transceivers architectures for mobile & wireless · pdf filetransceivers architectures for...
TRANSCRIPT
2
Wireless Transceiver Challenges
• Stringent linearity, phase noise, selectivity on RX
• Stringent mask, far-out noise on TX
WirelessReceiver
Desired TX
Interferer TXIn-band
N×200kHz in GSM
Select the desired channel
10
0d
B
Ou
t-o
f-b
and
3
Receiver Requirements
• Small signal linearity:
IIP2, IIP3
Y
XABAD
Operates here
GGD
fB1 fB22fB1-fB2• Large signal linearity:
Gain compression
• Reciprocal mixing: BNF = 174 – PB -PN
• Harmonic mixing
• Receiver NF sets the sensitivity, range:
Set by the standard10Log(KT), dBm/Hz
Sensitivity = -174 + NF + 10Log(BW) +SNRRS=50ΩΩΩΩ
RX
Blo
cke
rs: I
n/O
ut-
Ban
d
4
Transmitter Requirements
• Far-out noise
• Modulation quality: PE, EVM
10
dB
/
100kHz/
–Sets the modulator quality
–Sets the TX linearity
–Determines TX phase noiseGSM Mask
• Output power
• Spectrum Mask
• Phase noise intransceivers:
NF: Range
Blockers:Co-existence
SNR: Thru-put
Ph
ase
No
ise
Frequency
SNR, EVM
Mask, Blockers
Far-out noise
RX
Eq
uiv
alen
t
5
• CDMA less sensitive to jammers:
Code: 3.84MHz
TX Out
Code
RX IN
GSM/EDGE/WCDMA
Modulation Bands Thru-put Challenges
GSM GMSK: 270kbps 4 0.08 Blockers, TX mask, phase noiseEDGE 8PSK: 810kbps 4 0.236
3G+1 QPSK … 64QAM 19 0.384-21 TX leakage
• GSM/EDGE is TDMA, while 3G is CDMA. Both FDD.
1 HSPA, HSPA+, up to 21Mbps thru-put
Slot0 Slot1 Slot7 Frame: 8×577µµµµS…
6
RF IC
Large Blocker
Duplexer
RX Desired
TX Leakage
3G Full-Duplex Problem
• Duplexer is a dual-band filter
TX RX
7
Concept of MIMO
• Signals received on multiple antennas combined to create a more benign composite channel
– Combined signal energy increases SNR
– Variability across frequency reduces fading
• Trading robustness for rate
S1
S2 Dem
od
ula
tor
DiversityCombiner
8
• Flexible bandwidth and modes
–Over 40 bands
–1.4 -20 MHz variable bandwidth
–Flexible FDD, TDD, & FDD half duplex
• High mobility up to 120 km/h
• Carrier aggregation in LTE-A: 1Gbps
LTE Features
Band A Band B
• High data rate: DL 300Mbps, UL 75Mbps
RX1
RX2
RX3
RX4
LO1
LO2
Intra-Band
Inter-Band
Contiguous Non-Contiguous
9
WLAN Design Challenges
• Bandwidth variability and detection
– 20MHz up to 160MHz
– OFDM
• Receiver SNR and IQ imbalance
– LO (fixed) & filters (variable over frequency)
– Sensitivity: -67dBm for .11g, SNR = 25dB: NF =9dB
• TX PAR of 12dB for 64QAM
• High fidelity transmitters
– -28dB EVM for 64QAM
– IQ and in-band phase noise
0f
To help multi-path
I
Q
.3125M
10
Ideal Transceiver
AD
CD
AC
SDRBB
fIF = fIN - fLO
IF
fLO
RF
• Filter only rejects out-of-band blockers
• In-band blockers require a high-resolution ADC
• Power hungry
• Invented by Armstrong in 1918
• Frequency-conversion
relaxes IF signal processing
• Lower power
-99dBm
0dBm-26dBm
AD
CD
AC
7 Armstrong, 1924 10 Mitola, 2005
11
LO Harmonic Mixing Issue
• In a perfectly linear RX, blockers still problematic
fLO nfLO
fIFfIF
• Sets LNA IIPk , filter attenuation, range of acceptable IF
PD
PB
(nf L
O±f
IF)/
k
fLO nfLO3fLO
…
Hard Switching Mixer
IIPk
15 Cijvat, TCAS 2002
• Image: n=k=1, Only removed by filtering
• Half-IF: n=k=2, Differential helps fLO
fIF/2
(2f L
O±f
IF)/
2
12
Super-Heterodyne Receiver
• Down-conversion relaxes the ADC
• High IF avoids DC offset & low frequency noise
• Needs external filters for image and other blockers
On-Chip
Pre-Select
Channel-Select
fLOIF
Image Reject
7 Armstrong, 1924
13
Zero-IF Receiver
• Less severe image issue
• Channel selection on-chip
• Suitable for WB: LTE, WiFi
Channel-SelectIQ
• Requires quadrature LO
• DC Offset, 1/f noise, IIP2
• In band IRR
Pros Cons
1 Abidi, JSSC 1995
ωωωωLO-ωωωωLO
0
Complex
f1
/f n
ois
e OFDM
14
• For 12.2kbps reference measurement, SF=128, -117dBm sensitivity, required NF = 9dB
DPCH_Ec = -117dBm/3.84Mz
I^or = -106.7dBm/3.84MHz
Noise Floor = -108dBm(-174,3.84MHz)SF+CG=25dB
SNR = 7dB-99dBm
• For a given duplexer, NF depends on:
–RX thermal noise
–RX 2nd-order nonlinearity
–TX and PA noise at RX band
NF
3G RX NF Requirement
15
• TX leakage amplitude demodulated at zero IF
• In order not to affect sensitivity IIP2 > 50dBm
A(t)Cos(ωωωωt+φφφφ(t))
A(t)2
3G IIP2 Requirements
fTX fRX
0
+24dBm
…
+28dBm
I ≈50dB
IIP2 = 2×(28dBm-I) – 13dB – -99dBm -10dB = +52dBm
Desensitizes by 0.5dB
16
• Stringent IIP3 due to TX leakage:
TX
3G Out-of-Band IIP3 Requirements
PTX PB
(fRX-fTX)/2
28-50 -15-30-99+3-5
• IIP3 = -5.5dBm at the LNA input
• Need room for TX noise, 2nd-order nonlinearity …
Duplexer Isolation: 50dBDuplexer Filtering: 30dB
…
Signal = PTX + 2×PB - 2×IIP3
17
Low-IF ReceiverLow-IF
fLOI
fLOQ Act
ive
Po
lyp
has
e• IIP2, 1/f less problematic
• Image in-band
• Suitable for NB: GSM, BT
• Requires quadrature LO
• Higher IF, higher power
• Tighter IRR
Pros Cons
0 IF∝∝∝∝BWIR
Digital IR
18
GSM Noise Figure
• Most advanced receivers target for <-109dBm
• NF of < 3dB assuming 3dB loss at front-end
• The standard requires -102dBm
• Receiver NF sets the sensitivity:
≈ 59dB for GMSK
Sensitivity = -174 + NF + 10Log(BW) + SNR…
≈3dB
Band X
19
Choice of IF: Adjacent Blockers
•Trade-off between 1/f noise, IIP2, … vs. image rejection
12
0
14
0
16
0
18
0
IF, kHz
IMR
R, d
Bc
30
35
40
45
Desired
20
0k
60
0k
40
0k
-73dBm
-41dBm
-82dBm
-33dBm
fLOIF
n×200k
IF
n×200k-2×IF
Mostly 200k
Tail of 400k
400k
20
GSM In-Band Blocking Requirements
Desired
60
0k
1.6
M
3M
-43dBm
-99dBm
-23/-26dBm
• LO PN of -140dBc/Hz at 3MHz
• BNF = 174 – 26 – 140 = 8dB
• RX NF of 6dB: Total BNF ≈ 10dB
• 3GPP NF requirement: 13dB
• Compression degrades the BNF further
– -26/-23dBm blocker can heavily compress the front-end
PD
PBSNR
BWfLO
∆∆∆∆fB
Noisy LO
BNF = 174 + PB + PN
21
Dual-Conversion Receiver
• DC offset and 1/f noise issues less severe
• No high frequency quadrature LO
• Higher power due to 2 mixers in signal path
• First LO image
Zero or Low-IF
fLO2I
fLO2Q
fLO1
fLO2 = fLO1/N
Sliding 1st IF = fRF / N+1
13 Zargari, JSSC 2002
22
Gain Control & Receiver SNR
• SNR determines through-put at high input powers
P1
dB, d
Bm
-11
2
-97
-82
-67
-52
-37
-22
Input, dBm
• Front-end gain reduction improves P1dB
• But degrades SNR
0
10
20
30
40
-112
-97
-82
-67
-52
-37
-22
Input, dBmSN
R, d
B PN, IQ& linearity
Gai
n, d
B
In-bandAdjacent
Ideal Slope
23
RF PGA1 PGA2
-90
-60
-30Blocker
Sign
al, d
Bm
AD
C
ADC
Q-Noise
ADC and Filtering
U/D fading
• GC keeps desired signal
close to ADC full scale
• Trade-off between filter
& ADC
• ADC DR >> receiver SNR
Signal
24
Example: 2G ADC Requirements
• -400kHz blocker dominant in low-IF
ADC Full Scale
Up Fading
Blocker
Down Fading
87dB
Margin: GC
10dB
6dB
41dB
10dB
ADC Q Noise
20dB Noise Margin
Blocker
Signal(57dB below FS)
(SNR ≈ 6dB)
25
Out-of-Band Blocking Issue
External SAW filters attenuate out-of-band blockers
The in-band blocker as high as -23dBm
19
90
20
10
20
70
-23/6dBm
-12dBm
0dBm
19
30
-99dBm
PCS Band1
91
0
18
30
GSM out-of-band blocker profile:
Frequency, MHz
26
Narrow-Band Filtering Concerns
• Large blockers compress the receiver
• Impose stringent far-out phase noise
• External filtering is narrow-band and costly
Mu
lti-
Ban
d R
ece
ive
r
Swit
ch
…
-99dBm
0dBm
15-20dB
6.6V p-p
27
Passive Mixers as N-Path Filters4
)()(2
)(2 LOBBLOBBSWin jsZjsZRsZ ωω
π++−+≈
LO1
LO2
LO3
LO4
0
IRF
VRF
0
High-Q BPF from low-Q LPF
fLO
fLO
LO1
LO4
0
LO2
LO3
VRF
IRF
ZBB
4 L. Franks, Bell Syst. Tech. J., 196014 Mirzaei, TCAS 2010
Blo
cke
r
28
Current-Mode Receivers
• Passive mixers to achieve high-Q filtering
• Current mode LNA: LNTA
• Enhance the blocker tolerance
21,22 Mikhemar, VLSI 2012
GM
0fRF
fRF
0
BUF
BUF
29
Mixer-First Receivers
• At high frequency noise aliasing degrades NF significantly
• NF > 4dB in practice at GHz frequencies 18,19 Andrews, JSSC 2010
LO1
LO2
LO3
LO4TI
A_I
I
TIA
_Q Q
RS
VS
SW Resistance, ΩΩΩΩ
NF,
dB
0
1
2
3
4
5
0 10
20
30
40
50
gAlia
S
bb
S
DS NKTRM
v
R
ONrF
sin
2
)4(
)(1 ×
++=• For M phases:
30
Noise Cancelling Receivers
• Low noise and linear
• No Balun required
RS
RS = RSW + 2RBB/ππππ2
-+
BW << ∆∆∆∆fB
-+
GM αααα = -GM×RLA+
VOUT
-
Low resistance
Noise and linearity bottleneck
RLA
RLM
rm = RLM
26 Murphy, ISSCC 2012
31
Over-Sampling Mixer Architecture
• Square-wave LO harmonically rich
• Synthesizes arbitrary 8-phase LO:
TIA K0
TIA K1
TIA K7
ΣΣΣΣVOUT
…
IRF
8-P
has
e
√2
Harmonic Combination
∑=
−=7
0
)8
()(x
xRFOUTT
xtswKtiV
1
1/7
17 Weldon, JSSC 2001
32
Case Study: NC SDR Receiver
GM
10ΩΩΩΩ
50ΩΩΩΩ
I
Q
We
igh
tin
g &
Re
com
bin
atio
n
…LO0LO1
LO7
4fLO ÷4
NF ≈ 1 + γγγγ/GMRS
26 Murphy, ISSCC 2012
• 1.9dB NF, 4dB BNF
33
Direct-Conversion Transmitters
• Low power
• Versatile
• Highly integrated
• Suffers from pulling
• LOFT, IQ matching
• Far-out noise
Pros Cons
÷2/4
I Q
fLO
34
Dual-Conversion Transmitters
fLO2IfLO2Q
fLO1
• No pulling
• LOFT/IR less problematic
• Sliding IF
• Higher power
• More complex filtering needed
Pros Cons
÷N
13 Zargari, JSSC 2002
35
Third Harmonic Folding
fLO 3fLO
A
fLO +f1 fLO +f2
A/3
3fLO -f13fLO –f2
f
a3vd 3(t)
f
9a3A 3/43a3A 3/4
3a3vd 2(t)vu(t)
fa3A 3/4
a3A 3/2
fLO +2f1-f2 fLO +2f2-f1
fLO -2f1-f2fLO -2f2-f1
3a3vd (t)vu2(t)
f
a3A 3/2a3A 3/6
fLO +2f1-f2 fLO +2f2-f1
PA driver nonlinearity: y=a1x+a3x3
un
des
ired
co
mp
on
ents
vd (t) vu (t)
f LO
-3
f 1
…
…
…
23 Mirzaei, TCAS 2011
36
• -160dBc/Hz RX-band noise
results in 0.5dB NF degradation
RF IC
RX Desired
TX Leakage
WCDMA TX General Requirements
• ACLR1 at 5MHz: -33dBc
• ACLR2 at 10MHz: -43dBc
• EVM: 19%
Noise Floor = 28-50-160 = -182dBm/Hz
37
• In linear TX, IQ imbalance, LOFT & PN dominate
WCDMA TX EVM
101010 101010(%)
PNLOFTIQ
EVM ++=
• 40dBc equal contribuXon results in √3 = 1.7% EVM
• Baseband filter ripple adds further
• Typical RF IC EVM around 3%
1
Error+1.9
Mf L
O
-1.9
MNoise Floor
38
• Nonlinearity results in ACLR
– Depends on PAR, modulation
WCDMA TX ACLR
+1.9
M
0
-1.9
M
5±1.92MHz
ACLR1, dBc
IM3
• ACLR1 requirement of -33dBc at the antenna
– PA WC -37dBc (optimized for efficiency)
–2dB production margin
–Leaves RFIC WC of -40dBc
+3dBm
0d
Bm
-31
dB
c
OIP
3=
15
.5d
Bm
39
Folding Impact on PAD Linearity
• With third harmonic present: ACLR = -38 dBc
• With third harmonic removed: ACLR = -44 dBc
• Makes the PAD linearity requirements more stringent
DAC
DAC
f
1
≅≅≅≅ 1/3 ≅≅≅≅ 1/5
vRF,1(ωωωω)
vRF,5(ωωωω)vRF,3(ωωωω)
fLO 3fLO 5fLO
PAD
-6 -4 -2 0 2 4 6
-80
-60
-40
-20
Output 3G Signal
Frequency, MHz
Po
we
r, d
Bc
PAD Ideal
w/ 3rd
w/o 3rd
40
Case Study: Direct-Conversion 3G TX
• Passive 25% mixer
• LOFT scales w/ RF gain
I
…
10-20dB> 70dB
+24
~ -
50
dB
m
Q
GC
PAD Units
1
41
GSM RX-Band Noise Requirements
TX noise in RX-band -79dBm not to mask adjacent RX
Corresponds to a 20MHz phase noise of:
88
0
91
5
93
5
96
0
EGSM TX
+33dBm
-79dBm
Typical PA noise ≈ -83dBm, leaving -165dBc for RF IC
-112dBc spur, five exceptions allowed
EGSM RX
≈≈ ≈≈ ≈≈ ≈≈
Freq, MHz20M
PN = -79dBm – 10Log(100kHz) – 33dBm = -162dBc/Hz
42
Linear TX vs. Translational Loop
PFD/CP
LoopFilter
fLO
fLO+fIF
Modulator fIF
• Sensitive to Pulling
• 20M noise an issue
• Simple
• Generic TX
• No pulling issue
• Relaxed filtering
• More complex
• Suitable for PM only
2 Erdogan, ISSCC 2005
43
MMD
PFD/CP
∆Σ∆Σ∆Σ∆ΣModulator/
Pre-distortion
Reference
• Mixer/LO, analog modulator eliminated
• More sensitive to analog impairments
• Trade-off between BW and phase noise
PLL-Based Transmitters
3 Bonnaud, ISSCC 2006
Lowpass100kHz/
40
0kH
zPLL
44
GSM Mask & Phase Error Calculations
• -60dBc at 400kHz
– 3dB production margin
– 2dB PVT margin
• 5⁰ RMS phase error
– 2⁰ production margin
– 2⁰ PVT, BW1
00
kHz-88dBc
-88 + 10Log(200kHz) = -35dBc = 1⁰
Frequency
PN
Power integrated in 30kHz BW≈ 10Log (200k/30k)
PN = -65 – 10Log(30kHz) – 9 = -118.8dBc/Hz
0.0178 Radian
45
MMD
PFD/CP
∆Σ∆Σ∆Σ∆Σ
Mo
du
lato
r
Reference
PM
AM
A(t)
Sin(ωωωω0t+φφφφ(t))
∆Σ∆Σ∆Σ∆Σ
A(t)Sin(ωωωω0t+φφφφ(t))
• Lower power consumption
• Compatible with GMSK TX
• Very sensitive to nonlinearities
Basics of Polar Transmitters
I + jQ = rejθθθθ
EDGE Constellation
46
-20
-60
-40
-80
0
-40
0
-80
0
40
0
80
0
Frequency, kHz
Am
pli
tud
e, d
B
PMAM
8PSK
EDGE AM & PM Signals Spectrum
Ideal20nS40nS80nS160nS
0
-40
0
-80
0
40
0
80
0
Frequency, kHz
• AM & PM stand-alone signals much wider
47
MMD
PFD/CP
∆Σ∆Σ∆Σ∆Σ
REF
PM
• Phase noise
• VCO & CP nonlinearity due to large swing, wide BW
• PLL BW needs to be accurately controlled
BW ∝∝∝∝ KVCO × R × ICP : Measure KVCO and adjust ICP
PM Path Concerns
200 300 400
-40
-20
0
20
40
Time, µµµµS
VC
TRL
Swin
g, m
V
5,9 Darabi, JSSC 2011
Sin(ωωωω0t+φφφφ(t))
φφφφ/KVCO
Only function of KVCO
48
-70
-65
-60
-55
-50
-30
-15 0
15
30
45
AM-PM Relative Delay, nS
±40
0kH
z M
od
, dB
c
Corrected
AM-PM Uncorrected
AM Path Concerns
• AM-AM/PM distortion in PAD
• Phase feed-through
Mask Limiting Factors:
AM
Static: AM-AMDynamic: AM-PM
PA or PAD
40
0kH
z
20
dB
/
200kHz/
PFT
49
MMD
PFD/CP
∆Σ∆Σ∆Σ∆Σ
Mo
du
lato
r
REF
PM
AM∆Σ∆Σ∆Σ∆Σ
• HP path to give a flat response
• Accurate KVCO needed
• Accurate AM-PM matching needed
2-Point PLL Based Polar Transmitters
+
10 20 30 40 50
-80
-40
0
Frequency, MHz
Mag
nit
ud
e, d
Bc
AM
PM
3G
3G AM & PM Signals
LP
HP
×1
/KV
CO
KVCO
50
• KVCO = dωωωω/dV = 0.5×ωωωω00003×L× (dCVAR/dV)
Impact of VCO Nonlinearity on 3G SignalEV
M, %
-0.8
-0.4
0 0.4
0.8
VCO Gain Error, %
0.5
1.0
1.5
• KVCO accuracy/linearity of better than 2% needed
• LC and CVAR not modeled accurately
AC
LR, d
Bc
1 2 3 4 5
VCO Nonlinearity, %
-60
-50
-40
ACLR2
ACLR1
PVT Dependent
PLL BW: 250kHz
51
Case Study: 3G Polar TX
• < -40dBc ACLR, < 3% EVM
6 Youssef, JSSC 2011
MMD
PFD/CP
Mo
du
lato
r
REF
PM
AM
∆Σ∆Σ∆Σ∆Σ
+ ÷2
÷4
÷2
÷2
F/V OUT
LB
HB
HP Path
Matched20M10b
Linear VCO
∆∆∆∆T
PLL (250kHz)
AM Path
52
Summary & Conclusions• Standards set design parameters, radio architecture
• On receiver side:
Architecture 1/f, IIP2 IMR Linearity Limitations
Zero-IF Poor Self Modest WB
Low-IF Modest In-band Modest NB
• CM receivers improve linearity, relax filtering
• On transmitter side:
Architecture Power Noise Pulling Limitations
Direct Modest Poor Poor Generic
PLL Lowest Good None Only PM, NB
Polar Lowest Good Good Complex, NB
53
References1. A. A. Abidi, “Direct-Conversion Radio Transceivers for Digital Communications,” IEEE J. of Solid-State Circuits, vol. 30, no. 12, pp. 1399-1410, 1995
2. O. Erdogan, et. al., “A single-chip quad-band GSM/GPRS transceiver in 0.18mm CMOS,” ISSCC Digest of Technical Papers, pp. 318-319, Feb. 2005.
3. O. Bonnaud, et. al., “A fully integrated SoC for GSM/GPRS in 0.13mm CMOS,” ISSCC Digest of Technical Papers, pp. 482-3, Feb. 2006.
4. L. Franks and I. Sandberg, “An alternative approach to the realizations of network functions: N-path filter,” in Bell Syst. Tech. J., 1960, pp. 1321–1350.
5. H. Darabi, at. el., “A Quad-Band GSM/GPRS/EDGE SoC in 65nm CMOS,” IEEE J. of Solid-State Circuit, no. 4, April 2011.
6. M. Youssef, et. al., “A Low-Power Wideband Polar Transmitter for 3G Applications,” ISSCC Digest of Technical Papers, Feb. 2011.
7. E. H. Armstrong, “The Super-Heterodyne-Its Origin, Development, and Some Recent Improvements ” Proc. of the Institute of Radio Engineers
Volume: 12 , Issue: 5, 1924, pp. 539-552
8. Z. Boos, et. al., “A Fully Digital Multi-Mode Polar Transmitter Employing 17b RF DAC in 3G Mode,” ISSCC Digest of Technical Papers, Feb. 2011.
9. H. Darabi, et. al., “Analysis and Design of Small Signal Polar Transmitters for Cellular Applications,” IEEE J. of Solid-State Circuit, 2011.
10. J. Mitola, “The software radio architecture,” IEEE Communications Magazine, vol. 33, no. 5, pp. 26-38, May 1995.
11. A. Mirzaei, et. al., “Analysis and optimization of current-driven passive mixers in narrowband direct-conversion receivers,” IEEE J. of Solid-State Circuit,
no. 10, pp. 2678-2688, October 2009.
12. D. Kaczman, et. al., “A single-chip 10-band WCDMA/HSDPA 4-band GSM/EDGE SAW-less CMOS receiver with DigRF 3G interface and 90dBm IIP3”, IEEE
J. of Solid-State Circuit, no. 3, pp. 718-739, 2009.
13. M. Zargari, et. al., “A 5-GHz CMOS transceiver for IEEE 802.11a wireless LAN systems”, IEEE J. of Solid-State Circuit, no. 12, pp. 1688-1694, 2002.
14. A. Mirzaei, et. al, “Analysis and optimization of direct-conversion receivers With 25% duty-cycle current-driven passive mixers ,” IEEE Trans Circuits &
Systems I, No. 9, Vo. 47, 2010, pp. 23530-2366.
15. E. Cijvat, et. al, “Spurious mixing of off-channel signals in a wireless receiver,” IEEE Trans Circuits & Systems II, No. 8, Vo. 49, 2002, pp. 539-544.
16. A. Mirzaei, et. al., “A 65nm CMOS quad-band SAW-less receiver for GSM/GPRS/EDGE,” IEEE J. of Solid-State Circuit, no. 4, April 2011.
17. J. Weldon, et. al., “A 1.75-GHz highly integrated narrow-band CMOS transmitter with harmonic-rejection mixers ,” IEEE J. of Solid-State Circuit, no. 12,
pp. 2003-2015, Dec 2001
18. C. Andrews, et. al., “A passive-mixer-first receiver with Digitally Controlled and Widely Tunable RF Interface,” IEEE J. of Solid-State Circuit, Vol. 45, no.
12, pp. 2696-2708, Dec 2010
19. M. Soer, et. al., “A 0.2-to-2GHz 65nm CMOS receiver w/o LNA achieving >11dBm IIP3 and <6.5dB NF,” ISSCC Digest of Tech. Papers, pp. 222-3, Feb. 09
20. X. He, et. al, “A Low-Power, Low-EVM, SAW-Less WCDMA Transmitter Using Direct Quadrature Voltage Modulation,” IEEE J. of Solid-State Circuit, vol.
44, no. 12, pp. 3448–3458, 2009.
21. A. Mirzaei, et. al., “A Frequency Translation Technique for SAW-Less 3G Receivers,” Proc. of IEEE Symp. VLSI Circuits, pp. 280-281, 2009.
22. M. Mikhemar, et. al., “A 13.5mA Sub-2.5dB NF Multi-Band Receiver,” Proc. of IEEE Symp. VLSI Circuits, 2012.
23. A. Mirzaei, et. al, “Analysis of Direct-Conversion IQ Transmitters With 25% Duty-Cycle Passive Mixers,” IEEE Trans Circuits & Systems, No. 10, Vo. 58, pp.
2318-2331, 2011.
24. A. Abidi, “Linearization of Voltage-Controlled Oscillators using Switched Capacitor Feedback,” IEEE J. of Solid-State Circuits, vol. 22, no. 3, pp. 494–496,
June 1987.
25. F. Broccoleri, et. al., “Wide-band CMOS low-noise amplifier exploiting thermal noise canceling,” IEEE J. Solid-State Circuits, vol. 39, no. 2, pp. 275–282,
Feb. 2004.
26. D. Murphy, et. Al., “A blocker-tolerant wideband noise-cancelling receiver with a 2dB noise figure ” ISSCC Digest of Tech. Papers, pp. 74-6, Feb. 12.