Professional Documents
Culture Documents
Low Cost Narrowband Network Analyzer: Bachelor Degree Project Report
Low Cost Narrowband Network Analyzer: Bachelor Degree Project Report
Low Cost Narrowband Network Analyzer: Bachelor Degree Project Report
Bachelor Degree Project Report Submitted by: Kaja Ameeruddin Mohammad (09007035)
Abstract
Network analyzers have become one of the most important measurement tools for characterizing the performance of high-frequency components and devices. They can provide a wealth of knowledge about a device under test (DUT), including its magnitude, phase and group-delay response etc. But high costs of such instruments makes them unaffordable, especially in engineering institutes, where the use will only be limited to testing of devices for a specific application range. This project aims to design a low cost computer interfaced network analyzer to make it more affordable and portable, operating in the frequency range 800MHz 1000MHz. The design has been started with the idea of keeping the cost low without significantly affecting the performance of the system as a whole. Since laptops, computer desktops are extremely common nowadays, their computation power can be used for analysis. A microcontroller is used for system control and the laptop/computer desktop is used for computation and display purposes. The performance of both hardware and software has been followed very closely so as to keep the performance optimal. The following report covers not only the logic involved but also the detailed working, results and prototype hardware and software used.
Table of Contents
Abstract .. i List of Figures .... iii List of Tables iv Nomenclature . iv
Introduction 1 Networks and their Properties 1 Scattering Parameters .. 2 Measurement of S-parameters .. 3 Network Analyzers Background and Cost Survey .. 5 Motivation . 6 Outline of Report .. 6 Network Analyzer Overview . 7 Network Analyzer Architecture .. 7 2.1.1 Transmission/Reflection (T/R) Test Set .. 8 2.1.2 S-parameter Test Set 9 Proposed Design Technique . 10 Hardware Specifications of SNA 12
2.2 Chapter 3
3.1 Computer Microcontroller Interface 12 3.1.1 Microcontroller Unit .. 12 3.2 Voltage Controlled Oscillator (VCO) . 14 3.3 Power Divider .. 15 3.3.1 Wilkinson Power Divider . 15 3.4 Directional Coupler . 17 3.5 Power Detectors 19 Chapter 4 Software Specifications of SNA 21
4.1 PC Software and Graphical User Interface (GUI) . 21 Chapter 5 Conclusions and Future Work .. 23 Bibliography . 24
ii
List of Figures
Figure 1.1 Two port network showing incident and reflected waves . 2 Figure 2.1 Transmission/Reflection (T/R) Test Set Network Analyzer . 8 Figure 2.2 S-parameter Test Set Network Analyzer . 9 Figure 2.3 Proposed PC-based Network Analyzer Architecture .. 11 Figure 3.1 PIC18F4550 pin Configuration . 13 Figure 3.2 Functional Block Diagram of ADF4350 . 14 Figure 3.3 T-junction Power Divider 15 Figure 3.4 Wilkinson Power Divider 16 Figure 3.5 Directional Coupler 18 Figure 3.6 Power Detector - MAX4003 pin configuration 19 Figure 3.7 MAX4003 Output Voltage vs Input Power . 20
iii
List of Tables
Table 1.1 Cost Survey of Network Analyzers 5 Table 4.1 Advantages and disadvantages of different Coding Platforms .. 21
Nomenclature
VSWR Voltage Standing Wave Ratio Vmax, Vmin Maximum and Minimum Voltage Values Reflection Coefficient T Transmission Coefficient ZL Impedance of Load ZO Impedance of the Transmission Medium IL Insertion Loss RL Return Loss
Efficiency
S11 Input port Voltage Reflection Coefficient S12 Reverse Voltage Gain S21 Forward Voltage Gain S22 Output port Voltage Reflection Coefficient
iv
Chapter 1
Introduction
Microwave and RF Networks are used in a large variety of applications today, and their uses will keep increasing in the future with the coming of next generation of networks. It is important to understand the properties of these networks as well as the transmission and receiving devices. Each of these would have parameters we can theoretically calculate, but their practical applications can only be understood once their parameters are measured.
Efficiency
|2| 1| |2
1+| | 1| |
The above equations exhaustively describe the quantities of a given network we would want to analyse. In order to simplify the calculation and measurement of these quantities we use the two-port method calculating the scattering parameters for the signal. We use matched and unmatched loads unlike a normal electric circuit. Quantities are measured in terms of power of voltage or travelling waves.
Figure 1.1: Two Port network showing incident and reflected waves
For a case like the above two-port, the relations between the input and the output ports can be shown as the following equation
VSWR VSWRin =
1+|11| 1|11|
Hence, |a1|2 = Power incident on input of the network; Power available from the source |a2|2 = Power incident on the output of the network; Power reflected from load |b1|2 = Power reflected from the input port of the network; Power available from source minus the power delivered to the network |b2|2 = Power reflected from the output port of the network; Power incident on the load
|S11|2 =
S11 =
1 1
(if a2 = 0)
|S22|2 =
2 2
S22 = |S21|2 =
(if a1 = 0)
2 1
S21 =
(if a2 = 0)
(given a1 = 0)
Considering we have found the modulus of the S-parameters, we know the values of most of the properties of the given device under test (DUT). Since we are working on a scalar network analyzer, we plot only the magnitudes of the reflection and transmission coefficients and VSWR.
Apart from its importance, the cost of a network analyzer affects its availability. Most commercially available network analyzers are costly, which make them unaffordable, especially in colleges and universities. Table 1.1 shows the cost survey of network analyzers available commercially
Manufacturer Agilent Technologies Agilent Technologies Rohde & Schwarz Rohde & Schwarz Rohde & Schwarz
Frequency Range 100kHz 3GHz 30kHz 4.5GHz 9kHz 3GHz 9kHz 6GHz 9kHz 13GHz
The cost of a network analyzer is affected by the following parameters: 1. Scalar or Vector Network Analyzer: Since a Vector VNA measures both magnitude and phase, its cost is more than that of an SNA. 2. Type of test set: There are two kinds of test sets which may be used in a network analyzer a T/R test set and an S-parameter test set. A T/R test set offers single-port calibration, while the latter offers two-port calibration. These are more costly. 3. Frequency Range: Many network analyzers work on the lower range of the microwave spectrum while others work up to a larger frequency. Consequently, the large bandwidth instruments are of higher cost. 4. Display and Accessories: High-resolution displays with accuracy, availability of markers, dynamic range, internal memory, external memory addressing, etc. add to the instrument cost.
1.5 Motivation
Cost survey suggests that most network analyzers available commercially are very expensive, which demand the need for a low-cost design. Usually, the work done in colleges, universities and even at the industry level in many places, involves a small range of frequencies. For example, cellular operators as in GSM900, CDMA (IS-95) are interested in only 800MHz-950MHz band. Bluetooth and cordless applications are in the 1900MHz2500MHz band. Students, on the other hand, require narrowband use and their requirements are not as high-end as may be required in industry. The motivation behind this project is to design a low-cost, affordable (especially for engineering colleges) yet accurate network analyzer, in the 800MHz-1000MHz band.
Chapter 2
The reference signal generated is split into two parts by means of a power splitter. Once of the arms is the unchanged reference signal, fed to the reference receiver. The other output of the splitter is routed through to the DUT, through port 1. The reflected wave from the DUT travels back through port 1 and is coupled by the directional coupler to receiver 2, while the transmitted wave travels through port 2 and is sent directly to receiver 3. Each of the receivers measures the respective incident power and sends the data to the processor.
Clearly, an incident wave on port 1 results in two wave components a reflected wave on port 1, and a transmitted wave through port 2. The measurements will give S11 and S21. To measure S12 and S22, the user simply needs to reverse the DUT connections to the two ports. The computation to calculate various factors such as reflection coefficient, VSWR etc. is done by the processor, and displayed accordingly. Note that, for a T/R test set network analyzer, calibration is possible through port 1 only. In such an architecture, port 1 is called the reflection port and port 2 is called the transmission port. Two port calibration may be done in an S-parameter test set.
In an S-parameter test set there are four receivers, two of which are for reference power measurement. At a given position of the control switch, the outputs of only one of the two reference receivers are read by the processor. The two power splitters and directional couplers are identical.
10
1. The user enters the start and stop frequencies in the PC-based GUI and initiates the data acquisition process. 2. The GUI starts the USB communication by sending a start byte, which is received by the microcontroller (MCU) 3. The MCU starts the VCO sweep. The VCO is to be controlled by the SPI (Serial Peripheral Interface) bus of the MCU 4. The VCO is the test signal generator it generates a frequency sweep, i.e. sine waves with frequency increasing linearly in the specified range, as a function of time. It is controlled to generate frequencies from 800MHz 1000MHz. 5. The output RF signal is passed through a power splitter (50%). Half the power is used as a reference power and is measured by the reference power detector. 6. The other output of the power splitter is sent to the DUT. The reflected power is captured by power detector 2, through a directional coupler (DC), while the transmitted power through the DUT is measured by power detector 3. 7. Each power detector gives a DC voltage output, proportional to their input power. The analog signals are fed to A/D converters (ADCs), which are internal to the MCU. 8. The MCU then uses these digitized values to compute the reflection/transmission coefficients and sends them to the PC through the USB. 9. When the acquisition is complete, the GUI displays the results accordingly.
11
Chapter 3
12
Advantages: Large range of power supply from 2V 5.5V On-chip USB v2.0 compliant Serial Interface Engine (SIE) Large on-chip program memory of 32kB Master Synchronous Serial Port (MSSP) which incorporates SPI 13-channel, multiplexed 10-bit A/D converter (ADC)
It is important to note that the microcontroller has several tasks to handle. These are 1. Microcontroller Programming: Even before the MCU can handle the application program tasks it need to, there must be means to burn the application program binary code into the program memory of the MCU. This is usually done by means of a debugger or a hardware JTAG emulator in case of most microcontrollers. It is important to make the programming process independent of the emulator, so as to reduce the total system cost. A USB-based programming, has therefore been developed. 2. Application Deployment: Once the MCU has been programmed, it needs to communicate over USB to obtain commands from the user (PC) to start acquiring data, control the rest of the hardware and send the measured power values back to the PC.
13
14
15
2-way Wilkinson Power Divider: Wilkinson power divider concept can be used not only as a 2-way system but can also be expanded for an N-way system. It uses quarter wave transformers to split the input signal to provide two output signals that are in phase with each other. The resistor between the two output ports enables the two outputs to be matched while also providing isolation. The resistor does not dissipate any power, and as a result the Wilkinson power divider can theoretically be lossless. In practice there are some losses, but these are generally low.
The values within the two way Wilkinson divider / combiner can be calculated: R = 2 x Zo Zmatch = SQRT 2 x Zo = 1.414 x Zo Where: Zo = characteristic impedance of transmission line 2Zo = the value of the terminating resistor connected between the two ports Zmatch = the impedance of the quarter wave transformers in the legs of the power divider
16
Working of Wilkinson Power Divider: In order to see how the Wilkinson power divider works, consider a signal entering the left hand port, port 1 in the diagram above. The signal reaches the physical split and passes to both outputs, ports two and three of the Wilkinson divider. As the two legs of divider are identical, the signals appearing at the outputs will have the same phase. This means that ports 2 and 3 will be at the same potential and no current will flow in the resistor. As the power is being split, it is necessary to ensure that the impedances within the Wilkinson divider are maintained. To achieve this, the two output ports must each appear as an impedance of 2 x Zo - the two output ports of 2 Zo in parallel will present an overall impedance of Zo. The impedance transformation is achieved by placing a quarter wave transmission line between the split and the output - the transmission line has an impedance of 1.414 x Zo. In this way, the impedance within the system is maintained. The Wilkinson divider is an ideal form of divider for many RF applications. It provides a low level of loss and maintains a high level of isolation between the output ports. A further advantage is that it can often be made very cheaply when used at microwave frequencies because the transmission line elements can be printed on the circuit board. This means that the only component required for the Wilkinson divider is a resistor.
17
Directional Coupler basics: A directional coupler is a four port device. The four ports are generally termed:
Input (Port 1, Incident) Transmitted (Port 2, Output) Coupled (Port 3, Forward coupled port) Isolated (Port 4, Reverse coupled port)
Typically the main line is the one between ports 1 and 2. Normally this may be more suited to carry high power levels and it may have larger RF connectors, if it is a unit with RF connectors. The other ports are normally more suited for lower powers as they are only intended to carry a small proportion of the main line power. Ports 3 and 4 may even have smaller connectors to distinguish them from the main line ports of the RF coupler. Often the isolated port is terminated with an internal or external matched load which would typically be 50 ohms. While specific ports are given labels on a device, this is normally more of a physical constraint as some ports will be manufactured to carry higher powers than others. In fact any port can be the input, and this will result in the directly connected port being the transmitted port, the adjacent port being the coupled port, and the diagonal port being the isolated port.
Directional Coupler Specifications: Common properties desired for all directional couplers are wide operational bandwidth, high directivity, and a good impedance match at all ports when the other ports are terminated in matched loads. The following three quantities are generally used to characterize a directional coupler: 1. Coupling Factor: The coupling factor indicates the fraction of the input power that is coupled to the output port Coupling = C = 10 log (P1/P3)
18
2. Directivity: The directivity is a measure of the couplers ability to isolate forward and backward waves Directivity = D = 10 log (P3/P4) 3. Isolation: Power level difference between port1 and port4 (related to directivity) Isolation = I = 10 log (P1/P4) These quantities are related as I = D + C dB The ideal coupler would have infinite directivity and isolation.
19
Features of MAX4003: Complete RF Detector Frequency range from 100MHz to 2500MHz Input Range of -45dBm to 0dBm Fast response Low-current consumption: 5.9mA at Vcc = 3V Vcc = 3V 5V Linear output characteristics
Below is the output voltage to input power trend of MAX4003. From the figure we see that the output voltage is a linear function of input power.
20
Chapter 4
The software used to develop the above interface has a few pre-requisites needed by the specifications mentioned above and for the ease of development: Should be able to build applications and installation files Should be easy to code different parts for easy expansion of functionality Should be able to develop a graphical user interface Deal with mathematical analysis and graph plotting
21
Language Labview
Visual Basic
MATLAB
SciLogic
Java
Advantages Easily manageable user interface Intuitive connection between the front panel and the logic blocks Technical software: easy to plot smith charts and bode plots Controls on the user interface panel can be easily customized to the real world controls Easy to develop a user interface Functions and coding similar to C++ Easy application builder Event driven programming language Integrated development environment Great with signal analysis Has very good imaging software for plotting of data Can build applications and a GUI too Technical software Has plotting devices and imaging Signal handling and analysis Freeware Freeware coding platform Universally used for application in all windows computers Signal handling too will be good User interface can be made Applications and installation files are easy to build Libraries for USB access available Libraries for plotting available Free software
Disadvantages Licensed Software, which has high investment cost Not very freely available, in case of future expansion or modification, the same version would be required Not all programmers are familiar with it Does not have mathematical plotting functions very easily Handling of signals from a USB port input requires intensive coding
It is a licensed software
wxGlade
Not a structured application builder Requires a real time LINUX platform Complex program structure Understanding a new language altogether, not similar to any of the previous Data handling is different from an of the other languages Difficult to find applets with RF handling capabilities GUI software utilities are somewhat limited Difficult to debug
Table 3.1: Advantages and disadvantages of different Coding Platforms It was decided to choose Python-based software wxGlade to design the first GUI. Considerable development of a GUI for a spectrum analyzer has been done using National Instruments LabVIEW, however, the licensing issue prevents distribution and ability for easy individual use. Hence, it seemed a worthwhile choice to attempt a design using freely available Python-based software for a GUI.
22
Chapter 5
23
Bibliography
[1] Agilent Technologies Electronics Measurement Group, Network Analyzer Basics, 2006 [2] Agilent Technologies, AN1287-2: Exploring the architectures of Network Analyzers, 2000 [3] David M. Pozar, Microwave Engineering, Wiley Student Edition, 2005 [4] Wikipedia, http://en.wikipedia.org/wiki/Main_Page [5] Debpratim Ghosh, Low Cost Narrowband Network Analyzer, IIT Bombay, October 2012 [6] Neha Rambhia, Network Analyzers, IIT Bombay, June 2011 [7] Pinguino, http://wiki.pinguino.cc/index.php/Main_Page [8] wxGlade, http://wxglade.sourceforge.net [9] PyUSB, http://sourceforge.net/apps/trac/pyusb [10] Jan Axelson, USB Complete: The Developers Guide, Penram Publications, 2009 [11] Maxim, 100MHz to 2500MHz, 45dB RF Detector in a UCSP, MAX4003 [12] ADF4350 Power Divider, http://www.analog.com/static/importedfiles/data_sheets/ADF4350.pdf
24