fft analysis device of digital oscilloscope

Abstract

The invention discloses an FFT analysis device of a digital oscilloscope, which buffers parallel data collected by a multi-channel ADC, then measures and obtains the signal frequency of each channel, calculates and obtains frequency multiplication parameters according to the extraction coefficient, and calculates the frequency of each channel data The signal is frequency multiplied to obtain a frequency multiplied signal. The synchronous extraction module reads the parallel data of each channel from the cached acquisition data, performs data extraction according to the frequency multiplied signal, and averages the parallel sampling data of a single channel to obtain the average value of each channel. Data, perform FFT operation and cache on the average data, read and display by the host computer. The invention takes the average value after synchronously extracting multi-channel and multi-channel parallel ADC acquisition signals, reduces the channel noise, and makes the frequency spectrum of the input signal more favorable for observation.

Description

technical field [0001] The invention belongs to the technical field of digital oscilloscopes, and more specifically relates to an FFT analysis device of a digital oscilloscope. Background technique [0002] In traditional digital oscilloscopes, the FFT analysis method is: convert the input analog current and voltage into discrete digital quantities through the AD sampling link, and then perform fast Fourier transform to calculate and obtain the complex values ​​of the fundamental wave and each harmonic. Existing digital oscilloscopes have a high sampling rate, traditional FFT is implemented by software, and the serial processing speed is slow, which cannot meet the real-time FFT requirements of high-speed digital oscilloscopes; the traditional oscilloscope FFT analysis data point extraction signal cannot be synchronized with the original acquisition signal, and because it cannot meet There is a truncation error due to the fixed-point sampling of the whole cycle, so the actua...

AI Technical Summary

Problems solved by technology

Existing digital oscilloscopes have a high sampling rate, traditional FFT is implemented by software, and the serial processing speed is slow, which cannot meet the real-time FFT requirements of high-speed digital oscilloscopes; the traditional oscilloscope FFT analysis data point extraction signal cannot be synchronized with the original acquisition signal, and because it cannot meet There is a truncation error due to the fixed-point sampling of the whole cycle, so the actual accuracy is not high; the noise spectrum component of the traditional FFT analysis method is too large, which affects the observation of the spectrum of the measured signal. The specific analysis is as follows:

[0003] (1) The traditional FFT analysis method takes a long time, resulting in a decrease in the waveform capture rate:

[0004] Traditional FFT analysis is realized by software. The high-speed digital oscilloscope has a fast transmission rate and multi-channel synchronous sampling. Because the software is serially processed, the processing speed is slow, and signal processing needs to consume more time. The data processing time and waveform capture rate Inversely proportional to the waveform capture rate of the digital oscilloscope will decrease

[0005] (2) The precision of traditional FFT analysis method is low:

The sampling algorithm cannot realize the fundamental frequency f of the measured signal 0 of 2 q ,q∈[1,2,3…] times the sampling point, and the sampling point enable signal and the fundamental frequency of the measured signal f 0 The phase is random, which leads to data truncation error and phase error reduces the analysis accuracy of FFT

[0007] (3) The traditional FFT analysis method cannot effectively reduce the noise floor of the channel:

[0008] Traditional FFT analysis cannot suppress channel noise, resulting in the spectrum of the measured signal with a low signal-to-noise ratio being easily overwhelmed by the noise spectrum, resulting in inaccurate analysis results, which is not conducive to reproducing the spectrum of the original signal, and brings great difficulties to the observation of the spectrum of the original collected signal. great disturbance

Images