Skip navigation
All Places > Keysight Blogs > Better Measurements: The RF Test Blog > Blog > 2017 > March
2017

  Good news about RBW, noise floor, measurement speed, and finding hidden signals

Some things never change, and others evolve while you’re busy with something else. In recent years, RF measurements provide good examples of both, especially in signal analysis fundamentals such as resolution bandwidth filtering, measurement noise floor, and speed or throughput. It can be a challenge to keep up, so I hope this blog and resources such as the one I’ll mention below will help.

One eternal truth: tradeoffs define your performance envelope. Optimizing those tradeoffs is an important part of your engineering contribution. Fortunately, on the measurement side, that envelope is getting substantially larger, due to the combined effect of improved hardware performance plus signal processing that is getting both faster and more sophisticated.

Perhaps the most common tradeoffs made in signal analysis—often done instinctively and automatically by RF engineers—are the settings for resolution bandwidth and attenuation, affecting noise floor and dynamic range due to analyzer-generated distortion. Looking at the figure below, the endpoints of the lines vary according to analyzer hardware performance, but the slopes and the implications for optimizing measurements are timeless.

Graphic shows how signal analyzer noise floor varies with RBW, along with how second and third order distortion vary with mixer level. Intersection of these lines shows mixer level and resolution bandwidth settings that optimize dynamic range

Resolution-bandwidth settings determine noise floor, while attenuation settings determine mixer level and resulting analyzer-produced distortion. Noise floor and distortion establish the basis for the analyzer’s dynamic range.

This diagram has been around for decades, helping engineers understand how to optimize attenuation and resolution bandwidth. For decades, too, RF engineers have come up against a principal limit of the performance envelope, where the obvious benefit of reducing resolution bandwidth collides with the slower sweep speeds resulting from those smaller bandwidths.

That resolution bandwidth limit has been pushed back substantially, with dramatic improvements in ADCs, DACs, and DSP. Digital RBW filters can be hundreds of times faster than analog ones, opening up the use of much narrower resolution bandwidths than had been practical, and giving RF engineers new choices in optimization. As with preamps, the improvements in noise floor or signal-to-noise ratio can be exchanged for benefits ranging from faster throughput, to better margins, to the ability to use less-expensive test equipment.

Improvements in DSP and signal converters have also enabled new types of analysis such as digital demodulation, signal capture and playback, and real-time spectrum analysis. These capabilities are essential to the design, optimization and troubleshooting of new wireless and radar systems.

If you’d like to know more, and take advantage of some of these envelope-expanding capabilities, check out the new application note Signal Analysis Measurement Fundamentals. It provides a deeper dive into techniques and resources, and you’ll find it in the growing collection at our signal analysis fundamentals page.

A few months ago, Keysight’s Brad Frieden and I both wrote about downconversion and sampling, related to wireless and other RF/microwave signals. Brad's article in Microwaves & RF appeared about two weeks before my blog post, though I somehow missed it.

His main focus was on oscilloscopes and improving signal-to-noise ratio (SNR) in measurements of pulsed RF signals. He described the use of digital downconversion, resampling, and filtering to trade excess bandwidth for improved noise floor.

Inside Keysight, debates about scopes versus signal analyzers can become quite animated. One reason: we have slightly different biases to how we look at signals. Engineers in some areas reach first for oscilloscopes, while others have always leaned on spectrum and signal analyzers. It’s more than a general preference for time or frequency domain analysis, but that’s a start.

In test, those distinctions are fading among manufacturers and end users. On the supply side, oscilloscopes are extending frequency coverage into the microwave and millimeter ranges, and signal analyzers are expanding bandwidth to handle wider signals in aerospace/defense and wireless. Happily, both platforms can use the same advanced vector signal analyzer software that provides comprehensive time-, frequency-, and modulation-domain measurements.

On the demand side, frequencies and bandwidths are expanding rapidly in wireless and aerospace/defense applications. That’s why both types of instruments have roles to play.

But if they run the same software and can make many of the same measurements, how do you choose? I’ll give some guidelines here, so that your requirements and priorities guide your choice.

Bandwidth: In the last 15 years, this change has pulled oscilloscopes into RF measurements because they can handle the newest and widest signals. In some cases they’re used in combination with signal analyzers, digitizing the analyzer’s IF output at a bandwidth wider than its own sampler. That’s still the case sometimes, even as analyzer bandwidths have reached 1 GHz, and external sampling can extend the bandwidth 5 GHz! It’s telling that, in his article on oscilloscopes, Brad speaks of 500 MHz as a reduced bandwidth.

Accuracy, noise floor, dynamic range: In a signal analyzer, the downconvert-and-digitize architecture is optimized for signal fidelity, at some cost in digitizing bandwidth. That often makes them the only choice for distortion and spectrum emissions measurements such as harmonics, spurious, intermodulation, and adjacent-channel power. Inside the analyzer, the processing chain is characterized and calibrated to maximize measurement accuracy and frequency stability, especially for power and phase noise measurements.

Sensitivity: With their available internal and external preamps, narrow bandwidths, noise subtraction and powerful averaging, signal analyzers have the edge in finding and measuring tiny signals. Although Brad explained processing gain and some impressive improvements in noise floor for narrowband measurements with oscilloscopes, he also noted that these gains did not enhance distortion or spurious performance.

Multiple channels: Spectrum and signal analyzers have traditionally been single-channel instruments, while oscilloscope architectures often support two to four analog channels. Applications such as phased arrays and MIMO may require multiple coherent channels for some measurements, including digital demodulation. If the performance benefits of signal analyzers are needed, an alternative is a PXI-based modular signal analyzer.

Measurement speed: To perform the downconversion, filtering and resampling needed for RF measurements, oscilloscopes acquire an enormous number of samples and then perform massive amounts of data reduction and processing. This can be an issue when throughput is important.

With expanding frequency ranges and support from sophisticated VSA software, the overlap between analyzers and oscilloscopes is increasing constantly. Given the demands of new and enhanced applications, this choice is good news for RF engineers—frequently letting you stick with whichever operating paradigm you prefer.