EECS 452 Digital Signal Processing Design Laboratory Winter 2008 Homework 3b ? January 4, 2008 All work must be individual. Hand written homework solutions will not be graded. Due January 23rd. No homework accepted late. Homework may be turned in early. Homework is to be submitted in the form of a PDF file sent by email to email@example.com. The file should be named hw3b_your_name. Where your is replaced by your last name and name is your first name. For example I would use hw3b_metzger_kurt. Include your full name in the text of your submission. Do not include your student ID number or any other identifying information. The purpose of this homework is to generate a pool of project ideas that can be used when defining projects and forming project teams. The homework solutions will be collected together and printed copies will distributed to the class on the Friday following the due date. The printed copies are intended to get people to start thinking about what projects they would like to work on. We will have an evening class meeting (Tuesday, January 29 starting at 8 PM). where projects will be defined and group membership will be decided. The projects involve: ? design specification; ? application of material learned from earlier courses (say EECS 306, EECS 451, EECS 452, and others such as EECS 270, EECS 373 and the like); ? real world issues involving cost, reliability, safety, and production; ? evaluation and/or testing regard to the design specification and the real world issues. Basically a project is to involve the creation of a product starting with an idea and taking it to the proof-of-concept or prototype stage. Because this is EECS 452 a primary component of the project is the application of digital signal processing. Some things to keep in mind are: ? Digital signal processing is typically an enabling technology rather than an end goal itself. That is, DSP makes products possible or better than they would otherwise be. ? Because DSP is an enabling technology this means that the projects will mostly be shaped by other technologies. For example if a project is to de- sign a configurable software radio then an understanding of the following areas would be essential: Homework 3b 1 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 ? modulation theory ? basic radio design practice If you do not have a background appropriate for a particular project then will have to either acquire it or look into another project. ? Be careful in setting the scope of your projects. A project needs to be more than trivial but still within reach of available resources. Projects need to be doable with about 8 week?s part time effort by a small team. Start early!!!! ? A project does not have to be targeted to end users. A project can create somethingthat is to be incorporated into other products. An example is de- velopment of a speech recognition chip or technology. For example, there is a voice controlled robot on the toy market that uses speech recognition technology that was purchased rather than created by the toy maker. In the cell phone industry Qualcom sells its CDMA technology to cell phone system creators. The Fraunhofer Institut sells its MP3 technology in the form of actual code for a variety of DSP processors. The economics asso- ciated with such intellectual property (IP) are a bit different than those of end products but the goal is still the same, to make money. 1. (20pts) Suggest two projects that can be done using the C5510 DSK or some other hardware such as a FPGA. There should be a strong DSP aspect to your suggestion. Lookingatthelistofapplication notesavailable fortheTMS320C55x and/or the older TMS320C54xx family on TI?s web site should provide an abun- dance of ideas. You might also review the hardware resources available for the TI C5510 DSK. We can add daughter boards. Last semester a project designed and built a printed circuit board that connected to the C5402 DSK. Try to suggest a project that appears feasible using the DSK either to implement a proof-of-concept model or actually to implement a prototype. Try to estimate the hardware and software resources necessary and relate them to the resources available on the DSK or whatever device suggested. Each suggestion should (i.e, must) address the following: a: What the basic product that is being proposed. b: What are competing products? How would yours differ? Why would one purchase it? c: Scope of the EECS 452 effort for this project. d: If all went well, what would the demonstrable result be at the end of the semester? Homework 3b 2 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 The suggestions should not require much more than a few paragraphs each. I (KM) got carried away on my suggestions. Keep a hand on reality when setting the scope of the work to be done. Starter ideas Here are some ideas for projects that being put forth in order to start things going. These are developed in much greater detail than expected of your project suggestions. If you have a particular project that you want to pursue you might want to put in some effort into your suggestion in order to interest others in working with you. Frequency Domain Reflectometry Appliation This semester there is a small start-up located near Lansing that would like to partner with an EECS 452 project group in order to develop a patent they have the rights to into a working product. The patent is 7,071,7000 titled Characterizing analog and digital telephone circuits and other types of wiring systems using frequency domain reflectometry (FDR). A possibly related patent assigned to Harris Corp is 6,744,854 titled Detection of bridge taps by frequency domain reflectometry-based signal processing with precursor signal conditioning. This patent includes a good discussion about how FDR works. It would nice to form a group early for this project and have them meet with the principals of the start-up in order to learn about what would be wanted and expected. Through-the-earth communications system There is interest by Dr. Marlin Ristenbatt and myself to revive work on a low data rate through a the earth communications for use by cave explorers and trapped miners. This would pick up at the point documented in A New Post- Disaster Mine Communication System, Ristenbatt, Holland-Moritz and Metzger, IEEE Transactions of Industry Applications, Vol. 24, No. 2, March/April 1988. This is a sort of a Over-the-Hill Gang effort that might spring to life with the proper motivation. Homework 3b 3 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 An idea seed This first item isn?t really a project idea but more an idea starter. As pre- sented, use of the C5510 is NOT assumed. It is relatively straight forward to implement a text and graphic display device in what today is a relatively small, low-cost FPGA. Indeed such a demonstration of concept implementation has been created. The unit uses a ($149) Digilent Spartan-3 Starter Board and an inexpensive ($100) 17 inch color LCD monitor as the display device. Resolution is 1024 by 768 pixels using a 70 Hz refresh rate. Approximately 8% of the FPGA fabric (the -1000 part) is used. This leaves lots of room for other tasks. By adding A/D and D/A converters it should be (look out for ?should be?) possible to design a universal signal processing device, at least at audio frequen- cies including display. The USPD should be able to: ? generate arbitrary waveforms including standards such as the sine wave, square wave, etc. ? function as an oscilloscope. ? Make filter transfer function measurements, amplitude as well as phase. ? generate and display spectra in real time. Xilinx includes a FFT core in it?s free WebPACK version of its development tools. ? do much more. The Xilinx Spartan 3 starter board is very inexpensive ($99 ? $149 depend- ing on FPGA size) board sold by Digilent Inc. (www.digilentinc.com). This board contains a 4 digit numeric display, push button switches, slide switches, connectors for a PS2 peripheral and a VGA display (user figures out how to im- plement the logic in the FPGA), and an RS-232 DB9 connector. Digilent also sells peripheral boards that connect to the starter board. These include dual channel 1 MHz A/D converters, dual channel 1 MHz D/A converters, H-bridges and the like. Multiple peripheral boards can be used on a single FPGA starter board. Xilinx is nominally also has available a Spartan 3E starter board ($149). This boards also has a 20x2 LCD display. Digilent also have a number of interesting boards that might be used. Of particular interest is the FX2 board that boasts a Xilinx Virtex-4 FPGA. This device includes a PowerPC in addition to the FPGA fabric. This board is sold by NuHorizons and costs $300. Homework 3b 4 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 Software Defined Radio (SDR) Application: Location Dependent Services (LDS) Goal Product: In mobile communications, an important class of applications involves the location dependent services (LDS) of a mobile user. Two major issues must be addressed to deliver appropriate services to mobile users. One is the availability of a good location-positioning system and the second is an infrastructure that enables the positions to be tracked for the perusal of users? application. The primary goal of this project is to apply DSP techniques to achieve LDS as an SDR application. Competing Products: Location dependent services can be provided to mobile users in different ways, depending on the mobile device. Currently the most popular way to pro- vide LDS is with the Global Positioning System (GPS). However, GPS devices are expensive. Besides, GPS signal can easily be blocked if there is no direct line- of-sight communication link between the device and the satellite. Another way is to use the cellular network. With the current infrastructure, cellular carriers can provide their customers with LDS by software with extra hardware cost. LDS can also be provided for devices with wireless connection such as laptops or pocket pcs in a Wi-Fi environment. This is a similar concept as SDR. Possible applications include the following (and many more): ? Optimum driving directions (avoid traffic, road work, etc.) ? Finding the nearest parking space ? Finding restaurants, stores, theaters, etc. Scope of the EECS 452 Efforts: ? Develope/implement signal processing algorithms for wireless positioning. ? Develope/implement signal processing algorithms for wireless services (eg. route selection). ? Test the algorithms by simulation (Matlab/C programming). Homework 3b 5 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 ? Add wireless capability to the TI DSK. ? Implement the algorithms using the TI DSK. Demonstrable Results: Several DSK?smight be needed in this project. The firstpart is to demonstrate the positioning capability of the system. The second part is to demonstrate how the services can be delivered once the location of the mobile device is known. Discussion: In the first part, the real problem is how to get access to the location infor- mation describing the position of the target device. The accuracy of the location estimation may affect how precisely the services can be delivered. There are several location estimation algorithms, and many of them are based on the mea- surement of RSSI (received signal strength indication). Here is the description of RSSI from WIKIPEDIA: ?...RSSI is generic radio receiver technology metric, which usually is invisible to the user of device containing the receiver, but is directly known to users of wireless networking of IEEE 802.11 protocol family. RSSI is often done in the IF stage before the IF amplifier. In zero-IF systems, it is done in the baseband signal chain, before the baseband amplifier. RSSI output is often a DC analog level. It can also be sampled by an internal ADC and the resulting codes available directly or via peripheral or internal processor bus....? The are off-the-shelf RF module products that can provide RSSI measurement for a wireless link. Assuming we can obtain the RSSI, the next question is how to use the RSSI measurements to estimate the location of the mobile device. You should do an extensive survey and figure out which algorithm is most suitable for implementation using the TI DSK in this project. As to the second part, algorithms should be studied/developed depending on the services you want to provide in the system. Note: Currently, there is a student team in EECS working on a SDR competition project related to this topic. Please see ?Problem 3: Traffic Management? on the followingwebpage fordetails. (http://www.radiochallenge.org/SampleProblems.html) Homework 3b 6 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 If you are interested, please join the team and make your 452 project as part of the competition. Chih-Wei Wang Audio Equalizer Goal Product: This project would design a complete Audio Equalizer for mobile phones using TI DSP that could be sold as a black box feature for the post-processing of audio signals obtained from any audio decoder. It should be able to do the following: ? provide control to the user to specify the numberof bands for the equalizer ? provide control to set the cut-off frequencies, gain and Q factor for each band ? should accommodate on the fly changes to the equalizer settings ? If possible, provide with a graphical display of the spectrum thats been played out. Competing Products: I (Suresh) am not aware of any products that sell their equalizer block sep- arately for mobile phones. There are other products like Winamp, real player and windows media player which have their own audio equalizer. But almost all of these are for use on general purpose PC and hence mostly floating point implementation. It is not possible to have floating point operations on a mobile phone. Hence these are not suitable for the market we are targeting. Scope of EECS 452 effort: ? research the various algorithms available for audio equalization ? select the best that can be implemented on the DSK ? MATLAB simulations to prove the working of the selected algorithm ? DSP implementation of the algorithm to take care of the requirements metioned above. Homework 3b 7 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 Demonstrable results: ? an audio equalizer suite that can be used in the post-processing of any audio decoder ? ability to respond to the equalizer changes on the fly. ? if possible, display the audio spectrum on the oscilloscope or LCD display DSP based LORAN navigator Goal product: The ultimate goal is to develop a lightweight, battery operated, hand holdable navigation device using the LORAN positioning system and the TI C5510 DSP chip. This product would be used by hikers and other outdoor people who find use of similar GPS devices to be unreliable in heavy foliage and in urban canyons. Competing products: There are a number of hand held LORAN units available. Most of these are older designs no longer in production. Our design will differ in that it ex- ploits DSP theory and hardware. Our most likely competition is from a company named, Locus Incorporated. Scope of the EECS 452 effort: The scope of the EECS 452 effort consists of: ? the development of accurate positioning algorithms using MATLAB, ? the implementation of these algorithms in the C5510 DSK, ? a block diagram system design. ? a survey of possible antenna technology. This project is in reasonable reach of the EECS 452 DSP systems. We have an DSK compatible A/D converter that will sample at 1 MHz. There is a VLF antenna located on the roof of EECS that can pickup the LORAN signals. We have access to the LORAN RF waveform using an Austron 2000 LORAN receiver. Hardware and software is available that can be used to spool data sets into PC files for use in MATLAB studies. Demonstrable results: Homework 3b 8 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 ? The results of MATLAB test runs made using with digitized actual LORAN receptions, ? A C5510 prototype will be used to determine the location of a LORAN an- tenna located on the EECS building roof. This will be compared with the location of a GPS antenna located a few feet away. Discussion: Before GPS there was LORAN. GPS denotes the space based global positioning system and LORAN denotes the land based (long range navigation) system. GPS LORAN transmitter power low (about 50 W) high (400 kW and 800 kW) transmitters based in space on land carrier frequency 1575.42 MHz 100 kHz dimensions 3 (x, y and z) 2 (x and y) accuracy cesium locked cesium locked time of day yes no annual cost $ 400×106 $ 28×106 position accuracy very good not quite as good? blockage easily blocked robust The GPS system was to have eliminated the need for the LORAN system. The LORAN system was to have been closed down by the year 2000. However, events warranted continuing its operation until the year 2008 (at the least). The primary events causing this were ? Someone at an air show in Moscow in 1997 jammed GPS over an area ap- proximately 300 square kilometers. This illustrated how easily this could be accomplished. ? In 1998 the FAA, contrary to prior statements, declared that commercial aircraft navigating using GPS must have a back up system. LORAN is prob- ably the only logical candidate. The U.S. government is to make and announce a decision on the fate of the LORAN system sometime in the near future. Even though the LORAN system had been planned to be shut down congress has kept providing funds for mod- ernization. As a result the system has become much less costly to operate and more accurate. It has been claimed that the limiting factors in LORAN position accuracy are associated with the transmitters. Using modern processors and DSP algorithms the receivers are not the limiting factor. Homework 3b 9 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 This project is a bit speculative. Because of the planned shutdown(s) and the increased use of GPS many manufacturers of LORAN equipment have stopped producing LORAN receivers and, at the least, stopped improved receiver devel- opment. But not all. This would be a good project for a grad student having a background of statistics and random processes to participate in. Realtime Spectrum/Network Analyzer Goal product: To develop a DSK based real time spectrum analyzer and network analyzer. The system is intended for use in DSP teaching labs for making real time spec- trum plots, measuring transfer functions, and making delay measurements. The software will be open source facilitating it?s modification for other applications and for teaching use. The proposed product will be stand alone and will include a small LCD graph- ical display to display results. It will also be possible to communicate results and data between a PC and the unit using the Code Composer Studio interface. The operating code will reside in flash EPROM allow ready updating. The goal product is intended for use in teaching DSP. The hardware and soft- ware could be bundled along with a text. The range of educational topics in- cludes program design, realtime systems, digital signal processing basics, spec- trum analysis, hardware design, and much, much more. Competing products: Competing products include: 1. SigLab. A SigLab unit consist of: ? a hardware box that contains A/D and D/A converters, memory and a floating point DSP chip, ? a SCSI interface used to communicate values to a PC running Windows and MATLAB, ? a set of MATLAB scripts that implement various processing algorithms and generate interactive graphical interfaces. Homework 3b 10 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 In order to use SigLab you not only have to purchase the SigLab hardware you also need to purchase or license a PC, an Adaptec SCSI interface, Win- dows and MATLAB. There have been significant problems migrating the PC based software to updated versions of Windows and MATLAB. 2. National InstrumentsLabview/Dynamic signalacquisition hardware bundle (about $7K). This consists of ? An National Instruments A/D and D/A card that plugs into a PC?s PCI bus. ? LabView data acquisition software. ? Application specific support running over LabView. 3. Agilent Technologies 35670A (about $18K) Dynamic Signal Analyzer. This is a very versatile top-end product but not inexpensive. Scope of the EECS 452 Effort: A significant part of this project would be learning how to properly compute the real time spectra and the transfer function magnitudes and phases. Resources available for this project include: ? the TI C5510 DSK ? the Tektronix 2642A manuals. ? the SigLab manuals ? Agilent application notes discussing spectrum analysis and network ana- lyzer operation. ? National Instruments LabView tutorials and application notes. ? Product information from manufacturers selling products that could be use in this application. The cost of a C5510DSK is under $400. Demonstrable results: ? generation and display of real time spectra ? ability to apply spectral windows ? generation and display of filter transfer functions ? generation and display of group delay Discussion: The focus will be on the realtime software for data acquisition and process- ing. Homework 3b 11 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 The basic software design should be modular. Interfaces should be defined between modules that will simplifying upgrading the software as needs change. Forexampleencapsulating the display supportwill allow(hopefully)easily chang- ing from generating a display using a terminal emulator to driving a color LCD panel. The final product could be enhanced by: ? Using higher speed and perhaps higher resolution (i.e., more bits) sigma- deltaconverters. Thesecouldbe mountedonadaughter boardandplugged onto the DSK. This would expand the usable frequency range. ? Include network support so the system can be operated remotely over the internet. ? Add a flat panel display and make the unit completely stand alone. (How to save output?) ? Other ideas??? RLC measurement system Goal product: The primary goal of this project is to apply DSP techniques to create a RLC (resistance, inductance, and capacitance) metering hardware. Competing products: Hewlett-Packard and Stanford Research sell high precision RLC measurement hardware. A number of other manufacturers sell much lower cost hand held units. The small hand hand held unit that I have is made by Almost All Digital Electronics (one man company?) (www.aade.com) is supposed to be about 1% accurate but it only measures L and C. The cost is only about $100 in kit form and about $130 assembled. Over 2000 units have claimed to have been sold. Scope of the EECS 452 effort: ? develop signalprocessingbasedalgorithmsforreducing measurementnoise, ? test the above algorithms using MATLAB, ? implement the algorithms using the TI DSK. Homework 3b 12 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 Demonstrable results: The DSK will be used to measure the component values for a number of R, L, and C parts. Demonstrate the range of values that can be accurately measured. Discussion: There exists an excellent Hewlett-Packard application note on LRC measure- ment. In this application it shows how given a precision resistor measurements of values of inductance, resistance, and capacitance can easily be made. The (dangerously) simple circuit is shown in Figures 1 a & b. Z R Ein Eout+ - (a) Z Ein Ez (b) R Figure 1: Basic LRC meter measurement network. In Figure 1-a the input voltage is a sine wave (or other shape?..how about a square wave and measure at multiple harmonics..any reasons for or against this?) generated using a DAC driven by DSP hardware implementing a direct digital synthesis algorithm. A two channel A/D is used to measure the input and the output voltages magnitudes and phases. The complex impedance is calculated as Z = REinE out and the component value determined knowing the input frequency. The posi- tions of Z and R might be interchanged. One worries a bit about what happens at the summing node and stability may be a problem. In Figure 1-b the measurement point is buffered by a high input impedance unity gain (or maybe a switched gain) amplifier. I have a magazine article de- scribing an instrument designed around this configuration and using a PIC mi- croprocessor. The DSP aspect would be to make highly accurate measurements of magni- tude and phase rejecting random and perhaps incidental structured noise. A direct digital synthesizer would be implemented to produce the test waveform. Homework 3b 13 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 This should be readily doable without much extra hardware using the TI C5510 DSK. The goal would be developing and applying DSP algorithms to mini- mize the required hardware and yet have a reasonably broad measurementrange and good accuracy (say 0.1%). If components are switched in and out of the circuit the quality of the switch- ing device can have a significant effect on the measurement. Reed relays in par- ticular are said to have excessive capacitance when open. Learning about such things is part of designing an instrument. One component that would need to be added the lab EVMs or DSKs would be an LCD display to show the measurement result. This can be done. A range of designs might be investigated. A low end system would likely make only work at a single frequency. A mid system might work using a number of frequencies using a digital synthesizer and perhaps using sampling or radio type demodulators. A high end system would compete with the HP and SRT systems. WWVB Time and Frequency Calibrator Goal product: This project would design a DSP based VLF radio receiver and hardware that uses WWVB radio transmissions at 60 kHz to: ? accurately determine time and to create a reference frequency standard could be used to calibrate oscillators to an accuracy of better than one part in 1011. ? to determine time of day to a high degree of accuracy. This would allow the unit to be used for critical timing applications such as a network time source. Competing products: WWVB has long been used as a frequency reference. A number of manufac- turers have produced products for this application. Among the manufacturers and products are: ? HP (now Agilent), the 117A ? Truetime (we have access to a unit) ? Spectracom, a number of various models. Homework 3b 14 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 Unfortunately I do not know of any products for this application currently in production. GPS is typically used today for this application. There is at least one manufacturer making a WWVBbasednetwork clock. This is Spectracom. Scope of the EECS 452 effort: ? digitized values of a WWVB reception will be acquired and and used to test algorithms using MATLAB. ? we have a A/D converter that will operate upto 1 MHz using the C5510 DSK. This can be used to acquire data values for real time processing. ? the lessons from the MATLABinvestigations will be used to develop code for the C5510 DSK. We have access to: filtered WWVB carrier based reception, a highly stable crystal frequency source, a GPS precision time and frequency source. Demonstrable results: ? generation of a display showing the current time, ? generate a one second waveform which is compared with a similar wave- form obtained using a GPS receiver. ? locking a reference clock to WWVB and a comparison of the frequency to that of a GPS based precision clock. Discussion: In 1956 low-frequency station WWVB (Ft. Collins, Colo.) began broadcasting at 60 kHz. It offers several features: ? more predictable propagation (at 60 kHz) than that of WWV and WWVH, ? digital time code only, ? time and accuracy of 0.1 ms to 1 ms, ? frequency accuracy of 1?10?11 for measurements over 1 day, and ? good reception in most areas of the continental United States.? The time information is placed onto the carrier using amplitude shift keying, essentially at a one bit per second rate. WWVB recently had its antenna system updated and the transmitter power increased to 50 kW. WWVB lacks the time accuracy of a GPS based time clock but does have the advantages of lower cost and the ability to work indoors (though I?m not sure how well in the EECS building). Chips that receive WWVB exist but do not make significant use of digital sig- nal processing. Can a DSP based receiver be designed that will work in environ- ments (like this building) where the existing devices will not. One use of WWVB Homework 3b 15 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 is to serve as a frequency reference. How can DSP be used to use to make better measurements (time and/or frequency) than do existing equipment. Perhaps at lower cost or at higher cost but with higher quality. The competition to WWVB is GPS. There are similar low frequency transmission broadcasts made in Europe and other parts of the world. The goal of the time processing to use DSP to extract the time information and do it better, more reliably, more accurately, etc than existing WWVB clock chips. WWVB based clocks can be purchased from a number of sources includ- ing Radio Shack. This project would do its own design and likely compare the resulting performance (including cost) to that of a commercial unit. One difficulty is that propagation from Ft. Collins to Ann Arbor is significant and needs to be corrected for. There are a number of ways this an be done: ? Simultaneously process the VLF LORAN waveform to locate the position of the receiver and then use a propagation model to estimate the propagation time. ? Use the know locations of the WWVB antenna and the antenna on the EECS roof to estimate the propagation time. ? Cheat. Use a GPS clock and measure the difference between the GPS clock and the WWVB clock and then correct by this amount. ANTENNA AMP FILTER & GAIN SA602A A/D 10 MHz OSC COUNT DOWN 10 MHz 1 MHz 50 KHz 8 KHz PROCESSOR 1 SEC SQUARE WAVE LCD DISPLAY 60 KHz CARRIER 10 KHz IF I/Q DEMODULATION USING SUBSAMPLING REF ? KHZ Figure 2: One possible WWVB receiver block diagram. Figure 2 shows the block diagram of one possible implementation. This cir- cuit was something I was experimenting with. This circuit hetrodynes the 60 kHz carrier to 10 kHz in order to avoid problems with self oscillation and to reduce the bandwidth requirements on the A/D converter input. Specific frequencies were chosen because of ease of generation. Homework 3b 16 January 4, 2008 EECS 452 Digital Signal Processing Design Laboratory Winter 2008 I?ve only built part of this circuit at this point. The existing portion consists of a 60 kHz ferrite rod antenna, a pre-amp, a SA602A demodulator, the 10 MHz oscillator, and the clock count down circuit. The IF filters need to be designed. Subsampling is planned to be used to digitize the IF waveform. With a bit higher quality thinking maybe the design could be simplified. I decided to start with something that I figured would work and then see where the next generation would go. This circuit is a natural for the Analog Devices ADUC812 with it?s built in 12-bit A/D converter. It should also go well with the EVMs and DSKs available in the lab. LCD support would have to be added in any case. The DSP connection is developing algorithms and associated code to, in some sense, optimally extract the time code and synchronize the locally one second generated output to WWVB as accurately as possible. There is a powered 60 kHz antenna on the roof which could be run into the EECS 452 lab for use. Homework 3b 17 January 4, 2008 Kurt Metzger EECS 452 DSP Design Laboratory DSP, CCS, DSK, TMS320VC5510, C5510 Homework 3b - Winter 2008
Want to see the other 17 page(s) in Homework 3b?JOIN TODAY FOR FREE!