2011 International Conference on Communication Engineering and Networks
IPCSIT vol.19 (2011) © (2011) IACSIT Press, Singapore
A 1.2V HighLinearity Down-Conversion Mixer for LTEadvancedApplications
Hung-Che Wei
Dept. of Electronic Communication Engineering, NationalKaohsiungMarineUniversity, Kaohsiung, Taiwan,
R.O.C. 81143
e-mail: hcwei@mail.nkmu.edu.tw
Abstract. Awideband high-linearity down-conversion mixer for LTE-advanced applications is described.
The gm-boosted and the current-bleeding techniques are employed to improve the linearity of the proposed
mixer.The mixer implemented by tsmc 0.18 μm CMOS process achievesmaximuminput third-order intercept
point (IIP3) of 2.6dBm, power conversion gainsof 6.9 dB,and single side-band noise figure (SSB NF) 20.8dB.
The mixer operates over the entire 1.4–3.6 GHz LTE-advancedbands and consumes 6.5 mA of current from a
1.2 V power supply.
Keywords- LTE-advanced;mixer; linearity
1. Introduction
Due to the evolution of the modern communication, the demands for mobile internet access have grown in
recent years significantly. Long Term Evolution (LTE) evolved from the High-Speed Packet Access cellular
standards is proposed by the 3rd Generation Partnership Project (3GPP). The LTE provides high data rates
and wide coverage. LTE release 10 (LTE-advanced) is depicted in September 2009. The operating bands for
LTE-advanced are 698–960 MHz, 1710–2025 MHz, 2110–2200 MHz, 2500–2690 MHz and 3400–3600 MHz,
respectively[1]. In the radio frequency (RF) receiver design, the linearity of a receiver is calculated by
3
1
3
3
3
3
1
whereAndenotes the n-th loaded voltage gain, and IIP3nrepresents the IIP3 magnitude of the n-th stage. It
shows the linearity of a receiver is dominated by the ones located following the first stage of the receiver[2].
The more IIP3n increases, the more IIP3total can be improved. IIP3 of a receiver is usually dominated by the
subsequent circuits such as mixers.
2. Design Methodology of Mixers
Fig.1 shows the gm-boosted common gate (CG) topology. M1 realizes a CG transistor and the inverting
amplifier between gate and source of the CG transistor is employed for thegm-boosted technique. Owing to the
noisy inverting amplifier, gm-boosted architecture can be realized by a capacitor. The CG architecture with the
capacitor cross-coupling technique is suitable for the impedance matching to 50 Ω by the equivalent
impedance
102
Fig. 1.gm
m-boosted CG topology.
1
1
2
Due to the gm-boostted CG and the
t capacitorr cross-couplling techniquue, the CG aand common source (CS))
c be realizedin a gm-bboostedcapaccitor cross-ccouplingtranssconductor sttage and thee third-orderr
topologies can
intermodulaation (IM3)oof a transconnductor staggecan be sup
ppressed[3,4]]. However, the overall gain of thee
topology is attenuated by
b the CG toppology. The gain and lin
nearity of thee gm-boosted CG can be improved
i
byy
adopting thee current-bleeding techniqque[5].
The prooposed high-linearity mixxer with a gm-boosted currrent-bleedinng transconduuctor stage iss depicted inn
Fig. 2. Thee gm-boosted current-bleeedingtransconnductor stag
ge consists of
o M1–M4, C1–C4, L1–L2 and R1–R2.
M1–M4and operatein thhesaturationrregion. C1 and
a
C2 are the cross-coouplingcapaccitors.M3 and
d M4are thee
current-bleeeding transisttors and imprrove the convversion gain and IIP3 of the proposedd mixer. R1 and
a R2are thee
wideband innputmatchingg network of RFport. Duue to the feeedback matcchingnetworkk of the RF port,
p
the RF
F
port does not
n require thhe external biasvoltage.T
b
Thegm-boosteed current-blleedingtranscconductor staage convertss
the input RF
R voltage siignals into sm
mall output current sign
nals to the coommutating stage. The commutating
c
g
stage is ofteen driven by the power froom local osccillator(LO) in
i the RF front-end. The ccommutating
gstage whichh
consists of M5–M8acts as ideal sw
witches whenn the input LOpower
L
is large. If thhe LOpower issmall, thee
commutatinngstageacts as an amplifieer. The DC operation poin
nt and the asppect ratio of
Fig. 3..Proposed mixxer with the gm
m-boosted currrent-bleedinggtranscoductorr stage.
the com
mmutatingstagge will influeence the requuirement of thedriven
t
LO
O power. Thee transistors are
a biased inn
the boundarry between the
t saturationn region andd the triode reegion to makke M5–M8 acct as ideal sw
witches withh
lower drivenn power. The RF signal is injected frrom the sourrce of M5–M8, then the IF
F current signal is down-103
converted by the commutating stage with the multiplied function. is translated into the voltage signal by The
load stage consisted of M9, M10, R3 and R4translates the down-converted current signal into the voltage signal.
R1 and R2 can provide output impedance and M9 and M10can provide appropriate voltage swing headroom.The
proposed gm-boosted current-bleeding transconductor stage compensatesthe IM3 phenomenon by adjusting
the bleeding current and cross-coupling capacitors. However, the bleeding current also increases the power
consumption of the mixer. In the mixer design, the trade-offs between IIP3, conversion gain and NFare the
main design considerations.Due to the NF of the receiver is dominated by the NF of the low noise amplifier,
the proposed mixer can be optimized inthe aspects of linearity and conversion gain.
3. Simulation Results
The simulator for the circuit simulation is Agilent Advance Design System (ADS) 2009. The proposed
mixer is realized by tsmc 0.18 μm Mixed Signal CMOS RF model. The RF is from 1.4 GHz to 3.6 GHz in the
simulation of the proposed mixer, Due to the fixed IF of 10 MHz, the LO frequency is from 1.39 GHz to 3.59
GHz. The active current of the mixer is about 6.5 mA from a 1.2 V supply voltage. In order to the mixer
performance,the three terminals of RF, LO and IF ports are matching to 50-Ω. As shown in Fig. 4, the
maximum conversion gain reached the peak value of 8.3 dB at the RF of 2.3 GHz when the LO power is –5
dBm. To optimize the overall performance of the proposed mixer, the LO power is –8 dBm . The conversion
gain of the proposed mixer is 4.3–6.9 dB and the SSB NF is 20.8–23.2 dB.
IIP3 of the mixer is calculated by using a two-tone testing. The frequencyspacing in the two-tone test is
set to be 300 kHz which is thechannel spacing in a LTE-advanced system. Fig. 5 illustrates CG and IIP3versus
RF. The conversion gainsare 4.3–6.9dB.Theextrapolation plot of IIP3 is illustrated in Fig. 6 andthe maximum
measured IIP3
Fig. 4.Conversion gain versus LO power at different frequencies.
Fig. 5.Conversion gain (CG) and IIP3 versus RF.
104
Fig. 6.IIP3 characteristic at 3.6 GHz.
achieves 2.6dBm at 3.6 GHz andthe minimum IIP3 is –0.6 dBm at 1.4GHz.The simulation results of other
mixers are compared by thesame design consideration in optimizing the linearity. Table Isummarizes the
simulation results of the proposed mixer. Theproposed mixer reveals excellent properties oflinearity and
power consumption.
TABLE I.
PERFORMANCE SUMMARY OF MIXERS.
Ref.
This work
[6]
[7]
0.18
0.18
0.18
Supply Voltage (V)
1.8
1.8
1.2
fRF(GHz)
1.9
2.4
1.4–3.6
fLO(GHz)
1.95
2.25
1.39–3.59
Currentconsumption (mA)
23.56
14.5
6.5
LO Power (dBm)
–1.2
4
–8
IIP3 (dBm)
20.45
8.6
–0.5–2.6
P-1dB (dBm)
12.8
1
–12.9––13.9
Conversion Gain (dB)
1.65
-5.3
4.3–6.9
SSB Noise Figure (dB)
17.2
17.5
20.8–23.2
Process (μm)
4. Conclusion
A 1.2 V high-linearityCMOS mixer with the gm-boosted current-bleedingtranscondutor stage is presented.
The proposedmixer operates at the RF of 1.4–3.6 GHz, the LO frequency of1.39–3.59 GHz, and IF of 10 MHz,
respectively. The mixer is realizedby adopting the linearity compensation method based on the gm-boosted
current-bleedingtransconductor. Both of the conversion gain and the linearity are improved. The
mixerconsumes 7.8mW from a 1.2 V power supply. The simulation results of the proposed mixer
exhibitsmaximum power conversion gain of 6.9 dB, IIP3 of 2.6dBm, and singleside-band noise figure of
20.8dB. The proposed mixer revealshigh conversion gain and IIP3 and is suitable forLTE-advanced
applications.
5. Acknowledgment
This work was supported by National Science Council, Taiwan, underthe Grant NSC99-2218-E-022-002.
The chip fabricationwas supported by the National Chip Implementation Center of Taiwan,R.O.C.
6. References
105
[1] M.J. Chang, Z. Abichar and C.-Y. Hsu, “WiMAX or LTE: Who will Lead the Broadband Mobile Internet?,” IT
Professional, vol. 12, no. 3, pp. 26–32, 2010.
[2] W. H. Hayward, Introduction to Radio Frequency Design, Upper Saddle River, NJ: Prentice-Hall, 1982.
[3] J. Jeong, J. Kim, D. S. Ha and H.-S. Lee, “A reliable ultra low power merged LNA and Mixer design for medical
implant communication services,” IEEE Life Science Systems and Applications Workshop (LiSSA), pp. 51–54,
Apr. 2011.
[4] Z. Lei, M. Kim and Y. Yang, “A common-gate down-conversion mixer with capacitive cross-coulping technique
for UHF RFID applications,” IEEE Microwave Technology & Computational Electromagnetics (ICMTCE), pp.
277–280, May 2011.
[5] S. G. Lee and J. K. Choi, “Current-reuse bleeding mixer,” IEE Electronics Letters, vol. 36, no. 8, pp. 696–697, Apr.
2000.
[6] N. Islam, S. K. Islam and H. F. Huq, “High performance CMOS converter design in TSMC 0.18-μm process,”
IEEE Proceedings of Southeast Conference, pp. 148–152, Apr. 2005.
[7] S.-Y. Chao and C.-Y. Yang, “A 2.4-GHz 0.18-μm CMOS doubly balanced mixer with high linearity,” IEEE
International Symposium onVLSI Design, Automation and Test, pp. 247–250, Apr. 2008.
106