How to improve the sensitivity of a spectrum analyzer?

The spectrum analyzer will display the noise level when no signal is added. The noise generated by the spectrum analyzer itself comes mostly from the first stage of the intermediate frequency amplifier. The sensitivity of the spectrum analyzer is defined as the average noise level (DANL) it displays. This indicator is related to the instrument's ability to detect weak signals. If the level of a signal is equal to the displayed average noise level, it will be displayed above the average noise level with an approximate 3dB protrusion. This signal is considered to be the minimum measurable signal level, but this phenomenon cannot always be seen if the video filter is not used to average the noise.

The sensitivity of a spectrum analyzer is defined as the average noise level displayed under a certain resolution bandwidth. "Average" means that the amplitude of the noise signal changes randomly with time and frequency. To quantitatively test the noise power, only its average value can be obtained. The sensitivity of a spectrum analyzer is an important indicator of the instrument. The sensitivity of a spectrum analyzer is related to its RBW; VBW; and attenuator settings.

The impact of the internal noise of the spectrum analyzer on the test can be reflected from different aspects.

1. When the input signal power level is less than the instrument noise level, the signal will not be displayed and the instrument has no testing capability for this small signal.

2. When the input signal amplitude is greater than the instrument noise, the instrument noise will be superimposed on the input signal, that is, the final displayed signal level is the power sum of the input signal level and the instrument noise.

When the power of the tested signal is more than 10~20dB greater than the internal noise power of the instrument, the influence of the internal noise of the spectrum analyzer can be ignored.

The reasons why spectrum analyzers generate noise and the impact of noise on instrument testing have been clarified above. The following is an analysis of the factors that affect the noise level in the following instrument settings.

Factors affecting the noise level of frequency spectrum analyzer 1: Input attenuation setting.

For every 10dB increase in attenuation of the attenuator, the noise level displayed by the spectrum analyzer increases by 10dB.

Analysis of the influence of attenuator setting value on the sensitivity of spectrum analyzer:

The input signal level does not decrease as the attenuation increases. This is because when the attenuation reduces the signal level applied to the detector, the intermediate frequency (IF) gain increases by 10 dB to compensate for this loss, resulting in the signal amplitude displayed by the meter remaining unchanged. However, the noise signal is only greatly affected by the amplifier, and its level is amplified by 10 dB.

Since the internal noise is mainly generated by the first stage of the IF amplifier, the input attenuator does not affect the internal noise level. However, the input attenuator affects the signal level of the mixer and reduces the signal-to-noise ratio.

Methods for Improving the Sensitivity of Spectrum Instruments

1: Use the smallest possible input attenuation to get the best sensitivity.

The noise generated inside the meter is broadband white noise. That is, its level is flat random noise over the entire frequency range, and its bandwidth is wide compared to the resolution bandwidth filter. Therefore, the resolution bandwidth filter only passes a small part of the noise energy to the envelope detector. If the resolution bandwidth is increased (or decreased) by a factor of 10, then 10 times more (or less) noise energy reaches the detector, and the displayed average noise level will increase (or decrease) by 10dB.

The relationship between the displayed noise level and the resolution bandwidth RBW is:

Noise level change (dB) = 10log (resolution bandwidth 2/resolution bandwidth 1)

The RBW is changed from 100kHz (resolution bandwidth (old)) to 10kHz (resolution bandwidth (new)), resulting in a noise level change of Noise level change = log (10 kHz / 100 kHz ) = 10dB.

The spectrum analyzer's intermediate frequency filter has a band suppression function for the broadband white noise generated by the intermediate frequency amplifier. Therefore, the smaller the RBW, the smaller the noise energy passing through the intermediate frequency filter, and the lower the noise level displayed after detection.

The noise of a spectrum analyzer is defined within a certain resolution bandwidth.

The lowest noise level (and slowest sweep time) of a spectrum analyzer is obtained at the smallest resolution bandwidth.

Method 2 to improve the sensitivity of spectrum meter:

Use the smallest possible RBW setting for best sensitivity.

A spectrum analyzer displays signal plus noise, so when the signal approaches the noise level, additional noise is superimposed on the scan lines, making it more difficult to read the signal.

The video filter is a low-pass filter after detection. Since the amplitude of the acoustic signal fluctuates randomly with time and frequency, it is output as an AC signal through detection processing. These AC signals are reflected on the display as the jitter of the track line. The low-pass processing of the video filter is used to average (smooth) the noise fluctuations. Although it cannot improve sensitivity, it can improve discrimination and repeatability of measurements under low signal-to-noise ratio conditions.

Reducing VBW will not affect the displayed CW signal spectrum, because the CW signal detection output is a DC signal, and when the DC signal is processed by low-pass filtering, it will not be affected by the filter bandwidth.

It should be noted that reducing VBW can smooth the noise signal, but it does not mean that the power average value of the noise signal is obtained.

Finally, let’s summarize the technical methods to improve the test sensitivity of the spectrum analyzer:

1. The narrowest resolution bandwidth;

2. Minimum input attenuation;

3. Make full use of video filters (video bandwidth < 0.1-0.01 resolution bandwidth)

4. Preamplifier (internal or external). The internal preamplifier requires an option, and the operating frequency range is generally <3GH. The preamplifier switch is controlled by [Amplitude] Int Amp: on/off.

The improvement of spectrum analyzer sensitivity by external amplifier = amplifier gain - amplifier noise factor.