Kazuo Tsubouchi

Tohoku University, Sendai-shi, Miyagi, Japan

Are you Kazuo Tsubouchi?

Claim your profile

Publications (122)88.48 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: High speed wireless LAN prototype of 324 Mbit/sec has been developed. Six channel binding of 802.11 a signal on frequency domain was employed for increasing PHY data rate. Frame aggregation was employed to improve MAC-SAP throughput. Individual adaptive data rate setting on each channel, so called adaptive dynamic channel assign was implemented in order to achieve maximum MAC-SAP throughput at any channel condition. In this paper, multiple channel interleaving over all frequency channels for randomization of burst error was presented. Interleaving can be carried out over all channels or inside individual channel. Bit and packet error rate were measured using implemented WLAN equipements with/without multiple channel interleaving. It was confirmed that improvement using multiple channel interleaving were measured to be 2 dB.
    Preview · Conference Paper · Oct 2008
  • [Show abstract] [Hide abstract]
    ABSTRACT: A 5 GHz differential power amplifier (PA) module with a dipole antenna has been implemented. The PA integrated circuit (IC) was mounted on a fiuoroplastic substrate on which the dipole antenna was fabricated. The differential PA IC was fabricated using 0.18 μm Si-CMOS process. For the differential operation, an n/nMOS push-pull circuit was employed. The PA IC was directly connected to the balance input of dipole antenna without a balun using stud bump bonding on a fiuoroplastic substrate. The antenna was designed using a 3-dimensional electromagnetic simulator including a feed line. The measured gain of the PA module was 12.6dB at 5 GHz.
    No preview · Article · Apr 2008 · IEICE Electronics Express
  • [Show abstract] [Hide abstract]
    ABSTRACT: A 5 GHz-band SiGe-MMIC transceiver for 324 Mbps transmission is developed. The broadband transceiver characteristics are obtained by utilizing a 5 GHz- band Q-MOD with a circular polyphase filter in the transmitter, and a broadband IFVGA and Q-MIX in the receiver. A developed SiGe-MMIC performs EVM of 3.3%rms and EVM of 5.9%rms in the transmitter and the receiver, respectively using 64-QAM OFDM signal with 120 MHz bandwidth, which corresponds to the data rate of 324 Mbps. Utilizing the developed SiGe-MMIC, the Hi-Vision streaming video data is transmitted without interruption, and is clearly projected onto the TV.
    No preview · Conference Paper · Feb 2008
  • Tran Ngoc Hao · Suguru Kameda · Tadashi Takagi · Kazuo Tsubouchi
    [Show abstract] [Hide abstract]
    ABSTRACT: In this work, delay profiles of indoor multipath characteristics at 60 GHz were measured using a pulse compression method based on M-sequences. Packets included preamble of two M-sequences were generated by a software modulator and a 4 GSample/s dual channel arbitrary waveform generator. Baseband signal was converted to 60 GHz band by transmitter. At receiver, a 4 GSample/s sliding correlator was built into a demodulation software, resulting in 0.25 ns delay time resolution. Arrival angels of multipaths were measured by rotating a narrow beam received antenna with angle resolution of 5 degrees. From measurement results, the multipaths were found to be separable. The multipaths can be combined by using a rake receiver.
    No preview · Article · Jan 2008
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have proposed an ultra-small and low-cost wireless 3-dimentional system-in-package that includes 60-GHz-band radio frequency (RF) components. 60-GHz-band integrated circuits (ICs) of a 60-GHz-band transmission (TX) module were mounted onto an any layer interstitial via hole (ALIVH) substrate using stud bump bonding (SBB) technology. A mixer using antiparallel diode pair, power amplifier and driver amplifier ICs were fabricated on GaAs substrates. The characteristics of the fabricated TX module in a package were measured. The measured conversion gain and output power at 1-dB gain compression point of the TX module using a single balance mixer were 13 dB and 12 dBm, respectively. The 3-dB bandwidth at lower-side-band (LSB) output was found to be 3 GHz. The measured conversion gain and output power at 1-dB gain compression point of the TX module using a double balance mixer were 11 dB and 10.5 dBm, respectively. The 3-dB bandwidth at LSB output was found to be 3.6 GHz. The measurement results showed sufficient performance for a 60-GHz-band TX module.
    No preview · Conference Paper · Jan 2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: High frequency characteristics of any layer interstitial via hole (ALIVH) substrate have been evaluated for an application to the RF modules on 60GHz. A dielectric constant and loss tangent up to 60GHz was measured using strip line resonator designed on the ALIVH substrate. Measured dielectric constant 3.85-3.47 and loss tangent of 0.032-0.043 were almost constant from 2.5GHz to 60GHz. Equivalent lumped component circuit of via hole was modeled from measurement of frequency of strip line resonator with via hole. T-type LC circuit was employed for the circuit. The evaluated inductance was 0.090-0.128 nH and capacitance was 0.129-0.195 pF. It was confirmed that the high frequency characteristics of ALIVH were sufficient for the application to RF module and 3D SiP with no critical degradation.
    No preview · Conference Paper · Jan 2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: We developed a high speed wireless LAN prototype in the 5GHz band. Maximum transmission rate of the developed WLAN equipment was 324Mbit/sec using 6 multi-channels of 802.11a (1). We implemented a frame aggregation scheme to improve MAC efficiency (2-3). The evaluated system throughput was achieved more than 170Mbit/sec using the implemented frame aggregation. The HDTV transmission using 324Mbit/sec WLAN equipment was demonstrated. We compared the developed prototype and the conventional 802.11g (4).
    No preview · Article · Jan 2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have proposed hotspot network based on layer 2 forwarding for seamless handover. A mobile terminal sends movement notification packet to new access point when the terminal changed connection from previous access point to the new access point. All nodes which received the movement notification packet change address table. The previous access point sends buffered packets to the new access point. We have evaluated performance of the proposed network using network test bench. IEEE802.11a modems were used as wireless interfaces. Measured delay time depending on the buffered packet forwarding was confirmed negligibly small. No packet loss was measured. The degradation of throughput at handover was suppressed to compare with that of conventional method without packet forwarding
    No preview · Conference Paper · Nov 2006
  • [Show abstract] [Hide abstract]
    ABSTRACT: For a high-density multilayer substrate ALIVH, balanced pair transmission lines on which 60-GHz signals can be transmitted are proposed. First, it is shown that transmission of millimeter wave signals is difficult due to impedance mismatch of the vias in the conventional unbalanced transmission lines. Further, it is shown that transmission at up to 90 GHz is possible by a design such that the microstrip line on the surface, the strip line in the inner layer, and the vias connecting each layer have the same transmission line spacing and the same characteristic impedance. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 2, 89(10): 1–8, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjb.20249
    No preview · Article · Oct 2006 · Electronics and Communications in Japan (Part II Electronics)
  • [Show abstract] [Hide abstract]
    ABSTRACT: New frequency compensation circuit has been proposed using preamble symbols and some pilot symbols. The proposed circuit had a performance of compensation with the offset of 50 ppm at 2.4 GHz carrier. Initial phase detection and course frequency compensation was achieved from 16-symbol preamble. Minimum mean square scheme was used for evaluation of initial phase and frequency error from phase information of preamble symbols. Accurate compensation of residual frequency offset was operated from the phase information of pilot symbols inserted placed in information payload. Wireless packet modem has been implemented on field programmable gate array (FPGA). Conversion from rectangle coordinates function to polar coordinate function was employed for simple signal processing. Division calculator by bit shifts and lookup table by read only memory (ROM) were used for compact circuit design. Degradation of measured BER from ideal performance was confirmed to be less than 2.5 dB
    No preview · Article · Jan 2006
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have proposed hotspot network based on layer 2 forwarding using movement notification packet for seamless handover. Descriptions of address table for the layer 2 forwarding were renewed at the time of receiving the movement notification packet, which is sent from mobile terminal at changing connection. In order to receive several packets buffered in previous access point, we propose the mobile terminal with two receivers; main and sub receivers. The mobile terminal receives buffered packets in the previous access point using sub receiver. We evaluated throughput of the mobile terminal with dependence of handover frequency. IEEE802.11b modems were used as wireless interfaces. From the experimental results, degradation of throughput with handover was negligibly small compared with throughput without handover
    No preview · Article · Jan 2006
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have investigated the surface acoustic wave (SAW) properties of atomically flat-surface (0001)aluminum nitride on a (0001)sapphire (AlN/alpha-Al2O3) combination. SAW propagated along [1\bar{2}10]AlN/[1\bar{1}00]alpha-Al2O3 and [10\bar{1}0]AlN/[11\bar{2}0]alpha-Al2O3. SAW velocity was measured to be approximately 1.0% higher than that of a conventionally calculated curve. The dispersion of SAW velocity as a function of normalized thickness (kH) was as low as 1.3%. The measured temperature coefficient of delay was 9 ppm/°C at kH values of 5.9 and 9.9. The propagation loss in the case of using an atomically flat-surface AlN film was lower by one order of magnitude than that using a conventional AlN film. The propagation loss at 5.172 GHz was measured to be 0.0053 dB/lambda.
    No preview · Article · Jun 2005 · Japanese Journal of Applied Physics
  • [Show abstract] [Hide abstract]
    ABSTRACT: A high c-axis-oriented and atomically flat-surface aluminum nitride (AlN) film has been successfully deposited on a (0001)sapphire substrate by metalorganic chemical vapor deposition. We evaluated the dependences of surface roughness, tilted mosaicity and twisted mosaicity on the conditions of AlN deposition. It was found that the atomically flat-surface AlN film was deposited under the diffusion-limited area with suppression of vapor phase reaction at a substrate temperature of 1200°C and a V/III ratio of 800. It was also recognized that surface roughness was controlled by gas flow velocity, which is determined by both gas flow rate and pressure in reactor. Mean surface roughness (Ra) of the deposited AlN films was approximately 1 Å. The full width at half maximum of X-ray rocking curve for (0002) and (10\bar{1}2)AlN were approximately 100 and 2300 arcsec, respectively.
    No preview · Article · May 2005 · Japanese Journal of Applied Physics
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have proposed new carrier frequency offset compensation scheme for wireless communication system with burst transmission. Key functions of the proposed scheme are pilot signals insertion into information block and polar-coordinates conversion scheme using read only memory. The proposed compensation scheme was performed by single packet under burst transmission. The result of computer simulation showed that degradation of Eb/No performance from theoretical limit at the bit error rate (BER) of 103 was found to be less than 0.2 dB. We implemented modem with the proposed phase compensation scheme. Measured result of the implemented modem showed that degradation from theoretical limit at the BER of 103 was found to be less than 2 dB
    No preview · Article · Jan 2005
  • Tadashi Takagi · Kazuo Tsubouchi
    [Show abstract] [Hide abstract]
    ABSTRACT: The activities of the IT-21 Center and the set up of the IT-program RR2002 of the national project for the development of mobile wireless communication technologies are introduced, and the successful results are shown as well. Then, we discuss some issues related to realizing millimeter wave broadband mobile communication systems and we make some proposals to facilitate the development of device technologies.
    No preview · Conference Paper · Jan 2005
  • Hiroyuki Nakase · Tsuyoshi Sagitani · Kazuya Masu · Kazuo Tsubouchi
    [Show abstract] [Hide abstract]
    ABSTRACT: We have proposed a theoretical design method for the lower boundary of supply voltage using the universal E-b/N-0-bit error rate (BER) characteristics from the system viewpoint. In order to achieve the system error rate of less than 10(-10) with the conditions of 10 years term, 10 GHz clock and 10(6) gate, the bit error rate of a gate should be less than 10(-34). For a BER of 10(-34), V-DD should be larger than 0.1 V for 0.13 mum complementally metal-oxide-silicon (CMOS) LSI. For 0.05 mum CMOS of next generation, VDD should be larger than 0.2 V under the ideal thermal noise environment. The measured E-b/N-0-BER characteristics of CMOS circuits as a function of supply voltage have agree well with the theoretical value.
    No preview · Article · Oct 2003 · Japanese Journal of Applied Physics
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have proposed a switched-current (SI) analog programmable finite-impulse-response (FIR) filter for software-defined radio. The programmable tap coefficient circuit for programmable operation of the SI FIR filter has also been proposed. The 4-bit-resolution tap coefficient circuit has been designed and fabricated using 0.35 mum complementary-metal-oxide-semiconductor technology. The current error between measured and calculated tap coefficient circuits has been less than 2%. The radio frequency/intermediate frequency SI band-pass filter has been designed using the measured tap coefficients. The designed SI FIR filter has a frequency response as well as calculated tap coefficients. The maximum operating frequency of the proposed SI FIR filter using 0.13 mum CMOS technology has been estimated to be greater than 10 GHz with circuit simulation.
    No preview · Article · Apr 2003 · Japanese Journal of Applied Physics
  • Seong-Kweon Kim · Jae-Sang Cha · Hiroyuki Nakase · Kazuo Tsubouchi
    [Show abstract] [Hide abstract]
    ABSTRACT: We propose a design method of fast Fourier transform (FFT) large-scaled integration (LSI) for an orthogonal frequency division multiplexing (OFDM) system. The proposed FFT LSI is designed with simple current mode circuits, wired-OR connection and gate-width-ratioed current mirror. For a low power consumption, two design methods are proposed, current-cut (CC) and rounding process. The CC operation is performed with on-off operation of a current source. The rounding process is adopted for the component of the FFT matrix. Bit error rate (BER) in the OFDM system is simulated from “0.1” rounding step to “1.0”. The BER performance of the FFT matrix with the “0.2” rounding step has little degradation from that of the original FFT matrix. The 8-point FFT LSI with the “0.2” rounding step is designed and implemented using a 0.8 μm complementary metal-oxide semiconductor (CMOS) technology. On the basis of measurement of the 8-point FFT LSI, the power consumption of the 64-point FFT LSI using CC can be estimated as being less than 10 mW.
    No preview · Article · Apr 2001 · Japanese Journal of Applied Physics
  • [Show abstract] [Hide abstract]
    ABSTRACT: A silicon Class-B complementary push-pull power amplifier for a code-division-multiple-access IS-95 cellular system has been proposed for the first time. 0.8-µm complementary metal-oxide-semiconductor (CMOS) devices are fabricated and evaluated using RF measurement and simulation. The RF-CMOS balanced design between n- and p-MOS devices has been investigated. From RF simulation, a power-added efficiency (PAE) of over 60% at an input power of 5 dBm and an adjacent-channel-power-ratio of -42 dBc are obtained. It is found that both high efficiency and high linearity are obtained using the RF-CMOS balanced design with optimized channel length and width. Furthermore, drain voltage control for IS-95 criteria has been proposed in addition to the balanced design. The overall PAE, which is defined as the PAE multiplied by the probability distribution function of the used power-level and integrated over the entire used power range, has been improved up to 49%.
    No preview · Article · Apr 2000 · Japanese Journal of Applied Physics
  • Kenji Togura · Koji Kubota · Hiroyuki Nakase · Kazuya Masu · Kazuo Tsubouchi
    [Show abstract] [Hide abstract]
    ABSTRACT: We have proposed the current-cut switched-current matched filter (CC-SIMF) with a low power consumption of less than 10 mW for mobile terminals in direct-sequence code-division-multiple-access. The parallel SIMF configuration is proposed for the reduction of the current transfer error accumulation in the conventional serial SIMF. In order to evaluate the fundamental operation of the parallel SIMF, we have designed and fabricated a 32-chip parallel SIMF using 0.8-µm complementary-metal-oxide-semiconductor (CMOS) technology. The autocorrelation performance of a 32-chip orthogonal m-sequence has been clearly observed to be more than 5 Mcps. The measured dynamic range is improved to 12 dB as compared with that of the conventional SIMF, even though each current memory cell has a current transfer error of over 10% per chip. Since the current sources in the parallel SIMF can be reduced to be one-third of those in the serial SIMF, the power consumption of the parallel SIMF can be reduced to one-third of that of the serial SIMF.
    No preview · Article · Mar 2000 · Japanese Journal of Applied Physics