Chapter 4 Transmitter Circuit Design
4.5 Summary of the Transmitter
Like the receiver, there are also power switches in the transmitter for saving power consumption. The PA power switch is in the bias circuit of PA, as shown in Fig. 4.12 (a). The pre-amplifier power switch is in the pre-amplifier core circuit because of the self-bias architecture, as shown in Fig. 4.12 (b).
The designed transmitter is achieve the simulated performances of 12.4 dB conversion power gain, 6 dBm OP1dB, 37 dB LO leakage rejection, and 27.5 dB RF image rejection, as shown in Fig. 4.13. Table 4.1 makes the summary of these performances of the designed transmitter.
1dB
Fig. 4.13 Simulated transmitter performance (a) Conversion gain (b) Output power
M
PA switchFig. 4.12 (a) PA power switch (b) Pre-amplifier power switch
38
Table 4.1 Summary of the simulated transmitter performance This Work Specification
Conversion Power Gain 12.4 dB N/A
OP1dB 6 dBm 4 dBm
Efficiency 16.4% 10%
Power Consumption 24.3 mW 25 mW
39
Chapter 5
Chip Implementation and Measurements
5.1 Chip Implementation
This transceiver is fabricated by the UMC 90-nm CMOS low-power process.
The top metal layer M9 of this process is comparatively thinner than the top metals of the normal RF process. In the RF integrated circuits implementation, the inductances are realized on the top metal layer because it’s the thickest metal layer in all metal layers. The inductance of thinner metal layer has larger parasitic resistor, and this resistor would reduce the quality factor Q of the inductances.
In this work, the inductance Lg are implemented in 3 metal of top metal M9, M8, and M7, and connected together by the contacts to form a thicker inductance, as shown in Fig. 5.1.
The Lg of 3 metal layers has higher Q value than the Lg of 1 or 2 metal layers has.
The inductance with higher Q value would restrain the noise figure of the receiver, as shown in Fig. 5.2. Moreover, the RF choke inductance LD of the PA is also implemented by 3 metal layers for higher Q value. The transformer is analyzed without loss like an ideal transformer. Therefore, both the primary and secondary coils of the transformer are realized by two metal layers in order to reduce the parasitic resistor.
M9M8
M7
contect
Fig. 5.1Inductance implemented in 3 metal layers
40
The whole transceiver is implemented on a single chip by the UMC 90-nm CMOS technology, as shown in Fig. 5.3. The chip size is .
Fig. 5.3 Chip photo of the low-power 1.4-GHz transceiver
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
Fig. 5.2 (a) Q factor of the inductances of 1, 2, and 3 metal layers (b) Noise figure of the receiver with the inductances of 1, 2, and 3 metal layers
41
5.2 Chip Measurement
The transceiver is measured by the way of chip on board. The measurement setups of the receiver for conversion gain, IIP3 by two tone test, and noise figure are shown in Fig. 5.4(a), (b), and (c), respectively.
Agilent E8257D
Fig. 5.4Measurement setups of the receiver for (a) Conversion gain (b) IIP3 by tone test (c) Noise figure
42
Because there are some problems in the tape out procedure, the digital pads of power switching don’t work. In order to solve this problem, we use laser to burning some nodes of the transceiver to short these nodes and VDD, as shown in Fig. 5.5. Using laser burning we successfully turn on the receiver, but there is no appropriate node of pre-amplifier for laser burning to turn it. Consequently, we can turn ON the receiver and measure it on the non optimal bias, and we measure the transmitter on the unable condition.
The receiver frequency response measurement results shows that the additional capacitor C1’ is a little bigger than the appropriate capacitance. It
Fig. 5.5 Laser burning
43
causes the peak of the frequency response shift from 1.4 GHz to lower frequency, as shown in Fig. 5.7. Therefore, we used the laser cut to cut out C1’ and used only the parasitic capacitances of the LNA and the primary coils of the transformer to form the C1 of the RCN, as shown in Fig. 5.6. Finally, we made the peak of the frequency shift back the higher frequency, as shown in Fig. 5.7.
The receiver measurement results of conversion gain versus LO power,
44
The conversion gains of the receiver in Fig. 5.5 (a), (b), and (c) are measured after the output buffer with 7 dB buffer loss. Both in the simulations and the measurements of frequency response, noise figure, P1dB, and IIP3, the LO power is chosen as -5 dBm for the maximal conversion gain in Q path and smaller gain mismatch consideration.
Fig. 5.8 Measurement results of (a) Conversion gain versus LO power, (b) Frequency response, (c) Noise Figure, (d) P1dB, (e) IIP3_I, and (f) IIP3_Q
45
Fig. 5.5 (a) shows the conversion gain of the receiver versus LO power as RF frequency is 1.4 GHz. The simulated LO power of the maximal conversion gain of I path is -6 dBm, and of Q path is -5 dBm. The measured LO power of the maximal conversion gain of I path is -3 dBm, and of Q path is -5 dBm. The measurement results prove that the conversion gain of the receiver using the RCN technique has a peak at low LO power. Compared to the normal receivers, the designed receiver only needs comparative low LO power to achieve high conversion gain. The low LO power characteristic of the receiver is very suitable for the low power application.
Fig. 5.5 (b) shows the frequency response of the receiver. The simulated maximal conversion gains of I path and Q path are both at 1.58 GHz. The measured maximal conversion gains of I path and Q path are both at 1.4 GHz. The peak of the conversion gain is at the designed resonant frequency ωL. The measurement results approve that the receiver using the RCN technique can get high conversion gain at low power operation as long as the operation frequency ω0 is designed at one of the resonant frequencies of the RCN ωL and ωH.
Fig. 5.5 (c) shows the noise figure of the receiver. The simulated noise figures at 1.4-GHz RF frequency of I path and Q path are both 3.32 dB. The measured noise figure at 1.4-GHz RF frequency of I path is 5.86 dB, and of Q path is 5.32 dB. The noise figures of I path and Q path are both meet the specification of noise figure less than 7 dB.
Fig. 5.5 (d), (e), and (f) show the linearity measurement results. The P1dB of I path is -19 dBm, and of Q path is -24 dBm. The IIP3 of I path is -11 dBm, and of Q path is -13 dBm. The linearity of the receiver also meets the specifications.
Table 5.1 makes the summary of the measured receiver performance.
46
The measurement setup of the transmitter is shown in the Fig. 5.6.
The measurement results of the transmitter are shown in Fig 5.7.
-15 -12 -9 -6 -3 0 3 6 -50
-40 -30 -20 -10 0
O u tp u t P o w e r (d B m )
Input IF Power (dBm) RF
LO
RF Image
Fig. 5.10 Measurement results of the transmitter
Agilent E8257D Signal Generator
Hybrid BalunBalun
LOQ
LOI
Agilent E4448A Spectrum Analyzer
R&S Vector Signal Generator
Fig. 5.9 Measurement setup of the transmitter
Table 5.1 Summary of the measured receiver performance
Measured performances Specification Conversion Gain I: 16.44 dB, Q: 15.53 dB * 20 dB
Noise Figure I: 5.86 dB, Q: 5.32 dB 7 dB
DC Power Consumption 6 mW 6 mW
P1dB I: -19 dB, Q: -24 dB -25 dBm
*The conversion gain is measured after output buffer with 7.6 dB buffer loss.
47
The switch of the pea-amplifier Mswitch can’t be turned ON. Under this condition, the pre-amplifier doesn’t work and can’t provide the RF signal voltage gain. Consequently, the input signal of the PA is too small the output power of the transmitter is also on the low power level. Even though the transmitter doesn’t entirely work, the LO leakage rejection of the transmitter still can be observed in the measurement results. From the measurement results the LO-RF leakage is -25 dB. Table 5.2 males the summary of the measured transmitter performance
5.3 Summary of the Measurement Results
Table 5.3 is the performances comparison table. The voltage conversion gain with 7.6 dB buffer loss of this 1.4-GHz receiver is 16.44 dB is close to the gain of [12]. The power consumption and the linearity of this receiver are at the same order of other three works. The noise figure is the best among these works.
The linearity and the LO leakage of this 1.4-GHz transmitter are at the same order of other three works. The power consumption of this transmitter is the critical part we must improve.
Table 5.2 Summary of the measured transmitter performance
Measured performances Specification
Peak transmitter power -6 dBm 4 dBm
Average transmitter power -13 dBm -3 dBm
Power consumption 22 mW 25 mW
Efficiency 1.2% 10%
48
*Including 8.4 dB buffer loss
49
Chapter 6 Conclusion and Future Work
6.1 Conclusion
In this work, we fabricated a low-power 1.4-GHz transceiver for wireless sensor work on a single chip by UMC 90-nm CMOS technology.
Our analysis of the RCN provides an efficient way to design an optimal transformer. The resonator coupling network with 2-metal-layer transformer provides the receiver high transfer current gain in the low power operation. The 3-metal-layer on chip inductance Lg really complete the input 50-Ω matching and noise matching of the receiver. The noise figure of this receiver is comparative lower in the low power operation. The needed LO power of this receiver with the RCN is lower than the LO power of other common receivers. The low LO power characteristic can reduce the loading of VCO, reduce the VCO power consumption, and further achieve the target of low power consumption.
The problems of the digital pads cause the transmitter unable to work regularly. However, based on the simulation results, we know that the low power target could be achieved by this transmitter architecture.
The measured results of the receiver and simulated results of the transmitter all meet the specifications which are based on the 1.4-GHz band in Wireless Medical Telemetry Service.
The whole chip size of the RF transceiver is only 1210μm*1360μm. The small chip size endues the transceiver with low cost advantage.
50
6.2 Future Work
Nowadays, the low-power application is becoming more and more important application. In Table 5.3, the noise performance is better than other works. The noise figure is the most flexible part we can alter in the next design. The power consumption of the receiver can be reduced more by sacrificing some noise performance. In the transmitter, the power consumption is the critical part we must improve. Using other classes of amplifier would raise the efficiency of the PA, and further reduce the power consumption.
In the chip layout of this 1.4-GHz transceiver, the pads of LO signal and analog baseband signal are set up under the consideration of the integration of this transceiver, a 1.4-GHz frequency synthesizer, and the analog baseband circuits.
This chip layout contributes convenience to the circuit integration in the future.
51
Reference
[1] Yen-Lin Liu, “Low-Voltage Low-Power 5-GHz Receiver Front-End Circuit Design for Wireless Sensor Networks,” Mast Thesis, NCTU Hsinchu Taiwan, July 2007.
[2] Chun-Hsing Li, Yen-Lin Liu, and Chien-Nan Kuo, “A 0.6-V 0.33-mW 5.5-GHz Receiver Front-End Using Resonator Coupling Technique,” IEEE Transactions
on Microwave Theory and Techniques, vol. 59, pp. 1629-1638, Jun. 2011.
[3] Chun-Hsing Li, Chia-Yu Hsu, and Chien-Nan Kuo, “Low Power 2.4-GHz Receiver Front-End Using Resoantor Coupling Technique,” 54th IEEE Int.
Midwest Symp. Circuits Syst., Aug. 2011.
[4] Chien-Chung Tsai, “Low-power Circuits Design in RF Transmitter Applications,”
Mast Thesis, NCTU Hsinchu Taiwan, Sep. 2009.
[5] Chang-Tsung Fu and Chien-Nan Kuo, “3~11-GHz CMOS UWB LNA Using Dual Feedback for Broadband Matching,” IEEE Radio Frequency Integrated
Circuits Symposium, pp. 53-56, Jun. 2006.
[6] Derek K. Shaeffer and Thomas H. Lee, “A 1.5-v, 1.5-GHz CMOS Low Noise Amplifier,” IEEE Journal of Solid-State Circuits, vol. 41, pp. 2359-2359, Oct.
2006.
[7] Thomas H. Lee, The design of CMOS radio-frequency integrated circuits, 2nd ed.
Cambridge, UK ; New York: Cambridge University Press, 2004.
[8] Hooman Darabi and Asad A. Abidi, “Noise in RF-CMOS Mixers: A Simple Physical Model,” IEEE Journal of Solid-State Circuits, vol. 35, pp. 15-25, Jan.
2000.
[9] Behzad Razavi, RF microelectronics, Upper Saddle River, NJ: Prentice Hall, 1998.
52
[10] Chiung-An Chen; Ho-Yin Lee; Shih-Lun Chen; Hong-Yi Huang; Ching-Hsing Luo, “Low-Power 2.4-GHz Transceiver in Wireless Sensor Network for Bio-Medical Applications,” IEEE Biomedical Circuits and Systems Conference, pp. 239-242, Nov. 2007.
[11] Trung-Kien Nguyen, Vladimir Krizhanovskii, Jeongseon Lee, Seok-Kyun Han,Sang-Gug Lee, Nae-Soo Kim, and Cheol-Sig Pyo, “A Low-Power RF Direct-Conversion Receiver/Transmitter for 2.4-GHz Band IEEE 802.15.4 Standard in 0.18-μm CMOS Technology,” IEEE Transactions on Microwave
Theory and Techniques, vol. 54, pp. 4062-4071, Dec. 2006.
[12] D. Zito, D. Pepe and B. Neri, “Low-Power RF Transceiver for IEEE 802.15.4 (ZigBee) Standard Applications,” IEEE International Conference on Electronics,
Circuits and Systems, pp. 1312-1315, Dec. 2006.
[13] R. van Langevelde, M. van Elzakker, D. van Goor, H. Termeer, J. Moss, and A. J.
Davie, “An Ultra-Low-Power 868/915 MHz RF Transceiver for Wireless Sensor Network Applications,” RFIC: 2009 IEEE Radio Frequency Integrated Circuits
Symposium, pp. 99-102, Jun. 2009.
[14] Carsten Hermann, Marc Tiebout, and Heinrich Klar, “A 0.6-V 1.6-mW Transformer-Based 2.5-GHz Downconversion Mixer With +5.4-dB Gain and 2.8-dBm IIP3 in 0.13-m CMOS,” IEEE Trans. Microw. Theory Tech., vol. 53, no.
2, pp. 488–495, Feb. 2005.
[15] C. K. Alexander and M. N. O. Sadiku, Fundamentals of electric circuits, 4th ed.
Boston: McGraw-Hill, 2009.
[16] X. He and J. van Sinderen, "A Low-Power, Low-EVM, SAW-Less WCDMA Transmitter Using Direct Quadrature Voltage Modulation," IEEE Journal of
Solid-State Circuits, vol. 44, pp. 3448-3458, Dec. 2009.
53
VITA
黃暉舜 Hui-Shun Huang
Birthday: 1987/05/30
Birthplace: Taoyuan County, Taiwan Education:
2005/09 ~ 2009/06
B.S. Degree in Department of Electrophysics, National Chiao Tung University 2009/09 ~ 2011/09
M.S. Degree in Department of Electronics Engineering & Institute of Electronics, National Chiao Tung University