Spectrum Analyzer Key Performance Indicators

Spectrum Analyzer Key Performance Indicators

As an analytical instrument, the basic performance requirements of a spectrum analyzer include:

1. Frequency indicators:

Measurement frequency range: reflects the spectrum analyzer’s ability to measure signal range;

Frequency resolution: reflects the ability of the spectrum analyzer to resolve signals separated by two frequencies.

2. Amplitude index:

Sensitivity: the ability of a spectrum analyzer to detect small signals;

Internal distortion: reflects the ability of the spectrum analyzer to measure large signals;

Dynamic range: A spectrum analyzer’s ability to analyze large and small signals simultaneously.

3. In addition, the performance of the spectrum analyzer also includes its analysis accuracy and measurement speed.

Measuring harmonic distortion or searching for signals requires the frequency range to extend from below the fundamental to beyond the multiple harmonics. Measuring intermodulation distortion requires a narrow span to observe adjacent intermodulation distortion products. Therefore, the first step is to select a spectrum analyzer with sufficient frequency and span. The second requirement is what frequency resolution? Measuring two-tone intermodulation places strict requirements on resolution.

The measurement frequency range of a spectrum analyzer is determined by its local oscillator range. The analysis frequency range of the spectrum analyzer can be expanded by using the harmonics of the local oscillator, and the external mixing method can also be used to extend the analysis frequency range to higher (75GHz; 110GHz; 325GHz, etc.).

frequency resolution

This example reflects the impact of the measurement resolution of the spectrum analyzer on the test results. The input physical signal is a signal with two frequency intervals. Only when the resolution of the spectrum analyzer is high enough will the characteristics of the signal be correctly reflected on the screen.

Many signal test applications require spectrum analyzers to have the highest possible frequency resolution.

Figure 1 Frequency resolution

The frequency resolution of a spectrum analyzer is related to its internal IF filter and local oscillator performance.

Among them, the influencing factors of the IF filter include:

Filter type; bandwidth; shape factor.

Local oscillator residual FM and noise sidebands are also factors that should be considered when determining useful resolution.

Analyze each item in turn. One of the first things to note is that an ideal CW signal cannot appear as an infinitely thin line on a spectrum analyzer, it has a certain width itself. When a passing signal is tuned, its shape is a representation of the shape of the spectrum analyzer's own resolution bandwidth (IF filter). Thus, if you change the bandwidth of the filter, you change the width of the displayed response. The data sheet for technical specifications specifies a 3 dB bandwidth, and other applications (EMC) define the filter bandwidth as a 6 dB bandwidth.

The local oscillator performance has an impact on the resolution because the intermediate frequency signal comes from the mixing of the input signal and the local oscillator signal, and the noise in the two signals is the sum of the powers.

The input signal phase noise performance is: 10kHz offset –110dBc/Hz;

The mixing local oscillator phase noise performance is: 10kHz offset –110dBc/Hz,

Then the phase noise performance of the mixed output intermediate frequency signal is: 10kHz offset –107dBc/Hz.

The test result of a single-point frequency signal on the spectrum is the frequency response shape of the intermediate frequency filter.

The shape of the filter is defined by its bandwidth (3dB or 6dB) and squareness factor. Both parameters affect the frequency resolution capability of the spectrum analyzer.

Figure 2 Definition of IF filter bandwidth and shape factor (rectangle coefficient)

In the two-tone test, the two signals are 10kHz apart and when RBW=10KHz, the instrument test can show two signal peaks. Obviously there is no problem in using a 10kHz filter to distinguish two-tone signals of equal amplitude.

The RBW of a spectrum analyzer is its ability to resolve signals of equal amplitude.

The conclusion drawn from the above analysis is:

The smaller the RBW of a spectrum analyzer, the higher its frequency resolution.

The IF filter 3dB bandwidth tells us how close signals of equal amplitude can be to each other and still be separated from each other (based on 3dB drop). Generally speaking, if the distance between the two signals is greater than or equal to the 3dB bandwidth of the selected resolution bandwidth filter, two signals of equal amplitude can be distinguished. Two signals in the two-tone test illustrate this meaning. When two signals are separated by 10 kHz, they are easily separated using a resolution bandwidth of 10 kHz. However, with a wider resolution bandwidth, the two signals appear as one.

Note: When two signals appear within the resolution bandwidth, due to the interaction of the two signals, a video bandwidth approximately 10 times smaller than the resolution bandwidth can be used to smooth their response.

Usually we need to measure signals with unequal amplitudes. Since the two signals describe the shape of the filter in our example, it is possible for the small signal to be buried in the filter skirt of the large signal filter. For two signals that differ by 60dB in amplitude, they must be separated by at least half the 60dB bandwidth (with an approximate 3dB drop). Therefore, the shape factor (the ratio of the filter's 60dB to 3dB bandwidth) is the key to determining the resolution of unequal amplitude signals.

Examples of frequency analyzers resolving unequal amplitude signals:

Testing of distortion products whose amplitude drops by 50dB at intervals of 10kHz.

If the shape factor of the 3kHz filter is 15:1, then the bandwidth of the filter down 60dB is 45kHz, and the distortion products will be hidden under the skirt of the test signal response. If switched to another narrowband filter (such as a 1kHz filter) with a 60dB bandwidth of 15kHz, the distortion product is easily observed (because half of the 60dB bandwidth is 7.5kHz, which is smaller than the sideband spacing). Therefore, the resolution bandwidth required for this measurement should not be greater than 1kHz (<>

Range of filter shape factors:

Analog filter: 15:1 or 11:1

Digital filter: 5:1

Conclusions from the above analysis:

The smaller the rectangular coefficient of the spectrum analyzer, the higher its frequency resolution for unequal amplitude signals.