It is very important to analyze the frequency components of the signal, because they often cause noise in the design. Once it exceeds the allowable tolerance, it may lead to device failure and dysfunction. Serious may also lead to voltage spikes, damage devices. If we don’t do the right test at the time of design, then the above problems are likely to occur. So how to analyze the frequency component of the signal?

You may think that only spectrum analyzer can do this job, but in fact oscilloscopes can also do it. Besides time domain analysis, oscilloscopes also have FFT function, which can be used to do this job. FFT is the abbreviation of fast Fourier transform. In short, FFT is actually an algorithm, which can help us separate the time domain signals, and then convert these separated signals to the frequency domain. At this time, the oscilloscope will convert from the time domain to the frequency domain, showing the relationship between the signal amplitude and frequency.

As shown in the GIF diagram below, you can clearly see how the oscilloscope converts the signal from time domain to frequency domain.

The menu bar of FFT contains the selection of FFT operation spectrum type. You can select the line or decibel as the amplitude to be drawn on the oscilloscope display in v-hz or DB Hz respectively. When FFT is turned on, you can see that the time base of horizontal axis changes from time to frequency, and the unit of vertical axis changes to V or dB.

Below the spectrum type is the selection of trigger source, which is easy to understand. To carry out FFT operation on which channel, we choose which channel as the source.

Below the source are four different FFT windows, namely rectangular window, Hamming window, Blackman window and Hanning window. So why do FFT have different window choices?

Because FFT algorithm can only get the information of the sampling point when calculating the spectrum signal sampling, it is impossible to measure and calculate the infinite signal, but take its limited time segment for analysis, so it ignores the data information in the sampling interval, which is inevitable, also known as the fence effect. Oscilloscopes transform finite length time records by FFT. FFT algorithm assumes that time domain waveform is repeated. In this way, when the period is an integer, the amplitude of the waveform at the beginning and end of the time domain waveform is the same, and the waveform will not be interrupted. However, if the period of the time-domain waveform is not an integer, the amplitude of the waveform at the beginning and end of the waveform will be different, resulting in high-frequency transient interruption at the junction. In the frequency domain, this effect is called leakage. Therefore, in order to avoid leakage, the original waveform is multiplied by a window function to force the values at the beginning and end to be zero.

Different window functions adopt different algorithms and have their own advantages in different situations. Window function can change the frequency-domain waveform and make the spectrum form a convenient appearance for us to observe, but it will not eliminate the spectrum leakage in essence. Different window functions have their own unique characteristics. We only need to choose according to the measurement needs.

At the same time, the following points should be paid attention to during the measurement:

1. Because FFT is a mathematical function, the more data it processes, the more accurate it is. Therefore, when measuring, we should enlarge the storage depth and time base as much as possible, so that the frequency resolution can be higher. As shown in the following two figures, we can see clearly that the FFT effect is much better when the time base is set to 200 μ s and 2ms respectively.

But it should also be noted that the longer the time domain signal length is, the better, because the storage depth of the oscilloscope is limited, the longer the waveform recording time is, the lower the sampling rate is, which may lead to source waveform distortion. Generally speaking, it is appropriate to have at least 4 to 8 waveform periods in the time domain diagram.

2. The signal with DC component or deviation will lead to the error or deviation of FFT waveform component. In order to reduce DC component, we can choose AC coupling mode.

3. When acquiring periodic signals, the average sampling mode should be used to reduce the signal noise. It is suggested that the average should not be less than 16.

FFT can help to find the noise interference source in electronic measurement, test the impulse response of filter and system, jitter analysis, harmonic power analysis, electromagnetic interference analysis, frequency response analysis, etc.

Editor in charge: GT