Browsing by Author "Wu, TH"

Now showing 1 - 20 of 22

A 0.18-mu m CMOS CMFB downconversion micromixer with deep N-well technology for LO-RF and LO-IF isolation improvements
(2005-04-20)
Meng, CC
;
Hsu, SK
;
Wu, TH
;
Huang, GW
;
電信工程研究所
CMOS deep N-well technology can eliminate the physical effects of NMOS transistors and reduce substrate noise and coupling in order to reach the NMOS channel. These properties result in better LO-IF and LO-RF isolations in a Gilbert micromixer. Two identical 0.18-mu m CMOS downconversion micromixers (except at the RF input stage) with deep N-well or without deep N-well are fabricated in adjacent areas of the same wafer for the purpose of isolation comparison. A -37-dB LO-IF and -38-dB LO-RF isolation downconversion micromixer with 19-dB conversion gain and IP1dB = -20 dBm and IIP3 = -13 dBm when RF = 2.4 GHz and LO = 2.25 GHz is demonstrated here using 0.18-mu m-deep N-well CMOS technology. On the other hand, a downconversion micromixer without deep N-well has almost identical power performance but achieves only -20-dB LO-IF isolation and -21-dB LO-RF isolation. (c) 2005 Wiley Periodicals, Inc.
A 1.2-V fully integrated 2.4-GHz low-noise amplifier in 0.35-mu m CMOS technology
(2003-01-20)
Meng, CC
;
Chiang, MH
;
Wu, TH
;
電信工程研究所
;
Institute of Communications Engineering
A 1.2-V fully integrated 0.35-mum inductively degenerated common source CMOS low-noise amplifier has been demonstrated at 2.4 GHz in this paper. A simple common source configuration can be operated at lower voltage and has lower output impedance when compared with a conventional high output impedance cascode LNA circuit topology. The input matching inductance, output matching inductance, and source degeneration inductance are integrated on a single chip. The fully integrated 2.4-GHz CMOS LNA has 5.27-dB power gain, 17-dB input return loss, 15413 output return loss, 4-dB noise figure, and -1-dBm IIPIdB and 8-dBm IIP3 at V-dd = 1.2 V and I-dd = 10 mA, respectively. (C) 2003 Wiley Periodicals, Inc.
A 5.7 GHz Gilbert upconversion mixer with an LC current combiner output using 0.35 mu m SiGe HBT technology
(2005-06-01)
Wu, TH
;
Meng, CC
;
Wu, TH
;
Huang, GW
;
電信工程研究所
This paper demonstrates a small compact 5.7 GHz upconversion Gilbert micromixer using 0.35 mu m SiGe HBT technology. A micromixer has a broadband matched single-ended input port. A passive LC current combiner is used to convert micromixer differential output into a single-ended output and doubles the output current for single-ended-input and single-ended-output applications. Thus, a truly balanced operation of a Gilbert upconversion mixer with a single-ended input and a single-ended output is achieved in this paper. The fully matched upconversion micromixer has conversion gain of -4 dB, OP(1 dB) of -9 dBm and OIP(3) of 4 dBm when input IF=0.3 GHz. LO=5.4 GHz and output RF=5.7 GHz. The IF input return loss is better than IS dB for frequencies up to 20 GHz while RF output return loss is 25 dB at 5.7 GHz. The supply voltage is 3.3 V and the current consumption is 4.6 mA. The die size is 0.9 X 0.9 mm(2) with 3 integrated on-chip inductors.
A fully integrated 5.2 GHz SiGe HBT upconversion micromixer using lumped balun and LC current combiner
(2005)
Wu, TH
;
Meng, CC
;
Wu, TH
;
Huang, GW
;
電信工程研究所
A small and compact 5.2 GHz upconversion Gilbert micromixer using 0.35 mu m SiGe HBT technology is demonstrated in this paper. The upconversion micromixer has a broadband matched single-ended IF input port. A lumped-clement rat-race hybrid is inserted in the LO input stage to maintain LO phase accuracy. A passive LC current Combiner is used to convert the micromixer differential output into a single-ended output and double the output current. Thus, the IF port, LO port and RF port of the upconverter are single-ended, and the stand-along upconverter is suitable for hybrid RF system applications. The fully matched high linearity micromixer has the conversion gain of -1 dB, OpidB of -10 dBm, OIP(3) of 6 dBm when input IF=300 MHz, LO=4.9 GHz and output RF=5.2 GHz. The IF input return loss is better than 10 dB for frequencies up to 20 GHz while the RF output return loss is better than 13.5 dB for frequencies up to 20 GHz. The supply voltage is 3.3 V, the current consumption is 11.5 mA, and the die size is 0.98x0.83 mm(2) with 6 integrated on-chip inductors.
A high isolation CMFB downconversion micromixer using 0. 18-urn deep N-well CMOS technology
(2003)
Meng, CC
;
Xu, SK
;
Wu, TH
;
Chao, MH
;
Huang, GW
CMOS deep N-well technology can eliminate body effects of NMOS transistors and improve LO-IF and LO-RF isolation in a Gilbert micromixer. A 37 dB LO-IF and 38 dB LO-RF isolation downconversion micromixer with 19 dB conversion gain, IP1dB=-19.5 dBm and IIP3=-12.5 dBm when RF=2.4 GHz and LO=2.25 GHz is demonstrated in this paper by using 0.18 mum deep N-well CMOS technology. e input return loss and output return loss are better than 15 dB for frequencies up to 6 GHz. On the other hand, a downconversion micromixer without deep n-well has almost identical power performance but achieves only 20 dB LO-IF isolation and 21 dB LO-RF isolation even if two kinds of mixers are fabricated in adjacent areas of the same wafer. The downconversion micromixer used here has intrinsically single-to-differential input stage and active differential PMOS loads to increase IF differential gain while CMFB is used to stabilize bias points. An IF differential amplifier converts differential output into a single-ended output. Finally, an off-chip rat-race coupler provides balanced LO signals to facilitate isolation measurement.
A high isolation CMFB downconversion micromixer using 0.18-mu m deep N-well CMOS technology
(2003)
Meng, CC
;
Xu, SK
;
Wu, TH
;
Chao, MH
;
Huang, GW
CMOS deep N-well technology can eliminate body effects of NMOS transistors and improve LO-IF and LO-RF isolation in a Gilbert micromixer. A 37 dB LO-IF and 38 dB LO-RF isolation downconversion micromixer with 19 dB conversion gain, IP1dB=-19.5 dBm and IIP3=-12.5 dBm when RF=2.4 GHz and LO=2.25 GHz; is demonstrated in this paper by using 0.18 mum deep N-well CMOS technology. The input return loss and output return loss are better than 15 dB for frequencies up to 6 GHz. On the other hand, a downconversion micromixer without deep n-well has almost identical power performance but achieves only 20 dB LO-IF isolation and 21 dB LO-RF isolation even if two kinds of mixers are fabricated in adjacent areas of the same wafer. The downconversion micromixer used here has intrinsically single-to-differential input stage and active differential PMOS loads to increase IF differential gain while CMFB is used to stabilize bias points. An IF differential amplifier converts differential output into a single-ended output. Finally, an off-chip rat-race coupler provides balanced LO signals to facilitate isolation measurement.
Behaviors of negative resistances and its influences on VCO design
(2001-02-01)
Kao, YH
;
Wu, TH
;
傳播研究所
;
Institute of Communication Studies
The features of the negative resistance in common source and common gate FET configurations for wideband VCO are studied. They are also explained by the simplified three-capacitor model. A design procedure is then developed. The results are applied to a design of wide band oscillator at the several gigahertz region.
Compact 5.2-GHz GaInP/GaAs HBT gilbert upconverter using lumped rat-race hybrid and current combiner
(2005-10-01)
Meng, CC
;
Wu, TH
;
Lin, MC
;
電信工程研究所
;
Institute of Communications Engineering
A compact 5.2-GHz Gibert upconversion mixer is demonstrated using the 2-mu m GaInP/GaAs heterojunction bipolar transistor (HBT) technology. A miniature lumped-element rat-race hybrid and an LC current combiner are used in the LO port and the RF port of the upconversion Gilbert mixer, respectively. The effect of the LC current combiner with inductor loss is established in this letter. An active IF balun is incorporated in. the Gilbert upconverter with no extra power consumption. The fully integrated Gilbert upconverter has conversion gain of 1 dB, OP (1 dB) of -10 dBm, and OIP3 of 2 dBm when input IF = 300 MHz, LO = 4.9 GHz, and output RF = 5.2 GHz.
DC to 6-Ghz high-gain low-noise GaInP/GaAs HBT direct-coupled amplifiers with and without emitter-capacitive peaking
(2004-10-05)
Meng, CC
;
Wu, TH
;
Lu, SS
;
電信工程研究所
;
Institute of Communications Engineering
High-gain shunt-series shunt-shunt wideband amplifiers with and without emitter-peaking capacitors are demonstrated by using GaInP/GaAs HBT technology. Experimental results show that the power gain is 28 dB front DC to 6 GHz for a wideband amplifier without emitter-peaking capacitors. On the other hand, a wideband amplifier with emitter-peaking capacitors can increase the gain bandwidth up to 8 GHz at the cost of lower input/output return loss. Both circuits have similar power and noise performance. The noise figures of both designs are less than 2.8 dB for frequencies below 6 GHz,. Op(1dB) and OIP3 are 7 and 20 dBin at 2 GHz, respectively. Total current consumption is 67 mA at 5-V supply voltage for both wideband amplifiers. (C) 2004 Wiley Periodicals, Inc.
DC-12-GHz 10-dB gain shunt-series shunt-shunt wideband amplifiers by commercially available 0.35 mu M SiGe HBT technology
(2004-03-20)
Meng, CC
;
Wu, TH
;
Huang, GW
;
電信工程研究所
;
Institute of Communications Engineering
The realization of wideband amplifiers with shunt-series shunt-shunt dual-feedback configuration with commercially available 0.35-mum SiGe BiCMOS technology is reported. The SiGe HBT used here has f(t) of 67 GHz and BVceo of 2.5 V. The experimental results show that power gain is 10 dB from DC to 12 GHz for a shunt-series shunt-shunt wideband amplifier, with the help of the emitter capacitive peaking technique. Input- and output-return losses are better than 10 dB for the same frequency range. Noise figure increases from 8 to 12 dB for frequencies from 1 to 18 GHz. OP1db and OIP3 are 0 dBm and 12 dBm at 1 GHz, respectively. Total current consumption is 11 mA at 3.3 V supply voltage. (C) 2004 Wiley Periodicals, Inc.
Exchange coupling coefficient and domain wall mobility of (Dy,Tb)FeCo magneto-optical films
(1997-06-01)
Hwang, WK
;
Wu, TH
;
Shieh, HPD
;
交大名義發表
;
光電工程學系
The magnetic properties and recording characteristics of (Dy,Tb)Fr,Co films are dependent upon the anisotropy dispersion of rare earth constituents and the exchange integral between rare earth and transition metal sub networks. The exchange coupling coefficient lambda and the domain wall mobility mu(eff) are found to be interrelated by both measurements and calculations. The higher the lambda or the lower the mu(eff), the higher bias field that is required for magneto-optical recording. The temperature independent lambda is demonstrated to be an important parameter in the assessment of magnetic properties and recording characteristics of sperimagnetic films. (C) 1997 American Institute of Physics.
General purpose RFICs using 0.35 um SiGeBiCMOS technology
(2005)
Meng, C
;
Wu, TH
;
Wu, TH
;
電信工程研究所
;
Institute of Communications Engineering
This paper presents an overview of several 5 GHz general purpose RFICs implemented using 0.35 urn SiGe BiCMOS technology by our research group. These RFICs include emitter-inductively-degenerated cascode LNA, shunt-series shunt-shunt dual feedback wideband amplifier, common mode feedback Gilbert down-converter, Gilbert up-converter with LC current mirrors and Gilbert down-converter with polyphase filters for image rejection.
Geometric shape dependence of coercivity for patterned magnetic thin films
(2000-02-01)
Wu, TH
;
Wu, JC
;
Wu, CS
;
Chen, BM
;
Shieh, HPD
The micro-strips with patterned magnetic domains using electron beam lithography have been made to study the geometric shape dependence of coercivity. The size of the micro-strip is 10 mu m x 30 mu m with 0.5 mu m periods hole arrays pattern. Arrays with different types of geometry, such as square- circle-, and ellipse-shapes have been made. amorphous rear-earth transition-metal (RE-TM) thin films with magnetic perpendicular anisotropy were deposited on the microstrips. The extraordinary Hall effect has been employed to measure the hysteresis loop and the coercivity of the sample. We have found that for the same deposited material, the magnitude of coercivity for various shapes are dissimilar. For example, the coercivity of the ellipse-shaped hole arrays is much larger than the coercivity of the square-shaped hole arrays. In addition, we observed that the coercivity of the patterned sites was larger than the sites without patterns for RE-dominated compounds and the coercivity of the patterned sites was smaller than the sites without patterns for TM-dominated compounds. The magnetic moments canting between RE and TM subnetworks will be used to explain the observed phenomena. (C) 2000 Elsevier Science B.V. All rights reserved.
Magnetic domain pinning in lithographically patterned amorphous magnetic layer
(1999-03-01)
Wu, JC
;
Huang, YW
;
Chen, BM
;
Wu, TH
;
Shieh, HPD
Artificial pinning sites have been fabricated for the magnetic domain pinning in perpendicular anisotropy magneto-optical (MO) thin film media. The pinning sites were radial hole array made by patterning a layer of radial gold grid on the silicon nitride (SiN) coated silicon (Si) substrate using a standard electron beam lithography. Various magnetization and demagnetization procedures were performed to investigate the magnetic domain pinning behavior. A polar Kerr microscope was used in situ to monitor the magnetic wall motion and a magnetic force microscope was employed to scan the magnetic domain structures. Magnetic domains were found to be pinned inside the hole array and resembled to the geometric shape of the holes. The coercivity in the patterned MO layer is much higher than in thr unpatterned MO layer. Moreover, the coercivity in the larger pinning sites area is higher than in the smaller pinning sites area.
Magnetic domain pinning in patterned magneto-optical material
(1999-04-15)
Wu, TH
;
Wu, JC
;
Huang, YW
;
Chen, BM
;
Schieh, HPD
Pinning magnetic domains with defined shapes have been made for magneto-optical thin film material. The pinning array of holes on the substrate was fabricated using electron beam lithography. The domain pinning behaviors were studied by applying an external magnetic field perpendicular to the film plane and imaged by atomic force microscopy in their remanent state. Magnetic domains were found to be pinned inside the hole arrays and reproduced the shapes of the holes. The coercivity in the patterned region is larger than in the nonpatterned region. Moreover, the coercivity inside the hole sites is larger than the land sites. (C) 1999 American Institute of Physics. [S0021-8979(99)38708-9].
Magnetic domain pinning in patterned perpendicular magnetic anisotropy material
(1999-07-01)
Wu, TH
;
Wu, JC
;
Chen, BM
;
Shieh, HPD
;
光電工程學系
A method for pinning magnetic domains with prescribed shapes has been developed for the magneto-optical thin film material Dy(x)(FeCo)(1-x). The pinning array of holes on the substrate was fabricated using electron beam lithography. The thin film of Dy(x)(FeCo)(1-x) sandwiched between SiN layers was deposited onto a patterned polymethyl methacrylate layer. The pinned domains acquire the shape of the holes, while the sharpness of their boundaries depends on the size of the holes. The stability of the pinned domains is affected by the sample's magnetization. In our experiments, the minimum stable domain size pinned within the hole array was found to be around 50 nm in diameter. (C) 1999 Elsevier Science B.V. All rights reserved.
Magnetoresistance anomaly in DyFeCo thin films
(2001-06-01)
Wu, JC
;
Wu, CS
;
Wu, TH
;
Chen, BM
;
Shieh, HPD
Microstructured rare-earth-transition-metal DyFeCo films have been investigated using magnetoresistance and extraordinary Hall-effect measurements. The Hall loops reveal variation of coercive fields depending on the linewidth and the composition of the films. The magnetoresistance curves, with changes up to as high as 1.3%, show positive/negative magnetoresistance peaks centered on the coercive fields depending on the linewidth of the films only. The variation of the coercivity can be attributed to the magnetic moment canting between the Dy and FeCo subcomponents and the existence of the diverged magnetization on the edges, and the anomalous magnetoresistance peaks observed are discussed with the existing theories. (C) 2001 American Institute of Physics.
Mapping spatial variations of the coercivity on patterned magneto-optical materials
(2003-05-15)
Wu, TH
;
Ye, LX
;
Lee, JM
;
Wu, JC
;
Shieh, HPD
The spatial variation of coercivity on patterned magneto-optical (MO) materials has been studied with a Kerr microscope. Our results show that the coercivity values on arrays made by holes are larger than those observed on the land regions. In contrast with nonpatterned MO media, which present a sharp Gaussian-like distribution, the coercivity distribution in the patterned material is much wider and shows an exponential-type distribution. This article maps the spatial variation of the coercivity on patterned magnetic materials with a spatial resolution of 0.1x0.1 mum(2). (C) 2003 American Institute of Physics.
Pinning magnetic domain via patterning artificial lattice under amorphous magnetic layer
(1999-03-01)
Wu, TH
;
Wu, JC
;
Chen, BM
;
Shieh, HPD
;
光電工程學系
A novel method for pinning magnetic domains on pre-formatted pinning sites has been developed for perpendicular magnetic anisotropy magneto-optical (MO) thin film media. The pinning sites were artificial lattices made by patterning a layer of gold grid on the substrate using electron beam lithography. Compared with the work proposed using photolithography, our method showed the ability of making smaller pinning domains. Moreover, compared with our previous work using the same electron beam lithography method and creating hole arrays on polymethyl methacrylate, this new method could stand higher film deposition temperatures and avoid the MO films rapid deterioration on a PMMA substrate. In order to investigate the domain pinning behavior we did several different procedures and observed their corresponding domain pinning layouts. Magnetic domains were found to be pinned inside the lattice and resembled the geometric shapes of the lattice. The pinned domains acquired the shape of the lattice. In this article, the star-shaped type of geometry of grid arrays is presented. (C) 1999 Elsevier Science B.V. All rights reserved.
Relaxation of pinned domains in patterned magnetic thin films
(2000-02-01)
Wu, TH
;
Wu, JC
;
Huang, YW
;
Ye, LX
;
Chen, BM
We have shown in previous papers that the magnetic domains could be pinned inside the artificially patterned hole arrays under suitable geometry aspect ratio. Nevertheless, we have found that if we reverse the magnetization directions through domain wall motion, the confined domains expand from smaller territory into larger territory for some samples. In addition, the pinned domains maintained the same moment's orientation after the domain expansion. The possible reason for the pinned domains to retain the same moment's orientation maybe those pinning holes that act as high anisotropy defects. Thus, domain wall motion was around the high anisotropy sites and only peeled away the domain in the land area while the enclosed domain of the hole area maintained the same orientation. Moreover, the feasible reason for the expansion domains is the coercive force, which is perpendicular to the side-walls and pinning the domains inside the holes, are relaxed and thus causing the domains growth. This phenomenon is called the relaxation of pinning domains. (C) 2000 Elsevier Science B.V. All rights reserved.