Brief Introduction of Spectrum Analyzer
Introduction to Spectrum Analyzer
Like an oscilloscope, a spectrum analyzer is an instrument used for signal observation and analysis. Spectrum analyzer is mainly used for frequency domain measurement, which can measure signal power, occupied bandwidth, adjacent channel power ratio, etc. The spectrum analyzer is mainly the state of the receiver. If the receiver is highly sensitive, it can also be used for electromagnetic interference testing. Of course, it must be equipped with an inductive probe, which is mainly the ability to measure some small signals.
There are many spectrum analyzer parameters, related to frequency parameters and functions: start frequency, stop frequency, center frequency Center Freq, sweep width Span, resolution bandwidth RBW, video bandwidth VBW, sweep time Sweep, sweep points, etc. . The parameters related to amplitude include: reference level Ref Level, input attenuation Input Att., maximum mixing MaxMixL, dynamic range, etc.
In the time domain test, the horizontal axis is time and the vertical axis is amplitude. In the frequency domain, the horizontal axis represents the frequency, the vertical axis represents the effective power, and the spectrum analyzer measures the collection of signal frequency and power. The waveform of an arbitrary signal in the Time Domain state changes with time. Frequency Domain (Frequency Domain) describes the characteristics of a signal in a certain frequency range.
Time domain measurement and frequency measurement of cables
For example, when a signal (as shown in the figure below) is observed in the time domain, it is a continuous curve that changes with time, but when converted to the frequency domain, it is similar to pulses. I remember reading an article before. The example cited above is intuitive and easy to understand: many people can play the piano. If the concept of converting the piano into the time domain and frequency domain is that the music played by the piano is the time domain information, and the staff It is frequency information, so it should be easy to understand that human society is a very broad time domain environment.
When measuring signals in the time domain, some signals not only change with time, but also with frequency and phase. This requires further frequency domain analysis, which uses the Fourier transform. I vaguely remember when I was in school. When the teacher explained the Fourier transform to us, the teacher said: one or more sine waves with appropriate frequency amplitude and phase can superimpose any electrical signal in the time domain. That is to say, any time domain signal can become frequency domain signal. Of course, frequency domain signal can also be converted into time domain signal. Low-pass or band-pass conversion, converted to time domain mode), the time domain mode of the vector network is mainly the following three:
BAND PASS
LOW PASS，step response
LOW PASS，pules response
Main parameters of spectrum analyzer
Sensitivity
When we just turn on the spectrum analyzer, it will produce a background noise. The sensitivity index of the spectrum analyzer is related to the instrument’s ability to detect weak signals. under a certain resolution bandwidth, we display average noise level (DANL) is the sensitivity of a spectrum analyzer. “Average” means that the amplitude of the noise signal is constantly changing with time and frequency. This indicator is related to the settings of VBW, RBW and input attenuation.
When the input signal power level is less than the noise level of the spectrum analyzer, we will not display the signal; when the input signal amplitude is greater than the noise level, we will display and superimpose its noise on the input signal, and the result on the spectrum analyzer is an external signal. The sum of the level and the noise power of the spectrum analyzer; if the measured signal power is more than 10-20dB greater than the noise power of the spectrum analyzer itself, we can not ignore the noise power of the spectrum analyzer.
Example
Take Rigol’s DSA875 spectrum analyzer as an example: set the attenuation to 0dB under the button [AMPT], the resolution bandwidth of RBW and VBW and the video bandwidth under the button [BW/Det] are both 10 Hz, and the detection type is [Sampling Detection] or [Effective] The noise cursor measurement is more accurate in the mode of [Trace/P/F] button, the average times of the trace under the [Trace/P/F] button is more than 50, [Reference level] is set to -100dBm, [SPAN] Span is set to 500Hz, the tracking source is off (some If you don’t have a TG tracking source, you don’t need to set it up), turn on the preamplifier under [AMPT] (this one does not have this option), and the input impedance is 50 Ω from 20°C to 30°C.
we can only set the minimum value of the DSA875 RBW and VBW of my Rigol spectrum analyzer to 10Hz, not 1Hz. In short, the sensitivity of a spectrum analyzer is the ability detect small signals. Due to the reduction of the resolution bandwidth, the scanning time will slow down. And the input standing wave ratio (VSWR) will increase due to the input attenuation of 0dB. It increases the measurement uncertainty. If you want to test, you need a spectrum analyzer with a tracking source. At the same time cooperate with a VSWR bridge to achieve a voltage standing wave ratio test.
Reference level and input attenuation
The reference level refers to the amplitude level indication of the vertical scale position on the display, which is generally displayed with the coordinate value of the top grid, which is set in [AMPT] –> [Reference Level].
Input attenuation refers to the attenuation of the attenuator at the signal input end of the spectrum analyzer, which is set in [AMPT] –> [Input Attenuation]. When the input attenuation increases by 10dB, the noise level of the spectrum analyzer increases by 10dB, and the scale value displayed on the meter also increases by 10dB, but the level of the input signal will not change with the increase in attenuation, that is to say, the attenuation setting is not It will affect the authenticity of the input signal. Take the following figure as an example, the amplitude value changes when the input attenuation is 10dB and 20dB.
But if you look at the Mark function, the power level has not changed in fact (I use the antenna to transmit and receive, which will slightly affect it and it is normal).
The reason is: when the attenuator reduces a certain power level, the IF also increases the corresponding level to compensate, so that the signal amplitude displayed by the meter does not change. However, the increase in attenuation also increases the noise level, reduces the signal-to-noise ratio, and reduces the sensitivity.
Resolution bandwidth
Resolution characterizes the ability of a frequency analyzer to separate two input signals in response. The commonly used resolution bandwidth (RBW) represents the bandwidth of the IF filter. The smaller the resolution bandwidth, the smaller the noise. The lower the noise level displayed after passing through the detector, the figure below shows the change of noise level.
At present, the new-type spectrum analyzer adopts a fully digital intermediate frequency circuit. The intermediate frequency filter completely adopts a digital filter. Which can achieve a small RBW change step (1Hz~3MHz), and has good measurement accuracy.
Video bandwidth
The video bandwidth is actually the bandwidth of the video filter. It is a low-pass filter that provides the filtering or average value of the detector output. When there is a combination of noise and signal, we need the video filter. Its function can reduce the noise change of the detector output and reveal some signals. Which is close to the noise floor. If we measured the noise power, it can also help stabilize the measurement.
If we set VBW to 100KHz, it means to sample once every 100KHz. So you can see that the smaller the VBW setting, the smoother the test curve. VBW is the bandwidth of the filter after peak detection. It is mainly to make the test signal more smooth. We should note that reducing VBW can smooth the noise signal. But it doesn’t get the average power of the noise signal. Generally set the video bandwidth to be less than 0.1~0.01 times the resolution bandwidth.
The relationship between the bandwidth of the video filter and the resolution bandwidth is: a narrow resolution bandwidth can reduce the noise before the detection, thereby reducing the noise output level of the detector; the narrowband video filter Fluctuations can smooth and reduce the noise after the detection, but cannot reduce the average power level of noise.
Preamplification
The preamplifier (PreAmp) is generally provided as an option, and its operating frequency is generally lower than 3GHz, and can usually be selected to be turned on in the [AMPT] button. For higher frequencies, to increase the sensitivity of the spectrum analyzer we need an external amplifier. I have tried to connect the spectrum analyzer to the earphone and antenna to receive FM broadcast. But the noise interference is too strong, so the definition is not high. We can receive it and turn on the pre-amplification.
There are several ways to improve the sensitivity of the instrument:
1–>Minimum resolution bandwidth
2–>Minimum input attenuation
3–>Minimize the video filter as much as possible
4–>Open pre-amplification
The above settings to increase sensitivity may conflict with other measurements:
1–>Small resolution bandwidth will greatly increase scanning time
2–>0dB attenuation will increase the standing wave ratio
3–>Preamplifier will affect the dynamic range of the spectrum
Detection method
RMS (effective average): To average the linear power of the signal we use the average detection method. This method is suitable for the average power of the modulated signal and the ACPR (adjacent channel power ratio) test.
Avg (Average Detector): Perform average processing on the sampled point values in each frequency period. This method is suitable for spectrum testing of randomly changing signals.
Neg Peak (negative peak detection): extract the minimum value of each envelope voltage, this method is suitable for low-power signal detection.
Sample (sampling detection): extract the voltage value of each envelope at equal intervals. If the spectrum analyzer does not have an average detection method, this method is suitable for spectrum testing applications of randomly changing signals.
Peak (maximum value detection): extract the maximum value of each envelope voltage, this method is suitable for point frequency signal test and signal search test.
Average way
In order to reduce the noise signals, the spectrum analyzer needs to average the measured signal. There are usually three methods: log average, video average (also called log average) and voltage average.
Video averaging is suitable for testing weak CW signals in a noisy environment. Logarithmic averaging will not affect the power of CW signals, but will smooth the noise signals.
we mainly using narrowband signals to test Voltage average under the background of broadband interference in EMI testing.
Power average only by using this test can the average power value of the measured noise signal be obtained.
The latest one is already a real-time spectrum analyzer, which can measure pulse signals.
