DANGER DETECTING HEADPHONES

Transcription

1 DANGER DETECTING HEADPHONES By Tae Hun Ahn Daniel Bang Yoon Mo Yang Final Report for ECE 445, Senior Design, Fall 2016 TA: Zipeng Wang 07 December 2016 Project No. 47

2 Abstract This report describes the details of the design in the Danger Detecting Headphones, which detects the car horn sound from the outside environment and mutes the speaker for three seconds. The end product was also able to filter out noises from the car horn signal. Hardware and software design implementation will be discussed along with the verification of the modules, followed by the cost analysis and future works. ii

3 Contents 1. Introduction Statement of Purpose Objectives Benefits and Features Goals and Functions Design Block Diagrams and Flowchart Hardware Block Diagram Software Flowchart Block Description Power Module Microphone Module Volume Control Module Microcontroller Module Design Verification Power Module Microphone Module Microcontroller Module (ATMega328p-pu) Volume Control Module Costs Parts Labor Grand Total Conclusion Accomplishments Uncertainties Ethical considerations Future work References iii

4 1. Introduction 1.1 Statement of Purpose Nowadays, there are many consumer electronics companies focusing mainly on the audio equipment. Many of these companies produce noise canceling audio systems which enhance general sound quality of music in all kinds of situation. Such products become main production lines of several renowned audio companies, and this reflects a high demand of consumers in the audio market. We, however, found a very hazardous feature of this noise canceling system. And this precarious feature appears in not only the noise canceling headphones but also most of regular headphones in use throughout the daily life. The problem is that the noise canceling headphones and regular headphones with high volume lead users to isolate themselves from surroundings by blocking all sounds but music. For the pedestrians who use headphones regularly on the street, this blocked sound could lead them directly to death. To prevent hazardous situations for the pedestrians, we decided to develop headphones that detect car horns and automatically mute music. This mute system will help pedestrians to perceive surround situation immediately. We successfully developed headphones that work properly as we expected. Even though overall functionality was satisfactory, there were some uncertainties like reliability in maximum detectable distance. 1.2 Objectives Benefits and Features Portable Detect car horns Mute the music player once the car horn is detected Filter out noises other than the car horns Goals and Functions Collect sound data from microphone Convert the analog signal to the digital waveform Fast Fourier transform the digital wave Provide safety for the users via the mute logic 1

5 2 Design 2.1 Block Diagrams and Flowchart Hardware Block Diagram Figure 2.1 depicts the hardware block diagram of the entire circuit. First, the power module regulates voltage from the battery and supplies 5V to all the other modules. The microphone module receives sound signal from the outside environment and amplifies it through an op-amp circuit. The amplified signal is still an analog signal and needs to be converted to a digital waveform in order for us to run the detection algorithm in software. The microcontroller module does the job through an A/D converter inside the chip. It also applies Fast Fourier Transform combined with the Digital band-pass filter, so that it filters out the unnecessary noises other than car horn signal. If the filtered signal s frequency matches with that of the car horn, microcontroller will send a mute logic signal to the analog MUX inside the volume control module. The volume control module, which controls the volume in normal state, will mute the speaker for three seconds once it receives the mute logic signal. Figure 2.1 Hardware Block Diagram Software Flowchart Figure 2.2 shows a software flowchart. The input of this flow chart is an analog signal generated by the microphone. This analog signal will be converted into the digital signal through ADC by the ATMega328P chip. After the conversion, we will run the fast Fourier transform test. Since the converted digital signal 2

6 contains a lot of noises, we pass the FFT results to the band pass filter to get rid of some noises that are below the threshold. Figure 2.2 Software Flowchart With the results from the filter, we perform sampling to get the frequency. If the frequency matches the expected car horn frequency, then the music will be muted. If not, the entire process that we described above will be performed infinite times until it detects the proper frequency. 2.2 Block Description Power Module The power module of Danger Detecting Headphones has a simple structure but its role is significantly important. To obtain an accurate detection feature, it was required to make the power voltage signal as clean as possible. If the power signal has some noise, it will affect the output of the microphone module and it will disturb the entire circuit to detect the car horn sound precisely. Therefore, it was fatal when we did not filter out the noise from the power signal. Figure 2.3 shows the PCB that contains the power 3

7 module and the microphone module. And the part inside the orange square represents the power module. For the power source, we used a 9 V alkaline battery and for the voltage regulator, we used a LM7805. The role of the LM7805 is basically to convert 9 V from the battery to 5 V. The reason for this choice is that every IC chip being used in the entire circuit needs 5 V as its supply voltage. As you can see from Figure 2.4, two capacitors were used in the module: The first capacitor whose capacitance is 100 μf, is hooked up after the battery and before the input of the LM7805 regulator. This capacitor is there to filter out any noise coming from the voltage source by shorting the AC signal of the battery to ground and let only DC portion go into the regulator.[1] The second capacitor, hooked up after the regulator also filters out any noise or ac signals by the same mechanism. We hooked up the resistor and the LED at the end of the circuit to make it easier to check if the power module works fine. Figure 2.3 PCB (Power Module) Figure 2.4 Schematic of the Power Module Microphone Module The microphone module simply receives any audio signals from the outside of Danger Detecting Headphones. Please refer to Figure 2.5 for the PCB design of the microphone module which is inside the orange square. Since the voltage output of the microphone is relatively small, it s not possible to use it for the detection feature without amplifying the output of the microphone. Therefore, it s required to use an audio amplifier. We found a low voltage audio power amplifier, LM386. We designed the module as shown in Figure 2.6. We came up with this design by referring to the datasheet of the LM386. This module basically receives the input audio signal from outside with the microphone and amplifies the signal with the LM386. As you can see from Figure 2.6, lots of capacitors were used in this module. Each of them has its own role. For example, C3, and C7 in Figure 2.6, block the DC voltage so only the AC signal can be passed. We need to block the DC signal since audio signals are basically AC signals. Capacitors block the DC voltage because the reactance of a capacitor is 4

8 X c = 1 wc (2.1) where w is angular frequency and C is capacitance. Since DC voltage has no frequency, the reactance will be infinity for DC cases so the capacitor blocks DC voltage. C2 whose capacitance is 10μF sets the voltage gain to 200 V. C1 has a similar role as the capacitors in the power module since it acts as a V bypass filter. Lastly, C4 acts as a current bank for the output. When sudden surges of current occur, it will drain. Moreover, when there is the low demand for current, it refills with electrons. Figure 2.5 PCB (Microphone Module) Figure 2.6 Schematic of the Microphone Module [2] Volume Control Module The volume control module lets the user control the volume of his or her music player. The special feature of this module is that its analog multiplexer receives the logic bit from the microcontroller as its select bit. And according to the select bit, the analog multiplexer chooses its output between two inputs. Figure 2.7 shows the design of the module on the breadboard. As you can see from Figure 2.8, there are three major components in this module. The first component is a stereo audio jack (STX-3120), it takes a 3.5 mm stereo audio cable as an input. We chose this audio jack since every music player uses the 3.5 mm stereo audio cable. STX-3120 also controls the volume by receiving the electric signal from the music player of the user. The volume control module also uses the LM386 amplifiers since the output from the audio jack is too small so, without the amplifiers, the analog multiplexer will not output the signal. We set the voltage gain to 20 V of the amplifiers. V 5

9 The last component is the analog multiplexer, SN74F257. It receives the logic bit from the microcontroller. Based on the logic bit, it chooses one of the inputs and sends the output signal to the speaker. Please refer to Figure 2.9 and 2.10 to understand the operation of the analog multiplexer. Figure 2.7 Volume Control Module Figure 2.8 Schematic of Volume Control Module [3] Figure 2.9 Logic Symbol of SN74F257 [4] Figure 2.10 Function Table of SN74F257 6

10 2.2.4 Microcontroller Module The main component of the microcontroller module is ATMega328P-PU, which is also called an Arduino UNO. Through the software design inside the chip, it performs the key functions to detect the car horn frequency. Since we did not need to use all the I/O pins of the ATMega chip, we have built our own circuitry for the ATMega328P-PU instead of a complete Arduino Uno board. This way, we were able to not only save a lot of money, but also reduce the physical size of the entire PCB board as below. We chose our clock to be 16 MHz instead of 8 MHz, because faster clock signal speeds up the execution time of the microcontroller, letting the Arduino sample the analog signal more times in a same amount of time. We expected this to result in more sampling rate, increasing the accuracy of the detection. Since our design did not require the continuous communication with PC, we did not utilize WiFi communication. Instead, we decided to use FT232 USB component to communicate with PC. Thus, we connected the USB port just to upload the new sketch. The connection was simple enough because it only required Rx and Tx pin of the ATMega chip to be connected to the Tx and Rx pin of the FT232, respectively. However, we met a huge challenge during the communication. Initially, our circuit did not have a capacitor C1 in the Figure 2.11, when the reset pin of the ATMega chip was connected to the DTR pin of FT232 USB. This resulted in AC voltage, which is a noise, going into the DTR pin, arbitrarily resetting the microcontroller while uploading. Thus, we had to cut out the AC voltage through the capacitor C1. Figure 2.11 Microcontroller PCB Figure 2.12 Microcontroller Schematics 7

11 3. Design Verification 3.1 Power Module In the power module, we have used 9 V alkaline battery. The battery should supply 9 ± 1.35 V to the voltage regulator. This can be easily verified by connecting voltmeter to the battery. This voltage passes the voltage regulator LM7805, and output should be 5.0 ± 0.25 V. As you can see in the Figure 3.1, max voltage value is 4.98 V and min voltage value is 4.80 V. These both values are within the range of 5.0 ± 0.25 V. Figure V Voltage Regulator Verification 3.2 Microphone Module In order to verify the microphone module, we focused on outputting the result through the oscilloscope. This way, we were able to easily check if the output voltage matches with our expected voltage stated in the requirements. The first important requirement was if the voltage output from the microphone itself was from -2.5 V to 2.5 V. Using the oscilloscope, we were able to verify the output voltage was within the expected range. The second important requirement was the output voltage of the op-amp circuit. Since the signal was amplified and given offset through the op-amp, the output should have been from 0 V to 5 V. We successfully verified that the output voltage matches with the expected range. 3.3 Microcontroller Module (ATMega328p-pu) To verify the performance of ATMega328p-pu, we first supplied the ATMega328p-pu with the 5 V. When we connected the voltmeter with the pin 7, the result value was which is in the range of 5.0 ± V. ATMega328p-pu can also receive the analog signal between 0 V and 5 V. As you can see in the Figure 3.2 the voltage is within 0 V and 5 V. 8

12 Figure 3.2 Analog Input from Microphone [5] ATMega328p-pu performs FFT and digital filter, and these also needed to be verified. FFT was verified by comparing the results of the different coding (MATLAB and Arduino). As you can see Figure 3.3 shows the MATLAB code and Figure 3.4 is the result of FFT. Figure 3.5 is the result after running band pass filter. Figure 3.3 MATLAB Simulation Code Figure 3.4 Before Band Pass Filter Figure 3.5 After Band Pass Filter [6] 9

13 3.4 Volume Control Module The operation of volume control module was verified by connecting the voltmeter to the digital power supply pin and GND. The measured voltage is in the range of 5 V (±0.5%). The volume control module also need to mute the music when the Arduino detects the car horns. This was verified by checking whether the voltage of pin 8 output 0 V. 10

14 4. Costs Price is always one of the most important factors in the market since it might be the first or last thing that consumers consider before they buy the product. Table 4.1 outlines the cost of each component while Table 4.2 outlines the cost of labor. As you can see from the last row of Actual Cost column of Table 4.1, the total price is not that expensive compared to any normal headphones in the current market. 4.1 Parts Table 4.1 Parts Costs Parts Manufacturer Retail Cost ($) Bulk Purchase ($) Quantity Actual Cost ($) 9 V alkaline battery Digikey Voltage Regulator (LM 7805CT) Microphone (CMA-4544PF-W) Audio amplifier (LM386N-1/NOPB) USB Interface Module (DEV ) Microcontroller (ATMega328p-pu) Analog Multiplexer (SN74F257N) Digikey Digikey Digikey Digikey N/A Adafruit Digikey Speaker Digikey (CLS0261MAE-L152) Audio Cable Digikey (AK203/MM) Audio Jack Mouser (STX B) 16MHz Crystal Digikey (ECS XDN) PCB PCBWay Resistors, Capacitors Digikey Total *Note: The bulk-purchase costs are based on 1000 quantities for each component and the retail cost is based on 1 quantity for each component. 11

15 4.2 Labor Table 4.2 Labor Costs Name Hourly Rate Hours Invested (12 weeks * 13hrs/week) Labor Total (2.5 * Hourly Rate * Total Hours) Tae Hun Ahn $40/hr 156 $15,600 Daniel Bang $40/hr 156 $15,600 Yoon Mo Yang $40/hr 156 $15,600 Total 468 $46,800 As shown in Table 4.2, the total labor cost is $46,800. The overall cost of this project, therefore, is $ $46,800 = $46, It s important to compare the actual cost with the bulk-purchase cost in Table 4.1. The bulk purchase will reduce the cost by 72%. Therefore, if we mass-produce our headphones, the amount of money that costs to manufacture the PCB will significantly decrease. 4.3 Grand Total Table 4.3 Grand Total Type Total ($) Parts Costs Labor Costs 46,800 Grand Total 46,

16 5. Conclusion 5.1 Accomplishments At the end, we successfully built an entire circuit that could verify all the functionalities except for the changed subcomponent, volume control module, from the design review. The end product showed a feasible precision on the Fast Fourier Transform algorithm, compared to the MATLAB simulation. It also filtered out the noises other than 300 ~ 400 Hz so that the accuracy of the detection was reliable. 5.2 Uncertainties We are uncertain of the detail in range of the detection. Since the car horn sound in a real life emits much more decibel than the car horn sound recorded in a smartphone, it was rather difficult to test the range of functionality without an accurate decibel meter. Since we were able to confirm that the microphone module receives a signal of a normal conversation in two meters, we assumed our headphones to be able to detect car horn sound, which is much louder, in five meters. In order to prove the reliability of our headphones, we have to verify this requirement in a real life before bringing the product to the market. 5.3 Ethical considerations During the entire project, we followed the IEEE Code of Ethics. The following is from the IEEE Policies, Section 7 - Professional Activities (Part A - IEEE Policies). We also followed the detailed descriptions as below. 1. To accept responsibility in making decisions consistent with the safety, health, and welfare of the public, and to disclose promptly factors that might endanger the public or the environment. Since our project was all about detecting danger around us, we always kept in mind about safety issue while building the final product. We paid extreme care while we were soldering, and cleaned up the lab after we were done. While we took care of the power module, we always checked if any power supply was shorted, which might have burnt our entire circuit. Overall, we had made great decisions regarding the safety issues. 3. To be honest and realistic in stating claims or estimates based on available data. Most of the requirements on our project were verified through the multimeter, oscilloscope, Arduino software (IDE), and MATLAB. Since these programs and machines were only used to show the measurement of our result, there was no opportunity for us to change our data to seem more plausible. By the end, we had been truly honest and realistic in stating claims. 9. To avoid injuring others, their property, reputation, or employment by false or malicious actions. Throughout the semester, our major concern was the safety. When we made a mistake burning one of our laptops by sourcing the current in an opposite direction, we quickly pulled out the battery wire in case of injuring other people. After this incident, we always paid close attention on our every action. Also, we saw some number of people in other groups got their components stolen while they were out. 13

17 Since we knew how painstaking it is to order new components via online, we vowed not to cause any harmful action to hurt others in their project. Eventually, this ethical consideration was well kept in our case. 5.4 Future work There are four possible future works. First, we can improve our FFT running time by trying different types of FFT libraries. There are several FFT open libraries for the Arduino, but we could not test every single FFT libraries and check the runtime. Therefore, in the future, we would like to pick the best library and algorithm to optimize overall performance Second, we can improve our sound quality of headphones by using different MUX. We have researched a lot about the mux and found out that MUX requiring negative voltage can generate much better sound quality. We, however, could not generate negative voltage in our system so we decided to use an analog mux, SN74F257, which only required positive voltage. Third, we can apply our system to more broad industry. For example, we can apply our danger detecting mechanism on the car audio system. If the car detects car horns, the car can automatically mute the car audio system so that the driver can see what is actually going on around his or her car. 14

18 References [1] How to Connect a Voltage Regulator in a Circuit [Online]. Available: [2] How to Build a Microphone Amplifier Circuit [Online]. Available: [3] LM386 Low Voltage Audio Power Amplifier, National Semiconductor Corporation., Dallas, TX, [Online]. Available: [4] SN54F257, SN74F257, Texas Instruments Incorporated., Dallas, TX, [Online]. Available: [5] Arduino. ReadAnalogVoltage. Arduino. 21 Sept [Online]. Available: [6] Aasvik, Mads. Arduino Tutorial: Simple high-pass, band-pass and band-stop filtering. Norwegian Creations. 10 Mar [Online]. Available: 15