APFFT has "phase immobility" in the entire frequency spectrum. This method can accurately measure the phase value without adding any auxiliary correction measures. The IQ demodulation method has a fast measurement speed, but the DC signal is greatly interfered by current noise, and the noise is difficult to filter, thus affecting its phase measurement accuracy.

[0005] It can be seen that the above-mentioned phase comparison measurement methods are all for the case where the frequency of the signal to be measured is relatively fixed. By setting a specific measurement scheme and circuit function, the phase comparison measurement between the signal to be measured and the reference signal is realized, and they do not have the function of adaptive measurement of the phase of the signal to be measured in a large dynamic range. Summary of the Invention

[0006] The technical problem to be solved by the present invention is: to overcome the deficiencies of the prior art, provide a wide-band phase measurement method, adaptively track the frequency of the local oscillator signal by evaluating the frequency of the signal to be measured, so as to increase the dynamic range of the input frequency of the signal to be measured; use frequency compensation technology to realize the interference of the phase of the local oscillator signal in the phase difference measurement result, and finally realize the phase comparison measurement between the signal to be measured and the reference signal.

[0007] To solve the above technical problem, the present invention discloses a wide-band phase measurement method, including:

[0008] The reference signal Ref and the signal to be measured DUT respectively pass through an anti-aliasing filter and then use an analog-to-digital converter ADC to realize the digitization of the analog signal respectively;

[0009] The digitized reference signal Ref is respectively mixed and filtered with two first local oscillator quadrature signals generated by the first numerically controlled oscillator NCO1 to solve the pseudo-phase of the reference signal

[0010] The digitized signal DUT is measured by FFT and the frequency is estimated to obtain the frequency f of the second digital controlled oscillator NCO2 used to generate two second local oscillator orthogonal signals. nco2 ;

[0011] The digitized signal DUT is mixed and filtered with two second local oscillator quadrature signals to solve the pseudo phase of the signal.

[0012] According to the reference signal pseudo phase and the pseudo phase of the signal to be measured Determine the pseudo phase of the DUT relative to the reference signal Ref

[0013] Correcting and compensating for spurious phase Get the true phase of the measured signal relative to the reference signal

[0014] In the above phase measurement method, the digitized reference signal Ref is mixed and filtered with the two first local oscillator quadrature signals generated by the first digitally controlled oscillator NCO1 to obtain the reference signal pseudo phase. The specific method is:

[0015] (1) The digitized reference signal is Among them, f ref is the reference signal frequency, is the initial phase of the reference signal, t = n / f s , f s is the sampling clock frequency, n is the number of sampling points;

[0016] (2) The two orthogonal signals generated by NCO1 are I nco1 and Q nco1 , where I nco1 (t)=sin(2πf nco1 t), Q nco1 (t) = cos(2πf nco1 t), the two orthogonal signals are mixed with the reference signal respectively, and the formula is:

[0017]

[0018]

[0019] (3) The two mixed signals are filtered out by FIR filters to remove high-frequency signals, and two beat signals I are obtained. FIR1 , Q FIR1 , the formula is:

[0020]

[0021]

[0022] (4) Use CORDIC to calculate the phase information in the two-beat signals. The formula is:

[0023] Where, is the initial phase of the reference signal, f ref is the reference signal frequency, f nco1 is the frequency of the two local oscillator quadrature signals generated by NCO1. Among them, f ref = f nco1 ,

[0024] In the above phase measurement method, the digitized device under test (DUT) signal is frequency estimated after FFT frequency measurement to obtain the frequency f nco2 of the second numerically controlled oscillator (NCO2) for generating two second local oscillator quadrature signals. The specific method is:

[0025] The digitized device under test (DUT) signal is:

[0026]

[0027] Where, f dut is the frequency of the device under test signal, is the initial phase of the device under test signal; t = n / f s , f s is the sampling clock frequency, and n is the number of sampling points;

[0028] Perform discrete Fourier transform on the device under test signal S dut (n). The formula is:

[0029]

[0030] Where, N is the total number of discrete Fourier transform points; Obtain the point with the largest amplitude in the Fourier sequence X(k) through peak search. The formula is:

[0031]

[0032] Where, mag is the maximum amplitude value in the sequence X(k), and index is the sequence number corresponding to the maximum amplitude value in the sequence X(k);

[0033] Calculate the frequency of NCO2 for generating two local oscillator quadrature signals. The formula is:

[0034]

[0035] where index is the sequence number corresponding to the maximum value of the sequence X(k), f s is the sampling clock frequency, and N is the number of sampling points.

[0036] In the above phase measurement method, the digitized device under test (DUT) signal is mixed and filtered with two second local oscillator quadrature signals respectively to solve the pseudo-phase of the DUT signal The specific method is as follows:

[0037] The digitized device under test signal is mixed and filtered with two quadrature signals I nco2 and I nco2 generated by NCO2 respectively, where I nco2 (t) = sin(2πft nco2 t), Q nco2 (t) = cos(2πft nco2 t). The phase information in the two filtered signals is calculated using CORDIC. The formula is:

[0038]

[0039] where, is the initial phase of the DUT signal, f dut is the frequency of the DUT signal, f nco2 is the frequency used by NCO2 to generate two local oscillator quadrature signals, t = n / f s , f s is the sampling clock frequency, and n is the number of sampling points.

[0040] In the above phase measurement method, the frequency f dut of the DUT signal has the following calculation formula:

[0041]

[0042] where, f nco2 is the frequency used by NCO2 to generate two local oscillator quadrature signals, is the mean value of the frequency change Δ ρ(dut) (f);

[0043] The mean value of the frequency change Δ ρ(dut) (f) has the following calculation formula:

[0044]

[0045] where, Δ ρ(dut) (f) represents the frequency deviation between the DUT signal and NCO2, and N is the number of averaging times;;

[0046] The frequency change Δ ρ(dut)(f), and the calculation formula is:

[0047]

[0048] Among them, f s is the sampling clock frequency.

[0049] In the above phase measurement method, according to the pseudo-phase of the reference signal and the pseudo-phase of the signal under test to solve the pseudo-phase of the signal under test DUT relative to the reference signal Ref The formula is:

[0050]

[0051] Among them, is the initial phase of the reference signal, is the initial phase of the signal under test, f dut is the frequency of the signal under test DUT, f ref is the frequency of the reference signal REF, f nco2 is the frequency for NCO2 to generate two orthogonal local oscillator signals, f nco1 is the frequency for NCO1 to generate two orthogonal local oscillator signals;

[0052] Convert the phase difference to a frequency deviation, and the formula is:

[0053]

[0054] Among them, respectively represent the initial phases of the reference signal and the signal under test at time t0, respectively represent t n the frequency deviations of the reference signal and the signal under test at time.

[0055] In the above phase measurement method, correct and compensate the pseudo-phase to obtain the true phase of the signal under test relative to the reference signal The specific formula is:

[0056]

[0057]

[0058] Among them, respectively represent the initial phases of the reference signal and the signal under test at time t0, are respectively the values after removing the frequency deviation outliers respectively represent t nFrequency deviation between the moment reference signal and the signal to be measured; is the frequency deviation between NCO2 and NCO1;

[0059] Frequency deviation between NCO2 and NCO1 The calculation formula is:

[0060]

[0061] Where, f nco2 is the frequency of the two local oscillator quadrature signals generated by NCO2, f nco1 is the frequency of the two local oscillator quadrature signals generated by NCO1, f s is the sampling clock frequency.

[0062] In the above phase measurement method, the elimination of frequency deviation abnormal values, the specific method is:

[0063] (1) Calculate the threshold γ = k1·std(Err) using the error signal Err(k), where k1 is an adjustment factor;

[0064] (2) Compare the size of the error signal Err(k) and the threshold γ;

[0065] (3) When Err(k) is less than or equal to the threshold γ, it means that the current frequency deviation is normal, then output the current frequency deviation and update the frequency deviation Reg stored at the previous moment to the current frequency deviation

[0066] (4) When Err(k) is greater than the threshold γ, it means that the current frequency deviation is abnormal, then output the frequency deviation Reg stored at the previous moment,

[0067] In the above phase measurement method, the calculation formula of the error signal Err(k) is:

[0068]

[0069] Where, is the frequency change amount of the frequency deviation data , k ∈ (1, N), where N is the number of points of the frequency change amount; is the new frequency deviation data generated by constructing a linear fitting model from the original frequency deviation data .

[0070] In the above phase measurement method, the calculation formula of the frequency deviation data is:

[0071]

[0072] Among them, is the initial value of the frequency deviation data, d is the frequency drift rate, and w y (t) is the frequency noise.

[0073] The beneficial effects of the present invention compared with the prior art are as follows:

[0074] (1) The present invention proposes a phase comparison measurement method with a large dynamic range. By using traditional frequency measurement and phase measurement techniques, a rough estimation of the frequency of the signal to be measured and a rough measurement of the phase are realized. Based on the estimated value of the frequency of the signal to be measured, the frequency of the local oscillator signal is adjusted so that there is a certain deviation between the frequency of the local oscillator signal and the frequency of the signal to be measured, thereby reducing the design complexity of the mixer filter and ensuring the generality of the hardware and software design.

[0075] (2) The present invention proposes a phase measurement method based on frequency compensation. During the phase measurement based on I-Q demodulation, the phase interference of the local oscillator signal is introduced. By using the integral relationship between frequency and phase, the phase compensation of the local oscillator signal is realized in the frequency domain, thereby realizing the phase comparison measurement between the reference signal and the signal to be measured. This not only greatly reduces the background noise of the measurement system but also effectively solves the problem of adaptive measurement of the phase of the signal to be measured in a large dynamic range. Description of the Drawings

[0076] Figure 1 is a schematic diagram of wide-band phase measurement;

[0077] Figure 2 is a schematic diagram of the phase relationship between the reference signal and the signal to be measured;

[0078] Figure 3 is a schematic diagram of the detection and elimination of abnormal frequency deviation values based on linear fitting;

[0079] Figure 4 is a flow chart of the detection and elimination of abnormal frequency values;

[0080] Figure 5 is a schematic diagram of the elimination of abnormal frequency values;

[0081] Figure 6 is the corrected true phase difference diagram when the reference signal and the signal to be measured are of the same frequency;

[0082] Figure 7 is the corrected true phase difference diagram when the reference signal and the signal to be measured are of different frequencies. Specific Embodiments

[0083] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings.

[0084] Aiming at the problem of the dynamic range limitation of the frequency of the signal to be measured in the existing phase measurement methods, a wide-band phase measurement method is proposed. This method uses FFT frequency measurement technology to roughly estimate the frequency of the signal to be measured, and then autonomously adjusts the frequency of the local oscillator signal according to the estimated value of the frequency of the signal to be measured to make the frequency of the local oscillator signal close to the frequency of the signal to be measured. The I-Q demodulation method is used to respectively realize the preliminary estimation of the phases of the reference signal and the signal to be measured, and then the phase comparison measurement between the reference signal and the signal to be measured is realized through the frequency compensation method.

[0085] As Figure 1 shown, a wide-band phase measurement method includes:

[0086] The reference signal Ref and the signal to be measured DUT respectively pass through the anti-aliasing filter and then use the analog-to-digital converter ADC to respectively realize the digitization of the analog signal. The digitized reference signal Ref is respectively mixed and filtered with two first local oscillator quadrature signals generated by the first numerically controlled oscillator NCO1 to solve the pseudo-phase of the reference signal The digitized signal to be measured DUT is frequency estimated by FFT and then the frequency f of the second numerically controlled oscillator NCO2 used to generate two second local oscillator quadrature signals is obtained nco2 ; The digitized signal to be measured DUT is respectively mixed and filtered with two second local oscillator quadrature signals to solve the pseudo-phase of the signal to be measured According to the pseudo-phase of the reference signal and the pseudo-phase of the signal to be measured Solve the pseudo-phase of the signal to be measured DUT relative to the reference signal Ref Correct and compensate the pseudo-phase Obtain the true phase of the signal to be measured relative to the reference signal

[0087] Step 1: Rough frequency estimation

[0088] Since the frequency of the signal to be measured has a wide range, in order to reduce the complexity of FIR design, FFT is used to quickly estimate the frequency f of the signal to be measured nco2 , and then NCO2 respectively generates two orthogonal signals with the frequency of f nco2 . The basic process of the rough frequency estimation of the signal to be measured:

[0089] Assume that the digitized signal to be measured is f dut is the frequency of the signal to be measured, is the initial phase of the signal to be measured, t = n / f s , f s is the sampling clock frequency, and n is the number of sampling points. For the signal to be measured S dut(n) Perform a discrete Fourier transform as shown in Equation (1).

[0090]

[0091] Where N is the total number of points of the discrete Fourier transform. Obtain the point with the largest amplitude in the Fourier sequence X(k) through peak search, as shown in Equation (2).

[0092]

[0093] Where mag is the maximum amplitude value in the sequence X(k), and index is the subscript of the maximum amplitude value in the sequence X(k). Combine the subscript of the maximum value of the sequence X(k) in Equation (2), and estimate the rough frequency of the signal to be measured using Equation (3).

[0094]

[0095] Step 2: High-precision phase measurement

[0096] (1) Reference signal pseudo-phase measurement

[0097] Assume that the digitized reference signal is f ref is the reference signal frequency, is the initial phase of the reference signal, t = n / f s , f s is the sampling clock frequency, and n is the number of sampling points. The two orthogonal signals generated by NCO1 are I nco1 and Q nco1 , where I nco1 (t) = sin(2πf nco1 t), Q nco1 (t) = cos(2πf nco1 t). These two orthogonal signals are respectively mixed with the reference signal as shown in Equations (4) and (5).

[0098]

[0099]

[0100] After the two mixed signals pass through the FIR filter respectively to filter out the high-frequency signals, two beat signals (I FIR1 , Q FIR1 ) are obtained, as shown in Equations (6) and (7).

[0101]

[0102]

[0103] The phase information in the two-beat signals is calculated using CORDIC, as shown in Equation (8). The phase information in Equation (8) includes the initial phase of the reference signal and the influence of the NCO1 frequency on the phase of the reference signal. When the frequency of the reference signal is equal to the frequencies of the two orthogonal signals generated by NCO1, the initial phase of the reference signal can be obtained from Equation (8).

[0104]

[0105] (2) Pseudo-phase measurement of the signal under test

[0106] The digitized signal under test is mixed and filtered with the two orthogonal signals I nco2 and Q nco2 generated by NCO2 respectively, where I nco2 (t) = sin(2πf nco2 t), Q nco2 (t) = cos(2πf nco2 t). Finally, the phase information in the two filtered signals is calculated using CORDIC, as shown in Equation (9).

[0107]

[0108] (3) Phase difference between the reference signal and the signal under test

[0109] The pseudo-phase difference between the reference signal and the signal under test is as shown in Equation (10).

[0110]

[0111] As can be seen from Equation (8) and Equation (9), when f ref ≠ f dds1 , f dut ≠ f dds2 , the pseudo-phases of the reference signal and the signal under test are linear functions of time t. In practice, the phase of any signal satisfies the interval [0, 2π]. Therefore, there is a cross-period phenomenon for the pseudo-phase difference . To address the phase cross-period problem, the phase difference is differentiated to obtain the frequency deviation, and after completing various calculations, the frequency deviation is converted back to the phase difference through integration. Another form expressed using the frequency deviation is given in Equation (11).

[0112]

[0113] where ]>[[]END]represent the initial phases of the reference signal and the signal under test at time t0 respectively, represent the frequency deviations of the reference signal and the signal under test at time t n respectively.

[0114] Figure 2 The phase relationship between the reference signal and the signal to be measured is given, where the frequency of the reference signal is f ref = 10 MHz, f nco1 = 10 MHz. At this time the frequency of the signal to be measured is f dut = 8 MHz, and the FFT rough estimate frequency f nco2 = 7.998 MHz. At this time is a linear function of time t. Figure 2 The phase difference between the reference signal and the signal to be measured is given in Since the initial phase of the reference signal is zero, therefore, and are in antiphase.

[0115] Step 3: Detection and rejection of frequency outliers

[0116] Due to the interference of uncertain factors such as the external environment and the characteristics of its own equipment during the measurement of the frequency of the signal to be measured, as well as the errors introduced by the phase cross-cycle compensation, there are abnormal data such as "glitches" in the frequency deviation. In order to improve the measurement accuracy of the frequency of the signal to be measured, it is necessary to focus on identifying the outliers in the frequency deviation data and performing scientific and reasonable compensation. Mathematically, it is generally considered that the frequency deviation can be expressed as shown in formula (12).

[0117]

[0118] where is the initial value of the frequency deviation, d is the frequency drift rate, and w y (t) is the frequency noise. Based on the relationship between the frequency deviation and time t in formula (12), in this embodiment, a linear fitting method is used to detect the outliers in the frequency deviation.

[0119] The basic process of detecting and rejecting frequency deviation outliers: Based on a set of frequency change amounts (k ∈ (1, N), where N is the number of points of the frequency change amount), the least squares method is used to solve the relevant parameters in formula (12) and construct a linear fitting model. In this embodiment, the influence of the frequency noise w y (t) factor is not considered; then the linear model is used to predict the new frequency deviation data The comparator is used to solve the original frequency deviation data and the difference Err(k). Finally, the frequency outliers are identified according to the error signal Err(k), and the outliers are rejected.

[0120] Calculate the threshold γ = k1·std(Err) using the error signal Err(k), where k1 is an adjustment factor, typically taking a value of 3. By comparing the magnitude of the error signal Err(k) with the threshold γ, when Err(k) is greater than the threshold γ, it indicates that there are outliers in the frequency. At this time, use the normal frequency deviation data Reg at the previous moment to replace the current abnormal frequency deviation; when Err(k) is less than or equal to the threshold γ, it indicates that the frequency data is normal, and then output the original frequency deviation data

[0121] Figure 3 shows a schematic diagram of the detection and elimination of frequency deviation outliers based on linear fitting Figure 4 shows the flow chart of frequency outlier detection and elimination Figure 5 shows the frequency deviation data after abnormal detection processing and the frequency deviation data after linearization

[0122] Step 4: Measurement of the frequency of the signal to be measured

[0123] According to the differential relationship between phase and frequency, taking the first derivative of time t in formula (9) can obtain the frequency change Δ(f), as shown in formula (13), where

[0124]

[0125] Since the pseudo-phase data range in formula (9) is [-π, π], therefore, when using the pseudo-phase data to solve the frequency change Δ(f), the problem of phase cross-period needs to be solved. When considering the case of phase cross-period, the frequency change Δ(f) can be characterized as shown in formula (14).

[0126]

[0127] Due to various noise interferences and errors in phase cross-period processing during the measurement of the frequency of the signal to be measured, therefore, by calculating the mean value of the frequency change reflects the degree to which the frequency of the signal to be measured deviates from the frequency of NCO2, as shown in formula (15).

[0128]

[0129] Among them, represents the frequency deviation between the signal to be measured and NCO2, N is the number of averaging times, generally N takes 100. Formula (16) gives the estimated value of the frequency of the signal to be measured obtained through measurement.

[0130]

[0131] Step 5: Phase Compensation

[0132] The purpose of phase compensation is to eliminate the phase difference between the reference signal and the signal to be measured. The influence of the local oscillator signal phase in formula (17). Since the phase range is [-π, π], when the reference signal and the signal to be measured have different frequencies, the phase must exist across cycles. It is obviously not advisable to offset the local oscillator signal phase in formula (17) by phase compensation. Therefore, the present invention adopts frequency compensation to achieve the elimination of the local oscillator signal phase in formula (17), as shown in formula (18).

[0133] When the reference signal and the measured signal have different frequencies than the local oscillator frequencies of the NCOs on the corresponding paths, the phase difference between the reference signal and the measured signal represented in formula (10) includes the phases of the two NCOs. Therefore, phase compensation is required to obtain the true phase difference between the reference signal and the measured signal. Formulas (17) and (18) give the true phase difference between the reference signal and the measured signal.

[0134]

[0135]

[0136] in is the frequency deviation between NCO2 and NCO1,

[0137] Figure 6 The reference signal and the measured signal have the same frequency (f ref =f dut ) The corrected true difference in Figure 7 The actual phase difference after correction when the reference signal and the measured signal have different frequencies is given in the figure, where f ref =10MHz, f dut =8MHz, Combine Figure 6 and Figure 7 From the phase between the reference signal and the signal to be measured, it can be seen that the phase measurement method based on frequency compensation proposed in the present invention is effective and feasible.

[0138] The principle of the present invention is as follows: using FFT frequency measurement technology to estimate the frequency of the signal to be measured, tracking the frequency of the local oscillator signal based on the frequency estimation value, so that there is a certain deviation between the frequency of the local oscillator signal and the frequency of the signal to be measured; respectively realizing the phase comparison measurement between the reference signal and its local oscillator signal and the phase comparison measurement between the signal to be measured and its local oscillator signal; realizing the influence of the phase of the local oscillator signal in the measured phase between the reference signal and the signal to be measured by frequency compensation technology, and finally realizing the phase comparison measurement within a large dynamic range of the frequency of the signal to be measured.

[0139] As described above, it is only the best specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

[0140] The content not detailed in the specification of the present invention belongs to the well-known technology of those skilled in the art.

Claims

1. A wide-band phase measurement method, characterized in that, Including: Digitize the anti-aliasing filtered reference signal Ref and the signal under test DUT respectively; The digitized reference signal Ref is respectively mixed with two first local oscillator quadrature signals generated by the first digital control oscillator NCO1, filtered, and the pseudo-phase of the reference signal is solved After the digitized signal under test (DUT) is frequency - measured by FFT, frequency estimation is performed to obtain the frequency f of the second digital control oscillator (NCO2) for generating two - path second local oscillator quadrature signals. nco2 ; The digitized device under test (DUT) signal is mixed and filtered with two second local oscillator quadrature signals respectively to solve the pseudo-phase of the DUT signal. According to the pseudo-phase of the reference signal and the pseudo-phase of the signal under test Solve for the pseudo-phase of the signal under test DUT relative to the reference signal Ref Correction and compensation of pseudo-phase Obtain the true phase of the signal to be measured relative to the reference signal The specific formula is as follows: Among them, respectively represent the initial phases of the reference signal and the signal to be measured at time t0, are the values after removing the frequency deviation and outliers, respectively represent the frequency deviations of the reference signal and the signal to be measured at time t n ; is the frequency deviation between NCO2 and NCO1.

2. The broadband phase measurement method according to claim 1, characterized in that: The digitized reference signal Ref is respectively mixed with two first local oscillator quadrature signals generated by the first digital control oscillator NCO1, filtered, and the pseudo-phase of the reference signal is solved The specific method is as follows: (1) The digitized reference signal is where f ref is the reference signal frequency, φ ref is the initial phase of the reference signal, t = n / f s , f s is the sampling clock frequency, and n is the number of sampling points; (2) The two orthogonal signals generated by NCO1 are I nco1 and Q nco1 , where I nco1 (t) = sin(2πft nco1 ), Q nco1 (t) = cos(2πft nco1 ). These two orthogonal signals are respectively mixed with the reference signal, and the formula is: (3) After the two-way mixing, the signals respectively pass through FIR filters to filter out high-frequency signals, and two beat signals I FIR1 and Q FIR1 are obtained. The formula is as follows: (4) Use CORDIC to calculate the phase information in the two beat signals, and the formula is: where φ ref is the initial phase of the reference signal, f ref is the reference signal frequency, f nco1 is the frequency at which NCO1 generates two local oscillator quadrature signals, where f ref = f nco1 , 3. A broadband phase measurement method according to claim 1, characterized in that: The digitized signal under test (DUT) is frequency-estimated after FFT frequency measurement to obtain the frequency f of the second numerically controlled oscillator (NCO2) for generating two-way second local oscillator quadrature signals nco2 , and the specific method is as follows: The digitized signal under test DUT is: Among them, f dut is the frequency of the signal to be measured, is the initial phase of the signal to be measured; t = n / f s , f s is the sampling clock frequency, and n is the number of sampling points; For the signal S to be measured dut (n) is subjected to discrete Fourier transform, and the formula is: where N is the total number of points of the discrete Fourier transform; Obtain the point with the largest amplitude in the Fourier sequence X(k) through peak search, and the formula is: where mag is the maximum amplitude value in the sequence X(k), and index is the serial number corresponding to the maximum amplitude value in the sequence X(k); Calculate the frequency of NCO2 for generating two local oscillator quadrature signals, and the formula is: where index is the sequence number corresponding to the maximum value of the sequence X(k), f s is the sampling clock frequency, and N is the number of sampling points.

4. A broadband phase measurement method according to claim 1, characterized in that: The digitized signal under test (DUT) is mixed and filtered with two second local oscillator quadrature signals respectively to solve the pseudo-phase of the signal under test. The specific method is as follows: The digitized signal to be measured is S dut (t) = sin(2πf dut t + φ dut ) is mixed and filtered with the two orthogonal signals I nco2 and Q nco2 respectively. Among them, I nco2 (t) = sin(2πf nco2 t), Q nco2 (t) = cos(2πf nco2 t). The phase information in the two filtered signals is calculated using CORDIC, and the formula is: Among them, is the initial phase of the signal to be measured, and f dut is the frequency of the signal to be measured, and f nco2 is the frequency used by NCO2 to generate two local oscillator quadrature signals, t = n / f s , and f s is the sampling clock frequency, and n is the number of sampling points.

5. A broadband phase measurement method according to claim 4, characterized in that: The frequency f of the signal to be measured dut , and the calculation formula is: where f nco2 is the frequency for the NCO2 to generate two local oscillator quadrature signals, is the frequency change Δ ρ(dut) (f) mean value; The frequency change amount Δ ρ(dut) (f) mean value, and the calculation formula is: where Δ ρ(dut) (f) represents the frequency deviation between the signal to be measured and NCO2, and N is the number of averaging times; The frequency change amount Δ ρ(dut) (f), and the calculation formula is: Among them, f s is the sampling clock frequency.

6. A broadband phase measurement method according to claim 1, characterized in that: According to the pseudo-phase of the reference signal and the pseudo-phase of the signal under test Solve for the pseudo-phase of the device under test (DUT) relative to the reference signal Ref The formula is: Among them, is the initial phase of the reference signal, is the initial phase of the signal to be measured, f dut is the frequency of the DUT (Device Under Test) of the sequence number to be measured, f ref is the frequency of the reference signal REF, f nco2 is the frequency for NCO2 to generate two local oscillator quadrature signals, f nco1 is the frequency for NCO1 to generate two local oscillator quadrature signals; Convert the phase difference into a frequency deviation, and the formula is: wherein, respectively represent the initial phases of the reference signal and the signal to be measured at time t0, respectively represent n the frequency deviations of the reference signal and the signal to be measured at time t.

7. A broadband phase measurement method according to claim 1, characterized in that: Correction and compensation of pseudo-phase Obtain the true phase of the signal to be measured relative to the reference signal Frequency deviation between NCO2 and NCO1 The calculation formula is as follows: Among them, f nco2 is the frequency for the NCO2 to generate two local oscillator quadrature signals, f nco1 is the frequency for the NCO1 to generate two local oscillator quadrature signals, f s is the sampling clock frequency.

8. A broadband phase measurement method according to claim 7, characterized in that: The removed frequency deviation Outliers, and the specific method is as follows: (1) Calculate the threshold γ = k1·std(Err) using the error signal Err(k), where k1 is an adjustment factor; (2) Compare the magnitude of the error signal Err(k) with the threshold γ; when Err(k) is less than or equal to the threshold γ, it represents that the current frequency deviation is normal, then output the current frequency deviation and update the frequency deviation Reg stored at the previous moment to the current frequency deviation When Err(k) is greater than the threshold γ, it represents that the current frequency deviation is abnormal, then output the frequency deviation Reg stored at the previous moment, 9. A broadband phase measurement method according to claim 8, characterized in that: The calculation formula of the error signal Err(k) is: Among them, is the frequency deviation data of the frequency change amount, k ∈ (1, N), where N is the number of points of the frequency change amount; is the new frequency deviation data generated by constructing a linear fitting model from the original frequency deviation data .

10. A broadband phase measurement method according to claim 9, characterized in that: The frequency deviation data has the following calculation formula: Among them, is the initial value of the frequency deviation, d is the frequency drift rate, and w y (t) is the frequency noise.

Images