Selecting The Best Resolution Bandwidth - Rbw - X - Series signal analyzers - Agilent

Like its name suggests, the resolution bandwidth, or RBW, plays an important role in resolving signals which are close together in the frequency domain. In this scenario, we will attempt to measure a two-tone input signal with a tonal spacing of 20 kHz. We would expect to see two strong peeks on our display -- here and here. However, we only see one broad peek which leads us to suspect that the RBW may be set too wide. Our RBW is set too wide for signals spaced 20 kHz apart. This causes the signals to blend together on the display. We can easily fix this problem, but first we will save this trace on the screen for comparison later. Now, we will turn on trace 2 and reduce the resolution bandwidth. As a general rule, we can resolve two tones of equal power if the RBW is equal to the tonal spacing. Since our signals are spaced by 20 kHz, we will check this rule by setting the RBW to 20 kHz also. Now, we have just resolved the two major peaks of our input signal. The two-tone signal should also produce a third-order distortion peak 20 kHz out from either input tone. The skirt around our principle tones is masking these low level close-in peaks, so we will need to reduce the RBW even more if we want to measure them. When we reduce the RBW, our principle peaks became more sharply defined. However, we still do not see the third-order distortion peaks. Let us try letting the instrument automatically select the RBW. Now we have revealed the third-order peaks. Typically, the automatic settings for RBW will provide enough resolution for your selected span. The RBW setting has a secondary effect on the displayed average noise level or DANL. Let us pull up the traces from the previous demo while focusing on the noise floor. As you can see, narrowing the resolution bandwidth also causes the noise level to drop. As a general rule, you can expect the DANL to drop by 10 dB when RBW is reduced by a factor of 10. Resolution bandwidth also has a major impact on sweep time. With an RBW of 51 kHz, our sweep time is set to 1 ms, which is the minimum sweep time for most spans. After stepping down to 20 kHz, the sweep time has not changed from the minimum. At 10 kHz, the sweep time increased to about 4 ms. When we step down to 2.7 kHz, there is a steep increase in the sweep time to nearly 50 ms. RBW and sweep time have an inverse square relationship. That is, for every factor of 10 reduction in RBW, there is generally a factor of 100 increase in sweep time. This relationship is one reason why choosing a very narrow RBW is not always ideal. So, what guidelines can we follow to help select the best RBW for a measurement? For closely-spaced or low-power signals, we generally want to reduce the RBW. To get a fast sweep time over a wide span, increase RBW or change from swept analysis to the FFT measurement mode. If you want to balance RBW and sweep time, set your span to cover your desired range and then reduce the RBW until all the signals of interest are resolved.