Segmented frequency parameter correction method for dual-channel voltage comparator

Abstract

The invention discloses a segmented frequency parameter correction method for a dual-channel voltage comparator, belongs to the technical field of voltage comparator correction, and solves the technical problem that an existing voltage comparator is poor in correction accuracy and applicability. The method comprises the following steps: constructing a minimum error path and a hardware system with self-calibration; constructing a broadband sine wave measurement backlash compensation algorithm system and realizing the system; constructing a segmented correction method for broadband phase absolute calibration and channel amplitude consistency calibration; performing sub-ppm level and sub-ns level phase shift K coefficient compensation, audio testing and harmonic verification; the segmented frequency parameter correction method and device for the dual-channel voltage comparator are high in accuracy, and can be widely applied to electric parameter measurement and AC-DC bridge measurement by matching with various voltage and current converters.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of voltage comparator calibration, in particular to a segmented frequency parameter calibration method for a dual-channel voltage comparator. BACKGROUND

[0002] The field of metrological detection is an important field supporting the progress of knowledge in today's society, and is the cornerstone of the development of the country in various fields such as industrial production, scientific research, semiconductor manufacturing, aerospace, and military industry, especially in today's era of explosive knowledge development.

[0003] The present application relates to a tool for value transfer, which covers a measurement range including direct current measurement, AC-DC conversion, AC sampling core algorithm and traceable parameter correction technology, and covers the manufacturing process and detection process method.

[0004] Current problems of electric parameter detection:

[0005] 1. Accuracy - the accuracy of the AC-DC comparator is 8ppm, and it is difficult to obtain effective direct value transfer.

[0006] 2. Applicability - the equipment of the existing metrological institutions has slow detection speed and can only meet the transfer of tens of mV.

[0007] 3. Limitations - electric parameter traceable calibration has strict requirements for detection tools, detection personnel and detection process.

[0008] 4. Stability - short calibration period and unable to determine the change amount of frequency parameter segmentation.

[0009] Simple description of prior art method:

[0010] Value transfer method: each metrological authority has mature detection equipment for detection, and the detection uncertainty is about (6-15) ppm at power frequency and about (15-150) ppm at 3000Hz.

[0011] Indirect transfer method: calibrate the DC reference voltage of the analog-to-digital converter, and the default DC measurement index is the AC measurement accuracy, and the uncertainty is about (30-300) ppm. SUMMARY

[0012] The technical problem to be solved by the present application is to solve the above-mentioned deficiencies of the prior art, and the purpose of the present application is to provide a segmented frequency parameter calibration method for a dual-channel voltage comparator with high accuracy and wide applicability.

[0013] The technical solution of the present application is: a segmented frequency parameter calibration method for a dual-channel voltage comparator, comprising the following steps:

[0014] Step 11. Construct a minimum error path and a hardware system with self-calibration:

[0015] The minimum error path measurement channel is composed of an operational amplifier buffer and an AD sampling chip;

[0016] The input buffer operational amplifier is low-noise and low-drift, the A / D chip has a dynamic self-calibration function, a DC reference chip and a temperature compensation crystal are used, and a double-channel electrical isolation design is used to trace 4V AC voltage to DC voltage 7V layout and wiring optimization. The AC-DC conversion accuracy is equivalent to the DC reference of the AD chip manufacturer, which eliminates the gain conversion resistance error source of the input channel and the DC reference, thereby ensuring the accuracy and long-term stability of the AC-DC comparator for AC voltage measurement.

[0017] Step 12. Construct the optimal equivalent frequency parameter and innovative AC sampling algorithm:

[0018] The amplitude component of the equivalent frequency parameter of the hardware system is 10E-14, and the phase shift component is 10E-10.

[0019] The innovative AC sampling algorithm uses real-time polynomial data reconstruction technology, which covers the audio frequency range of 40Hz-3500Hz and the power frequency range of 0-256 harmonics.

[0020] As a further improvement, the following steps are also included:

[0021] Step 21. Construct the gap compensation algorithm system for wideband sinusoidal wave measurement and implement it. The innovative design of the "90°+conjugate" filter window is that the sampling sequences are 90° to each other and the overall is in a symmetrical position on the coordinate axis. This solves the problem of rapid increase of gap error with increasing signal frequency and solves the problem of amplitude and phase algorithm for non-integer period sampling of AC signals using existing filtering methods.

[0022] Further, the following steps are also included:

[0023] Step 31. Sub-ppm magnitude error correction and sub-ns phase delay correction: i represents the segment number, ωi represents the characteristic frequency point parameter signal frequency of the segment, ωi=2*π*signal frequency, fi represents the characteristic parameter measurement error amplitude component of the segment, δi represents the characteristic parameter measurement phase shift component of the segment, Kfi and Kti represent the frequency band calculation coefficients, and fi=ωi^2*Kfi(ωi) and δi=ωi*Kti(ωi) are defined. That is, the segmented constant is used to solve the continuous frequency working condition measurement, and when the FFT transform is integrated into the frequency parameter K coefficient compensation, the harmonic frequency parameter correction problem can be solved.

[0024] Further, it is characterized by constructing a wideband phase absolute calibration and channel amplitude consistency calibration segmented correction method, which includes:

[0025] 1) calibration frequency points include: 50.0032, 122.0703, 244.1406, 488.2813, 976.5625, 1953.125, 3004.808;

[0026] Frequency segments include:

[0027] F0: 40<=F0<100; F1: 100<=F1<200;

[0028] F2: 200<=F2<400; F3: 400<=F3<800;

[0029] F4: 800<=F4<1800; F5: F5>=1800;

[0030] 2) calibration phase points, the wide frequency low distortion single voltage input two channels are used to carry out absolute calibration on phase error;

[0031] 3) since the frequency point selection "gap" tends to be zero, the signal source waveform distortion power frequency is 0.01%, and the signal source waveform distortion high frequency is less than 0.025%, so that the accuracy and stability required for calibration are ensured, and the actually measured amplitude parameter is 10E-13, and the phase parameter is 10E-10;

[0032] 4) after the amplitude frequency parameter consistency calibration is completed, absolute calibration is implemented.

[0033] Further, the double-channel voltage comparator can be used as a precision audio bridge in quantity transmission application, including:

[0034] 1) the double-channel wide frequency zero degree phase absolute calibration and amplitude consistency calibration of the comparator meet the technical requirements of two precision detection arms of the bridge, the resolution of the double-channel voltage comparator is better than 0.1ppm, and good stability is achieved, and the knowledge of metrology "transfer precision" can be used in the technical field of non-standard voltage proportional comparison, impedance measurement and transducer measurement;

[0035] 2) the precision test principle of the standard impedance is used to solve the precision measurement of the AC resistance time constant.

[0036] Advantages

[0037] Compared with the prior art, the present application has the advantages of:

[0038] 1) meet the high precision and wide frequency technical requirements of design;

[0039] 2) has a long-term stability of more than ten years without recalibration;

[0040] 3) the AD dynamic self-calibration resolution is higher than the index given by the manufacturer;

[0041] 4) AC / DC conversion traceable to the highest standard center value of the manufacturer;

[0042] 5) Accuracy can be checked at any time;

[0043] 6) Can be combined with other standard components into various precision measurement systems. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 Equivalent K coefficient diagram for dual-channel voltage comparator;

[0045] Figure 2 System principle diagram for dual-channel voltage comparator;

[0046] Figure 3 Application diagram of high-precision electrical parameter traceability system for dual-channel voltage comparator;

[0047] Figure 4 High-precision wideband bridge diagram for dual-channel voltage comparator. DETAILED DESCRIPTION

[0048] The application will be further described below with reference to specific embodiments in the accompanying drawings.

[0049] Reference Figures 1-4 A segmented frequency parameter correction method for a dual-channel voltage comparator, comprising the following steps 11-12:

[0050] Step 11. Construct a minimum error path and a hardware system with self-calibration.

[0051] The hardware system is designed based on the minimum error path and an equivalent distributed parameter model to realize wideband measurement of AC voltage and realize 10 -6 order precision traceability (see Figure 1 、 Figure 2 ). The AC / DC voltage conversion of the dual-channel voltage comparator (referred to as comparator) is based on the AC sampling principle. The AC signal is discretely sampled by a high-precision A / D converter, and various characteristic quantities of the AC signal are obtained through mathematical analysis and calculation. The AC voltage precision traceability is obtained through the "optimal A / D converter group AC measurement probability distribution model"; the "90° filter window" gap compensation sampling algorithm is adopted, the "D / A time division based on circular (0°~360°) phase linear verification technology" is adopted, and the linearity test result is better than FFT (Fast Fourier Transform) analysis by one order of magnitude under sinusoidal working condition. The "zero degree phase absolute calibration technology" and "AC amplitude consistency calibration" are adopted, the "equivalent frequency distribution parameter model" and the K coefficient multi-section correction model are adopted, and the amplitude and phase accuracy of the AC parameter target frequency band 40Hz~3500Hz under working condition is ensured.

[0052] The minimum error path measurement channel is composed of an operational amplifier buffer and an AD sampling chip.

[0053] Low-noise and low-drift input buffer operational amplifier, A / D chip with dynamic self-calibration function, DC reference chip, temperature compensation crystal, rated 4V AC voltage traceable to DC voltage 7V layout and wiring optimization, AC / DC conversion precision equivalent to AD chip manufacturer's highest DC reference, such as DC reference better than 5ppm, eliminates input channel and DC reference gain conversion resistor error source, ensuring AC voltage measurement precision and long-term stability of AC / DC comparator.

[0054] Equivalent temperature drift of A / D chip is less than 0.1ppm, AC sampling resolution is better than 0.1ppm. Annual stability of DC reference chip is 0.8ppm, temperature drift is 0.05ppm / ℃. Annual variation of temperature compensation crystal is less than 10E-6, phase jitter is less than 100ps.

[0055] In this embodiment, A / D chip uses AD677 / AD676 A / D chip with dynamic self-calibration function (equivalent temperature drift <0.1ppm / ℃, AC sampling resolution <0.1ppm) (see Table 1).

[0056] Measurement error caused by resolution:

[0057] Table 1 AD677 2s AC sampling 80000 point resolution

[0058]

[0059] Dynamic calibration error:

[0060] AD manufacturer gives a DC calibration index of 1 / 4LSB offset, if calibrated every 2 seconds, the calibration discreteness is obviously observed, statistical method (DC calibration 9 times every two seconds) reduces the converter to 1 / 12LSB (1.25x10 -6 ) error, experimental results prove this result.

[0061] DC reference chip uses LTZ1000ACH, annual stability is 0.8ppm, temperature drift is 0.05ppm / ℃. LTZ1000ACH about 7V DC voltage and AD input rated AC voltage 4V 120% match; thus eliminating the influence of voltage conversion gain resistor's time drift and temperature drift.

[0062] Dual-channel analog-to-digital device isolation power supply, separate electromagnetic shielding and device layout and wiring optimization.

[0063] Step 12. Construct the optimal equivalent frequency parameter and innovative AC sampling algorithm.

[0064] Define the frequency-related comprehensive K distribution coefficient: including clock, AD, input buffer, DC reference, shielding, ground, spatial leakage electromagnetic field, power supply, PCB device layout and wiring, which are equivalent to virtual frequency-sensitive R, C, L devices connected in series to the circuit, and it is particularly important to establish an equivalent correction model for this hidden characteristic error source.

[0065] The actual measurement temperature drift of the comparator is less than 0.1 ppm / ℃. The amplitude component of the equivalent frequency parameter of the hardware system is 10E-14, and the phase shift component is 10E-10;

[0066] The innovative AC sampling algorithm uses real-time polynomial data reconstruction technology, and the polynomial data reconstruction covers the audio frequency range of 40Hz~3500Hz and the power frequency range of 0~256 harmonics.

[0067] AC-DC conversion error:

[0068] The AC-DC conversion error is an inherent error of the AD conversion device, which is determined by design and production process and cannot be eliminated by statistical method on an AD converter. The present application eliminates the AC-DC conversion error by testing 9 AD samples traced to the highest DC reference of the manufacturer (better than 5ppm) to obtain the correction value from the probability distribution center value.

[0069] As described above, the error introduced by AD conversion is composed of AC sampling resolution error and AC sampling system error, and the two errors are not related. In the limiting case, the two errors are combined linearly, and the error introduced by the AD converter is: 4.37ppm. Multiple comparators have passed the test of more than ten years of operation and do not need to be calibrated again.

[0070] The correction method further includes the following step 21:

[0071] Step 21. Construct the wideband sinusoidal wave measurement "gap compensation" algorithm system and implement it. Specifically as follows:

[0072] 1. Power frequency error simulation calculation and measurement;

[0073] The gap compensation scheme algorithm simulation test for the amplitude measurement of the comparator shows that: by adding the decimal weight of the gap to the effective value integral formula, this leakage type error can be effectively suppressed. The maximum amplitude measurement error is not more than 0.088ppm under the test condition of 80000 points of power frequency test, and the maximum measurement standard deviation is not more than 0.6ppm; the maximum phase measurement error is not more than 1×10-6°, and the measurement standard deviation is not more than 0.5×10-4°.

[0074] 2. Audio error simulation calculation and measurement;

[0075] In order to solve the problem of small number of AD sampling points in signal period and sharp increase of gap error, innovative (conjugate +90° window) must be introduced. For example, 1000 points per cycle at 50Hz, 16.6667 points per cycle at 3000Hz, simple weight compensation is not enough. The comparator designed a high-efficiency compensation model for measuring error under audio working condition, which greatly reduced the influence of "equivalent distribution parameters" to 1 / 10~1 / 100 of ordinary measurement method.

[0076] 3. AD sampling signal initial phase will affect amplitude and phase calculation accuracy periodically. Follow this rule to add digital phase shift technology: a. The error of two times amplitude calculation of sampling sequence with 90° interval is reduced by 1~2 orders of magnitude than single calculation; b. After the influence of effective value is greatly reduced, the original phase of two channels is in any state, and digital phase shift is used to make them 90° and symmetric to adjacent coordinate axes. Through low-power calculation, it is converted to small phase, and the small phase plus digital phase shift phase realizes accurate phase measurement.

[0077] 4. Gap compensation example: simulation found that the gap error of initial phase of sampling has periodicity: 0° or 180°, 90° or 270°, the maximum effective value calculation error, the standard deviation is small, 45°, 135°, 225°, 315°, the calculation error is the smallest, but the standard deviation is large. If the initial phase calculation is taken twice with 90° interval, the average result can reduce the gap error. The arrangement of gap compensation sampling experiment results are as follows (algorithm: 0° and 90° product calculation results are averaged to eliminate gap error):

[0078] 50Hz: 0° sequence AD code square product calculation result is 185525276.74;

[0079] 90° sequence AD code square product calculation result is 185525060.99;

[0080] The calculation of discrete error is 0.58ppm;

[0081] 2000Hz: 0° sequence AD code square product calculation result is 185393111.54;

[0082] 90° sequence AD code square product calculation result is 185387813.47;

[0083] The calculation of discrete error is 14.28ppm;

[0084] 3000Hz: 0° sequence AD code square product calculation result is 185631832.38;

[0085] 90° sequence AD code square calculation result is 185630245.32;

[0086] The calculated discrete error = 8.55 ppm.

[0087] Considering the square root 1 / 2 weight, the measurement gap error correction amounts of 50Hz, 2000Hz, 3000Hz are 0.29ppm, 7.14ppm, 4.27ppm respectively. The corrected gap error tends to zero.

[0088] 5. Asynchronous sampling compensation simulation based on symmetry theory:

[0089] 1) Data analysis of power phase test, four zero values appear in 16 initial phase data, better than 10E-8 urad, and there are four maximum and minimum values.

[0090] 2) After arranging the data, the similar situation of 0.5L and 0.5C conjugate symmetry characteristic quantity when standard electric energy meter is found, that is, any small phase difference will show the characteristics of the same magnitude and opposite direction of power impact at these two characteristic points, while the above item 3 small power algorithm changes dynamically and slightly with sampling frequency and point number.

[0091] 3) Two-channel sequence arrangement is 90° to each other and symmetric with adjacent coordinate axes, at this time, the calculation of measurement error is two orders of magnitude higher than the simple item 3 small power algorithm, which is much lower than the accuracy, resolution and noise limit level of the signal source audio phase, and the effective value after 90° filtering is measured as follows:

[0092] 3000Hz: 90° calculation result is 0.17°; conjugate + 90° calculation result is 359.9988496°.

[0093] *Note: The limit accuracy and resolution of the signal source 3000Hz double-channel output, that is, the result of the innovative algorithm is the true value (although the sampling working condition at this time is 16.66667 points per cycle).

[0094] 6. Double-channel phase absolute calibration and amplitude consistency calibration (see Table 2).

[0095] Table 2 6100A signal source single voltage output to comparator double-channel self-check

[0096]

[0097] In summary, the application innovatively designs a "90°+conjugate" filter window, that is, the sampling sequences are 90° to each other and the whole is in a coordinate axis symmetrical position. Due to the characteristics of the function y=acos(x), that is, when y=90 degrees, the derivative (slope) of the function is maximum, in view of this, the hardware system innovatively calculates the phase by digitally shifting the arbitrary angle signal sequence to 90 degrees to each other, thereby improving the calculation accuracy of the phase. The simulation of the project on the algorithm shows that the 0~360° phase linearity is only 0.0005° and changes periodically (this phenomenon is "cosine leakage") under the ideal AD (according to 8000 points, the actual 0.88ppm resolution) condition calculated according to the definition. The application adopts the innovative 90-degree filter window, that is, the so-called "reactive measurement method", that is, after the adjacent 90° instantaneous values of the two channels are calculated and the preset time is added back, the simulation result shows that the non-linear error is 0.00001°. The problem of rapid increase of gap error with the increase of signal frequency is solved, and the problem of amplitude and phase algorithm for non-integer period sampling of alternating current signal by the existing filtering method is solved.

[0098] The correction method further includes the following step 31:

[0099] Step 31. Amplitude error correction and phase delay correction in sub-ppm level: use i to represent the segment number, ωi represents the characteristic frequency point parameter signal frequency of the segment, ωi=2*π*signal frequency, fi represents the characteristic parameter measurement error amplitude component of the segment, δi represents the characteristic parameter measurement phase shift component of the segment, Kfi and Kti represent the calculation coefficients of the frequency segment, and fi=ωi^2*Kfi(ωi) and δi=ωi*Kti(ωi) are defined; that is, the continuous frequency working condition measurement is solved by using segmented constant. When 3000Hz, the amplitude component is calculated as 0.56ppm, and the inverse calculation Kfi=0.56E-6 / ((2*π*3000)^2)=1.576*10E-15 is obtained, and the phase shift component is calculated as 0.000017°, and the inverse calculation Kti=0.000017 / 180*π / (2*π*3000)=1.574E-11 is obtained. The above K coefficient expression is obtained by inversely calculating a plurality of frequency point errors, and the frequency segment K coefficient is used as a constant to calculate the error compensation amount: f=ω^2*Kfi(ωi), δ=ω*Kti(ωi). When the FFT transform is integrated with the frequency parameter K coefficient compensation, the harmonic frequency parameter correction problem can be solved.

[0100] 1. According to the self-checking calculation in Table 2, the amplitude is consistent after the correction from 1000Hz reaches the sub-ppm level accuracy, the maximum delay of the phase absolute zero degree is 0.0001° / 2500Hz, and the time delay is 0.11ns, and the correction will further reduce.

[0101] 2. Influence of authoritative agency uncertainty:

[0102] Multiple comparison instrument respectively multiple NIM calibration, comparison, respectively to NRC (Canada National Institute of Metrology), PTB (Germany National Institute of Metrology) calibration, comparison. After data research: the power frequency measurement of each authority showed considerable consistency, but there is a non-consistent problem of high frequency working condition, obviously the research of equivalent distribution parameter in this direction still has a larger space.

[0103] 3. Final data correction decision: (see table 3, table 4, table 5).

[0104] Table 3 NIM authority comparison data

[0105]

[0106] Table 4 NIM authority audio and harmonic part calibration data

[0107]

[0108] Note: NIM audio detection system voltage current uncertainty is 20ppm, power uncertainty 50ppm

[0109] Table 5 comparison instrument reference table 3 authority frequency parameter linear correction result

[0110]

[0111] Note: amplitude high frequency calibration number is no 5790A reference real-time comparison, using B, C phase (the same sampling sequence as A phase) as 0.0ppm reference, correcting A phase parameter calibration number.

[0112] Through self-checking and authority calibration and comparison, the detection model and algorithm verification are obtained.

[0113] The correction method also includes constructing wide frequency phase absolute calibration and channel amplitude consistency calibration segmented correction method, including:

[0114] 1) Calibration frequency (Hz) includes: 50.0032, 122.0703, 244.1406, 488.2813, 976.5625, 1953.125, 3004.808.

[0115] Frequency segmentation (Hz) includes:

[0116] F0: 40≤F0<100; F1: 100≤F1<200;

[0117] F2: 200≤F2<400; F3: 400≤F3<800;

[0118] F4: 800 ≤ F4 < 1800; F5: F5 ≥ 1800.

[0119] Calibration frequency point calculation: F0 is 50.0032 Hz

[0120] Crystal 100M / 512 (frequency division number control signal source DA point number about 4000 points or so) / 50 (preliminary selection frequency point)=3906.25 (DA waveform synthesis point with decimal), then F0=100 / 512 / 3906=50.0032Hz, the time gap error is determined by the frequency measurement error (the frequency measurement error is about 10E-7, and the high frequency will increase), the actual measured waveform distortion is about 0.012%, and the distortion is 0.038% when F0=50Hz, at this time the gap error changes with the initial phase of the comparator sampling sequence. In order to solve the influence of frequency measurement error and gap error on other measured electrical parameters, gap compensation is introduced to improve the measurement accuracy.

[0121] 2) Calibration phase point, absolute calibration of phase error by inputting two channels of wide frequency low distortion single voltage.

[0122] 3) Since the "gap" of frequency point selection tends to zero, the signal source waveform distortion is 0.01% at power frequency, and the signal source waveform distortion is less than 0.025% at high frequency, which ensures the accuracy and stability required for calibration. The actual measured amplitude parameter is 10E-13, and the phase parameter is 10E-10.

[0123] The actual segmented frequency parameters are as follows:

[0124] Kf[6]={ -1.0E-13,-1.0E-13,-1.0E-13,-1.0E-13,-8.0E-14,-5.0E-14};

[0125] Kt[6]={ 1.99E-10,1.99E-10,1.99E-10,3.99E-10,3.99E-10,3.99E-10};

[0126] Note: The amplitude component of the above frequency parameters shows a long linear segment in the low frequency band, indicating that the L component of the comparator in the low frequency band is very small and can be ignored. The 1000Hz and 2000Hz are corrected for capacitive error. Similarly, the time delay component of the frequency parameter is small in the low frequency band, and the 500Hz, 1000Hz and 2000Hz are fine-tuned and corrected, proving that there is a small difference in time delay between the two channels. Correction amount calculation (2000Hz):

[0127] Amplitude correction increment= - (2*π*1953.125)^2*(-5E-14 + 1E-13)= 7.53ppm;

[0128] Phase correction increment = -1953.125 * (3.99E-10 - 1.99E-10) * 360 = -1.41E-4°;

[0129] Corrected time delay (2000Hz) = 0.00001 / 360 / 1953.125 = 1.42E-11 = 14 (ps).

[0130] 4) After the amplitude frequency parameter consistency calibration is completed, absolute calibration is implemented. That is, after the amplitude frequency parameter consistency calibration is completed, the absolute calibration is implemented by referring to the data of the uncertainty authority. Multiple comparators have undergone a high stability test of more than ten years of operation and free recalibration.

[0131] In the present application, the double-channel voltage comparator can be used as a precision audio bridge in the quantity transfer application (precisely reaching 15ppm), including:

[0132] 1) The double-channel wide-frequency zero-phase absolute calibration and amplitude consistency calibration of the comparator meet the technical requirements of the two precision detection arms of the bridge. The resolution of the double-channel voltage comparator is better than 0.1ppm and has good stability, and can be applied to the technical field of non-standard voltage proportional comparison, impedance measurement, and transducer measurement by using the knowledge of metrology "transfer precision".

[0133] 2) The precision test principle of the standard impedance is used to solve the precision measurement of the AC resistance time constant (the precision of the constant τ = 10E-8~10E-10).

[0134] Of course, the following conditions are required when applying:

[0135] 1. Auxiliary measurement component manufacturing: a. The comparator needs external analog, digital interface, automatic calibration program, etc. as the core of the precision bridge; b. Sectional equipotential shielded inductive voltage divider; c. Sectional equipotential inductive current shunt; d. Standard capacitor bank and standard resistance bank; e. When the comparator is used as a bridge, its absolute error does not affect the measurement result; f. By using the concept of metrology "transfer precision", under the condition that the test conditions and environment do not change in a short time interval, the "transfer precision" of the comparator tends to its calculated resolution of 0.1ppm or even higher (data quantity is converted into data quality typical example).

[0136] 2. Special connection shielded cable: The shielded connection cable adopts double shielding to prevent various electromagnetic interferences.

[0137] 3. Wide-frequency high-stability low-distortion power source.

[0138] 4. Standard voltage, current transformer, and standard 10mA traceable special high-power standard resistance and active impedance compensator.

[0139] The following is a specific application case

[0140] a. Segmented equal potential shielded current transformer: the working principle of segmented equal potential shielded current transformer is equivalent to that of segmented equal potential shielded voltage divider, and the output of the voltage divider can be used as the input of the current transformer for processing. Its calibration method is also to measure the segment voltage by using the impedance segment (see Table 6).

[0141] Table 6 Standard impedance method for calibrating segmented equal potential shielded voltage divider

[0142]

[0143] Note: In the table, the main sensing part 100 is 10V, and the auxiliary sensing part 10 is 4V. Since the accuracy of the object to be calibrated is very high, four measurements are required, i.e. comparing the reference of channel A and the measurement of each segment voltage of channel B to obtain 10 sets of data, and then calculating the average value and segment error. After comparing the reference of channel B and the measurement of each segment voltage of channel A, the average error of each segment is calculated. The above two operations are repeated by reversing the high and low ends of the power supply. Finally, the tap error of the voltage divider is synthesized. The process algorithm eliminates the errors of the comparator, power supply, and wiring.

[0144] b. Composition of electrical parameter traceability system Figure 3 )

[0145] In the figure, the input A channel of the comparator is the secondary voltage of the voltage transformer 4V, the primary voltage is 400V~50V, the B channel is the secondary current of the current transformer 0.5A~0.00005A, and the primary current is 100A~0.01A. The measurement uncertainty of the system is 15ppm.

[0146] c. Composition of impedance measurement system Figure 4 )

[0147] In the figure, the input A channel of the comparator is a standard impedance transformer, which can be a standard capacitor, a standard resistor, a standard U / U, I / U transformer, etc.

[0148] In the figure, the input B channel of the comparator is the impedance to be tested, such as R, C, L, etc. The measurement object is the time constant of AC resistance, the dielectric loss of capacitor, and the value of inductor, etc. The measurement uncertainty of the system is 15ppm.

[0149] The above is only a preferred embodiment of the present application, and it should be noted that for those skilled in the art, without departing from the structure of the present application, a number of modifications and improvements can be made, which will not affect the effect and practicality of the present application.

Claims

1. A method for calibrating segmented frequency parameters of a dual-channel voltage comparator, characterized in that: The following steps are involved: Step 11. Construct the minimum error path and the hardware system with self-calibration: The minimum error path measurement channel is composed of an op amp buffer and an AD sampling chip; The system utilizes low-noise, low-drift input buffer op amps, an A / D chip with dynamic self-calibration, a DC reference chip, and a temperature-compensated crystal. A dual-channel electrical isolation design optimizes the layout and routing for traceability from a rated 4V AC voltage to a 7V DC voltage. Statistical methods are used to trace the AC / DC conversion accuracy back to the AD chip manufacturer's highest DC reference. By eliminating the gain conversion resistor error source in the input channel and DC reference, the AC / DC comparator ensures the accuracy and long-term stability of AC voltage measurements. Step 12. Construct the optimal equivalent frequency parameters and innovative AC sampling algorithm: The amplitude component of the equivalent frequency parameter of the hardware system is 10E-14, and the phase shift component is 10E-10; The innovative AC sampling algorithm uses real-time polynomial data reconstruction technology, which covers the audio frequency range of 40Hz~3500Hz and the power frequency harmonic measurement range of 0~256.

2. A dual-channel voltage comparator segment frequency parameter calibration method according to claim 1, characterized in that: The following steps are also included: Step 21. Construct and implement a "gap compensation" algorithm system for broadband sinusoidal wave measurement. Innovatively design a "90°+conjugate" filter window, where the sampling sequences are 90° apart and symmetrical about the coordinate axis. This solves the problem of rapidly increasing gap error as the signal frequency increases, and addresses the algorithmic difficulties of existing filtering methods in sampling amplitude and phase of AC signals with non-integer periods.

3. The method for calibrating segmented frequency parameters of a dual-channel voltage comparator according to claim 1, characterized in that: The following steps are also included: Step 31. Sub-ppm amplitude error correction and sub-ns phase delay correction: Use i to represent the segment number, ωi to represent the characteristic frequency parameter signal frequency of the segment, ωi=2*π*signal frequency, fi to represent the characteristic parameter measurement error amplitude component of the segment, δi to represent the characteristic parameter measurement phase shift component of the segment, Kfi and Kti to represent the frequency band calculation coefficients, and define fi=ωi^2*Kfi(ωi), δi=ωi*Kti(ωi); that is, use segmented correction constants to solve the continuous frequency working condition measurement, and when the FFT transform is integrated into the frequency parameter K coefficient compensation, the problem of harmonic frequency parameter correction can be solved.

4. A method for calibrating segmented frequency parameters of a dual-channel voltage comparator according to any one of claims 1 to 3, characterized in that: Construct a segmented correction method for broadband phase absolute calibration and channel amplitude consistency calibration, including: 1) Calibration frequency points include: 50.0032, 122.0703, 244.1406, 488.2813, 976.5625, 1953.125, 3004.808; Frequency segments include: F0: 40≤F0<100; F1: 100≤F1<200; F2: 200≤F2<400; F3: 400≤F3<800; F4: 800≤F4<1800; F5: F5≥1800; 2) Calibrate the phase point by inputting a broadband, low-distortion single voltage into the two channels to perform absolute calibration of the phase error; 3) Since the "gap" of the frequency selection tends to zero, the signal source waveform distortion is 0.01% at the power frequency and <0.025% at the high frequency, ensuring the accuracy and stability required for calibration. The measured amplitude parameter is 10E-13 and the phase parameter is 10E-10. 4) After the amplitude and frequency parameter consistency calibration is completed, perform absolute calibration.

5. A dual-channel voltage comparator segment frequency parameter calibration method according to claim 4, characterized in that: The dual-channel voltage comparator can be used as a precision audio bridge in value transfer applications, including: 1) The comparator's dual-channel, wideband, zero-degree phase absolute calibration and amplitude consistency calibration meet the technical requirements for the two precision detection arms of a bridge. The dual-channel voltage comparator boasts a resolution better than 0.1ppm and excellent stability. Leveraging metrology's "transfer accuracy" knowledge, it can be applied to non-standard voltage ratio comparison, impedance measurement, and transformer measurement. 2) Use the standard impedance precision test principle to solve the precise measurement of AC resistance time constant.

Images