SlideShare a Scribd company logo

S-functions Presentation: The S-parameters for nonlinear components - Measure, Model, Verify, Simulate-

NMDG NV, profile picture
NMDG NV

The document discusses S-functions, which are proposed as behavioral models for nonlinear components and applications, analogous to how S-parameters are used for linear components and applications. S-functions aim to simplify design and testing of nonlinear RF/microwave circuits by providing a uniform characterization approach, as S-parameters do for linear circuits. By extracting S-functions from a component, its nonlinear behavior can be modeled and its performance can be simulated, enabling more efficient system-level design and easier comparison to measurements during manufacturing testing. The document outlines benefits of the S-function approach and similarities to the established S-parameter methodology.

Technology
1 of 76
Downloaded 44 times
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
S-functions “The S-parameters for nonlinear components” - Measure, Model, Verify, Simulate - An application in ICE March 2011
Outline ● Why S-functions? What is the impact? ● S-functions, the “S-parameters” for all nonlinear applications? ● S-functions: Key Benefits ● Theory: from S-parameters to S-functions ● Applicability and Assumptions of S-functions ● Confidence in S-functions ● S-function Extraction and Verification Tool ● Deployment of S-functions in ADS and MWO / VSS ● Strengths and Key Capabilities ● References ● Conclusion © 2011 2
Why S-parameters? DESIGN ● Complete characterization of a component in linear mode of operation ● Derive insertion loss, return loss, gain, … ● S-parameters enable system-level interpretation of behavior of the component ● Low – pass filtering, high reflective, ... ● S-parameters enable design in conjunction with other circuits TEST in Manufacturing ● S-parameters can be extracted for the designed circuit ● The S-parameters can be measured for the manufactured circuit and can be compared S-parameters close the characterization, design and test loop © 2011 3
S-parameter Design- and Test-Cycle for Linear Applications At the Foundry Design kit Design library Passive [S-parameter based] S-parameters Devices Semiconductor Manufacturer © 2011 4
S-parameter Design- and Test-Cycle for Linear Applications At the Semiconductor Manufacturer Design Chips S-parameters Design kit Design library [S-parameter based] Iterations almost completely eliminated Design Houses System Manufacturers © 2011 5
Beyond S-parameters??? ● S-parameters, the behavioural model for “linear” applications ● Components in linear mode of operation only ● Filters ● Transistors under small signal of excitation ● ... ● Their success is based on its uniform approach for linear RF and microwave problems both to measure and to simulate ● What about components in nonlinear mode of operation? ● No uniform approach for nonlinear RF and microwave problems ● There is a lot of “trial and error” or “measure – tweak” ● S-parameters are used mainly during device modelling in conjunction with a lot of model expertise to go from small-signal to large-signal © 2011 6
“Trial-Error” Design- and Test-Cycle for “Nonlinear” Applications At the Foundry Design kit Design library Modelling Devices (6 months) Small-Signal Large-Signal Load-Pull Mathematics... Semiconductor Manufacturer © 2011 7
“Trial-Error” Design- and Test-Cycle for “Nonlinear” Applications At the Semiconductor Manufacturer ● Models do not meet the needs of application ● To time costly to develop own models ● Wafer fabs cannot worry about specific problems Initial Chips DataSheet Design Eval Board Trial and Error Design Houses System Manufacturers © 2011 8
The Magic ... S-Parameters S- S-functions © 2011 9
The S-function Design- and Test-Cycle for Active Devices At the Foundry Design kit Design library [S-functions based (*)] Devices S-Functions Semiconductor Manufacturer (*) S-functions for different applications © 2011 10
The S-function Design- and Test-Cycle for Active Devices At the Semiconductor Manufacturer Improving S-functions with application-specific information Design Chips S-functions Design kit Design library [S-functions based] Reducing the number of iterations Design Houses System Manufacturers © 2011 11
Why S-functions? Adopting the S-parameter paradigm DESIGN ● “Complete” characterization of a component in nonlinear mode of operation for specific applications under a relevant set of conditions ● Derive “insertion loss”, “return loss”, “gain”, “mismatch”, conversion coefficients ● S-functions enable system-level interpretation of behaviour of the component ● Power- dependent Low – pass filtering, power conversions, … ● Ideal for dividers, multipliers and mixers ● S-functions enable design in conjunction with other circuits ● When the signals are limited to the relevant set of conditions TEST in Manufacturing ● S-functions can be extracted for the designed circuit ● The S-functions can be measured for the manufactured circuit and can be compared S-functions close the characterization, design and test loop © 2011 12
S-Functions, the “S-parameters” for nonlinear applications? ● S-Functions, the behavioural model for nonlinear applications ● Deal with a subset of nonlinear RF and microwave phenomena in a uniform way as a natural extension of S-parameters ● Will not solve “all” nonlinear problems S-parameters Linear Applications All Other Applications S-functions NonLinear (*) Applications ● Can be “measured” How to define the application boundary? ● Can be used for design in simulators ● Can be used for test to compare with realizations in simulator (*): Nonlinear behaviour determined by a small number (e.g. 2) of tones © 2011 13
S-Functions – Key Benefits S-Functions are for nonlinear applications ... ... what S-parameters are for linear applications ● Simplify the use of HF components and circuits ● Complement limited data sheets with more complete system-level models ● Complement evaluation boards, enabling upfront more realistic simulations ● Improve and speed up the design and test process ● Adequate replacement when classic models fail ● Simulate with a behavioural model, optimized for your design problem ● Same Look and Feel as S-parameters: measure, model, verify and simulate ● Verify the realized circuit with S-functions against the simulation during test ● Shorter time to market for component manufacturers and buyers © 2011 14
From S-parameters to S-functions S-parameters measured at fixed DC bias point S f  VDC f f IDC Keep signal small to f stay in a linear mode of operation S-parameters measured at different DC bias points Linear S V DC ; f  Nonlinear I DC V DC  Remark: for the sake of intuitive explanation, mathematics is not 100% correct © 2011 15
From S-parameters to S-functions f0 2f 0 3f 0 f0 A1  f 0  VDC B2 k f 0 k : 1... 3 ... f0 B1 k f 0 A2=0 50 k : 1... 3 ... f0 2f 3f 0 0 Simple model and “easy” to measure Nonlinear Bm n f 0=F mn V DC , f 0 ,∣a 1  f 0∣ Nonlinear I DC =H V DC , f 0 ,∣a 1  f 0 ∣ © 2011 16
From S-parameters to S-functions Simple model and “easy” to measure Bm n f 0=F mn V DC , f 0 ,∣a 1  f 0∣ I DC =H V DC , f 0 ,∣a 1  f 0 ∣ ● AM-AM ● AM-PM ● Harmonic Distortion But not useful in the real world ● S-functions should be able to predict cascades Input contains harmonics !! “Easy” to measure .... not in reality f0 A1  k f 0 2f ● 0 3f 0 ● Harmonic distortion of source A2 l f 0 ● Imperfect match at input and output © 2011 17
From S-parameters to S-functions A1  k f 0 f0 2f 3f f0 0 0 B2 k f 0 2f 0 3f VDC 0 f0 k : 1... 3 ... B1 k f 0 A2 k f 0  ZL f0 f0 2f 3f 2f 3f 0 0 0 0 Bm n f 0=F mn V DC , f 0 ,∣a 1  f 0∣ I DC =H V DC , f 0 ,∣a 1  f 0 ∣ “Simple” model extension Bm n f 0=F mn V DC , f 0 ,∣a i  j f 0 ∣ , all phase comb of ai  j f 0 I DC =H V DC , f 0 ,∣a i  j f 0∣ , all phase comb of  ai  j f 0  © 2011 18
From S-parameters to S-functions Bm n f 0=F mn V DC , f 0 ,∣a i  j f 0 ∣ , all phase comb of ai  j f 0 I DC =H V DC , f 0 ,∣a i  j f 0∣ , all phase comb of  ai  j f 0  Huge number of combinations with sweeps in amplitude and phase © 2011 19
From S-parameters to S-functions What now ??? f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B 2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 3f 0 0 0 0 © 2011 20
From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B 2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 3f 0 0 0 0 In many cases : A1  k f 0 , A2 k f 0  with k 1 SMALL Linearize equations in A1  k f 0 , A2 k f 0  with k 1 © 2011 21
From S-parameters to S-functions The intuitive approach f0 f0 2f 0 3f 0 VDC A1  f 0  B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 © 2011 22
From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 © 2011 23
From S-parameters to S-functions x2 f0 f0 2f 3f 2f 0 0 0 x2 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 x2 2f 3f 0 0 LINEAR But with frequency conversion, like a mixer © 2011 24
From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 LINEAR Superposition © 2011 25
From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 0 0 0 © 2011 26
From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 3f 0 0 0 0 © 2011 27
From S-parameters to S-functions The Mathematical Approach Bm n f 0=F mn V DC , f 0 ,∣a i  j f 0 ∣ , all phase comb of ai  j f 0 I DC =H V DC , f 0 ,∣a i  j f 0∣ , all phase comb of  ai  j f 0  Linearization The S-functions * I DC =Sf 0001 V DC , A11 , A21 Sf 00ij V DC , A11 , A21  Aij Sfc 00ij V DC , A11 , A21  Aij * Bmn =Sf mn01 V DC , A11 , A21 Sf mnij V DC , A11 , A21 Aij Sfc mnij V DC , A11 , A 21 Aij “LSOP” Sf mnij “Tickle tone” m: output port with j 1 n: frequency at output port m i: input port with B mn≡B m n f 0  j: frequency at input port i with Aij ≡ Ai  j f 0  LSOP: large-signal operating point Sweep in strongly reduced LSOP © 2011 28
Sfc ??? f0 f0 2f 0 3f 0 2f 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 © 2011 29
Sfc ??? f0 f0 2f 0 3f 0 2f + df 0 VDC A1  k f 0 B2  k f 0 f0 df B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 df © 2011 30
Sfc ??? * A kl f0 f0 2f 0 3f 0 2f + df 0 VDC A1  k f 0 B2  k f 0 f0 df B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 df © 2011 31
How to extract S-functions? Select a LSOP and keep it constant * Bmn =Sf mn01 V DC , A11 , A21 Sf mnij V DC , A11 , A21  Aij Sfcmnij V DC , A11 , A 21  Aij constant constant constant Change the tickle tones Measure each time Aij , B mn Solve for the S-functions : Sf mnij LSOP , Sfc mnij  LSOP Select a new LSOP and repeat © 2011 32
Extract S-functions for a real device DC Bias f0 f0 or Large-Signal Tickling Source k f0 ZL Source DC Bias ● Repeat the following for all LSOPs of interest ● Select tickle tones Large-Signal ● Large enough to be detectable Source Tickling Source ● Small enough not to violate linearity assumption ● Measure incident and reflected waves for different tickle tones ● Model by solving for all Sf and Sfc ● Resulting into S-functions © 2011 33
S-functions for Real Devices with ZVxPlus Bias ZVxPlus Tickling Source Large-Signal Source © 2011 34
Extract S-functions for a simulated device Bias Settings Large-Signal Source Settings Tickling Source Settings Tickling Source Bias Bias Large-Signal Source © 2011 35
Assumptions of S-functions VDC v3 i3 a1 a2 DUT b1 b2 ` a 1 k f 0  , a 2 l f 0  with l , k ≠0,1 Is causing only a LINEAR perturbation on the NONINEAR behaviour Large-Signal Operating Point (LSOP) Tickle or probing tones a 1  f 0 , a 2  f 0 and v dc a 1 k f 0  , a 2 l f 0 with l , k ≠0,1 © 2011 36
The crucial question for S-functions S-parameters Linear Applications All Other Applications Some S-functions NonLinear Applications Mathematically well-defined To what applications does it apply? a 1 k f 0  , a 2 l f 0 with l , k 1 SMALL 20 dBc down from main tone BUT © 2011 37
Applicability of S-functions a1 a2 a3 DUT X DUT Y b1 b2 b3 Components Prediction ● Transistors ● Harmonic distortion ● Amplifiers ● AM – AM and AM – PM ● Dividers ● Source-pull ● Multipliers ● Load-pull ● Modulation behaviour (*) ● Intermodulation (*) : The component is assumed to be pseudo-static © 2011 38
S-functions in ICE(*) ● Sweepable Large-Signal Operating Point (LSOP) ● Auto or user-defined tickle signal level ● From simple push-the-button solution to access to expert-level details ● Visualization of component behaviour during data collection ● Easy model verification (no EDA tool required) ● Sanity checks included ● Easy export to and integration in Agilent™ ADS and AWR™ MWO ● Support for mismatched environments ● Harmonics generated by RF sources don't cause any problem ● Simple output prediction does not require EDA tool ● Export to and integration in other EDA tools on request ● Possible to extend the LSOP variables, e.g. temperature (*) Integrated Component Characterisation Environment © 2011 39
Confidence in S-functions ● Constantness of LSOP ● All Sf, and Sfc are assumed to be extracted at fixed LSOP ● e.g. variation in DC drain voltage due to changing current violates this assumption ● Interpolation capability of all Sf, and Sfc ● LSOP interleaving verification measurements ● Linearity assumption of tickle tones ● Model verification for different amplitude and phases of tickle tones Confidence Extract Measure through Simulate S-functions Verification © 2011 40
Confidence in S-functions through Verification ADS MWO ADS: Agilent Technologies Advanced Design Systems design software MWO: AWR Corporation's Microwave Office design software © 2011 41
S-functions for Real and Simulated Components / Circuits Measurements Simulations Data Collection on Real Devices On Virtual Devices using ICE (*) in Simulation Tool (ADS - MWO) Extraction S-functions S-functions Verifications Simulators S-functions Deployment (ADS - MWO) © 2011 42
S-functions Verification – Constantness of LSOP Select LSOP variable Variation on Filter for a LSOP variable drain bias voltage and value Variation on fixed input power © 2011 43
S-functions Verification – Interpolation Capability verify interpolation of b2 using independent set of measurements © 2011 44
S-functions Verification – Linearity Assumption of Tickle Tones Large-Signal DC Bias f0 f0 or Large-Signal Tickling Source k f0 ZL Source Small-Signal for harmonics DC Bias DC Bias a 2 k f 0 Switching Amplifier design  L 2 f 0   L  k f 0  L 3 f 0 a 2 2 f 0  Still small signal? a 2 3 f 0 © 2011 45
ICEBreaker – Option S-function Verification Tool ICEBreaker Predicted Compare Dataset ICEBreaker Datasets Dataset ● Measured Data ● Using S-function ● Derived Quantities ● Realistic Sweep Plan ● Incident / Reflected Waves ● Under multi-harmonic Load-Pull ● Voltage and Current ● Absolute Error ● Relative Error © 2011 46
Generation of “S-function Predicted Dataset” in ICEBreaker Generate S-function based Dataset Compare Dataset and S-function based Dataset © 2011 47
Absolute Comparisons with S-function Predictions Quantity for which to compare measurement and Idc_out prediction ● Derived quantities ➢ Idc_out ➢ Gain ➢ Pdel_in ➢ Pdel_out ➢ PAE ➢ .. ➢ Frequency selection ● Basic quantities ➢ Incident waves ➢ Reflected waves ➢ Voltages / Currents ➢ Frequency selection Measurements S-function Prediction Absolute Error © 2011 48
Relative Comparisons with S-function Predictions b_out(3f0) ● Versus ➢ Harmonic index ➢ Harmonic refl Relative Error (dB) © 2011 49
Case Study(*): EPA120B-100P  EPA120B-100P • high efficiency heterojunction power FET • power output: + 29.0dBm typ. • power gain: 11.5dB typ. @ 12 GHz Real Device Virtual Device (*) : all results in this slide set are based on S-function extraction and deployment for this device © 2011 © 2009 - NMDG NV 50
Data Collection with ZVxPlus LSOP “tickle” sweep settings detailed detailed feedback feedback of LSOP on tickle for actual for actual measurement measurement © 2011 51
Expert Details If Desired @ Data Collection Summary Monitor RF and DC source settings Observe what happens at the DUT © 2011 52
S-functions Extraction S11 S12 S22 S21 meaningful complex ~3 dB compression conjugate © 2011 53
S-functions Verifications – Residual Error zoom in on zoom in tickling on result Number of measurements required to extract S-functions amplitude of complex error for measured and predicted b2 using extraction data © 2011 54
S-functions Verification – Interpolation Capability verify interpolation of b2 using independent set of measurements © 2011 55
S-functions Verifications – Constantness LSOP Select LSOP variable Variation on Filter for a LSOP variable drain bias voltage and value Variation on fixed input power © 2011 56
S-functions in ADS Acknowledgement: With the support from Agilent Technologies providing ADS licensing © 2011 57
S-functions Useage in Agilent ADS EPAClassAMediumPower.ael EPAClassAMediumPower.dsn Set of equations using FDD Values of all Sf, Sfc © 2011 58
S-functions Verification in ADS - Schematic ● Example: Harmonic Balance simulation with measured a1 and a2 spectra ● Excitation: Sweeping through a subset of LSOP points + tickle tones Measured incident waves S-function to be verified Measurements for model verification Access to measured data using DAC Sweeping through subset of LSOP and tickle tones © 2011 59
S-functions Verification in ADS - Frequency Domain Frequency = 2 GHz, VDC1= -1.3 V, VDC2= +2.0 V Measurements at different input powers than those which were used for S-functions model extraction! circles – measurements solid lines - simulations © 2011 60
S-functions Verification in ADS - Time Domain circles – measurements Frequency = 2 GHz, VDC1= -1.3 V, VDC2= +2.0 V, Pa1 = +10.0 dBm solid lines - simulations Saturation Pinch-off © 2011 61
S-functions in ADS – Modulation Prediction - Schematic Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pa1 = +0.0 dBm – Class C © 2011 62
S-functions in ADS – Modulation Prediction - Display Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pa1 = +0.0 dBm – Class C © 2011 63
S-functions in ADS – Source-Pull - Schematic Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pav = +1.0 dBm – Class C Reflection : -0.27 © 2011 64
S-functions in ADS – Source-Pull - Schematic Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pav = +1.0 dBm – Class C Complex Conjugate Measured Input Impedance © 2011 65
S-functions in MWO Acknowledgement: With the support from AWR providing MWO / VSS licensing © 2011 66
S-functions Usage in AWR MWO MDF file SUBCKT ID=S1 BIASTEE NET="SFUNC_Model" ID=X2 1 RF 2 & RF BIASTEE 2 DC DC ID=X1 PORT 3 P=2 Z=50 Ohm Model data 2 RF 1 1 RF & DC DC PORT_PS1 3 DCVS .. P=1 Z=50 Ohm ID=V2 .. V=6 V PStart=PinMin dBm 1.16226472890446E-16,-1.35308431126191E-16 PStop=PinMax dBm PStep=PinStep dB -3.15025783237388E-15,4.72191730160887E-15 DCVS ID=V1 9.71445146547012E-17,-2.37310171513627E-15 V=-1.8 V 1.51267887105178E-15,1.07899800205757E-15 -2.98198965520413E-15,1.59594559789866E-15 -5.759281940243E-16,-1.97758476261356E-16 -1.83880688453542E-15,1.92901250528621E-15 -5.06539254985228E-16,2.94209101525666E-15 -2.50494069931051E-15,-6.73072708679001E-16 -8.60422844084496E-16,2.65412691824451E-16 -2.019218126037E-15,2.7373936450914E-15 -2.28289609438548E-15,1.70870262383715E-15 1.08246744900953E-15,-6.78276879106932E-16 -1.02955838299224E-15,-1.31578775652841E-15 5.45570533194706E-16,3.99333344169861E-15 -3.62904151174348E-15,-1.74860126378462E-15 -2.00100352953925E-15,-5.68989300120393E-16 2.14975606760426E-15,-5.53376788836601E-16 -1.16226472890446E-15,-2.44942954807925E-15 2.08166817117217E-16,2.4980018054066E-15 1.04372118124342E-05,-1.36061371223614E-05 2.85882428840978E-15,3.11556336285435E-15 -2.15105711021124E-16,1.37390099297363E-15 9.71445146547012E-16,9.95731275210687E-16 -1.69309011255336E-15,2.33146835171283E-15 4.27435864480685E-15,3.3584246494911E-15 -1.08246744900953E-15,-8.60422844084496E-16 … 2D and 3D package data © 2011 67
S-functions Verification in MWO - Frequency Domain HBT UNER2 ID=X1 Mag1=0 Ang1=0 Deg Mag2=0 Ang2=0 Deg Mag3=0 Ang3=0 Deg Fo=2 GHz Zo=50 Ohm PORT _PS1 P=1 1 2 Z=50 Ohm PStart=-20 dBm BIAST EE 2 PStop=8 dBm ID=X2 PORT PStep=1 dB 3:Bias P=2 SUBCK T 2 RF 1 1 Z=50 Ohm RF & ID=S1 DC NET ="EPA_120_B " DC 3 DCVS ID=V1 V=DC2val V DCVS ID=V2 V=DC 1val V Po with swept Pin Gain and Phase with swept Pin 40 20 19.5 20 19 F0 18.5 DB(|LSSnm(PORT_2,PORT_1,1,1)|)[1,X] 0 18 Swept Power.AP_HB Ang(LSSnm(PORT_2,PORT_1,2,1))[1,X] (Deg) 17.5 Swept Power.AP_HB -20 17 150 2F0 -40 100 DB(|Pcomp(PORT_2,1)|)[1,X] (dBm) Swept Power.AP_HB 50 3F0 DB(|Pcomp(PORT_2,2)|)[1,X] (dBm) -60 Swept Power.AP_HB DB(|Pcomp(PORT_2,3)|)[1,X] (dBm) Swept Power.AP_HB 0 -80 -50 -20 -15 -10 -5 0 5 10 -20 -15 -10 -5 0 5 10 Power (dBm) Power (dBm) © 2011 68
S-functions Verification in MWO - Time Domain Vtime(V_METER.VM1,1)[*,T] (V) Itime(I_METER.AMP1,1)[*,T] (mA) HBTUNER2 Waveform Tests.AP_HB Waveform Tests.AP_HB ID=X1 Mag1=0 Ang1=0 Deg Mag2=0 Ang2=0 Deg Mag3=0 IV Ang3=0 Deg 250 Fo=2 GHz SUBCKT Zo=50 Ohm PORT_PS1 ID=S1 200 P=1 NET="EPA_120_B" 1 2 Z=50 Ohm PStart=-20 dBm BIASTEE 2 150 PStop=8 dBm ID=X2 PORT PStep=1 dB I_METER 3:Bias P=2 2 RF 1 1 ID=AMP1 Z=50 Ohm 100 RF & DC DC 50 3 V _METER DCVS ID =VM1 ID=V1 V=DC2val V 0 15 DCVS ID=V2 V=DC1val V 10 5 DB(|Pcomp(PORT_2,1)|)[1,X] (dBm) DB(|Pcomp(PORT_2,2)| )[1,X] (dBm) DB(|Pcomp(PORT_2,3)|)[1,X] (dBm) Waveform Tests.AP_HB Waveform Tests.AP_HB Waveform Tests.AP_HB0 0 0.2 0.4 0.6 0.8 1 Harmonics 40 Time (ns) 20 0 -20 -40 -60 -80 -20 -10 0 8 Power (dBm) © 2011 69
S-functions in MWO – Modulation Prediction - Schematic DB(PWR_SPEC(TP.AMP Out,1000,0,10,0,-1,0,-1,1,0,4,0,1,0)) (dBm) DB(PWR_SPEC(TP.AMP In,1000,0,10,0,-1,0,-1,1,0,4,0,1,0)) (dBm) 16QAM System 16QAM System HBTUNER2 ID=DCBlock2 Mag1=0 Ang1=0 Deg Mag2=0 Ang2=0 Deg Mag3=0 ACPR SUBCKT Ang3=0 Deg Fo=2000 GHz 50 ID=S1 Zo=50 Ohm ISOL8R NET="EPA_120_B" ID=U1 1 2 R=50 Ohm LOSS=0 dB BIASTEE 2 ISOL=30 dB ID=X1 PORT 3:Bias P=2 2 RF RF & 1 1 Z=50 Ohm 0 DC DC PORT_PS1 3 DCVS P=1 ID=V3 Z=50 Ohm V=DC2val V PStart=-20 dBm PStop=8 dBm PStep=0.5 dB DCVS -50 ID=V1 V=DC1val V -100 -150 1.92 1.94 1.96 1.98 2 2.02 2.04 2.06 2.08 Frequency (GHz) VS A ID=M1 VA RNA ME="" VA LUE S=0 NA VG= PBA SE = -76 V e c to r S w e p t v a r ia b le S ig n a l SRC MEAS A n a ly z e r QAM_S RC G e n e r a l R e c e iv e r ID=A1 MO D=16-QA M (Gray) OUT LV L=-0.2 OLVLT Y P=Avg. Power (dB m) NL_ S AW GN RAT E=_DRA T E ID=S 1 BER ID=A3 CTRFRQ=2 GHz TP NET ="AMAM A MPM.AP_HB " PW R=-170 ID=5 SIMT YP=Harm onic Balanc e TP PLST Y P=Root Rais ed Cos ine ID=AMP In PW RT YP=Avg. Po wer, Sym bol (dBW ) VARNAME="PBA SE" ID=AMP Out ALPHA =0.35 NO SE=RF Budget only I VALUES=s tepped(-76,-40,2) LOSS=0 dB PLSLN= RFIF RQ= QAM_RX OUT FL="" ID=A 2 1 R D 2 BER IQ 3 C irc u it 5 4 B E R D e te c to r B a s e d A m p lifie r C h a n n e l N o is e 1 6 Q A M S o u rc e TP ID=RX Cons tellation 1 6 Q A M (G r a y ) S y s te m © 2011 70
S-functions in MWO – Load-Pull HBTUNER2 ID=DCBlock2 Mag1=0.8 Load Tuner (and bias) Ang1=40 Deg LP_Data_Low_Power Mag2=0 Ang2=0 Deg 1.0 Swp Max Mag3=0 0.8 -0.89 Ang3=0 Deg Fo=2 GHz 6 0. Zo=50 Ohm 0 . 2 4 . 1 2 PORT1 0 0 . 3 P=1 2 p8 0 p1: Pcomp_PORT_2_1_M_DB = -2.87 Z=50 Ohm PORT p7 4. Pwr=10 dBm 3:Bias P=2 p9 p2: Pcomp_PORT_2_1_M_DB = -2.65 SUBCKT 5. 0 2 RF 1 Z=50 Ohm 2 p6 1 ID=S1 0. RF & p10 p3: Pcomp_PORT_2_1_M_DB = -2.43 DC NET="SFUNC_Model" p5 DC 10. 0 BIASTEE p4: Pcomp_PORT_2_1_M_DB = -2.21 3 DCVS ID=X1 p4 ID=V2 10.0 V=7 V 4.0 p5: Pcomp_PORT_2_1_M_DB = -1.99 0.2 0.4 0.6 0.8 1.0 2.0 3.0 5.0 0 p3 p6: Pcomp_PORT_2_1_M_DB = -1.77 DCVS p2 ID=V1 .0 -10 p7: Pcomp_PORT_2_1_M_DB = -1.55 V=-0.6 V p1 -0 . 2 0 . p8: Pcomp_PORT_2_1_M_DB = -1.33 -5 0 . -4 p9: Pcomp_PORT_2_1_M_DB = -1.11 p10: Pcomp_PORT_2_1_M_DB = -0.89 0 . 3 4 - . 0 - 0 . 2 6 - . -0 8 Swp Min -1.0 -0. -2.87 LP_Data_High_Power S M wp ax 1.0 0 .8 25.93 6 0. .0 2 .4 0 p1: Pcomp_PORT _2_1_M_DB = 21.93 .0 3 p9 p8 p7 p6 p5 p2: Pcomp_PORT _2_1_M_DB = 22.33 p4 4 .0 2 p10 p3 5. 0 p3: Pcomp_PORT _2_1_M_DB = 22.73 0. p2 0 p4: Pcomp_PORT _2_1_M_DB = 23.13 1 0. p11 p5: Pcomp_PORT _2_1_M_DB = 23.53 10.0 0.4 0.8 3.0 0.2 0.6 1.0 2.0 4.0 5.0 p1 p6: Pcomp_PORT _2_1_M_DB = 23.93 0 0 . p7: Pcomp_PORT _2_1_M_DB = 24.33 - 10 p8: Pcomp_PORT _2_1_M_DB = 24.73 .2 -0 .0 .0 -5 p9: Pcomp_PORT _2_1_M_DB = 25.13 -4 0 p10: Pcomp_PORT _2_1_M_DB = 25.53 3. 4 - 0. - p11: Pcomp_PORT _2_1_M_DB = 25.93 0 2. .6 - -0 8 S M wp in -1.0 - 0. 21.93 © 2011 71
S-functions Strengths ● S-functions are completely public ● S-functions are transparent ● S-functions are connected into ADS and MWO and can be connected in other tools on request, e.g. Matlab ● S-functions are closely integrated with load-pull in general and especially with tuners from Focus Microwave ● S-functions extraction tool has a verification capability for LSOP constantness and interpolation capability, investigating quality of model, without needing a simulation tool ● The source- and load-pull software ICEBreaker from NMDG allows S- function verification, related to the linearity assumption ● Customers can explore the capabilities of S-functions via NMDG services © 2011 72
S-Functions - Key Capabilities ● Natural extension of S-parameters ● Reduce to S-parameters for small-signal excitation ● S-parameters are cascadeable, S-Functions are cascadeable too ● e.g. Transistors, amplifiers, dividers, multipliers ● Predict harmonic behaviour of components under different impedances ● Source – Pull ● Load – Pull ● Harmonic distortion, Waveforms ● Predict modulation behaviour of components under different impedances ● When no long-term memory effects ● Valid for multi-ports, applicable to differential components © 2011 73
References ● F. Verbeyst and M. Vanden Bossche, “VIOMAP, the S-parameter equivalent for weakly nonlinear RF and microwave devices”, published in the Microwave Symposium Digest of IEEE 1994 MTT-S International and published in the 1994 Special Symposium Issue of the MTT Transactions, vol. 42, no. 12, pp. 2531 – 2535. ● F. Verbeyst and M. Vanden Bossche, “VIOMAP, 16QAM and Spectral Regrowth: Enhanced Prediction and Predistortion based on Two-Tone Black-Box Model Extraction”, published in the Proceedings of the 45th ARFTG Conference, Orlando, June 1995 and winner of the “Best Conference Paper Award”. ● J. Verspecht and P. Van Esch, “Accurately characterizating of hard nonlinear behaviour of microwave components by the Nonlinear Network Measurement System: introducing the nonlinear scattering function,” Proc. International Workshop on Integrated Nonlinear Microwave and Millimiterwave Circuits (INMMiC), October 1998, pp.17-26. ● J. Verspecht, “Scattering functions for nonlinear behavioral modeling in the frequency domain,” IEEE MTT-S Int. Microwave Symp. Workshop, June 2003. ● J. Verspecht and D.E. Root “Polyharmonic Distortion Modeling,” IEEE Microwave Magazine, vol.7 no.3, June 2006, pp.44-57. ● D.E. Root, J. Horn, L. Betts, C. Gillease, and J. Verspecht, ”X-Parameters: The new paradigm for measurement, modeling, and design of nonlinear RF and microwave components,” Microwave Engineering Europe, December 2008, pp. 16-21. © 2011 74
Conclusion ● S-functions are a natural extension to S-parameters for nonlinear behaviour ● S-functions aren't much more complex than S-parameters ● S-functions are accurate when assumptions are not violated ● Linearity assumption ● S-function extraction is supported on R&S network analysers ● S-functions can be coupled into ADS and MWO ● S-functions can be used to compress or hide specifics of nonlinear circuits in ADS and MWO For more information info@nmdg.be www.nmdg.be © 2011 75
Explore the power of S-functions Send your device or circuit to NMDG ... ` … and NMDG sends you the S-functions Please contact NMDG at info@nmdg.be © 2011 76

More Related Content

PPT
Technosoft modeling and simulation 2006 03-27 from craig
keyven
 
PDF
Standard application for Value Stream Simulation
jhjsmits
 
PDF
MADES Seminar @ Laboratory of Model-Driven Engineering Applied to Embedded Sy...
Alessandra Bagnato
 
PDF
Kahuna Corporate Profile
kahunasystems
 
PPT
Kutulu: A Domain-specific Language for Feature-driven Product Derivation
Orçun Dayıbaş
 
PPTX
Migrating Legacy Waveforms to the Software Communications Architecture (SCA)
ADLINK Technology IoT
 
PDF
Are good SharePoint solutions only a myth?
Adis Jugo
 
PDF
PCB Virtual Prototyping with PSpice
EMA Design Automation
 
Technosoft modeling and simulation 2006 03-27 from craig
keyven
 
Standard application for Value Stream Simulation
jhjsmits
 
MADES Seminar @ Laboratory of Model-Driven Engineering Applied to Embedded Sy...
Alessandra Bagnato
 
Kahuna Corporate Profile
kahunasystems
 
Kutulu: A Domain-specific Language for Feature-driven Product Derivation
Orçun Dayıbaş
 
Migrating Legacy Waveforms to the Software Communications Architecture (SCA)
ADLINK Technology IoT
 
Are good SharePoint solutions only a myth?
Adis Jugo
 
PCB Virtual Prototyping with PSpice
EMA Design Automation
 

Viewers also liked (10)

PDF
Bild presentation - Stephen Hunt
Kenneth Geris
 
PDF
Contemporary trends
Diana V. Carriço
 
PPTX
Art and craft Movment Architecture Style.
Zill Dave
 
PPTX
Customer Experience and the 3 F's
Shane Morris
 
PDF
Ancient Greece Interior Design & Furniture
Omaima Al-Ansari
 
PPTX
anthropometry uses and applications
geo michael
 
PPTX
PLYWOOD
Abhishek Mewada
 
PPT
Ergonomics Presentation Final
krisazavache
 
DOCX
Fs 3
Florlyn Cabang
 
PPTX
Furniture styles according to time.
aestheticDinterior
 
Bild presentation - Stephen Hunt
Kenneth Geris
 
Contemporary trends
Diana V. Carriço
 
Art and craft Movment Architecture Style.
Zill Dave
 
Customer Experience and the 3 F's
Shane Morris
 
Ancient Greece Interior Design & Furniture
Omaima Al-Ansari
 
anthropometry uses and applications
geo michael
 
PLYWOOD
Abhishek Mewada
 
Ergonomics Presentation Final
krisazavache
 
Fs 3
Florlyn Cabang
 
Furniture styles according to time.
aestheticDinterior
 
Ad

Similar to S-functions Presentation: The S-parameters for nonlinear components - Measure, Model, Verify, Simulate- (20)

PDF
Presentation of NMDG nv: Your Solution Provider for Nonlinear Component Chara...
NMDG NV
 
PDF
Verteilte SoftwareEntwicklung 2011 - von klassischen Modellen bis Scrum und S...
Intland Software GmbH
 
PDF
Distributed Software Development with Scrum and Social Coding
Intland Software GmbH
 
PDF
The SENSORIA Development Environment
Istvan Rath
 
PPTX
Simulation Data Management using Aras and SharePoint
Aras
 
PDF
Imaginea Service Sheet - Performance Engineering
Imaginea
 
PDF
The challenges and opportunities in open source reuse
Ivica Crnkovic
 
DOCX
CV Ranan Fraer Apr 2016
Ranan Fraer
 
PDF
A comprehensive formal verification solution for ARM based SOC design
chiportal
 
PDF
Best practices for building and deploying predictive models over big data pre...
Kun Le
 
DOC
Bharath Chinthamani_W1
Bharath Chary
 
PDF
Swindon the making of an asic
SwindinSilicon
 
PDF
Swindon the making of an asic
SWINDONSilicon
 
PDF
FAC14_Vlach.pdf
SamHoney6
 
PDF
Multi domain product architecture: start integrated, stay integrated
Obeo
 
PDF
F1270089476650
Anil Kumar
 
PPT
Software Factories in the Real World: How an IBM WebSphere Integration Factor...
ghodgkinson
 
PDF
ChipDesignEssentials_6weeks_MOSartLabs.pdf
178r1a04l3
 
PDF
Mentor vlsi lab btech_4_1
Vijay Kannamalla
 
PDF
Multi-physics with MotionSolve
Altair
 
Presentation of NMDG nv: Your Solution Provider for Nonlinear Component Chara...
NMDG NV
 
Verteilte SoftwareEntwicklung 2011 - von klassischen Modellen bis Scrum und S...
Intland Software GmbH
 
Distributed Software Development with Scrum and Social Coding
Intland Software GmbH
 
The SENSORIA Development Environment
Istvan Rath
 
Simulation Data Management using Aras and SharePoint
Aras
 
Imaginea Service Sheet - Performance Engineering
Imaginea
 
The challenges and opportunities in open source reuse
Ivica Crnkovic
 
CV Ranan Fraer Apr 2016
Ranan Fraer
 
A comprehensive formal verification solution for ARM based SOC design
chiportal
 
Best practices for building and deploying predictive models over big data pre...
Kun Le
 
Bharath Chinthamani_W1
Bharath Chary
 
Swindon the making of an asic
SwindinSilicon
 
Swindon the making of an asic
SWINDONSilicon
 
FAC14_Vlach.pdf
SamHoney6
 
Multi domain product architecture: start integrated, stay integrated
Obeo
 
F1270089476650
Anil Kumar
 
Software Factories in the Real World: How an IBM WebSphere Integration Factor...
ghodgkinson
 
ChipDesignEssentials_6weeks_MOSartLabs.pdf
178r1a04l3
 
Mentor vlsi lab btech_4_1
Vijay Kannamalla
 
Multi-physics with MotionSolve
Altair
 
Ad

More from NMDG NV (20)

PDF
Stability under Large-Signal Conditions Presentation
NMDG NV
 
PDF
Contactless Probing Presentation
NMDG NV
 
PDF
In-Circuit Probing Presentation
NMDG NV
 
PDF
S-Functions in MATLAB Presentation
NMDG NV
 
PDF
Primary Mixer S-Functions Presentation
NMDG NV
 
PDF
ICE Remoting Presentation
NMDG NV
 
PDF
NM310S Product Brochure: Nonlinear Starter Kit For R&S VNA
NMDG NV
 
PDF
NM310S Product Brochure: Nonlinear Starter Kit For Agilent VNA
NMDG NV
 
PDF
ICEBreaker Option: S- functions Verification
NMDG NV
 
PDF
ICEBreaker Presentation: Complex Sweep Plans for Automatic Component Characte...
NMDG NV
 
PDF
ZVxPlus Presentation: Pulsed DC & RF Characterization
NMDG NV
 
PDF
ZVxPlus Application: Transistor Characterization, Reliability and Model Verif...
NMDG NV
 
PDF
ZVxPlus Application: Power Amplifier Design
NMDG NV
 
PDF
ZVxPlus Technology Update: Pulse And Modulation
NMDG NV
 
PDF
ZVxPlus Presentation: Beyond S-parameters
NMDG NV
 
PDF
S-functions Paper Presentation: Switching Amplifier Design With S-functions
NMDG NV
 
PDF
S-functions Paper: Switching Amplifier Design With S-functions
NMDG NV
 
PDF
NMDG Services Brochure
NMDG NV
 
PDF
FSLP Presentation: Fast Source- and Load-Pull using your VNA at its full powe...
NMDG NV
 
PDF
ZVxPlus Product Brochure: Nonlinear Starter Kit For R&S VNA
NMDG NV
 
Stability under Large-Signal Conditions Presentation
NMDG NV
 
Contactless Probing Presentation
NMDG NV
 
In-Circuit Probing Presentation
NMDG NV
 
S-Functions in MATLAB Presentation
NMDG NV
 
Primary Mixer S-Functions Presentation
NMDG NV
 
ICE Remoting Presentation
NMDG NV
 
NM310S Product Brochure: Nonlinear Starter Kit For R&S VNA
NMDG NV
 
NM310S Product Brochure: Nonlinear Starter Kit For Agilent VNA
NMDG NV
 
ICEBreaker Option: S- functions Verification
NMDG NV
 
ICEBreaker Presentation: Complex Sweep Plans for Automatic Component Characte...
NMDG NV
 
ZVxPlus Presentation: Pulsed DC & RF Characterization
NMDG NV
 
ZVxPlus Application: Transistor Characterization, Reliability and Model Verif...
NMDG NV
 
ZVxPlus Application: Power Amplifier Design
NMDG NV
 
ZVxPlus Technology Update: Pulse And Modulation
NMDG NV
 
ZVxPlus Presentation: Beyond S-parameters
NMDG NV
 
S-functions Paper Presentation: Switching Amplifier Design With S-functions
NMDG NV
 
S-functions Paper: Switching Amplifier Design With S-functions
NMDG NV
 
NMDG Services Brochure
NMDG NV
 
FSLP Presentation: Fast Source- and Load-Pull using your VNA at its full powe...
NMDG NV
 
ZVxPlus Product Brochure: Nonlinear Starter Kit For R&S VNA
NMDG NV
 

Recently uploaded (20)

PDF
Approach and Philosophy of On baking technology
30anthuong17
 
PDF
Empathic Computing: Creating Shared Understanding
Mark Billinghurst
 
PDF
Optimiser vos workloads AI/ML sur Amazon EC2 et AWS Graviton
Julien SIMON
 
PDF
Building Integrated photovoltaic BIPV_UPV.pdf
SandeepSingh286037
 
PDF
NewMind AI Weekly Chronicles - August'25-Week II
NewMind AI
 
PDF
Per capita expenditure prediction using model stacking based on satellite ima...
IAESIJAI
 
PDF
A comparative analysis of optical character recognition models for extracting...
IAESIJAI
 
PDF
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
Akar16
 
PDF
cuic standard and advanced reporting.pdf
SimoneTitaniumTomase
 
PDF
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
SOFTTECHHUB
 
PDF
Dropbox Q2 2025 Financial Results & Investor Presentation
Dropbox
 
PPTX
Group 1 Presentation -Planning and Decision Making .pptx
MatthewLewis227954
 
PPTX
SOPHOS-XG Firewall Administrator PPT.pptx
AgnesMunisi
 
PDF
Accuracy of neural networks in brain wave diagnosis of schizophrenia
IAESIJAI
 
PDF
Agricultural_Statistics_at_a_Glance_2022_0.pdf
SandeepSingh286037
 
PDF
Mobile App Security Testing_ A Comprehensive Guide.pdf
flufftailshop
 
PDF
Encapsulation_ Review paper, used for researhc scholars
gurumoop
 
PDF
Diabetes mellitus diagnosis method based random forest with bat algorithm
IAESIJAI
 
PPT
Teaching material agriculture food technology
LiaRayya
 
PDF
Spectral efficient network and resource selection model in 5G networks
IAESIJAI
 
Approach and Philosophy of On baking technology
30anthuong17
 
Empathic Computing: Creating Shared Understanding
Mark Billinghurst
 
Optimiser vos workloads AI/ML sur Amazon EC2 et AWS Graviton
Julien SIMON
 
Building Integrated photovoltaic BIPV_UPV.pdf
SandeepSingh286037
 
NewMind AI Weekly Chronicles - August'25-Week II
NewMind AI
 
Per capita expenditure prediction using model stacking based on satellite ima...
IAESIJAI
 
A comparative analysis of optical character recognition models for extracting...
IAESIJAI
 
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
Akar16
 
cuic standard and advanced reporting.pdf
SimoneTitaniumTomase
 
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
SOFTTECHHUB
 
Dropbox Q2 2025 Financial Results & Investor Presentation
Dropbox
 
Group 1 Presentation -Planning and Decision Making .pptx
MatthewLewis227954
 
SOPHOS-XG Firewall Administrator PPT.pptx
AgnesMunisi
 
Accuracy of neural networks in brain wave diagnosis of schizophrenia
IAESIJAI
 
Agricultural_Statistics_at_a_Glance_2022_0.pdf
SandeepSingh286037
 
Mobile App Security Testing_ A Comprehensive Guide.pdf
flufftailshop
 
Encapsulation_ Review paper, used for researhc scholars
gurumoop
 
Diabetes mellitus diagnosis method based random forest with bat algorithm
IAESIJAI
 
Teaching material agriculture food technology
LiaRayya
 
Spectral efficient network and resource selection model in 5G networks
IAESIJAI
 

S-functions Presentation: The S-parameters for nonlinear components - Measure, Model, Verify, Simulate-

  • 1. S-functions “The S-parameters for nonlinear components” - Measure, Model, Verify, Simulate - An application in ICE March 2011
  • 2. Outline ● Why S-functions? What is the impact? ● S-functions, the “S-parameters” for all nonlinear applications? ● S-functions: Key Benefits ● Theory: from S-parameters to S-functions ● Applicability and Assumptions of S-functions ● Confidence in S-functions ● S-function Extraction and Verification Tool ● Deployment of S-functions in ADS and MWO / VSS ● Strengths and Key Capabilities ● References ● Conclusion © 2011 2
  • 3. Why S-parameters? DESIGN ● Complete characterization of a component in linear mode of operation ● Derive insertion loss, return loss, gain, … ● S-parameters enable system-level interpretation of behavior of the component ● Low – pass filtering, high reflective, ... ● S-parameters enable design in conjunction with other circuits TEST in Manufacturing ● S-parameters can be extracted for the designed circuit ● The S-parameters can be measured for the manufactured circuit and can be compared S-parameters close the characterization, design and test loop © 2011 3
  • 4. S-parameter Design- and Test-Cycle for Linear Applications At the Foundry Design kit Design library Passive [S-parameter based] S-parameters Devices Semiconductor Manufacturer © 2011 4
  • 5. S-parameter Design- and Test-Cycle for Linear Applications At the Semiconductor Manufacturer Design Chips S-parameters Design kit Design library [S-parameter based] Iterations almost completely eliminated Design Houses System Manufacturers © 2011 5
  • 6. Beyond S-parameters??? ● S-parameters, the behavioural model for “linear” applications ● Components in linear mode of operation only ● Filters ● Transistors under small signal of excitation ● ... ● Their success is based on its uniform approach for linear RF and microwave problems both to measure and to simulate ● What about components in nonlinear mode of operation? ● No uniform approach for nonlinear RF and microwave problems ● There is a lot of “trial and error” or “measure – tweak” ● S-parameters are used mainly during device modelling in conjunction with a lot of model expertise to go from small-signal to large-signal © 2011 6
  • 7. “Trial-Error” Design- and Test-Cycle for “Nonlinear” Applications At the Foundry Design kit Design library Modelling Devices (6 months) Small-Signal Large-Signal Load-Pull Mathematics... Semiconductor Manufacturer © 2011 7
  • 8. “Trial-Error” Design- and Test-Cycle for “Nonlinear” Applications At the Semiconductor Manufacturer ● Models do not meet the needs of application ● To time costly to develop own models ● Wafer fabs cannot worry about specific problems Initial Chips DataSheet Design Eval Board Trial and Error Design Houses System Manufacturers © 2011 8
  • 9. The Magic ... S-Parameters S- S-functions © 2011 9
  • 10. The S-function Design- and Test-Cycle for Active Devices At the Foundry Design kit Design library [S-functions based (*)] Devices S-Functions Semiconductor Manufacturer (*) S-functions for different applications © 2011 10
  • 11. The S-function Design- and Test-Cycle for Active Devices At the Semiconductor Manufacturer Improving S-functions with application-specific information Design Chips S-functions Design kit Design library [S-functions based] Reducing the number of iterations Design Houses System Manufacturers © 2011 11
  • 12. Why S-functions? Adopting the S-parameter paradigm DESIGN ● “Complete” characterization of a component in nonlinear mode of operation for specific applications under a relevant set of conditions ● Derive “insertion loss”, “return loss”, “gain”, “mismatch”, conversion coefficients ● S-functions enable system-level interpretation of behaviour of the component ● Power- dependent Low – pass filtering, power conversions, … ● Ideal for dividers, multipliers and mixers ● S-functions enable design in conjunction with other circuits ● When the signals are limited to the relevant set of conditions TEST in Manufacturing ● S-functions can be extracted for the designed circuit ● The S-functions can be measured for the manufactured circuit and can be compared S-functions close the characterization, design and test loop © 2011 12
  • 13. S-Functions, the “S-parameters” for nonlinear applications? ● S-Functions, the behavioural model for nonlinear applications ● Deal with a subset of nonlinear RF and microwave phenomena in a uniform way as a natural extension of S-parameters ● Will not solve “all” nonlinear problems S-parameters Linear Applications All Other Applications S-functions NonLinear (*) Applications ● Can be “measured” How to define the application boundary? ● Can be used for design in simulators ● Can be used for test to compare with realizations in simulator (*): Nonlinear behaviour determined by a small number (e.g. 2) of tones © 2011 13
  • 14. S-Functions – Key Benefits S-Functions are for nonlinear applications ... ... what S-parameters are for linear applications ● Simplify the use of HF components and circuits ● Complement limited data sheets with more complete system-level models ● Complement evaluation boards, enabling upfront more realistic simulations ● Improve and speed up the design and test process ● Adequate replacement when classic models fail ● Simulate with a behavioural model, optimized for your design problem ● Same Look and Feel as S-parameters: measure, model, verify and simulate ● Verify the realized circuit with S-functions against the simulation during test ● Shorter time to market for component manufacturers and buyers © 2011 14
  • 15. From S-parameters to S-functions S-parameters measured at fixed DC bias point S f  VDC f f IDC Keep signal small to f stay in a linear mode of operation S-parameters measured at different DC bias points Linear S V DC ; f  Nonlinear I DC V DC  Remark: for the sake of intuitive explanation, mathematics is not 100% correct © 2011 15
  • 16. From S-parameters to S-functions f0 2f 0 3f 0 f0 A1  f 0  VDC B2 k f 0 k : 1... 3 ... f0 B1 k f 0 A2=0 50 k : 1... 3 ... f0 2f 3f 0 0 Simple model and “easy” to measure Nonlinear Bm n f 0=F mn V DC , f 0 ,∣a 1  f 0∣ Nonlinear I DC =H V DC , f 0 ,∣a 1  f 0 ∣ © 2011 16
  • 17. From S-parameters to S-functions Simple model and “easy” to measure Bm n f 0=F mn V DC , f 0 ,∣a 1  f 0∣ I DC =H V DC , f 0 ,∣a 1  f 0 ∣ ● AM-AM ● AM-PM ● Harmonic Distortion But not useful in the real world ● S-functions should be able to predict cascades Input contains harmonics !! “Easy” to measure .... not in reality f0 A1  k f 0 2f ● 0 3f 0 ● Harmonic distortion of source A2 l f 0 ● Imperfect match at input and output © 2011 17
  • 18. From S-parameters to S-functions A1  k f 0 f0 2f 3f f0 0 0 B2 k f 0 2f 0 3f VDC 0 f0 k : 1... 3 ... B1 k f 0 A2 k f 0  ZL f0 f0 2f 3f 2f 3f 0 0 0 0 Bm n f 0=F mn V DC , f 0 ,∣a 1  f 0∣ I DC =H V DC , f 0 ,∣a 1  f 0 ∣ “Simple” model extension Bm n f 0=F mn V DC , f 0 ,∣a i  j f 0 ∣ , all phase comb of ai  j f 0 I DC =H V DC , f 0 ,∣a i  j f 0∣ , all phase comb of  ai  j f 0  © 2011 18
  • 19. From S-parameters to S-functions Bm n f 0=F mn V DC , f 0 ,∣a i  j f 0 ∣ , all phase comb of ai  j f 0 I DC =H V DC , f 0 ,∣a i  j f 0∣ , all phase comb of  ai  j f 0  Huge number of combinations with sweeps in amplitude and phase © 2011 19
  • 20. From S-parameters to S-functions What now ??? f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B 2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 3f 0 0 0 0 © 2011 20
  • 21. From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B 2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 3f 0 0 0 0 In many cases : A1  k f 0 , A2 k f 0  with k 1 SMALL Linearize equations in A1  k f 0 , A2 k f 0  with k 1 © 2011 21
  • 22. From S-parameters to S-functions The intuitive approach f0 f0 2f 0 3f 0 VDC A1  f 0  B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 © 2011 22
  • 23. From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 © 2011 23
  • 24. From S-parameters to S-functions x2 f0 f0 2f 3f 2f 0 0 0 x2 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 x2 2f 3f 0 0 LINEAR But with frequency conversion, like a mixer © 2011 24
  • 25. From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 LINEAR Superposition © 2011 25
  • 26. From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 0 0 0 © 2011 26
  • 27. From S-parameters to S-functions f0 f0 2f 0 3f 0 2f 3f 0 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2 k f 0  ZL f0 2f 3f 2f 3f 0 0 0 0 © 2011 27
  • 28. From S-parameters to S-functions The Mathematical Approach Bm n f 0=F mn V DC , f 0 ,∣a i  j f 0 ∣ , all phase comb of ai  j f 0 I DC =H V DC , f 0 ,∣a i  j f 0∣ , all phase comb of  ai  j f 0  Linearization The S-functions * I DC =Sf 0001 V DC , A11 , A21 Sf 00ij V DC , A11 , A21  Aij Sfc 00ij V DC , A11 , A21  Aij * Bmn =Sf mn01 V DC , A11 , A21 Sf mnij V DC , A11 , A21 Aij Sfc mnij V DC , A11 , A 21 Aij “LSOP” Sf mnij “Tickle tone” m: output port with j 1 n: frequency at output port m i: input port with B mn≡B m n f 0  j: frequency at input port i with Aij ≡ Ai  j f 0  LSOP: large-signal operating point Sweep in strongly reduced LSOP © 2011 28
  • 29. Sfc ??? f0 f0 2f 0 3f 0 2f 0 VDC A1  k f 0 B2  k f 0 f0 B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 © 2011 29
  • 30. Sfc ??? f0 f0 2f 0 3f 0 2f + df 0 VDC A1  k f 0 B2  k f 0 f0 df B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 df © 2011 30
  • 31. Sfc ??? * A kl f0 f0 2f 0 3f 0 2f + df 0 VDC A1  k f 0 B2  k f 0 f0 df B1 k f 0  f0 A2  f 0  ZL f0 2f 3f 0 0 df © 2011 31
  • 32. How to extract S-functions? Select a LSOP and keep it constant * Bmn =Sf mn01 V DC , A11 , A21 Sf mnij V DC , A11 , A21  Aij Sfcmnij V DC , A11 , A 21  Aij constant constant constant Change the tickle tones Measure each time Aij , B mn Solve for the S-functions : Sf mnij LSOP , Sfc mnij  LSOP Select a new LSOP and repeat © 2011 32
  • 33. Extract S-functions for a real device DC Bias f0 f0 or Large-Signal Tickling Source k f0 ZL Source DC Bias ● Repeat the following for all LSOPs of interest ● Select tickle tones Large-Signal ● Large enough to be detectable Source Tickling Source ● Small enough not to violate linearity assumption ● Measure incident and reflected waves for different tickle tones ● Model by solving for all Sf and Sfc ● Resulting into S-functions © 2011 33
  • 34. S-functions for Real Devices with ZVxPlus Bias ZVxPlus Tickling Source Large-Signal Source © 2011 34
  • 35. Extract S-functions for a simulated device Bias Settings Large-Signal Source Settings Tickling Source Settings Tickling Source Bias Bias Large-Signal Source © 2011 35
  • 36. Assumptions of S-functions VDC v3 i3 a1 a2 DUT b1 b2 ` a 1 k f 0  , a 2 l f 0  with l , k ≠0,1 Is causing only a LINEAR perturbation on the NONINEAR behaviour Large-Signal Operating Point (LSOP) Tickle or probing tones a 1  f 0 , a 2  f 0 and v dc a 1 k f 0  , a 2 l f 0 with l , k ≠0,1 © 2011 36
  • 37. The crucial question for S-functions S-parameters Linear Applications All Other Applications Some S-functions NonLinear Applications Mathematically well-defined To what applications does it apply? a 1 k f 0  , a 2 l f 0 with l , k 1 SMALL 20 dBc down from main tone BUT © 2011 37
  • 38. Applicability of S-functions a1 a2 a3 DUT X DUT Y b1 b2 b3 Components Prediction ● Transistors ● Harmonic distortion ● Amplifiers ● AM – AM and AM – PM ● Dividers ● Source-pull ● Multipliers ● Load-pull ● Modulation behaviour (*) ● Intermodulation (*) : The component is assumed to be pseudo-static © 2011 38
  • 39. S-functions in ICE(*) ● Sweepable Large-Signal Operating Point (LSOP) ● Auto or user-defined tickle signal level ● From simple push-the-button solution to access to expert-level details ● Visualization of component behaviour during data collection ● Easy model verification (no EDA tool required) ● Sanity checks included ● Easy export to and integration in Agilent™ ADS and AWR™ MWO ● Support for mismatched environments ● Harmonics generated by RF sources don't cause any problem ● Simple output prediction does not require EDA tool ● Export to and integration in other EDA tools on request ● Possible to extend the LSOP variables, e.g. temperature (*) Integrated Component Characterisation Environment © 2011 39
  • 40. Confidence in S-functions ● Constantness of LSOP ● All Sf, and Sfc are assumed to be extracted at fixed LSOP ● e.g. variation in DC drain voltage due to changing current violates this assumption ● Interpolation capability of all Sf, and Sfc ● LSOP interleaving verification measurements ● Linearity assumption of tickle tones ● Model verification for different amplitude and phases of tickle tones Confidence Extract Measure through Simulate S-functions Verification © 2011 40
  • 41. Confidence in S-functions through Verification ADS MWO ADS: Agilent Technologies Advanced Design Systems design software MWO: AWR Corporation's Microwave Office design software © 2011 41
  • 42. S-functions for Real and Simulated Components / Circuits Measurements Simulations Data Collection on Real Devices On Virtual Devices using ICE (*) in Simulation Tool (ADS - MWO) Extraction S-functions S-functions Verifications Simulators S-functions Deployment (ADS - MWO) © 2011 42
  • 43. S-functions Verification – Constantness of LSOP Select LSOP variable Variation on Filter for a LSOP variable drain bias voltage and value Variation on fixed input power © 2011 43
  • 44. S-functions Verification – Interpolation Capability verify interpolation of b2 using independent set of measurements © 2011 44
  • 45. S-functions Verification – Linearity Assumption of Tickle Tones Large-Signal DC Bias f0 f0 or Large-Signal Tickling Source k f0 ZL Source Small-Signal for harmonics DC Bias DC Bias a 2 k f 0 Switching Amplifier design  L 2 f 0   L  k f 0  L 3 f 0 a 2 2 f 0  Still small signal? a 2 3 f 0 © 2011 45
  • 46. ICEBreaker – Option S-function Verification Tool ICEBreaker Predicted Compare Dataset ICEBreaker Datasets Dataset ● Measured Data ● Using S-function ● Derived Quantities ● Realistic Sweep Plan ● Incident / Reflected Waves ● Under multi-harmonic Load-Pull ● Voltage and Current ● Absolute Error ● Relative Error © 2011 46
  • 47. Generation of “S-function Predicted Dataset” in ICEBreaker Generate S-function based Dataset Compare Dataset and S-function based Dataset © 2011 47
  • 48. Absolute Comparisons with S-function Predictions Quantity for which to compare measurement and Idc_out prediction ● Derived quantities ➢ Idc_out ➢ Gain ➢ Pdel_in ➢ Pdel_out ➢ PAE ➢ .. ➢ Frequency selection ● Basic quantities ➢ Incident waves ➢ Reflected waves ➢ Voltages / Currents ➢ Frequency selection Measurements S-function Prediction Absolute Error © 2011 48
  • 49. Relative Comparisons with S-function Predictions b_out(3f0) ● Versus ➢ Harmonic index ➢ Harmonic refl Relative Error (dB) © 2011 49
  • 50. Case Study(*): EPA120B-100P  EPA120B-100P • high efficiency heterojunction power FET • power output: + 29.0dBm typ. • power gain: 11.5dB typ. @ 12 GHz Real Device Virtual Device (*) : all results in this slide set are based on S-function extraction and deployment for this device © 2011 © 2009 - NMDG NV 50
  • 51. Data Collection with ZVxPlus LSOP “tickle” sweep settings detailed detailed feedback feedback of LSOP on tickle for actual for actual measurement measurement © 2011 51
  • 52. Expert Details If Desired @ Data Collection Summary Monitor RF and DC source settings Observe what happens at the DUT © 2011 52
  • 53. S-functions Extraction S11 S12 S22 S21 meaningful complex ~3 dB compression conjugate © 2011 53
  • 54. S-functions Verifications – Residual Error zoom in on zoom in tickling on result Number of measurements required to extract S-functions amplitude of complex error for measured and predicted b2 using extraction data © 2011 54
  • 55. S-functions Verification – Interpolation Capability verify interpolation of b2 using independent set of measurements © 2011 55
  • 56. S-functions Verifications – Constantness LSOP Select LSOP variable Variation on Filter for a LSOP variable drain bias voltage and value Variation on fixed input power © 2011 56
  • 57. S-functions in ADS Acknowledgement: With the support from Agilent Technologies providing ADS licensing © 2011 57
  • 58. S-functions Useage in Agilent ADS EPAClassAMediumPower.ael EPAClassAMediumPower.dsn Set of equations using FDD Values of all Sf, Sfc © 2011 58
  • 59. S-functions Verification in ADS - Schematic ● Example: Harmonic Balance simulation with measured a1 and a2 spectra ● Excitation: Sweeping through a subset of LSOP points + tickle tones Measured incident waves S-function to be verified Measurements for model verification Access to measured data using DAC Sweeping through subset of LSOP and tickle tones © 2011 59
  • 60. S-functions Verification in ADS - Frequency Domain Frequency = 2 GHz, VDC1= -1.3 V, VDC2= +2.0 V Measurements at different input powers than those which were used for S-functions model extraction! circles – measurements solid lines - simulations © 2011 60
  • 61. S-functions Verification in ADS - Time Domain circles – measurements Frequency = 2 GHz, VDC1= -1.3 V, VDC2= +2.0 V, Pa1 = +10.0 dBm solid lines - simulations Saturation Pinch-off © 2011 61
  • 62. S-functions in ADS – Modulation Prediction - Schematic Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pa1 = +0.0 dBm – Class C © 2011 62
  • 63. S-functions in ADS – Modulation Prediction - Display Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pa1 = +0.0 dBm – Class C © 2011 63
  • 64. S-functions in ADS – Source-Pull - Schematic Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pav = +1.0 dBm – Class C Reflection : -0.27 © 2011 64
  • 65. S-functions in ADS – Source-Pull - Schematic Frequency = 2 GHz, VDC1= -1.37 V, VDC2= +4.0 V, Pav = +1.0 dBm – Class C Complex Conjugate Measured Input Impedance © 2011 65
  • 66. S-functions in MWO Acknowledgement: With the support from AWR providing MWO / VSS licensing © 2011 66
  • 67. S-functions Usage in AWR MWO MDF file SUBCKT ID=S1 BIASTEE NET="SFUNC_Model" ID=X2 1 RF 2 & RF BIASTEE 2 DC DC ID=X1 PORT 3 P=2 Z=50 Ohm Model data 2 RF 1 1 RF & DC DC PORT_PS1 3 DCVS .. P=1 Z=50 Ohm ID=V2 .. V=6 V PStart=PinMin dBm 1.16226472890446E-16,-1.35308431126191E-16 PStop=PinMax dBm PStep=PinStep dB -3.15025783237388E-15,4.72191730160887E-15 DCVS ID=V1 9.71445146547012E-17,-2.37310171513627E-15 V=-1.8 V 1.51267887105178E-15,1.07899800205757E-15 -2.98198965520413E-15,1.59594559789866E-15 -5.759281940243E-16,-1.97758476261356E-16 -1.83880688453542E-15,1.92901250528621E-15 -5.06539254985228E-16,2.94209101525666E-15 -2.50494069931051E-15,-6.73072708679001E-16 -8.60422844084496E-16,2.65412691824451E-16 -2.019218126037E-15,2.7373936450914E-15 -2.28289609438548E-15,1.70870262383715E-15 1.08246744900953E-15,-6.78276879106932E-16 -1.02955838299224E-15,-1.31578775652841E-15 5.45570533194706E-16,3.99333344169861E-15 -3.62904151174348E-15,-1.74860126378462E-15 -2.00100352953925E-15,-5.68989300120393E-16 2.14975606760426E-15,-5.53376788836601E-16 -1.16226472890446E-15,-2.44942954807925E-15 2.08166817117217E-16,2.4980018054066E-15 1.04372118124342E-05,-1.36061371223614E-05 2.85882428840978E-15,3.11556336285435E-15 -2.15105711021124E-16,1.37390099297363E-15 9.71445146547012E-16,9.95731275210687E-16 -1.69309011255336E-15,2.33146835171283E-15 4.27435864480685E-15,3.3584246494911E-15 -1.08246744900953E-15,-8.60422844084496E-16 … 2D and 3D package data © 2011 67
  • 68. S-functions Verification in MWO - Frequency Domain HBT UNER2 ID=X1 Mag1=0 Ang1=0 Deg Mag2=0 Ang2=0 Deg Mag3=0 Ang3=0 Deg Fo=2 GHz Zo=50 Ohm PORT _PS1 P=1 1 2 Z=50 Ohm PStart=-20 dBm BIAST EE 2 PStop=8 dBm ID=X2 PORT PStep=1 dB 3:Bias P=2 SUBCK T 2 RF 1 1 Z=50 Ohm RF & ID=S1 DC NET ="EPA_120_B " DC 3 DCVS ID=V1 V=DC2val V DCVS ID=V2 V=DC 1val V Po with swept Pin Gain and Phase with swept Pin 40 20 19.5 20 19 F0 18.5 DB(|LSSnm(PORT_2,PORT_1,1,1)|)[1,X] 0 18 Swept Power.AP_HB Ang(LSSnm(PORT_2,PORT_1,2,1))[1,X] (Deg) 17.5 Swept Power.AP_HB -20 17 150 2F0 -40 100 DB(|Pcomp(PORT_2,1)|)[1,X] (dBm) Swept Power.AP_HB 50 3F0 DB(|Pcomp(PORT_2,2)|)[1,X] (dBm) -60 Swept Power.AP_HB DB(|Pcomp(PORT_2,3)|)[1,X] (dBm) Swept Power.AP_HB 0 -80 -50 -20 -15 -10 -5 0 5 10 -20 -15 -10 -5 0 5 10 Power (dBm) Power (dBm) © 2011 68
  • 69. S-functions Verification in MWO - Time Domain Vtime(V_METER.VM1,1)[*,T] (V) Itime(I_METER.AMP1,1)[*,T] (mA) HBTUNER2 Waveform Tests.AP_HB Waveform Tests.AP_HB ID=X1 Mag1=0 Ang1=0 Deg Mag2=0 Ang2=0 Deg Mag3=0 IV Ang3=0 Deg 250 Fo=2 GHz SUBCKT Zo=50 Ohm PORT_PS1 ID=S1 200 P=1 NET="EPA_120_B" 1 2 Z=50 Ohm PStart=-20 dBm BIASTEE 2 150 PStop=8 dBm ID=X2 PORT PStep=1 dB I_METER 3:Bias P=2 2 RF 1 1 ID=AMP1 Z=50 Ohm 100 RF & DC DC 50 3 V _METER DCVS ID =VM1 ID=V1 V=DC2val V 0 15 DCVS ID=V2 V=DC1val V 10 5 DB(|Pcomp(PORT_2,1)|)[1,X] (dBm) DB(|Pcomp(PORT_2,2)| )[1,X] (dBm) DB(|Pcomp(PORT_2,3)|)[1,X] (dBm) Waveform Tests.AP_HB Waveform Tests.AP_HB Waveform Tests.AP_HB0 0 0.2 0.4 0.6 0.8 1 Harmonics 40 Time (ns) 20 0 -20 -40 -60 -80 -20 -10 0 8 Power (dBm) © 2011 69
  • 70. S-functions in MWO – Modulation Prediction - Schematic DB(PWR_SPEC(TP.AMP Out,1000,0,10,0,-1,0,-1,1,0,4,0,1,0)) (dBm) DB(PWR_SPEC(TP.AMP In,1000,0,10,0,-1,0,-1,1,0,4,0,1,0)) (dBm) 16QAM System 16QAM System HBTUNER2 ID=DCBlock2 Mag1=0 Ang1=0 Deg Mag2=0 Ang2=0 Deg Mag3=0 ACPR SUBCKT Ang3=0 Deg Fo=2000 GHz 50 ID=S1 Zo=50 Ohm ISOL8R NET="EPA_120_B" ID=U1 1 2 R=50 Ohm LOSS=0 dB BIASTEE 2 ISOL=30 dB ID=X1 PORT 3:Bias P=2 2 RF RF & 1 1 Z=50 Ohm 0 DC DC PORT_PS1 3 DCVS P=1 ID=V3 Z=50 Ohm V=DC2val V PStart=-20 dBm PStop=8 dBm PStep=0.5 dB DCVS -50 ID=V1 V=DC1val V -100 -150 1.92 1.94 1.96 1.98 2 2.02 2.04 2.06 2.08 Frequency (GHz) VS A ID=M1 VA RNA ME="" VA LUE S=0 NA VG= PBA SE = -76 V e c to r S w e p t v a r ia b le S ig n a l SRC MEAS A n a ly z e r QAM_S RC G e n e r a l R e c e iv e r ID=A1 MO D=16-QA M (Gray) OUT LV L=-0.2 OLVLT Y P=Avg. Power (dB m) NL_ S AW GN RAT E=_DRA T E ID=S 1 BER ID=A3 CTRFRQ=2 GHz TP NET ="AMAM A MPM.AP_HB " PW R=-170 ID=5 SIMT YP=Harm onic Balanc e TP PLST Y P=Root Rais ed Cos ine ID=AMP In PW RT YP=Avg. Po wer, Sym bol (dBW ) VARNAME="PBA SE" ID=AMP Out ALPHA =0.35 NO SE=RF Budget only I VALUES=s tepped(-76,-40,2) LOSS=0 dB PLSLN= RFIF RQ= QAM_RX OUT FL="" ID=A 2 1 R D 2 BER IQ 3 C irc u it 5 4 B E R D e te c to r B a s e d A m p lifie r C h a n n e l N o is e 1 6 Q A M S o u rc e TP ID=RX Cons tellation 1 6 Q A M (G r a y ) S y s te m © 2011 70
  • 71. S-functions in MWO – Load-Pull HBTUNER2 ID=DCBlock2 Mag1=0.8 Load Tuner (and bias) Ang1=40 Deg LP_Data_Low_Power Mag2=0 Ang2=0 Deg 1.0 Swp Max Mag3=0 0.8 -0.89 Ang3=0 Deg Fo=2 GHz 6 0. Zo=50 Ohm 0 . 2 4 . 1 2 PORT1 0 0 . 3 P=1 2 p8 0 p1: Pcomp_PORT_2_1_M_DB = -2.87 Z=50 Ohm PORT p7 4. Pwr=10 dBm 3:Bias P=2 p9 p2: Pcomp_PORT_2_1_M_DB = -2.65 SUBCKT 5. 0 2 RF 1 Z=50 Ohm 2 p6 1 ID=S1 0. RF & p10 p3: Pcomp_PORT_2_1_M_DB = -2.43 DC NET="SFUNC_Model" p5 DC 10. 0 BIASTEE p4: Pcomp_PORT_2_1_M_DB = -2.21 3 DCVS ID=X1 p4 ID=V2 10.0 V=7 V 4.0 p5: Pcomp_PORT_2_1_M_DB = -1.99 0.2 0.4 0.6 0.8 1.0 2.0 3.0 5.0 0 p3 p6: Pcomp_PORT_2_1_M_DB = -1.77 DCVS p2 ID=V1 .0 -10 p7: Pcomp_PORT_2_1_M_DB = -1.55 V=-0.6 V p1 -0 . 2 0 . p8: Pcomp_PORT_2_1_M_DB = -1.33 -5 0 . -4 p9: Pcomp_PORT_2_1_M_DB = -1.11 p10: Pcomp_PORT_2_1_M_DB = -0.89 0 . 3 4 - . 0 - 0 . 2 6 - . -0 8 Swp Min -1.0 -0. -2.87 LP_Data_High_Power S M wp ax 1.0 0 .8 25.93 6 0. .0 2 .4 0 p1: Pcomp_PORT _2_1_M_DB = 21.93 .0 3 p9 p8 p7 p6 p5 p2: Pcomp_PORT _2_1_M_DB = 22.33 p4 4 .0 2 p10 p3 5. 0 p3: Pcomp_PORT _2_1_M_DB = 22.73 0. p2 0 p4: Pcomp_PORT _2_1_M_DB = 23.13 1 0. p11 p5: Pcomp_PORT _2_1_M_DB = 23.53 10.0 0.4 0.8 3.0 0.2 0.6 1.0 2.0 4.0 5.0 p1 p6: Pcomp_PORT _2_1_M_DB = 23.93 0 0 . p7: Pcomp_PORT _2_1_M_DB = 24.33 - 10 p8: Pcomp_PORT _2_1_M_DB = 24.73 .2 -0 .0 .0 -5 p9: Pcomp_PORT _2_1_M_DB = 25.13 -4 0 p10: Pcomp_PORT _2_1_M_DB = 25.53 3. 4 - 0. - p11: Pcomp_PORT _2_1_M_DB = 25.93 0 2. .6 - -0 8 S M wp in -1.0 - 0. 21.93 © 2011 71
  • 72. S-functions Strengths ● S-functions are completely public ● S-functions are transparent ● S-functions are connected into ADS and MWO and can be connected in other tools on request, e.g. Matlab ● S-functions are closely integrated with load-pull in general and especially with tuners from Focus Microwave ● S-functions extraction tool has a verification capability for LSOP constantness and interpolation capability, investigating quality of model, without needing a simulation tool ● The source- and load-pull software ICEBreaker from NMDG allows S- function verification, related to the linearity assumption ● Customers can explore the capabilities of S-functions via NMDG services © 2011 72
  • 73. S-Functions - Key Capabilities ● Natural extension of S-parameters ● Reduce to S-parameters for small-signal excitation ● S-parameters are cascadeable, S-Functions are cascadeable too ● e.g. Transistors, amplifiers, dividers, multipliers ● Predict harmonic behaviour of components under different impedances ● Source – Pull ● Load – Pull ● Harmonic distortion, Waveforms ● Predict modulation behaviour of components under different impedances ● When no long-term memory effects ● Valid for multi-ports, applicable to differential components © 2011 73
  • 74. References ● F. Verbeyst and M. Vanden Bossche, “VIOMAP, the S-parameter equivalent for weakly nonlinear RF and microwave devices”, published in the Microwave Symposium Digest of IEEE 1994 MTT-S International and published in the 1994 Special Symposium Issue of the MTT Transactions, vol. 42, no. 12, pp. 2531 – 2535. ● F. Verbeyst and M. Vanden Bossche, “VIOMAP, 16QAM and Spectral Regrowth: Enhanced Prediction and Predistortion based on Two-Tone Black-Box Model Extraction”, published in the Proceedings of the 45th ARFTG Conference, Orlando, June 1995 and winner of the “Best Conference Paper Award”. ● J. Verspecht and P. Van Esch, “Accurately characterizating of hard nonlinear behaviour of microwave components by the Nonlinear Network Measurement System: introducing the nonlinear scattering function,” Proc. International Workshop on Integrated Nonlinear Microwave and Millimiterwave Circuits (INMMiC), October 1998, pp.17-26. ● J. Verspecht, “Scattering functions for nonlinear behavioral modeling in the frequency domain,” IEEE MTT-S Int. Microwave Symp. Workshop, June 2003. ● J. Verspecht and D.E. Root “Polyharmonic Distortion Modeling,” IEEE Microwave Magazine, vol.7 no.3, June 2006, pp.44-57. ● D.E. Root, J. Horn, L. Betts, C. Gillease, and J. Verspecht, ”X-Parameters: The new paradigm for measurement, modeling, and design of nonlinear RF and microwave components,” Microwave Engineering Europe, December 2008, pp. 16-21. © 2011 74
  • 75. Conclusion ● S-functions are a natural extension to S-parameters for nonlinear behaviour ● S-functions aren't much more complex than S-parameters ● S-functions are accurate when assumptions are not violated ● Linearity assumption ● S-function extraction is supported on R&S network analysers ● S-functions can be coupled into ADS and MWO ● S-functions can be used to compress or hide specifics of nonlinear circuits in ADS and MWO For more information info@nmdg.be www.nmdg.be © 2011 75
  • 76. Explore the power of S-functions Send your device or circuit to NMDG ... ` … and NMDG sends you the S-functions Please contact NMDG at info@nmdg.be © 2011 76