Large-signal Device Model Research Articles

A 5.9-GHz Fully Integrated GaN Frontend Design With Physics-Based RF Compact Model

This paper presents the design of a fully integrated high-efficiency and high-power RF frontend for the IEEE 802.11p standard in GaN HEMT technology. An embedded transmitter/receiver (Tx/Rx) switching scheme and a dual-bias power amplifier linearization technique are used to improve Tx efficiency and linearity. An accurate physics-based nonlinear large-signal device model is developed and used for the design, providing insight into the impact of the behavioral nuances of the GaN HEMTs on RF circuit performance. The fully integrated RF frontend is fabricated in 0.25- $\mu{\hbox{m}}$ GaN-on-SiC technology and occupies only 2 mm $\times$ 1.2 mm. The Tx branch achieves 48.5% drain efficiency at 33.9 dBm, $P{\rm sat}$ with 28-V supply. With orthogonal frequency-division multiplexing modulated signals, it achieves 30% average efficiency at 27.8-dBm output power while meeting the $-$ 25-dB error vector magnitude limit without predistortion. The Rx branch achieves $+$ 22-dBm output third-order intercept point with 3.7-dB noise figure at 12-V supply. The fully integrated high-efficiency and linear RF frontend designed with physics-based RF GaN compact models is demonstrated for the first time for future device-to-device applications.

A $W$-Band On-Wafer Active Load–Pull System Based on Down-Conversion Techniques

A new W-band active load-pull system is presented. It is the first load-pull system to implement a 94 GHz load by means of an active loop exploiting frequency conversion techniques. The active loop configuration demonstrates a number of advantages that overcome the typical limitations of W-band passive tuners or conventional active open-loop techniques in a cost-effective way: load reflection coefficients Γ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">L</sub> as high as 0.95 in magnitude can be achieved at 94 GHz, thus providing a nearly full coverage of the Smith chart. Possible applications of the setup include technology assessment, large-signal device model verification at sub-terahertz frequencies, and W-band monolithic microwave integrated circuit design and characterization. The availability of direct and accurate load-pull measurements at W-band should prove an asset in the development of sub-terahertz integrated circuits. First measurements performed on high-performance InP double heterojunction bipolar transistors and GaN high electron-mobility transistors are presented.

An improved nonlinear model of HEMTs with independent transconductance tail-off fitting

We present an improved large-signal device model of GaAs/GaN HEMTs, amenable for use in commercial nonlinear simulators. The proposed model includes a new exponential function to independently control the transconductance compression/tail-off behaviors. The main advantage of this model is to provide a simple and coherent description of the bias-dependent drain current (I—V) that is valid in all regions of operation. All aspects of the model are validated for 0.25-μm gate-length GaAs and GaN HEMT processes. The simulation results of DC/pulsed I—V, RF large-signal power and intermodulation distortion products show excellent agreement with the measured data.

Comprehensive First-Pass Design Methodology for High Efficiency Mode Power Amplifier [TC Contests

In this paper, such a first-pass design methodology is presented, taking advantage of the arrival of highly accurate large signal device models. Using the method described, an inverse Class F power amplifier (PA) with minimum tuning is realized, measuring 71% PAE at 3.27 GHz at the 2009 IMS Student Design Competition.

Temperature-dependent nonlinearities in GaN/AlGaN HEMTs

Temperature-dependent nonlinearities of GaN/AlGaN HEMTs are reported. The large-signal device model of the transistor is obtained by using a physics-based analysis. The model parameters are obtained as functions of bias voltages and temperature. The analysis of the device has been carried out using a time-domain technique. f/sub max/ for a 0.23 /spl mu/m/spl times/100 /spl mu/m Al/sub 0.13/Ga/sub 0.87/N/GaN FET is calculated as 69 GHz at 300 K, while at 500 K, f/sub max/ decreases to 30 GHz, which are in agreement with the experimental data within 7% error. f/sub max/ as obtained from calculated unilateral gain, decreases monotonically with increasing temperature. For shorter gate lengths irrespective of the operating temperature f/sub max/ is less sensitive to bias voltage scaling. For longer gate length devices, f/sub max/ becomes less sensitive to the bias voltage scaling at elevated temperatures. 1-dB compression point (P/sub 1-dB/) at 4 GHz for a 1 /spl mu/m/spl times/500 /spl mu/m Al/sub 0.15/Ga/sub 0.85/N/GaN FET is 13 dBm at 300 K. At 500 K, P/sub 1-dB/ decreases to 2.5 dBm for the same operating frequency. Similar results for output referred third intercept point (OIP3) are reported for different gate length devices.

Large‐signal modelling and measuring go hand‐in‐hand: Accurate alternatives to indirect S‐parameter methods (invited paper)

Classical large-signal device models are indirectly derived from small-signal S-parameter measurements. Due to the availability of the nonlinear network measurement system (NNMS), models can be based directly upon full two-port large-signal measurements, resulting in higher model accuracy. In this article, we discuss three large-signal measurement-based modelling approaches. ©2000 John Wiley & Sons, Inc. Int J RF and Microwave CAE 10: 6–18, 2000.

Large-signal modelling and measuring go hand-in-hand: Accurate alternatives to indirect S-parameter methods (invited paper)

Classical large-signal device models are indirectly derived from small-signal S-parameter measurements. Due to the availability of the nonlinear network measurement system (NNMS), models can be based directly upon full two-port large-signal measurements, resulting in higher model accuracy. In this article, we discuss three large-signal measurement-based modelling approaches. ©2000 John Wiley & Sons, Inc. Int J RF and Microwave CAE 10: 6–18, 2000.

Method to determine forward gate capacitance of HFETs using resonance effect

A new method to determine gate capacitance under forward bias is presented. The method uses the frequency domain resonance of the real part of the Z11 parameter at microwave frequencies. This method extends the range of conventional CV profiling to higher forward gate bias regions. The result of this method helps to resolve the ambiguity of forward bias gate capacitance in large signal device models.

Millimeter wave QWITT diode oscillator

The quantum well injection transit time (QWITT) diode charecteristics, small signal and large signal device models are used. A peak output power density of 35–5 KW/Cm2 in the frequency range of 5–8 GHZ hasx been obtained from a planar QWITT Oscillator

Microwave and millimeter-wave QWITT diode oscillators

The authors present DC, microwave, and millimeter-wave characteristics of different quantum-well-injection transit-time (QWITT) devices. Small-signal and large-signal device models are used to provide physical design parameters to maximize the output power density at any desired frequency of operation. A peak output power density of 3.5-5 kW/cm/sup 2/ in the frequency range 5-8 GHz has been obtained from a planar QWITT oscillator. This appears to be the highest output power density obtained from any quantum-well oscillator at any frequency. This result also represents the first planar circuit implementation of a quantum-well oscillator. Good qualitative agreement between DC and RF characteristics of QWITT devices and theoretical predictions based on small-signal and large-signal analyses is achieved. The device efficiency has been increased from 3% to 5% by optimizing the design of the drift region in the device through the use of a doping spike with optimized concentration, without compromising the output power at X-band. Self-oscillating QWITT diode mixers are also demonstrated at X-band in both waveguide and planar circuits. The self-oscillating mixer exhibits a conversion gain of about 10 dB in a narrow bandwidth and a conversion loss of about 5 dB if broadband operation is desired.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Large-signal FET simulation using time domain and harmonic balance methods

Lumped-element circuit models, modified to include RF voltage-dependent elements may be used to describe the large-signal performance of a MESFET. Several solution methods exist for the analysis of amplifier circuits incorporating the large-signal device model. The paper reports the results of a comparative study of three solution methods: harmonic balance, modified SPICE and time domain methods. A simple amplifier, with 50 Ω generator and load impedances is used to show the importance of the nonlinear gate-source capacitance and avalanche breakdown in predicting harmonic generation in the MESFET. An amplifier, with cascaded transmission-line input and output matching sections, illustrates the effect on the large-signal analysis, of using the three alternative solution methods with their associated descriptions of the distributed external microwave circuits. It is concluded that, for this application, the harmonic balance method is faster and more accurate than the other solution methods.