Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 3(6): 929-933 Scholarlink Research Institute Journals, 2012 (ISSN: 2141-7016) jeteas.scholarlinkresearch.org Journal of Emerging Trends in Engineering and Applied Sciences (JETEAS) 3(6):929-933(ISSN: 2141-7016) A Impulse Radio Ultra-Wideband Transceiver for Inter/Intra-chip Wireless Interconnection Nhan Nguyen 1, Nghia Duong 2, Anh Dinh 3 1 Faculty of Physics and Engineering Physics, University of Science-VNU Hochiminh City, Vietnam; 2 Faculty of Electrical and Electronics Engineering, University of Technology, VNU Hochiminh City, Vietnam; 3 Dept. of Electrical and Computer Engineering, University of Saskatchewan, Saskatoon, Canada. Corresponding Author: Nhan Nguyen Abstract This paper presents a impulse radio ultra-wideband (IR-UWB) transceiver designed and implemented using IBM 0.13um technology for inter-chip and intra-chip wireless interconnec-tion. The IR-UWB transmitter is based on the high order derivative Gaussian pulse using LC-Tank oscillator. A pulse for the 6.3-10GHz IR-UWB transmitter generated with low power, high peak amplitude, and high frequency. This pulse PSD fully complies with the FCC spectrum mask. The IR-UWB receiver is based on the non-coherent architecture which removes the complexity of RF architecture (such as DLL or PLL) and reduces power consumption. The receiver consists of three blocks: a low noise amplifier (LNA) with active balun, a correlator, and a comparator. Simulation results of the IR-UWB transmitter show a pulse width of 1.1ns, a peak to peak amplitude pulse of 90mV and an energy consumption of 39pJ/pulse at 200MHz pulse repetition rate (PRR). The receiver provides a power gain (S21) of 12.5dB, a noise figure (NF) of 3.05dB, an input return loss (S11) of less than -16.5dB, a conversion gain of 18dB, a NF DSB of 22, and a third order intercept point (IIP3) of -1.3dBm. The transceiver uses an area of 0.38mm 2 and consumes 30.7mW of power on the 1.4V power supply. The results show a promising transceiver for wireless communications applicable to 3-D ICs due to its low complexity and small chip area. Keywords: IR-UWB transmitter, IR-UWB receiver, inner-chip/inter-chip wireless interconnection, Gaussian pulse, non-coherent receiver INTRODUCTION Traditional wire interconnect systems using metal are projected to be limited in their ability to meet the future interconnect needs (Brian A. Floyd et al., 2001). The limitation of the traditional wire interconnects can be viewed from four aspects, namely, interconnect resistance, interconnect capacitance, interconnect inductance, and bit-rate capacity (James D. Meindl et al., 2001; Jeffery A. Davis et al., 2001; A. Deutsch et al., 2001). The interconnect resistance is calculated as R interconnect = R wire + R contact + R via (1) With interconnect resistance increasing, more and more interconnect wires are becoming RC transmission lines with RC delay constant, which can be seen as τ delay = R interconnect x C interconnect (2) To address this problem, inter-chip and intra-chip wireless interconnects using microwaves are being evaluated (Takahide Terada et al., 2006; Yuanjin Zheng et al., 2006;). In this paper, we propose a transceiver architecture to be used as wireless interconnect system using ultra-wideband (UWB) technology. The technology employs ultra-wide bandwidth and very low emission power density. UWB can be potentially used in low cost, low power, and short range high-speed communication applications. In 2002, the Federal Communications Commission (FCC) has released unlicensed 3.1-10.6GHz frequency band for UWB related applications. The UWB transmission is defined as the occupied fraction band-width of >20% or larger than 500MHz of the absolute bandwidth and the power spectrum density (PSD) of the pulse must be below - 41.3dBm/MHz (FCC, 2002). In recent years, a number of research works focused on inter-chip and intra-chip wireless interconnections by using IR-UWB communi-cations due to its low complexity and low power (Tiuraniemi et al., 2005; Yuanjin Zheng et al., 2005; S. Jeong-Bae et al., 2008; Yuanjin Zheng et al., 2006). In this work, a fully integrated transmitter and receiver based (TX- RX) on the UWB technique were designed and implemented in a 0.13um technology. The implemented TX-RX can achieve a transmission rate 929
of 100-200Mbps in noncoherent mode and consumes a minimum amount of power. The layout area is as small as 0.38mm 2. IR-UWB TRANSCEIVER OVERVIEW Figure 1 is a simplified block diagram of the proposed IR-UWB transceiver. for the loss in the tank. The oscillation frequency of this circuit can be found as: 2 1 1 R C (3) LC L Figure 2. LC-tank model. To generate a negative conductance, a balanced NMOS circuit topology shown in Figure 3 is chosen for this work. In practical integrated LC-tank, the inductors are on-chip spiral inductors with low quality factor that dominates the loss of the LC-tank. Figure 1. Block diagram of the proposed transceiver In IR-UWB systems, the receiver is an important component and is more complex than the transmitter. A high order derivative Gaussian pulse generator using a LC tank is proposed for the transmitter. This generator is intended to be used in the upper UWB frequency band of 6-10GHz. PSD of the generated pulse fully complies with the FCC spectrum mask for indoor applications. In addition to the spectrum requirement, wireless applications require outstanding noise (phase noise and jitter) performance at high frequency, LC tank is the component of choice in the design. The receiver consists of a low-noise amplifier (LNA) with an active balun, a correlator, and a comparator. There are basically two different types of receivers employed in IR-UWB systems namely coherent receiver and non-coherent receiver. An IR-UWB based non-coherent single chip receiver was presented (Tiuraniemi et al., 2005; Yuanjin Zheng et al., 2005). The receiver covers the high frequency band 6.0-10GHz of the UWB system. The received UWB signal is amplified through a low-noise amplifier (LNA) with an active balun. This amplifier provides a constant gain optimized around the operation frequency of 8.0 GHz. After being first amplified by the LNA, the received pulse is then selfcorrelated by a correlator. The received signal is then sent to a comparator for digital quantization. Figure 3. Balanced NMOS Figure 4. Schematic diagram of the LC-tank based high order derivative Gaussian pulse generator IR-UWB TRANSMITTER CIRCUIT DESIGN The proposed high order derivative Gaussian pulse generator was designed using a LC-tank oscillator and implemented in IBM 0.13μm technology with 1.4V supply voltage. Figure 2 shows the LCtank model, where the conductance g tank represents the tank loss and -g active is the effective negative conductance of the active devices that compensates 930 A completed circuit of the high order derivative Gaussian pulse generator is shown in Figure 4. This circuit comprises three stages. The first stage is a chain of inverters that is used to delay the input signal in the order of picosecond. The second stage is a XOR gate that used to generate a short pulse. The final stage is a LC-tank that is used to generate a high order derivative Gaussian pulse. The frequency
selective tank consists of two inductors in parallel with capacitor C 4, C 5. Resistor R represents the inductor equivalent series resistance. The crossedcouple pair of nmos M 11,12 is biased with the current I/2, implemented with nmos M 13 in the saturation region. The crossed-couple pair generates a negative resis-tance of -2/gm 11,12 to compensate for the resistance R, where g m11,12 are the transconductances of the transistors M 11,12. There are two complementary short pulses generated by the inverter chain M 1 -M 10 and the XOR gate. Capacitors C 1 -C 3 and the XOR gate control the switches SW 1,2. When SW 1 is open (i.e., SW 2 is closed), the LC tank starts oscillate due to noise. The amplitude of the oscillating signal increases gradually within a certain period. When SW 1 is closed (i.e., SW 2 is open), the energy inside the LC tank is released through SW 1, and the oscillation signal gradually disappear. IR-UWB RECEIVER CIRCUIT DESIGN The proposed IR-UWB receiver employs the noncoherent receiver architecture. The receiver consists of a low-noise amplifier (LNA) with an active balun, a correlator, and a comparator. A. LNA and Active Balun: The UWB LNA design uses resistive feedback current reuse technique. The design was implemented in the IBM 0.13um technology with appropriate impedance matching and noise/power optimizations. The LNA achieves up to 12.5dB power gain with a noise figure of 3.05dB over the UWB 3.1-10.6GHz frequency range. Figure 5 shows the schematic of the designed LNA. Figure 5. Schematic diagram of the LNA The input inductor L 1 is added to compensate for the parasitic capacitors of M 1 and M 2 transistors at high frequencies. A resistive feedback buffer with peaking inductor load is coupled with the reuse UWB stage through capacitor C 3 to drive the 50Ohm output load. In addition, in the resistive feedback current reuse configuration (the first stage), loading the NMOS transistor M 1 with the PMOS transistor M 2 allows the circuit to operate under lower supply voltage than resistive load configuration. and the input of the second cascode. Since v gs5 =-v gs6, two balanced differential outputs can be achieved on the condition of g m5 =g m6 Figure 6. Schematic diagram of the active balun B. Correlator: The output of the LNA must be correlatively multiplied and then integrated in order to detect the energy of the received signal. Figure 7 shows the block diagram of correlator and equation (4) shows the output signal of correlator. Figure 7. Block diagram of the correlator t 0 0 ( ). template ( ) t0 T r t r t S t dt (4) where r 0 (t) is the output signal of the correlator, r(t) is the input signal of the correlator and S template (t) is the template signal. Figure 8 shows the schematic diagram of the correlator. The correlator employs the Gilbert mixer topology, and the integrator is realized by capacitors C1 and C2. After the pulse is mixed with itself, the integrator begins to integrate; between the pulses intervals, the integrator discharges energy and is ready for the next integration. The capacitances of C1 and C2 should be large enough to hold the integrated voltage for the comparator and yet small enough to discharge between pulses intervals in order to be ready for the next integration cycle. The transmission data rate relies on the speed of the integration process. A two-cascode stage active balun of Figure 6 is used to convert the single-ended output of the LNA to differential signals. The output of M4 connects to M6 931 Figure 8. Schematic diagram of the correlator
C. Comparator: After the received signal is squared and integrated by the correlator, a comparator compares it with a reference voltage and performs digital quantization. However the comparator output is a return-to-zero (RZ) signal which needs to be converted to a non-return-to-zero (NRZ) signal that can synchronize with the baseband clock. In a coherent receiver, a DLL/PLL is usually introduced to perform synchronization between the received pulse and the local pulse. This requires a precision on the order of several tens of picoseconds. However, in a noncoherent receiver, the RZ signal quantized by the comparator exhibits a duty cycle on the order of nanoseconds. Therefore, a low jitter DLL/PLL is no longer necessary and a sliding correlator is employed. The technique reduces complexity of the receiver. The topology of the comparator consists of differential amplifier with buffering inverter and common-source amplifier. The block diagram and the schematic diagram of the comparator are shown in Figure 9 and Figure 10, respectively. Figure 10. Gaussian pulse FCC Mask Figure 11. Normalized PSD of Gaussian pulse Figure 9. Block diagram of the comparator SIMULATION RESULTS The IR-UWB transceiver was designed and simulated using Cadence tools applicable in IBM 0.13um technology. Figure 10 and Figure 11 show the simulation results in time domain and the normalized PSD of the generated high order derivative Gaussian pulse with 50Ω load (for antenna matching purpose). The peak-to-peak swing is 90mV. For wireless communications at short distances, this pulse amplitude is sufficient to deliver the pulse to the antenna without the need for a wide band amplification. The pulse duration is 1.1ns exhibits the high spectral allocation in the 6.3 10GHz range, shown in Figure 10. Figure 11 shows the power spectral density of this pulse. Figure 12 shows the layout view of the IR-UWB transceiver. The area of the transceiver is about 0.38mm 2. The power consumption of the trans-ceiver is 30.7mW for the 1.4V supply voltage. Table 1 shows the performance parameters of the designed IR-UWB transceiver compared with other published works. Figure 12. Layout view of the IR-UWB Transceiver 932
Table 1: Performance parameters summary and comparison with the other transceivers Parameters Technology This work 0.13um Yuanjin Zheng (2005) Bandwidth (GHz) 6.3-10 3.1-5 Takahide Terada (2006) 250MHz F c=400 MHz Pulse Width (ns) 1.1 0.8 2 0.7 V pp (mv) 90 35 100 20 Energy(pJ/pulse) 39 - - - S11 (db) < -16.5 - - - Gain Max(dB) (S21) 12.5 18 16 - NF (db) 3.05 7.5 5.6 - Conversion Gain(dB) 18 5 - - IIP3 (dbm) -1.3-14 - - M. Anis (2007) 6-8.5 Data rate(mbps) 200 50 1 4MHz Power (mw) 30.7 41.4 ~9.2 2.6 Chip Area(mm 2 ) 0.38 4.67 0.42 3.15 CONCLUSION In this work, a IR-UWB transceiver was designed and simulated using Cadence tool with IBM 0.13um technology for inter-chip and intrachip interconnection. The results show a promising transceiver for wireless communications applicable to 3-D ICs due to its low complexity and small chip area. The simulation shows a data rate of 200Mbps can be achieved. However, the transceiver circuit also has high power consumption due to the tail current needed in the transmitter architecture. REFERENCES A. Deutsch, Paul W. Coteus, Gerard V. Kopcsay (2001): On-chip wiring design challenges for Gigahertz operation, Proceedings of the IEEE, Vol. 89, No. 4., 2001. Federal Communications Commission (2002): First report and order, revision of part 15 of the commission s rules regarding ultra wideband transmission system, Washington, DC, ET Docket 98-153, 2002. J.A. Davis, R. Venkatesan, A. Kaloyeros, M. Bylansky, S.J. Souri, K. Banerjee, K.C. Saraswat, A. Rahman, R. Reif, and J.D. Meindl, (2001): Interconnect Limits on Gigascale Integration (GSI) in the 21st Century, Proceedings of the IEEE, vol. 89, no. 3, pp. 305-324, March 2001. S. Jeong-Bae, K. Jong-Ha, S. Hyuk, and Y. Tae- Yeoul, (2008): "A Low-Power and High-Gain Mixer for UWB Systems," Microwave and Wireless Components Letters, IEEE, Vol. 18, pp. 803-805, 2008. Takahide Terada, Shingo Yoshizumi, Muhammad Muqsith, Yukitoshi Sanada, and Tadahiro Kuroda (2006): A Ultra-Wideband Impulse Radio Transceiver for 1-Mb/s Data Communications and ±2.5-cm Range Finding, IEEE Journal of Solid-State Circuits, Vol. 41, No. 4, 2006. Tiuraniemi, S., Stoica, L.; Rabbachin, A., and Oppermann, I., (2005): Front-end receiver for low power, low complexity non-coherent UWB communication system, IEEE International Conference on Ultra-Wideband, pp. 339-343, 2005. Yuanjin Zheng, Yan Tong, Jiangnan Yan; Yong-Ping Xu, Wooi Gan Yeoh, and Fujiang Lin, (2005): A low power non-coherent UWB transceiver ICs, IEEE Radio Frequency Integrated Circuits (RFIC) Symposium, Digest of Papers, pp. 347-350, 2005. Yuanjin Zheng, Yueping Zhang, and Yan Tong, (2006): A Novel Wireless Interconnect Techno-logy Using Impulse Radio for Interchip Communications, IEEE Transactions On Micro-wave Theory And Techniques, Vol. 54, No. 4, 2006. M. Anis and R. Tielert (2007): Design of UWB Pulse Radio Transceiver Using Statistical Correlation Technique In Frequency Domain, Advances in Radio Science, 5, 297-304, doi:10.5194/ars-5-297- 2007. James D. Meindl, Raguraman Venkatesan, Jeffrey A. Davis, James Joyner, Azad Naeemi, (2001): Interconnecting Device Opportunities for Gigascale Integration (GSI), Electron Devices Meeting 2001 (IEDM '01) Technical Digest International. 933