Phase Noise & Jitter Specifications
- details of the phase noise or phase noise specifications used to define the levels of phase noise or jitter on an RF signal.
As phase noise is a particularly important parameter for many RF applications, phase noise measurements and phase noise measurement techniques are of great interest to many RF design engineers.
Phase noise measurement can be quite difficult under some circumstances. Fortunately many new spectrum analysers make phase noise measurement very much easier.
However there are still a number of precautions to be observed, and when measurements of very low levels of phase noise are to be made, sometimes, specialised techniques are required.
Phase noise specification
It is often necessary to specify the levels of phase noise. In this way the performance of an oscillator or signal source can be quantified and compared against the required specification. The phase noise of phase jitter may be specified in a number of ways, but the one that is most commonly used is to look at the noise level, i.e. the single sideband phase noise at a given point.
Single sideband phase noise measurement
When looking at the noise level there are three elements to the specification:
- The phase noise level itself: this is expressed in dB relative to the carrier, i.e. dBc. This method is adopted because the phase noise normally varies in line with the carrier level. The level of phase noise relative to the carrier is also the important factor. Where the phase noise varies with the carrier level, the specification can state that the phase noise is - n dBc at a given carrier level.
- The offset from the carrier: An essential part of the phase noise specification is the offset from the carrier at which the phase noise was a certain level. This is because the noise level varies according to the frequency offset from the carrier, the frequency offset must be given. Typically the phase noise rises much faster closer in towards the carrier and falls away until it ultimately reaches a noise floor.
- The measurement bandwidth: The noise power is proportional to the bandwidth and therefore it is necessary to state the bandwidth that has been used. Obviously the wider bandwidth that is used, the greater the level of noise that will pass through the filter and be measured. Technically the most convenient bandwidth to use is 1 Hz because it is easy to relate the level to other bandwidths. As a result this phase noise specification format has been almost universally adopted. Spectrum analyzers are unable to measure in a 1 Hz bandwidth directly because this would require a very narrow filter bandwidth - therefore they measure the signal in a wider bandwidth and mathematically adjust the level to that of a 1 Hz bandwidth. Using current signal processing technology this is simply a further calculation that is added into the routine.
Thus a typical phase noise specification for a signal generator or other oscillator may be -100 dBc / Hz at a 100 kHz offset. For a complete phase noise specification several points will be specified to give an indication of the phase noise at different points, typically at points varying by a factor of ten: 10 Hz, 100 Hz, 1 kHz, etc.
Typical SSB phase noise profile
For very approximate calculations the various points at which the phase noise has been specified can be joined to give an indication of the performance of the signal source.
Phase jitter specification
Apart from the more usual method of quoting phase noise as the level of the single sideband phase noise at a given offset, it is also possible to look at the phase jitter specification.
It is possible to quote a phase jitter specification in two formats:
- Phase jitter: When actual phase jitter is required, this is measured in radians (or less commonly degrees). Normally it is quoted in RMS radians. This format for phase jitter is more widely used in applications where the angular or phase element is important - typically for RF modulation, etc.
- Time jitter: This format is often used for clocks were the phase jitter in terms of a time difference is required. It is may be specified in a number of ways according to what is of interest to the designer - it can be RMS, peak or peak to peak time difference, often measured in picoseconds. This is used where the time element of the phase jitter is important.
Both formats for phase jitter are seen in various specifications - the type used depends upon the application envisiaged.
By Ian Poole
Share this page
Want more like this? Register for our newsletter