Professional Documents
Culture Documents
Dr. Larry Dunleavy President & CEO Modelithics Inc. Tampa, FL Professor Department of Electrical Engineering University of South Florida Tampa, FL
1
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Acknowledgments
I would like to acknowledge various contributions and collaborations :
Rick Connick, Byoungyong Lee, Dr. Jiang Liu, Modelithics, Inc. Bill Clausen, formerly with Modelithics (now with RFMD) Dr. W.R. Curtice, W.R. Curtice Consulting Dr. David Snider, Univ. South Florida Ray Pengelly and Simon Wood, Cree, Inc. Dr. Steve Maas, Non-linear Technologies, Inc. Dr. Peter Aaen, Freescale Dr. Yusuke Tajima, Auriga Measurement Systems
2
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Overview
Nonlinear Modeling Thermal and Trap Issues MESFET and PHEMT Modeling MOSFET Modeling HBT Modeling Behavioral Modeling of Amplifiers References
3
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Accurate models can predict precisely the performances of RF circuit designs yet challenges remain! PA Design has become more complex in terms of competing multidimensional requirements of BW, efficiency, linearity and power performance.
Electro-thermal effects often a critical issue for accurate HPA modeling Requirement of Isothermal measurements
Self-heating effects held constant Some applications (GSM, radar) required pulsed operation.
Advance model testing
Wireless systems use various digital modulation signals. Are currently available models adequate for emerging requirements?
4
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Nonlinear Modeling
Device behavior is different under largesignal conditions than for small-signal conditions.
5
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Source of PA Nonlinearities
Example BJT Device
6
PRECISION MEASUREMENTS AND MODELS YOU TRUST
7
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Appropriate Model
(Angelov , EEHEMT CFET, etc.)
Parameter Extraction
(IC-CAP, ADS, etc.)
Model Validation
Optimization / Tuning
Ic (A)
Vce (V)
Advance testing
Acceptable Model
Load pull Pulse RF measurements Time domain Digital modulation
8
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Qg
Ig
Id
Source
Qd
Accurate simulation and measurements are required. Shell representation of packaged entire transistor.
Used with permission from Peter Aaen of Freescale June 15, 2008
Freescale and the Freescale logo are trademarks of Freescale Semiconductor, Inc. All other product or service names are the property of their respective owners. Freescale Semiconductor, Inc. 2006.
TM
MOS capacitors
500 mil
Flange
This packaged transistor operates at 2.1 GHz and is capable of producing 170 W (CW) output power.
Used with permission from Peter Aaen of Freescale
TM
GPIB connection Keithley 4200 DC Parameter Analyzer sense Anristu 37369 C VNA Bias force sense
Agilent ICCAP
Bias force
Bias tee
Bias tee
11
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Extraction of ID Equation
IC-CAP Setup Measured (Solid Lines) and Simulated (Dashed Lines) IV Data:
The quality of the IV extraction plays a large part in determining the path of the large-signal swing in the IV plane as well as gain and output conductance.
12
PRECISION MEASUREMENTS AND MODELS YOU TRUST
80 W MESFET
Legend Model 18 16 14 12 Measured
Ids (A)
10 8 6 4 2 0 0 1 2 3 4 5 6
Vds (V)
13
PRECISION MEASUREMENTS AND MODELS YOU TRUST
14
PRECISION MEASUREMENTS AND MODELS YOU TRUST
15
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Tuner controller
( MT 986A)
RF source
( Agilent E4438C)
Bias tee
Tuner
DUT
(InGaP HBT)
Tuner
Bias tee
Power meter
(Anritus ML 2438A)
Note: Can also be performed under pulsed RF conditions with minor modifications to setup.
16
PRECISION MEASUREMENTS AND MODELS YOU TRUST
80 W MESFET
1850 MHz loadpull results (Source = 13.5-j25.0 Ohms) Load Pout Device 1 Device 2 Model
S ys tem Re ference Impeda nce
6.6-j18 5.23-j18.04
6.86-j17.91 4.78-j16.86
P AE (thick) and Delivere d P ower (thin) Contours Maximum P owe r Delivere d, dBm
p p _ _ s r r s u u o t t o n n o o c c _ _ l eE dA PP
10.5 10 45
46.04
Maximum P ower-Added Effic ie ncy, % m1 m2
58.35
inde p(P AE_c ontours _p) (0.000 to 31.000) inde p(P de l_c ontours _p) (0.000 to 56.000)
17
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Measurement
100 80 60 40 20 0 P A E ( % )
Pin = 23dBm, Freq = 1200 MHz, Vds = 28 V, Vgs = 1.65 V. S= 0.83 < 98 . In measurement pulse width = 200us, pulse separation = 2ms.
18
PRECISION MEASUREMENTS AND MODELS YOU TRUST
What are the main considerations for nonlinear Non-linear transistor models?
Transistor model parameters
Active components
Resistances
Pulsed IV Measurement
Measurements are performed during brief (~0.2 s) excursions from a quiescent bias. The pulses are usually separated by at least 1 ms. Thermal and trap conditions during the measurement are those of the quiescent bias, as in high-frequency operation.
20
PRECISION MEASUREMENTS AND MODELS YOU TRUST
AU4550
Pulser
DUT
Pulser
For a demo, visit the room at the Embassy From please Yusuke Tajima, usedhospitality with permission. Suites, across the street, from Tuesday to Thursday.
June 16, 2008 2008 IMS Workshop
21
PRECISION MEASUREMENTS AND MODELS YOU TRUST
23
24
Electrothermal Circuit
TC = Z th PD + T A
+ Pd Rth Cth
Tc
1 Z th = Rth // j C th
Ta -
Cth =
th
Rth
23
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Trapping Effects
Trapping Effects in MESFETs*:
Substrate Traps Surface Traps
Electron Flow
Substrate Traps
*C. Charbonninud, S. DeMeyer, R. Quere, J. Teyssier, 2003 Gallium Arsenide Applications Symposium, October 6-10, 2003, Munich.
24
PRECISION MEASUREMENTS AND MODELS YOU TRUST
JFET [1] Curtice3 [2] CFET [3] EE HEMT1 [4] Angelov [5] CMC (Curtice/Modelithics/Cree) [6] MET(Motorola Electro-Thermal) [7] MOS Level 1/2/3 [1] BSIM3 (v3.24) [8] BSIMSOI3 [9]
No/No Yes/No Yes/Yes Yes/No Yes/Yes Yes/Yes Yes/Yes Yes/Yes Yes/Yes Yes/Yes
25
PRECISION MEASUREMENTS AND MODELS YOU TRUST
26
PRECISION MEASUREMENTS AND MODELS YOU TRUST
28
PRECISION MEASUREMENTS AND MODELS YOU TRUST
29
PRECISION MEASUREMENTS AND MODELS YOU TRUST
MOSFET Modeling
Motorolas Electro-Thermal (MET) Model Curtice-Modelithics-Cree (CMC) Model Both of these models possess traditional electrothermal subcircuits. Used for Si LDMOSFET, VDMOSFET devices No traps Electrothermal subcircuit and temperature dependence extraction are much simpler!
30
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Curtice-Modelithics-Cree Topology
The CurticeModelithics-Cree (CMC) model is a proprietary electrothermal LDMOS model Four region (4R) current model based on work of Fager et.al. (see IEEE Trans. MTT, Dec. 2002) The model provides accurate predictions of power, efficiency and distortion performance over a wide range of devices sizes.
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Gate
Cdg
Rd Cdd
Drain
Rdd Rs
Thermal Circuit
See W. Curtice, L. Dunleavy, W. Clausen, and R. Pengelly, ,High Frequency Electronics Magazine, pp18-25, Oct.. 2004. The CMC model is copyright Cree, Inc. 2004-2008 all rights reserved.
31
2.5
6
2.0
5
Gm (S )
1.5
Ids (A)
1.0
0.5
1
0.0 0 1 2 3 4 5 6 7 8
0 0 1 2 3 4 5 6 7 8
V G (V )
V g (V )
Sub-threshold I
5 4 3 Ids (A)
2 1 0
32
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Ids (A)
0.00
0.00 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0
Vds (V)
Pout (dBm)
Gain (dB)
11
43 42
41 40
7 39 5 24 26 28 30 32 34 36 38 38
Pin (dBm)
The CMC model is copyright Cree, Inc. 2004-2008 all rights reserved.
34
PRECISION MEASUREMENTS AND MODELS YOU TRUST
45
Gain (dB)
11 40 9 35 7 30 25 15 20 25 30 35 40 45
5
S 21 S 22 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0
Model
-20
The CMC model is copyright Cree, Inc. 2004-2008 all rights reserved.
35
PRECISION MEASUREMENTS AND MODELS YOU TRUST
37
39
BJT models GP (1970) VBIC (1985) Mextram (1987) HICUM (1995) Agilent (2003) FBH (2005) Curtice (2004)
substrate effect / self heating No/No Yes/Yes Yes/Yes Yes/Yes Yes/Yes No/Yes No/Yes
Original Device Context Si BJT SiGe BJT SiGe HBT GaAs HBT InP/GaAs HBT GaAs HBT GaAs HBT
36
PRECISION MEASUREMENTS AND MODELS YOU TRUST
15 10 5 0 -5 -1 0 P o u t ( d B m ) ) A ( c I
0.020
0.015
0.010
0.005
RFpo we r
5.5 GHz power sweep results at vc= 3V and ib= 100uA. Source reflection coefficient Gms= .06522<148.97 (mag<deg); L Load reflection coefficient Gml= .07354<36.24 (mag<deg).
) m B -20 d ( t u -40 o P
-60 -80
-22
-20
-18
-16
-14
-12
-10
-8
-6
-4
P in (dBm)
37
PRECISION MEASUREMENTS AND MODELS YOU TRUST
- Constant Base Current vs. Constant Base Voltage (See B. Lee, L. Dunleavy ,, High Frequency Electronics, May 2007.)
-o- line: Mextram 504 model and solid line: measurements
20 10 5
0.020 0.018
15 0
0.016 0.014 0.012 0.010 0.008 0.006 -34 -32 -30 -28 -26 -24 -22 -20 -18 -16 -14 -12 -10
ga in ga in_db
P out p1_db
10
-5 -10
5 -15 0 -34 -32 -30 -28 -26 -24 -22 -20 -18 -16 -14 -12 -10 -20
0.010
5 15
0.008
ga in ga in_db
P out p1_db
10
-5 -10
0.006
0.004
Behavioral Models
Empirical models (behavioral models, black-box models) Requires no knowledge about the internals of the PA Based on the observation of the input-output signal relationships Its simulation performance heavily depends on the dataset used for the extraction of the model It fits well to the given datasets and requires small simulation time; However it may suffer when trying to extrapolate the PA performance or fit to different datasets (by that means different PA topologies)
39
PRECISION MEASUREMENTS AND MODELS YOU TRUST
PA Modeling Techniques
Circuit Level Models (Physical Models) Based on the knowledge of the amplifiers circuit structure Require accurate active device models and other components The simulation results can be accurate, however, timeconsuming
40
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Am plifie rS 2D AMP 1 S 2DFile ="s2dfile .s 2d" S S fre q=a uto Inte rpMode =Line a r Inte rpDom =Da ta Ba s e d
41
PRECISION MEASUREMENTS AND MODELS YOU TRUST
42
PRECISION MEASUREMENTS AND MODELS YOU TRUST
43
PRECISION MEASUREMENTS AND MODELS YOU TRUST
44
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Simulated output spectrum shows the correlation between the spectral regrowth and the PA performance at different frequencies.
45
PRECISION MEASUREMENTS AND MODELS YOU TRUST
46
PRECISION MEASUREMENTS AND MODELS YOU TRUST
-10C
25C 60C
47
PRECISION MEASUREMENTS AND MODELS YOU TRUST
A11 A12
A1N
A21 A22
A2N
Port 1
DUT
Port 2
B11 B12
B1N
48
B21 B22
B2N
D.E. Root, J. Verspecht, D. Sharrit, J. Wood, A. Cognata, Broad-band poly-harmonic distortion (PHD) behavioral models from fast automated simulations and large-sinagl vectorial network measurements, IEEE Trans. Microw. Theory Tech., vol. 53, no. 11, pp. 36563664, Nov. 2005.
49
PRECISION MEASUREMENTS AND MODELS YOU TRUST
* 2
T22 = C5 + jC6
12/16/2005
PRECISION MEASUREMENTS AND MODELS YOU TRUST
J. Liu, L.P. Dunleavy and H. Arslan, Large Signal Behavioral Modeling of Nonlinear Amplifiers Based on Loadpull AM-AM and AM-PM Measurements, IEEE Trans. Microw. Theory Tech., vol. 54, no. 8, pp. 31913196, Aug. 2006.
51
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Gain (dB)
4 3
2 30
Frequency: 900 MHz Pin: -30 dBm to 5 dBm Bias: Vagc, 1.3875 V; Vcc, 2.775 V 25 20 15 10 5 0
Pin (dBm)
11/07/2005
PRECISION MEASUREMENTS AND MODELS YOU TRUST
10 Pout (dB) 0 10 20 25 Meas. LargeS21 Model New Beh. Model 20 15 10 Pin (dBm) 5 0 5
b d a
0 IM3 (dBm) 20 40 60 80 25 20 15 10 Pin (dBm) Meas. LargeS21 Model New Beh. Model 5 0 5
Note: the Large S21 model neglects the last conjugate term.
11/07/2005
53
Summary
Non-linear device measurement/modeling requires Careful attention to measurement setup/accuracy
Pulsed multi-temperature testing High current/high power instrumentation and components Advanced non-linear instrumentation (e.g. load-pull) Large signal modeling requires Advanced models (templates) and extraction techniques. Focused expertise that can pull together the varied aspects of IV, S-parameter and non-linear test results into an effective modeling extraction and validation.
57
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Summary (contd)
A Good Behavorial Model
Needs be created based on measurement datasets through instruments available to the modelers. Good News! More advanced non-linear test instruments/software are becoming available. Model should be easy to use and no more complex than necessary. Powerful enough to present multiple dimensional datasets for designers to inspect the amplifiers performance in a system view (Ideally) Model should be supported in popular CAE software packages.
58
PRECISION MEASUREMENTS AND MODELS YOU TRUST
Useful References
P. Ladbrooke and J. Bridge, The Importance of the Current-Voltage Characteristics of FETs, HEMTs, and Bipolar Transistors in Contemporary Circuit Design, Microwave Journal, March 2002. P. Winson, An Investigation of Linear and Nonlinear Modeling of MESFET Characteristics as a Function of Temperature Doctoral Dissertation: University of South Florida, Tampa, Florida:, 1997. K. Jenkins, and K. Rim, Measurement of the Effect of Self-Heating in Strained-Silicon MOSFETs, IEEE Electron Device Letters, Vol. 23, No. 6, June 2002, pp. 360-362. Accent DIVA Models D210, D225, D225HBT, D265 Dynamic i(V) Analyzer, User Manual, Issue 1.0, (P/N 9DIVA-UM01), 2001. Accent Optical Technologies, 131 NW Hawthorne, Bend, OR 97701. C.P. Baylis II, L.P. Dunleavy, Understanding Pulsed IV Measurement Waveforms, 11th IEEE Intl Symposium on Electron Devices for Microwave and Optoelectronic Applications, Nov. 17-18, 2003, Orlando FL. L. Dunleavy, W. Clausen, T. Weller, Pulsed I-V For Nonlinear Modeling, Microwave Journal, March 2003. C. Baylis, L. Dunleavy, and J. Daniel, Thermal Correction of IV Curves for Nonlinear Transistor Modeling IEEE Wireless and Microwave Technology Conference 2004, April 15-16, Clearwater, FL . C. Baylis, L. Dunleavy, and J. Martens, Constructing and Benchmarking a Pulsed-RF, PulsedBias S-Parameter System, Automatic RF Techniques Group Conference, December 2005, Washington, D.C.
59
PRECISION MEASUREMENTS AND MODELS YOU TRUST
60
PRECISION MEASUREMENTS AND MODELS YOU TRUST
61
PRECISION MEASUREMENTS AND MODELS YOU TRUST
62
PRECISION MEASUREMENTS AND MODELS YOU TRUST
63
PRECISION MEASUREMENTS AND MODELS YOU TRUST
64
PRECISION MEASUREMENTS AND MODELS YOU TRUST