You are on page 1of 3

0.

4 mW wideband LNA with double gm The input impedance of the LNA is


enhancement and feed-forward noise 2
cancellation Zin = (3)
2gm,NM1 + (1 + Gboost ) × gm,NM3

Zhi Li, Liguo Sun and Lu Huang where gm,NM1, gm,NM3 and gm,PM1 are the transconductances of NM1,
NM3 and PM3 respectively, Rl is the load impedance at the drain of
A low-power wideband common-gate (CG) low-noise amplifier (LNA) NM1.
presented . The CG LNA uses double gm enhancement to provide input Equation (2) indicates that a smaller gm value is required for a good
matching under low-power consumption. Feed-forward noise cancella- input matching because gm2 is boosted by a coefficient 1 + Gboost.
tion (FFNC) is employed in the LNA to suppress the noise from the CG Therefore, the required DC bias current is reduced.
transistor. The LNA is designed and fabricated in TSMC 130-nm The balanced ports of the off-chip balun are directly coupled to the
CMOS technology. This LNA can achieve a maximum gain of
sources of CG transistors. The balun has a DC-feed port to ground the
14 dB with a 3 dB bandwidth from 350 to 950 MHz. The LNA con-
sumes 0.5 mA current under a 0.8-V supply. The average noise
source of the transistor, so the voltage headroom can be saved and the
figure of the LNA is 4.0 dB. The core area of the LNA is 0.06 mm2. supply voltage can be reduced.
The double gm enhancement reduces the current and the direct-
coupled balun reduces the supply voltage, therefore this architecture
Introduction: Multiband multistandard concepts have gained impressive guarantees a low power.
interest in modern wireless communication. Today’s trend is to focus on The noise contribution from the main CG amplifier can be sup-
a single wideband front-end to accommodate all the standards to save pressed, because the signal at the gate of the main CG amplifier is
area. As the first signal processing block in the receiving chain, a already amplified by the gm-boost CG amplifier.
wideband low-noise amplifier (LNA) with low power, low cost and However, the noise contribution from the gm-boost amplifier is sig-
low noise is desirable. nificant, because the signals are coupled to NM1, 2 without being ampli-
The presented wideband common-gate (CG) LNA covers frequency fied and the gm value of NM1, 2 is small.
bands from digital video broadcasting (DVB) at 450–850 MHz to the To reduce the noise contribution of the gm-boost amplifier, a pair of
global system for mobile communications (GSM) at 900 MHz, provid- pmos amplifiers is stacked upon NM1, 2 as a feed-forward noise cancel-
ing a practical solution for multistandard applications. This LNA is able lation circuit.
to relieve the power and input matching limitations of traditional CG The feed-forward noise cancellation scheme is shown in Fig. 2.
LNAs with double gm enhancement. Feed-forward noise cancellation
(FFNC) is used to reduce the noise figure (NF) and consumes no
VDD
extra power.

Proposed wideband LNA: The proposed wideband LNA is depicted in Z1 Z2


Fig. 1; it is an all-differential architecture with an off-chip balun.
R5 vb3 R6

VDD
PM1 PM2
Vout
Z1 Z2
R3 R4
R5 vb3 R6
vb2
PM1 PM2 C5 NM3 NM4 C6
Vout C3 C4
R3 R4 main CG
amplifier
vb2
C5 NM3 NM4 C6 vb1
C3 C4 In,M1 R1 R2

NM1 NM2
vb1
R1 R2 gm-boost C1 C2
amplifier
NM1 NM2
C1 C2

Vin
Vin

Fig. 2 Feed-forward noise cancellation scheme of proposed LNA


Fig. 1 Schematic diagram of proposed wideband LNA
The noise current of NM1 produces noise voltage with different
Double gm enhancement is applied to this LNA. The gm-boost ampli- polarity at the drain and the source of NM1. One path is from the drain
fier is a pair of capacitor-cross-coupled CG transistors. The main CG of NM1 to the drain of NM3, the noise voltage at the drain of NM1 is
amplifier is enhanced by the gm-boost amplifier. amplified by NM3. The other path is from the source of NM1 to the
On considering the gm-boost amplifier and the feed-forward amplifier drain of NM4, the noise voltage at the source of NM1 is amplified by
PM1, the gain from the source of NM1 to the gate of NM3 is PM2 and then by NM4. As a result, the noise voltage at the drain of
  NM3 and NM4 has the same polarity. Therefore, the output noise
Gboost = gm,PM1 + 2gm,NM1 × Rl (1) voltage produced by NM1 is partially cancelled with feed-forward
noise cancellation. Similarly, the noise of NM2 is also partially cancelled.
The effective transconductance of NM1 and NM3 is The feed-forward noise cancellation circuit shares DC bias current
with the gm-boost amplifier. This current reuse structure is used to
gm,eff = 2gm,NM1 + (1 + Gboost ) × gm,NM3 (2) achieve a lower power.

ELECTRONICS LETTERS 27th February 2014 Vol. 50 No. 5 pp. 400–401


Measurement results: The proposed LNA was fabricated in the 130-nm Conclusions: A 130-nm inductorless wideband LNA has been pre-
TSMC CMOS technology. Fig. 3 shows the die photo of the fabricated sented. Double gm enhancement and a direct-coupled balun are used
circuit. The core area is only 0.06 mm2. to reduce the current and supply voltage, respectively. Feed-forward
noise cancellation is employed to reduce the NF without extra power
consumption. This LNA achieves a much lower power than the other
wideband LNAs while also providing a comparable gain and an NF
to the other wideband LNAs, as shown in Table 1.

Acknowledgment: This work was supported by the National Science


and Technology Major Project of the Ministry of Science and
Technology of China (2011ZX03004-002-01).

© The Institution of Engineering and Technology 2014


25 October 2014
doi: 10.1049/el.2013.3537
Zhi Li and Liguo Sun (Department of Electronic Engineering and
Information Science, University of Science and Technology of China,
Hefei, People’s Republic of China)
E-mail: liguos@ustc.edu.cn
Fig. 3 Die photograph of LNA
Lu Huang (Department of Electronic Science and Technology,
University of Science and Technology of China, Hefei, People’s
Measured S-parameters and NF are shown in Fig. 4. The LNA has a Republic of China)
maximum gain of 14 dB with a 3 dB bandwidth from 350 to 950 MHz.
The input return loss stays below −8 dB from 350 to 950 MHz. The NF References
is between 3.7 and 4.6 dB from 350 to 950 MHz and the average NF is
4.0 dB. The LNA consumes 0.5 mA under a 0.8 V supply. 1 Belmas, F., Hameau, F., and Fournier, J.: ‘A low power inductorless
LNA with double Gm enhancement in 130 nm CMOS’, IEEE
J. Solid-State Circuits, 2012, 47, pp. 1094–1103
16 6.0 2 Sanghyun, W., Woonyun, K., Chang-Ho, L., Hyoungsoo, K., and
14 Laskar, J.: ‘A wideband low-power CMOS LNA with positive-negative
12 5.5 feedback for noise, gain, and linearity optimization’, IEEE Trans.
10
S21 Microw. Theory Tech., 2012, 60, pp. 3169–3178
S-parametesr, dB

8
noise figure, dB

S11 5.0
6 3 Sobhy, E.A., Helmy, A.A., Hoyos, S., Entesari, K., and
4 NF
2 4.5 Sanchez-Sinencio, E.: ‘A 2.8-mW sub-2-dB noise-figure inductorless
0 wideband CMOS LNA employing multiple feedback’, IEEE Trans.
–2 4.0
–4 Microw. Theory Tech., 2011, 59, pp. 3154–3161
–6 3.5 4 Vidojkovic, M., Sanduleanu, M., van der Tang, J., Baltus, P., and van
–8
–10 Roermund, A.: ‘A 1.2 V, inductorless, broadband LNA in 90 nm
3.0
–12 CMOS LP’. IEEE Radio Frequency Integrated Circuits (RFIC)
–14
–16 2.5 Symposium, Honolulu, HI, June 2007, pp. 53–56
–18
–20 2.0
300 400 500 600 700 800 900 1000
frequency, MHz

Fig. 4 Measured S-parameters and NF of LNA

Table 1: Performance comparison of wideband inductorless LNAs


Process Frequency Gain NF Power Core area
Ref.
(nm) (GHz) (dB) (dB) (mW) (mm2)
[1] 130 0.1–2.7 20 4.0a 1.32 0.007
[2] 180 0.3–0.92 21 2–3.7 3.6 0.04b
[3] 90 0.1–1.77 23 1.8–2.3 2.8 0.03
[4] 90 0.4–1 16 3.5–5.3 9.8 0.002
TW 130 0.35–0.95 14 3.7–4.6 0.4 0.06
a
Average NF
b
Estimated core area

ELECTRONICS LETTERS 27th February 2014 Vol. 50 No. 5 pp. 400–401


Copyright of Electronics Letters is the property of Institution of Engineering & Technology
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.

You might also like