ME 224 Final Project AUDIO CONTROL DECEMBER 2001 ME 224 PROFESSOR ESPINOSA NORTHWESTERN UNIVERSITY RACHEL FINE GARETH HAYES JOSHUA HIGGINS JEHANA RAY TABLE OF CONTENTS TABLE OF CONTENTS .................................................................................................... 1 INTRODUCTION .............................................................................................................. 2 BACKGROUND INFORMATION.......................................................................... 2 EXPERIMENTAL PROCEDURE ........................................................................... 7 TESTING RESULTS AND ANALYSIS ......................................................................... 12 SUMMARY ...................................................................................................................... 16 BIOGRAPHIES ................................................................................................................ 17 REFERENCES ................................................................................................................. 20 APPENDIX A: BILL OF MATERIALS .......................................................................... 21 1 INTRODUCTION This project hopes to explore the mechanisms used to manipulate sound waves. Specifically we will show how a limiting circuit can be used to create a sort of “Smart Sound” system commonly found in televisions. This technology can be achieved through fairly basic circuitry. The results can be very useful in amplifying quieter, or lower amplitude, outputs and minimizing louder, or higher amplitude, outputs. Both types of outputs are corrected by the circuitry to fit within a set voltage range set by the user. In this lab we will use LabView to control the voltage range and also to gather the necessary data to show how our circuit is working. We will show that regardless of how high or low the input frequency is, the output voltage will not go beyond a certain range specified by the user. We will also test the upper and lower ranges of the limiter to see if performance is affected at these extremes. We will show that a limiter can be utilized to control output and test the effectiveness of our limiter in this capacity. BACKGROUND INFORMATION The “Smart Sound” system we have based our project upon is a fairly common and widely used application, especially in televisions. The desire to limit the volume of loud television shows and commercials resulted in the creation of this technology. It is made possible by the fairly basic premise of limiting circuitry. This is a simplified explanation of the “Smart Sound” system but is valid. Limiting circuit: Limiting circuits can be very simple, as shown in Figure 1, or fairly complex, as shown in Figure 2, depending on the needs called for by the 2 user. Due to the needs for audio limiting we required we used the circuit shown in Figure 2. Both circuits do show the basic properties of a limiting circuit, however. Figure 1 The circuit above limits the output voltage through the use of a pair of diodes, placed in parallel and in reverse polarity, and a resister in series with these. The nature of the diodes results in the output voltage never getting higher than ±0.7V. In many cases this sort of circuit is used to protect amplification circuits. In the case of our lab, we are using the more complex circuit of Figure 2 to actually limit the output voltage based on an input from a microphone circuit. The result of the more complex circuit is similar to that of the simple circuit shown in Figure 1 in that both limit the output voltage over a certain range. A major difference is that while the output voltage of the first circuit is set by the diodes and cannot be changed, the range allowed by the second circuit can be manipulated. This allows the limiting circuit we have utilized to incorporate a LabView initiated volume knob to control the amplitude of the output voltage 3 Figure 2 The circuit uses an FET to control the voltage and two TL072 op-amps. We replaced these op-amps with a single TL074 op-amp as it allows for virtually the same results. R14 and C5 determine the attack time, about 5ms and R12 and C5 determine the recovery time, approximately 1 second. R11, C3 C4 and R13 form the distortion cancelling circuit that is necessary to avoid compromise of the circuit. Ideally the system should be used above 500 Hz, as the distortion rises with decreasing frequency. This circuit does limit distortion very well within the threshold voltage. The threshold voltage is the variable limit to which the voltage is to be limited to. All of these values work well with audio signals but we were at liberty to change them to see how each helped to control the system. Some approximate specifications of the limiting circuit we used are shown in Figure 3 4 Limiter Specifications Maximum Attenuation Noise Level (unweighted) Typical Max. Output Level Gain FET Voltage (at max. o/p) Distortion 40dB -80dB (ref. 1.65V RMS output) 1.65V RMS 6.8 (16dB) < 45mV typical) < 0.5% typical Figure 3 This circuit that we utilized is a comparably cheap model of what is used in more complex systems. It allows us to show the basic results found in the more expensive and elaborate systems in an easy to visualize and manipulate circuit. We decided to focus on the limiting circuit as opposed to other methods for electronic control we investigated because we felt it most closely resembled the results that are gained in an actual commercial level “Smart Sound” systems. While other methods, such as a band-pass filtering systems, clipping circuits, compression devices, and simple RC or RLC filters, would have also given us insight into the workings of electronic control systems, we felt the limting circuit would give us more to work with than these other alternatives. Microphone: We also built a condenser microphone onto our bread board. Two views of the microphone, both the actual microphone component and the correlating circuit that we used to utilize it are shown in Figure 4 & 5. The condenser microphone works by the movement of a diaphragm in relation to a rigid backplate. This movement creates a change in an electric field between the two. This, in turn affects the current that goes to an amplifier and thus, the output is amplified or attentuated. We are using the microphone to pick up an input 5 signal, anything from a human speaking voice to a tap onto the actual microphone and then to translate it to our limiting circuit. Figure 4 Figure 5 This particular microphone creates a 2.2 KΩ impedence. This allows for a lowering of the necessary voltage bias. The fairly simple microphone circuit also utilizes an FET. 6 EXPERIMENTAL PROCEDURE The experimental apparatus for this experiment is quite simple. The main step in setting up this experiment is building the circuits, shown in Figures 2 & 4, on a breadboard. Once the circuit has been built, it needs to be connected to the DAQ so that the computer can control the experiment and compile the results in a more convenient manner. An oscilloscope need also be attached to the outputs so that the results can be viewed in real time. The final piece of equipment needed is a speaker that will produce sound into the microphone. The easiest way to provide a controlled sound is to output audio files through the computer’s speakers, which must be placed in close proximity to the microphone to eliminate outside noise. Building the Circuit: The microphone circuit (Figure 4) as well as the limiter (Figure 2) should be built on the breadboard and the outputs connected to the oscilloscope. The breadboard with the complete circuit can be seen in Figure 6 below. 7 Figure 6 LabVIEW Construction: There are two main purposes of the LabVIEW program: 1. Control the system by producing exact, user-defined voltage inputs to the circuit 2. Compile the resultant data in a clear, convenient fashion. To execute purpose (1), the program must utilize the DAC stg function to output a value through the DAQ and into the circuit. Then a knob, designated as a control, should be placed in the Front Panel of LabVIEW to allow the user to select a desired voltage by “turning” the knob. To execute purpose (2), the program must first go through the above step. Thus, a sequence function with two frames should be used to ensure that the voltage input occurs before the output is generated. In the second step of the sequence, an ADC stg function is employed to retrieve data from the DAQ’s input port. This data is then outputted to a Waveform Graph for real-time viewing of the results, as well as to a text file for later manipulation. An example program can be seen in Figure 7 on the next page. 8 9 a) b) c) Figure 7 10 Connecting the DAQ: The DAQ should be connected such that its input voltage comes from Channel 2 (if using the LabVIEW function in Figure 7 on the output side of the DAQ, and the resultant output comes from Channel 0 on the input side of the DAQ. The input and output sides are shown in Figure 8 below. Ground Output from circuit Ground Input voltage Figure 8 Running the Experiment: Once the experiment is set up, performing the actual experiment is quite simple. The LabVIEW program is run continuously, and the user controls the voltage input using the knob on the Front Panel. The resultant output can be seen in real time on the oscilloscope as well as the Waveform chart. The results can also be viewed in spreadsheet format and the data can be manipulated as chosen. The important factor to observe is the amplitude. We will manipulate the system to not only see how the limiter works with ideal inputs, but also with very high and very low values of input. 11 TESTING RESULTS AND ANALYSIS As described in the previous sections, the purpose of this device is to actively limit an audio signal within a specified amplitude based on a set input voltage. Software interfaces through the LabVIEW program were used to provide an interface with the electronics in an effort to illustrate how in-software signal control can be accomplished. This general objective has been accomplished through the course of the experiment, and the circuit - as was originally intended limits signal output actively based on the manipulation of a graphical volume knob. Throughout the experiment, we ran into a number of hurdles due to the fact that we had substituted a few unavailable components with similar model parts. The main concern occurred during initial conditioning of the circuit prior to LabVIEW implementation. The fact was that the part substitution had produced a considerably larger amplification of the initial signal than expected. This and other less significant problems were accounted for, and both the microphone conditioning circuit and the audio limiter were found to be in operating condition. The only process matter that remained was to facilitate an active control of the circuit. In initial testing, it was revealed that a control voltage range of approximately zero to 2V at a current on the range of 1.5A was enough to make the circuit transition from its maximum amplification to a behavior where the circuit was overloaded. The overload was such that the signal reached a point of 12 producing near constant voltage, which in effect masked the microphone output signal. Eventually the circuit voltage ranges were found and applied to the programming, which allowed the group to produce constantly controlled amplitude for our given signals. Figure 9 illustrates the effect of our circuit on the waveform information. Figure 9 The top signal [ch1] is the primary amplified signal output from the microphone conditioning circuit, while the lower signal [ch2] represents the limited output. Both signals are on the same voltage/div scale on the oscilloscope window, to eliminate any confusion about scaling of the signals. By setting the scales as such, the direct amplitude relation between the two is representative of the amount of signal limiting, and is therefore of primary importance in this diagram. The control voltage at this point is barely above 1V, and its result can be seen as the relative compression of [ch2] as compared to the initial signal. It should be noted that the limited signal has fully maintained its original frequency data in the path of signal processing. This information is not 13 significantly affected by the limiter circuit in any way, as can be clearly demonstrated by a comparison of the two signals, principally when one looks at the corresponding peaks of each signal. Peaks of the two waves are coincident through the oscilloscope view, which was given a slight DC offset to separate the signals enough to represent them without interference of the two graphical representations. Given this association, it is possible to set the oscilloscope’s math function to give a direct visual representation of the amount of limiting applied to the signal as opposed to the direct comparison method just discussed. The concept is as follows: When the second signal has been unprocessed, it is an exact copy of the original signal, and by applying the function [ch1]-[ch2] a related function is graphed. When the limiter circuit is not active, the function produces a graph of near zero value – essentially a flat line aside from a given amount of background noise. As the limiter is turned to an active state, the graph begins to increase its signal and reaches a maximum output as the limiter approaches its maximum capability. In essence this resultant signal appears as a modified version of the original microphone conditioner output, albeit with a slight offset. 14 CONCLUSIONS Given a second look at the conditioning process, it would be possible to improve upon one major aspect in our limiter software interface concerning the relationship between voltage and audible volume. This is due to the fact that audio signals are not linearly related to input voltage, but instead are associated by means of a logarithmic relation. LabVIEW has the ability to create outputs based on mathematic relations, which can ameliorate this discrepancy. Given enough time to condition the circuit in this regard, the volume control could be improved to provide more practical control over audible volume. In this case, the visual method used in the experiment is characterized by a linear relationship between voltages and therefore must also undergo the same transformation to be a completely accurate representation of the audible signal. Be that as it may, the current setup has certainly served its purpose throughout this experiment. 15 SUMMARY A microphone is a type of transducer that takes a mechanical input and converts it to give an electrical output. To eliminate surrounding interference, a limiter is needed. A limiter attenuates a high amplitude output by limiting the current that goes into the amplifier set by a particular voltage. In the case of our project, the voltage is manipulated and stabilized using a designed program in LabView. Physically speaking, the microphone and limiter work together to clarify the direct input without creating a drastic damper on the amplitude, i.e. the loudness. This is a basic version of commercially available “Smart Sound” systems that are commonly found in televisions. We were able to implement a version of this and test it. Our circuits gave us insight into the workings of the limiting circuit and the microphone. LabVIEW was able to act as both a data collector and a threshold controller for the limiter. The threshold is basically the voltage set by the user that the output voltage is maintained within. Ideally the limiter does not affect the shape of the input waveform, just the amplitude. This would result in volume control without sacrificing sound quality through distortion. Our results showed that the circuit we used does a good job of completing this desired task. 16 BIOGRAPHIES Fine, Rachel Presently a fourth-year mechanical engineering student at Northwestern University, Ms. Fine plans to graduate June 2002. Past experience includes satellite design for the Space Dynamics Laboratory (Logan, UT), energy management for Goodyear Tire and Rubber Company (Akron, OH), and computer consulting for Academic Technologies (Evanston, IL). She plans to further her education by obtaining a Master of Science degree in Aerospace Engineering. Potential career vocations include astronautics and systems/spacecraft engineering. Hayes, Gareth Gareth is currently a junior mechanical engineering student in the McCormick School of Engineering and Applied Science at Northwestern University. He graduated from Watertown High School in Watertown, NY in 1999. He is currently involved in Northwestern’s co-op program as a mechanical design engineer for Cummins-Allison Corp. located in Mount Prospect, IL. He plans on specializing in design at Northwestern and graduating in June of 2004. After college, Gareth plans on obtaining employment in the field of mechanical engineering. 17 Higgins, Joshua Joshua Higgins is a Junior Mechanical Engineering major in McCormick. His focus is on design engineering. Joshua is from Rockdale, IL and graduated from Joliet Central High School in 1999. Past projects at Northwestern University include website development for Norris University Center through the first-year Engineering Design and Communication (EDC) course, much of which was subsequently implemented into their actual site improvement the following year. He has also worked with United Parcel Service through the EDC program. Involvement in the latter project was centered at their Midwest ground-based package processing facility, and was focused on an effort to increase working life of pulley machinery in their conveyor belt systems. Joshua involves himself in a number of design projects on his own time; currently he is in the process of designing and building an alternate scale-length electric guitar. This coming year he will compete in Design Competition for the first time. Ray, Jehana Jehana Ray was born in Silver Spring, Maryland and raised in DeSoto, Texas. Today, Jehana is a junior in the McCormick School of Engineering at Northwestern University in Evanston, Illinois. Her focus is in mechanical engineering with a specialization in design and a theme in economics and art. Post graduation, Jehana wants to work for a well-known medical company in their diagnostic division. Her goals are to design surgical instruments for animals and human beings that will facilitate surgery, and possibly design prosthetics, as well, 18 with hopes of bettering mankind. After two to four years working as an engineer, Jehana would like to enroll in a graduate school to become a veterinarian, her dream career goal. As a veterinarian, she hopes to be able to use her skills as an engineer to improve veterinary medicine and veterinary instrumentation, in order to keep all animals safer, healthier, and happier. 19 REFERENCES [1] Allison, Phil. “Fast Audio Peak Limiter.” 2000. http://sound.westhost.com/project67.htm [2] Alspector, Professor J. “Limiting Circuits.” http://eceweb.uccs.edu/alspector/Limiting%20Circuits%20web/limiting%20ci rcuits.htm [3] Elliot, Rod. “Musical Instrument (Expandable) Graphic Equaliser.” http://sound.westhost.com/project64.htm [4] "LabVIEW Student Edition 6i", Robert H. Bishop, Prentice Hall, Inc., 2001. [5] Plonus, Martin. Electronics and Communications for Scientists and Engineers. Harcourt/Academic Press, Burlington, Mass; 2001. 20 APPENDIX: BILL OF MATERIALS Hardware Proto-Board 203A (“Bread Board”) TDS 200 Series Digital Oscilloscope Servo to Go DAQ Personal computer Computer Speakers Circuitry as in Table I Software National Instruments LabVIEW Microsoft Excel STG driver Table I : Elements of the circuit Description Type Quantity Microphone 9.7MM Diameter/18MM Leads 1 2.2 Kohm 10K 2K2 39K 1M 4K7 39M 3K 220K 100K 2K7 1.0 micro F 10 micro F 250 micro F .05 micro F 33 micro F 100 micro F 47 micro F 10 pico F 15 nino F 33 micro F LM386, 2xTL072 1 2 1 1 1 3 1 1 1 2 1 3 2 1 1 1 1 1 1 1 1 3 10 K ohms 1 2 Resistors Capacitors OP AMP Potentiometer Diode 21