1 of 5

http://www.embedded.com/print/4402377

How to test modern radars

Darren McCarthy - November 29, 2012
Using a spectrum analyzer for pulsed measurements has always required a careful understanding and knowledge of the parameters of the pulse signal, as well
as the subsequent operation of the spectrum analyzer to make valid measurements. One of the primary uses of pulsed measurement techniques has been the
accurate measurement of pulsed radar signals. Advances in the type of pulse and pulse train information can greatly improve the utility of the radar; however,
these advances will increase the complexity of the measurements. Automatic pulse measurements have been introduced on modern spectrum analyzers to
greatly simplify challenging new measurements.
This article looks at traditional pulse measurements using a typical spectrum analyzer with a traditional swept-tuned architecture, discusses some of the
advances in radar waveforms, and reviews some of the important considerations for reducing the measurement uncertainty of these advanced pulse
measurements on more modern spectrum analyzer architectures.
Basic Pulse Measurements
The main advantage of using a traditional spectrum analyzer is that it can be used to test frequency-dependent power components with a wide dynamic range.
Simple measurements, such as checking the symmetry of the pulse spectrum, can validate the radar transmitter. An asymmetrical spectrum can waste power,
generate unwanted spurious emissions, and degrade the overall performance of the radar.
When making measurements using the spectrum analyzer, particularly on signals with low duty cycles, one needs to be familiar with the parameters of the
expected pulse and the important spectrum analyzer settings such as resolutions bandwidth (RBW), span, and sweep time in order to produce informative
results.
Figure 1 shows the traditional swept-tuned architecture of the spectrum analyzer. A signal is filtered and downcoverted to an IF frequency where various
resolution bandwidth and video filters are applied to the signal while the local oscillator is swept across a frequency span. The resulting energy and frequency
is plotted on the display.

Figure 1: Traditional Swept Tuned Spectrum Analyzer Architecture.

Since a pulsed signal is not on at all times, the energy will not completely fill the spectrum on a single sweep. Figure 2 shows the spectral
characteristics of a simple pulsed RF signal with a pulse width and the pulse repetition interval . The amplitude of the spectral lines are determined by the
envelope curve about the center frequency f0.

Figure 2: Typical display of pulse signal showing pulse width and pulse interval
When measuring the spectrum using a spectrum analyzer, it is possible to display the individual spectral lines or the envelope curve of the pulse spectrum
depending on the instrument settings. The RBW should be set to a value significantly less than the pulse repetition frequency (= 1/). The line spacing is
equal to the pulse period (pulse repetition interval) and is independent of the setting for the sweep time on the analyzer. The height of the individual spectral
lines is also independent of the RBW.
The largest spectral line displayed in the spectrum display is below the pulse amplitude of the actual pulse by the pulse desensitization factor(PDF). The
PDF is dependent upon the pulse width to the pulse period ratio:
PDF = 20 * log(/)

8/4/2016 4:59 PM

2 of 5

http://www.embedded.com/print/4402377

Using the line spectrum, the peak power of the pulse signal can be calculated when placing a marker on the tallest spectral line (as shown in Figure 3):
Peak power = marker reading PDF = marker reading 20*log(/).

Figure 3: Shows the line spectrum using 50 Hz RBW for 1000 Hz pulse interval
When using the maximum peak detection method, if the RBW of the analyzer is increased such that it is greater than the reciprocal of the pulse period (but
still smaller than the reciprocal of the pulse width), the spectrum analyzer will display the spectrum envelope. The amplitude of the envelope increases
linearly with the RBW, thus doubling the RBW produces a 6dB increase in the amplitude.
By continuing to increase the RBW until the RBW is greater than the reciprocal of the pulse width, the spectrum analyzer can approximate the peak power
of the pulse signal within the limitations of the resolution bandwidth of the traditional spectrum analyzer.
To demonstrate this limitation, Figure 4 shows the zero-span capture of three different pulse widths using a 10MHz RBW filter. While accurately
representing the pulse widths for 500 and 200ns durations, when the pulse width of the signal is decreased to 100ns, the peak amplitude becomes reduced due
to the filter bandwidth of the RBW filter. As the pulse width gets shorter, the limitations of the traditional spectrum analyzer impact the measurement
uncertainty.

Figure 4: Zero-span measurements of three different pulse widths using the zero-span mode of a typical spectrum analzyer (RBW = 10 MHz)
Increasing Complexities of Radar
Many modern types of radar have advanced beyond the simplistic traditional functions of range detection to improve range resolution, mitigate operational
limitations, and improve function. The impact of these modern radar types increases the complexity and performance requirements of the traditional
spectrum analyzer.
Pulse-Doppler radar provides radial speed information about the target in addition to range and direction. Using a typical coherent transmitter and receiver,
the speed information can be derived from the pulse-to-pulse variations in the received signal. Pulse-to-pulse transmitter stability verification test has
become much more demanding on the performance of the measurement equipment as phase information is not collected in a traditional spectrum analyzer.
Pulse compression radars are used to improve range resolution. Frequency modulation on pulse (FMOP) and phase modulation on pulse (PMOP) can
substantially improve the ability to resolve multiple targets at greater distances. A 2GHz bandwidth FM chirp can resolve target variances less than 10cm

8/4/2016 4:59 PM

3 of 5

http://www.embedded.com/print/4402377

apart. Some of the typical FMOP/PMOP techniques include: Linear frequency modulation (FM Chirp); non-linear frequency modulation; encoded pulse
phase modulation (e.g. Barker codes); and polyphase modulation and time-frequency coded modulation. Not only do pulse compression radars tend to
increase the need for analysis bandwidth with fast risetimes and reduced pulse widths, but to check for transmitter stability requires different types of
waveform analysis not available on traditional spectrum analyzers. To collect the necessary phase information, a baseband I/Q conversion is required to
perform the type of analysis needed to validate these types of radar transmitters.
An example of an advanced radar technique and associated measurement might be the use of a staggered pulse repetition interval (PRI). A staggered PRI
technique is used in most modern radars to overcome the limitations of a constant PRI. Constant PRI frequency radars are susceptible to self-jamming, blind
speeds, false target recognition due to double echo returns, and can also be susceptible to jamming or spoofing.
Figure 5 shows the analysis of a multi-rate PRF transmitter. Not only is the PRI varied in this case, but the pulse widths are also varied on a pulse-to-pulse
basis. This type of analysis on a traditional spectrum analyzer would not be possible as many of the measurements that were described earlier require a
constant and stable PRI for timing and spectrum measurements.

Figure 5: Measurements of Varying PRI Pulse Train Signal using the Rohde & Schwarz FSW Spectrum Analyzer and K6 Pulse Analysis Software
>>Modern spectrum analyzer architectures for testing advanced radars Modern spectrum analyzer architectures for testing advanced radars
Modern spectrum analyzer architectures for testing advanced radars
Performing analysis of modern radar signals requires a spectrum analyzer architecture that goes beyond the traditional swept-tuned spectrum analyzer.
Modern spectrum analyzers now incorporate an FFT acquisition or vector signal analysis mode of operation. This class of spectrum analyzer is also called a
signal analyzer, and real-time spectrum analyzers also use this similar acquisition architecture.

Figure 6: Architecture of vector signal analyzer

As shown in Figure 6, the vector signal analyzer has a similar front-end to a traditional swept spectrum analyzer with filtering and downconversion. However,
once the signal is placed at an IF frequency, the entire spectrum of the downconverted signal is digitized by an A/D converter and recorded into memory.
The time sampled data can then converted through FFT and waveform processing where the spectrum, time, and phase information can be preserved for
analysis.
With the vector signal analyzer, the bandwidth of the analysis in not limited by the maximum resolution bandwidth like the swept-analyzer, but the maximum
IF bandwidth in the analyzer. The IF bandwidth is defined by the A/D converter, sampling rate, and associated IF filtering. Typical vector signal analyzers
have 40 MHz, 80 MHz, and up to 160 MHz bandwidths for analysis enabling analysis of much faster risetime and pulse widths with a higher degree of
certainty as shown in Figure 7.
Acquisitions are captured into memory and subsequent FFT and waveform analysis can be performed on the acquired signals.

8/4/2016 4:59 PM

4 of 5

http://www.embedded.com/print/4402377

Figure 7: Wide IF bandwidths enable accurate peak-power measurements of narrow pulsewidths

Important considerations for pulse-pulse measurements
Representing information on individual pulses and trending information requires advanced analysis in modern spectrum analyzers. Figure 8 shows an
individual pulse analysis of a linear FM chirp. The displays show the pulse train for several pulses and tabulated analysis (green) for the individual pulses.
The individual pulse analysis (blue) for the linear FM chirp is expressed in frequency, amplitude, and phase vs time displays using the pulse analysis.

Figure 8: Advanced Pulse Analysis overview of linear FM chirp sequence

Defining the pulse parameters
As shown in Figure 8, the vector signal analyzer provides a series of time domain views and measurement results that are not available on a swept tuned
spectrum analyzer. To understand the importance of the measurement uncertainty associated with the results, it is important to define the parameters of the
pulse.

Figure 9: Pulse parameters

The rise and fall times are typically measured between the 10-90% values exclusive of the overshoot and droop associated with the pulse. The pulse width is

8/4/2016 4:59 PM

5 of 5

http://www.embedded.com/print/4402377

typically a measure of the 50% values of the rising and falling edge measured in linear units. The selection of the trigger point and phase reference point for
pulse-to-pulse measurements has a direct bearing on the measurement results.
When measuring the pulse-to-pulse performance of a radar transmitter, it is important to understand the variables that can impact the uncertainty of the
measurement system for accurate Doppler measurements, which include:
-

Signal-to-noise ratio
Signal bandwidth and filtering
Reference (or timebase) clock stability and trigger jitter
Phase noise accumulation

The same variables can also contribute to the uncertainty in the signal generator when testing the receiver circuit and Doppler measurement accuracy.
Analyzing the variables
Signal-to-Noise Ratio (SNR):
As a general rule, the higher the level of SNR, the lower the uncertainty will be due to noise contribution. While this is not typically an issue when measuring
a stable pulse, the uncertainty can increase if the pulse train is going through a power ramp mode of operation. One might also observe a power ramp making
an over-the-air measurement if the measurement is located in a fixed position while the radar antenna rotates (ex. air traffic control radar). One also must be
concerned about the measurement bandwidth of the instrument with respect to the bandwidth of signal of interest. Too much bandwidth can increase the
noise power with respect to the signal.
Signal Bandwidth and Filtering:
The bandwidth of the IF acquisition system must be sufficient to accurately represent the risetime of the pulsed signal. As mentioned previously, too much IF
bandwidth can increase the noise. However, if the bandwidth is artificially reduced to filter the pulse width of the signal, the pulse-to-pulse measurements
could be artificially reduced by the measurement instrumentation. Applying filtering to prevent overshoot and preshoot of the rising and falling edges of the
pulse can substantially improve measurement reproducibility as any ringing on the edges can impact the measurement points on a pulse-to-pulse basis. It is
important that the selection of the pulse-to-pulse measurement point, or set of measurement points, is sufficiently far away in time from the edges of the
pulse. Applying a Gaussian filter to smooth the pulse ringing can improve the measurement uncertainty of the pulse-to-pulse measurements. It should be
noted that the filtering used to make stable pulse measurement will impact other measurements such as risetime and spectrum occupancy. Care should be
taken to assure proper parameters are set for each measurement to assure reproducibility and accuracy of the results.
Reference (Timebase) Clock Stability and Trigger Jitter:
When measuring a radar signal, it is important to lock the timebase of the radar synthesizer to the measurement equipment. However, this is not always
possible, especially when measuring signals over-the-air. Some radars, such as bistatic or multistatic radars, have receivers located a great distance from the
transmitter and require synchronization via a reference clock (GPS). Between the reference clock short-term stability and the errors associated with trigger
circuit synthesis, phase ambiguity can result in pulse-to-pulse errors.
Phase Noise Accumulation:
The impact of phase noise on the measurement uncertainty is directly proportional to the measurement time and the phase noise performance at different
frequency offsets. Phase noise accumulation occurs in the interval between the reference measurement pulse and the pulse being measured. The longer the
period, the longer the accumulation of the close-in phase noise. Therefore, the performance of the phase noise at close-in offset frequencies can be one of the
most important variables to pulse-to-pulse measurements.
Summary
Simple radar signals have been traditionally measured using swept-tuned spectrum analyzers. However, modern radar signals that now include phase and
frequency modulation techniques or staggered PRI can no longer use a simple architecture to make meaningful results. Modern spectrum analyzer
architectures, such as the vector signal analyzer, are now required to make measurements of advanced pulsed radar signals. It is important to select the vector
signal analyzer with enough bandwidth and performance to make reproducible results. Vector signal analyzers, such as the Rohde & Schwarz FSW with
Option K6, are now available with advanced signal analysis software to provide accurate scalar and vector measurement of pulse signals.
If you liked this feature, and would like to see a weekly collection of related features delivered directly to your inbox, sign up for
Test&Measurement World (www.tmworld.com) newsletters here
Related Content
1 Test sets verify and calibrate radar altimeter
2 FSW analyzer shows multiple measurements in single window
3 Agilent debuts VNA and radar test capabilities
4 R&S to unveil signal and spectrum analyzer at EuMW 2012
About the Author
Darren McCarthy is the Aerospace and Defense Technical Marketing Manager for Rohde & Schwarz America. He has worked extensively in various Test
and Measurement positions for the over 20+ years including R&D engineer, R&D project manager, Product Planning, Business and Market Development.
During his career, he has also represented the US as a Technical Advisor and Working Group Member for eight years on several IEC Technical Committees
and Working Groups for international EMC standards, and currently represents R&S in several industry associations.
His published contributions include a wide variety of subjects from non-linear device characterization, real-time technology, pulse measurements and Ultrawideband measurement challenges.
He holds a BSEE from Northwestern University in Evanston, Illinois.

8/4/2016 4:59 PM