Research On No-interval Frequency Measurement Which Is Equivalent To Phase-comparison

As one of the seven international standard physical quantities,time is a component of many derived physical quantities,and it is also the most accurate measurement unit in the current basic measurement units.With the continuous development of science and technology,more and more high requirements are put forward for accurate frequency measurement in the fields of national defense,science and technology,frontier scientific research,industrial and high-speed communication.This paper analyzes the deficiency of the interval frequency measurement in characterizing the long-term stability of the frequency source.Among Advanced arbitrary frequency measuring equipment,the measurement principle of many devices is frequency measurement with intervals.This type of measurement method omits the phase information between two sampling periods,resulting in the measurement results can not fully reflect the long-term frequency stability of the measured signal.Use the Allan variance in time domain on the characterization of the measurement,The stability curve is 1/τ^1/2.When the method of phase comparison is used for measurement,no phase information is omitted between two measurements,and the stability curve of the measurement is 1/τ.However,the phase comparison is usually only used to measure the same frequency standard,which cannot meet the demand of measuring any frequency source.In this paper,two no-intervals frequency measurement methods are proposed.Because the spaceless frequency measurement method does not omit any phase information,the measurement result can be equivalent to the phase comparison,and the measurement range is large,which can measure any frequency signal between 1 and 80MHz.The first method is frequency measurement method based on phase tracking.In this method,DDS is used as a frequency synthesizer.Firstly,the frequency of the measured signal is measured to the magnitude of Hz through the gate method,and the frequency of DDS is adjusted close to the measured signal,so that the frequency is less than 10 Hertz.Then phase detector is used to detect the DDS signal and the measured signal,and the detector whose voltage is linearly related to the phase of the two signals is obtained.Then use ADC to measure the signal,calculate the frequency of the signal according to the voltage change speed of the signal,and then adjust the DDS frequency further close to the measured signal frequency to 50 m Hz.Then into the precision measurement,the use of operational amplifier to amplify the signal,and phase compensation is carried out for the DDS continuously,so that the voltage of the discriminator is always in the sampling range of ADC.The voltage sampling and DDS phase compensation are carried out in a period of one second.The phase change speed of the detector is calculated by using the ADC sampling value of two adjacent measurements and the phase of the DDS,that is,the frequency difference between the DDS signal and the measured signal.Then the frequency of the measured signal can be calculated according to the frequency of the DDS.Experimental results show that the second stability of this method can be achieved3.3×10^-13/swhen measuring 10MHz signal.The second method is frequency measurement method based on frequency tracking.In this method,the frequency difference between the output frequency of the DDS and the measured signal is kept about 1kHz.The phase relation between the DDS signal and the measured signal is converted into the amplitude change of the sawtooth wave by using the phase detector with linear characteristics,and the frequency of the sawtooth wave obtained is also1kHz.Through theoretical analysis,different phase detectors are selected for experiments,so that the probability of the reference signal and the measured signal colliding along the edge can be reduced as much as possible,and the jitter of sawtooth wave jumping can be reduced.Then the high-speed clock is used as the signal source of FPGA and the time of edge jumping of each sawtooth wave is recorded.The reasons for the large jump of the sampling value in FPGA are analyzed,and the optimization scheme is proposed in the hardware circuit design and software algorithm.On hardware by adjusting the cut-off frequency of the low pass filter,the jump edge of the signal is steeper.On the software by using the least squares curve fitting method to map the time for each jump along the curve.Reducing single trigger instability of the measurement on the final measurement results,the influence of the curve of the slope is the frequency of discriminator.Finally,the frequency of the measured signal is calculated according to the frequency of discriminator and the frequency value of DDS.Experimental results show that the second stability of this method can be achieved5.7×10^-13/swhen measuring 10MHz signal.