Here, the A/D’s Time Data has been transformed from the Time Domain into the Frequency Domain using the Fast Fourier Transform (FFT) provided by ScopeDSP™. We have transformed all 8192 data points of a 1.953125 MHz signal which was sampled at 5.0 MHz, and have not windowed the signal. The 1.953125 MHz signal falls exactly in the center of an FFT bin. The FFT bin spacing is: 5 MHz/8192 = 0.000610352 MHz, and the bin number is: 1.953125 / 0.000610352 = 3200. (We have deliberately chosen to feed the A/D a signal whose frequency falls in the exact center of an FFT bin.)

The spectrum is:

 

ADC Signal Spectrum
ADC Signal Spectrum

This Time Data can be called a “sine” since there is one just major component. Since this signal was captured live (not generated mathematically) it shows random “grass” across the board, as expected. The eye also spots a few other things. We see what we DSP professionals call a distinct hump of noise at the lowest frequencies. Also, a few distortion components poke out distinctly above the grass, most notably near 1.1 MHz.

Let’s zoom in around the fundamental:

ADC signal spectrum

Around the fundamental we see what we DSP Professionals call a “Christmas Tree*”. This is not unexpected. Phase noise on the input signal and jitter on the clock of the A/D can produce this. The system is actually performing quite well if the jitter components are below the levels of other distortion components. (As we’ll see later, the “Christmas Tree” would be a lot worse if the signal source had not been tuned to the center of an FFT bin.)

But what about that other component near 1.1 MHz?

Next, let’s zoom in there…