1. The Self Tuning Guitar Kit
Group 5
ECE 445: Senior Design
Tom DaMario & KyungJin Lee
December 4, 2008

2. Introduction
Kit provides fast, easy way to tune almost any guitar automatically
Combines TI DSP, a simple microcontroller, and a stepper motor
Gives easy user interface to make anyone could tune the guitar

3. Features Universal Design 3 different tuning configurations
Standard, Drop-D, Open-G
Simple operation
Accurate to +/ Hz

4. Objectives
Portability: 100% Battery Powered Device and should be able to fit in guitar case
Universal: Device should fit on any guitar
Functionality: Device should have multiple tuning modes

5. Tuning Configurations
String Number
Standard Tuning
Drop-D Tuning
Open-G Tuning
Note
Frequency [Hz] 6 E
82.4 *D
73.4 5 A
110 *G 98 4 D
146.8 3 G
196 2 B
246.9 1 e
329.6
293.6
* Strings that change from standard tuning.
Source:

6. Hardware Overview DSP: C6713 DSK PIC 16F877A
Takes analog signal from guitar and runs ADC
Runs FFT on digital signal
Takes difference between desired signal and received signal
PIC 16F877A
Takes string number from DSK and outputs to Hex display
Receives two signals from DSK to turn motor
Motor: Parallax #
Turns guitar tuning peg until desired frequency is reached

7. Software Overview Code Composer Studio v3.3
Used to connect to DSK
Program/Run DSK
Flashburn
PIC C compiler and MPLAB IDE
Used to convert .c program to assembly to load program on PIC 16F877A

8. Project Flowchart

9. Output from Guitar Without Amplification: Noise Level:
Ave noise level = 22.8mVp-p
Ave signal level = 150.3mVp-p
Noise % = 1- { ( ) / } = 15.16%
Noise Level:
String 3 without Amplification:

10. Signal After Amplification
Amplifier Circuit:
After Amplification:

11. Original plan: Use C5509A Has built-in A/D converter
Design to draw low power (Easy to make Portable)
Problems:
Emulator could not detect target DSP chip
Back current between CVdd and DVdd

12. C6713 DSK Board

13. C6713 Board DIP switches for user input Built-in LED for tuning config
Analyze the audio input from guitar to find different frequency in real time
Send the motor control output to PIC

14. Flowchart of DSP Software
Holding DIP Switch 3 Analyzes the Signal in Real Time
To ignore any input when guitar is not plucked, we add this part in the code.
Simplify the output to PIC to control the motor
Only two bits of GPIO are required

15. Only Run FFT When String is Plucked
Without any audio input
With guitar audio input
String 6
Max ≈ 4
String 6
Max ≈ 90

16. Calculating the Frequency
Lowest sampling frequency of A/D converter is 8kHz
# of sample = 2048
Step size in frequency = 8000/2048 ≈ 3.906Hz
Frequency = (step)(3.906Hz)
1st Peak: 84
2nd Peak: 168
3rd peak: 252
Mean difference ≈ 84
Freq = (84)(3.906) = 328.1Hz
Frequency difference = 1.496Hz

17. Testing the DSP
To test the user input from DIP switch, use LED on the board and check desired frequency from the memory
To test the FFT, compare graphs from CCS to make sure the peak frequency of FFT is same as the frequency of input signal

18. DSK Challenges
Find the common ground – to solve this problem we use TP32 as common ground from DSP board
Find I/O from DSP – Use HPI to output as GPIO
Cannot find the way to send any input to DSP – Use DIP switch as the user input
Get the right speed for motor to turn, and prevent any delay

19. Increasing the Accuracy
Increase the number of samples taken by the DSP
Would make the running time longer
Requires more memory space
Formula that filters out unclear guitar inputs
When the first peak of FFT is smaller than other peaks of FFT

20. Daughter Card

21. Wire Wrapping Technique
Disadvantages:
Wires not soldered in place
Time requirement
Messy

22. Voltage Regulator for Battery Power

23. Motor Specifications Parallax Continuous Rotation Servo #900-00008
Stepper Motor
Powered with 5VDC
Torque = 3.40 kg-cm (47 oz-in)
Need about 37 oz-in to turn peg

24. Motor Power Consumption
String Number
Average Current [A]
Max Current [A]
Average Power [W]
Max Power [W]
Clockwise
Counter Clockwise 6
0.35
0.17
0.45
0.18
1.75
0.85
2.25
0.9 5
0.33
0.36
1.65
1.8 4
0.31
0.19
0.34
0.2
1.55
0.95
1.7 1 3 2
0.29
1.45
0.3
1.5
Max Power Consumed: 2.25 Watts (String 6 Clockwise)
Average Power Consumed: 1.25 Watts

25. Motor Resolution Find Hz/Degree for each peg using motorRes.c
while (input(PB1)==1) //Push Button 1 is pressed, rotate clockwise {
delay_ms(500); //wait for button to be depressed so that exact
//number of pulses measured
for (k=0; k<85; k++) //send 85 pulses to motor
output_high(motor);
delay_ms(1);
output_low(motor);
delay_ms(20); }
Adjust k until motor turns exactly 360⁰

26. Data Taken from Pic.c String Number Clockwise 360⁰ Turn [Steps]
Counter Clockwise 360⁰ Turn [Steps]
Clockwise Resolution [⁰/Step]
Counter Clockwise Resolution [⁰/Step]
No Load 51 49
7.06
7.36 6 56 84
6.43
4.29 5 60 81
4.44 4 3 2 62
5.81 1 63 86
5.71
4.19

27. Motor Resolution (continued)
String Number
Frequency at θ = 0⁰ [Hz]
Frequency at θ = 360⁰ [Hz]
Frequency at θ = 720⁰ [Hz]
Frequency at θ = 1080⁰ [Hz] 6
52.63
82.65
105.26
125 5
108.7
129.87
142.86 4
119.05
144.93
166.67
181.82 3
156.25
196.08
223.21
263.16 2
211.86
250
280.9
333.33 1
297.62
362.32
384.62
Θ = Standard Tuning – Frequency of 360⁰ Peg Turn

28. Motor Resolution (continued)
String Number
Equation 6
f = 0.067θ+79 5
f = 0.056θ+110 4
f = 0.058θ+140 3
f = 0.097θ+190 2
f = 0.110θ+250 1
f = 0.081θ+330
Observations:
Strings 6-4 have different slopes than strings 3-1 due to manufacturing process

29. Motor Resolution Results
String Number
Frequency/Rotation [Hz/⁰]
Rotation/Step [⁰/Pulse]
Resolution [Hz/Pulse]
Clockwise
Counter Clockwise 6
0.067
6.430
4.290
0.431
0.287 5
0.056
6.000
4.440
0.336
0.249 4
0.058
0.348
0.258 3
0.097
0.582
0.416 2
0.110
5.810
0.639
0.488 1
0.081
5.710
4.190
0.463
0.339
Average Clockwise Resolution = .467 Hz/Pulse
Average Counter Clockwise Resolution = .340 Hz/Pulse

30. Motor Challenges Motor designed to run on up to 6VDC
Running on 5VDC reduced torque
Running daughter card and motor on same power supply reduced current to motor
Gave motor separate battery
More power supplies tied to common ground
Needed to slow motor & still provide adequate torque

31. Motor Bracket

32. Future Improvement Increase Frequency Accuracy Flashburn
Create PCB with DSP chip embedded
No need to have unused board components
Power consumption at a minimum
Convert entire project to battery power
Reduces size
Increase Number of Tuning Configurations

33. Ethical Considerations
Complies with IEEE Code of Ethics
Tried to make kit as universal as possible
One guitar maker will not benefit over others

34. Capacitors/Resistors/Diodes
Cost Analysis
Parts:
Item
Manufacturer
Cost Per Item
Quantity
Total Cost
C6713 DSK Board
Spectrum Digital
$415.00 1
PIC 16F877A
Microchip
$3.71
Voltage Regulator
Estek
$0.90 2
$1.80
Crystal Oscillator
Fox
$0.40
Capacitors/Resistors/Diodes
N/A
$0.10 15
$1.50
Hex Display HP
$10.56
Battery 9V
Duracell
$4.00
$8.00
Protoboard
Silicon Labs
$15.00
DaughterCard
$99.99
Code Composer Studio v3.3
Texas Instruments
$995.05
Guitar Tuner TU-80
Boss
$24.95
Audio Cable 20'
Peavy
$19.95
Stepper Motor #
Parallax
$12.99
Mounting Bracket
Machine Shop
$50.00
TOTAL
$1,658.90
Labor:
$40.00/Hr x 2.5 x 200Hr x 2 People = $40,000
Total = $40,000 + $1, = $41,658.90

35. Special Thanks Professor Gary Swenson Ben Graf
Texas Instruments University Division
ECE Machine Shop
ECE Parts Shop

36. Questions
???