Why and How to Use Spectrum Analyzers

Resource Description

Introduction

A Spectrum Analyzer lets you see signal problems. If you can't see it, you can't fix it.

Why use a spectrum analyzer?

Spectrum analyzers display RF signals from base stations and other emission sources. They find rouge signals, measure carriers and distortion, and verify base stations. Unlike a power meter, they validate carrier frequency and identify desired and undesired signals.

Spectrum Analyzer Setup

RF signals have a frequency, bandwidth, and power. To best view an RF signal, three things need to be set:

1. Adjust the center frequency to center the desired signal. Entering a known carrier frequency is a common way to do this. Other common methods use the knob or arrow buttons, a channel number, or the Marker-to-Center function.
2. Adjust the span so that the desired signal covers about half of the screen.
3. Adjust the reference level to bring the top of the signal near the top of the screen.

It is a flexible process. Often, it is best to start with a close center frequency, an approximate reference level, and narrow the span (span down) in stages, re-centering the signal as you go. Otherwise, the signal may move off the side of the screen as you span down.

A Users’ Spectrum Analyzer Block Diagram Description

Spectrum analyzer users need to understand major analyzer functional blocks. While most of the choices described here are made automatically, manual selections are possible and can be helpful.

The DC block prevents DC voltage from entering the instrument and allows measuring lines with DC power.

The step attenuator helps to prevent overdriving the first mixer; the excessive level at the first mixer causes distortion visible on the display. Spectrum analyzers normally enable the step attenuator when the reference level is over -20 dBm or so. The step attenuator raises the noise floor when enabled but allows measuring higher signal levels. You can think of this as matching the dynamic range of the analyzer to the input signal level.

A preamplifier is often the first block in a spectrum analyzer. Use it for weak signals lower than about -50 dBm. When on, it reduces the noise floor by about 10-15 dB. Turn it off if the RF signals are higher than -40 dBm or so. As shown in the picture to the right, the preamplifier reduces the noise level, allowing lower-level signals to be seen in the red trace, where they couldn't be seen without the preamp in the white trace.

The first mixer is the first down-conversion stage, the stage that changes the RF frequency to an intermediate frequency (IF). The step attenuator and preamplifier allow adjusting the input signal to match what the first mixer needs to see.

IF section contains further mixers and creates more IF frequencies.

Resolution Bandwidth (RBW) sets the level of signal detail displayed. RBW also affects the noise floor. Lower RBW settings show more detail and create a lower noise floor but cause slower sweeps. Higher RBW settings are good for wider spans since they sweep faster, at the cost of less detail and a higher noise floor. The picture to the right illustrates the two traces showing the effect of two different resolution bandwidths (30 kHz in red and 300 kHz in white).

The Envelope detector converts the signal into a video signal that can be displayed. The video filter (VBW) averages the envelope, giving more stable answers for noisy or noise-like signals (such as digital modulation). It is also useful when looking for small consistent signals near the noise floor, such as spurious CW signals, as well as when measuring noise or Signal-to-Noise Ratio (SNR).

The detector samples the IF signal and converts it to digital samples. The input to the detector normally gets many more samples than can be displayed on the screen. The detector groups samples, creating one group for each display point. It then chooses one number related to each group for display.

• Peak displays the largest sample of each group. This is the default detector and is useful when looking at CW signals, as well as an intermittent or bursty signal.
• RMS calculates the average power of the grouped samples and displays that. This is useful when measuring noise or noise-like signals (many cellular signals are noise-like).
• Negative displays the smallest sample of each group. This is useful when looking for dropouts of a signal, mostly in a zero span.
• Sample picks just one of the input samples
• Quasi Peak is useful for EMC (Electromagnetic Compatibility) measurements. This is a special detector type that emulates the response of the human ear to impulses.
• Sweep speed is normally set automatically. It can be helpful to set lower sweep speeds manually when working with intermittent signals. In this case, the Max-Hold trace setting is also helpful.
• Units converts the trace into other units, such as Watts or Volts.
• Trace Modes include normal, average, max-hold, and min-hold. Average reduces the variation of noise and noise-like signals. Max-hold and min-hold creates a record of the signals’ highest and lowest excursions, respectively, and is useful for finding transients and measuring the power of wideband burst signals - ones that can’t be measured in a zero span due to the bandwidth.
• Limits and Measurements examine the trace to see if a limit line has been exceeded, and create numerical answers such as channel power.
• Sweep Generator tunes the Local Oscillator (LO) over frequency (Center – Span/2 to Center + Span/2) and provides a reference for the x-axis of the display. The rate of tuning is controlled by the sweep time control. Normally this is left to run as fast as possible for spectrum displays.

Spectrum Analyzer Setup (Part 2)

The first mixer level is critical for proper performance. If an Over-the-Air (OTA) signal is overdriving the first mixer, false signals, or spurs, may appear. The noise floor may also rise and hide signals of interest. This can happen even if the OTA signal is outside of the current span.

Spectrum analyzers display an “overload” message if an overload occurs. The first response is often to raise the reference level, either directly or by manually increasing the attenuation of the step attenuator.

Coupled mode is the default operating mode for spectrum analyzers. In this mode, attenuation, RBW, and VBW values adjust automatically, set by the user’s choice of reference level and span. In coupled mode, both span and reference levels affect the noise floor.

Manual mode allows attenuation, RBW, and VBW. Manual mode allows making finer tradeoffs in noise level, distortion, trace variation, and sweep speed.

Manual attenuation capability is helpful when checking for distortion. A quick check to see if a signal is spurious is to change the step attenuator 10 dB. If the signal doesn’t change, all is well; it is a real signal. If it changes by 10 dB or more, it is an internally generated intermodulation product. Sometimes the distortion can simply be ignored; in other cases increase the attenuation until the distortion level doesn’t change.

Manual RBW capability is helpful when a lower noise floor, even at the expense of a slower sweep speed, is desirable, or faster sweep speed is needed, at the expense of a higher noise floor and poorer frequency resolution. Every 10 times change in RBW means a 10 dB change in the noise floor.

Some spectrum analyzers will have a lower noise floor, at any given RBW, than others. The noise figure specification determines this, or you can estimate noise figure from NF ≈ (DANL – 10*log(RBW for DANL))-174. where DANL is the Displayed Average Noise Level specification, and RBW for DANL is the RBW used for the DANL spec - often this is 1 Hz or the spec is normalized to 1 Hz, but this is not always the case.

Low-level signals are easier to see if attenuation is minimized and the preamplifier is on. Selecting lower RBW directly reduces the noise floor, and makes CW signals easier to see; unfortunately, many cellular communication signals are also noise-like and do not benefit from reduced RBW. Selecting lower VBW smoothes both the noise and noise-like signals, and can make it easier to see signals near the noise. Trace Averaging can also be used to further smooth the noise.

Intermittent signals can be viewed or monitored by max hold traces, gated sweep, Save on Event with Limit Lines, and spectrograms.

Using Markers

Markers make it easy to verify amplitude & frequency points on the spectrum; the marker table lets you see up to 6 markers and delta values at a glance.

Instrument setup is faster when using Markers. The marker functions “Peak Search” and “Marker Freq to Center” can be very helpful, as they will find and center the strongest signal.

Unique marker measurement capabilities include fixed and tracking delta measurements (“Delta” or “∆” is shorthand for difference). Fixed markers are useful for delta measurement when the signal is intermittent. In this case, the reference marker can be fixed, or frozen, at a specific amplitude, even if the signal goes away. A tracking marker follows changes in trace amplitude. Marker delta measurements can either use marker 1 as the reference, allowing six delta measurements, or use each of the six markers to have their own delta marker. This will allow up to 6 delta measurements and 12 total markers on one screen. This is a good way to measure and document signals.