Topic 2 - Vector Signal Analysis Fundamental, Keysight
Topic 2 - Vector Signal Analysis Fundamental, Keysight
January 2017
Hewlett-Packard Journal article December 1993
details 89400 Series Vector Signal Analyzer
The Worlds First
Vector Signal
Analyzer (VSA)
Page 2
89600 VSA Software
A comprehensive toolset for demodulation & vector signal
analysis
Time Frequency Modulation
Page 3
Explore virtually every facet of today’s most
complex signals
– Measure your signal: supports >75 standards and modulation types
– Analyze proprietary signals with flexible tools for custom APSK, IQ & OFDM
Cellular Wireless Audio/Video Detection, Positioning, General
Communications Connectivity Broadcasting Tracking & Navigation RF & MW
Page 4
Agenda
Introduction to Vector Signal Analysis
Time Capture
Recordings
Page 5
Swept Spectrum Analyzer Compared to Vector Signal Analyzer
Key Characteristics
• Swept Analyzers do not measure all signal characteristics (i.e. no phase)
• Very long sweeps times required for narrow Resolution-Bandwidths
• Provides accurate amplitude vs. frequency for time stationary signals
• Gating provides accurate amplitude .v. freq. for repetitive burst signals
• Measures RF Spectrum by sweeping a signal past a fixed I -F filter
Carrier
Frequency Resolution
Bandwidth
I-F Filter
Sweep Span
Start Frequency Stop Frequency
Page 6
Vector Signal Analyzers
A new Class of Analyzer
• Digitizes complete signal, magnitude and phase
• Measures RF signals IF signals or baseband signals
• Measures Time Domain, Frequency Domain and Modulation Domain
• Accurately measures time varying signals
• Record and Playback functionality
• Transforms the digitized signal data into information via Digital Signal Processing
• Measures Vector Modulated signals
• General purpose user defined Digital demodulator
t
Time Domain Frequency Domain Modulation Domain
Page 7
Modern Signal Analyzer Block Diagram
Display
Mixer IF filter
Memory
DSP
RF signal A/D
Signal Processor
VCO
Page 8
Modern Keysight Signal Analyzer Block Diagram
Digital Detectors
Digital IF Filter •Normal
• 160 Settings •Peak
• 1 Hz to 8 MHz RBW •Minimum Peak
Wideband PreAmp • CISPR and MIL BW •Sample
Improve 1GHz • 1 Hz to 50 MHz VBW •RMS
DANL from –153 Analog IF Filter • Min Switching •Quasi Peak
dBm to –167 dBm (Single Pole) Uncertainty •CISPR RMS Average
FFT
ADC Log Detectors
Attenuator Filters
2 dB Step
Page 9
Vector Signal Analyzer Block Diagram
Mixer IF filter
Memory
A/D
DSP
RF signal
Signal Processor
VCO
Page 10
Background information - IF Processing Past
Analog
RBW filter Amplitude Only - Spectrum Analyzer (HP856x)
set here
RBW IF2 Bits
RF Log Amp/
ADC
Detector
Detected and logged level
LO1 LO2 called video, mostly dc or slow moving
Bits
ADC
Audio rates
Cos(ωt) Analog
RBW filter at best
RF/IF Sin (ωt)
set here <12bits
Gain/Phase match errors
Bits
ADC
Analog -- Digital
Page 11
Background information – IF Processing Today
ASIC
Push the digitizers up the RF chain
I
Bits
IF (MHz) Cos(ωt)
RBW filters
RF ADC Sin (ωt) done digitally
Q
LO • ADC digitizes IF!
•not detected amplitude Q
Page 12
Vector Modulation Basics
Page 13
The Advantages of Digital
Transmission
• Data compression
• Encryption
Page 14
Polar Display - Magnitude and Phase Represented Together
• Magnitude is an absolute value
Phase
0 deg
Page 15
Phasors and Signal Space
Phase
0 deg Phase
0 deg
Magnitude Change Phase Change
0 deg
0 deg
Both Change Frequency Change
Page 16
Polar vs. "I-Q" Format
• Project signal to "I" and "Q" axes
"Q"
• Polar to Rectangular Conversion
Q-Value
0 deg
"I"
I-Value
Page 17
Vector Modulation Basics
BPSK QPSK
One Bit Per Symbol Two Bits Per Symbol
Page 18
Digital, or Vector, modulation- Getting bits on
to a carrier.
State or Constellation diagram
Q Example QPSK
11 10
Page 19
Digital, or Vector, Modulation.
States, Bit Rate & Symbol Rate
Bit Rate
is the number of bits transferred per second in the
01 00
system
Symbol Rate
determines the minimum system bandwidth
requirement
Page 20
Quadrature Amplitude Modulation
16 QAM 32 QAM
Q
Page 21
Using Less Bandwidth for the Same Data
Same Bits/sec
Modulation Format Number of bits Transmission Bandwidth
per symbol
BPSK 1
QPSK 2
F/2
16QAM 4
F/4
Page 22
Major Modulation Goal: Spectral Efficiency
• Theoretical Bandwidth Efficiency Limits:
Modulation Format Spectral Efficiency
BPSK 1 bit/second/Hz
QPSK 2 bits/second/Hz
8PSK 3 bits/second/Hz
16QAM 4 bits/second/Hz
32 QAM 5 bits/second/Hz
64 QAM 6 bits/second/Hz
128 QAM 7 bits/second/Hz
256 QAM 8 bits/second/Hz
512 QAM 9 bits/second/Hz
1024 QAM 10 bits/second/Hz
2048 QAM 11 bits/second/Hz
4096 QAM 12 bits/second/Hz
8192 QAM 13 bits/second/Hz
16384 QAM 14 bits/second/Hz
Note: The figures are theoretical limits and CAN NOT be achieved in practical radios
Page 23
Transmission Bandwidth
• The spectrum of a digital signal is very wide.
• Theoretically infinite.
-20dB/decade
f Log f
Roll Off rate is a function of pulse risetime, for 0 risetime pulses
the power spectrum rolls off at –20dB/decade
Page 24
Pulse Spectrum (baseband)
NRZ baseband signal
0 Hz
Sinc(x) - the envelope of the
spectral energy
Sin( x)
Sinc( x) =
– Freq + Freq x
1
= Symbol (Bit) Rate
Sample Interval
Page 25
The Spectrum Analyzer View
• Sinc(x) Magnitude • Sinc(x) Magnitude One Sided
Page 26
The Nyquist Bandwidth
In a radio transmitter the filtering
is done at baseband.
frequency
fn= Nyquist Frequency
= Symbol Rate/2
This condition gives zero ISI (Inter Symbol Interference)
Page 27
Inter Symbol Interference
Response of two successive impulses
time
0 TS
Page 28
Filter Bandwidth Parameter "∝“
Practical Filter Shapes
1
0.8
α= 1.0
0.6
α = 0.5
α = 0.3
0.4
α=0 brickwall
0.2
0
0 0.2 0.4 0.6 0.8 1
Fs : Symbol Rate
Alpha describes the "sharpness" of the filter.
Occupied bandwidth is approximately: Symbol rate X (1 + ∝)
Page 29
Frequency and impulse response of raised
cosine (RC) filters
Page 30
Modulation Quality Error Vector Magnitude (EVM)
32QAM Which
symbol is
it???
Magnitude
Q
Error
Error Vector
Measured
Vector
Reference Q:
Vector What if the signal doesn’t land
on a constellation point?
Phase Error A:
Symbol error!
I
Page 31
Demo!
Page 32
Error Vector Magnitude Measurements
Constellation Errors Signatures
• Ideal Symbol Point 64 QAM Constellation
Q
• Random Noise
• Phase Noise
• AM/AM Distortion
I
• AM/PM Distortion
• Delay Distortion/ISI
• Spur / Interference
Page 33
Why do you need a Window ?
Implied Periodicity of the Fourier Transform assumes x(t) is periodic in T.
Implied Periodicity
Implied Periodicity
One spectral
line no leakage Spectrum
Log Log with leakage
dB dB
f
f
Exact Result
Page 34
Windows Reduce Leakage
Original
signal
Window Modified
Function Waveform
True
spectrum
Log Reduced leakage
Log
Spectrum dB spectrum
dB
with leakage
f f
Page 35
Window Functions Frequency
dB
Uniform 1 Sin x/x first side lobe -12 dB
0 ∆f
0 T/2 T
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1
Hann
dB
0
∆f
0 T/2 T 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 0
-20
-40
Gaussia
dB -60
n
-100
0 -120
-140
0 T/2 T
1
dB
Flattop
-95 dB
0
∆f
0 T/2 T 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Page 36
Window Comparison
Page 37
Window Characteristics Comparison
Time Domain Shape Freq. Domain Shape
WINDOW SUMMARY
Parameter Uniform Hanning Gaussian Flat Top (Default)
Leakage Performance: Poor Relatively Good Best Good
Frequency Resolution Poor** Relatively Good Best Better
main length fixed :
Frequency Resolution: Good Good Best Better**
RBW fixed
Normalized Equivalent 1.00 Hz-sec 1.50 Hz-sec 2.215 Hz-sec 3.8194 Hz- sec
Noise BW:
3 dB BW 0.8844 Hz-sec 1.438 Hz-sec 2.091 Hz-sec 3.767 Hz-sec
Window Shape Factor 716:1 9.1:1 4.0:1 2.45:1
Maximum Amp. Error* 3.92 (dB) 1.42 (dB) 0.68 (dB) < 0.01 (dB)
Highest Side lobe -13 (dB) -31 (dB) -125 (dB) -95 (dB)
Side lobe Fall Off -20 dB /dec. -60 dB /dec. -20 dB /dec. -20 dB /dec.
* The maximum amplitude error occurs when a signal is halfway between frequency bins.
** Frequency resolution is best for equal amplitude signals (e.g. two closely spaced sine waves), Flat Top
window side lobes are well below the dynamic range & noise floor of the VSA.
Page 38
Vector Modulation Measurements
Page 39
VSA Digital Demodulation Block Diagram
Result
Search Pulse Length
Length Detector Carrier
Freq Error
Page 40
Theory of Operation
Block diagram
Page 41
What about real-time?
• The real-time bandwidth is the frequency span at
which the FFT processing time equals the time
record length-this means all input data is included
in the average (in other words, there is no gap
between the end of one time record and the
beginning of the next).
Page 42
REAL-TIME Bandwidth
• The widest bandwidth (span) that can be selected such that every sample from
the digital filter is used in the Spectrum.
T1 T2 T3
T1 T2 T3
Real Time Processing
CALC1 CALC2 CALC3
gaps!
NOT Real Time
T1 T2 T3
CALC1 CALC2 CALC3
Calculation time includes windowing, FFT, power spectrum, Linear to Log conversion, and averaging calculation plus display update!
Page 43
Overlap Processing
• When in real time operation the analyzer can do additional signal processing.
0% Overlap
T1 T2 T3
CALC1 CALC2 CALC3
Processor idle
Tn -nth Time Record
T1 CALCn -Processing of nth Time Record
T2
T3
Window Function T4
T5
50% Overlap
T6
CALC1 CALC2 CALC3 CALC4 CALC5 CALC6
Page 44
Recording signals
Page 45
Gap free analysis
• Overlap Processing
• Recording playback requires that the recording buffer be broken into time records
for processing and display. Shown in blue below, is the normal way to do that.
This method results in the fastest playback speed.
• Overlap processing advances through the recording buffer by "sliding" the time
record window through the buffer. You controls the rate at which data is replayed
(0% to 99.99%). The result is a slower playback. Data lost by the FFT windowing
process is now analyzed as well.
Page 46
Recording signals
To record a signal
Page 47
Playback recordings
• Recording playback can be started by clicking the Play key or Control > Restart.
• Or
• Re-record
Page 48
Playback Control
To view all or a large portion of the Recording Buffer, increase the number of Frequency Points..
Expanded view of Time Gated Signal
Anti-alias Anti-alias
filter (analog)
ADC
filter (digital)
– The Start Time for Analysis can be set by typing in a Time in the Player dialog box or click and dragging
the slide bar of the time Record buffer data to be used for a measurement.
Page 49
Keysight 89601B Player Control Window
Detail
Control Icons and Indicators on Player Window Interface
Set Start time Set Stop Time
to Current to Current Graphical Time
Position Position Advance Indication
Start
Page 50
Summary Recording Post-processing
Changes in Measurement Parameters
• After data has been Recorded, many measurement parameters can be
changed, for each pass through the Recording.
Changes can be made to:
• Center Frequency & Span
• Resolution BandWidth
• Number of Frequency Points
• Window - ResBW
• Time length
• Gate Delay & length
• Average types, overlap processing
• Sweep mode, Single or Continuous
• Demodulation type
• Analog - Demodulation
• Digital - Demodulation
Page 51
Keysight Vector Signal Analyzers
Page 52
Vector Signal Analysis Block Diagram
Analog RF
Mixer IF filter
Memory
Signal DSP
A/D
input Signal Processor
VCO
Page 53
Vector Signal Analysis Block Diagram
Analog IF or IQ baseband
Instrument Hardware or PC and software
Memory
Signal DSP
A/D
input Signal Processor
Page 54
Vector Signal Analysis Block Diagram
Digital IF or IQ baseband
Instrument Hardware or PC and software
Memory
Signal DSP
input Signal Processor
Page 55
Keysight Vector Signal Analyzers
16900 Series Logic Analyzer M9391 PXI VSA
Page 56
First-to-Market in Wireless
– IEEE 802.11ac WLAN
• All signal BWs: 20, 40, 80, 80+80, 160 MHz
• All modulation formats: BPSK to 256QAM
• Up to 8x8 MIMO & multi-user MIMO
Page 57
Robust Analysis in A/D, Automotive & Satellite
Page 58
Industry-Standard Tools for Proprietary Signals
– Custom APSK
• APSK constellations, staggered
• High-order PSK (e.g. 16PSK, 32PSK)
• Amplitude Shift Keying (on-off keying)
– Custom IQ
• IQ constellation editor for custom,
proprietary, non-standard, unique,
asymmetric signals
• Longer symbol length with
synchronization robustness
– Custom OFDM
• Demodulation of custom, proprietary
OFDM-based signals
• Support for multiple streams and users
Page 59
Apply Vector Signal Analysis Across the Lifecycle
Produce consistent, comparable results from simulation to production
Page 60
Thoroughly Characterize Designs
Complex stimulus-response measurements
– Measure device distortion with the signals it will see during real-
world operation
• Wide bandwidth, high PAPR signals, like OFDM, stress devices
much more than CW tones.
Page 61
Verify Signal Problems Quickly
Powerful display increases clarity
– Isolate unexpected
behavior with unlimited
markers and marker
coupling between traces
– Intuitively analyze
today’s most complex
signals with color-coded
results by channel and
user
Page 62
Analyze & Display Multiple Signals—Simultaneously
Multi-measurements
– Perform multiple measurements at once;
display all results together
Page 63
Multi-Measurements
Signal acquisition styles
Simultaneous
Shared: Acquire once, then perform
multiple, truly simultaneous meas.
F1 F2 F3 F4 within the max span width.
Unlimited BW
Independent, multi-box:
Synchronized measurements at
same/different freqs, spans, formats,
etc. using multiple HW front ends and
triggering.
Page 64
Pinpoint the Answers to Signal Problems
Advanced troubleshooting tools for capture & playback
– Easily record signals with familiar controls
Trace
format
Trace
data
Hotspot
Page 66
89600 VSA Software
See through the complexity
Page 67
89600 VSA Ordering Information
89601B (transportable), 89601BN (floating)
Page 68
Try 89600 VSA Software Today
Free 30-day trial
– Play recorded demo signals – Learn more with Dynamic Help
• >110 signals available • Click any feature to view text
• File > Recall > Recall Demo • Help > Dynamic help
www.keysight.com/find/vsa_trial
Page 69
Additional Resources
– Literature
• Vector Signal Analysis Basics, application note, literature number
5990-7451EN
• Brochure, literature number 5990-6553EN
• Configuration guide, literature number 5990-6386EN
Page 70
Thank You
Page 71