It is part of the performance of almost every signal source, even active components, such as amplifiers. It is phase noise, and, at sufficiently high levels, it can prevent a radar or communications system from operating effectively. Fortunately, it can be measured quickly and accurately constant improvements in test hardware and software are making the task of determining a high-frequency source’s phase noise easier even as signal bandwidths continue to reach into millimeter-wave frequency bands.
What is phase noise?
Phase noise consists of fluctuations in the phase of an oscillator, frequency synthesizer, or other signal source as measured in the frequency domain. It is also known as jitter when it is associated with timing deviations in a system clock. A sine wave without noise or change in frequency would go from one period to the next without variations in phase or amplitude. However, real-world signals exhibit some deviations in phase and amplitude which can occupy more of the spectrum than originally intended. Too much noise can be problematic, for example, in a crowded communications spectrum or an operating environment with many closely spaced channels or communication signals, where the noise can block or obscure some signals from reaching an intended receiver. The phase/frequency instability can also cause errors in communication systems that transmit and receive signals based on phase or frequency modulation.
How can it be measured?
Fortunately, the phase noise of a signal source can often be measured directly, with one of the most versatile of high-frequency instruments, the spectrum analyzer. The phase noise of a source is usually expressed as the amount of signal power within a 1-Hz bandwidth at some offset frequency from the carrier frequency (dBc/Hz). Because a real-world signal will deviate above and below a nominal tuned carrier frequency with some variations in amplitude and phase, it can fill more of the frequency spectrum than intended. On a spectral plot, phase noise appears as sidebands above and below the carrier and is measured and calculated as the single-sideband phase noise—the amount of noise contained within a 1-Hz bandwidth lying at a specific spectral distance from the carrier frequency.
Limiting the amount of measured noise to a 1-Hz bandwidth requires some filtering, or calculations, or a combination of the two. Any filter, such as a bandpass filter, will have a response shape that is not strictly rectangular and will have some effect on what is measured as the phase noise for that bandwidth, but the processing power of modern spectrum analyzers and other signal analyzers can usually compensate for the less-than-ideal shape of an instrument’s filtered response. Phase noise typically drops abruptly with distance from the carrier so that phase noise measured closer to the carrier, such as 10 kHz from the carrier, will be higher than phase noise further from the carrier, such as 1 MHz from the carrier.
Phase noise can be determined with different measurement approaches, such as residual and cross-correlation techniques. Residual measurements pass signals from a device under test (DUT) through a frequency mixer for analysis, using a reference oscillator with low phase noise (lower than the DUT) to aid in the frequency translation of the DUT’s signals. The phase noise of the single-channel test setup includes the phase noise of additional components required for the measurements, such as a phase detector and phase shifter for phase tuning. In contrast, dual-channel cross-correlation phase-noise measurements, on the other hand, correlates and averages the noise of two channels of a single DUT created by a power divider. Without the added noise of the frequency mixer and associated components, this approach can achieve lower noise floors than the residual measurement approach.
Finding the Right Fit
Not every spectrum or signal analyzer is meant for SSB phase-noise measurements and the best fit for the job must meet a minimum set of requirements. Direct measurements of SSB phase noise call for a signal or spectrum analyzer with frequency range exceeding the frequency of the DUT. A suitable analyzer should also have a wide enough dynamic range (the minimum to maximum input signal level) to handle signals from a DUT, ideally without additional external attenuation or gain. The analyzer should have a noise floor that is considerably lower (10 dB or more) than the phase noise of the DUT so that measurement accuracy is not degraded by instrument noise. A limited dynamic range can be overcome, but it requires adding components to the test setup, such as attenuators and amplifiers; they can add their own noise contributions to the phase noise test system.
Because of the low noise levels being recorded in many SSB phase-noise measurements, any noise sources near the test site should be a concern. Test ports on the analyzer should be shielded to minimize the effects of any noisy nearby components. Any electromagnetic (EM) noise within the environment, such as from broadcast stations and cellular antennas, often indicates a need for phase-noise testers to perform their measurements within a fully EM-shielded room.
An earlier blog Measuring Phase Noise and Why It Matters, from 2016, reviewed phase-noise testing and some of the instrument options for making those measurements at that time. But highly integrated instrument solutions and highly effective software are making present-day phase-noise measurements much easier to make. For lower-frequency carriers or even baseband testing of frequency-downconverted carriers, few test instruments are easier to use than the Microsemi 53100A Phase Noise Analyzer Although limited in frequency range (1 to 200 MHz) and dynamic range (-5 to +15 dBm), this is a low-cost solution that can make Allan deviation measurements at low noise levels. With a pair of input ports, it can even perform cross-correlation phase-noise measurements at offsets of 0.001 Hz to 1 MHz from the carrier. The straightforward instrument achieves a noise floor that is essentially at the thermal noise floor of a test instrument (kTB), or -174 dBm/Hz. With external frequency mixing capable of downconverting a DUT’s signals within the 200-MHz baseband range of the model 53100A, it can function as a quite effective yet affordable phase-noise measurement system.
Another easy-to-use phase-noise tester, the Holzworth HA7062D real-time phase-noise analyzer features a greatly expanded frequency range. Standard units test signals from 10 MHz to 26 GHz; an option extends the upper-frequency limit into the millimeter-wave region, to 40 GHz. It handles input test signal levels from -5 to +20 dBm with a single 2.92-mm female input connector and is supplied in a fully shielded 1U-high instrument chassis. The compact analyzer performs cross-correlated phase-noise measurements by means of an internal low-noise 10-MHz reference oscillator. It can make phase-noise measurements at offsets from 0.1 Hz to 100 MHz per ANSI z540.1 standards and boasts noise floors of better than -100 dBc/Hz offset 10 Hz from the carrier, -149 dBc/Hz offset 10 kHz from the carrier, and better than -155 dBc/Hz offset 40 MHz from the carrier.
The Rohde & Schwarz FSWP26 Phase Noise Analyzer and VCO Tester is housed within a traditional laboratory benchtop spectrum-analyzer configuration and is equipped to perform both residual and cross-correlation phase-noise measurements close to the thermal noise floor for carriers from 1 MHz to 26.5 GHz. It exhibits the normal trend of increasing phase noise with increasing carrier frequency, achieving typical SSB phase noise of -172 dBc/Hz within a 1-Hz bandwidth offset 10 kHz from a 1-GHz carrier and -158 dBc/Hz within a 1-Hz bandwidth offset 10 kHz from a 10-GHz carrier. For those seeking even higher carrier frequencies, the R&S FSWP50 extends the frequency range to 50 GHz.
Different measurement options from Keysight Technologies also provide the capabilities for measuring the SSB phase noise of millimeter-wave signal sources. For example, the model E5052A signal analyzer offers a basic frequency range of 3 Hz to 3.6 GHz with options that extend the measurement bandwidth to 8.4, 13.6, and 26.5 GHz and as high as 50 GHz. An optional internal low-noise preamplifier increases sensitivity and extends the dynamic range and is available for all the optional frequency ranges.
For those requiring laboratory-grade performance levels, the Keysight N9030A PXA signal analyzer also has versions starting at 3 Hz through 3.6 GHz and up through 26.5 GHz and can be equipped with an optional preamplifier for each of the frequency ranges. It provides phase-noise measurements at the thermal noise floor with a typical displayed average noise level (DANL) on the front-panel screen of -172 dBm. It is an extremely accurate group of the company’s X-Series signal analyzers that can be operated manually or automatically, under the control of the Windows XP operating system (OS) from Microsoft.
While the emphasis in this blog has been on the spectrum analyzer as the tool of choice for phase-noise measurements, other instruments can be used, including network analyzers such as the two- and four-port model N5242A network analyzers from Keysight Technologies.
In any case, those in need of phase-noise test capabilities should evaluate their basic requirements of frequency range, offset range, and noise floor before starting the search. As these examples show, many test options are available, becoming even “smarter” with embedded software for automated report generation and for coordinating the phase-noise measurements according to an operator’s required parameters and routines.
More equipment information, like data sheets, for these and other units are always available at www.axiomtest.com by using our searchbar to find the unit you need. If you need help finding the right equipment for your project, you can contact us at either firstname.lastname@example.org, or 760-806-6600.