High-Frequency CMOS LNA circuit

The LNA is the first block of the RF receiver front-end. Thus, LNA must provide sufficient gain to suppress the noise distribution form the subsequent stages and 50Ohm input impedance matching. Gain can be provided by a single transistor. There are three topologies for a single transistor, as shown in Fig. 2. Each one of the basic amplifiers has many common uses and each is particularly suited to some tasks and not to others.

Fig. 2 Single transistor amplifier [3].

The common-source amplifier is most often used as a driver for LNA to provide gain. The common-drain amplifier, with high input impedance and low output impedance, makes an excellent buffer between stages or before the output driver. The common-gate amplifier is often used as a cascade in the combination with the common-source to form an LNA stage with gain to high frequency, but it can be used by itself as well. Since the common-gate amplifier has low input impedance when it is

driven from a current source, it can pass current through it with near unity gain frequency. Therefore, with an appropriate choice of impedance levels, it can also provide voltage gain [3]. The cascode LNA is shown in Fig. 3 and Fig. 4 is the small signal analysis of cascode LNA.

Fig. 3 Common-gate amplifier used a cascode transistor in the LNA [4] .

Zin

Ls Lg

C_gs ^gmVgs +

Vgs

-Fig. 4 The input impedance analysis of cascode LNA.

From the Fig. 4, we can easily straightforward analyze the input impedance of the cascode LNA.

s gs m gs g

s

in L

C g L sC

L s

Z 1 ¹

)

( + + +

= . (1 ) Note that Ls contributes a real term to the input impedance through interaction with Cgs

and gm1. By choosing Lg+Ls to resonate with Cgs to create conjugate matching at the input. The inductor LD provides significant voltage gain. The common-gate transistor of the cascode LNA, M2, plays two important roles by increasing the reverse isolation of LNA: (1) it lowers the LO leakage produced by the follower mixer and (2) it improves the stability of the circuit by minimizing the feedback from the output to input [2]. But the topology of the common-source with inductive degeneration will degrade it performance substantially at higher frequency comparable to ωT due to the noise factor, Fmin and effective transconductance, Gm, are linearly related to the working frequency, ωo and 1/ωo, respectively [5].

Lg

Ls M1 M2 Vb

V_DD

Vout

(a)

Fig. 5 The high frequency effect of the cascode LNA topology [6].

Above 20GHz, the pole at the drain of M1 of the cascode LNA showing in Fig. 5 shunts a considerable portion of the RF current to ground, thereby lowering the gain and raising the noise contributed by M2. Furthermore, the small degeneration and gate series inductances (50-150 pH) required for the input matching make the circuit very sensitive to package parasitic. The above observations suggest that the LNA must contain a single transistor before voltage amplification occurs [6]. Therefore, all of the high frequency LNA circuit design for applications at high-gigahertz range must use a single stage transistor as a first stage amplifier to provide sufficient gain amplification.

For example, the 3-stage common-source amplifier is also a popular topology of LNA [7].

Fig. 6 The simplified schematic of 3-stage common-source LNA [7].

Because the gate-source and gate-drain parasitic capacitances of the common-gate (CG) LNA are absorbed into the LC tank and resonated out at operation

frequency. Due to the constraints of input matching, the CG LNA has a lower bound of 1+γ for perfect input match, where γ is the channel thermal noise coefficient. Therefore, to the first order, the noise and gain performance of the common-gate stage are independent of the operation frequency, which is a desirable feature for high frequency design [6], [8]. For example in the [8], a 24-GHz CMOS LNA is designed with common-gate with resistive feedback (CGRF) topology. It adds an external resistor, Rp, to the traditional CG LNA in parallel with the input transistor to improve its noise performance, as Fig. 7 shows.

Fig. 7 Common-gate with resistive feedthrough LNA. (a) Schematic. (b) Small-signal equivalent circuits [8].

Fig. 8 shows the 24-GHz CMOS three stage LNA with a CGRF topology as the first stage [8]. The first stage employs CGRF topology, where shunt inductor L2

resonates the capacitive coupling while introduces a feed-through resistance between drain and source of M1. A capacitor C2 isolates the dc level of source and drain. The second and third stages are both common-source with inductive degeneration amplifiers which are used to enhance the overall gain.

Fig. 8 Three stage LNA with a CGRF topology as the first stage [8]

In recent report about high frequency LNA topology, all of them use a single transistor as the first stage and are realize by voltage-mode or partial-voltage operation. Therefore, we make an attempt to use the current-mirror amplifier to perform a LNA. The current-mirror is also a single stage transistor, not a cascode topology. Because it can bias itself, it doesn’t need extra biasing voltage point.

Therefore, we use two stage current-mirror amplifiers to realize a LNA. The first stage is a current-mirror with an inductor shunt feedback to provide an input matching and lower the noise figure. The second stage is also a current-mirror amplifier which is

used to enhance the overall gain level. The current-mode LNA with a current mirror topology would be described in chapter 2 particularly. And it has a power gain of 17.1 dB and a noise figure of 3.4dB consuming 9mA from 1.2V supply voltage in the post-simulation result.