Superheterodyne Spectrum Analyzer

General Instructions

This spectral simulation is an interactive Java applet. You can change parameters by clicking on the vertical arrow keys. The five control buttons at the lower right are used to start (triangle) and pause (square) the simulation, to skip forward or back one section at a time (double triangles), and to change speed (+ and -).

After the simulation is complete, the start button takes you back to the beginning of the simulation. You may experience a delay at this point.

A spectrum analyzer is an extremely useful measurement tool that can show, in the frequency domain, information not readily recognizable with a time domain instrument such as an oscilloscope. For example, the frequency content of a signal (e.g. a fundamental sinusoid and one or more harmonic components) is readily identified with a spectrum analyzer, something difficult to identify from an oscilloscope display.

The most common type of spectrum analyzer, especially at high radio frequency and microwave frequencies is the Superheterodyne Spectrum Analyzer, shown in the simplified block diagram below. It operates in a fashion similar to a superheterodyne AM radio receiver, with the output in this case going to a CRT display rather than a speaker.

 


This experiment requires a Java-enabled Web Browser.

 

The illustrated Superheterodyne Spectrum Analyzer uses a fixed IF [Intermediate Frequency] filter and a sweeping LO [Local Oscillator] signal. Once the input power is limited by an attenuator (so the instrument isn't damaged), the analyzer combines the input signal and the LO through a device called a "mixer". The mixer output contains the input signal (Fsig), the LO signal (Flo), the sum and difference between these two signals, and various other frequency components related to these two signals.

If we know the LO frequency exactly, then by sending these frequency components through a narrow filter - the IF -we can measure both the amplitude and the frequency of the unknown. Whenever any of these components fall within the IF filter bandwidth, an ac voltage is produced that is related to the input signal's amplitude. This ac voltage is converted to a dc voltage by an envelope detector, and the results are displayed on the y-axis of the screen.

The response is typically displayed in units of decibels referenced to a milli-watt, or dBm. The log scale (dB scale) is used because it is a good mathematical tool to cover an extremely large dynamic range.

The narrower the bandwidth of the IF, the more information we can get from the signal. But a narrow bandwidth filter has one major drawback: sweeping the LO too fast will result in display distortion, one of the many tradeoffs in engineering.

The simulation shows a linear dc voltage from the sweep generator that drives the X-axis of the display, and at the same time varies the oscillator frequency of the LO.

Try it:

  • Using the DOWN ARROW key, set the signal frequency to 2.7 GHz.
  • START the simulation by hitting the FORWARD ARROW key with the mouse.
  • Notice the CRT display. Why are there TWO peaks when we only have ONE signal??

The answer:

Look at the LO frequency and the IF frequency. In this analyzer, the LO starts at 3.6 GHz, and that's the same as the IF center frequency. So at a signal frequency of dc, or zero Hz, the full LO signal will pass through the IF filter, and it will be displayed on the screen. In fact, this "LO feedthrough" is always present at 0 Hz and can be used as an accurate dc frequency reference. As the LO begins to sweep from 0 Hz to a higher value, the display will actually begin to show the filter characteristic of the IF.

As the mixer output of Flo-Fsig approaches the IF frequency, we begin to see the amplitude of the unknown (Fsig). The signal and LO frequencies are such that the display responds when the LO signal itself is within the IF filter bandwidth (LO feedthrough) or when Flo - Fsig = the IF frequency. When Flo - Fsig = Fif, we can find the unknown Fsig simply by changing the equation around: Fsig = Flo- Fif. Since Flo and Fif are known very accurately, it's easy to determine Fsig.

With a 3.6GHz IF frequency, and an LO frequency sweeping from 3.6 (Flo1) to 6.6GHz (Flo2), the x-axis of the display corresponds to input signals between: (Flo1 - 3.6) = 0 Hz and (Flo2- 3.6 =) 3.0 GHz.

  • Now set the signal frequency on the simulator to 0.6 GHz. Why are the two peaks interfering with each other?
  • What could you do to keep the peaks from overlapping?

The answer:

The IF filter bandwidth is beginning to interfere with the measurement. We could set the IF filter bandwidth to a smaller value, but we would have to remember to sweep slower to avoid amplitude distortion.

A high bandwidth IF means fast sweeping, but poor resolution. A low bandwidth means high resolution, but slow measurements. The term "resolution bandwidth" is often used to specify a spectrum analyzer.

The above simulation is for an all-analog instrument. Today, the digital analyzer has taken over the market. A modern spectrum analyzer contains a digital display, several microprocessors, fft or similar digital filters, log amps and compression amps for better dynamic range, firmware compensation for better accuracy and a host of other innovations.