Frequency Domain Windowing (Part 2)

        In the previous blog post ( http://bbs.ednchina.com/BLOG_ARTICLE_3005040.HTM ), I introduced the window factor used in the oscilloscope FFT transform, including why the window factor is needed and how to choose the appropriate window factor. In this blog post, we will continue to introduce the key points related to the oscilloscope FFT function and topics that are often confusing and difficult to understand.

      One of the biggest problems when performing an FFT is aliasing. Aliasing occurs when an oscilloscope does not sample a signal fast enough to accurately capture the high frequency components of the signal. Its higher frequencies then appear as lower frequencies, or aliasing, when performing an FFT. Below is an example of an aliasing waveform in the time domain. 

        The maximum frequency that can be input into a sampler (such as an oscilloscope) without generating aliasing is 1/2 the sampling frequency. Even if the fundamental frequency of the signal is less than 1/2 the oscilloscope sampling rate, you still need to pay attention to the harmonics of the fundamental frequency and, if it is a complex waveform, the high frequency components of the signal. These high frequency components may be greater than 1/2 the sampling rate and will generate aliasing. In the FFT, this appears as frequencies folded into the picture.

        When a false signal occurs, there are three ways to identify it:

        1. The fast rising edges in the waveform generate many high frequency harmonics. The amplitude of these harmonics generally decreases as the frequency increases. The figure below shows these harmonics folded back into the picture at the Nyquist point, making them easy to identify. 

[Image content:]

FFT display: FFT display screen

The harmonics reflect back into the display at Nyquist point:

The aliases are actually higher frequency components displayed as lower frequencies:

        2. The second way to identify aliased signals is to select the channel to which the FFT is applied and rotate the horizontal scale knob to increase the sampling rate. This will increase the Nyquist frequency point, causing the aliased signals to spread out and no longer exist. The following figure illustrates how the signal in the above example is spread out by adjusting the sampling rate. 

[Image content:]

FFT display: FFT display screen

        3. Finally, adjust the frequency of the input signal if possible. As the input frequency increases, harmonics without aliasing will move to the right of the screen, and harmonics with aliasing will move to the left, as shown in the following figure.

        Note that the aliasing may also move to the right. As the fundamental frequency of the input signal increases, the aliasing will move to the left side of the screen. When they reach the edge, they will reflect back into the display and begin moving to the right again.

        Here’s another tip: You can use the oscilloscope’s bandwidth-limiting filter to filter out (or at least attenuate) the higher frequencies and minimize aliasing.