Low Noise Amplifier Design
Amplifier Specification:
Center Frequency Bandwidth Noise Figure Gain Source Impedance Load Impedance Relative Permittivity of Substrate () Thickness of the Substrate (h) Thickness of the conductor Transistor
2.4GHz 5% < 3 dB As High as Possible 50 ohms 50 ohms 2.31 31.5 mil 1.2 mil Fujitsu FHX35LG HEMT
Design procedure for a given Noise Figure
A single stage transistor amplifier with matching networks
at the input and output terminals of the transistor are shown
below:
Zo
input matching circuit
transistor [S]
output matching circuit
Zo
in
out
Figure 4.1 a) Check stability performance by calculating Rolette
Stability Factor K using the S-parameter of the transistor at the given frequency and plot respective stability circles to determine the potentially unstable
region in the Smith Chart for and .
S L
b) If the transistor is potentially unstable, it can be stabilized by adding a ballast resistor at the drain or a feedback resistor from drain to the source. But this
would add to the noise figure. Otherwise we can go for a conditionally stable design.
c) Calculate unilateral figure of merit for choosing
unilateral or bilateral design.
d) Calculate NFmin, opt of the transistor from the SParameter. Then choose a desirable noise figure above the NFmin. Calculate the Centre and radius of
the Noise circle. Plot it in the smith chart for plane.
in
e) For a unilateral case use the constant gain circles for the desired or maximum gain. For a bilateral case use the available power gain circle for the desired or maximum gain. Plot any of these circles accordingly in the plane.
in
f) Choose a value which is in the stable region as
S
well as with in the Noise circle, the corresponding circle gain circle. We dont use operating power gain circles as it has to be plotted in plane. So it is
out in
difficult to correlate with the noise circles in plane. g) Calculate the corresponding . From these values
L
obtain the corresponding impedance values Z and Z .
S L
h) These are the impedance values for a given noise
figure, gain etc for the transistor. Now these have to be matched with their corresponding Source and load
impedance Z (usually 50 ohms).
0
i) The impedance matching networks (Input and Output) are designed using the smith chart. For e.g. in the input side, the matching network must transform Z to Z . This can be done using lumped elements like
S 0
LC based network or distributed elements like open or short circuited stubs combined with a length of a transmission line. j) Provide DC bias for the Q point taken from the data sheets and take care that it is transparent with the RF operation.
Calculated parameters and design:
The S-Parameter file FHX35LG.s2p for the given transistor is downloaded from Fujitsu website. This file is given as an input to a RF utility software called Appcad. Using Appcad the following parameters are calculated at 2.4 GHz as:
4
S11 = 0.925 -53.0 S21 = 4.200 132.5 S12 = 0.050 52.5 S22 = 0.500 -45.0 K = 0.26 < 1 Stability: The transistor is potentially unstable. The stability circles are as given below
The unilateral figure of merit is found to be -5.53dB < (Gt/Gtu) < 19.7dB The error range is 25dB which is very high. So we should use bilateral approach. Noise and Gain circles: The Noise parameters at 2.4 GHz are calculated as NFmin = 0.42dB opt Rn = 0.8 38 = 27.4
The desired Noise Figure and the Gain is selected as 1.5 dB and 12 dB
6
The corresponding noise and available gain circles are plotted as given below
Figure 4.4
The corresponding , , Z , Z are found as given
S L S L
below
Figure 4.5
Impedance Matching networks: The Source impedance matching network should transform 50 ohms to Z and it is done as given below.
S
The load impedance matching network is as given below
The circuit is still not stable at certain frequencies. So a
ballast resistor of value 150 ohms was added at the drain.
This value was obtained by tuning and optimization in
ADS.
But this increases the noise figure to 3.128 dB which is not
acceptable. Therefore the entire circuit was treated as a 2 port network and the corresponding parameters were calculated once again. The circuit was matched once again at the source and load. The performance in terms of noise figure, gain and stability was very good after the second
matching network. Calculations for second matching: S11 = 0.806 -128.4 S21 = 4.569 -140 S12 = 0.038 146.3 S22 = 0.031 66.3 K = 1.085 > 1 NFmin = 1.177dB opt Rn = 0.761 163.86 = 2.017
S L
Then the , values are chosen for a noise figure of 2dB
and gain 11.5dB
The corresponding impedance values are Z = 5.3 + j6.643
S
Z = 33.815 + j6.211
L
The second matching networks are as given below
DC Bias Network: The operating conditions for FHX35LG are read from the data sheet as: Ids = 10 mA Vgs = -0.4V Vds = 3.0V The DC biasing network is designed as given below
11
Amplifier Simulation and Results
Schematics and Layout:
The Schematics for the amplifier drawn using ADS is shown in the next page.
The initial simulations to test the RF performance were done with the S-Parameter file of the transistor. This model doesnt need biasing circuits. But later the transistor model with a foot print was used, needs inclusion of biasing
network.
The transmission line dimensions for the Impedance Matching Networks and DC Bias Circuit was calculated
using Linecalc a utility of ADS.
After initial Simulations the schematic was transferred in to a layout specific schematic means some modifications to
ensure a proper layout were done. These were 1) Adding Tee networks for branching in the layout. 2) Adding vias for grounding.
3) Adding a small length of transmission line to connect
the pads of devices to other transmission lines.
All these modifications were done without affecting the
performance. This is shown in figure 5.2 below
12
13
14
The Layout of the amplifier was done using the layout conversion tool of ADS and it is as shown below
15
Simulation and results
The circuit was simulated for S-parameter simulation using
ADS and various parameters and their performances were
noted as given below. S21 and gain at 2.4 GHz
Figure 5.4
16
S11 and S22 (return losses):
17
18
Noise Performance:
Stability:
19
Measured results:
S11 and gain at 2.4 GHz
Figure 5.10 The measured gain at 2.4 GHz was 13.986 dB
20
Return loss S11:
Figure 5.11 At 2.4 GHz -13.306dB
Figure 5.12
21
