In these modern days where wireless applications have become important communication devices have caused the current market to require more and more wireless devices to be available for customers. Different wireless devices operate at different frequency band.For example the common basis for 3G networks will be CDMA (1.9 GHz) and Bluetooth Technology that is operating at band (2.4 GHz). These applications are required in the latest cellular phones in the market[1]. Thus, having a dual band device will enhance the mobility of the user as the device can be used in many countries as it covers more frequency bands[2].
LNA Introduction
Figure 1: LNA ina direct conversion receiver architecture.
The pre-selection filter removes out of band signals and partially rejects image band signals received by the antenna. These filtered signals will then be passed to the LNA which is a critical component in wireless devices as it amplifies the weak signal received. [3]
1.2.1 Dual Band LNA
Designing a dual band LNA will increase the usability of the wireless devices so that it can work on 2 frequency bands, thus making it more versatile and relevant to the new trend market of multiple usage devices.The dual band device will increase the mobility of the users. Many countries have different wireless standard for its cellular system, which uses different frequency bands. As explained in the introduction earlier, each application utilizes different frequency band hence the dual band LNA is required to be able to meet multiple (dual) frequency bands.Having a dual band device will also enhance the mobility of the user as the device can be used in many countries as it covers more frequency bands.
1.3 Project Aim
By doing this project, the low noise amplifier used in the microwave field can be studied and explored deeper. This project is also done to overcome the single band amplifier. The purpose of designing this dual band low noise amplifier is to open up a new range of frequencies ranging from 1.9 GHz to 2.4GHz. This opens up a new route to the academic field and also the industrial field as most of the technology in this era uses frequency bands to transmit information. The dual band low noise amplifier enables the particular equipment to receive more than one band at a time compared to the single band amplifier. The information transferred from a further destination might get lost halfway and noise might also occur when a single band chip is used together to receive multiple bands. It can also help save space in boards by using fewer chips. The design of dual band amplifier will overcome all the problems faced by currents users.
1.4 Technical Objectives
Technical Objectives
1.
To be able to know the operational theory of a single band and dual band receivers
2.
To be able to know the theory of low noise amplifier (LNA) amplifier
3.
To be able to create a dual band receiver to overcome the disadvantage of single band receivers
4.
To be able to match the impedance and noise at two frequency ranges
5.
To be able to use Advanced Design System (ADS) to simulate the results
6.
To be able to meet the specifications for the parameters in the table below:
Parameters
Specifications
Gain
≥ 16 dB
Noise Figure (NF)
≤ 2.4 dB
Unconditional stability, (K)
> 1
Input ( ) and output ()
≤ - 10 dB
Low range frequency
(1.8 GHz - 2.0 GHz)
High range frequency
(2.3 GHz - 2.5 GHz)
Chapter 2: Literature Review
The dual band LNA designing have been attempted by people with different frequency bands used for different applications. This chapter intends to give an overview of some basic principles used in the analysis and design of dual band low noise amplifier. The most challenging task is the selection of the most suitable devices for application such as the transistor in a LNA circuit. It is normally based on its characteristics and S-parameters that correlate with its performance in the whole system at a later stage. Researches on dual band LNA previously challenged by other people are tabulated in Table 1 below:
Frequency Bands
(174Hz -240 Hz) and (1450 Hz -1490 Hz)
(480MHz - 880MHz) and (1.4GHz -1.67GHz)
(868-915) MHz and 2.4 GHz)
(2.4GHz and 5.2GHz)
(2.45GHz) and (5.25GHz)
Table 1: Dual Band LNA challenged by other people
2.1 Dual Band Low Noise Amplifier Design
The topology adopted in this project to achieve the dual band frequency is to design a wideband LNA that is able to receive frequencies ranging from 1.9GHz and 2.4 GHz. Designing a dual band LNA presents a considerable challenge because of its simultaneous requirement for high gain, low noise figure, good input and output matching and unconditional stability at the lowest possible current draw from the amplifier operating at both frequencies. Although gain, noise figure, stability and input and output match are all equally important, they are interdependent and do not always work in each other's favor. The selected transistor should exhibit high gain, low noise figure, and offer high third-order intercept point (IP3) performance at the lowest possible current consumption, while preserving relatively easy matching at frequency of operation. [4]
2.2 Scattering Parameters (S-parameters) of a two port network
Figure 1: Incident and reflected waves for a two-port
As seen on the figure shown above, any travelling wave present in the circuit is made up of two components. For instance, the total traveling-wave component flowing from the output of the two-port device to the load is actually made up of that portion of which is reflected from the output of the two-port device together withthe portion of that is transmitted through the two-port device. Similarly, the total traveling wave flowing from the input of the two-port device back toward the source is made up of that portion of that is reflected from the input port plusthat fraction of that is transmitted through the two-port device. S-parameters relate to the traveling waves that are scattered or reflected when a network is inserted into a transmission line of certain characteristic impedance, (Zo).
When these observations are set in equation form, the following formula will be obtained:
Eqn. 1
Eqn. 2
where, =the input reflection coefficient,
=the reverse transmission coefficient,
=the forward transmission coefficient,
=the output reflection coefficient. [5]
From eqn. 1 , when is set to zero, then the can be obtained by the formula below,
which is a reflected wave divided by an incident wave and, therefore, it is equal to the input reflection coefficient. So now, the can be plotted on a Smith Chart and the input impedance of the two-port device can be found immediately. The same condition can be applied on the eqn. 2 by setting the equals to zero,
The value for the transmission coefficient can be obtained by the formulas below:
The and for all the above equation can be set to zero by forcing and to be equal to the characteristic impedance of the measuring system. Therefore, any wave that is incident upon or is totally absorbed and none is reflected back toward the two-port device. [5]
The input return loss ( is input reflection coefficient () expressed in decibels (dB).
The gain () is forward transmission coefficient () expressed in (dB).
The output return loss ( is output reflection coefficient () expressed in (dB). [5]
2.2.1 Stability
A major factor in the amplifier design is the potential stability of the transistor. A transistor is considered stable if there is no output signal when there is no input signal. In LNA, stability check must be done on the device before the design can be started. The stability of the Low Noise Amplifier or its tendency to oscillate at a range of frequency can be calculated using the Rollett stability factor (K) equation.Before calculating the stability of a transistor with S-parameters, the intermediate quantity must first be calculated:
=
The Rollett Stability Factor (K) is then calculated as:
If K > 1, then the device will be unconditionally stable for any combination of source and load impedance. If K < 1, the device is potentially unstable and will most likely oscillate with certain combinations of source and load impedance.[5]
The K-factor represents a quick check for stability at given biasing condition. A sweep of the K-factor over frequency for a given biasing point should be performed to ensure unconditional stability outside of the band of operation. The goal is to design an LNA circuit that is unconditionally stable for the complete range of frequencies where the device has a substantial gain. [4]
2.3.2 Noise Figure
The degradation measurement of the signal-to-noise ratio (SNR) caused by components in a radio frequency (RF) signal chains is known as the noise figure (NF). The factor is calculated from the ratio of the input SNR over the output of SNR. [6]
Where = signal-to-noise ratios of input power
= signal-to-noise ratios of output power
The noise figure (NF) is the noise factor, given in decibels (dB):
2.3.3 Power Gain
The term gain used in RF transistors is normally referred to the power gain of the device rather than just the voltage or current gain because of the myriad of impedance levels which abound in RF circuitry. When an impedance level changes in a circuit, the voltage and current gains no longer mean anything. [5]
Figure 2: A two-port network driven by a signal source and terminated by a load
The transducer power gain, is a function of the two-port network scattering parameters and of both signal source port reflection coefficient and the load port reflection coefficient . [7]
The power gain, G is a function of only the and two-port network scattering parameters. Power gain does not depend on the . [7]
where
The available power gain is a function of only the and two-port network scattering parameters. Available power gain does not depend on . [7]
where and
2.3.4 Third Order Input Intercept Point (IIP3)
When an amplifier is assumed to be linear, the intersect point between the power in the 3rd order product and fundamental tone is known as the third order input intercept point (IIP3). [8] The output of the 3rd intercept point is known as OIP3.
Figure 3: IIP3 concept Figure 4: IIP3 measurement
The diagram shown in Figure 3 is the concept of IIP3 while Figure 4 is the IIP3 measurement diagram. The IIP3 concept is the result of a two-tone test done to measure the IIP3. Two fundamental tones with same frequencies are supplied to the circuit and due to non-linearities, the 3rd order intermodulation products will be present at the output, beside the fundamental frequencies [9]. The equation used measure IIP3 derived from the diagram in Figure 4 is shown below:
where= desired output, difference between fundamental and 3rd product output.
IIPs and OIP3 is not critical in this dual band LNA design as it is located at the receiver section, the power it receives at the input is always low therefore it will always stay at the linear region.
2.3.5 Bandwidth
The bandwidth can be defined as the difference between the upper and lower operating frequencies which is typically measured in Hertz (Hz). Bandwidth is the frequency range that a signal contains. An LNA usually works between a minimum and maximum frequencies, for example the lower frequency band for this project is 1.9 GHz while the upper frequency band is 2.4 GHz therefore the bandwidth is 500 MHz.
2.5 Smith Chart
Smith Chart is one of the most useful graphical tools available to the RF circuit designer. The chart was originally conceived back in the 1930s by a Bell Laboratories engineer named Phillip Smith, who wanted an easier method of solving the tedious repetitive equations that often appear in RF theory. His solution, appropriately named the Smith Chart, is still widely in use. The chart can be used to represent many parameters including impedances, reflection coefficients, scattering parameters, noise figure circles, constant gain contours and regions for unconditional stability. [11] Figure 5 below is a diagram of the Smith chart used for impedance matching.
http://upload.wikimedia.org/wikipedia/commons/c/c3/Smith_chart_bmd.gif
Figure 5: Smith Chart
2.4 Impedance Matching
The main objective of doing impedance matching is to maximize the power transfer and minimize reflections from the load. [5] For a two-port network, is usually done by adding an input matching network towards the source and an output matching network towards the output. These networks consist of an inductor (L) and a capacitor (C) known as the lumped elements. An example of input and output matching for a two-port network is shown in Figure 6.
Figure 6: Example of Input and Figure 7: Placing Lumped Components Out Matching Network [12] Basedaccording to Smith Chart Lines [12]
The series or shunt placement for lumped components in a matching network is based on the direction where the current point is matched towards the center (50Ω) in the smith chart. The placing of components is shown in Figure 7. Matching procedure in this project will be done through an optimization feature which is available in the simulation software itself. Due to the rule of thumb set in the industrial standard, the specification set for and in this project is
2.5 Transistors Selection
A transistor is a semiconductor device used to amplify and switch electronic signals. It is made of a solid piece of semiconductor material, with at least three terminals for connection to an external circuit. Transistors made of different materials have different names such as those prior developed that are made of silicon materials are categorized as (Si) and germanium materials are categorized as (Ge) while the modern microwave transistors such as those made of gallium arsenide are categorized as (GaAs) and indium materials; (InP). [13]
Microwave field-effect transistors (FET) escpecially GaAsFET have better advantages as it also include metal semiconductor FETs (MESFET) thus they're also called high electron mobility transistors (HEMT). Therefore, the GaAs FET transistor is chosen to be used in this project in designing a LNA with wide bandwidth that is able to cover the dual band (1.9 GHz and 2.4 GHz) desired.
The Avago ATF-55143 GaAs FET transistor has been chosen as it is an Enhancement Mode Pseudomorphic HEMT (E-PHEMT) which can provide high gain, high linearity, low noise and low power output that is suitable for wireless applications.
Chapter 3: Design Methodology
This chapter depicts on the design development process based on the concept explained in chapter 2. The software used in designing and simulating the result of this project is the Advanced Design System (ADS).
3.1 AdvancedDesign System Software Overview
Advance Design System (ADS) is electronic design automation software system which is created by Agilent EEsofEDA, a Development team of Agilent Technologies. It supports RF design engineers in developing all types of RF designs from simple to the most complex of RF microwave modules. ADS let designers fully characterize and optimize designs.[14] Post processing capabilities available allow data manipulating using custom expressions, data viewing on different plots and specifications changing, all without re-simulating. The software library contains examples of template, pre-configured schematics, data displays, and test benches in helping design verification against measurements defined in the wireless standards specifications. Thus the ADS software is the preferred design tool selected for this project.
3.2Dual Band LNA Design Procedure
Transistor Selection Based on Specifications
Design Biasing Circuit
Measure S-parameters using AC circuits
Check Stability Conditions of Transistor
Design Matching Network
Simulate Results Using ADS software
Optimization
3.1 Transistor Biasing
The first procedure starts with biasing the selected transistor to ensure that when it is DC biased at a proper operating point, it is able to achieve the current requirements that will later be supplied to the whole circuit.
Figure 8: FET_Curve_Tracer Connection with ATF-55143
The FET_Curve Tracer from ADS template is used to perform the transistor biasing procedure. By simulating this circuit, the I-V curve characteristics of the biased transistor will be shown. The purpose of this I-V curve is to determine the operating point for the selected and .
3.2 DC Biasing Network Design and Simulation
A biasing network is needed for this LNA circuit in order to produce a constant value of voltage and current to the transistor. The passive biasing method is chosen to bias the ATF-55143 E-PHEMT transistor due to its simpler design because additional transistor or not needed as the current course. An active will require extra transistor, which will be more expensive to the design.
This biasing is accomplished by using a voltage divider consisting of two resistors, R1 and R2. The voltage for the divider is derived from the drain voltage which provide a form of voltage feedback through the use of R3 to help keep drain current constant. The values of resistors R1, R2 and R3 are calculated using the formula shown below [15]:
where
is the power supply voltage
is the device drain to source voltage
is the desired drain current
is the current flowing through the resistor voltage divider network
The is chosen to be at least 10X the normal expected gate leakage current (rule of thumb). In this case, it was conservatively chosen to be 0.5 mA. Calculations are done by replacing the values of obtained from the result of FET_Curve_Tracer shown in Chapter 4, Figure 16 into the formulas shown above.
With the resistors values obtained from the calculation above, the transistor bias network is constructed as shown in Figure 9 below.
Figure 9: DC biasing network design using lumped components.
The DC biasing performance is done using DC simulator while the results can be checked by the DC Annotation feature that is available in the Simulate drop down in ADS software itself. Additional tuning was done by reducing the resistor and to achieve the required bias condition of .
3.3 Parameter Analysis
Additional circuitry using capacitors and inductors are combined to the DC biasing network to form a typical ATF-55143 LNA with passive biasing circuit as shown in Figure 10 below. According to the datasheet in Appendix A, the high pass impedance matching network consisting of L1/C1 and L4/C4 contribute to noise figure, gain, matching while capacitor C2 and C5 provide low impedance in-band RF bypass for the matching networks. Capacitors C3 and C6 provide low frequency RF bypass for R3 and R4 that provide low frequency termination. The value of C3 and C6 also provide termination for low frequency mixing products that will affect the IIP3 results. All these components are set to Discrete Optimization mode from the component properties.
Figure 10: Typical ATF-55143 LNA with passive biasing circuit.
Two ports instead of terms were connected to the circuit and the terms were deactivated in order to simulate the noise figure. The whole circuit was compressed into a schematic symbol by using the Create/Edit Schematic Symbol function. This compression will turn the whole circuit into a component and stored in the software library.
In order to simulate the S-parameters of this circuit, the component must be connected to the S-param template which can be obtained from the ADS software template library. The templates and components are as shown in Figure 11 below.
Figure 11: Parameter Analysis Circuit.
Goal for each specification were set in the template after the circuit in Figure 11 above is constructed. The optimization and goal templates were taken from Optim/Stat/Yield/DOE palette while S-parameter and stability factor templates were taken from the Simulation-S_Param palette. The specification settings are shown in Figure 12 below.
Figure 12: Specifications Settings for Dual Band LNA circuit.
3.4 Impedance Matching Design
The output matching network consisting of a LC tank (an inductor and a capacitor) was added into the circuit as the simulation results for output return loss () shown in Chapter 4, Figure 18is far from reaching the specification set.
Figure 13: Output Impedance matching circuit added into the circuit.
A combination of shunt inductor and series capacitor was chosen to match the frequency to the center point therefore this matching network was added to the circuit as shown in Figure 13 above. The values for these two components were set to be optimized to get the value of. The simulation results for the output matching network added to the circuit are shown in Chapter 4, Figure 19.
3.5 OIP3 Simulation
The HB2Tone template from ADS software template library is used to simulate the OIP3 value. The settings for harmonic balance template were set as shown in Figure 15 below. Two frequency simulations will be done from this circuit. The first simulation will be set with frequency 1.9 GHz while the second with 2.4 GHz as these frequencies are the main specifications for this project. The frequency settings will be keyed in the variable equation template which is high-lighted in red in the diagram below.
Figure 14: IIP3 Simulation Circuit.
3.6 Current Drain Measurement
The current drain is very important as it is the main current that ensuring the transistor produces a gain to the signal. Gate current is to turn on the LNA, but the Drain current is the one that amplifies the input signal. Without the current drain, it won't be called an amplifier. Therefore, a probe is placed in the circuit as shown in Figure 16 to measure the current drain of the whole circuit. The node which connects both voltage supply and the probe must be named. In the circuit shown, the node is named "a". At the same time, the three nodes from transistor are named VG, VD and VS respectively for simulation in order to obtain the of the final circuit.
Figure 15: Current Drain Simulation Circuit.
As explained in the DC biasing network previously, the results can be done by simulating the circuit using DC simulator and results checked using DC Annotation. The results will be shown in Chapter 4.
Chapter 4: Results & Discussions
4.1Transistor Biasing
Figure 16: FET_Curve_Tracer Simulation Result (ATF-55143)
The results obtained from the FET_Curve_Tracer showed that when the ATF-55143 transistor is biased at and , the selected .
4.2 DC Network Biasing
The DC Annotation simulated results were plotted into a table in the simulation results shown in Figure 17 below while a comparison between the simulated results and biasing requirement were tabulated in Table 2.
Figure 17: DC Network Biasing Results
Comparison
Transistor Biased
2 V
20 mA
0.55 V
Network Biased
2.008 V
20.60 mA
0.515 V
Table 2: Comparison of Network Biasing Results
As shown in the comparison table above, the results from biasing network has met the transistor biasing requirement thus this circuit is suitable to be used as a biasing network for the Dual Band LNA circuit.
4.3 Parameter Analysis Results
Figure 18: Parameter Analysis Results
Simulated results for the parameter analysis are shown in Figure 18 above. Results obtained from simulation are tabulated into Table 3 below to make a comparison between the results and specifications set for this project.
Specifications
Aim
1.9 GHz
2.4 GHz
Gain (S21)
≥ 16 dB
17.524 dB
15.836 dB
Input Return Loss (S11)
≤ -10 dB
-8.085 dB
-11.780 dB
Output Return Loss (S22)
≤ -10 dB
-6.060 dB
-7.344 dB
Noise Figure, NF
≤ 2.4 dB
0.268 dB
0.290 dB
Stability, K
>1
0.903 dB
1.021 dB
Table 3: Simulated results comparison towards specification.
Based on the comparison shown in Table 4, most of the parameters have or near to achieving the specifications set. The only parameter which is far from reaching the specification is the output return loss (). Thereforean output impedance matching network would be needed for the circuitin order to improve the poor result.
4.2: Output Impedance Matching Results
Figure 19: Impedance Matching Results
Simulated results for the output impedance matching network added to the circuit are shown in Figure 18 above. Results obtained from simulation are tabulated into Table 5 below to make a comparison between the results and specifications set for this project.
Parameters
Specifications
1.9 GHz
2.4 GHz
Gain (S21)
≥ 16 dB
17.337 dB
16.147 dB
Input Return Loss (S11)
≤ -10 dB
-11.343 dB
-13.194 dB
Outputt Return Loss (S22)
≤ -10 dB
-16.204 dB
-18.327 dB
Noise Figure, NF
≤ 2.4 dB
0.268 dB
0.290 dB
Stability, K
>1
1.229 dB
1.165 dB
Table 4: Simulated results comparison towards specification.
Based on the comparison shown in Table 5, all parameters have achieved the specifications set. The insertion loss (); for both frequency is more than 16 dB and input and output return loss () and () are lower than -10 dB while NF for both frequencies are less than 2.4 dB. This circuit is unconditionally stable as the stability factor; K for both frequencies is more than 1.
Figure 20: Smith Chart Results
The smith chart on the left side shown in Figure 19 above is the simulated result before adding the output matching network while the smith chart on the right shows the result after the matching network has been added to the circuit. Theoretically, the () points should be matched to the center point which is 50Ω but since the output return loss () specification set is ≤ -10 dB, and the optimization process set for the two components ( LC tank) have already met the required specification, there is no need to further match the towards the center point.
4.3 OIP3 Simulation Results
Figure 21: OIP3 result with fundamental tone of 1.9 GHz.
From the OIP3 results at fundamental 1.9 GHz shown in Figure 20 above, the value were substituted into the formula shown in Chapter 2 for IIP3 measurement. The calculations are shown below:
Figure 22: OIP3 result with fundamental tone of 2.4 GHz.
From the OIP3 results at fundamental 2.4 GHz shown in Figure 21above, the value were substituted into the formula shown in Chapter 2 for IIP3 measurement. The calculations are shown below:
. 4.4 Current Drain Results
The current drain, results are plotted in table shown in Figure 23 below while a comparison table for the transistor biased values, network biased value and final circuit simulation value are tabulated in Table 5.
Figure 23: Current drain and operating point results.
Comparison
Transistor Biased
2 V
20 mA
0.55 V
Network Biased
2.008 V
20.60 mA
0.515 V
Final circuit simulation
1.999 V
20.80 mA
0.513 V
Table 5: Comparison of Network Biasing Results
As seen in the comparison table shown above, the value of , and slightly varies a from one another. These changes did not affect the overall circuit as all the gain and other specifications are still met.
4.5 Performance Comparison
Parameter
Performance at 1.9 GHz
Performance at 2.4 GHz
Ideal
Discrete
Ideal
Discrete
Table 6: Performance Comparison between Ideal and Discrete Component Value
Chapter 5: Conclusion
The low noise amplifiers (LNAs) are important components as they are the first amplifying stage that the signal will meet once received from the antenna in a receiver chain. During the process of doing this project, it is learnt that microwave field-effect transistors are useful in designing a dual band LNA aimed for wireless applications as they produce low noise, high gain and most importantly low power consumption. It is important for the LNA to be able to reduce noise as much as possible and at the same time amplifies the desired signal to the following stage in a receiver chain as unwanted noise are major hindrance in the receiver system.
As for the software, ADS is a powerful simulation tool due its many template examples and pre-configured schematics available in the software library, the calculation and designing process has been made easier. A lot of new tools in this software which were very useful throughout the design process have also been learnt.
The final circuit simulated performance has been tabulated in Table 7 below:
Parameters
Specifications
1.9 GHz
2.4 GHz
Achievement
Gain (S21)
≥ 16 dB
17.337 dB
16.147 dB
Achieved
Input Return Loss (S11)
≤ -10 dB
-11.343 dB
-13.194 dB
Achieved
Outputt Return Loss (S22)
≤ -10 dB
-16.204 dB
-18.327 dB
Achieved
Noise Figure, NF
≤ 2.4 dB
0.268 dB
0.290 dB
Achieved
Stability, K
>1
1.229 dB
1.165 dB
Achieved
Table 7: Final Results
As a conclusion, the dual band LNA design for 1.9 GHz and 2.4 GHz can be considered as a successful design since all the parameters given in the specifications are satisfied.
There were problems faced while doing the design for the LNA. The simulations results for the LNA design with ideal components and discrete components were different. The simulation result for the output return loss () were not satisfied that is caused of the tradeoff between gain, noise figure and stability. There are times at the starting stage, often when one specification is met; the others will be out of specifications. For example, when the components are optimized for Noise Figure, the other parameters such as gain, stability and input/output return loss will be affected. A lot of time was spent on optimization of the components in the dual band LNA design.
Chapter 6: Recommendation
There are a few recommendations for future work such as:
Inclusion of the harmonic balance or IP3 (third order interception) in the design.
The input return loss matching network can be added to the circuit and both input and output return loss ( can be further matched to 50 Ohm.
Adding another stage in order to cover a wider range of frequency bands.
The final layout of the design also can be fabricated to a PCB board by adding microstirp transmission lines and tested at real time using spectrum analyzer and then the result of the end product and at the simulation level can be compared.
