The delay correction in the
frequency domain calculates the delay through the phase difference. The phase difference calculation method is ambiguous in the calculated phase difference because of the periodicity of the trigonometric function, so that the range of the corrected delay can only be limited to one
signal cycle; The
delay calculation in the
time domain is to calculate the phase difference through the signal expression, and to calculate the delay through the phase difference, so calculation errors will be introduced
[0004]The traditional
processing method is to calculate the phase difference between channels at a single frequency point, and calculate the time between channels according to the relationship between phase difference and frequency delay, and correct and compensate the corresponding channels. The published literature mainly includes: "
Phase Difference Method Corrects Multi-Channel AD Acquisition High-precision
Time Error" in "Electronic World", extracts the phase difference and then calculates the synchronization time through the principle of arcsine transformation Error, due to the periodicity of the sine function, the calculated phase difference is ambiguous. Therefore, the calculation principle of arcsine determines that it can only calculate the sampling delay within a single signal period. For delays greater than one signal period, its The method will fail; the "
Phase Difference Accurate Measurement Method Applicable to
Asynchronous Sampling" in "
Power Grid Technology", the phase difference measurement method based on the
correlation analysis method, also has the periodic problem of the sinusoidal function, and is based on the existence of
voltage and current It is carried out under the premise of transformers, and its application range is very limited due to high hardware requirements; "Design and Implementation of Multi-channel Synchronous
Data Acquisition System" in "University of Electronic Science and Technology", calculates the time delay between channels through the DFT
algorithm, There is also the
ambiguity problem of the phase difference, and the sampling frequency must be an integer multiple of the
signal frequency; the "
Sine Wave Fitting Method to Evaluate the Inter-channel
Delay of the
Data Acquisition System" in the "Journal of Testing Technology" uses an
oscilloscope to observe the inter-channel waveform Rising edge delay,
naked eye observation, the error is relatively large; "Information and Computer (Theoretical Edition)" "Synchronization
Time Difference and
Phase Difference Analysis Caused by Small Phase Difference Calculation" calculates the phase difference through the theoretical formula of the
time domain, and introduces the calculation error, and
engineering practice requires multi-stage filters to filter out redundant calculation items; "Phase Difference Measurement Method and
Error Analysis Based on Digital
Orthogonal Transformation" in "Journal of Circuits and Systems" requires that the sampling frequency be an integer multiple of the
signal frequency, and the same There is a problem of phase difference
ambiguity and calculation errors are introduced; "PhaseMeasurement and Adjustment of Digital Signals Using Random SamplingTechnique" of "IEEE International Symposium on Circuits and Systems" improves the
relative phase accuracy through random sampling statistical techniques, but it needs to design complex hardware circuit
[0005]The calculation on the time domain waveform requires a relatively high sampling rate, and there are errors caused by the calculation; the phase difference calculation in the frequency domain has the problem of phase difference
ambiguity, which leads to When correcting a channel with a delay greater than one signal cycle, the channel delay cannot be solved correctly; using an
oscilloscope to observe the waveform delay in the time domain, the delay accuracy cannot be guaranteed
These methods can directly or indirectly calculate the time delay of sampling out-of-
sync between channels, but the process is relatively complicated, the requirements for hardware design are relatively high, and the
correction method is relatively poor in versatility and has great limitations.