Focus on Adaptive Equalization Option AYA Flexible

Focus on Adaptive Equalization Option AYA Flexible
The 89600 VSA software shown in this document has been replaced by the
new 89600B VSA software, which enables more simultaneous views of virtually every
aspect of complex wireless signals. The instructions provided herein can be used with the 89600B;
however, some of the menu selections have changed. For more information, please reference the 89600B
software help:
Help > Getting Started (book) > Using the 89600B VSA User Interface (book) > VSA Application Window Illustration
Option AYA Flexible Modulation Analysis
89600 Vector Signal Analysis Software
Focus on Adaptive Equalization
Self-Guided Demonstration
Adaptive equalization is an important tool for today’s receivers. It is required by
standards such as IEEE 802.16e (Mobile WiMAX™). Adaptive equalization is also
useful for analyzing signals as it allows you to remove propagation effects from
measurements and separates linear errors from non-linear errors. Determining
whether your errors are linear or non-linear can provide useful troubleshooting
information in and of itself. Adaptive equalization is a useful tool for identifying
linear errors.
Linear errors include:
Group delay distortion
Frequency response errors (tilt, ripple)
Reflections or multi-path distortion from IQ modulated signals
DSP errors in miscoded bits, or incorrect filter coefficients
Non-linear errors include:
Spectral regrowth or adjacent channel interference
Harmonic distortion
The most common types of equalizers are feed-forward and decision feedback.
Figure 1, below, shows the process that each of these equalizers uses to compute the filter coefficients.
Feed-forward equalization uses an equalization filter, before applying any other
static measurement filters that are in place. The signal is then demodulated and
measured. After measurement, the equalization filter is modified to adjust for
additional changes.
In decision feedback equalization, the static measurement filters are applied
first, then combined with the output of the previous cycle. In this method, the
modulator will ideally output a noise-free signal, so that the equalization process
itself does not add any additional noise to the signal.
Feed-forward equalization
Decision feedback qualization
Figure 1. Predominant types of equalization.
There are also two types of equalizer training techniques: one uses training
sequences embedded in the signal, and the other, called blind equalization, does
not. Figure 2, below, shows the block diagram representation for these equalizer
training techniques. While blind equalization works without any prior knowledge
of the signal, it can be slow to converge if a lot of bit errors are present. Slow
convergence makes the equalizer less responsive to rapidly changing channel
conditions. Most equalizers use training sequences. The signal must include this
sequence, and throughput is therefore reduced. The trade-off is that the equalizer
is able to converge quickly, even in the presence of many bit errors.
Blind equalization
IQ measurement
Bits (as detected)
Least mean
EQ filter coefficients
Equalization using training sequences
IQ measurement
Training sequence
Least mean
EQ filter coefficients
Figure 2. Equalizer training techniques.
The 89600 VSA software uses a feed-forward equalizer type, with a blind
equalization training technique. This provides a universal equalizer setup that
accommodates a wide range of signal types. This demonstration guide discusses
the equalizer in the 89600 VSA Option AYA vector modulation analysis software.
Additional demodulation options which use equalization are available for: WLAN,
WLAN-MIMO, WiMAX, and Mobile WiMAX. There are however, several key
differences in how equalization is used in Option AYA digital demodulation and
these other options:
• Option AYA computes Channel Frequency Response (explained in more detail below) by comparing the IQ Meas Time and IQ Ref Time data, while the formats listed above compute it from a training sequence, the preamble/data.
• Option AYA uses averaging when computing the equalizer frequency response. The format options compute a new equalizer response for each data burst.
• Option AYA allows the equalization filter to be disengaged. The other format options cannot be disabled.
Further information about these topics can be found in the Help section of the
89600 VSA Software.
Table 2. Software requirements
89600 version 1.0 or higher (89601A, 89601AN, 89601N12) – Note: substantial improvements to the options listed below
have occurred since version 1.0, and customers are encouraged
to use the latest version. Existing 89600 VSA customers can
order the 89601AS/ASN software update subscription service to
upgrade to the latest version.
• 200
• 300
(89601A, 89601AN only)
Basic vector signal analysis
Hardware connectivity
(required only if measurement hardware will be used)
Vector modulation analysis
This note is written as a guide to using the 89600 Series VSA software for
making adaptive equalization measurements. More information about the topics
discussed can be found in the help text documentation included with the
software by clicking the Help button on the main toolbar menu.
Recall sample signal
The instructions listed in Table 3 recall the signal used in the first portion of this
demonstration. Your display should look similar to Figure 3. This is a QPSK signal
with a center frequency of 5 MHz and a nominal span of 156.25 kHz.
Table 3. Recall recorded sample signal
Toolbar menus
Preset the software
File > Preset > Preset All
Note: Using Preset All will cause all saved
user state information to be lost. If this is
a concern, save the current state before
using Preset All.
Click File > Save > Setup
Go to the default signal directory
(C:\Program Files\Agilent\89600VSA\
File > Recall > Recall Recording
Select the Qpsk.dat signal
Select Qpsk.dat
Click Open
Start the measurement
Click ► (toolbar, left side)
Figure 3. Recalled recorded signal.
For best results, the center frequency, span, range, etc. should be set to their
correct values for the signal you are using.
Starting Digital
To work with adaptive equalization, the analyzer must be in digital demodulation
mode. Table 4, below, describes the method for setting up digital demodulation,
and Figure 4 shows the resulting display.
Table 4. Setting up digital demodulation
Toolbar menus
Select demodulation type
MeasSetup > Demodulator > Digital Demodulation
Change display setting
Display > Layout > Grid 2x2
Set up the demodulator
MeasSetup > Demod Properties
Click on the Format tab
Set the format parameters
Click the drop-down arrow for the Format and select QPSK
In the Symbol Rate box type 50 and select kHz from
the drop down box
Set the search parameters
Click on the Search tab
Uncheck the box next to Pulse Search
Figure 4. Digital demodulation for Qpsk.dat signal.
Filter parameters
Table 5 describes turning off the measurement filter for the Qpsk.dat signal,
and replacing it with the equalization filter. Figure 5 shows the resulting display.
The 89600 VSA software provides several standard measurement and reference
filters that can be used with your signals. The current correct measurement
filter for the Qpsk.dat signal is a Root Raised Cosine or Nyquist filter. This filter
produces excellent EVM results. However, for the purpose of this demonstration,
we will replace the standard filter with the adaptive equalizer.
Figure 5. Qpsk.dat signal with measurement filter turned off.
Figure 6 shows the Digital Demodulation Properties window. The lower half of
this display is for the equalization filter. Detailed descriptions for each of these
parameters can be found on the following pages. To reset the equalization filter
at any time, check the Reset Equalizer button.
Figure 6. Setting equalization filter parameters for Qpsk.dat signal.
Filter length
The Filter Length parameter sets the filter length for the equalizer in terms of
symbols. The shorter the filter length, the faster the filter will converge. Filter
length should be kept as short as possible for your measurement needs. If you
are measuring directly at the transmitter, filter lengths as short as a few symbols can be used. However, if you are measuring at the receiver and multi-path
effects are present, much longer filter parameters may be needed to compensate for delays in the signal path. The preset value of 21 symbols, which is used
in this demonstration, is a good starting point. It can be adjusted depending on
your needs.
The Convergence parameter sets the size between steps of the filter coefficients when the filter is being shaped. In practicality, this dictates how quickly
the filter will converge to the correct shape. Values that are too small cause
the filter to converge very slowly, making the wait time for the filter impractical.
However, if values are too large you will not be able to minimize EVM error.
The best approach is to start with a larger value and step down to reasonable
values. Typically, a good convergence value is approximately 1E-7. Figure 7,
below, shows the resulting ‘blow up’ effect of a convergence value set to 1E-5,
which is too large. This same effect will occur even when using a reasonable
convergence value if the equalizer is left to continuously run for an extended
amount of time (10+ minutes for this signal with the given parameters). This is
a phenomenon that occurs with all blind equalization filters, as the filter is not
using a training sequence, and begins to ‘over-correct’ the filter coefficients.
Figure 7. Qpsk.dat signal with convergence value set to 1E-5. Note that the same ‘blow up’ effect
occurs if the equalizer runs for a long period of time.
The Adaptive parameter determines whether or not the adaptive equalizer is
actively computing new values for the filter. If the Run selection is chosen, new
values are actively being computed, and then applied to the filter. In this manner,
the filter is continuously adapting to changes it “sees” in the output. If the Hold
selection is chosen, the last computed filter coefficients are being used in the
equalization filter, but new values are not being computed.
Table 5. Turning on the adaptive equalizer
Toolbar menus
Turn off the Measurement Filter.
Note that the constellation is now
poorly formed.
Click the Filter tab
Click the drop-down arrow for Measurement
Filter and select Off
Turn on the Equalization Filter
Click the Compensate tab
Click in the box next to Equalization Filter
Set the Equalization Filter parameters
In the Filter Length box type 21
In the Convergence box type 1E-7
Click the drop-down arrow for Adaptive and
select Run
Click the Reset Equalizer button
Figure 8 and Figure 9, on page 11, are screen shots of the equalizer at two different
times. Figure 8 was taken immediately after the equalizer began running, while
Figure 9 was taken a little over a minute later. In these figures, Trace C shows
the error vector time display, which is error vector magnitude (EVM) vs. time,
and Trace D shows the EVM error, along with several other measurements.
Notice that in Figure 8 the error magnitude is larger than that of Figure 9, both
graphically in Trace C and the numeric EVM measurement in the Error Summary
Trace of Trace D. You can observe this phenomenon by clicking the Reset Equalizer
button again and watching Trace C. Over time you will see the effects of the
adaptive equalizer take effect and significantly decrease this error. You should
also notice in these figures that ‘EQ’ appears in the upper right hand corner of
each display. This indicates that the equalization filter is active.
Figure 8. Note the high EVM vs. Time in Trace C (upper right) at start of equalization.
Figure 9. Low EVM vs. Time after allowing equalizer to run (Trace C, upper right).
Trace displays
There are several trace displays that can be used to view the filtering that the
adaptive equalizer provides. The steps to view these traces are outlined in Table
6. The data from these traces can be captured for export to a spreadsheet by
selecting the trace (by left clicking anywhere in the trace), then clicking Edit >
Copy Trace Data. Use this to create your own filters by copying the equalizer
impulse response and using its coefficients.
Channel frequency response
The Ch Frequency Response display shows the channel frequency response
of the channel being measured. It is computed by taking the inverse of the
of the equalization filter’s frequency response. It is typically used to evaluate
transmitter and/or receiver multi-path effects. Trace B in Figure 11 gives an
example of the channel frequency response.
Equalizer impulse response
The Eq Impulse Response display shows the time domain response of the
equalization filter. Initially, or on reset, the Eq Impulse Response display shows a
unit impulse response. Once the filter has been turned on, however, the adaptive
equalizer modifies the coefficients and the impulse response is changed to
“clean up” the signal. Table 6. Equalization displays
Toolbar menus
Channel Frequency
Response display
Double click the Trace B title (B: Ch1 Spectrum)
Select Channel 1 from the Type menu
Select Ch Frequency Response from the Data menu
Click OK
Symbol Table display
Double click the Trace C title (C: Ch1 QPSK Err Vect Time)
Select Channel 1 from the Type menu
Select Syms/Errs from the Data menu
Click OK
Equalizer Impulse
Response display
Double click the Trace D title (D: Ch1 QPSK Syms/Errs)
Select Channel 1 from the Type menu
Select Eq Impulse Response from the Data menu.
Click OK
Reset the equalization filter
Click the Reset Equalizer button
Auto scale Trace B
Right click anywhere in Trace B
Select Y Auto Scale
Auto scale Trace D
Right click anywhere in Trace D
Select Y Auto Scale
Figure 10 and Figure 11 show several different stages of the equalizer response
as it modifies the filter coefficients. Figure 10 shows the initial response of the
equalization filter when it first starts running. Figure 11 shows the equalizer’s
response after a little over a minute, which is approximately the optimal value.
Figure 10. Beginning of digital demodulation for Qpsk.dat signal. Note that the equalizer channel
frequency response (trace B, lower left) is only slightly varied from its initial flat response.
Figure 11. Digital demodulation for Qpsk.dat signal after about a minute, with excellent EVM (see
Trace C table), and well-formed channel frequency response.
Making measurements on digitally modulated, burst, or hopping signals can be
difficult. Agilent’s 89600 VSA software can help aid you in troubleshooting your
signals so you can make your measurements quickly and easily. The adaptive
equalizer that is a standard feature in the vector modulation analysis Option
AYA is an invaluable tool to help measure and troubleshoot your signal. With
preset modes, and table summaries of common measurements, the 89600 VSA
software is the perfect tool for any designer wanting a competitive edge in
getting their product to market faster.
Related Literature
Publication title
89600 Series Vector Signal Analysis
Technical Overview
89600 Series Vector Signal Analysis
Data Sheet
89600 Vector Signal Analyzer
demo software
Equalization Techniques and OFDM
Troubleshooting for Wireless LANs
Application Note
Hardware Measurement Platforms for
the Agilent 89600 Series Vector Signal
Analysis Software
Data Sheet
89600 Series Vector Signal Analyzers, VXI
Configuration Guide
89607A WLAN Test Suite Software
Technical Overview
Web Resources
For additional information, visit:
For more information on Agilent Technologies’ products, applications or services,
please contact your local Agilent office. The
complete list is available at:
Agilent Email Updates
Get the latest information on the
products and applications you select.
AdvancedTCA® Extensions for Instrumentation and Test (AXIe) is
an open standard that extends the
AdvancedTCA® for general purpose
and semiconductor test. Agilent
is a founding member of the AXIe
LAN eXtensions for Instruments puts
the power of Ethernet and the Web
inside your test systems. Agilent is a
founding member of the LXI consortium.
PCI eXtensions for Instrumentation
(PXI) modular instrumentation
delivers a rugged, PC-based high-performance measurement and automation system.
Agilent Channel Partners
Get the best of both worlds: Agilent’s
measurement expertise and product
breadth, combined with channel
partner convenience.
Agilent Advantage Services is committed to your success throughout
your equipment’s lifetime. We share
measurement and service expertise
to help you create the products that
change our world. To keep you competitive, we continually invest in tools
and processes that speed up calibration and repair, reduce your cost of
ownership, and move us ahead of
your development curve.
WiMAX, Mobile WiMAX, and WiMAX Forum
are trademarks of the WiMAX Forum.
Windows and MS Windows are U.S. registered
trademarks of Microsoft Corporation.
Pentium is a U.S. registered trademark of
Intel Corporation.
FireWire is a registered trademark of the
Apple Computer, Inc.
(877) 894 4414
Brazil (11) 4197 3500
01800 5064 800
United States
(800) 829 4444
Asia Pacific
Australia 1 800 629 485
800 810 0189
Hong Kong 800 938 693
India 1 800 112 929
0120 (421) 345
080 769 0800
Malaysia 1 800 888 848
Singapore 1 800 375 8100
0800 047 866
Other AP Countries(65) 375 8100 Europe & Middle East
Belgium 32 (0) 2 404 93 40
45 70 13 15 15
358 (0) 10 855 2100
0825 010 700*
*0.125 €/minute
49 (0) 7031 464 6333
1890 924 204
39 02 92 60 8484
31 (0) 20 547 2111
34 (91) 631 3300
0200-88 22 55
United Kingdom 44 (0) 118 9276201
For other unlisted countries:
Revised: October 14, 2010
Product specifications and descriptions
in this document subject to change
without notice.
© Agilent Technologies, Inc. 2011
Printed in USA, February 17, 2011
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF