Differential structures such as backplanes and cables are the primary means for transmitting high speed serial data signals. Signal integrity of these systems is determined by the characteristics of the media such as insertion loss, crosstalk, and differential to common mode conversion.
Complete measurement of the mixed mode s-parameters is often performed by transforming single-ended s-parameters and assuming that the system is linear. In some cases, linearity cannot be assumed such as where active components are used.
This presentation describes how to measure true differential s-parameters which can be measured even in the presence of non-linear elements.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Characterization of Balanced Devices and ChannelsDr. Chris ScholzProduct Manager, Vector Network [email protected](817) 422 2512
True Differential Measurements
PORT 3
Reflectometer 3
Meas. Receiver
Ref. Receiver
Meas. Receiver
Ref. Receiver
Meas. Receiver
Ref. Receiver
PORT 2
Reflectometer 2Meas. Receiver
Ref. Receiver
Reflectometer 4
Reflectometer 1
PORT 1
PORT 4
Errorcorrected
Mag Phasedetection
and controlby software
BalancedDUT
LogicalPORT 1
LogicalPORT 2
Outline
ı Introduction to Signal Integrity Timing Signal Qualityı Balanced Architectures
The need for balanced architectures Ideal vs. non-ideal devicesı Measurement Techniques for Balanced Architectures
Single mode vs. Differential Mode Mixed mode S-Parameters TureDifferential vs. Virtual Differential TruDifferential Vector Network Analyzerı Experimental Examples
PORT 3
Reflectometer 3Meas. Receiver
Ref. Receiver
Meas. Receiver
Ref. Receiver
Meas. Receiver
Ref. Receiver
PORT 2
Reflectometer 2Meas. Receiver
Ref. Receiver
Reflectometer 4
Reflectometer 1
PORT 1
PORT 4
Errorcorrected
Mag Phasedetection
and controlby software
BalancedDUT
LogicalPORT 1
LogicalPORT 2
-24.0
-23.5
-23.0
-22.5
-22.0
-21.5
-21.0
-20.5
-20.0
-23.0
2 of 16 (Max)
FreqFreq
1 GHz1 GHz
Ch1Ch2
StartStart
-25 dBm-28 dBm
——
StopStop
0 dBm-3 dBm
Trc18Trc21
Sdd21Sdd21
dB MagdB Mag
0.5 dB /0.5 dB /
Ref -23 dBRef -23 dB
Ch1Ch2
Cal intCal int
Sdd21
3/8/2007, 1:10 PM1/29/2013 2
Introduction – Signal Integrity
ı What is Signal Integrity? Signal integrity or SI is a set of measures of the quality of an electrical signal. If the PCB or package already exists, the designer can also measure the
impairment presented by the connection using high speed instrumentationsuch as a vector network analyzer. For example, IEEE P802.3ap Task Forceuses measured S-parameters as test cases[9] for proposed solutions to theproblem of 10 Gbit/s Ethernet over backplanes. (Source: Wikipedia, last accessed 01/29/2013)
ı Two Key Aspects of SI:
Timing Signal Quality
1/29/2013 3
Introduction – Signal Integrity
ı Timing
Jitter RJ, DJ, SJ, PJ, DDJ, DCD, ISI, etc.
interconnect flight time vs bit period chip-to-chip vs on-chip packaging
1/29/2013 4
Introduction - Signal Integrity
ı Signal Quality Ringing Cross talk Distortion Ground Bounce Ground Noise Signal Loss Power Supply Noise
Noise = (S+N)-S
-6
-4
-2
0
2
4
6
8
10
time
ampl
itude
1/29/2013 5
Introduction – Signal Integrity
ı Reflection Noise Caused by impedance mismatch, vias, interconnect discontinuities
ı Crosstalk Noise Caused by electromagnetic coupling between traces and vias
ı Power and Ground Noise Caused switching noise of the power and ground delivery systems
ı Signals with equal amplitude but 180° phaseshift
• Also supports a common-mode (in-phase)signal
• Virtual ground
1 21 2
a
b
c
d
• Signals referring to ground
ı Advantages:ı High noise immunity
Minimizes Power andground plane noise
Minimizes EMIsusceptibility
Minimizes Cross talkı Low radiated noiseı High integration densityı Lower power
consumption
Balanced devices - Why balanced design?
Components withbalanced design:
• Amplifiers• Mixers• Filters (e.g. SAW filters)• PCB layout in mobile phones• LAN adapters, converters, filters• PC components (HDD control, etc)• Almost all signals high-speed serial data signals
Ideally, balanced devices transmit differential and reject common-mode signals
Ideal Balanced Device Characteristics
Differential-mode signal
Common-mode signal(EMI or ground noise)
Gain = 1
Fully balanced
Differential-mode signal
Common-mode signal(EMI or ground noise)
Gain = 1
Balanced tosingle ended
1/29/2013 10
Non-Ideal Balanced Device Characteristics
ı Non-ideal balanced devices convert modes
+
Differential to common-mode conversion
Common-mode todifferential conversion
Generates EMI
Susceptible to EMI
1/29/2013 11
Non-Ideal Balanced Device Characteristics
ı Non-ideal balanced devices convert input modes
Differential to common-mode conversion
Common-mode todifferential conversion
1/29/2013 12
1/29/2013 13
Measurement Techniques for BalancedDevices
Parameters to Test for a Balanced Device
ı Performance in pure differential modeı Performance in pure common modeı Conversion from differential mode to common mode (in both directions)ı Conversion from common mode to differential mode (in both directions)
1/29/2013 14
Balanced Devices Characteristics
ı Real Devices Propagation of both common mode and differential mode signals Mode conversion due to non-symmetric design Susceptability of noise (mainly common mode)
Detailed insight of differential/common mode response required
ı Description of balanced devices via special type of S-Parameters: Mixed-Mode S-Parameters
1/29/2013 15
Challenge when measuring balanced devices
ı Network analyzers are unbalancedı Classic Network Analyzers are 2-port instrumentsı No balanced calibration standardsı No standard reference impedance (Z0) for balanced deviceı Characterization of common and differential transmission model
1/29/2013 16
Balun
DUT
Unbalancednetwork analyzer
Measurement with Physical BalUns
Measurement with differential modesignals at a balanced device
1/29/2013 17
DUT
Unbalancednetwork analyzer
Measurement with commonmode signals at a balanceddevice
Measurement with Physical Transformers
1/29/2013 18
Balanced Device CharacteristicsBalun setup for Mixed-Mode-Characterization
Each 2-port combination between balanced and unbalanced ports isnecessary for complete mixed mode characterization.
DUTbal bal
unbal unbal
Balanced Device CharacteristicsPhysical BalUns: Disadvantagesı Calibration plane different from desired measurement planeı Degradation of measurement accuracy due to poor RF performanceı Different configurations for different modes necessary (e.g. differential to
common-mode conversion)ı Limited in frequency range
ı Solution: Us ideal (virtual) transformer to characterize mixed mode S-parameters of the
Basic Architecture: Definition of DifferentialMeasuremets
ı VirDi = Virtual differential Mode Characterization of balanced DUT as single ended DUT with mathematical
calculation of mixed-mode S-Parameters form single ended S-Parametersı TruDi = True differential Mode Stimulation of DUT with true differential and common mode signals with
calculation of mixed-mode S-Parameters from error corrected mixed modewave quantities
Measurement Principle
1/29/2013 21
Virtual Differential Measurement
ı Subsequent single ended measurements with post processing using linearsuperpositionı Applicable for all passive devices and active devices operating in their linear
regionı Large deviations compared to TruDi in large signal operation, especially in
terms of compression curve characteristics Nonlinear behavior of the DUT forbids linear superposition
1/29/2013 22
ı The DUT is stimulated using a real differential mode or real common modesignalı Better accuracy in small signal operationı Accurately measure compression under large signal operation
True Differential Measurement
1/29/2013 23
ZVA – True Differential Mode
ı Coherent sources Generation of true differential and common mode stimulus signals At least one signal output can be adjusted in amplitude and phase with
respect to the otherı Simultaneous measurement of two reference signals (a waves) and two
measurement signals (b waves)ı Four-port calibration in the reference plane Vector-corrected measurement of a single ended waves or voltages (a and b
waves)ı Calculation of true differential S-Parameters from vector corrected wave
quantities
1/29/2013 24
A True Differential Network Analyzer
PORT 3
Reflectometer 3
Meas. Receiver
Ref. Receiver
Meas. Receiver
Ref. Receiver
Meas. Receiver
Ref. Receiver
PORT 2
Reflectometer 2
Meas. Receiver
Ref. Receiver
Reflectometer 4
Reflectometer 1
PORT 1
PORT 4
Errorcorrected
Mag Phasedetection
and controlby software
BalancedDUT
LogicalPORT 1
LogicalPORT 2
1/29/2013 25
True Differential Measurements with R&S NetworkAnalyzers ZVA and ZVT
1/29/2013 26
differential mode 180° common mode 0°
Coherent signals of arbitrary phase and amplitude imbalance are possible
Sweep modes (R&S®ZVA-K6)
Sweep Modes: Frequency Phase (Phase of the stimulating signal can be swept from 0° to 180° ) Magnitude (Variation of the relative magnitude of the differential signals) “Classical” calibration techniques sufficient (full two port) Investigation of the DUT under real conditions
1/29/2013 27
Typical measurements quality parameters
ı Differential and common mode insertion lossı Differential and common mode return lossı NEXT-Measurements (Near End Crosstalk)ı FEXT-Measurements (Far End Crosstalk)ı Amplitude-Imbalanceı Phase-Imbalanceı Common-Mode Rejection Ratio (CMRR)
1/29/2013 28
1/29/2013 29
True Differential vs. Virtual DifferentialTruDi vs. VerDi
Modal Decomposition Method Principle
ı Calculation of the mixed mode S-parameters using unbalaced S-Parametersand virtual transformers
ı Describes fundamental performance in pure differential-mode operation
22212221
12111211
22212221
12111211
cccccdcd
cccccdcd
dcdcdddd
dcdcdddd
SSSSSSSSSSSSSSSS
input reflection
output reflectionforward transmission
reverse transmission
1/29/2013 35
input reflection
output reflectionforward transmission
reverse transmission
22212221
12111211
22212221
12111211
cccccdcd
cccccdcd
dcdcdddd
dcdcdddd
SSSSSSSSSSSSSSSS
Mixed Mode S-Matrix: CC Quadrant
ı Describes fundamental performance in pure common-mode operation
1/29/2013 36
Mixed Mode S-Matrix: DC Quadrant
ı Describes conversion of a common-mode stimulus to a differential-moderesponseı Terms are ideally equal to zero with perfect symmetryı Related to the generation to EMI
22212221
12111211
22212221
12111211
cccccdcd
cccccdcd
dcdcdddd
dcdcdddd
SSSSSSSSSSSSSSSS
input reflection
output reflectionforward transmission
reverse transmission
1/29/2013 37
Mixed Mode S-Matrix: CD Quadrant
ı Describes conversion of a differential-mode stimulus to a common-moderesponseı Terms are ideally equal to zero with perfect symmetryı Related to the susceptibility of EMI
input reflection
output reflectionforward transmission
reverse transmission
22212221
12111211
22212221
12111211
cccccdcd
cccccdcd
dcdcdddd
dcdcdddd
SSSSSSSSSSSSSSSS
1/29/2013 38
3-Port devicesingle ended common / differential-mode
Diff.-modeStim.
SingleEndedStim.
Com.-modeStim.
Port 2Port 1 Port 2
Port 1Differential -Mode Responsecommon-modeResponse
Single-endedResponse
Port 2
Port 2
Port 1(unbalanced)
Port 2(balanced)
differential-modecommon-mode
Single-ended
DUT
ss11 sd12 sc12
ds21 dd22 dc22
cs21 cd22 cc22
S S SS S SS S S
1/29/2013 39
1/29/2013 40
Measurement Examples: TrueDi vs.VirDi
Instrument Control of TruDi
1. Apply full n-port calibration,e.g. with CalUnit
2. Configure balancedMeasurement
3. Switch to True differentialMode
1/29/2013 41
Special Features of TruDi
ı Simultaneous display of VirDiand TruDi S-Parametersı Same “calibration” for VirDi and
TruDiı Measurement of error corrected
S-Parameters and wavequantities
(measure diff/comm powerwith diff/comm stimulation)ı Phase imbalance sweep P, f : fixed (max) = -180° to +180°ı Magnitude imbalance sweep f : fixed P(max) = -10 dB to + 10 dBm
(max)
ZVA Coherent Sources
ı Coherence Mode allows to set an arbitrary phase
and amplitude between theR&S®ZVA’s signals sources
ı R&S®ZVA-K6 True DifferentialOption
ı Applications: Modulators Antenna beam formingı Realtime measurementı In R&S®ZVA67 four individual
phase shifts
Example 1: Tunable Active Filter
-24.0
-23.5
-23.0
-22.5
-22.0
-21.5
-21.0
-20.5
-20.0
-23.0
2 of 16 (Max)
FreqFreq
1 GHz1 GHz
Ch1Ch2
StartStart
-25 dBm-28 dBm
——
StopStop
0 dBm-3 dBm
Trc18Trc21
Sdd21Sdd21
dB MagdB Mag
0.5 dB /0.5 dB /
Ref -23 dBRef -23 dB
Ch1Ch2
Cal intCal int
Sdd21
3/8/2007, 1:10 PM
Gain compressiontrue differential
virtual differential
True differential power axis has been shifted by -3 dB to equalize voltage amplitudes
1/29/2013 44
Tunable Active Filter
-18
-16
-14
-12
-10
-8
-6
-4
-2
-10
2 of 16 (Max)
PwrPwr
-20 dBm-23 dBm
Ch1Ch2
StartStart
10 MHz10 MHz
——
StopStop
2 GHz2 GHz
Trc18Trc21
Sdd21Sdd21
dB MagdB Mag
2 dB /2 dB /
Ref -10 dBRef -10 dB
Ch1Ch2
CalCal int
Sdd21
3/8/2007, 1:20 PM-18
-16
-14
-12
-10
-8
-6
-4
-2
-10
2 of 16 (Max)
PwrPwr
-10 dBm-13 dBm
Ch1Ch2
StartStart
10 MHz10 MHz
——
StopStop
2 GHz2 GHz
Trc18Trc21
Sdd21Sdd21
dB MagdB Mag
2 dB /2 dB /
Ref -10 dBRef -10 dB
Ch1Ch2
CalCal int
Sdd21
3/8/2007, 1:19 PM
• No difference between modes at low power (left),• Higher gain for true mode at high power (right)
1/29/2013 45
S-Parameters vs. Input Power
-20
-10
0
10
20
0
1
Freq 1 GHzCh3 Start -30 dBm Stop 11 dBm
Trc18 Sdd21 dB Mag 5 dB / Ref 0 dB Ca?Mkr 1Mkr 2
6.96-29.76
dBmdBm
-4.89316.145
dBdB
Sdd21
Mkr 1
Mkr 2
-20
-10
0
10
20
0
2
Freq 1 GHzCh4 Start -30 dBm Stop 11 dBm
Trc19 Sdd21 dB Mag 5 dB / Ref 0 dB Cal intMkr 1Mkr 2
6.96-29.76
dBmdBm
-0.57116.577
dBdB
Sdd21
Mkr 1
Mkr 2
-25
-15
-5
5
15
-5
3
Freq 1 GHzCh3 Start -30 dBm Stop 11 dBm
Trc20 Scd21 dB Mag 5 dB / Ref -5 dB Ca?•Mkr 1Mkr 2
6.96-29.76
dBmdBm
-16.7356.742
dBdB
Scd21
Mkr 1
Mkr 2
-25
-15
-5
5
15
-5
4
Freq 1 GHzCh4 Start -30 dBm Stop 11 dBm
Trc21 Scd21 dB Mag 5 dB / Ref -5 dB Cal intMkr 1Mkr 2
6.96-29.76
dBmdBm
-12.1837.575
dBdB
Scd21
Mkr 1
Mkr 2
-13
-11
-9
-7
-5
-9
5
Freq 1 GHzCh3 Start -30 dBm Stop 11 dBm
Trc22 Scc21 dB Mag 1 dB / Ref -9 dB Ca?Mkr 1Mkr 2
6.96-29.76
dBmdBm
-6.733-9.186
dBdB
Scc21 Mkr 1
Mkr 2
-13
-11
-9
-7
-5
-9
6
Freq 1 GHzCh4 Start -30 dBm Stop 11 dBm
Trc23 Scc21 dB Mag 1 dB / Ref -9 dB Cal intMkr 1Mkr 2
ı Approach: A model based analysis Analytical calculations using
MATLAB Experimental Verification Measurements
ı DUT The most simple bipolar
differential amplifier
ı The two inputs / outputs can be regarded as common / differential inputs andoutputs
ı a1 & afb determine the CMRR ratio between differential mode and common mode voltage gain
Gain of the individual amplifier
Compression and 3rd
order intermodulation
21
21
dd
cc
SCMRRS
Modeling the DUT
1/29/2013 50
Modeling the DUT
ı Feedback factor afb
1/29/2013 51
ı Does not include a shared feed back (CMRR →0)ı A system of two independent, ideally identical single-ended amplifiersı VirDi leads to underestimation!
Input referred 1-dB compression point
Modeling the DUT
1/29/2013 52
TruDi ↔ VirDiideal differential pair
I current sourced differential pair (CMRR →∞)I VirDi leads to overestimation!
ı Passive Devices/Linear operation TruDi and VirDi give exactly the same resultsı Active Devices/Non-linear operation Significant difference between TruDi and VirDi TruDi represents the real operating conditions of a device
ı TruDi Measurements Requires two (or more) phase coherent sources Ability to scan amplitude and phase independently Relative phase stability of VNA sources is crucial for reproducible results
ı Challenges of Fixtures Mask true device behavior No well characterized
ı Disadvantages of Physical Matching Networks: Poor reproducibility Narrow band Restricted to low frequencies Inflexible (one network for one frequency range)
ı Use of theoretically Embedded Matching Networks: Both Embedding and Deembedding Highest degree of flexibility to integrate networks No frequency restriction Possible disadvantages just with active devices
1/29/2013 65
DU
T
Introduction: Embedding
Matching Networks not present as hardware but represented by calculation
DUT+ Test Fixture+ Matching Networks
network analyzer
DUT1
Port
1
Port
2
DUT+ Test Fixture
DU
T
Available Networks:
• Import of arbitrary S-parameter files• Use of predefined matching networks
Response of testfixture, strip lines etc.
DU
T
Introduction: Deembedding
Response of networks not corrected by calibration corrected by calculation
network analyzer
Refe
renc
eplan
eat
POR
T1
w/o
calib
ratio
nor
deem
bedd
ing
Refe
renc
eplan
eat
POR
T2
w/o
calib
ratio
nor
deem
bedd
ing
Shift of reference planeby Deembedding
(or alternatively via calibration)
• Import of S-parameter files(gained e.g. using a SW design tool)
• Use of predefined networks
(De)Embedding Networks(single ended DUTs)
ı 8 predefined networks Import of *.s2p files
Single Ended Port
(De)Embedding
(De)Embedding Networks(single ended DUTs)
Pre-defined matching networks
(De)Embedding Networks(differential DUTs)
ı Import of *.s4pfiles
12 predefinednetworks
Balanced Port
(De)Embedding
(De)Embedding Networks(differential DUTs)
Pre-defined matching networks
Appendix 2: Measurement Wizard
1/29/2013 72
Measurement Wizard
Step 1 : Selection of testconfiguration
Step 2 : Impedance Settings
Measurement Wizard
Step 3 : Selection of S-Parameters
Measurement Wizard
Step 4 : General Settings
Measurement Wizard
Step 5 : Bandwidth and PowerSetting
Measurement Wizard
Step 6 : Calibration
Measurement Wizard
Measurement result of SAW Filter
Measurement Result (1)
Automatic Amplitude and Phase ImbalanceMeasurement