Results of pulsed IV, CV and loadpull measurements performed on a GaN HEMT are presented. A model is developed for this device in the framework of the ASM-HEMT model. The developed model utilizes the advantage of measuring beforehand several parameters of the model; hence, reducing the number of parameters to be optimized for a reasonable fit to the measured data. A new approach for modelling the OFF-state capacitances of GaN HEMT is also presented which accounts for the details of the structure of the device, while at the same time maintaining the speed of the model.
We report a fully integrated 272 GHz direct-conversion transmitter in 250nm InP HBT technology. The transmitter has more than 18dB conversion gain over 264-285GHz, consumes 1. 6W, and has 14. ldBm output power at 272 GHz. Its -3dB bandwidth is 18 GHz, while its -6dB bandwidth is 36 GHz. An internal 8:1 active frequency multiplier generates the local oscillator. To the authors’ knowledge, the IC demonstrates record saturated output power for an integrated transmitter operating near 272 GHz.
In this paper, we designed a CMOS active phase shifter that operates in the Ka-band. In the active phase shifter, the differential signals with high phase accuracy were generated using active balun to alleviate complexity, thereby reducing the insertion loss of I/Q generator. In addition, we proposed a layout technique that can secure symmetry between the output signals of poly phase filter (PPF). To verify the feasibility of the proposed design technique, we designed an Ka-band active phase shifter using 65 nm RFCMOS process. The core size of the designed phase shifter was 0.506× 0.257m$\mathrm{m}^{2}$. At the target frequency of 28 GHz, the measured RMS phase and gain errors were lower than 0.35° and 0.07 dB, respectively.
Next generation communication standardization is calling for more flexibility with an access to higher carrier frequencies and wider instantaneous bandwidth. We propose to tackle these challenges by introducing wideband data converter solutions capable to directly access the microwave spectrum up to Q-band $(40 \mathrm{GHz})$. After a proof of concept presented at EuMW 2021, a product has been developed based on a $2^{\text {nd }}$ generation $T / H$ die, and a specially developed ADC die in $28 \mathrm{~nm}$ process, thus allowing for a dramatic power reduction, cutting down the overall power to $2.5 \mathrm{~W}$. These two dice have been assembled in a dedicated FCBGA package to offer optimal performances with a tight power budget in a reduced footprint. With a $B_{-3 d B}$ of 32GHz at $\mathrm{T} \mathbf{j}=100^{\circ} \mathrm{C}$, the measurement frequency has been extended up to the beginning of the Q-band $(40 \mathrm{GHz})$. Performance of SFDR, $\mathrm{SNR}, \mathrm{THD}$ and ENOB are shown in X-, Ku-, and Ka-band, at a sample rate of 12.8GSps.
This paper describes the design and implementation of a 113.5-127 GHz D-band LNA using GlobalFoundries’ 22 nm FD-SOI technology. The proposed design achieves 6.6 dB of minimal noise-figure in-band for a peak gain of 17.5 dB at its nominal bias for a 27.5 mW power-consumption using a low supply voltage of 0.75 V, and can achieve an NF down to 6.1 dB and gain up to 19.1 dB in a high-performance state. On top of that, the LNA can be configured as a variable-gain device by making use of the unique back-gate bias node of the 22FDXO technology. In that case, a 9. 7dB gain-control is achieved over a 2 V bias range at the back-gates. The design achieves a low area of 0.039 mm 2 thanks to its layout based on compact transformers.
Future 5G telecommunication system will require advanced circuit components operating above 100GHz. A critical component is the Low-Noise Amplifier (LNA) that should concurrently feature low noise figure, high gain, and high linearity. In this context, a demonstrator LNA is developed featuring typical $3.7 \mathrm{~dB}$ noise figure, $22 \mathrm{~dB}$ gain, +5 dBm output power at $1 \mathrm{~dB}$ compression in the 90 to 115GHz bandwidth using an industrial grade Gallium Arsenide 100-nm technology. Gain control functionality is also applied to allow operation in heavily congested EM scenarios. The concurrent noise-linearity performance is one of the best reported in the open literature above 100GHz.
In this work, normally-off, aluminium nitride (AlN)/ gallium nitride (GaN) high electron mobility transistors (HEMTs) were successfully fabricated with T-gate structures. GaN HEMTs with A1N barriers and 70 nm T-gates exhibited +0.62V of threshold voltage ($\mathrm{V}_{\mathrm{t}\mathrm{h}}$), 1.15mV/V of drain induced barrier lowering (DIBL), and $\mathrm{f}_{\mathrm{T}}$/f max of 89/232 GHz. This is a fundamental step towards realising W-band amplifiers for wireless communications applications.
This paper presents a clever solution for a SiGe MMIC that is capable of supplying a selectable output signal from two common frequency bands. The Ka-Band signal is directly generated by a fundamental LC-VCO while the Ku-Band signal is generated on-chip with a frequency division by 2 of the Ka-Band signal. The selectability between both bands is realized by using wideband antiparallel SPDT RF switches with a series-shunt design. For further selectability two wideband output buffers are integrated on the MMIC. The total power consumption of the MMIC is 212 mW from a 3.3V supply. The achieved output levels are higher than -10 dBm over a relative tuning range of 32%. The phase noise is as low as -107.2dBc/Hz and -101.9 dBc/Hz for the Ku- and Ka-Band, respectively.
In this paper, a 100-GBd linear modulator driver is presented for short-reach intensity modulation and direct detection (IM/DD) links with low-drive-voltage Mach-Zehnder modulators (MZMs) or elctro-absorption modulators (EAMs). The fabricated driver features differential gain of 19.3 dB with >67-GHz bandwidth. To improve the quality of a 100-GBd eye diagram, a distributed load network (DLN) is demonstrated. Eye diagrams are measured, showing 100-GBd NRZ signal is supported. For 75-GBd PAM4 (150-Gb/s) signal, differential 2.6 V pp with a ratio of level mismatch (RLM) of 91% is achieved. The driver has a small footprint of 0.26 mm 2 with 336-mW DC power consumption.
In this paper, the experimental verification of suitable input harmonic injection to achieve class F operating conditions in class $\mathbf{C}$ (under pinch-off) biased active devices is discussed. In detail, it is demonstrated that by generating a third harmonic component $V_{g, 3}$ with a proper phase relationship with respect to the fundamental one $V_{g, 1}$, it is possible to change the phase of the resulting output harmonic components, and in particular of the third one $I_{D, 3}$, thus generating the proper squaring of the output voltage. Experimental results at device level have shown an improvement in terms of output power and efficiency from 2.24W to 3.00W and from 72.2% to 79.3%, respectively. The proposed approach could play a key role in the maximization of the achievable efficiency of Doherty power amplifiers, where the auxiliary branch needs to be biased in class C for a exploitation of the architecture.
Aggressive duty-cycling is an approach to reduce the DC power demand of wireless networks. Turning off the power hungry blocks as often and as long as possible requires them to have fast operating point switching (OPS) times with a low energy consumption. This paper presents a V-band SiGe power amplifier (PA) with ultra-fast transition times. An approach of deriving the turn-on/off times is proposed and a method to measure the energy consumption during the OPS is presented. The single-stage inverse class-E PA achieves a power-added-efficiency (PAE) of 24.8% and an output power of 8.5 dBm at 54 GHz and 2 dBm available source power. It reaches these continuous-wave characteristics within 7.3 ns and a single OPS cycle consumes about 100 pJ.
This paper presents a fully integrated true time delay (TTD) circuit for beamforming applications based on the delay sum principle operating at a center frequency of 144 GHz with a 3 dB bandwidth of 26 GHz. It features a measured delay range of 1.75 ps whose resolution is only limited by the integrated digital-to-analog converter (DAC). A demonstrator has been manufactured in a 130 nm bipolar complementary metal-oxide-semiconductor (BiCMOS) technology, occupying 0.53 × 0.3mm 2 of chip area including serial peripheral interface (SPI), bandgap reference and four DACs.
This paper presents a broadband balanced power amplifier (PA) implemented in a 28 nm CMOS FD-SOI process with a compact size of 0.82 m$\mathrm{m}^{2}$ for mm-wave 5G application. Through the use of a dual-resonant transformer, the power amplifier operates over a relative bandwidth of 62.5% between 22 and 42 GHz. Thanks to its balanced architecture, the PA shows a great robustness to active voltage standing wave ratio (VSWR). The PA exhibits 20.7 dBm$\mathrm{P}_{\mathrm{s}\mathrm{a}\mathrm{t}}$ with 34.2% peak PAE and 16% PA$\mathrm{E}_{6\mathrm{d}\mathrm{B}}$PBO at 28 GHz. In the 22-42 GHz frequency band, the PAE max and $\mathrm{P}_{\mathrm{s}\mathrm{a}\mathrm{t}}$ of the PA remain above 20% and 18 dBm, respectively. Over the entire frequency band, the peak PAE and $\mathrm{P}_{\mathrm{s}\mathrm{a}\mathrm{t}}$ drop by only 12% and 2.0 dB between the optimal case 50 $\Omega$ and the worst cases at 3:1 VSWR. Using a 200 MHz 64-QAM 3GPP NR FR2 signal at 28 GHz, the proposed PA achieves 12.2 dBm average output power and 11.8% average PAE. For a 0.72 Gb/s(120MHz) single carrier (SC) 64-QAM signal at 28 GHz, the PA achieves a 15.2 dBm$\mathrm{P}_{\mathrm{a}\mathrm{v}\mathrm{g}}$ and 18% PA$\mathrm{E}_{\mathrm{a}\mathrm{v}\mathrm{g}}$.
This paper presents a highly integrated, digitally reconfigurable K-band transceiver suitable for cancer detection in a large diversity of breast tissues. First, we present a thorough system level analysis on how the statistical spread of breast density and water content affects radar system parameters. Then, we discuss how an FMCW radar with a reconfigurable analog baseband (ABB) can be used to ensure that the hardware can cover a large variety of breast tissues. Based on these considerations, we analyze some key specifications. Next, we present a highly integrated FMCW transceiver operating from 18–26 GHz that can provide sufficient gain and corner frequency variability to support tumor detection over various types of tissues. The chip integrates a TX, an I/Q RX with programmable ABB and an 8-bit SAR ADC with an SPI interface, which is used for configuration and data readout. The gain of the receiver can be varied in 16 steps between 15-75dB, while the transmitter achieves a maximum output power of 12.5 dBm. Measured receiver DSB noise figure including ABB is 6.7 dB at 19.5 GHz. The supply voltage is 1.5 V. The I/Q RX channel draws 30 mA and the TX channel 45 mA. The chip size including pads is 2150 $\mu$mx 1448 $\mu$m.
According to IPCC 1 , “global warming is likely to reach 1.5°C between 2030 and 2052 if it continues to increase at the current rate“ [1]. Meanwhile, global data traffic continues to rise significantly [2], spurring the development of next generation communication systems. Therefore, it is critical to quantify and decrease environmental impacts of new technologies starting from R&D stage before industrialization. This study aims at comparing die-to-wafer (DtW) 3D interconnection technologies (hybrid bonding and copper pillar) required for 6G application using Life Cycle Assessment (LCA). This study is carried out on CEA-Leti R&D assembly process of silicon dies on $300 \mathrm{~mm}$ silicon wafer with CEA cleanroom data. Results show that hybrid bonding has 1,3 to 2,1 more impacts than copper pillars on relevant selected impact categories in R & D scenario and almost similar impacts in prospective industrial scenario. Future investigations are needed to consolidate results, including whole life cycle. 1 IPCC: Intergovernmental Panel on Climate Change
This paper describes the verification of the phase- and amplitude noise suppression property of a Josephson Arbitrary Waveform Synthesizer (JAWS) circuit. Simulations are performed to gain insight in the improvement in signal-to-noise ratio (SNR) delivered by the JAWS circuit when random amplitude andjitter noise are introduced. Measurements show an improvement for both integrated phase- and amplitude noise of 15 to 40 dB for synthesized frequencies in the range of 0.5-2MHz.
This paper describes a D-band phase shifter for a phased array antenna system implemented in a 22-nm FD-SOI CMOS process. The phase shifter employs a vector modulator-type circuit that synthesizes gain-controlled quadrature signals. The gain control circuit is realized by controlling the path of the unequal distribution synthesis circuit ON/OFF by an amplifier with an enable function. Measurements of this circuit, including temperature characteristics, confirmed that it is capable of 360° phase control in approximately 5° steps at 140-170GHz, with a pass-through gain of approximately -12dB at 150 GHz and room temperature. At 150 GHz for all phase states, the RMS phase errors and the RMS gain errors were confirmed to be 4.1° and 0.28 dB, respectively. Even in the temperature range from 0 to 50°C and the frequency range of 140-170GHz, the RMS phase error was less than 6.8° and the RMS gain error was less than 0.81 dB. The implementation of the circuit with only ON/OFF control of the amplifier resulted in a phase shifter with excellent stability over temperature, with little phase and gain error even when measured at room temperature ±25°C.
Physics-based simulations allow for an accurate insight into the impact of trap dynamics on GaN HEMT performance. In particular, traps are responsible for the low-frequency dispersion of AC performance, e.g. the Y parameters. In this paper we present an in-house TCAD simulator implementing the trap rate equations coupled to the drift-diffusion physical model and solved through the Harmonic Balance algorithm. The developed TCAD allows for the extraction of the trap rate equations Green’s Functions (GFs) in the frequency domain. GFs are then used to compute the sensitivity of the AC Y parameters towards variations of the trap physical parameters (e.g. the trap energy) and to extract the local sensitivity, showing the parts of the device where traps influence most the HEMT AC parameters.
This work describes the use of a triple barrier resonant tunneling diode (TB-RTD) as a zero-bias detector for terahertz (THz) radiation detection. The device is grown using the InP (Indium Phosphide) material system, which is well suited for high-speed electronic devices operating in the THz range due to its very high electron mobility. The TB-RTD device exhibits a highly nonlinear I-V curve around zero bias, along with a high current density (above 380 kA/c$\mathrm{m}^{2}$), enabling efficient THz signal rectification. The detector test structure used for characterization is presented, along with the responsivity measurement procedure. The TB-RTD’s static IV characteristic and responsivity measurements are evaluated, indicating strong nonlinear characteristics and high sensitivity (up to 2,587 V/W and 1.56 pW/$\sqrt{}$Hz) in the 330-500 GHz band, making it suitable for direct detection applications in THz systems.
In this paper, we report on developing W-band GaAs $p-i-n$ diode phase shifters (PSs) with a wideband phase and amplitude response. The PSs employ the reflection-type architecture with reflective loads providing discrete phase tunability through diode dc control. We address the design challenge of having a stable phase shift and a minimal insertion loss imbalance for the diodes with a relatively low commutation quality factor at high millimeter-wave frequencies. Two PS examples, namely 180- and 90-degree bits, are designed for the targeted (85-105) GHz operating band. The circuits were fabricated in the commercial PIN-pHEMT GaAs process with $8 \times 8 \mu \mathrm{m}^{2}\,p-i-n$ diodes. Both simulated and measured results are in good agreement demonstrating wideband frequency performance.
