A 16-GHz CMOS differential Colpitts VCO for DS ... - Wiley Online Library
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Microwave Millimeter-Wave Monolithic Circuits Symposium, Albuquerque, NM 1992, pp. 79 – 82. M. Gillick, I.D. Robertson, and J.S. Joshi, Design analysis of novel coupling structures for multilayer MMIC’s, IEEE Trans Microwave Theory Technol 41 (1993), 346 –349. S. Banba and H. Ogawa, Multilayer MMIC directional couplers using thin dielectric layers, IEEE Trans Microwave Theory Technol 43 (1995), 1270 –1275. H. Okazaki and T. Hirota, Multilayer MMIC broad-side coupler with a symmetric structure, IEEE Microwave Guided Wave Lett 7 (1997), 145–146. K. Price, Differential evolution: A fast and simple numerical optimizer, In: Proceedings of the IEEE North American Fuzzy Information Process Conference, Berkeley, CA, 1996, pp. 524 –527. R. Storn and K. Price, Differential evolution: A simple and efficient heuristic for global optimization over continuous spaces, J Global Optim 11 (1997), 341–359. K.A. Michalski, Electromagnetic imaging of circular cylindrical conductors and tunnels using a differential evolution algorithm, Microwave Opt Technol Lett 27 (2000), 330 –334. K.A. Michalski, Electromagnetic imaging of elliptical cylindrical conductors and tunnels using a differential evolution algorithm, Microwave Opt Technol Lett 28 (2001), 164 –169. A. Qing, Electromagnetic inverse scattering of multiple two-dimensional perfectly conducting objects by the differential evolution strategy, IEEE Trans Antennas Propagat 51 (2003), 1251–1262. K. Guney and D. Karaboga, New narrow aperture dimension expressions obtained by using a differential evolution algorithm for optimum gain pyramidal horns, J Electromagn Waves Appl 18 (2004), 321–339. X.F. Luo, P.T. Teo, A. Qing, and C.K. Lee, Design of double-squareloop frequency-selective surfaces using differential evolution strategy coupled with equivalent-circuit model, Microwave Opt Technol Lett 44 (2005), 159 –162. X.F. Luo, A. Qing, and C.K. Lee, Application of the differentialevolution strategy to the design of frequency-selective surfaces, Int J RF Microwave CAE 15 (2005), 173–180. C. Yildiz, A. Akdagli, and M. Turkmen, Simple and accurate synthesis formulas obtained by using a differential evolution algorithm for coplanar strip lines, Microwave Opt Technol Lett 48 (2006), 1133–1137.
© 2007 Wiley Periodicals, Inc.
A 16-GHZ CMOS DIFFERENTIAL COLPITTS VCO FOR DS-UWB AND 60GHZ DIRECT-CONVERSION RECEIVER APPLICATIONS C.-C. Lee,1 H.-R. Chuang,1 and C.-L. Lu2 Institute of Computer and Communication, Department of Electrical Engineering National Cheng Kung University, Tainan, Taiwan, Republic of China, Corresponding author: chuang [email protected] 2 Department of Computer and Communication, Kun Shan University, Tainan, Taiwan, Republic of China
1
Received 13 March 2007 ABSTRACT: A 16-GHz CMOS differential Colpitts VCO fabricated with the 0.18 m 1P6M process is presented. The 16-GHz VCO is a good choice for the local oscillator (LO) circuit of the UWB or 60-GHz WPAN direct conversion receiver. The VCO is composed of a PMOS transistor-pair core circuit and two source follower output buffers. The VCO can operate at 16.5 GHz, and the measured phase noise at 1-MHz offset is ⫺115 dBc/Hz. The power consumption of the VCO core is 12.6 mW. Compared with previous reported works, this VCO has an output power of ⫺0.9 dBm and about 800-mV output peak-to-peak voltage swing of the VCO core at 16.5 GHz. © 2007 Wiley Periodicals, Inc.
DOI 10.1002/mop
Microwave Opt Technol Lett 49: 2489 –2492, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop. 22742 Key words: 0.18 m; 16 GHz; 60-GHz WPAN; CMOS; differential Colpitts VCO; UWB
1. INTRODUCTION
Recently, ultra-wideband (UWB) radio has been proposed for physical layer standard of the future high-speed wireless personal area networks (WPANs). There are two main proposals, the multiband orthogonal frequency division multiplexing (MB-OFDM), and the direct-sequence ultra-wideband (DS-UWB). DS-UWB supports two independent bands of operation. They are the lower band that occupies the spectrum from 3.1 to 4.85 GHz, and the upper band that occupies the spectrum from 6.2 to 9.7 GHz. The lower band offers the piconet channel number 1– 6 and the upper band offers that of 7–12 [1]. Moreover, for dense local communications, the 60-GHz band with a band-width of about 8 GHz is of special interest to the short range (⬍1 km) communications [2]. It is due to its attenuation characteristic of 10 –15 dB/km by the atmospheric oxygen. For the UWB or 60-GHz WPAN direct conversion receiver, the 16-GHz frequency is a good choice for the local oscillator (LO). For the DS-UWB direct conversion receiver, a 16-GHz LO will be transformed into an 8-GHz or 4-GHz LO simply by a divide-by-2 ordivide-by-4 circuit. Also, a 16-GHz LO with a four-time multiplier can be used for a 60-GHz WPAN receiver application. For a CMOS mixer, the LO voltage swing directly affects the conversion gain and the noise figure performance. The switching pair of the mixer requires much greater voltage swing to complete switching. If the VCO output voltage swing is not high enough, differential LO amplifiers should be added to the VCO output, and will consume additional DC power. However, in order to achieve a high frequency operation, the value of the inductor of the tank has to be reduced. While the inductance is decreased, the parallel resistance (Rp) of the tank also decreases for a given quality factor (Q) and also the degrade of the output swing with a fixed bias current [3]. Moreover, the output swing of the VCO core and the phase noise performance are usually related. Therefore, it is desired to design a VCO with a moderate output swing at high frequency and with a low phase noise. The most popular topology of a VCO is the cross-coupled structure [4 – 6]. Compared with the NMOS- or PMOS-only and complementary cross-coupled oscillators, the Colpitts oscillator presents a smaller RMS and DC value of its effective impulse sensitivity function (ISF) [7, 8]. In [7], a 1.8-GHz NMOS differential Colpitts VCO with a phase noise of ⫺138.2 dBc/Hz at 3-MHz offset was first presented. It also shows that the Colpitts oscillator possesses a larger output voltage swing and higher energy transfer efficiency while comparing with the cross-coupled ones. In [9], a Colpitts VCO shows that by adopting only the PMOS to construct the core circuit of the oscillator would achieve a better phase noise performance since the PMOS flicker noise is lower than that of the NMOS. The measured phase noise of the demonstrated Colpitts VCO is ⫺120.4 dBc/Hz at 1-MHz offset while working at 5-GHz. It also achieves a good FOM value of about 190 and a DC-to-RF conversion efficiency of better than 28% with the open-drain buffer. In [10], a reported 20-GHz VCO in a 0.18-m standard CMOS process exhibits an output power of ⫺3 dBm, phase noise of ⫺111 dBc/Hz at 1-MHz offset, and FOM value of 182, respectively. The frequency tuning range is 510MHz and the power consumption is 32 mW.
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
2489
Figure 2 Simulation and measurement of the 16-GHz VCO output frequency versus control voltage Figure 1 VCO
Circuit schematic of a 16-GHz PMOS differential Colpitts
In this article, a 16-GHz PMOS differential Colpitts VCO (see Fig. 1) is designed by using bias-T source followers [11]. The VCO is fabricated in a 0.18-m standard CMOS process. The design goals of this high frequency VCO are to have a simple circuit scheme without needing a complicated EM-simulation, and achieve a high output voltage swing and a good FOM value. Although the small inductor and the parallel resistance in a high frequency VCO tank circuit will degrade the output voltage swing, this VCO with a source follower buffer still achieve a moderate output power of ⫺0.9 dBm and about 800-mV output peak-to-peak voltage swing of the VCO core. 2. CIRCUIT DESIGN
Figure 1 shows the schematic of a 16-GHz PMOS differential Colpitts VCO. Only the PMOS transistors are used in the circuit. The core of the VCO consists of a current switching pair and two symmetrical single-ended Colpitts VCO which is individually connected to a varactor diode. The PMOS cross-coupled pair of the current switching also provides the negative resistance to the Colpitts oscillators to improve the oscillating conditions, especially the start-up condition. Traditionally, a transistor constant current source is necessary for the bias in a differential structure. The current source works as a DC short circuit to offer a fixed bias current to the differential pair circuit, while offering a high AC impedance to avoid a loading effect. However, in order to avoid the reduction of the voltage headroom, the transistor constant current source is replaced by the on-chip RF choke and the by-pass capacitors to enlarge the voltage swing. Without the transistor constant current source, the bias current was directly controlled by choosing the transistor size. To filter out the high frequency noise, such as the 2nd harmonic (⬃32 GHz) at the common mode VDD node, the on-chip by-pass capacitors are used. Two on-chip symmetrical inductors with a centertap are used as the RF choke and the tank. Generally, the open drain circuits are used as the buffers of the VCO for measurement purpose. It can offer not only a sufficient ` load of the measurement instrument, and current for driving 50- U also provide a voltage gain to achieve a higher output voltage swing. However, for a 16-GHz VCO with the 0.18-m standard CMOS technology (unit-gain frequency fT is 60 GHz), it is not easy to amplify the large voltage swing from the VCO core without any matching network. Since the source follower output
2490
buffer does not provide any voltage gain, the output voltage swing of the VCO core circuit should be higher than the measured one through the buffer. Also, the inductive choke inside the bias-T provides a high ac impedance and good stability for the buffer. Therefore, small-size transistors can be used to provide enough output current drive without loading the tank heavily. However, long metal trace on the lossy CMOS substrate also degrades the output power seriously. Hence, the connecting metal trace between VCO core and output buffer should be as short as possible to reduce the parasitic effect. 3. MEASURED PERFORMANCE
As shown in Figure 2, the measured oscillation frequency of the VCO covers from 15.63 to 16.5 GHz with control voltage from ⫺1.8 to 1.8 V. Figure 3 shows the measured output power which ⫺0.9 dBm at 16.5 GHz after deembedding the cable and bias-T loss. The free running measurement of the phase noise is ⫺115 dBc/Hz at 1-MHz offset from a 16.5-GHz carrier as shown in Figure 4. The VCO core and two output buffers dissipate 12.6 and 30 mW, respectively. Here it should be noted that the bias-T source-follower output buffer is only for 50-⍀ system measurement purpose. The measured output power is ⫺0.9 dBm at 16.5 GHz. Since the source
Figure 3 Simulation and measurement of the 16-GHz VCO output power versus control voltage
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
DOI 10.1002/mop
Figure 4 Measured free running phase noise. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]
follower buffer has a voltage loss of 3 dB, the output peak-to-peak voltage amplitude of this VCO core will be about 800 mV from the measured power of ⫺0.9 dBm at the 50-⍀ load of the source follower buffer. Figure 5 shows the chip micrograph with a die area of 0.558 mm2. A widely used figure of merit (FOM) for VCO is defined as [12] FOM共dB兲 ⫽ 10 log
冉冉 冊 f0 fm
2
䡠
冊
1 , L共 fm兲 䡠 P
(1)
where L(fm) is the measured phase noise from the oscillator frequency at a frequency offset (fm), and P is the DC power consumption of the VCO. The FOM value is about 188. Table 1 summarizes the measurement results and the comparison with the previously reported VCOs. Again it is noted that the bias-T sourcefollower output buffer is only for 50-⍀ system measurement and has a voltage loss of 3 dB. The output peak-to-peak voltage swing of the VCO core is determined from the measured output power of the buffer by considering the buffer voltage loss. Hence the DC current consumption of the bias-T output buffer is excluded from total DC consumption of the VCO core circuit.
4. CONCLUSION
Figure 5 Chip micrograph (0.6 ⫻ 0.93 mm2). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley. com] TABLE 1
[4] [5] [6] [7] [11] This work
This article presents the design, fabrication, and measurement of a 16-GHz differential Colpitts VCO fabricated in a TSMC 0.18-m standard CMOS process. This 16-GHz VCO can be used for the LO circuit of the UWB or 60-GHz WPAN direct conversion receiver. The core circuit of the VCO consists of a PMOS differential Colpitts oscillator pair and a switching pair. The small DC and RMS value of the ISF and high energy transfer efficiency of the VCO core helps this work achieving a good phase noise performance. An LC tank circuit is used to replace the traditional current source to avoid the headroom limitation of voltage swing. Two bias-T source followers are used to for the measurement of the real output voltage swing. Although the small inductor and the parallel resistance in a high frequency VCO tank circuit will degrade the output voltage swing, this VCO still achieves a moderate output power level and high output voltage swing. The measured output frequency is from 15.63 to 16.5 GHz with a tuning range of about 5.4%. The measured free running phase
Summary of the Performance of a 16-GHz Diffential Colpitts VCO and Comparison with Previous Works Tech.
OSC. Frequency
Phase Noise
Power Consumption
0.18 m CMOS 0.13 m CMOS 0.13 m p-HEMT 0.18 m CMOS 0.18 m CMOS 0.18 m CMOS
10.7 GHz 18 GHz 12.2 GHz 12 GHz 19.9 GHz 16.5 GHz
⫺118.7 dBc/Hz @ 1 MHz ⫺117 dBc/Hz @ 1 MHz ⫺109 dBc/Hz @ 1 MHz ⫺84.8 dBc/Hz @ 100 KHz ⫺111 dBc/Hz @ 1 MHz ⫺115 dBc/Hz @ 1 MHz
11.8 mW 14.4 mW 3 mW 2 mW 32 mW 12.6 mW (VCO core)
DOI 10.1002/mop
Output Power/Core Voltage Swing (p-p)
FOM
⫺6 dBm/⫺10 dBm/⫺7 dBm/-/⫺3 dBm/⫺0.9 dBm/800-mV (with a 3-dB voltage-loss source follower)
188.5 190.5 190.5 183.4 182.0 188.4
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
2491
noise at 1-MHz offset is ⫺115 dBc/Hz at 16.5 GHz. The core circuit consumes 12.6 mW under 1.8-V power supply. The FOM value is about 188. The bias-T source-follower output buffer used for 50-⍀ system measurement has a voltage loss of 3 dB. From the measured output power of ⫺0.9 dBm at 16.5 GHz, the output peak-to-peak voltage amplitude of this VCO core will be about 800 mV by considering the 3-dB output buffer voltage loss. By a careful circuit design, this differential PMOS Colpitts VCO has a simple circuit scheme without needing a complicated EM-simulation, and achieve a high output voltage swing and a good FOM value. ACKNOWLEDGMENT
The authors thank the Chip Implementation Center (CIC) of the National Science Council, Taiwan, ROC, for supporting the TSMC CMOS process. REFERENCES 1. R. Fisher et al., DS-UWB physical layer submission to 802.15 task group 3a, IEEE P802.15– 04/0137r4 (2005). 2. P. Smulders, Exploiting the 60 GHz band for local wireless multimedia access: Prospects and future directions, IEEE Commun Mag 40 (2002), 140. 3. B. Razavi, RF microelectronics, Prentice Hall, Upper Saddle River, NJ 1997. 4. S. Ko, J.-G. Kim, T. Song, E. Yoon, and S. Hong, 20 GHz integrated CMOS frequency sources with a quadrature VCO using transformers, In: IEEE radio frequency integrated circuits (RFIC) Symposium, Fort Worth, TX June 2004, pp. 269 –272. 5. V. Manon, and S.I. Long, A low power and low noise p-HEMT Ku band VCO, IEEE Microwave Compon Lett 16 (2006). 6. T.K.K. Tsang and M.N. El-Gamal, A hign figure of merit and areaefficient low-voltage 12 GHz CMOS VCO, In: IEEE radio frequency integrated circuits (RFIC) symposium, Philadelphia, PA 2003, pp. 89 –92. 7. R. Aparicio and A. Hajimiri, A noise-shifting differential colpitts VCO, IEEE J Solid-State Circuits 37 (2002). 8. A. Hajimiri and T.H. Lee, A general theory of phase noise in electrical oscillators, IEEE J Solid-State Circuits 33 (1998), 179 –194. 9. M.-D. Tasi, Y.-H. Cho, and H. Wang, A 5-GHz low phase noise differential colpitts CMOS VCO, IEEE Microwave Wireless Compon Lett 15 (2005), 327–329. 10. H.-H. Hsieh and L.-H. Lu, A low-phase-noise K-band CMOS VCO, IEEE Microwave Compon Lett 16 (2006). 11. N.H.W. Fong, J.-O. Plaouchart, N. Zamdmer, D. Liu, L.F. Wagner, C. Plett, and N. Garry Tarr, A 1-V 3.8 –5.7-GHz wide-band VCO with differential tuned accumulation MOS varactors for common-mode noise rejection in CMOS SOI technology, IEEE Trans Microwave Theory Tech 51 (2003), 1952–1959. 12. E. Hegazi, H. Sjoland, and A.A. Abidi, A filtering technique to lower LC oscillator phase noise, IEEE J Solid-State Circuits 36 (2001), 1921–1930. © 2007 Wiley Periodicals, Inc.
2492
CALCULATION OF PATTERN IN UTD METHOD BASED ON NURBS MODELING WITH THE SOURCE ON SURFACE Wang Nan, Liang ChangHong, and Yuan HaoBo School of Electronic Engineering, Xidian University, Shaanxi, People’s Republic of China, Corresponding author: [email protected] Received 20 March 2007 ABSTRACT: The UTD method with the source on convex surface which is usually used in the analysis and design of antennas especially conformal arrays on electrically large platforms is studied in this article where NURBS is introduced as way of modeling. Together with the information from ray-tracing done before, the calculation of fields is presented which includes direct field in the lit region and creeping ray field in the shadow region. The method can be applied to arbitrary electrically large targets and the usefulness can be seen in the examples given. © 2007 Wiley Periodicals, Inc. Microwave Opt Technol Lett 49: 2492–2498, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.22727 Key words: NURBS; UTD; source on surface; pattern
1. INTRODUCTION
Within computational electromagnetics, electrically large targets are usually modeled either by meshes of planar patches or by typical objects. Modeling by planar meshes which are usually used in method like MoM is to use lots of planar patches (usually triangular) to simulate the target and the precision of the approached model depends on the number of the planar patches used but the requirement of computational resource raises greatly with the scale of the mesh at the same time. Modeling by typical objects which are usually used in method like UTD is easy to construct and is less consuming but obviously the approached models may not simulates the real ones precisely. In the area of modeling, STEP, the international standard for the geometrical definition of industrial products, was published by the international organization for standardization (ISO) where the nonuniform rational B-spline (NURBS) technique was accepted as the only mathematical way to define shapes of products. While the study of this technique becomes deeper and deeper, it has been already included by more and more commercial CAD/CAM softwares like CATIA, UG, etc. As introduced above, in UTD method, models are approached by typical simple geometric objects but not by arbitrary surfaces, which makes the use of UTD method limited, like the simplified model of an aircraft (Figure 1) that follows, whose wings are always approached by boards and head by cone which may not precisely simulate the real ones. Therefore, UTD method based on arbitrary surfaces modeling is worth studying both practically and theoretically. When analyzing EMC problems of electrically large platforms, UTD method is widely used and reasonably effective. On the other hand, practically the radiation from sources on perfectly conducting convex surfaces is of interest in the design of antennas mounted on aircraft and other electrically large platforms and in the design of conformal arrays, then it is necessary to study UTD under the fact that the source is on surface. The NURBS technique has been introduced into the area of calculational electromagnetics where many focus on the PO method based on NURBS models [1–3]; Perez et al. presented
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
DOI 10.1002/mop
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Microwave Millimeter-Wave Monolithic Circuits Symposium, Albuquerque, NM 1992, pp. 79 – 82. M. Gillick, I.D. Robertson, and J.S. Joshi, Design analysis of novel coupling structures for multilayer MMIC’s, IEEE Trans Microwave Theory Technol 41 (1993), 346 –349. S. Banba and H. Ogawa, Multilayer MMIC directional couplers using thin dielectric layers, IEEE Trans Microwave Theory Technol 43 (1995), 1270 –1275. H. Okazaki and T. Hirota, Multilayer MMIC broad-side coupler with a symmetric structure, IEEE Microwave Guided Wave Lett 7 (1997), 145–146. K. Price, Differential evolution: A fast and simple numerical optimizer, In: Proceedings of the IEEE North American Fuzzy Information Process Conference, Berkeley, CA, 1996, pp. 524 –527. R. Storn and K. Price, Differential evolution: A simple and efficient heuristic for global optimization over continuous spaces, J Global Optim 11 (1997), 341–359. K.A. Michalski, Electromagnetic imaging of circular cylindrical conductors and tunnels using a differential evolution algorithm, Microwave Opt Technol Lett 27 (2000), 330 –334. K.A. Michalski, Electromagnetic imaging of elliptical cylindrical conductors and tunnels using a differential evolution algorithm, Microwave Opt Technol Lett 28 (2001), 164 –169. A. Qing, Electromagnetic inverse scattering of multiple two-dimensional perfectly conducting objects by the differential evolution strategy, IEEE Trans Antennas Propagat 51 (2003), 1251–1262. K. Guney and D. Karaboga, New narrow aperture dimension expressions obtained by using a differential evolution algorithm for optimum gain pyramidal horns, J Electromagn Waves Appl 18 (2004), 321–339. X.F. Luo, P.T. Teo, A. Qing, and C.K. Lee, Design of double-squareloop frequency-selective surfaces using differential evolution strategy coupled with equivalent-circuit model, Microwave Opt Technol Lett 44 (2005), 159 –162. X.F. Luo, A. Qing, and C.K. Lee, Application of the differentialevolution strategy to the design of frequency-selective surfaces, Int J RF Microwave CAE 15 (2005), 173–180. C. Yildiz, A. Akdagli, and M. Turkmen, Simple and accurate synthesis formulas obtained by using a differential evolution algorithm for coplanar strip lines, Microwave Opt Technol Lett 48 (2006), 1133–1137.
© 2007 Wiley Periodicals, Inc.
A 16-GHZ CMOS DIFFERENTIAL COLPITTS VCO FOR DS-UWB AND 60GHZ DIRECT-CONVERSION RECEIVER APPLICATIONS C.-C. Lee,1 H.-R. Chuang,1 and C.-L. Lu2 Institute of Computer and Communication, Department of Electrical Engineering National Cheng Kung University, Tainan, Taiwan, Republic of China, Corresponding author: chuang [email protected] 2 Department of Computer and Communication, Kun Shan University, Tainan, Taiwan, Republic of China
1
Received 13 March 2007 ABSTRACT: A 16-GHz CMOS differential Colpitts VCO fabricated with the 0.18 m 1P6M process is presented. The 16-GHz VCO is a good choice for the local oscillator (LO) circuit of the UWB or 60-GHz WPAN direct conversion receiver. The VCO is composed of a PMOS transistor-pair core circuit and two source follower output buffers. The VCO can operate at 16.5 GHz, and the measured phase noise at 1-MHz offset is ⫺115 dBc/Hz. The power consumption of the VCO core is 12.6 mW. Compared with previous reported works, this VCO has an output power of ⫺0.9 dBm and about 800-mV output peak-to-peak voltage swing of the VCO core at 16.5 GHz. © 2007 Wiley Periodicals, Inc.
DOI 10.1002/mop
Microwave Opt Technol Lett 49: 2489 –2492, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop. 22742 Key words: 0.18 m; 16 GHz; 60-GHz WPAN; CMOS; differential Colpitts VCO; UWB
1. INTRODUCTION
Recently, ultra-wideband (UWB) radio has been proposed for physical layer standard of the future high-speed wireless personal area networks (WPANs). There are two main proposals, the multiband orthogonal frequency division multiplexing (MB-OFDM), and the direct-sequence ultra-wideband (DS-UWB). DS-UWB supports two independent bands of operation. They are the lower band that occupies the spectrum from 3.1 to 4.85 GHz, and the upper band that occupies the spectrum from 6.2 to 9.7 GHz. The lower band offers the piconet channel number 1– 6 and the upper band offers that of 7–12 [1]. Moreover, for dense local communications, the 60-GHz band with a band-width of about 8 GHz is of special interest to the short range (⬍1 km) communications [2]. It is due to its attenuation characteristic of 10 –15 dB/km by the atmospheric oxygen. For the UWB or 60-GHz WPAN direct conversion receiver, the 16-GHz frequency is a good choice for the local oscillator (LO). For the DS-UWB direct conversion receiver, a 16-GHz LO will be transformed into an 8-GHz or 4-GHz LO simply by a divide-by-2 ordivide-by-4 circuit. Also, a 16-GHz LO with a four-time multiplier can be used for a 60-GHz WPAN receiver application. For a CMOS mixer, the LO voltage swing directly affects the conversion gain and the noise figure performance. The switching pair of the mixer requires much greater voltage swing to complete switching. If the VCO output voltage swing is not high enough, differential LO amplifiers should be added to the VCO output, and will consume additional DC power. However, in order to achieve a high frequency operation, the value of the inductor of the tank has to be reduced. While the inductance is decreased, the parallel resistance (Rp) of the tank also decreases for a given quality factor (Q) and also the degrade of the output swing with a fixed bias current [3]. Moreover, the output swing of the VCO core and the phase noise performance are usually related. Therefore, it is desired to design a VCO with a moderate output swing at high frequency and with a low phase noise. The most popular topology of a VCO is the cross-coupled structure [4 – 6]. Compared with the NMOS- or PMOS-only and complementary cross-coupled oscillators, the Colpitts oscillator presents a smaller RMS and DC value of its effective impulse sensitivity function (ISF) [7, 8]. In [7], a 1.8-GHz NMOS differential Colpitts VCO with a phase noise of ⫺138.2 dBc/Hz at 3-MHz offset was first presented. It also shows that the Colpitts oscillator possesses a larger output voltage swing and higher energy transfer efficiency while comparing with the cross-coupled ones. In [9], a Colpitts VCO shows that by adopting only the PMOS to construct the core circuit of the oscillator would achieve a better phase noise performance since the PMOS flicker noise is lower than that of the NMOS. The measured phase noise of the demonstrated Colpitts VCO is ⫺120.4 dBc/Hz at 1-MHz offset while working at 5-GHz. It also achieves a good FOM value of about 190 and a DC-to-RF conversion efficiency of better than 28% with the open-drain buffer. In [10], a reported 20-GHz VCO in a 0.18-m standard CMOS process exhibits an output power of ⫺3 dBm, phase noise of ⫺111 dBc/Hz at 1-MHz offset, and FOM value of 182, respectively. The frequency tuning range is 510MHz and the power consumption is 32 mW.
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
2489
Figure 2 Simulation and measurement of the 16-GHz VCO output frequency versus control voltage Figure 1 VCO
Circuit schematic of a 16-GHz PMOS differential Colpitts
In this article, a 16-GHz PMOS differential Colpitts VCO (see Fig. 1) is designed by using bias-T source followers [11]. The VCO is fabricated in a 0.18-m standard CMOS process. The design goals of this high frequency VCO are to have a simple circuit scheme without needing a complicated EM-simulation, and achieve a high output voltage swing and a good FOM value. Although the small inductor and the parallel resistance in a high frequency VCO tank circuit will degrade the output voltage swing, this VCO with a source follower buffer still achieve a moderate output power of ⫺0.9 dBm and about 800-mV output peak-to-peak voltage swing of the VCO core. 2. CIRCUIT DESIGN
Figure 1 shows the schematic of a 16-GHz PMOS differential Colpitts VCO. Only the PMOS transistors are used in the circuit. The core of the VCO consists of a current switching pair and two symmetrical single-ended Colpitts VCO which is individually connected to a varactor diode. The PMOS cross-coupled pair of the current switching also provides the negative resistance to the Colpitts oscillators to improve the oscillating conditions, especially the start-up condition. Traditionally, a transistor constant current source is necessary for the bias in a differential structure. The current source works as a DC short circuit to offer a fixed bias current to the differential pair circuit, while offering a high AC impedance to avoid a loading effect. However, in order to avoid the reduction of the voltage headroom, the transistor constant current source is replaced by the on-chip RF choke and the by-pass capacitors to enlarge the voltage swing. Without the transistor constant current source, the bias current was directly controlled by choosing the transistor size. To filter out the high frequency noise, such as the 2nd harmonic (⬃32 GHz) at the common mode VDD node, the on-chip by-pass capacitors are used. Two on-chip symmetrical inductors with a centertap are used as the RF choke and the tank. Generally, the open drain circuits are used as the buffers of the VCO for measurement purpose. It can offer not only a sufficient ` load of the measurement instrument, and current for driving 50- U also provide a voltage gain to achieve a higher output voltage swing. However, for a 16-GHz VCO with the 0.18-m standard CMOS technology (unit-gain frequency fT is 60 GHz), it is not easy to amplify the large voltage swing from the VCO core without any matching network. Since the source follower output
2490
buffer does not provide any voltage gain, the output voltage swing of the VCO core circuit should be higher than the measured one through the buffer. Also, the inductive choke inside the bias-T provides a high ac impedance and good stability for the buffer. Therefore, small-size transistors can be used to provide enough output current drive without loading the tank heavily. However, long metal trace on the lossy CMOS substrate also degrades the output power seriously. Hence, the connecting metal trace between VCO core and output buffer should be as short as possible to reduce the parasitic effect. 3. MEASURED PERFORMANCE
As shown in Figure 2, the measured oscillation frequency of the VCO covers from 15.63 to 16.5 GHz with control voltage from ⫺1.8 to 1.8 V. Figure 3 shows the measured output power which ⫺0.9 dBm at 16.5 GHz after deembedding the cable and bias-T loss. The free running measurement of the phase noise is ⫺115 dBc/Hz at 1-MHz offset from a 16.5-GHz carrier as shown in Figure 4. The VCO core and two output buffers dissipate 12.6 and 30 mW, respectively. Here it should be noted that the bias-T source-follower output buffer is only for 50-⍀ system measurement purpose. The measured output power is ⫺0.9 dBm at 16.5 GHz. Since the source
Figure 3 Simulation and measurement of the 16-GHz VCO output power versus control voltage
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
DOI 10.1002/mop
Figure 4 Measured free running phase noise. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]
follower buffer has a voltage loss of 3 dB, the output peak-to-peak voltage amplitude of this VCO core will be about 800 mV from the measured power of ⫺0.9 dBm at the 50-⍀ load of the source follower buffer. Figure 5 shows the chip micrograph with a die area of 0.558 mm2. A widely used figure of merit (FOM) for VCO is defined as [12] FOM共dB兲 ⫽ 10 log
冉冉 冊 f0 fm
2
䡠
冊
1 , L共 fm兲 䡠 P
(1)
where L(fm) is the measured phase noise from the oscillator frequency at a frequency offset (fm), and P is the DC power consumption of the VCO. The FOM value is about 188. Table 1 summarizes the measurement results and the comparison with the previously reported VCOs. Again it is noted that the bias-T sourcefollower output buffer is only for 50-⍀ system measurement and has a voltage loss of 3 dB. The output peak-to-peak voltage swing of the VCO core is determined from the measured output power of the buffer by considering the buffer voltage loss. Hence the DC current consumption of the bias-T output buffer is excluded from total DC consumption of the VCO core circuit.
4. CONCLUSION
Figure 5 Chip micrograph (0.6 ⫻ 0.93 mm2). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley. com] TABLE 1
[4] [5] [6] [7] [11] This work
This article presents the design, fabrication, and measurement of a 16-GHz differential Colpitts VCO fabricated in a TSMC 0.18-m standard CMOS process. This 16-GHz VCO can be used for the LO circuit of the UWB or 60-GHz WPAN direct conversion receiver. The core circuit of the VCO consists of a PMOS differential Colpitts oscillator pair and a switching pair. The small DC and RMS value of the ISF and high energy transfer efficiency of the VCO core helps this work achieving a good phase noise performance. An LC tank circuit is used to replace the traditional current source to avoid the headroom limitation of voltage swing. Two bias-T source followers are used to for the measurement of the real output voltage swing. Although the small inductor and the parallel resistance in a high frequency VCO tank circuit will degrade the output voltage swing, this VCO still achieves a moderate output power level and high output voltage swing. The measured output frequency is from 15.63 to 16.5 GHz with a tuning range of about 5.4%. The measured free running phase
Summary of the Performance of a 16-GHz Diffential Colpitts VCO and Comparison with Previous Works Tech.
OSC. Frequency
Phase Noise
Power Consumption
0.18 m CMOS 0.13 m CMOS 0.13 m p-HEMT 0.18 m CMOS 0.18 m CMOS 0.18 m CMOS
10.7 GHz 18 GHz 12.2 GHz 12 GHz 19.9 GHz 16.5 GHz
⫺118.7 dBc/Hz @ 1 MHz ⫺117 dBc/Hz @ 1 MHz ⫺109 dBc/Hz @ 1 MHz ⫺84.8 dBc/Hz @ 100 KHz ⫺111 dBc/Hz @ 1 MHz ⫺115 dBc/Hz @ 1 MHz
11.8 mW 14.4 mW 3 mW 2 mW 32 mW 12.6 mW (VCO core)
DOI 10.1002/mop
Output Power/Core Voltage Swing (p-p)
FOM
⫺6 dBm/⫺10 dBm/⫺7 dBm/-/⫺3 dBm/⫺0.9 dBm/800-mV (with a 3-dB voltage-loss source follower)
188.5 190.5 190.5 183.4 182.0 188.4
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
2491
noise at 1-MHz offset is ⫺115 dBc/Hz at 16.5 GHz. The core circuit consumes 12.6 mW under 1.8-V power supply. The FOM value is about 188. The bias-T source-follower output buffer used for 50-⍀ system measurement has a voltage loss of 3 dB. From the measured output power of ⫺0.9 dBm at 16.5 GHz, the output peak-to-peak voltage amplitude of this VCO core will be about 800 mV by considering the 3-dB output buffer voltage loss. By a careful circuit design, this differential PMOS Colpitts VCO has a simple circuit scheme without needing a complicated EM-simulation, and achieve a high output voltage swing and a good FOM value. ACKNOWLEDGMENT
The authors thank the Chip Implementation Center (CIC) of the National Science Council, Taiwan, ROC, for supporting the TSMC CMOS process. REFERENCES 1. R. Fisher et al., DS-UWB physical layer submission to 802.15 task group 3a, IEEE P802.15– 04/0137r4 (2005). 2. P. Smulders, Exploiting the 60 GHz band for local wireless multimedia access: Prospects and future directions, IEEE Commun Mag 40 (2002), 140. 3. B. Razavi, RF microelectronics, Prentice Hall, Upper Saddle River, NJ 1997. 4. S. Ko, J.-G. Kim, T. Song, E. Yoon, and S. Hong, 20 GHz integrated CMOS frequency sources with a quadrature VCO using transformers, In: IEEE radio frequency integrated circuits (RFIC) Symposium, Fort Worth, TX June 2004, pp. 269 –272. 5. V. Manon, and S.I. Long, A low power and low noise p-HEMT Ku band VCO, IEEE Microwave Compon Lett 16 (2006). 6. T.K.K. Tsang and M.N. El-Gamal, A hign figure of merit and areaefficient low-voltage 12 GHz CMOS VCO, In: IEEE radio frequency integrated circuits (RFIC) symposium, Philadelphia, PA 2003, pp. 89 –92. 7. R. Aparicio and A. Hajimiri, A noise-shifting differential colpitts VCO, IEEE J Solid-State Circuits 37 (2002). 8. A. Hajimiri and T.H. Lee, A general theory of phase noise in electrical oscillators, IEEE J Solid-State Circuits 33 (1998), 179 –194. 9. M.-D. Tasi, Y.-H. Cho, and H. Wang, A 5-GHz low phase noise differential colpitts CMOS VCO, IEEE Microwave Wireless Compon Lett 15 (2005), 327–329. 10. H.-H. Hsieh and L.-H. Lu, A low-phase-noise K-band CMOS VCO, IEEE Microwave Compon Lett 16 (2006). 11. N.H.W. Fong, J.-O. Plaouchart, N. Zamdmer, D. Liu, L.F. Wagner, C. Plett, and N. Garry Tarr, A 1-V 3.8 –5.7-GHz wide-band VCO with differential tuned accumulation MOS varactors for common-mode noise rejection in CMOS SOI technology, IEEE Trans Microwave Theory Tech 51 (2003), 1952–1959. 12. E. Hegazi, H. Sjoland, and A.A. Abidi, A filtering technique to lower LC oscillator phase noise, IEEE J Solid-State Circuits 36 (2001), 1921–1930. © 2007 Wiley Periodicals, Inc.
2492
CALCULATION OF PATTERN IN UTD METHOD BASED ON NURBS MODELING WITH THE SOURCE ON SURFACE Wang Nan, Liang ChangHong, and Yuan HaoBo School of Electronic Engineering, Xidian University, Shaanxi, People’s Republic of China, Corresponding author: [email protected] Received 20 March 2007 ABSTRACT: The UTD method with the source on convex surface which is usually used in the analysis and design of antennas especially conformal arrays on electrically large platforms is studied in this article where NURBS is introduced as way of modeling. Together with the information from ray-tracing done before, the calculation of fields is presented which includes direct field in the lit region and creeping ray field in the shadow region. The method can be applied to arbitrary electrically large targets and the usefulness can be seen in the examples given. © 2007 Wiley Periodicals, Inc. Microwave Opt Technol Lett 49: 2492–2498, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.22727 Key words: NURBS; UTD; source on surface; pattern
1. INTRODUCTION
Within computational electromagnetics, electrically large targets are usually modeled either by meshes of planar patches or by typical objects. Modeling by planar meshes which are usually used in method like MoM is to use lots of planar patches (usually triangular) to simulate the target and the precision of the approached model depends on the number of the planar patches used but the requirement of computational resource raises greatly with the scale of the mesh at the same time. Modeling by typical objects which are usually used in method like UTD is easy to construct and is less consuming but obviously the approached models may not simulates the real ones precisely. In the area of modeling, STEP, the international standard for the geometrical definition of industrial products, was published by the international organization for standardization (ISO) where the nonuniform rational B-spline (NURBS) technique was accepted as the only mathematical way to define shapes of products. While the study of this technique becomes deeper and deeper, it has been already included by more and more commercial CAD/CAM softwares like CATIA, UG, etc. As introduced above, in UTD method, models are approached by typical simple geometric objects but not by arbitrary surfaces, which makes the use of UTD method limited, like the simplified model of an aircraft (Figure 1) that follows, whose wings are always approached by boards and head by cone which may not precisely simulate the real ones. Therefore, UTD method based on arbitrary surfaces modeling is worth studying both practically and theoretically. When analyzing EMC problems of electrically large platforms, UTD method is widely used and reasonably effective. On the other hand, practically the radiation from sources on perfectly conducting convex surfaces is of interest in the design of antennas mounted on aircraft and other electrically large platforms and in the design of conformal arrays, then it is necessary to study UTD under the fact that the source is on surface. The NURBS technique has been introduced into the area of calculational electromagnetics where many focus on the PO method based on NURBS models [1–3]; Perez et al. presented
MICROWAVE AND OPTICAL TECHNOLOGY LETTERS / Vol. 49, No. 10, October 2007
DOI 10.1002/mop
