The influence of the intermediate frequency filter of the spectrum analyzer on the frequency resolution

Frequency resolution refers to the ability of a spectrum analyzer to clearly separate two input sine waves.

The 3dB bandwidth or 6dB bandwidth of the intermediate frequency filter is usually expressed by the resolution bandwidth RBW (Resolution Bandwidth).

When we use a resolution bandwidth of 10kHz for two signals separated by 5kHz, their traces overlap at the top and look like a response; when we use a resolution bandwidth of 3kHz, they can be clearly separated from each other.

The impact of different resolution bandwidths on the resolution of equal-amplitude signals In our actual measurements, we often encounter signals with unequal amplitudes but close together, such as some distortion products.

In this case, the signal with a smaller amplitude is often submerged by the skirt of the signal with a larger amplitude. So, how can we ensure that small signals are not drowned by large signals? First, the intermediate frequency filter must be made narrow enough. Now some analyzers have intermediate frequency filters with a bandwidth of 1Hz. Secondly, it must have good selectivity, that is, the intermediate frequency filter should be as steep as possible. The smaller the shape factor (the ratio of the intermediate frequency filter's 60dB bandwidth to the 3dB bandwidth), the better its selectivity. For analog filters, it can reach 11:1, generally (12-15):1, and for digital filters, it can reach 5:1.

Let's take an example to discuss the influence of shape factor and resolution bandwidth on the resolution of unequal amplitude signals.

To illustrate this problem, we quote the following formula to calculate how much the edge of the filter drops under a given frequency offset and RBW: -3dB-[(△F-BW3/2)/(BW60/2-BW3/2)]×57dB In the formula, △F is the frequency difference between the two signals; BW3 is the 3dB bandwidth of the intermediate frequency filter; BW60 is the 60dB bandwidth of the intermediate frequency filter.

After the value of the filter edge drop is determined by the above formula, it means that signals less than this value can be distinguished, while signals greater than this value cannot be distinguished.

For example: when we choose a shape factor of 12:1 and a frequency deviation of 5kHz, calculate the minimum signal that can be distinguished under RBW of 3kHz and 1kHz.

When RBW=3kHz: -3-[(5-3/2)/(36/2-3/2)]×57=-15(DB) That is, when we use an intermediate frequency filter with a bandwidth of 3kHz and a shape factor of 12:1, the maximum two signals that can be distinguished are 5kHz deviation and 15dB amplitude difference.
Similarly, it can be calculated that when we use an intermediate frequency filter with a bandwidth of 1kHz and a shape factor of 12:1, the maximum two signals that can be distinguished are 5kHz deviation and 49dB amplitude difference.

The above calculations seem a little complicated, but they fully illustrate the fact that under the same shape factor, the smaller the RBW, the stronger the ability to resolve small signals; under the same RBW, the smaller the shape factor, the stronger the ability to resolve small signals.

In actual use, we do not need to perform such calculations, but it is helpful for us to understand the characteristics of the intermediate frequency filter.

When we use a spectrum analyzer to analyze and observe the signal, there are two main factors that affect the frequency resolution: one is the shape factor of the intermediate frequency filter used by the analyzer; the other is the resolution bandwidth we choose.

The shape factor of an analyzer cannot be changed, while the resolution bandwidth is generally adjustable, so we can achieve our goal by adjusting the resolution bandwidth during use.