RF Transconductance Research Articles

A study of DC and RF transconductance for different technologies of HEMT at low and high temperatures

A study of DC and RF transconductance for different technologies of HEMT at low and high temperatures

A 12.4–32 GHz CMOS Down-Conversion Mixer for 28 GHz 5G New Radio (NR)

We report a low voltage (VDD) and power (PDC) 12.4–32 GHz CMOS down-conversion mixer with high conversion gain (CG) for 28 GHz 5G communications. A quarter-wavelength (λ/4) transmission line (TL) and a coupling capacitor (Cc), named the λ/4-TL-C-based coupler, is proposed. This is the way to attain low-VDD, independent RF transconductance (gm)-stage bias, harmonic suppression, and near perfect coupling from the RF gm stage to the LO switch transistors. The body-self-forward-bias (BSFB) technique, i.e., connection of the gm-stage transistors’ body to drain via a large body resistance, is used for threshold voltage (Vth) and VDD reduction and substrate leakage suppression. CG and noise figure (NF) enhancement at the same or even a lower PDC is achieved because lower VDD and higher gm (due to larger bias current) are used. To facilitate the RF measurement, a compact Wilkinson-power-divider-based balun with small-phase deviation and amplitude imbalance is included at RF and LO inputs. The mixer consumes 6.5 mW and achieves a CG of 14.4 ± 1.5 dB for 12.4–32 GHz (i.e., 3 dB bandwidth (f3dB) of 19.6 GHz), a lowest noise figure (NFmin) of 7 dB, and figure-of-merit (FOM) of 0.023, which is one of the best results ever reported for millimeter-wave (mm-wave) down-conversion mixers with an f3dB larger than 10 GHz and PDC lower than 10 mW.

High Efficiency Over 70% at 3.6-GHz InAlN/GaN HEMT Fabricated by Gate Recess and Oxidation Process for Low-Voltage RF Applications

In this work, high-efficiency InAlN/GaN high electron mobility transistor (HEMT) is fabricated by recess and oxidation process under the gate (RAO) for low-voltage RF applications. Compared with conventional HEMT fabricated on the same heterojunction, RAO-HEMT exhibits a significant decrease in gate leakage, attributed to the thin oxide layer formed around and under the gate as well as adjustment of the location of the peak electric field and confinement of the strong electric field in the oxide. Due to the alleviated electric field as well as adoption of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\text{N}_{{2}}\text{O}$ </tex-math></inline-formula> for plasma treatment, suppressed RF dispersion including the extremely low current collapse and negligible knee voltage walkout is obtained for RAO-HEMT. Although RAO-HEMT delivers lower saturation current as the result of the partial oxidation of the barrier under the gate, it shows a similar ON-resistance to conventional HEMT, which contributes to the reduction in knee voltage. Besides, higher <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${f}_{T}/{f}_{\text {MAX}}$ </tex-math></inline-formula> is achieved for RAO-HEMT, as the result of the higher dc transconductance and suppressed RF transconductance collapse as well as larger output resistance. Eventually, benefiting from the improvement in gate leakage, RF dispersion, knee voltage, and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${f}_{\text {MAX}}$ </tex-math></inline-formula> by RAO, RAO-HEMT delivers both higher power-added-efficiency (PAE) and output power ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${P}_{\text {OUT}}$ </tex-math></inline-formula> ) than conventional HEMT. At 3.6 GHz, RAO-HEMT provides a high PAE of 71% (74%) at <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${V}_{{\text {DS}}}$ </tex-math></inline-formula> of 6 V (9 V), which performs better than most of the existing Si, GaAs, or GaN RF devices in PAE, revealing the great potential of InAlN/GaN RAO-HEMT in low-voltage low-power consumption RF applications.

Performance projection of multi‐bias and nonlinear distortion for gallium arsenides nano‐pHEMT

AbstractIntermodulation distortion tests using the two‐tone technique with an extensive bias range were used to demonstrate the nonlinearity in terms of frequency for active devices like nano pHEMT based on gallium arsenide semiconductor material. It is demonstrated that the intermodulation distortion components are directly connected to the RF transconductance derivatives calculated using the two‐tone test. Furthermore, it was discovered that the third‐order intermodulation distortion power is lowest when the second derivative of the transconductance Gm3 is zero. To emphasize the influence of RF transconductance, a two‐tone approach has been presented as an empirical, analytical explanation for this device's nonlinearity. The modeled data is demonstrated to agree with the results of the measurements, making it a helpful tool for analyzing these devices.

Low‐Voltage Low Noise Figure Down‐Conversion Mixer for Band #1 of MB‐OFDM System in 180 nm Complementary Metal Oxide Semiconductor Technology

Low‐voltage design is a challenge for Gilbert cell‐based mixers due to stacking transconductance and switching stage. This work addresses this issue by proposing a design of a low‐voltage down‐conversion mixer for band #1 of multiband orthogonal frequency division multiplexing (MB‐OFDM) system in 180 nm complementary metal oxide semiconductor (CMOS) technology. The mixer is tuned at band #1 at RF frequency of 3.432 GHz and IF frequency of 264 MHz. The proposed mixer uses folded cascode connection of LO and RF in order to increase headroom and to reduce the DC supply voltage for low‐voltage operation. Common gate configuration is used at RF transconductance stage to enhance the input bandwidth of the mixer. RF and LO ports are matched to 50 Ω using differential T and LC matching, respectively. The resistive source denegation technique is used to linearize transconductance with respect to the bias point. The common source stage is used at the IF port as a buffer cum matching circuit it. The simulation results of the mixer show the maximum conversion gain of 9.76 dB, 1 dB compression point (P1dB) of ‐16.25 dBm, the third‐order input intercept point (IIP3) of ‐4.70 dBm, an SSB noise figure of 9.036 dB, and S11of ‐19.490 dB at the supply voltage of 1.2 V. Excluding off chip components, proposed mixer records an active area of 926.35 μm2.

High gain and low noise wideband folded-switching mixer employing transformer and complimentary CS/CG hybrid topology

In this paper, a high gain and low noise wideband folded-switching down conversion mixer has been designed. A differential complimentary common source/common gate (CS/CG) hybrid topology is utilized as the input RF transconductor stage, where a transformer is subtly applied to the source nodes of the complimentary CS/CG transistors. With the transformer and complimentary CS/CG hybrid topology as the input RF transconductor for both RF signal amplification and chip area reduction, the conversion gain, noise figure, input impedance matching, and chip area-efficiency of the mixer can be signficantly improved. Under a 0.8 ​V supply voltage and 1 ​dBm LO power, the mixer exhibits a conversion gain of 12.2–17.9 ​dB and a double-sideband noise figure of 2.6–6.1 ​dB between RF frequencies from 3.0 to 7.0 ​GHz. The IIP3 is −6.5 ​dBm at the RF frequency of 5.0 ​GHz while consuming only 2.72 ​mW of power at an IF frequency of 100 ​MHz. The chip area is only 0.55 ​mm ​× ​0.65 ​mm.

A 5.4 GHz Fully-Integrated Low-Noise Mixer

An active mixer using a Gilbert-cell topology that incorporates a low-noise RF transconductance stage is presented in this paper. The chip operates at a frequency of 5.4 GHz and was fabricated in 180 nm CMOS technology. A current-bleeding circuit is used to provide different dc bias currents to the LO switching stage and the RF transconductors. The transconductor, designed using the power constrained simultaneous noise and input match technique, together with the bleeding circuit enables the mixer to have a measured single-sideband noise figure of 7.8 dB and a power conversion gain of 13.1 dB. The measured input-referred 1-dB compression point, IP1dB is -17.8 dBm while its OP1dB is -5 dBm. A two-tone test was carried out and the mixer exhibits an IIP3 of -6.2 dBm and an OIP3 of +6.9 dBm. All of the inductors are on-chip and the size of the mixer core is only 380 μm x 350 μm (0.133 mm2).

A high‐linearity Ka‐band CMOS down‐conversion mixer

AbstractThis article presents a high‐linearity Ka‐band complementary metal–oxide–semiconductor (CMOS) down‐conversion mixer. The mixer comprises an RF transconductance stage, switching stage, local oscillator buffer, and intermediate frequency (IF) buffer, and employs modified multiple gate transistors for high linearity. The intermodulation and gain characteristics are analyzed, revealing that the proposed IF buffer can enhance the linearity performance without a gain reduction. Implemented in 65‐nm CMOS technology, the proposed mixer consumes 23 mW and occupies a core chip area of 0.54 mm2. The mixer achieves a conversion gain of 9.3 to 11.1 dB, a noise figure of 8.4 to 9.2 dB, and an output third‐order intercept point of 15.6 to 23.6 dBm at an RF frequency of 37 to 41 GHz.

Design of 0.13 µm low power CMOS subharmonic mixer for DCR applications

ABSTRACT This paper presents a design of 2.4 GHz folded subharmonic active mixer in TSMC 130 nm CMOS process for direct conversion receiver (DCR). To achieve the low supply voltage, the folded topology is adopted for the proposed subharmonic mixer. The complementary current reuse structure is used as RF transconductance stage where the biasing for the PMOS and NMOS transistors are self-biased by a resistor to support high conversion gain at low supply voltage. Frequency doubling operation is done by the mixing core which is driven by a quadrature local oscillator (LO) signals. The proposed mixer is designed and simulated using ADS (Advanced Design System) simulator with Cadence software. It produces 10.5 dB conversion gain, a single sideband) noise figure of 9.5 dB, 2 LO-to-RF isolation of 53 dB, LO-to-RF isolation of 117 dB, RF-to-IF isolation of 55 dB and IIP3 of −7.5 dBm. The quadrature LO subharmonic mixer consumes 824μW power from a supply voltage of 0.8 V. The proposed mixer is compared with existing subharmonic mixers. The comparison shows the proposed subharmonic mixer gives a good Figure of Merit.

AV-Band Ultra Low Power Sub-HarmonicI/QDown-Conversion Mixer Using Current Re-Used Technique

In this brief, we present a low power sub-harmonic I/Q mixer using the current re-used technique for WiGig at 60 GHz. A transformer integrates the RF trans-conductance stage and LO switching stage by the current re-used technique to save dc power. We only adopt a single transistor at RF stage to replace differential RF amplifier to save die area as only half of the RF matching network is required. The trans-conductance of the RF stage is increased by the $g_{m}$ -boost technique. In addition, since the current re-used technique gathering the dc current from both ${I}$ and ${Q}$ mixer cores, the transconductance of the RF stage is also increased. In order to reduce the die area of passive components, we use mutually coupled inductors to implement lumped element Wilkinson power divider. We also add extra capacitors on all ports of the coupler to improve return loss. The measured peak gain is −7.9 dB at 61 GHz under LO power of 5 dBm at 30 GHz. The measured IP1dB is better than −9.5 dBm in four channels. The measured 3-dB RF bandwidth is from 56.5-67 GHz and 3-dB IF bandwidth is 1.5 GHz. Total dc power consumption is only 3.45 mW and each mixer consumes 1.73 mW for 1.1 V supplied voltage. To our best knowledge, this current re-used I/Q mixer has the lowest dc power and good conversion gain among active I/Q mixers at ${V}$ -band.

On the correlation between intermodulation distortion and RF transconductance for microwave GaN HEMT

This paper investigates the correlation between intermodulation distortion and RF transconductance in AlGaN/GaN high electron mobility transistor (HEMT) grown on SiC Substrate. Two-tone intermodulation distortion measurements were carried out across a comprehensive biasing range changing the frequency and ambient temperature and input power to point out this correlation. The main contribution is that, the correlation between intermodulation distortion and the behavior of transconductance and its derivatives (Gm, Gm2, and Gm3) is investigated and summarized. It is demonstrated that the nonlinearity is directly correlated to the RF transconductance behavior where the peak values of the fundamental IMD arise at the peak of Gm and the null points in IMD2 and IMD3 arises when Gm2 and Gm3 alters from positive to negative magnitudes. The changes in frequency, input power and temperature affect the threshold voltage VT which in turn impacts on that phenomenon. As the nulls/notches are crucial indicator for identifying minimum distortion level, an empirical model for null’s current is develop in correlation with transconductance. Finally, the measured and simulated IMDs as well as the null’s current shows very good agreement. In addition, the small-signal measurement and as well as the de-embedding of the device has been done before the IMD study. The power gain and its derivatives also compared to the transconductance. This analysis is important to know how the intermodulation distortion products behave and to identify a distortion free device at the particular operation point of interest.

A W-band CMOS down-conversion mixer using CMOS-inverter-based RF GM stage for conversion gain and linearity enhancement

A W-band (75–110 GHz) down-conversion mixer with CMOS-inverter-based RF transconductance (GM) stage for 94 GHz image radar sensors in 90 nm CMOS is reported. Due to the current bleeding and GM contribution of the upper PMOS transistors of the RF GM stage, a larger load resistance can be adopted and a larger GM can be obtained while keeps the same linearity. This leads to a conversion gain (CG) enhancement. In addition, the second-order term GM (g2) and third-order term GM (g3) of the main NMOS transistors of the RF GM stage can be cancelled by those of the upper PMOS transistors and the auxiliary NMOS transistors of the RF GM stage. This leads to a better linearity. The mixer consumes 8.56 mW and achieves an excellent RF-port input reflection coefficient of − 10 to − 18 dB for frequencies of 88.4–103.4 GHz. In addition, for frequencies of 70–100 GHz, the mixer achieves CG of 13.9–17.5 dB and LO-RF isolation of 40.6–43.5 dB, one of the best CG and LO-RF isolation results ever reported for a down-conversion mixer with operation frequency around 94 GHz. Furthermore, the mixer achieves prominent NF of 16.4 dB and output third-order intercept point of 14.1 dBm at 94 GHz. These results demonstrate that the proposed down-conversion mixer architecture is very promising for 94 GHz image radar sensors.

A High Linearity High Gain Double Balanced Down Conversion Mixer Using Current Bleeding for Mode 1 of UWB System

This work presents high linearity and high gain CMOS down conversion Gilbert cell mixer for ultra wide band application, designed in 0.18𝜇𝑚 technology. Inductive source degeneration is used at the RF trans-conductance stage of Gilbert cell mixer to improve the linearity of the mixer. Differential LC matching is used at RF and LO stage and current bleeding is used to increases the conversion gain and linearity of the mixer. Output buffer is used at the load to achieve 50 ohm impedance match. The proposed mixer shows low reflection coefficient less than –10 dB for entire band 1, band 2, and band 3 ranges from 3.168 GHz to 4.752 GHz frequency. The proposed mixer works at DC supply of 1.5 volt for low power consumption. Simulation results shows that the mixer achieves the conversion gain of 11.03 dB, 1 dB compression point of -14.65 dB, IIP3 is 0.257 dBm, input return loss is –24.292 dB and measured matched RF (input) port impedance of 48.79 ohm.

Flicker Noise Improved CMOS Mixer Using Feedback Current Bleeding

A CMOS mixer using a negative-feedback current-bleeding circuit is presented with low flicker noise and robustness to process–voltage–temperature variations. The proposed technique reduces the common-mode flicker noise generated by the RF transconductance stage and the current-bleeding circuit, thus reducing conversion to the differential-mode flicker noise caused by mismatch and process variations. In addition, the proposed technique achieves stability against the voltage and temperature variations due to the negative feedback. Measurement results show a voltage conversion gain of 19 dB, an IIP3 of −7 dBm, and a flicker-noise figure of 14.2 dB at 10 kHz. The proposed mixer dissipates 3 mW from a 1.3 V supply voltage.

94 GHz down-conversion mixer with gain enhanced Gilbert cell in 90 nm CMOS

A 94 GHz down-conversion mixer for image radar sensors using standard 90 nm CMOS technology is reported. The down-conversion mixer comprises a double-balanced Gilbert cell with an LC-tank-oscillator-based RF transconductance stage load for conversion gain (CG) enhancement and noise figure (NF) suppression, two miniature planar baluns for converting the single RF and LO input signals to differential signals, and an IF amplifier. The mixer consumes 9.5 mW and achieves excellent RF-port input reflection coefficient (S11) of −27.2 dB at 94 GHz, and S11 smaller than −10 dB for frequencies of 83.6–110 GHz. That is, RF-port −10 dB input matching bandwidth is greater than 26.4 GHz. In addition, for frequencies of 75–100 GHz, the mixer achieves CG of 4.9–7.9 dB (the corresponding 3-dB CG bandwidth is greater than 25 GHz), LO–RF isolation of 38.5–44.7 dB and NF of 15.4–21.2 dB, one of the best CG, LO–RF isolation and NF results ever reported for a down-conversion mixer with operation frequency around 94 GHz. Furthermore, the mixer achieves an excellent input third-order intercept point of 2.6 dBm at 94 GHz. These results demonstrate the proposed down-conversion mixer architecture is promising for 94 GHz image radar sensors.

A 90–96 GHz CMOS down-conversion mixer with high conversion gain and excellent LO–RF isolation

A 90–96 GHz down-conversion mixer for 94 GHz image radar sensors using standard 90 nm CMOS technology is reported. RF negative resistance compensation (NRC) technique, i.e. NMOS LC-oscillator-based RF transconductance (GM) stage load, is used to increase the output impedance and suppress the feedback capacitance Cgd of RF GM stage. Hence, conversion gain (CG), noise figure (NF) and LO-RF isolation of the mixer can be enhanced. The mixer consumes 15 mW and achieves excellent RF-port input reflection coefficient of −10∼−36.4 dB for frequencies of 85∼105 GHz. The corresponding −10 dB input matching bandwidth is 20 GHz. In addition, for frequencies of 90∼96 GHz, the mixer achieves CG of 6.3∼9 dB (the corresponding 3-dB CG bandwidth is greater than 6 GHz) and LO-RF isolation of 40∼45.1 dB, one of the best CG and LO-RF isolation results ever reported for a down-conversion mixer with operation frequency around 94 GHz. Furthermore, the mixer achieves an excellent input third-order intercept point (IIP3) of 1 dBm at 94 GHz. These results demonstrate the proposed down-conversion mixer architecture is very promising for 94 GHz image radar sensors.

Design and implementation of a 94 GHz CMOS down-conversion mixer with integrated miniature planar baluns for image radar sensors

A 94 GHz down-conversion mixer for image radar sensors using standard 90 nm CMOS technology is reported. The down-conversion mixer comprises a double-balanced Gilbert cell with peaking inductors between RF transconductance stage and LO switching transistors for conversion gain (CG) enhancement and noise figure suppression, a miniature planar balun for converting the single RF input signals to differential signals, another miniature planar balun for converting the single LO input signals to differential signals, and an IF amplifier. The mixer consumes 22.5 mW and achieves excellent RF-port input reflection coefficient of ź10 to ź35.9 dB for frequencies of 87.6---104.4 GHz, and LO-port input reflection coefficient of ź10 to ź31.9 dB for frequencies of 88.2---110 GHz. In addition, the mixer achieves CG of 4.9---7.9 dB for frequencies of 81.8---105.8 GHz (the corresponding 3-dB CG bandwidth is 24 GHz) and LO---RF isolation of 37.7---47.5 dB for frequencies of 80---110 GHz, one of the best CG and LO---RF isolation results ever reported for a down-conversion mixer with operation frequency around 94 GHz. Furthermore, the mixer achieves an excellent input third-order intercept point of ź3 dBm at 94 GHz. These results demonstrate the proposed down-conversion mixer architecture is promising for 94 GHz image radar sensors.

High-Field Stress, Low-Frequency Noise, and Long-Term Reliability of AlGaN/GaN HEMTs

A small degradation of threshold voltage and a significant reduction in dc and RF transconductance are observed for AlGaN/GaN high-electron-mobility transistors fabricated via three different process techniques and subjected to high voltage stress in the semi-ON bias condition. Low-frequency noise measurements as a function of temperature show peaks at ∼0.2, ∼0.5, ∼0.7, and/or ∼0.8 eV that increase with the magnitude and duration of high-field stress. The dehydrogenation of passivated Fe and O impurity defects and/or N antisite defects leads to increased charge trapping, noise, and transconductance degradation. Density functional theory calculations show that Ga vacancy motion and dehydrogenation of substitutional Fe complexes in the GaN buffer layer also contribute significantly to the poststress degradation of these devices. At higher stress voltages, these defects also likely contribute to device failure.

6.3 mW 94 GHz CMOS Down-Conversion Mixer With 11.6 dB Gain and 54 dB LO-RF Isolation

A 94 GHz CMOS down-conversion mixer is reported. RF negative resistance compensation (NRC) technique, i.e., PMOS LC-oscillator-based RF transconductance (GM) stage load, is used to increase the output impedance and suppress the feedback capacitance C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">gd</sub> of RF GM stage. As a result, conversion gain (CG), noise figure (NF) and LO-RF isolation of the mixer are enhanced. For frequencies of 80~110 GHz, the mixer consumes 6.3 mW and achieves excellent RF-port input reflection coefficient (S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sub> ) of -8.7~ -22 dB and LO-port input reflection coefficient (S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">22</sub> ) of -10.3~-19.4 dB. In addition, the mixer achieves excellent CG of 4.1~11.6 dB, NF of 15.8~18.1 dB, and LO-RF isolation of 42.1~54 dB for frequencies of 80~110 GHz, one of the best CG, NF and LO-RF isolation results ever reported for a W-band CMOS down-conversion mixer.

Using the Gate–Bulk Interaction and a Fundamental Current Injection to Attenuate IM3 and IM2 Currents in RF Transconductors

Two new linearization techniques to attenuate the second-order intermodulation (IM2) and third-order intermodulation (IM3) currents in transconductance stage of CMOS active mixers are proposed. In the first technique, the third-order Volterra kernel of the transconductance stage output current is cancelled using the interaction between the gate and bulk terminals of the input transistors. For this end, a IM2 voltage, with an adjustable magnitude and phase, is produced and applied to the bulk of these transistors. This produces an interaction term in the IM3 current of the stage and attenuates the total IM3 current. In the second technique, a fundamental component is added to the current of the tail source in the transconductance stage. This component produces an IM2 term in the drain current of the input transistors, caused by g’ m of the input transistors, with an equal magnitude and opposite phase related to the intrinsic IM2 current of the stage, and cancels the total IM2 current. Spectre-RF simulation results show that the proposed techniques simultaneously improve the third-order input intercept point and second-order input intercept point by $\sim 14.1$ and 16.4 dB, respectively, compared with the conventional active mixer, while 1.66-mA extra current is drawn from a single 1.2 V power supply.