Performing Better Spectrum Analyzer Measurements

Contributed by Hewlett-Packard (HP)

Spectrum analyzers are essential instruments in the test and measurement of any wireless system design. Spectrum analyzers are vital tools when evaluating the performance of a systems design. As a result, these instruments are a standard fixture on the bench top of many wireless design engineers.

The spectrum analyzer, like an oscilloscope, is a basic tool used for observing signals. Where the oscilloscope provides a window into the time domain, the spectrum analyzer offers a link into the frequency domain.

How It Works

In a swept-tuned spectrum analyzer, a signal at the input first travels through the attenuator and the low-pass input filter. The attenuator then limits the amplitude of the signal while the filter eliminates undesirable frequencies.

Past the input filter, the signal gets mixed with another signal generated by a voltage-controlled oscillator (VCO). The frequency of the VCO is controlled by a repeating ramp generator whose voltage also drives the horizontal axis of the display. As the frequency of the VCO changes, the mixed input signal sweeps through the resolution bandwidth filter (intermediate frequency (IF) filter), which is fixed in frequency. A detector then measures the power level of the signal passing through the IF filter, producing a DC voltage that drives the vertical portion of the display.

As the VCO sweeps through its frequency range, trace is drawn across the screen. This trace shows the spectral content of the input signal within a selected range of frequencies.

There are three essential steps in any spectrum analyzer measurement. First, engineers must prepare the input signal for measurement. Then they must prepare the spectrum analyzer for measurement. Finally, engineers must interpret and interrogate the results. Below are eight tips to help engineers execute these three steps and use these instruments more effectively.

1. Improving Amplitude Measurements
When making accurate amplitude measurements with a spectrum analyzer, it is crucial to cancel out any effects that degrade or alter the signal of interest in-between the device under test (DUT) and the analyzer. One method of accomplishing this is to use the analyzer's built-in amplitude correction (Ampcor) function in conjunction with a signal source and a power meter. Ampcor takes a list of frequency and amplitude pairs, linearly connects the points to make a correction waveform, and then offsets the input signal according to these corrections.

An example of this can be seen when the frequency response of a signal delivery system attenuates the DUT's signal and injects a noise spike. To cancel out the unwanted noise, engineers must first measure (with the source and power meter) the attenuation/gain of the signal delivery network at the troublesome frequency points in the measurement. For example, at 600 MHz, the engineer might send a zero dBm signal through the network to the power meter. In this case, the reading on the power meter indicates the attenuation or gain associated with that frequency point. Performing this same action at different points throughout the frequency range will yield the table of frequency and amplitude points that will be fed into the Ampcor table. Once the Ampcor filter is turned on, the unwanted attenuation and gain of the signal delivery network have been eliminated from the measurement.

2. Stimulus-Response Measurements
When combined with a tracking generator, the spectrum analyzer forms a stimulus-response measurement system. With the tracking generator as the sweep source and the spectrum analyzer as the receiver, this system operates similarly to a network analyzer.

Stimulus-response measurement systems are a good tool for analyzing a filter's return loss. To measure the return loss of a filter, an engineer needs a spectrum analyzer equipped with a tracking generator, a bridge (directional coupler) and a short circuit. When measuring return loss, the engineer must first set the analyzer to the desired frequency span (when setting the analyzer, the filter DUT must be in place). Once set, the DUT input must be connected to the output of the bridge. In addition, the unconnected port of the DUT must be terminated with a matched load. The next step is to connect the tracking generator output to the input port of the bridge and the spectrum analyzer input to the coupled port of the bridge.

Once the measurement is set up, the engineer needs to turn on the tracking generator output and adjust its amplitude. Now the DUT should be replaced with a short circuit that reflects all incident power and has a reflection co-efficient equal to one.

Next, engineers must normalize the display with the spectrum analyzer's normalize function. The normalize function eliminates the frequency response error of the test system, providing a convenient 0 dB return loss reference at the top of the display.

Finally, designers reconnect the filter in place of the short circuit without changing any of the settings on the spectrum analyzer. The marker can then be used to read the return loss at any frequency point.

3. Measuring Low-Level Signals
A spectrum analyzer's ability to measure low-level signals is limited by the noise generated inside the instrument. This sensitivity to low-level signals is affected by the measurement setup.

The spectrum analyzer's input attenuator and resolution bandwidth settings are the key factors that determine how small of a signal the spectrum analyzer can measure. The input attenuator, when activated, reduced the level of the signal at the input of the mixer. As amplifier at the mixer's output then re-amplifies the attenuated signal to keep the signal peak at the same point on the analyzer's display. In addition to amplifying the input signal, the noise present in the analyzer is amplified as well. This raises the displayed noise level of the analyzer.

The spectrum analyzer's resolution bandwidth filter affects how closely a small signal can be seen in the presence of a large one. By increasing the width of this filter, more noise energy is allowed to hit the envelope detector of the analyzer. This also raises the displayed noise level of the analyzer.

For maximum sensitivity, both the input attenuator and resolution bandwidth settings must be minimized. If a signal is still near the noise after adjusting the attenuation and resolution bandwidth, video averaging or video filtering the display can improve the stability of the displayed trace.

4. Identifying Internal Distortion Products
High-level input signals may cause internal spectrum analyzer distortion products that could mask the real distortion measured on the input signal. Using dual traces and the analyzer's RF attenuator, an engineer can determine which signals are internally generated distortion products.

To identify these products, engineers should tune the analyzer to the second harmonic of the input signal and set the input attenuator to 0 dBm. Next, they should save the screen data in trace B, select trace A as the active trace and activate marker D. The spectrum analyzer now shows the stored data in trace B and the measured data in trace A, while marker D shows the amplitude and frequency difference between the two traces. Finally, the engineer should increase the RF attenuation by 10 dB and compare the response in trace A to the response in trace B. If the responses in trace A and trace B differ, then the analyzer's mixer is generating internal distortion products due to the high level of the input signal.

5. Selecting The Best Display Detection Mode
Modern spectrum analyzers use digital technology for data acquisition and manipulation. In these analyzers, the analog signal at the input of the analyzer is segmented into bins that are digitally sampled for further data processing and display.

The question that naturally arises for engineers is: What point in the bin do we use for our data point? Spectrum analyzers generally have two to three detector modes that dynamically affect how the input signal is interrupted and displayed. These modes include peak detection, sample detection, negative peak detection and rosenfell.

Peak detection mode detects the highest power level in each bin. Peak detection is good for analyzing sinusoids, but tends to over-respond to noise when sinusoids are not present.

Sample detection mode displays the last point in each bin regardless of power. Sample detection is good for noise measurements. This mode accurately indicates the true randomness of noise. However, sample detection is inaccurate for measuring continuous wave (CW) signals with narrow resolution bandwidths and will miss signals that do not fall on the same point in each bin.

Negative peak detection mode displays the lowest power level in each bin. This mode is good for amplitude modulation (AM)/frequency modulation (FM) demodulation and distinguishing between random and impulse noise. Negative peak detection does not give the analyzer better sensitivity, although the noise floor may appear to drop.

Higher performance spectrum analyzers also have a detection mode, called rosenfell. This sampling mode dynamically classifies the data point as either noise or a signal, providing a better visual display of random noise than peak detection while avoiding the missed-signal problem of sample detection.

6. Measuring Burst Signals

Analyzing burst signals (pulses) with a spectrum analyzer is very challenging. In addition to displaying the information carried by the pulse, the analyzer also displays the frequency content of the shape of the pulse. The sharp rise and fall times of the pulse envelope can create unwanted frequency components that add to the frequency content of the original signal. These unwanted frequency components may be so bad that they completely obscure the signal of interest.

For example, in a pulse carrying an AM signal, the AM sidebands can be completely shrouded by the spectral noise of the pulse envelope. Time gated spectral analysis permits analysis of the contents of the pulse with the effect of the envelope on the pulse.

As the name implies, time gating is achieved by placing a gate (switch) in the video path of the spectrum analyzer. In time-gated measurement, the analyzer senses when the burst starts. It then triggers a delay so the resolution filter has time to react to the sharp rise time of the pulse. Finally, time gating stops the analysis before the burst ends. By doing this, only information carried by the pulse is analyzed.

7. AM Measurement Using Zero Span and FFT
In addition to the swept-tuned frequency mode, spectrum analyzers can also be used in the fixed-tuned mode (zero span) to provide time domain measurement capability much like that of an oscilloscope. One of the most powerful uses of zero span is demonstrated when making quick measurements of amplitude modulation.

To make AM measurements using zero span, the center frequency of the analyzer is set to the AM carrier frequency while the resolution bandwidth of the analyzer is set so that it is wide enough to pass the side-bands unattenuated. Then the analyzer is set to 0 Hz. This causes the analyzer to stop sweeping and act as a fixed-tuned receiver, displaying signal amplitude versus time as opposed to frequency versus time.

With the analyzer set to linear display mode, the display shows the sinusoidal variation in carrier amplitude modulation. The maximum modulation frequency which may be resolved using zero span is determined by the analyzer's maximum resolution bandwidth and its minimum sweep time.

While zero span provides engineers with the frequency of the modulation signal, it does not tell them about the quality of this signal. If the analyzer has a built-in Fast Fourier transform (FFT) function, engineers should perform an FFT on the zero span signal. The analyzer will now show the frequency content of the modulating signal.

8. Eliminating The Grease Pencil
In many situations, it is necessary to quickly test a signal to see whether or not it falls within a set of frequency, amplitude or time boundaries. For example, a radio transmitter manufacturer would want to make sure that the center frequency of a signal carrier falls within a certain amplitude and frequency mask and might tune a variable capacitor or resistor until it does. During this tuning, the manufacturer will require constant feedback from the spectrum analyzer indicating whether or not the carrier fits within the mask.

In some cases, grease pencils are used to sketch these limit lines right on the display of the analyzer. Modern spectrum analyzers provide electronic limit line capability, offering more precise and much cleaner ways of making these measurements.

Limit lines compare trace data to a set of amplitude and frequency (or time) parameters, while the spectrum analyzer is sweeping the measurement range. Similar to Ampcor tables, limit lines are entered into the analyzer's memory as sets of linearly corrected frequency and amplitude points. When the signal of interest falls within the limit boundaries, a display indicating LIMIT PASS (on Hewlett-Packard analyzers) appears. If the signal should fall out of the limit line boundaries, LIMIT FAIL appears on the display.

Hewlett-Packard (HP) Co., Test and Measurement Organization, Microwave Instruments Division, P.O. Box 4026, Englewood, CO 80155, Tel: (800) 452-4844.