Spectrum Analysis Basics - Part 3: Detector Types
In Part 2 of the spectrum analysis basics blog series, I went over the block diagram of a classic spectrum analyzer and introduced the concept of the envelope detector. This time, I’ll go over more detector types that help spectrum analyzers gather important data.
When gathering data, detectors save data in a bucket, as shown in Figure 1. A bucket helps you differentiate between the six detector types: sample, positive peak, negative peak, normal, average, and quasi-peak. Each bucket contains data from a frequency span and timeframe. These formulas determine the bucket width:
Frequency:
Time:
Figure 1: An example of data collected in one bucket width in the frequency domain.
To achieve greater accuracy, decrease the span or increase the sweep time. Doing this will increase the number of samples per bucket, bringing your measurement closer to continuous-time processing.
Detector types
Figure 1 shows sample, positive peak (also called peak), and negative peak detectors. These types of detectors are simple to understand. Normal, average, and quasi-peak detectors, however, are more complex.
Sample detection
Sample detection occurs when the spectrum analyzer selects the data point as the instantaneous level at the center of each bucket. The spectrum analyzer then draws vectors between the data points, making it appear continuous. Less data points create a trace that is not a good replication of the actual signal, as depicted in Figure 2. A display with more points, such as Figure 3, more accurately represents the signal.
Figure 2: A trace using 10 data points to display the signal shown in Figure 1
Figure 3: A trace using 40 data points to display the signal shown in Figure 1
For X-Series signal analyzers, you can choose any number of data points from 1 to 40,001. Sample detection mode helps you visualize random noise in your signal but it isn’t the preferred sampling method for sinusoidal signals. In this case, sample detection may not capture the true amplitude of signals if you set to too few data points. When the resolution bandwidth is narrower than the sample interval, or bucket width, sample mode can give erroneous results.
Peak (positive) detection
To ensure all sinusoids appear at their true amplitudes, you can display the maximum value of each bucket. Most signal analyzers use peak detection by default since it ensures you do not miss any sinusoids, regardless of the ratio between the resolution bandwidth and the bucket width. However, unlike sample mode, peak detection does not give a good representation of random noise because it only displays the maximum value in each bucket and ignores the true randomness of the noise.
Negative peak detection
Negative peak detection displays the minimum value in each bucket. Most signal analyzers offer this mode, although it is not as common as the others. Use negative peak detection in scenarios such as differentiating continuous waveform (CW) signals from impulsive signals in electromagnetic compatibility (EMC) testing.
Normal Detection
To provide a better visual display of random noise than peak detection and to avoid missing a signal like in sample mode, use normal detection mode. This algorithm classifies signals that both rise and fall (determined by positive peak and negative peak detectors) as noise. When noise occurs, an odd-numbered data point displays the maximum value encountered within the bucket. An even-numbered data point displays the minimum value within the bucket. When you encounter a sinusoidal signal, a mixing product sweeps past the intermediate frequency (IF) filter. The analyzer traces out the frequency response as the signal passes through the filter on the display as depicted in Figure 4.
Figure 4: Frequency response as the signal passes through the filter detected using normal mode.
If the resolution bandwidth is narrow relative to the bucket, some problems may occur. For instance, if the signal rises and falls during one even-numbered bucket, the analyzer only plots the minimum value encountered. If this had occurred in an odd-numbered bucket, you would have seen the maximum peak value, which is what you want to display. However, depending on the ratio of the resolution bandwidth to the bucket width, only the minimum value will be displayed. Figure 5 depicts the normal detection algorithm plotting a trace.
Figure 5: Trace points selected by the normal detection algorithm,
As you can see in Figure 5, when a signal both rises and falls in a bucket, normal detection only plots the positive or negative peak, depending on if it is an odd or even bucket.
Average Detection
Instead of taking samples and tossing out data that doesn’t represent a positive or negative peak, average detection uses all the data values collected within the time and frequency interval of a bucket. There are a few different options for averaging:
- Power (rms) averaging computes root-mean-square (rms) levels by taking the square root of the average of the squares of the voltage data measured during the bucket interval. The analyzer then squares this voltage and divides it by the characteristic input impedance of the spectrum analyzer, which is usually 50 ohms. This method is best for measuring the power of complex signals.
- Voltage averaging involves averaging the linear voltage data of the envelope signal measured during the bucket interval. It is often used in electromagnetic interference (EMI) testing for narrowband signal measurement.
- Log-power (video) averaging includes averaging the logarithmic amplitude values in dB of the envelope signal measured during the bucket interval. This method works best for observing sinusoidal signals near noise.
EMI Detectors: Average and Quasi-peak Detection
To characterize devices for EMI, you use voltage averaging for narrowband signals masked by the presence of broadband impulsive noise. In this case, a low-pass filter averages the higher-frequency components such as noise.
Quasi-peak detectors (QPD), a weighted form of peak detection, also play an integral role in EMI testing. The measured value of the QPD drops as the repetition rate of the measured signal increases. QPD is a way of measuring and quantifying the “annoyance factor” of a signal, such as listening to the radio and hearing a “pop” every few seconds due to noise.
Conclusion
Now that you know which types of detectors signal and spectrum analyzers use to gather data, you can move on to filtering to see how the analyzers smooth variations in envelope-detected amplitude. Join me next time for Part 4 of the spectrum analysis blog series. For more information, check out Application Note 150.
Want to learn spectrum analyzer basics in one place? Register for our on-demand Spectrum Analyzer Basics course to see an overview and hands-on demonstrations.