Spectrum analysis method for measuring phase noise

Phase noise and spectrum analyzer resolution bandwidth (RBW)

Phase noise measurements with a spectrum analyzer require (1) normalization and (2) shape correction.

Let’s start with normalization. Phase noise is specified as the noise power contained within the bandwidth of 1 Hz. Spectrum analyzers measure power using a resolution bandwidth (RBW) filter, and this filter is usually more than 1 Hz wide. Therefore, noise power has to be “normalized” to a 1 Hz bandwidth. This is done by reducing the measured noise power value by ‘N’ dB, where ‘N’ is the ’10 log of the RBW in Hz.’ For example, let’s say that you use a 3 kHz RBW filter to measure a noise power of -90 dBm. Applying the normalization formula, you would get N = 34.77 dB, so the normalized 1 Hz noise power is 124.77 dBm.

• Learn more about phase noise measurements

Shape correction

In the previous example, the RBW was a simple rectangle, which made the math very straightforward. However, real-world RBW filters are not perfectly rectangular. The shape is usually Gaussian or something similar. So, in addition to normalizing for bandwidth, we also have to correct for the filter shape.

For a given RBW, a Gaussian filter has a wider noise bandwidth than its nominal or 3 dB bandwidth. This means that you must multiply the filter bandwidth by a scaling factor before normalization. The correction factor is implementation dependent. In other words, it depends on how the specific filter was physically implemented – not all Gaussian RBW filters have the same shape.

For example, the shape correction for this particular 3 kHz filter is 1.165, so when calculating N, you can multiply the nominal filter width by 1.165 before taking the logarithm.

Note that most spectrum analyzers can automatically apply both bandwidth and shape corrections with the special noise marker function.

Noise marker and manual phase noise measurement

This noise marker can be used for manual phase noise measurements. To do this, you would simply place the marker at the offset of interest and read off the normalized and shape-corrected value. However, like most other manual processes, manually measuring phase noise is both time consuming and error prone.

Many spectrum analyzers have a phase noise personality or option that automates the process, with the results presented in both graphic and numeric formats. Spectrum analyzers are general-purpose instruments, so that biggest advantage of using a spectrum analyzer for measuring phase noise is that it provides additional useful functions for characterizing sources, such as measurements of spurious emissions, settling time measurements and many other.

In many cases, a traditional spectrum analyzer is sufficient for making phase noise measurements. However, it’s important to be aware of potential challenges and limitations:

• Dynamic range
• Instrument phase noise
• Close-in noise or drifting sources
• Amplitude (AM) noise

Dynamic range

Dynamic range is the difference between the largest and smallest signals that can be accurately measured.

With the spectrum analyzer method, you calculate the phase noise by measuring both the power of the carrier as well as noise powers at different offsets from the carrier. The difference between the measured carrier power and measured noise power is usually quite large, typically ranging from 80 to over 140 dB. For accurate phase noise measurements, you need to measure both very high and very low powers simultaneously. Therefore, dynamic range is an important specification to consider when choosing a spectrum analyzer for phase noise measurements.

• Learn more about dynamic range

Instrument phase noise

Another important specification is the internal phase noise of the analyzer. Spectrum analyzers usually contain multiple local oscillators (LO). Like all oscillators, each LO has a certain amount of phase noise, and this phase noise is added to the phase noise of the DUT signal as it moves through the analyzer.

One of the limitations of the spectrum analyzer method is that it is difficult to separate or distinguish the phase noise present in the original signal from the phase noise added by the instrument. The easiest way to avoid this issue is to use an analyzer that has a better noise specification than the DUT. At least 10 dB better is a good place to start, but the larger the margin, the more accurate the phase noise results.

Close-in noise or drifting sources

Close-in phase noise is phase noise measured at very small offsets from the carrier. Measuring close-in noise can be very difficult for two reasons:

• A very narrow RBW needs to be chosen to avoid measuring the carrier power as well as the noise power. The Gaussian shape of RBW filters further complicates this issue.
• The phase noise of a carrier can drift in frequency, although some analyzers do have the ability to track a small amount of drift and automatically compensate for it.

Modern spectrum analyzers can avoid some of these issues by measuring phase noise using IQ data, which is the digital representation of the spectrum, obtained by means of fast Fourier transform. Measuring with IQ data can improve both the stability and accuracy of phase noise measurements, particularly for close-in or drifting sources.

Amplitude (AM) noise

The IQ mode is also useful for dealing with AM noise.

When measuring phase noise, we assume that the noise sidebands around the carrier are mostly due to phase noise but with some amount of AM noise mixed in. In general, this is a valid assumption: the AM noise in real-world devices is usually much less than the phase noise. However, when a large amount of AM noise is present, the standard spectrum analyzer method may not produce accurate results. That’s because this method cannot distinguish between AM noise and phase noise.

A traditional spectrum analyzer can filter out some unwanted AM noise when it uses IQ data for measurements.

It's important to note that AM noise tends to have a bigger effect at higher frequency offsets from the carrier. Therefore, the advantages of using this method become clearer the further away from the main frequency we look.