Guitar Amplifier with Analog/Digital Effects Fall 2013 Senior Design Project

Group 5 Members

Alex Chen (EE) Audrey Hernandez (EE) Brittany Delose (EE) Robert John (EE)

1

Table of Contents

1. Executive Summary 1.1 Project Motivation 1 1.2 Project Objectives 2 1.3 Project Specifications 2 1.4 System Block Diagram 2

2. Research 2.1 Tubes vs. Transistors 2 2.2 Pre Amplifier 3 2.2.1 Why Need A Pre Amplifier 3 2.2.2 Op Amp or Discrete 3 2.2.3 Selection of IC Op Amp 4 2.3 Power Amplifier 2.3.1 Choice between discrete and IC power amplifier 5 2.3.2 Input Stage 5 2.3.3 Voltage Amplifier Stage ` 6 2.3.4 Output Stage 6 2.4 Analog Effects 2.4.1 Introduction 8 2.4.2 Fuzz Box 8 2.4.3 Octave Up 9 2.4.4 Orange Squeezer 9 2.4.5 Tremolo 11 2.4.6 Big Muff 11 2.4.7 Distortion with Stutter Effect 11 2.4.8 Combination of Effects 13 2.5 Digital Effects 2.5.1 Introduction 13 2.5.2 Guitar Output Characterization 14 2.5.3 Prototype Board Selection 21 2.5.4 Prototype Board Discussion 22 2.5.5 Guitar Input Processing 23 2.5.6 Audio Algorithms 24 2.6 Graphical User Interface 2.6.1 Java Programming Language 24 2.6.2 Java IDE’s 26 2.7 Low-Power Wireless Technologies 2.7.1 Introduction 27 2.7.2 Evaluation Boards 28 2.7.3 Bluetooth Protocol Architecture 28

2.8 Power Supply 2.8.1 Functions of a Power Supply 31 2.8.2 Components of a Power Supply 31

2i i

3. Design Details 3.1 Pre Amplifier 3.1.1 Block Diagram 32 3.1.2 Simulation in SPICE 33 3.1.3 Parts Required 35 3.2 Power Amplifier 3.2.1 Block Diagram 36 3.2.2 Simulation in SPICE 37 3.2.3 Heat Sink Selection 40 3.2.4 Parts Required 40 3.3 Analog Effects 3.3.1 Fuzz Box 41 3.3.2 Octave Up 43 3.3.3 Orange Squeezer 46 3.3.4 Tremolo 49 3.3.5 Big Muff 51 3.3.6 Distortion with Stutter Effect 54 3.4 Digital Effects 3.4.1 No effect 56 3.4.2 Reverberation 58 3.4.3 Fuzz 60 3.4.4 Echo 63 3.4.5 Tin Can 65 3.4.6 Phase 67 3.4.7 Robot 68 3.4.8 Fuzzy Tube 69 3.5 Graphical User Interface 3.5.1 Design Summary 72 3.5.2 DSP Bootloader 78 3.5.3 Accessing eZdsp USBSTK in Java 80 3.6 Bluetooth 3.6.1 Module Description 81 3.6.2 On-board Bluetooth Protocol Stack 83 3.6.3 Hardware Setup/Connection via USB 83 3.7 Power Supply 3.7.1 Block Diagram 86 3.7.2 Simulation in SPICE 87 3.7.3 Parts Required 89

ii3

4. Design Summary 4.1 Pre Amplifier 89 4.2 Power Amplifier 91 4.3 Power Supply 95 4.4 Analog Effects 4.4.1 Fuzz Box 96 4.4.2 Octave Up 97 4.4.3 Orange Squeezer 98 4.4.4 Tremolo 100 4.4.5 Big Muff 101 4.5.6 Distortion with Stutter Effect 102 4.5 Digital Effects 4.5.1 Reverberation 104 4.5.2 Fuzz 105 4.5.3 Echo 106 4.5.4 Tin Can 107 4.5.6 Phase 108 4.5.7 Robot 110 4.5.8 Fuzzy Tube 111 4.6 Bluetooth Module 114 4.7 Cabinet Housing 115

5. Testing 5.1 Pre Amplifier 116 5.2 Power Amplifier 116 5.3 Power Supply 117 5.4 Analog Effects 117 5.5 Digital Effects 117 5.6 User Interface 118 5.7 PCB Manufacturing and Assembly 118

6. Administrative Content 6.1 Milestone Chart Discussion 118 6.2 Budget and Finance 119

7. Conclusion 120

8. Appendices A. Copyright Permissions 121 B. Acknowledgements 122

iii4

1 Executive Summary 1.1 Project Motivation

Music is an art form that can be expressed in many forms and is derived from various origins. The choice to design a guitar amplifier is to explore both music and technology. When appreciating music, it is not always apparent what is producing the notes, chords, and rhythms that are enjoyed. Exploring the technological side of this art form is the overall goal of this project. The guitar amplifier project can be broken into four main systems these include the power system, analog effects, digital effects, and the user interface. These four systems encompass a wide variety of design concepts and challenges.

In order to research, test, and finalize a design requires numerous electrical and computer engineering skills. A main feature of the guitar amp is that it will have two built in effects systems. There will be an analog effects channel and a digital effects channel. These require the knowledge and application of many electrical engineering concepts that will be thoroughly explained throughout the design report. The purpose of the combination of the analog and digital implementations will give a nice contrast to each other. Typically, there is a distinction between the digital and the analog effect systems, the choice to include both will be a unique feature of the design.

The power system requires many electrical engineering concepts. It is a custom design and requires the knowledge of designing multiple stages that will come together to power the amplifier. The system includes the power supply, pre-amplifier, and post-amplifier. The system is vital to hearing the output of the guitar and to ensure there is low noise.

The last component of the guitar amplifier is a custom user interface. This will be used to control the digital effects system. The design will cover software engineering concepts at both the low and high levels. The musician will select the effects from the interface and will be able to control certain aspects of the effects, giving the user more control over the sounds.

The guitar amplifier housing contains two main elements. The first element is the cabinet, which contains the speakers. The second element is the head, which contains the main electrical system. The type of guitar amplifier that was chosen be designed was a “practice” amplifier. This would keep the overall cost down but still proved a high quality sound. A “half stack” amplifier system contains four speakers in the cabinet. When considering the budget and overall goals of the design, the “half stack” design was determined to not be needed.

Overall the guitar amplifier is a challenge that our group was willing to take on. It provided a medium to apply many different design concepts and examine how they could all come together into one.

1

1.2 Project Objectives

This design project will fulfill the following objectives:

● Head and cabinet one piece, easily moveable ● Controlled output power ● Quality output that is low noise over a wide volume range ● Wireless (low-power) connected user interface ● Complete user range on cabinet

1.3 Project Specifications

This design project will satisfy the follow specifications:

● Output power rating: 50 Watts ● Frequency Response: 5 Hz - 100 kHz ● Distortion THD: 0.1% (1 kHz at 30 Watts) ● Input impedance: 500 k ● Output impedance: 4 - 8 ● Signal to noise ratio (S/N): 80 dB ● Bluetooth Range: Less than 50 meters ● One and a half feet to two feet tall for cabinet ● One and a half to two feet wide for cabinet ● Eight analog effects ● Eight digital effects

1.4 System Block Diagram

A functional block diagram of the system was constructed to dictate the flow of information through the system and highlight how each system will interact with each other. It describes the big picture of the design that will be explored in detail throughout the report. The first step of the diagram shown in Figure 1.4.1 is the input; this comes from the guitar. After the guitar is played the signals goes to either the digital or the analog effects block. If the digital portion is utilized then the user interface comes into play. Both effect blocks require the power’s supply as well as a wireless component for the user interface. After the guitar input has been processed either by the analog or the digital block it is passed through the pre amplifier. This will have volume and tonal control. The next block is the power amplifier; it is what regulates the power that drives the last block, the speaker. The power supply provides the power to the pre amplifier, power amplifier, analog effect, digital effect, and the wireless component to share data between the PC and the digital effects.

Figure 1.4.1 System Block Diagram

2 Research

2.1 Tubes vs. Transistors

It is very often to hear people say that tube amplifiers have better tone than solid state amplifiers because of the more “natural” and “warm” sounds produced by tube amplifiers. Tube amps have a long standing reputation for being the definitive guitar sound. The first amplifiers built specifically for guitar were all tube and are still highly sought after today. Many people describe the overall sound as “warm” because of the softer highs and mids mixed with solid bass response. They also have a tendency to clip fairly easily, which became known as “overdrive”, and in today’s case, the almighty distortion. However, there are a few drawbacks about all tube amps. First, tube amps produce a lot of heat because their supply voltages are usually high and they consume much more power. Also, they are heavy due to the power rating requirements for the bigger transformers needed. Second, tube amps aren’t as reliable as solid state amps. The life expectancy of vacuum tubes is shorter than semiconductors. In addition, vacuum tubes are made with glass so they are easy to break. Finally, tubes amps are more costly compared to solid state amps. Today, since the solid state technology and manufacturing of electronics has become so advanced that the solid state amplifiers enjoy the same good reputation as some of the best tube amps. Weighing the pros and cons between vacuum tubes and semiconductors, this guitar amplifier design will be completely based on semiconductors only.

2

2.2 Pre Amplifier

2.2.1 Why Need a Pre Amplifier

Pre amplifier is another amplifier which precedes the power amplifier. As its name implies, it prepares the signal coming from the input for further amplification. While the main function of a power amplifier provides the current gain to drive the speaker, the main job of a pre amplifier is to provide the necessary voltage gain for the input signal. Here are some functions of a pre amplifier:

1. Increases the weak input signal from mV range to 1-2 volts range 2. Cleans up a signal such as eliminating high frequency noise 3. Controls the volume and equalization

2.2.2 Op Amps or Discrete

The operational amplifier (op amp) was invented in the 40’s. Since then, op amps have gained great popularity and have been widely used in many applications. One reason that op amps have become increasingly popular is the ease of use. Op Amps can be used as building blocks by focusing on its terminal behavior. One doesn’t even need to fully understand the operation of the electronic components inside the op amp to use it. Most op amps today have the almost ideal characteristics. These are very high input impedance, very low output impedance, and very high gain. For a pre amp circuitry, some people prefer discrete components while others prefer IC op amps. Here are some advantages for op amps over discrete designs in the following:

● Components matching --- Since components are made on a same small piece of silicon, they can be made closely matched. ● High gain --- Op Amps have very high open loop gain. The closed loop gain is easily designed by some simple resistor ratios. ● High CMRR --- Op Amps usually have high CMRR (common mode rejection ratio) due to the closely matched internal component. This allows op amps reject noise. ● Simplicity --- Designs with op amps are usually neat and simple because most of the required parts are inside the IC package already. Op Amps are easy to be used by following the manufacturer’s datasheets. ● Low power consumption --- Op Amps uses much less power than a normal discrete circuit. ● Low cost --- Today we can find a very high performance IC op amp with just a few dollars.

3

So, considering the above advantages, the design for the pre amplifier was chosen to implement with IC op amps. Although IC op amps are great for pre amplifier design, they are not suitable for power amplifier design. To drive speakers, it requires both high voltage and current for the power amplifier. The required current is usually out of the op amps handling range. This is why most high power amplifiers are designed with discrete components. We will see more in the following section.

2.2.3 Selection of IC Op Amps

There are vast numbers of op amps available in the market, from top performance to average. The selection of an op amp for the design project is based on the following criteria:

● Input Noise --- Noise measured at the output and referenced back to the input ● THD + N --- Total harmonic distortion plus noise, the ratio of all other frequency components to the fundamental one. ● Slew Rate --- The rate of change in the output voltage caused by a step input. ● CMRR --- Common mode rejection ratio, the ratio of the differential voltage amplification to the common mode voltage amplification ● GBW --- Gain bandwidth product ● Price

The following table in Figure 2.2.1 is the comparison between six different types of common audio op amps.

Type Noise THD + N Slew Rate CMRR GBW Price (nV/√H ) (%) (V/us) (dB) (MHz)

LME49990 0.90 0.00001 22 137 110 $3.10

OPA134 8 0.00008 20 100 8 $2.94

OPA314 14 0.001 1.5 96 3 $0.42

AD8606 8 0.0007 5 100 10 $2.70

LMP7701 9 0.02 1.1 130 2.5 $2.70

TL071 18 0.003 13 100 4 $0.49

Figure 2.2.1 Comparison of Some Op Amps

Picking up the right IC op amps is vital for the guitar pre amplifier performance. After comparing the specifications for these op amps, it was decided to use the number two least noise and distortion, TI’s OPA134 for the pre amplifier design.

4

The OPA134 series of precision operational amplifiers achieve very low voltage noise density with a supply current of 2.5mA. It also offers rail to rail output swing which helps to maximize dynamic range. OPA134 can be operated over a wide dual power supply range of 2.25V to 18V. OPA134 has very fast settling time which makes it very suitable for fast, high precision applications such as an audio amplifier.

2.3 Power Amplifier

The basic function of a power amplifier is to control the amount of current flow from the power supply to the load. A typical power amplifier usually consists of three parts: input stage, voltage amplifier stage, and output stage.

2.3.1 Choice between discrete and integrated circuit power amplifier

It is certain that the benefits of using ICs for circuit design make the circuit much more compact and simple. However, most good amplifiers today still use discrete components. There are reasons for this. On an IC, the transistors are formed on a small piece of silicon. It is very difficult to optimize all the transistors for where they are used in the circuit. Input stage, voltage amplifier stage, output stage, etc., all have different requirements. With a discrete design, it would be possible to pick transistors suitable for the design. Also, all the resistors and capacitors inside an IC are difficult to control. In addition, the thermal coupling of all the components on the IC is another issue. The design does not need the output stage thermally coupled to the input stage. As discussed in pre amplifier section, power amplifier requires large output current which IC’s normally cannot satisfy. Based on the above considerations, it was decided to design the power amplifier for the design project with discrete components.

2.3.2 Input Stage

The most important function of an input stage is subtracting the feedback signal from the input signal, a procedure having a significant amount of error correction. Another function of this stage is to define DC operating points and buffer the voltage amplifier stage. The simplest input stage is a single transistor input stage. The transistor is configured as a common emitter amplifier. The input signal and output feedback both come in at the transistor’s base. This simple input stage topology has the disadvantages of high noise gain, low input impedance, and narrow bandwidth.

The second option of the input stage configuration is a differential circuit, also named long tailed pair (LTP), which is the most common topology. The input signal is coming into the non-inverting terminal, while the output is being fed back into the inverting terminal. The circuit’s nature of amplifying the difference of its input signals provides a couple of useful features. It basically rejects all common mode noise and makes the amplifier resistive to fluctuations of supply voltage. This negative feedback technique provides a number of advantages such as stable closed loop gain, high signal to noise ratio, and reduction of non-linear distortion. A sample of such input stage configuration is shown in Figure 2.3.1. The total current at the emitter side of transistors remains nearly

5

constant. Unequal base voltages will cause collector currents to be unequal. Then, the bias to driver changes, thus the output will change accordingly.

Figure 2.3.1 Long Tail Pair (LTP) input stage

2.3.3 Voltage Amplifier Stage

The function of voltage amplifier stage (VAS) is just simply what its name implies. It amplifies the low amplitude input signal to a level enough to drive the high power output transistors. The most common form of VAS is a common emitter circuit which is directly coupled to the output stage.

2.3.4 Output Stage

There are three main classes of operation for a guitar amplifier output stage. These are: A, B, and AB. Each class of operation has certain applications and characteristics. Likewise, each class has its advantages and disadvantages.

Class A

Class A operation is defined as a power amplifier in which output current flows for the full cycle (360°) of the input signal. In other words, the transistor remains forward biased and conducted throughout the input cycle. A simple class A topology can be described as the output of the class A circuit which has the exact copy of input signal. It has the least distortion. However, the class A amplifier has the worst power efficiency. There is always current flowing through the transistor even when there is no input signal. Indeed,

6

maximum power dissipation occurs under zero signal condition for class A. Even under ideal conditions, the class A power amplifier has less than 25% efficiency.

Class B

Class B operation is similar to class A except the transistor’s quiescent point is at the cutoff point, thus there is no base current for zero input signal. The transistor conducts for only the positive cycle of the input signal. The power efficiency of class B is about 5%. However, severe distortion occurs because of total absence of negative half cycle from the output. So, the class B operation alone cannot do the job. It needs a push and pull configuration to achieve a full output cycle.

Class AB

Class AB operation is when the transistor is biased such that the quiescent point lies slightly above the cutoff point. During a small portion of the negative half cycle and for the completed positive half cycle of the input signal, the transistor remains forward biased and conducted. But during less than the whole negative cycle of input, the transistor is reverse biased and not conducted. Like class B, it needs a push and pull configuration to achieve a full output cycle.

Complementary Symmetrical Push-Pull Operation

The simplest complementary symmetrical push-pull configuration can be seen in Figure 2.3.2. Q1 and Q2 are a matched NPN and PNP pair. During positive or negative cycle of input signal, there is only one of the two output transistors conducted. It greatly improves power efficiency. This complementary pair of transistors can be biased either in class B or class AB operation. But, class AB is the preferred method. For class B, it suffers from the nonlinear turn-on and turn-off crossover distortion. For class AB, each transistor amplifies about 55% of the input signal and combines them afterward, resulting in a full output signal. The push-pull pairs are carefully biased just above their fully off state so that there is a small overlap region that both transistors are on between the transitions. This alleviates most of the cross-over distortion at the expense of efficiency. Class AB amplifiers are still very efficient compared to class A amplifiers. The risk of crossover distortion still exists, but it won’t be readily noticeable.

7

Figure 2.3.2 Complementary Symmetrical Push-Pull

2.4 Analog Effects

2.4.1 Introduction to Analog Effects

Effect pedals are a medium to make unique sounds, allowing the manipulation of the sound waves coming from a guitar. Analog effect pedals use multiple mediums to control the sound from a musical instrument. These are not only limited to guitars, but to any other electrical instrument. For the design of this guitar amplifier, 8 effects pedals have been implemented into it for the ease of the musician. Instead of incorporating each pedal individually, musicians are able to simply stomp on their preferred effect. They are also able to mix effects by keeping their stomp switch in the “on” position. This allows the user to mix effects together to reach an acquired sound. From clipping circuits to feedback loops, effects pedal bring the guitar to another level. The manipulation of the sound waves takes a clean guitar sound to something beyond ordinary. It’s easy to see what guitar effects have grown over the years.

2.4.2 The Fuzz Box

The Fuzz Box is a popular effects pedal that has been made famous through The Rolling Stones in their song “(I can’t get no) Satisfaction”. It is a simple distortion effect that uses NPN transistors and diodes to clip the signal coming in making it nearly an almost

8

perfect square waveform. In Figure 2.4.1 it is visible to see what a clipped sine wave looks like. For the Fuzz Box, a potentiometer will control the amount of clipping within the signal. Thus the signal will vary from normal, to soft and to hard clipping with a slight turn of a knob.

Figure 2.4.1 Variations in Clipping

2.4.3 Octave Up

Octave pedals have become famous for their ability to move the octave of a guitar either up or down. The octave up pedal uses a full-bridge rectifier to make the sign wave always positive, rather than oscillating into the negative y-axis. This synthesized sound, our newly acquired octave, then gets mixed back in with the original signal producing the desired sound.

2.4.4 Orange Squeezer

The orange squeezer is a popular type of compression pedal that has been used widely throughout time. A compression pedal works as a sort of volume adjustment for a guitar amplifier. Imagine plucking the strings on a guitar there are times where some strings are plucked harder or softer than the others. A compressor helps equalize those sounds coming out of the guitar amplifier. When the soft notes are played, the compressor amplifies the sound to mimic turning up the volume knob, and then turns it down for the louder notes. The compressor has many variables that control it: threshold, knee, ratio, gain, release, and attack time.

9

● Threshold controls the maximum level for the signal to be before it gets compressed ● Knee is the shape of the volume adjustment; it can be a quick adjustment or it can take its time to get to the threshold ● The ratio is how much power of the signal gets turned up or down (i.e. if the ratio is 4:1, for every 4 dB, the signal gets increased by 1 dB) ● Gain which could be called “make-up gain,” is the amount of gain added to the signal typically before compression ● Release is the amount of time the signal is actually being compressed ● The attack time is the amount of time it takes for the effect to kick in

Two signals are given in the Figure 2.4.2. The red line is a signal without any compression. The input is equal to the output volume. The green curve however shows how the signal gets compressed. For example, with an input volume of -80 dB, the output is at -60 dB. This is an amplification of 20 dB. At -20 dB input volume, the output volume is a respective -8 dB. It then gradually decreases, as the input volume decreases.

Figure 2.4.2 Compressed and Normal Signal

10

2.4.5 Tremolo

The tremolo is an effect that is centered on a signal’s frequency. One of the main factors for the tremolo is a positive feedback loop. As the signal comes back in phase, a low-pass filter is incorporated to produce frequencies at the subtonic level (below 20 Hz). This feedback loop followed by a low-pass filter produces a low frequency oscillator. Most tremolo low frequency oscillators work at frequency range of 3-10 Hz. The new signal coming from the low frequency oscillators then feeds into an opto-isolator. As the peaks of the sine wave hit the opto-isolator, the resistance within it decreases. As the sine wave gets closer to 0, the resistance in the opto-isolator then increases. This would have the same effect as turning the volume knob up and down. There is an intensity knob that helps controls the signal strength as to not cancel it out completely.

2.4.6 Big Muff

The Big Muff was created in the late 1960’s by Mike Matthews who partnered up with Bob Myer from Bell Labs. It was a hit in the 1970’s and most musicians had it in their pedal line-up. Although there are multiple versions of the Big Muff, the decision was made to go with the first original version, the Triangular Big Muff. The Big Muff consists of four transistor stages.

● The first stage is considered a clean boost, which then feeds and drives the next two states. ● The second and third stages are the clipping stages which provide the distortion effect of the Big Muff ● The final stage is a tone recovery stage that recovers the volume lost through the second and third stages

NPN silicon transistors are incorporated into the Big Muff pedal for the four stages. There are three knobs to the Big Muff: Volume, Sustain and Fuzz (Tone). All knobs are 100K potentiometers implemented into the design. The sustain knob is located after the first transistor stage. The tone is after the third, and finally the volume is on the last stage of the Big Muff circuit. Although the Big Muff is known for going through so many changes through its 40-year stance, it is one of the most well-known guitar pedals in history.

2.4.7. Distortion with Stutter Effect

One of the most common effects pedals is that of distortion. In fact, some guitar amps have the distortion effect built into it. But the further the distortion, it was decided to add a stutter effect as well to the pedal. Before diving into a stutter effect, distortion must be considered. If a guitarist cranks their tube amplifier all the way, an overdriven sound can be heard coming out from it; this sound comes from the vacuum tubes being pushed to their limits. This is considered overdrive. Now distortion overdrives that overdriven sound, giving the guitar more gain. At the guitar signal’s peak, compression will occur and overlay of overtones are then added to give it its grungy sound. There are several sounds of distortion that can be achieved. As stated before, tube amps help achieve a

11

warm sound from a guitar. Some pedals use germanium circuits to also get the warm tube amp sound. There are some now that use silicon that give a brighter, more vivid sound.

The final addition to the effect pedal is that of the stutter effect. A stutter effect is like a glitch. As audio signal comes in, a sample of it gets chopped up into small fragments. This stutter occurs when the signal is cut up into pieces. Every other piece then gets “muted” out of the final sound. An audio signal illustrated in Figure 2.4.3 reprinted with permission from FocalPress is divided into several pieces. The pieces that are in green are what are being played, while those that are grayed out are not. Within the distortion pedal, this is done with a 555 timer that creates a square wave. The square wave is then multiplied into the distorted signal coming through, and takes out some of the signal. The timer can then be programmed to have a desired duty cycle to give the wanted effect.

Figure 2.4.3 Input Signal with Stutter Effect

12

2.4.8 Combination of Effects

The layout of the effects pedal can play a big role on how the sound from the guitar comes out of the amplifier. From research, the order of the pedals really plays a vital part in the sound of the guitar. The first set of pedals will have the most effect, since they are going to be the first effects on the signal. The first set would be the distortion and compression pedals. These are the pedals that are going to push the gain of the signal. Next will be the modulation signals such as the octave up pedal. Final set of pedals are the ones that take away from the sound, such as the Tremolo pedal. The following Figure 2.4.4 represents a diagram of how the boards will be laid out.

Figure 2.4.4 Flow Chart of Pedal Layout

2.5 Digital Effects

2.5.1 Introduction

The first step in designing the digital effects system was to decide how the audio data would be input, processed, then outputted. A general purpose processor was not needed because the functionality needed for the system mainly included taking in audio data, processes it, then produces output. The goal was to select a DSP that included what we needed without unnecessary features. At first glance it there were a large selection to choose from, keeping in mind our application was the first step to narrow the window of possibilities.

A very basic view of what we needed is ADC, DSP processor, and DAC with built in input and output jacks. Specifically the DSP needed capable of real time audio processing. The input frequency is relatively low in comparison to the sampling rates of the processor, ADC, and DAC. Keeping the basic things in mind aided in the selection of the DSP that was to be implemented.

13

After DSP was selected the corresponding evaluation board was selected. The purpose of buying the evaluation board was to develop the programming logic without laying out the custom PCB. It was a requirement that the evaluation board have everything we needed to generate effects, including a IDE that was under friendly and at no cost. The evaluation board was selected based off both functionality and cost. The upper limit of the budget was one hundred and fifty dollars. There were a few boards in this price range and there are dozens of brands to choose from. Naturally, the first choice was the company that was going to be selected. After researching various companies, Texas Instruments was selected. We want to guarantee that the parts will be in production throughout the duration of the design course. TI also satisfied the requirement of a board that would come with an IDE. TI’s IDE is Code Composer Studio and is free to download on their website. Code Composer Studio is where the audio algorithms were designed. After the finalization to the effects the PCB parts were selected. The PCB component selection was inspired by the evaluation board layout, because it was working properly on the evaluation board, and it was best not to change something that already worked.

2.5.2 Guitar Output Characterization

The guitars output is depended by the player and changes dynamically depending on what string is plucked or strummed and the finger positions on the fretboard. When the player varies there finger position it controls the frequencies that are output. The mechanical vibrations of the strings are received through the pickups and converted to an electrical signal. The pickups are embedded on the guitar.

In order to make the digital guitar effects the input needed to be characterized. A mono fourth inch jack was used as the connection to the output side of the guitar cable. The goal was to see the output voltage and frequency without any amplification. The oscilloscope was connected directly to the output of the mono jack. The results from this test were inconclusive because the signal was too small and contain too much noise to distinguish between waveforms. Because of this fact the waveforms are omitted. A simple solution to this problem was to use an operational amplifier. The operational amplifier that was used was the TI TL084CN. Although the first set of data was inconclusive it gave an approximate rage of the input frequency and voltage. Based off the data, a simulation could be utilized. The input voltage signal was assumed to be about 5mV and the frequency seemed to vary between 100Hz and 1kHz. The following simulation was run in order to see the best way to amplify the guitar signal with minimal attenuation.

The operational amplifier was connected in the non-inverting configuration and the gain was 101. The input voltage was a 5mV 200Hz sine wave. After running the simulation the circuit was implemented on a breadboard. The conclusion that arrived from testing was the the gain was too high and the waveform was being clipped. Clipping occurred because the output voltage was exceeding the Vcc supply voltage which was 15V. If the output times the gain exceeds the supply voltage then clipping occurs. Although clipping is desired in our effects system in certain cases, for characterizing the input data it was not desired. The next step was to lower the gain of the system. The resistor labeled R5

14

above was changed from 100kOhms to 47kOhms it is show below in Figure 2.5.1. The output wave form is shown in Figure 2.5.2.

Figure 2.5.1 Non-inverting Gain Circuit

Figure 2.5.2 Output of Non-inverting Amplifier

15

After the simulation was run the circuit was implemented on the breadboard and the output was read off the oscilloscope. The decrease in gain in this configuration allowed for the waveform be displayed without any clipping. Noise was not accounted for in the simulation, two capacitors a .001uF and a 10uF capacitor were added once collecting the guitar output waveforms. The guitar signal severely attenuated by noise and the addition of the capacitors was a remedy to the solution.

Below are a few waveforms of single notes and chords in order to gain an understand of what signal is being modified to create the desired effect. The figures are notes and chords that are played on the Fender Telecaster.

The first three waveforms that were collected were from played the open E note. It is played simply by plucking the top string without pressing down on the fret board. The following Figure 2.5.3 and Figure 2.5.4 shows what happens as the note is let played. There is a point where there are two peaks to the periodic wave and there is a point to where the wave appears to be a sine wave. If you were to view the oscilloscope when the measurement was taken it could be noted that the wave cycles through the single peak and multiple peak waveforms.

Below is the first of the two waveforms. There is a point in the cycling of the note that the wave that it has a single peak. This waveform shows a small secondary peak.

Figure 2.5.3 First E String Waveform

16

The waveform below is the open E string with the multiple peaks. The difference in amplitudes of the peaks is smaller than Figure 2.5.3.

Figure 2.5.4 Second E Sting Waveform

The figure below is the low and high E plucked as the simultaneously with no fingers pressing on the fret board. It should be noted that this waveform is not as smooth as the Figure 2.4.5. This is because the waveform was collected soon after the strings were plucked, this cause the additional distortion to the waveform.

17

Figure 2.5.5 Low and High E Strings Waveform

Below in Figure 2.5.6 is the E minor chord. Because two string were pluck at the same time to produce this wave form.

Figure 2.5.6 E minor Chords Waveform

18

Below in Figure 2.5.7 is the F bar chord. This chord is not as smooth as the previous E minor chord. It is because it is more difficult to collect a stable picture when strumming all six strings at once.

Figure 2.5.7 F Bar Chord Waveform

The last waveform that was collected was the highest note that could be played on the Fender Telecaster using standard tuning, it is shown below in Figure 2.5.8. The amplitude is smaller than the low E string and the frequency is higher as noted when comparing the two figures.

19

Figure 2.5.8 Highest Note Waveform

2.5.3 Prototype Board Selection

In order to select the evaluation board that would be best for the testing of the digital portion of the guitar amplifier, the minimal requirements of the system needed to be established. The goal was to find a evaluation board that would have enough functionality so that no external parts were needed. This would enable the designing of the processing algorithms to be the focus and not the design of the board. An important aspect of the is that the quality of the output guitar signal not be degraded. The guitar input has the frequency range of about 0 to 4kHz. This is a relatively low signal, but the ADC and DCA needed to perform in a manner that would not degrade the respective input signals. The evaluation board needed a way to receive the guitar input then output it after it had be processed. The I/O ports needed to be included on the evaluation board. Finally the board needed to have a IDE that was free and user friendly. The IDE was essential to the design of the digital effects system. It is where all the processing algorithms were designed and we needed one that could satisfy our programming needs. Like stated previously we selected TI as a starting point to pick the evaluation board because of it’s IDE Code Composer Studio.

20

TI has a wide selection of DSPs, and once selected, a few choices of evaluation boards. The first choice that was made was to pick from the C55X family. It was discovered that the C55X family had been used previously in audio applications and even lists “Musical Instruments” when looking through its application descriptions on TI’s website. The next step was to select an evaluation board, and the first choice was the TI’s TMDXEVM5515. This board is the evaluation model for the TMS320C5515 DSP. The first concern about this board was not the functionality, but it was the cost. According to TI’s website this board retailed for 395 dollars. Although it greatly surpassed the minimal requirements for the digital effects system, the cost was a little over two times what was allocated in the budget. Due to the cost the board was not inconsideration.

The TMS320C5515 eZdsp USB Stick evaluation module was then next evaluation module to be considered. The first thing that was noted was the board was within the budget. According to TI’s website the total cost was 79 dollars. It was a relief to find an evaluation module that was in our price range and one that was actually under the maximum amount allocated in the budget. The features that we needed were all included on the eZdsp. The ADC and DAC are on the TLV320AIC3204 codec that is included on the evaluation board. The ADC is sampled at 48ksps and the DAC playback was at 48ksps as well. This sampling rate met our minimum requirement. The eZdsp also included two stereo I/O ports. It is also compatible with code composer studio and connects through USB 2.0 port. The board, like most, included a few buttons on board and a LCD display. Although these are not needed in the final design they were appealing for the debugging and prototyping design stage. After concluding that this board was under budget and included all the components critical to the prototyping design it was selected.

2.5.4 Prototype Board Discussion

The TMS320C5515 eZdsp It is a fixed point DSP that is low power and has a USB 2.0 interface. The board interfaces with TI’s Code Composer Studio development environment. For our purposes it was the best choice for an evaluation board. We utilized the eighth of an inch stereo jacks for the input and output on the TMS320C5515 eZdsp. The eighth of an inch jack is the standard size for headphones. The guitar input and output cable uses a fourth of an inch jack, two adapters were purchased for the input and output so no modifications were necessary on the evaluation board. When prototyping it was nice to have the ability to use eighth or fourth of an inch output on the eZdsp because the output could be heard through headphones or could be connected to an existing external amplifier. This ensured that volume control would not be a problem when testing and anyone in the audio testing zone would not be disturbed.

There are additional features on the board that were utilized for debugging. The LEDs were available to indicate when certain parts of a program were reached. Also there are two switches on board used to simulate the footswitch that will be implemented in the final design. Figure 2.5.9 shows the TMS320C5515 eZdsp.

21

Figure 2.5.9 TMS320C5515 eZdsp

2.5.5 Guitar Input Processing

The analog data from the guitar needed to be digitized in order for the signal processing to take place. The audio codec on the TMS320C5515 eZdsp is the TLV320AIC3204. The analog to digital information is converted at a rate of 48ksps. This is well above the nyquist rate of the input signal which is around a maximum of around 4kHz. The TLV320AIC3204 has programmable amplifiers that can have a gain of 0 to +47.5dB according to the data sheet, so no external amplification is needed in order to hear the output of the stereo jack during the prototyping phase. The there are two ADC built in to the TLV320AIC3204. The sampling rate can be from the range of 8kHz to 192kHz. According to the data sheet, the ADC processing blocks are the signal processing blocks available are First-order IIR,Scalable number of biquad filters,Variable-tap FIR filter, Automatic gain control. These are the recommended processing blocks in order to limit power consumption. These may be utilized during the algorithm design stage. According to TI’s TLV320AIC3204 data sheet “The processing blocks are tuned for common cases and can achieve high anti-alias filtering or low group delay in combination with various signal processing effects such as audio effects and frequency shaping.” The high anti- alias filtering is desired because aliasing would degrade the audio signal.

The waveforms above show the rage of the input. The signal produced from the guitar closely resembles a sine wave. The CODEC’s fast sampling rate allows for a clear input signal to be processed and therefore will allow for a nice sounding output.

22

The input data travels down the I2C DSP bus and is then processed. The data size of the bus can be 16, 20, 24, or 32 bits. The size of that was chosen was 16-bit

2.5.6 Audio Algorithms

The audio algorithms are developed in TI’s IDE Code Composer Studio v4.1.3. The software was included with the TMS320C5515 eZdsp and is available on TI’s website. The first challenge of the audio processing development was simply getting the input form the guitar, having the signal processed, and then be able to hear the guitar output. There are a few parameters and initializations that each algorithm needed in order for the audio information to be processed, they are going to be discussed below.

The maximum sampling frequency that was used was 48kHz. When programming the additional values that were allowed to be selected were 48kHz, 24kHz and 12kHz. The slowing of the sampling rate causes a change in how the effects sound and will be addressed in the design section of the report. The Nyquest criteria for sampling rates says the sample twice the highest frequency, which is around 4kHz.

There is also control of the gain in software. The output gain of the ADC could be set up to +47dB. When plugging the headphone jack into the output of the TMS320C5515 eZdsp the gain was set to around 10dB. If the gain is set too high it caused attenuation and was not a comforting audio level to listen to.

Following these two values, the audio CODEC which contains the ADC and DAC needs to be initialized along with the programmable phase locked loop. The phase locked loop was set to 100Mhz. The phased locked loop makes sure the input and output do not fall out of phase. The guitars signal will be between 0-4kHz so the criteria will be satisfied.

In each effect there was a CODEC read function. This took the analog data in and converted it to 16-bit input. The data was stored into left and right input variables. The main function would then read from the addresses of this stored information and do a stereo to mono conversion. In this function it would average the left and right input data and return to main.

Last the algorithm would process the mono left and right signal and passes it to the CODEC write. This sends the data to be converted back to analog and down the I2C bus.

2.6 Graphical User Interface

2.6.1 Java Programming Language

An overview of the Java language can be described as simple, object-oriented, independent of the host platform, and based on the language C++. The Java Platform (Standard Edition) includes various GUI libraries such as the SWT, Swing Toolkit, and the AWT. These libraries are considered usually either lightweight or heavyweight. Lightweight/heavyweight refers to how the widgets are “drawn”, where the former are

23

painted using its own API and the latter are painted using the OS API (where API refers to the description of a set of class definitions).

The SWT (Standard Widget Toolkit) is developed by IBM and relies on the native operating system to create the GUI. The Swing library (released by Sun Microsystems) is part of the Java Foundation Classes (which is the collection of API’s for creating GUIs for Java programs). The JFC contains Swing, AWT, Accessibility, Java2D, and Drag and Drop. Specifically, the Swing toolkit is lightweight, configurable, extensible, customizable, and platform independent. The Swing API includes public packages, such as javax.swing.JFrame, javax.swing.JButton, and javax.swing.SwingUtilities, which allow for a rich set of widgets and the abilities to make these widgets functional as a desktop application (all-Java language). The following figure 2.6.1 displays a table of the complete set of packages included in the Swing API. The AWT (Abstract Widget Toolkit, also released by Sun Microsystems) contains classes for painting graphics and images, such as java.awt.ActionEvent and java.awt.event.ActionListener and also provides the use of Drag and Drop (through importing java.awt.dnd).

Figure 2.6.1 Swing Toolkit API

The Java language structure begins with packages. Classes are within packages, and within these classes there are methods, variables, constants, etc. Since the Java language is considered an OOP language, it combines data and program instructions into objects. An object is a discrete entity which contains necessary attributes and behavior which is dependent upon other objects to perform its tasks. A parent object serves as the basis for creating more-complex child objects (more specialized). Objects can communicate with other objects by using the method call. An object’s state can be represented at any time (by specifying the value of its attributes). Principles of OOP include encapsulation, inheritance, and polymorphism. Encapsulation enforces the idea that an object can be represented as public or private access. Private access means that the object’s attributes are accessible only within the object itself. Inheritance is the concept of copying the data and logic of the source object without additional code. Changing attributes/behaviors can be done by overriding them in order to create specialized objects (reinforcing the idea behind parent/child objects). Polymorphism means that an object belongs to the same branch of a hierarchy (a more complex concept). In order to define an object, a class must be declared (since a class defines the structure of a thing of which the object is an instance). Class names should begin with a capital letter and should only contain letters and numbers.

24

The Java Development Kit includes a tool which is used to create Java Archive Files (JARs). Using an IDE, for example Eclipse, makes creating a JAR file quite easy. A JAR file allows other developers to use the file and configure it into their projects.

The java.io package offers tools to collect and manipulate data from a variety of sources. Using streams allows a program to receive bytes from a source or to send output to a destination. The Reader and Writer types handle all kinds of 16-bit characters, where InputStream and OutputStream types handle only 8-bit bytes. The main types of streams are byte streams and character streams. The following are two common byte streams which are considered “raw” type:

FileInputStream/FileOutputStream: Read/Write bytes to a file ByteArrayInputStream/ByteArrayOutputStream: Read/Write bytes to an in - memory array

The following character streams read and write 16-bit characters:

StringReader/StringWriter: Read/Write characters to and from memory InputStreamReader/InputStreamWriter: Bridge between byte/character streams BufferedReader/BufferedWriter: Buffer data while reading/writing another stream

Object serialization is the process where the state of object and its metadata are stored in a special binary format, which preserves the important information to reconstitute the object later on. Remote object serialization sends the object to another computer or system, whereas object persistence means storing in a database.

2.6.2 Java IDE’s

Using a Java Integrated Development Environment is an easy way to develop a rich, customized user interface that relies less on writing tedious code and implements drag and drop functionality. The most appropriate factors considered were cost, user ability for application, and quality. Investigating various free and open source GUI builders has led to a few options including Eclipse and Netbeans. Researching these two IDE’s has yielded the conclusion that choosing one over the other is preference based on the developer, especially since the most recent versions of each were very similar in specs and features (expected competitor behavior). Consequently, the table below summarizes the many similarities (and very few dissimilarities) between the two GUI builders. From a slight beginner’s perspective on the Java programming language, it would be sufficient to use either Eclipse or Netbeans to build a thick client application. Figure 2.6.2 displays certain parameters which were researched in order to choose an applicable IDE for the design project.

25

Figure 2.6.2 Open-source Java IDE Comparison Table

2.7 Low-Power Wireless Technologies

2.7.1 Introduction

Since there wasn’t a need to transfer large amounts of data, Wi-Fi wireless technology was omitted and the focus led to the research for wireless personal area networks (WPANs).

Bluetooth Low Energy (Bluetooth v4.0) provides extremely low power consumption within a range of about 30 meters and operates in the 2.4GHz frequency band. It is not compatible with standard Bluetooth, but can be implemented through single-mode or dual-mode devices.

ANT is also a low power, short-range wireless technology which operates in the 2.4GHz frequency band. Research demonstrates that this technology is prominent and growing in sports and fitness applications.

Zigbee (a low-power wireless specification based on an IEEE 802 standard), is targeted towards smart meters, home automation, and remote control units. Zigbee is not a frequency hopping technology, unlike the other two mentioned above. Figure 2.7.1 shows a table which organizes the three wireless technologies in a convenient way so as to help the process of choosing the best technology for the design project.

Figure 2.7.1 Table for Wireless Technology Comparison

26

2.7.2 Evaluation Boards

Panasonic offers the PAN1323ETU (an upgrade from the no longer available PAN1315ETU), a development board which plugs directly into the MSP430 experimenter board. It supports three standards (BT Classic, BLE, and ANT.) Its BT stack is provided by the third party MindTree. It also plugs directly into the DSP prototype board chosen for the digital effects (TI’s TMS320C5515 eZdsp), but further research concluded that there is no developed stack available (except for very high licensing fees). The remainder of the evaluation boards researched were explored based on this realization. STMicroelectronics offers a development board based on the Bluetooth v3.0 class 2 (short range) module (SBT2632C2A.AT2) which connects via USB. The Bluetooth stack is embedded with profiles such as GAP, SDP, and SPP. The device is configured through the AT command set. Roving Networks features the RN-42 evaluation board (Bluetooth v2.1 + EDR), which has ha range of 30 meters and can be connected via USB or through the TTL UART interface, and also includes embedded Bluetooth stack profiles. The device is programmed and controlled from a console using an ASCII command set.

ConnectBlue offers the OLS426, a BLE development board with UART logic level interface. The firmware is embedded in the module which is configurable via AT commands. It supports a range of 200 meters, far more range than the design project would need).

2.7.3 Bluetooth Protocol Architecture

The lowest layer of the BT specification is the radio. It operates in the 2.4GHz radio band with frequency hopping sequences with 79 RF channels each separated by 1MHz. It uses the Gaussian Frequency Shift Keying (GFSK) modulation for transmitting where the index must be between 0.28 and 0.35. The receiver has a -70dBm sensitivity level for a 0.1% bit error rate.

The next level lies on top of the radio layer known as baseband which is the physical layer that manages physical channels and links. One master and one or more slave(s) using the same physical channel form a piconet. The hopping sequence is determined by the BT device address (of the master) and the phase is determined by the BT clock (of the master). The baseband supports two types of links, the SCO and the ACL, where the synchronous connection-oriented link mainly carries voice information. All data on the piconet channel is transported in packets, where each packet consists of the access code (68/72 bits), the header (54 bits) and the payload (0-2745 bits). The access code is used to determine paging and inquiry, offset compensation, and timing synchronization. The header includes the information for packet acknowledgement, flow control, slave address, and error check. The payload includes a data field which also contains a payload header.

27

The next level on top of the baseband is the link manager which discovers other remote link managers and communicates with it through the LMP (link manager protocol). This protocol includes a number of protocol data units (PDUs) which are sent between devices and are determined by the 3-bit number defined from the active member address (AM_ADDR) in the packet header. DM1 packets are used to transport these link manager protocol data units.

The Host Controller Interface (HCI) provides an interface (usually via keyboard ASCII commands) to the baseband link controller and the link manager, as well as access to hardware status and control registers. The HCI firmware (located on BT device) access baseband commands, link manager commands, hardware status registers, control registers, and event registers. The HCI driver (located on the OS unit) receives asynchronous notifications of HCI events and parses the received event packet to determine the event which occurred. The HCI firmware and driver communicate through the host controller transport layer, which provides the ability to transfer data. Three layers are defined for Bluetooth known as USB, UART, and RS232. The USB transport layer uses a class code specific to all USB BT devices and allows proper driver stacks to load, allowing HCI commands to be different from USB commands. The UART transport layer allows the use of the BT HCI over a serial interface between two UARTs. Event and data packets flow through this layer; however they are not decoded in this layer. The RS232 is a permitted physical interface between the BT host and the BT host controller. It functions similar to the UART interface.

The logical link control and adaptation layer protocol (L2CAP) is situated over the baseband layer and provides protocol multiplexing, segmentation and reassembly, and group abstractions. Therefore, this protocol permits higher level protocols and applications to transmit and receive data packets (up to 64k bytes in length).

The RFCOMM protocol provides emulation of RS232 serial ports over the L2CAP protocol. It is based on the ETSI standard TS 07.10.

The service discovery protocol (SDP) allows for applications to discover available services and determines the characteristics of the corresponding services. It uses a request/response model where each transaction consists of one request PDU and one response PDU. Each SDP PDU has a header containing the PDU ID, the transaction ID, and the parameter length. The figure 2.7.3 illustrates how the fundamentals described above are arranged, creating a layered stack.

28

Figure 2.7.3 Bluetooth Protocol Stack

29

2.8 Power Supply

2.8.1 Functions of a Power Supply

It is needless to tell how important a power supply will be. The function of a power supply is to transfer the 115 VAC from the wall outlet into the necessary DC’s for different modules of the project. A simple input and output of the power supply is shown in Figure 2.8-1.

Figure 2.8.1 I/O of Power Supply

The power supply has a single input and 4 outputs. The outputs of 30V, 15V, + 9V, and + 5V are for the power amplifier, pre amplifier, analog effect, and digital effect modules, respectively.

2.8.2 Components of a Power Supply

A typical DC power supply contains four major sections: transformer, rectifier, filter, and regulator.

Transformer

A step down transformer is needed for converting the 115 AC into lower AC, in the range of DC’s required. There are usually three types of secondary winding for a transformer: single, dual, and center tap. Due to the dual polarity of DC’s needed, a center tap transformer was chosen.

Rectifier

Rectifier will be a simple diode circuit. Three types of rectifier circuit are: half wave, full wave, and bridge. Bridge rectifier will be the most efficient way. Plus, most bridge rectifiers are manufactured in a single piece component. The output of a bridge rectifier is 120 Hz positive cycles only sinusoidal wave.

30

Filter

A filter circuit is to remove as much AC ripple as possible. This project uses a regular RC filter due to its simplicity. The resistor and capacitor together form a low pass filter that smooths out the ripple. Technically, the larger the time constant RC is, the smoother the output would be. However, in practice, capacitance should be picked as long as it satisfies the requirement of an application, for the cost of higher capacitance.

Regulator

A voltage regulator is to maintain the output voltage at a steady level. There are four output voltages for the power supply. The 30 VDC are supplies for the class AB power amplifier. The practical design is not to use regulator for such power amplifier due to the large load current. The reason is consideration of both power loss and power handling capacity of the regulator. Also, a small percentage of AC ripple is tolerated in high power circuit. However, the 15 VDC, 9 VDC, and 5 VDC are the supplies for pre amplifier, analog, and digital effect modules. They are required to be precisely regulated. Because these circuits are low current and low power, regulator IC’s will not only handle the job also makes the circuit very simple.

3 Design Details

3.1 Pre Amplifier

3.1.1 Block Diagram

The pre amplifier will have two single channel IC op amps to ensure it has the adequate gain. The band pass filter is to filter out the extreme low and high frequency. Gain & tone control is the adjustment for bass, mids, treble, and volume. Figure 3.1.1 is the block diagram.

The electric signal coming out from an electric guitar is at the level of 100mV - 200mV. It first goes into the band pass filter. The band pass filter has a large bandwidth. Its function is to roughly eliminate the extreme high frequency noise. First stage op Amp is a single channel Op Amps IC. It not only amplify the input signal also forms an active filter, which again eliminate the extra high noise. The gain and tone control is the volume and equalization circuit. It control how much of high, mid, and low frequency of the signal that the circuit amplifies. Also, the volume control is just a simple resistor control that how much signal passed into second stage Op Amps. The second stage Op Amps is another single channel Op Amps IC that is identical to the first stage. Two Op Amps, instead of one, is used in this case because it would have more distortions to use just one Op Amps to amplify the input signal too much. So, it is rather to use two amplifying stages to minimize the distortions.

31

Figure 3.1.1 Pre Amplifier Block Diagram

3.1.2 Simulation in SPICE

The designed pre amplifier circuit was simulated in Multisim. The input was 1 kHz 100 mV peak sine wave since 100 mV is the average electric guitar output signal. All potentiometers are set to the mid way (50% value). The result of output vs. input waveform is illustrated in Figure 3.1.2. The output signal is shown as a clean enlarged input signal without visual distortion.

Figure 3.1.2 Output vs. Input of Pre Amp

For the following simulations to test all three individual frequency range band, the volume knot of VR4 is turned all the way up (the loudest). To test the frequency response of the treble, VR1 is turned all the way up. VR2 and VR3 are turned all the way down. The frequency response of the treble effect is illustrated by Figure 3.1.3. It shows the maximum gain about 45dB happened at the high frequency of 6 kHz.

32

Figure 3.1.3 Frequency Response of Treble

To test the frequency response of the bass, VR2 is turned all the way up. VR1 and VR3 are turned all the way down. The frequency response of the bass effect is illustrated by Figure 3.1.4. It shows the maximum gain about 38dB happened at the low frequency of 60 Hz.

Figure 3.1.4 Frequency Response of Bass

To test the frequency response of the mid, VR3 is turned all the way up. VR1 and VR2 are turned all the way down. The frequency response of the bass effect is illustrated by Figure 3.1.5. It shows the maximum gain about 35 dB happened at the mid range frequency of 10 Hz to 10 kHz.

33

Figure 3.1.5 Frequency Response of Mid

The total harmonic distortion (THD) as well as signal to noise ratio (S/N) are also obtained in Figure 3.1.6. For input of 1 kHz 100 mV signal, the THD is 0.001% and S/N is above 8 dB. These satisfy the requirements of an audio amplifier.

Figure 3.1.6 THD and S/N of Pre Amplifier

3.1.3 Parts Required

Having more precise values of components is critical for audio amplifiers. All resistors are chosen with metal film and ¼ Watt. Metal film resistors have much less variations in values comparing with carbon resistor. Coupling capacitor are polyester capacitor, while others are ceramic capacitors. Potentiometers VR1, VR2, and VR3 are linear scale, while VR4 is log scale.

● Op Amp: OPA-134 (2) ● Resistors: 10 ● Capacitor: 13 ● Potentiometer: 4

34

3.2 Power Amplifier

3.2.1 Block Diagram

In order to design a 50W output power amplifier, we are considering to use the following voltage and current specifications: Vrms = 20 V, Vpeak=28 V, Irms=2.5 A, Ipeak = 3.5 A, impedance of the speaker = 8 ohm. The concept of our design based on the above specifications is shown in Figure 3.2.1. For a negative feedback control system, close loop gain is equal to A / (1 + Ab), where A is open loop gain and b is feedback ratio. For an ideal amplifier with a very high open loop gain, the close loop gain is approximately equal to 1 / b. We selected b = 0.1, so we have close loop gain = 10. When the input is 2 Vrms, output is 20 Vrms so the maximum output power of 50 watts will be delivered.

Figure 3.2.1 Design Concept

The design is based on a common three stage audio power amplifier topology. The block diagram for such a design is shown in Figure 3.2.2.

The input stage is the differential amplifier, which are form by two identical PNP transistors. The stable bias is provided by a current source from a third PNP transistor. The second stage is to amplify the signal voltage. It is completed by a NPN transistor in the common emitter configuration. Also, the common emitter configuration works greatly between the input and output stages because it couples their input and output impedance. The output stage is the complementary push and pull configured with Darlington pairs. The Darlington pairs are used instead of simple pairs, for a single output transistor would not be sufficient to output the large power. This power amplifier also has the circuit protection. A short load circuit is the most often fault a power amplifier can happen. When this happens, there is very large current flowing through the power transistors. Eventually, it would burn those transistors. The purpose of the circuit protection is to eliminate the current at the power transistors when such situation occurs.

35

Figure 3.2.2 Power Amplifier Block Diagram

3.2.2 Simulation in SPICE

Using one of the SPICE program, Multisim, I simulated the circuit with 1 kHz sine input. The output of Vpeak = 20 V is a clean enlarged duplicate of input for a 2 V peak input signal. As the input increases to 3 V peak, the output starts clipped at Vpeak = 27 V. The clipping effect is due to the 30 V power supply limit. Through the simulation, 2.5 V peak is the maximum input level before the clipping happens. The waveforms that before and after the distortion are shown in Figure 3.2.3 and Figure 3.2.4 respectively.

Figure 3.2.3 No Distortion (2V I/P)

36

Figure 2.2.4 Distortion (3V I/P)

The frequency response of the simulated circuit is shown in Figure 3.2.5. In frequency range of 5 Hz - 100 kHz, magnitude response maintains steady at 20 dB and phase response maintains 0 degree. So it well covers frequency range of 100 Hz - 10 kHz for a typical electric guitar.

37

Figure 3.2.5 Frequency Response of Power Amplifier

The total harmonic distortion (THD) as well as signal to noise ratio (S/N) are also obtained in Figure 3.2.6. For output at 30 watts and input of 1 kHz, the THD is 0.1% and S/N is above 75 dB. These satisfy the requirements of an audio amplifier.

Figure 3.2.6 THD and S/N of Power Amplifier

38

3.2.3 Heat Sink Selection

The performance of a power transistor is dependent on its ability to dissipate the heat it generates. One way to remove the heat effectively is to attach a heat sink to the power transistor. The power transistor deliver not only output power to the speaker also the waste heat power the air through its case. The waste heat the power transistor generates in this project is about 10 W. The thermal resistance required for a heat sink is calculated by the following formula:

Rsa = (TJ - Ta) / P - Rjc – Rcs

The thermal resistance formula parameters are defined as:

Rsa - thermal resistance for heat sink Rjc - thermal resistance for junction to case Rcs - thermal resistance for case to heatsink P - waste heat power TJ - Max temp for junction Ta - ambient temp

Therefore, the thermal resistance required is calculated as:

Rsa = (150 C - 25 C) / 10 W - 1.56 C/W - 3 C/W = 8 C/W

This means that a heat sink with minimum thermal resistance of 8 C/W was chosen to ensure that the amplifier could deliver 50 W of power.

3.2.4 Parts Required

Using precise values of resistors is vital for the quality of amplifier. All resistors except two emitter resistors are metal film and ¼ watt. The two power transistor’s emitter resistors are wirewound 3 watts because their power handling requirement. All coupling capacitors are polyester. High frequency bypass capacitors are regular ceramic. The two diodes are very common fast switching 1N4148.

● Transistors: BC556 (3), MJE182, BC639, 2N3904, 2N3906, TIP29C, TIP30C, TIP33C, TIP34C ● Diodes: 1N4148(2) ● Zener: 1N5242B ● Capacitors: 10 ● Resistors: 23 ● Heat sink: 4

39

3.3 Analog Effects

3.3.1 The Fuzz Box

The main component of the Fuzz Box is the 2N2222A transistor. It is a NPN bipolar junction transistor that is ideal for low power amplification. The 2N2222A is designed for low current, low power and is great for high switching speeds. Through research it has been stated that silicon transistors deliver a much richer sound, thus this ongoing mentality was implemented. For the diode clipping circuit the 1N4149 was implemented. They are listed to be high speed switching diodes with hermetically sealed leaded glass. They have a high switching speed of 4 ns at max.

For more of a bass effect, 1 uF capacitors will be used rather than a .1 uF which is the norm for most Fuzz Box boards. Two 500 K ohm potentiometers will be used for volume and fuzz control on the box. The board will be powered by 9V power supply. The Fuzz Box serves as a type of small pre-amplifier that distorts the signal by amplifying it before it gets to the actual amp. There are two knobs associated with this effect: fuzz and volume. The fuzz knob controls the level of distortion, while the volume controls the volume that is coming out of your amp. A deeper sound to the fuzz box would be appeal, being that most fuzz boxes look for a normal distortion sound. Thus the use of 1 uF caps was implemented into the design.

Figure 4.4.1 in the Design Summary, is the simulated layout for the Fuzz Box. This layout was created and tested through Multisim. The guitar signal comes in through the first two stages of the 2N2222A NPN transistors. After the second stage a 10k ohm potentiometer is implemented for the fuzz effect. By turning the resistance up and down, the user has the advantage of controlling how much fuzz effect is implemented into their guitar sound. Also following the second stage, is the diode clipping circuit that has been added for further effect. Lastly is the 100k ohm potentiometer that controls the volume. When the potentiometer is completely open (at 0 ohms), the fuzz volume is at its maximum. When the potentiometer is completely closed, there will essentially be no fuzz effect applied.

The input signal and output signal in Figure 3.3.1 were simulated in Multisim. The white signal is the output and the turquoise signal is the input. Using a function generator with a sine wave of 300 Hz (guitars range from 100 to 1000 Hz) the simulated input signal imitates the oscillating waves coming from a guitar. In this figure there is no fuzz added, and the signal is coming in almost clip free. This means there is very little distortion to the signal and it fairly matches that of the guitar. As you can see the two signals have similar amplitude and frequency. The output sound then is very clean and not distorted in any way.

40

Figure 3.3.1 Input and Output signal with No Effect

Figure 3.3.2 is the simulated signal with a full fuzz distortion added. This is caused when the fuzz potentiometer is turned all the way closed. That is the potentiometer is at the complete 10k ohms. You can see that the signal now looks more like a square wave. This is what a clipping signal looks like, as stated earlier in the research section. This is an almost ideal hard clipped signal. The rounding of the sine wave makes the sound distorted which is appealing to what the effect needs. Once again the output signal is the white wave, although it is about a fourth of the height of the input wave, the sound can be controlled by the volume knob to get the desired sound.

41

Figure 3.3.2 Simulated Fuzz Effect on the Input Signal.

Parts required for this board layout are as follows:

● Transistors: (2) 2N2222A NPN transistors ● Potentiometers: (1) 100k ohm, (1) 10 k ohm ● Capacitors: (1) 2.2 uF electrolytic, (1) 20 uF electrolytic, (1) 0.1 uF ceramic ● Resistors: (1) 1M, (1) 33k, (1) 330, (1) 100k, (1) 8.2k ● All resistors are ¼ Watt, 5% ● Jacks: (2) ¼ inch female stereo jack

3.3.2 Octave Up

As stated before, the octave up effect changes the octave of the guitar sound by “folding” the signal up, as to double the signals frequency. In order to double the frequency a full- wave rectifier is incorporated into the design. 1N4149 are small signal fast switching diodes that are used for the full-wave rectifier.

Figure 3.3.3 is a screenshot of the full-wave rectifier used within the circuit. This is a Graetz bridge rectifier. This will double the waveforms polarity thus double its frequency. This is the same as going up an octave within a scale range. The full-wave rectifier is the beneficial factor to the octave-up pedal.

42

The board has been designed to run on a 9V supply. A voltage divider is integrated in the beginning of the circuit to provide 4.5V to the NE5532AI operational amplifiers. The NE5532 are high performance operation amplifiers that amplify the signal coming in from the guitar. They are listed to be very low noise, high output-drive capability, high unity gain, and are rail to rail output. This is ideal for the amplification purposes of the octave up pedal.

Figure 3.3.3 Full-Wave Rectifier within Octave Up

Figure 3.3.4 is an example of how the NE5532 is used in the circuit. The operational amplifier on the right is the first amplifier used in the circuit. It is a non-inverting amplifier with a gain that is set by the 100 ohm potentiometer. Looking at R1 and C1 on the right hand circuit, these two passive components serve as a low pass filter for the signal being fed back into the operational amplifier. The output signal then goes into the op-amp on the left hand side of Figure 3.3.4. The left sided op-amp is an inverting one with a unity gain. It should be noted that for the purpose of diagraming the op-amp, the LM358 was used, but not within the design.

43

Figure 3.3.4 Use of the LM386 within Octave Up Circuit

Figure 4.4.2 in the Summary Design is the completed simulated circuit in Multisim. The voltage divider, along with the NE5532 operational amplifiers is used. And of course the full-wave rectifier is close to the output of the circuit. Within simulation, it was visible within the oscilloscope that a reduction in R4 (100k ohm potentiometer) gave a much preferred output. If the resistance was too high, the output would then oscillate at a much greater frequency than that of the input. Through trial and error, a 100 ohm potentiometer was better than the 100 k ohm.

Figure 3.3.5 is the simulated output produced onto the oscilloscope. The turquoise signal is the inputted sine wave and the white signal is the output. The input signal was set to 300 Hz and at an amplitude of 1 Vpp. It is visible that the sine wave has been folded “down” to give double the frequency. In order to give the preferred “folded up” appearance and inverting amplifier would be used to give it the sound desired. Although the outputted wave is smaller than the input, a change in the volume potentiometer will easily fix that.

44

Figure 3.3.5 Simulated Output of Octave Up Circuit

Parts required for this board layout are as follows:

● Op-Amps: (1) NE5532AI ● Potentiometers: (2) 100 k ohm ● Capacitors: (2) 100 uF electrolytic, (2) 10 uF electrolytic, (1) .01 uF film, (1) .47 uF ceramic ● Resistors: (4) 10k, (1) 220k, (1) 220, (1) 100k, (1) 33k, (1)4.7k ● All resistors are ¼ Watt, 5% ● Jacks: (2) ¼ inch female stereo jack

3.3.3 Orange Squeezer

The orange squeezer is a compression pedal that amplifies low signals and compresses the louder ones in order to equali e a guitar’s sound. 2 N-channel RF amplifiers (J310)

45

are used within the first input of the circuit. The use of 1of the N-channel transistor is to mimic the sound of a tube harmonics.

For the first JFET transistor (Q2), in Figure 3.3.6, the source is tied to the gate. This allows for that transistor to act as a clipping diode. A clipping diode makes the sine wave clip at its highest points. The second transistor serves as an RF amplifier. The 10k ohm potentiometer controls the bias of Q1. This allows for the user to control how much compression is applied to the notes. With the potentiometer completely open biasing would be at its greatest. With it closed, there would be no biasing/compression in the sound of the guitar.

Figure 3.3.6 RF Transistor J310 within Circuit Layout

The TL072ACD is a JFET input operational amplifier with low input bias, offset current and fast slew rate. Figure 3.3.7 shows the op-amp within its design of the circuit. It is used as a non-inverting amplifier with a gain of 22. This will amplify the signal coming out of the RF amplifiers before it goes through the diode. This is suitable for the orange squeezer for its low harmonic distortion and low noise. The final stage of the circuit is through the 10k ohm volume potentiometer. As in all other circuits, this is simply to control the volume of the effect on the guitar signal.

46

Figure 3.3.7 Layout of Op-Amp TL072ACD

The figure 3.3.8 is the simulated signal of the orange squeezer laid out in Multisim. It is very visible that the outputted signal is being clipped as was for mentioned in the purpose of the first FET transistor. It is a little difficult to express the properties of the orange squeezer through simulation, because there is no way to measure the strength of a signal through simulation. But it is very apparent that the incoming signal is about the same amplitude of the outgoing.

47

Figure 3.3.8 Simulated Output of the Orange Squeezer

Parts required for this board layout are as follows:

● Op-Amps: (1) TL072ACD ● Transistor: MMBFJ310 ● Potentiometers: (2) 10 k ohm ● Capacitors: (2) 4.7 nF ceramic, (4) 4.7 uF ceramic, (1) 2.2 nF film ● Resistors: (1) 1M, (1) 82k, (1) 2.4k, (3) 470k, (1) 390k, (1) 10k, (1) 220k, (1) 100k, (1) 1.5k ● All resistors are ¼ Watt, 5%

48

3.3.4 Tremolo

Traditional tremolo circuits include the uses of positive feedback loops that are filtered to lower frequencies. Those are then used with opto-isolators to produce the tremolo sound most people are fond of. The circuit designed uses a square wave from a LM555C timer.

The square wave that is produced from the LM555C can be controlled through its duty cycle. That is the purpose of the 100-ohm potentiometer R1 in Figure 3.3.9. The change in resistance affects the duty cycle of the square wave being produced. The square wave is then fed into the LM386, changing the amplification of the guitar. This is equivalent to turning the volume control up and down. R1 is representative of a depth knob on an actual pedal. This controls the rate of the pulses, rather the volume modulation.

Figure 3.3.9 555 Timer Layout within Tremolo

Figure 3.3.10 (reprinted with permission from Abel Domingues of Mojo Music) shows what the tremolo effect does to a signal. The first plot shows a signal that has already been altered. This is with R1 at it’s potentiometer at the halfway position, essentially at 50-ohms. The second plot shows the tremolo effect at its minimum depth, with R1 at the closed position. Finally is the bottom plot with R1 at completely open position, 100- ohms giving the 555 timer a very small duty cycle. This makes the “blerps” within the signal to happen more frequently.

49

Figure 3.3.10 Waveforms of Tremolo Effect on a Sine Wave

Once the square wave signal has been produced from the 555 timer, a PNP transistor (Q1) inverts the -9V signal. This makes the guitar signal oscillate. The output of the emitter then feeds the voltage Vs of the LM386.

The LM386 is a low distortion power amplifier used for low voltage consumption. Figure 3.3.11 represents the power amplifier within the tremolo circuit. It has an eternal gain of 20, but can have its value changed to 200. For the purpose of the tremolo, the external gain is set to 200. This is done by adding the 10 μF capacitor to pins 1 and 8. Once the signal has been fed in after the PNP transistor, the LM386 amplifies the signal and produces the desired effect.

50

Figure 3.3.11 LM386 Circuit within Tremolo

Figure 4.4.4 in the Design Summary is the circuit designed for the Tremolo effect pedal. R1 within the design controls the volume of the inputted signal. R2, as stated early, controls the depth and R3 controls the speed. 2 5mm LEDs are implemented into the design to represent when power has been applied to the circuit and to show the tapping rate of the signal.

Parts required for this board layout are as follows:

● Op-Amps: (1) LM386 ● Timer: (1) LM555C ● Transistor: (1) 2N3906 ● Potentiometers: (1) 100 k ohm, (1) 1M ohm ● Capacitors: (2) 10 uF electrolytic (1) 0.1 uF ceramic, (1) 220 uF electrolytic ● Resistors: (1) 470, (1) 4.7k ● All resistors are ¼ Watt, 5%

3.3.5 Big Muff

The Big Muff was the most intricate circuit to be applied to the design of the guitar amplifier. With four stages of NPN transistors along with varies values of resistors and capacitors, the Big Muff raised a challenge.

The first stage uses a 2N5191G BJT NPN power-switching transistor. For the purpose of simulation a BC109BP was used, because unfortunately Multisim did not have the 2N5191G. The 2N5191G has a collector-emitter voltage of 60 volts and a collector-base voltage of 60 volts. The collector current is up to 4 amps making it ideal for the Big Muff layout. This stage is regarded as the clean booster stage. The signal is amplified within the 2N5191G and then is fed to the 100-kiloohms potentiometer. This controls the sustain of the signal. Sustain in distortion pedals works like a compressor pedal does. The difference is that, it does not compress the higher signals. It does amplifier lower signals, but does not keep them from clipping. It increases the signal amplitude very large, and allows for the signal to be cut off. The change in sustain changes the amount of “cutting off” the transistor does to the incoming signal.

51

The second and third stages are equivalent and are considered the clipping stages. The BC546CTA is used for the NPN switching transistors. The collector-base voltage is rated at 80 volts and the collector-emitter voltage at 65 volts. The collector current is rated at 100 mA. The second stage feeds the third stage, which then pour into the second potentiometer at 100-kiloohms. Both circuits have 1N4149 diodes for the purpose of clipping the circuit further. The potentiometer after stage three is the tone setting. The tone weakens the signal going into the fourth stage to give it the gritty and dirty tone distortion effects wish to achieve.

The final stage is the tone recovery stage. The last stage is used with the 2N5088BU NPN high gain amplifier transistor. It has a collector-emitter voltage of 30 volts and a collector-base voltage of 30 volts as well. Its continuous collector current is as 100 mA. The final stage helps recover some of the audio gain that was lost, at approximately 8 dB. Finally the signal is lead into the third potentiometer that controls the volume output of the distorted signal. Figure 3.3.13 below is the circuit of the Big Muff designed.

Figure 3.3.12 below shows the signal from the Big Muff with the sustain knob at half its position. The input is the turquoise signal at 300 Hz and 1 Vpp. The output is the white signal. It is visible that the white signal is being distorted. The negative portion of the wave is being removed while the positive part of the wave is being clipped.

Figure 3.3.12 Simulated Output with Half Sustain

52

This next simulation shows the signal with the tone knob is halfway position it is included below in Figure 3.3.13. It is visible that the signal is being distorted. There are small impulses like pulses where the original signal crosses the zero axes on its positive slope. The outputted signal is being clipped, making it look more as a square wave.

Figure 3.3.13 Simulated Output with Zero Sustain

Lastly Figure 3.3.14 is the Big Muff with full sustain and zero tone. The signal is more amplified and is still having its negative portions “cut-off”. The tone is no longer affecting the signal, thus not compressing it to a smaller signal as before. The signal is still in phase with the original signal, which is what is desired for the distortion effect of the Big Muff.

53

Figure 3.3.14 Simulated Output with Full Sustain

Parts required for this board layout are as follows:

● Transistor: (1) 2N5191G, (2) BC546CTA, (1) 2N5088BU ● Potentiometers: (3) 100k ohm ● Capacitors: (4) 1 uF electrolytic, (3) 500 pF ceramic, (4) .68 uF poly film, (1) .004 uF ceramic, (1) .01 uF ceramic ● Resistors: (1) 36k, (4) 100k, (3) 120, (4) 470k, (2) 39k, (1) 1k, (3) 10k, (1) 27k, (1) 2.2k ● All resistors are ¼ Watt, 5%

3.3.6 Distortion with Stutter Effect

A distortion effect wouldn’t be complete without a clipping circuit within the layout of the board. For the clipping of the signal the LM386 was chosen as the operational amplifier. The LM386 is an audio amplifier power amp. It’s a low voltage audio amplifier with an internal gain of 20. Although with external resistors the gain can range from 20-200. It has a voltage supply range of 4V-12V, which is perfect for the 9V design I have implemented for all board layouts. It has a low quiescent current of 4 mA with a low distortion of 0.2%.

54

Figure 3.3.14 illustrates the LM386 within the distortion circuit. On the LM386 pins 1 and 8 control the gain. If both pins are left open the eternal 1.35 kΩ resistor sets the gain to 20. For the purpose of this layout a 10 μF capacitor is placed to reach a gain of 200. Pins 2 and 4 are both tied to ground, and pin 3 is tied to the input signal from the guitar. Pin 6 is tied to the output of the NE555 timer (described below) and pin 5, is the output signal. With a gain of 200, this allows for the signal to be amplified through the IC and allows it to be clipped, giving it its distorted sound.

Figure 3.3.15 Op-Amp LM386 within the Circuit Layout

The NE555 timer is a precision timing oscillator that produces a square wave with a programmable duty cycle. The timer is compatible with voltage supplies between 5 to 15 V. Making it ideal for a 9 volt supply being sourced to the board. It is included below in Figure 3.3.16.

55

Figure 3.3.17 555 Timer layout with Distortion with Stutter

The duty cycle of the NE555 is controlled by the two resistors labeled RA and RB. This layout causes the capacitor to charge and discharged through the threshold voltage (0.67*Vcc) and the trigger-voltage level (0.33*Vcc). The supply voltage then affects the frequency and duty cycle of the square wave. For the purpose of the distortion pedal, a 470 Ω is used for RA and 10 kΩ potentiometer for RB. With the 10 kΩ potentiometer in place for RB this allows the user to change the duty cycle, thus changing the stutter effect. The remaining circuit for the NE555 is fairly the same except there is no capacitor between pins 8 and 4, and no resistor between 8 and 3.

A toggle switch is implemented that allows the user to choose whether or not they want to use the stutter effect in their sound. A 5mm LED is incorporated to allow the user to see when the effect is powered on. For the use of the gain, a 100 kΩ is placed directly in line with the input signal, which is then fed into the input line of the LM386. There is also another 100 kΩ potentiometer placed with the output signal to control the output volume of the guitar. The rest of the parts are filled with 25 V capacitors and resistors onto the board.

Parts required for this board layout are as follows:

● Op-Amp:(1) LM386 ● Timer: (1)LM555C ● Potentiometers: (2) 100k ohm, (1) 10 k ohm ● Capacitors: (3) 10 uF electrolytic, (1) 220 uF electrolytic, (1) 0.1 uF ceramic ● Resistors: (1) 470, (1) 1K ● All resistors are ¼ Watt, 5%

3.4 Digital Effects

3.4.1 No effect

The first challenge of designing the effect algorithms was to simply read the guitar input and hear it played as you would hear it naturally, just amplified. The output should contain only the inputted data and the lowest noise that can be achieved.

After the sampling rate, gain, phase locked loop, and CODEC have been initialized the algorithm follows the next set of steps. There is a CODEC read function that is called. It takes in a two 16-bit values that are a digital representation of the analog signal that has been covered by the ADCs. Main looks for the address of the two inputs and assigns a left and right value to the respective inputs. This stereo input is converted into on a single mono value by averaging the two inputs and storing them in two variables called left and right. Then these digital values are converted back to the analog signal and passed down the I2C bus and through the output jack on the TMS320C5515 eZdsp.

56

Figure 3.4.1 is a block diagram representation of the process.

Figure 3.4.1 Flow if Guitar Input

This waveform included below in Figure 3.4.2 is the waveform of the input. It is the low E played with no fingers pressing on the guitar’s fret board. Figure 2.5.4 is the analog signal through the operational amplifier. The preceding wave form has been processed by the TMS320C5515 eZdsp and through an additional operational amplifier stage, which provided gain for the output. The gain circuit was in the same configuration as Figure 2.5.1. The gain was adjusted from both the software and the operational amplifier. The output voltage was made high enough to distinguish between noise but lower that the voltage required to be clipped by the operational amplifier. Upon viewing this waveform which is simply the guitars output with no processing it looks just as smooth waveform in comparison. This is what was expected because of the specifications of the eZdsp board.

57

Figure 3.4.2 E string Waveform

3.4.2 Reverberation

Reverberation is a guitar effect that adds fullness to the sound. Reverb is implemented by taking a time shifted output and playing it over the newest input. This effect works well with the clean sounding effects.

One parameter that can change the effect is the sampling frequency. When the sampling frequency is set to 48kHz the effect is very subtle. The effect just fills the sound slightly and it creates a more interesting sound. When the sampling frequency is set to 24kHz the effect is much more apparent. If you were to play a chord that you abruptly stop the strings from vibrating you will hear the output delayed by a short amount and repeating while its amplitude decreases until it fades away. The lowest sampling frequency that is set is 12kHz. This creates an echo like effect and is an extremely exaggerated reverberation. The difference between it an echo is that the output is being time shifted then played over the input not a saved version of the input.

The next parameter that can be changed is the depth. This is the value that is multiplied with the output that is stored in the reverberation array. The first value it was set to was 30000. The effect at 30000 was mild and did not alter the original sound drastically. When the value was decreased to 15000 the parameter’s effect is easily noticed. When 58

the depth is larger the delay is larger so you will hear more of the reverberation of the output longer. When it is decreased the reverberation duration is shorter.

The last parameter that can be changed is “N” the si e of the reverberation array. This effect is similar to when we increase and decrease the sampling frequency. The first value of array that was used was 4800. This gives a short reverberation of the output and is relatively mild. If the reverberation array is expanded to make N = 9600 then the reverberation is very apparent because more repetitions of the output can be heard. When the array size is set to 9600 the reverberation sounds similar to the echo.

Below in Figure 3.4.3 a simplified version the digital network of the reverberation effect. The time shifted output is added to the input.

Figure 3.4.3 Reverberation Network

The Figure 3.4.4 below is the frequency response of the reverberation effect. It was simulated in Matlab. The si e of the echo array “N” value in the simulation was set to 1000. The sampling frequency was simulated to be 24kHz.

Figure 3.4.4 Reverberation Frequency Response

59

Below in Figure 3.4.5 is a simulation of the poles and zeros plot of the reverberation. The same parameters were used in the Figure 3.4.4 as 3.4.5.

Figure 3.4.5 Reverberation Poles and Zeros

3.4.3 Fuzz

Fuzz is an effect that distorts the input signal. In most cases this is not desired, but it is when creating certain types of music. It can fill the sound and complements some forms of music.

This effect looks at the input value and based off its value will scale it or set it to a fixed value. There are three values set in this code. The first is “X” this value is an integer multiple of the larger number ‘Y’ the ratio of the X to Y is the third ‘Z’. In the code there is a check of the input value. If the input value is between +X and -X the input value is multiplied by Z and this becomes the newest output. If the value is not between +X and - X but it is a negative value then the output is set to -Y. If the value is not within the bounds of +X and -X, and greater than zero it is set to +Y. There is a check to make sure that the output does not exceed a certain value because this could cause an overflow and cause the program to crash. By setting the appropriate value of the output to +Y or -Y insures that the output will be clipped.

60

The Fu ’s effect can be controlled by decreasing the X value. This allows for more clipping of the input signal. The X value that produces a nice mid-level of Fuzz is 250. For a much more Fuzzy sound decreasing it to 100 works well. Finally for a light Fuzz 500 is recommended.

Below in Figure 3.4.6 simplified Fuzz network that shows that the input is passed through the effective clipping block to the output, where the degree of clipping is dictated by parameters in the algorithm.

Figure 3.4.6 Fuzz Network

The following Figure 2.4.7 was simulated in Matlab’s Simulink. The circuit shows the simulated quantized sine wave input and the saturation that is brought on by the clipping in software.

Figure 3.4.7 Fuzz Simulation Circuit

Figure 3.4.8 shows the simulated guitar input. It is a sine wave at 100Hz that has been sampled at 48kHz.

61

Figure 3.4.8 Simulated Guitar Input

Below in Figure 2.4.9 is the output after the saturation block in the simulation. The noted feature of the simulation is the clipping in the picture.

Figure 3.4.9 Simulated Fuzz Output

The below in 3.4.10 is the waveform that the Fuzz effect creates after being processed by the eZdsp. The waveform matches the simulated waveform.

62

Figure 3.4.10 Fuzz Output Waveform 3.4.4 Echo

Echo is an effect that takes the guitars input, time shifts it, and plays it over the current input. This creates more layers to the sound. It can even give the impression that two players are playing when in fact there is only one.

When the sampling frequency is changed it determines how far apart each echo will take place. When the sampling frequency is set to 48kHz the echo sounds similar to a mild reverberation effect. But when decrease the echo becomes more delayed. The two other values that can be set are 24kHz and 12kHz.

The size of the echo array “N” also controls how far apart the echo is from the input. When N is set to 12000 the echo can be heard clearly. When it is decreased to 6000 the echo is very close to the input in time and sounds similar to a mild reverberation effect.

Below in figure 3.4.11 a simplified block diagram of the echo effect. Where the input is summed to the time shifted version to create the effect.

63

Figure 3.4.11 Echo Network

Below is Figure 3.4.12 is the frequency response of the echo effect. It was simulated in Matlab. The si e of the echo array “N” value in the simulation was set to 1000. The sampling frequency was simulated to be 24kHz.

Figure 3.4.12 Frequency Response Echo

64

Below in Figure 3.4.13 is the poles and zeros plot of the echo effect. The same parameters were used for Figure 3.4.12 as the Frequency response plot in 3.4.11.

Figure 3.4.13 Poles and Zeros Plot for Echo Effect 3.4.5 Tin Can

Tin can adds a highly treble sounding color to the guitar’s input. It works well as a clean effect. The effect works first by decreasing and increasing the sampling rate of the input depending where you are in the algorithm, then adding it to a treble component of the input. Decreasing the sampling rate consists of skipping sample values. For example we fix the main sampling rate at 48kHz, but we only used 1 out of 12 samples. This would being the sampling rate down from 48kHz to 4kHz. The only problem is that once you slow down the sampling rate you have to check to make sure you do not violate Nyquest criteria, which states you need to sample at twice the highest frequency. In the 4kHz case you would need to a low pass filter of 2kHz bandwidth.

Increasing the sampling rate from our example value 4kHz to 48kHz involves two parts. You first add zeros in the places that you were skipping, then you apply a smoothing filter to make the output more coherent.

Below in Figure 3.4.12 is a block diagram representation of the tin can effect. One observation that can be made is that the guitar input takes two paths, the up and down

65

sample rate and the high pass filter. The reason why the effect is called its name is because of how prevalent the treble sound is. The output always contains the high pass treble element, but as the code flowchart shows, the path of up and down sampling gets added depending on the parameters of the algorithm.

Figure 3.4.12 Block Diagram Tin Can

The following figure 3.4.13 is output waveform of the tin can effect. The notable characteristic of this waveform is how jagged it is. This is attributed to the filtering that emphasized the treble element.

Figure 2.4.13 Wave Form Tin Can

66

3.4.6 Phase

This effect adds a nice sound to the guitars input and is considered a clean effect. It works by altering the sampling rate like the tin can effect, but does so with delay that varies. The alteration of the sampling rate caused by decreasing and increasing the sample rate. Decreasing the rate is implemented by skipping samples and the speeded up is by filling in those gaps with zeros then smoothing out the output.

Unlike the tin can effect, phase does not switch between the accelerated and decelerated clock rate to create the auto effect. The effect derives from the initial step of the algorithm. That is the after clock rate is decelerated, the time delay of when the input is stored in its appropriate array and is varied. In the algorithm there are two counters. These counters are what dictate the time delay. After the input has been processed, the clock rate is returned to 48kHz by filling the places in the array where we skipped with zeros and smoothing the output but filtering it.

To vary the degree to which phase is implemented the parameter called interval can be changed. The interval parameter is compared to the in the at the start of the algorithm to the primary counter. If we make the interval size large this means that we exaggerate the time that the individual time delays are varied. Meaning that you get a more dragged out effect. Conversely if we decrease the interval parameter we will switch between the different time delays at a much faster pace and the effect loses the dragged out sounds and begins to have the characteristic sound of something like a mild reverberation.

Below in Figure 3.4.14 is a simplified block diagram of the phase effect.

Figure 3.4.14 Phase Block Diagram

67

The figure below is waveform of the phase effect. It should be noted that it is a smooth waveform it reflects the sound that is outputted.

Figure 3.4.15 Phase Waveform

3.4.7 Robot

This effect is simple yet creates a very interesting sound. It takes the guitar input and modulates it with a sine wave. To modify this effect the frequency of the sine wave can be increased or decreased. This effect is more in line with digital synthesizers than a traditional guitar effect. The modulation causes one note to have a much fuller sound.

Below in figure 3.4.16 is a block diagram that illustrates the effect. The guitar input is modulated by a sine wave.

Figure 3.4.16 Robot Network

68

The following Figure 3.4.17 is the waveform that was collected form the eZdsp implemented after implementing robot effect.

Figure 2.4.17 Robot Waveform

3.4.8 Fuzzy Tube

The fuzzy tube is implemented by modulating the input with a modified sine wave. Effect was derived from the Robot effect. The sine wave was modified by determining two cases. The first when the sine waves amplitude was positive it was fixed at a specified positive value. The next case was when the sine amplitude was negative; it was fixed at a constant negative value. This turned the sine wave from a smooth curve to a square wave that alternates between negative and positive amplitudes. Below in Figure 3.4.18 is the simplified block diagram of the effect.

Figure 3.4.18 Fuzzy Tube Waveform

69

To confirm that the modified sine wave theoretical shape the guitar the following simulation was included. The goal of the simulation was to prove that the logic used to modify the sine wave was correct and to have an ideal waveform for comparison. The guitar’s original input has multiple peaks the waveform below accounts for these peaks.

The following Figure 3.4.19 was developed using Matlab’s Simulink DSP simulation software. The inputted sine wave was at 500 Hz and was modulated by the square pulse wave generated by the clock. These two waveforms were multiplied then their output was measured.

Figure 3.4.19 Fuzzy Tube Circuit Simulation

The following Figure 3.4.20 is the simulated result of the circuit above.

70

Figure 3.4.20 Fuzzy Tube Simulation Results

Below in Figure 3.4.21 is the actual waveform of the fuzzy tube effect. This waveform matches expected result that was shown above. Although the wave has some expected curvature, it can be observed that it has be modulated by the square wave.

Figure 3.4.21 Fuzzy Tube Waveform

71

The final design of fuzzy tube was modulating the guitars input with the fuzz effect. The outcome is the same as explained above. The input is modulated with a clipped sine wave.

3.5 Graphical User Interface

3.5.1 Design Summary

The user interface has been chosen to be constructed in Netbeans IDE, which is growing in usability, stability, and features, making it ideal for a multi-platform, multi-language IDE. It would allow the user to select effect files, modify, and update the DSP using a Windows 7 Laptop. These effects would be provided on the main GUI window. The allowed modifications of the effects would be user entered integer values that correspond to the selected effect. Only one effect would be modified during each interval, meaning that the design would not incorporate the mixing of two or more effects at one time. Then, the DSP would be updated with the newly selected and modified effects. This was done via the PC serial connection to the DSP. The DSP was first initialized to set up interrupts, timers, and ports, as well as other DSP options. Then, the BT module was initiali ed. After changes were made to an effect, the “Update Current Parameters” button would tell the DSP boot loader to execute. The program would then wait for an interrupt generated by the user input.

After downloading and installing the free, open-source software (which took very little time and thought), it was time to begin constructing the GUI. The basic communication flow between the GUI and the DSP would be realized by the following main points. First, the user would select one of the effects from the main GUI upon which the corresponding window would pop up allowing the manipulation of one to three parameters (the user will indicate when they are ready to update the settings). Next, the parameter(s) would be serialized so that the data stream would be in binary format and be processed with the .bin file associated with the effect parameter. The following step required writing the data to the USB port in order to access and update the DSP memory. As long as the user did not exit the main GUI, they could exit corresponding effect windows and continuously update different effects and parameters. The following Figure 3.5.1 represents a flow diagram of the processes described.

72

Figure 3.5.1 Simplified Flow Diagram for GUI/DSP Communication

After launching Netbeans IDE 7.3.1, a new project (Java Application) was specified, named DigitalEffectsProject, and the “create main class” button was unchecked. Then, the source packages button in the Project window was expanded, and after right-clicking the “default package”, a new JFrame form was selected. Next, a new Java Package was created called “AppPackage” and the “default package” was deleted (in order to replace the package). The JFrame was specified to the AbsoluteLayout, GenerateResizeCode(Center) and correspondingly a JLabel was place to cover the same dimensions as the JFrame. Creating the menu bar was simple which required importing the javax.swing.JMenuBar, JMenuItem, SwingUtilities and such. The events such as clicking on the menu items (File, About) required importing java.awt.event.KeyEvent, ActionEvent, and ActionListener. The GUI was intended to manipulate certain parameters based on the following digital effects: reverberation, fuzz, echo, tin can, phase, and robot, where the Figure 3.5.2 illustrates the 6 clickable buttons which pertain to their own separate JFrames (essentially there was one main GUI, and 6 sub-GUI’s).

73

Figure 3.5.2 Main Menu GUI

For the reverberation effect, the user would be able to adjust three parameters: the sampling frequency, the depth, and the si e of the reverberation array “N”. The sampling frequency was intended to vary within the 48 – 12 KHz range which would allow the user a subtle to exaggerated reverb. Regulating the depth parameter from 30000 to 15000 would increase or decrease the reverb duration. Modifying the si e of the array “N” would allow more repetitions of the output to be heard and could range anywhere from 50 – 9600. The screen shot of the window is included in Figure 3.5.3

Figure 3.5.3 Reverb GUI

74

For the fu effect, the user would be able to control the “X” value, which upon decreasing, allows for more clipping of the input signal. A screen shot of the Fuzz effect user screen is illustrated below in Figure 3.5.4.

Figure 3.5.4 Fuzz GUI

For the echo effect, the user would be able to manipulate parameters similarly as discussed in the reverb effect, including the sampling frequency and the size of the echo array, “N”. The “N” parameter would allow the range of 50 – 12000. Figure 3.5.5 shown below is the user interface for the echo effect.

75

Figure 3.5.5 Echo GUI

Figure 3.5.6 is a screen shot of the user interface for the Tin Can effect. The tin can effect has a sampling rate of the input, which can be decreased and increased.

Figure 3.5.6 Tin Can GUI

For the phase effect, the interval parameter could be altered to exaggerate the time between the time delays, and the effect would be more drawn out. Similarly the interval parameter could be decreased, and the effect would be comparable to the reverberation effect. Figure 3.5.7 illustrates how the user is able to control the time delays to get the desired phase effect.

76

Figure 3.5.7 Phase GUI

The robot effect is altered by the sine wave frequency, which could be increased or decreased. This would modulate the guitar’s input. Figure 3.5.8 is a screen shot of the frequency change input.

Figure 3.5.8 Robot GUI

The fuzzy tube effect was similar to the robot effect in that its parameter included the manipulation of the sine wave frequency by typing in any integer. Figure 3.5.9 is the user interface for the fuzzy tube.

77

Figure 3.5.9 Fuzzy Tube GUI

The final design for the user interface was combined into one main window, which was simple, clear, and concise. It allowed the user access to several effects, as listed below in Figure 3.5.10, as well as no sound (mute) and no effect (none). It was changed from the previous design which was a great decision. The previous design was not complete and therefore the final design shows the compatibility for the digital effects.

Figure 3.5.10 Final Design for GUI

78

3.5.2 DSP Bootloader

The TMS320C5515 Bootloader was investigated first since it would be responsible for transferring user code from an external source and writing it into the on-chip DSP RAM. The C5515 bootloader checks for a valid boot-image on each supported boot device in the corresponding order: NOR Flash, NAND Flash, 16-bit SPI EEPROM, I2C EEPROM, and MMC/SD. The first valid boot image to be received by the bootloader will be used to load and execute user code. Although the C5515 bootloader supports both secure and unsecure boot image formats, our design focused on the unsecure boot image format.

The COFF file mentioned above is created from the assembler and linker. It follows a specific format which makes modular programming easier. The smallest unit of a .out file is called a section. These sections, which are separate and distinct from each other, are known as blocks of code or data which occupy space in the memory. The assembler and the linker provide the medium that allows the creation and manipulation of sections. The three default sections contained in the COFF object file are as follows:

.text section (contains executable code) .data section (contains initialized data) .bss section (reserves space for uninitialized variables)

Two basic types of sections, initialized sections and uninitialized sections can be used like the three sections above. The assembler allows linking various parts of code and data to the appropriate sections. Then, the assembler builds and creates an object file. The linker provides a function (known as allocation) which relocates sections into the target memory map. Using sections allows using target memory more efficiently, since all sections are independent. This is applicable for moving any section into any block of target memory. The Figure 3.5.10 below illustrates the relationship between sections in an object file and a basic target memory.

Figure 3.5.10 Object File and Memory Relationship

Uninitialized sections reserve space in processor memory. The sections have no content any are only placeholders. This allows a program to create and store variables which are allocated in the RAM. To build these data areas, the assembler must use the directives

79

.bss and .usect. These two directives do not tell the assembler to stop a current section and begin with the indicated section; rather they escape temporarily from the current section.

Initialized sections contain the code and data of a program which are stored in the object file and place in processor memory when the program is loaded. The .text, .data, and .sect directives tell the assembler to stop the current section and begin the indicated section. The assembler maintains a separate program counter for each section, which represents the current address within a section of code or data. The starting value of the section program counter must only be specified once which can be specified through the value parameter, if present, or when the section is first encountered. The default starting value is O. As the section fills up, the section program counter increments the corresponding section.

Named sections can be created using the .usect and .sect directives, where the .usect creates sections used like the .bss section and the .sect creates named sections with relocatable addresses.

The linker is responsible for combining input sections to create output sections in an executable COFF output module, known as the Sections directive, while it also chooses memory addresses for those output sections, known as the Memory directive. This executable object file has the same format as object files that are used as linker input, but the sections in an executable object file are combined and relocated in order to be loaded directly into target memory. The methods used for loading a program are described next.

Three ways to create a boot image were observed. The first was to create a boot image using the hex conversion utility (hex55 utility). The hex conversion utility is able to produce output file formats such as ASCII-Hex (supporting 12-bit addresses), Texas Instruments SDSMAC (supporting 16-bit addresses) as well as others. Since most EPROM programmers do not accept COFF object files as input, the hex utility can be used to convert the object file into a standard format. This utility supports the on-chip boot loader built into the C5515. This utility can be invoked by specifying options and filenames on the command line, or can be specified in a command file. The command file is useful since it contains options and filenames, ROMS directive, and the Sections directive (which can be used to identify specific sections which will be initialized by the on-chip bot loader.

The next method utilized a tool included in CCS (v3.0 or greater) called the Object File Display (OFD) utility based on XML and Perl. The last method introduced the DSP Boot Assist Tool which allowed a more straightforward and flexible way to generate a boot image from a common object file format (COFF) or .out file. The Windows DSP Boot Assist Tool was referenced by TI based on the C6000 DSP, but because the COFF format has not changed, it was realistic to use this tool for the C5515 DSP. Creating the boot image with this tool was simplified to first specifying the .out file for processing, after which the generated boot table could be saved as a C header file or a binary file. Depending on what the boot image table was saved as (header or binary file), the file was

80

included into the host project and corresponding code to load the boot image through the host port interface into the DSP memory was added. Then, the DSP was notified by the HPIC to begin running from the predefined address.

3.5.3 Accessing eZdsp USBSTK in Java

Research had provided the conclusion that Java no longer supports its own open-source software to access USB devices. There was another option considered called RXTX, however their current version does not support Windows 7. Finally after plenty of searching, the most recent option was to implement a package which includes the class LibusbJava. The Java wrapper (Java libusb) is for the libusb-win32 USB library, which allowed user applications to access any USB device on Windows, without writing any line of kernel driver code. This package, called ch.ntb.usb, includes several main classes; however, the few that are relative to the design are Classes Device, LibusbJava, and USB. The java.io package was featured in the code as well since it provides for system input and output through data streams, serialization and the file system. The following Figure 3.5.11 describes the Package ch.ntb.usb hierarchy, which was referenced to accomplish a USB serial port through Java, where the superclass java.lang.Throwable (implements java.io.serializable) represents errors and exceptions.

Figure 3.5.11 Package ch.ntb.usb Hierarchy Needed for USB Access in Java

3.6 Bluetooth

3.6.1 Module Description

81

The RN-42 evaluation boards, as illustrated in Figure 3.6.1, components included the RN-42 module, two connection status LED’s, a voltage regulator (TC1185), a USB to serial UART interface (FT232RQ), four configuration switches, two general purpose input/output headers, PCB trace antenna, and a serial peripheral interface bus master. The evaluation board was purchased to ensure proper configuration and understanding in order to apply correctly to the final design.

Figure 3.6.1 Purchased Bluetooth Evaluation Board from MicrochipDIRECT.com

The RN-42 module itself provides a class 2 BT radio which supports multiple interface protocols while delivering up to 3Mbps data rate at a maximum of 20m. It supports baud rate speeds from 1,200 - 921Kbps, operates in the 2.4GHz band, provides error correction, auto-discovery/pairing, auto-connect master, and can configure via the local UART or RF. In addition to the operational mode (on-board stack running on the module) there is the Host Controller Interface mode, where the on-board stack is bypassed and the module runs in a state which accesses the BT baseband capabilities. The HCI provides a command reference interface to the baseband controller and the link manager, providing access to the software status and control registers. The HCI data rates are 1.5Mbps sustained, and 3.0Mbps burst in HCI mode. The module contains a header, crystal, flash memory, CSR on-board stack, RF switch, and a balun (to achieve compatibility between two systems). The Figure 3.6.2 illustrates a simple block diagram, where the header specifies GPIO4, GPIO5, and GPIO (Factory Defaults, Status, and Set BT master) respectively.

82

Figure 3.6.2 Reprinted with Permission from Microchip Technology

3.6.2 On-board Bluetooth Protocol Stack

The RN-42 module includes embedded Bluetooth stack profiles from CSR Synergy supporting HID, SPP, DUN-DCE, DUN-DTE, the multi-profile SPP and DUN-DCE, GAP, SDP, RFCOMM, L2CAP, and APL. The device uses the HID profile to replace the USB cable. By default, the SPP is enabled. The command S~, 6 (while in CMD mode) enables HID protocol. The command S~,0 following a reboot switches back to SPP protocol.

The serial port profile defines the protocols used by devices using BT for serial cable emulation. There are two roles defined for this profile known as the Initiator and the Acceptor. This profile achieves setting up virtual serial ports on two devices connecting with BT. It supports one-slot packets only, meaning that data rates up to 128Kbps can be used. Only point-to-point configurations can take place, meaning that only one connection can be dealt with at a time. Link establishment is done through the Initiator. RFCOMM is used to transport the user data, control signals, and configuration commands. The SPP is built upon the generic access profile (GAP). The GAP states the requirements on names, values, and other such parameters on the user interface level. It describes the general procedures that may be used to create connections to other BT devices.

3.6.3 Hardware Setup/Connection via USB

Access to the GUI was performed on a laptop running Windows 7, which integrated an external Bluetooth dongle attached to a USB port. The stack included in this operating system does not support Bluetooth radio connections over PCI, I2C, Serial, Secure Digital I/O, CompactFlash, or PC Card interfaces. Since Windows 7 supports Bluetooth v2.1 +

83

EDR, the Kinivo BTD-300 Bluetooth 3.0 USB adapter was purchased (since newer Bluetooth specifications are compatible with older versions.)

The first step was to establish a USB connection via the BT evaluation board. This allowed a local configuration for the Bluetooth module over a USB serial port. Put simply, the Bluetooth 3.0 USB dongle was inserted into a USB port on a Windows 7 laptop and the corresponding BT software was installed. Next, the evaluation board was connected via the provided mini USB cable. The default operating mode for this device, called slave mode, allowed other BT devices to discover and connect to it. Therefore, the PC acted as the master and the Windows BT device manager handled the discovery, pairing, and connecting processes. After attaching to the laptop, the discovery of the module was indicated by the blinking (once per second) green LED. Similarly, the device manager assigned the USB serial port (COM4) to the BT device (RNBT-38A1), where 38A1 indicated the last four bits of the Bluetooth evaluation board MAC address. After double-clicking on this device, the pin code was verified which enabled pairing with the computer. The pairing process was finalized through the device manager which installed two standard serial over bluetooth links known as COM ports (refer to figure [ ] ). The COM5 serial port was labeled outgoing which was used when the BT module was in slave mode. There was also an incoming port created labeled COM6 which was used when the BT module was in master mode. The PC’s Bluetooth Radio MAC address, known as the host remote address, was identified under the USB adapter device properties shown in Figure 3.6.3. Next, the computer needed a reboot so that there wouldn’t be complications communicating with the terminal emulator.

Figure 3.6.3 Assigned COM Ports Located in Bluetooth Settings (top); MAC Address for Bluetooth USB Dongle (bottom)

84

Upon launching TeraTerm (Window’s terminal emulator), a default window to create a new connection was displayed in which the USB serial port COM4 was selected. The serial port settings under the setup tab were altered to reflect the module’s default settings (115,200 Kbps baud rate, 8 bit data rate, no parity, 1 stop bit, no flow control). The device was blinking a fast green LED to show it was connected. It was necessary to change from the default (data mode) to command mode which was done by entering the ASCII bytes $$$; correspondingly the module returned the string CMD indicating a connection. This was verified by typing the command x , which summarizes current module settings. Figure 3.6.4 illustrates some (but not all) of the set, action, and display commands used in the configuration process. The get help feature, which was simply displayed after typing H while in CMD mode, revealed a thorough list allowing for more manipulation of advanced settings. While still in CMD mode, the ASCII byte + was entered to enforce the echo effect while typing. Next, D was entered to show basic settings such as the Bluetooth Address, name, mode, and remote address. Similarly, E was then entered for advanced settings such as server name and configuration timer. Another command (SO,%) was performed to enable the status message for connect/disconnect/reboot conditions (please refer to Figure 3.6.4). Immediately after enabling the set command above, a reboot was required (using the R, command). Since the BT module remote address was not specified, the SR, was implemented and the string AOK was returned to verify the command. After establishing the COM4 connection, it needed to be configured to communicate with the outgoing Bluetooth link (COM5).

85

Figure 3.6.4 Basic Reference Tables for Set, Action, and Display Commands

A new connection for the serial port COM5 was selected and its serial port settings were changed to reflect the module’s default settings (as declared above). Soon after the settings were changed, the COM4 terminal was able to connect with the COM5 terminal and data was transmitted/received by both terminals. The K, command was used to disconnect the device from the current connection.

Next, the BT board was connected to the eZdsp board using the UART interface (where CTS was held low or tied to RTS) with the corresponding pin numbers off of the Bluetooth expansion connectors P1 and P2 for UART_TX, UART_RX, VDD, and GND. The 4th dipswitch was held low in order to utilize this feature of the eZdsp (held low). Since only one device can have control of the UART line at a time, the pairing between the devices were such that TX of the BT device was tied to RX of the eZdsp device (as shown in figure 3.6.5) which also summarizes the pin pairing for the two evaluation boards.

86

Figure 3.6.5 Pin Connections

3.7 Power Supply

3.7.1 Block Diagram

The design of the power supply for the guitar amplifier is implemented within the block diagram of Figure 3.7.1.

Figure 3.7.1 Power Supply Block Diagram

Transformer

Due to the availability of power transformers in the market, this power supply will be using single primary and single secondary transformer. Since there are dual voltages for the power and pre amplifiers, a secondary winding with center tap transformer will be appropriate for this project. The highest DC supply needed is 30V. So a transformer with the secondary output of 24V AC is required. Before a transformer is acquired, it’s

VA rating needs to be decided. In a practical way, the VA rating of a power transformer should be twice of the output power of the amplifier. This guitar amplifier has 50 Watts of output power, so a power transformer with VA rating of 100 VA is needed.

Rectifier

87

A full wave bridge rectifier is nothing more than a package with four diodes inside. Using a bridge rectifier instead of individual diodes provides simplicity and efficiency. Due to the power supply output requirements, the bridge rectifier needs to have at least 100V/4A peak reverse voltage and forward current.

Filter

A filter is to filter out the AC ripple in the full rectified waveform. The RC filter is a capacitor connected in parallel with the load resistor. The larger the RC time constant is, the more smooth output power will be. For a full wave rectifier circuit, the ripple percentage is calculated as r = 1 / (2 * f * R * C). f is the 60 Hz, R is the load resistance seen by the transformer. It is the paralleled combination of four voltage regulators and power amplifier input resistance. For the purpose of testing, a load resistance of 1 k was selected. For a 50 Watts amplifier, the filter capacitor usually is 4700uF. Thus, the ripple percentage is approximately:

r = 1 / (2 * 60 * 1000 * 0.0047) = 0.2 %

Regulator

There are four voltage regulators in the power supply. The outputs of these regulator are used by DSP in digital effect module, analog effect module, and pre amplifier. Since these modules are low voltages and low currents, regulated power supply with minimum fluctuation becomes more important. Due to the use with ease, LM78XX and LM79XX serials were selected for the regulators. These IC regulators are in TO-220 package with three terminals, including input, common, and output. They all have internal current limiting, thermal shutdown and safe operating area protection. LM78XX are for positive DC output, while LM79XX are for negative DC output.

3.7.2 Simulation is SPICE

By using the Multisim program, the input to the transformer is 115 VAC 60 Hz power signal. The output waveform of +30 VDC obtained on oscilloscope is shown in Figure 3.7.2. The +30 VDC output on the graph shows that it is a straight line without visual AC ripple.

88

Figure 3.7.2 Input and Output of Power Supply

In order to find out the approximate AC component in the 30 VDC output, multimeters are used for another simulation as illustrated in Figure 3.7.3. As it shows, the AC component is 32 mV. Then, ripple percentage is 0.032 / 30 = 0.1 %. This is very close to the calculated value above.

Figure 3.7.3 AC Ripple Measurement

89

3.7.3 Parts Required

The following parts are required for the power supply module.

● Power transformer: 2X22 VAC 100VA ● Bridge rectifier: 100V 4A ● Regulators: LM7805, LM7809, LM7815, LM7915 ● Capacitors: 8 ● Resistor: 1 ● LED: 2.4V/10mA ● Switch: DPST ● Fuses and holders: 3

4 Design Summary

4.1 Pre Amplifier

Pre Amplifier circuit design is vital to the success of the overall amplifier design because the Pre amplifier is the first stage of the whole system. Any noise generated in this stage will be amplified after signal getting into the second stage. First, using good quality of IC’s is important since the main components are two Op Amps IC’s. Secondary, low variations of resistors and capacitors should be used. For resistors, precise 1% metal films resistors and 5% capacitors will be used. Third, the power supplies to the Op Amps must be as stable and smooth as possible. AC ripples reduction is accomplished by parallel connected capacitors along the power supply pins of the two Op Amps. The circuit descriptions for the three stages of the pre amplifier are explained as the following.

Band Pass Filter

The input to the pre amplifier circuit is the signal coming out from either analog or digital effects modules. R1, C1, R2, and C2 complete an initial band pass filter in such a following way. R1 and C1 form a low pass filter. The cutoff frequency can be obtained by fc = 1 / (2 R1C1) = 1.6 MHz. C2 and R2 form a high pass filter. The cutoff frequency is fc = 1 / (2 R2C2) = 4 Hz.

First & Second Stage Op Amp

The first and second gain stages use two same op amps IC’s of OPA-134 from TI. They are connected in noninverting configuration for having large input impedance. Also the signal is further filtered by the active filters completed by both OPA-134’s configurations. The low pass filter is established by R4 and C3, as well as R9 and C8. The cutoff frequency for this LP filter is determined by fLC = 1 / (2 R4C3) = 19.5 kHz. The high pass filter is established by R5 and C4, as well as R10 and C9. The cutoff

90

frequency for this HP filter is determined by fHC = 1 / (2 R5C4) = 30 Hz. C10, C11, C12, and C13 are bypass capacitors to minimize the power supply noise.

Gain & Tone Control

The equalization circuit used in this pre amplifier is the “tone stack” used in early Fender tube amplifiers. The treble capacitor C5 is on top of the stack. C5 and the combination of three potentiometers VR1, VR2, and VR3 complete a high pass filter. The bass capacitor C6 is in the middle of the stack. R6, C6 and the combination of VR2 and VR3 complete a low pass then a high pass filter. The mid capacitor C7 is at the bottom of the stack. As the bass circuit, R6, C7, and VR3 complete first a low pass then a high pass filter to let the mid frequency through. Finally, adjusting the VR4 will determine the loudness by adjusting the input to the second Op Amp. Due to the human’s ears logarithm reaction to the loudness, a log scale of potentiometer VR4 should be chosen.

The complete final pre amplifier circuit schematic is shown in Figure 4.1.1.

Figure 4.1.1 Detailed Schematic of Pre Amplifier

91

Part Description Quantity Unit Cost Total Pre Amplifier U1, U2 OPA134 Low noise Op Amps 2 $2.94 $5.88 C1 100pF 50V Ceramic 1 $0.29 $0.29 C2 0.039uF 50V Ployester 1 $0.32 $0.32 C3, C8 1uF 50V Ceramic 2 $0.48 $0.96 C4, C9 120pF 50V Ceramic 2 $0.29 $0.58 C5 1nF 50V Polyester 1 $0.33 $0.33 C6 220nF 50V Polyester 1 $0.64 $0.64 C7 100nF 50V Polyester 1 $0.46 $0.46 R1 1k 0.25W Metal Film 1 $0.15 $0.15 R2, R7 1M 0.25W Metal Film 2 $0.15 $0.30 R3, R8 2.2k 0.25W Metal Film 2 $0.15 $0.30 R4, R9 5.1k 0.25W Metal Film 2 $0.15 $0.30 R5, R10 68k 0.25W Metal Film 2 $0.15 $0.30 R6 33k 0.25W Metal Film 1 $0.15 $0.15 VR1, VR2 50k Potentiometer Linear 2 $2.50 $5.00 VR3 10k Potentiometer Linear 1 $2.50 $2.50 VR4 500k Potentiometer Log 1 $2.50 $2.50 Sub Total $20.96

4.2 Power Amplifier

The successful design of power amplifier will depend on the stability of the circuit. In order to make it stable, the bias for transistors should be stable. It is accomplished by using negative feedback from output to input. Also, one must consider the temperature rises during the operation. The power transistors will generate great amount of heat. Attaching heat sinks is the most common method to dissipate the heat. Due to the amount of output power of this project, a fan is not necessary. In addition, the bias transistor for the power transistors will be attached on the same heat sinks. So its temperature parameters will have the same changes as those power transistors. The circuit descriptions of the four stages of the power amplifier are explained as following.

Input Stage

Transistor pair of Q2 and Q3 form a differential circuit input stage. The input signal (output from the pre amplifier) is fed into Q2, while the output signal is fed back into Q3 through R9. Transistor Q1 provides constant current for Q2 and Q2. The value of the constant current source can be calculated by (VD1-VBE) / R3 = (12 - 0.55) / 5.6 = 2 mA. C2 and R2 form a RF filter with the cutoff frequency equal to 154 kHz by using fc = 1 / (2RC). The input impedance of input stage is about 50 kohm.

92

Voltage Amplifier Stage (VAS)

Transistor Q5 is a common emitter amplifier to provide the effective voltage gain for the output stage. The output of Q2 from the differential amplifier input stage comes into the base of Q5. The output of Q5 goes into driver Q8 and Q9. C7 is for Miller compensation.

Output Stage

Transistor Q4 provides the biases for Q8 and Q9. The trim pot TR1 and a couple of resistors form an adjustable bias for Q4. Adjusting TR1 can change the Q-point of Q4, in turn change the DC bias for Q8 and Q9. Since power transistor Q10 and Q11 operate in high temperature, their parameters would change during operation. To minimize this temperature effect, Q4 will be mounted on the same heat sink with Q10 and Q11 so it has same temperature change effect. This will keep providing proper biases for Q8, Q9. Transistor Q8 and Q10 as well as Q9 and Q11 form two complementary Darlington pair, thus this is a push-pull operation. Power amplifier circuits are vulnerable to instability. It is mainly caused by both phase shifts introduced by the feedback loop and imperfections of components. The most common result of these unwanted effects is high frequency oscillation. This circuit can provide oscillation prevention in the following ways. A simple low pass RC filter (C2, R2) preceding the input stage is designed to filter out RF. C4 is a HF feedback capacitor used for decreasing gain at higher frequencies and therefore prevent oscillation. C8 and C10 are capacitors to eliminate noise from power supply. R23 and C9 form a Zobel network which used to compensate for the speaker’s inductance.

Circuit Protection

To protect a false condition, usually a short circuit load, there should be some types of circuit protection in output stage. A usual way to do this is by inserting a current limiting circuit before the drivers of power transistors. This allows time for a fuse to blow. Transistor Q6 and Q7 provide such a short circuit protection. Should a short circuit load occur and when the output current increase to 3.33A, the voltage drop across R21 reaches 1.10 V. Q6 will be conducted through R16 and R19 voltage division network. This will drain the current from the base of Q8, and the current at the emitter of Q10 will be minimized.

The complete final power amplifier circuit schematic is shown in Figure 4.1.2.

93

Figure 4.1.2 Detailed Schematic of Power Amplifier

94

Part Description Quantity Unit Cost Total Power Amplifier Q1, Q2, Q3 BC556 PNP 3 $0.20 $0.60 Q4 MJE182 NPN 1 $0.45 $0.45 Q5 BC639 NPN 1 $0.47 $0.47 Q6 2N3904 NPN 1 $0.20 $0.20 Q7 2N3906 PNP 1 $0.20 $0.20 Q8 TIP29C NPN 1 $0.56 $0.56 Q9 TIP30C PNP 1 $0.59 $0.59 Q10 TIP33C NPN 1 $1.95 $1.95 Q11 TIP34C PNP 1 $1.69 $1.69 D1 1N5242B 12V 1 $0.42 $0.42 D2, D3 1N4148 2 $0.23 $0.46 C1 1uF 50V Ceramic 1 $0.68 $0.68 C2, C4 22pF 50V Ceramic 2 $0.44 $0.88 C3 220uF 50V Aluminum 1 $0.68 $0.68 C5, C6 100uF 50V Aluminum 2 $0.68 $1.36 C7 100pF 50V Ceramic 1 $0.29 $0.29 C8, C9, C10 0.1uF 100V Ceramic 3 $0.29 $0.87 R1, R8 1k 0.25W Metal Film 2 $0.15 $0.30 R2 47k 0.25W Metal Film 1 $0.15 $0.15 R3 5.6k 0.25W Metal Film 1 $0.15 $0.15 R4 4.7k 0.25W Metal Film 1 $0.15 $0.15 R5, R12 2.2k 0.25W Metal Film 2 $0.15 $0.30 R6, R7 680R 0.25W Metal Film 2 $0.15 $0.30 R9 22k 0.25W Metal Film 1 $0.15 $0.15 R10 2.7k 0.25W Metal Film 1 $0.15 $0.15 R11 150k 0.25W Metal Film 1 $0.15 $0.15 R13 510R 0.25W Metal Film 1 $0.15 $0.15 R14, R15 330R 0.25W Metal Film 2 $0.15 $0.30 R16,R17,R18,R19,R20 220R 0.25W Metal Film 5 $0.15 $0.75 R21, R22 0.33R 3W Wirewound 2 $0.45 $0.90 R23 15R 0.5W Metal Film 1 $0.20 $0.20 TM1 2k Trim Pot 1 $1.63 $1.63 Heatsink 26C/W 2 $0.58 $1.16 Heatsink 2C/W 2 $6.40 $12.80 Sub Total $32.04

95

4.3 Power Supply

The primary winding of the transformer T1 is connected to 115 VAC through the double poles single throw (DPST) switch and line fuse F1. The reason to use double poles switch is guaranteed to cut off the hot line even when the plug is in backward. Line fuse F1 is a slow blow fuse. T1 is a single primary and single secondary windings with center tap. BR1 is the bridge rectifier for full wave rectification. C1 and C2 are filter capacitors. C3 and C4 are noise bypass capacitors. R1 is the bleeding resistor, which is in series with a LED. Fuse F2 and F3 are rail fuses. Regulator IC’s LM7805, LM7809, LM7815, and LM7915 are connected in parallel. The input to those IC’s is the 30 VDC. Each output is again filtered by C5, C6, C7, and C8. The finished circuit for the power supply illustrated below in Figure 4.3.1.

Figure 4.3.1 Detailed Schematic of Power Supply

96

Part Description Quantity Unit Cost Total Power Supply T1 24VAC CT 100VA 1 $30.00 $30.00 SW1 DPST Switch 1 $4.00 $4.00 BR1 100V 4A Rectifier Bridge 1 $1.50 $1.50 C1, C2 4700uF 50V Aluminum 2 $2.80 $5.60 C3, C4 0.1uF 100V Ceramic 2 $0.29 $0.58 C5, C6, C7, C8 100uF 50V Aluminum 4 $0.70 $2.80 D1 LED 2.4V 10mA Green 1 $0.50 $0.50 R1 4.7k 1W Carbon 1 $0.30 $0.30 Sub Total $45.28

4.4 Analog Effects

4.4.1 Fuzz Box

The final Fuzz Box circuit is listed below in Figure 4.4.1. The signal is driven by two 2N2222A transistors out to a volume control knob. A 10 k-ohm potentiometer is attached to Q2 to control the “fu ” effect.

Figure 4.4.1 Fuzz Box Finalized Schematic

97

Part Description Quantity Unit Cost Total Fuzz Box Transistor TRANS GP SS NPN LP 40V TO92 2 $0.34 $0.68 Potentiometer 100K-Ohm Linear-Taper Potentiometer 1 $3.49 $3.49 Potentiometer 10K-Ohm Linear-Taper Potentiometer 1 $3.49 $3.49 Capacitor CAP ALUM 2.2UF 50V 20% RADIAL 1 $0.04 $0.04 Capacitor CAP CER 0.1UF 50V 10% RADIA 1 $0.29 $0.29 Capacitor CAP ALUM 20UF 16V AXIAL 1 $3.29 $3.29 Resistor RES 1M OHM 1/4W 5% 1206 SMD 1 $0.10 $0.10 Resistor RES 330 OHM 1/4W 5% 0805 SMD 1 $0.14 $0.14 Resistor RES 33K OHM 1/4W 5% 0805 SMD 1 $0.15 $0.15 Resistor RES 8.2K OHM 1/4W 5% 0805 SMD 1 $0.14 $0.14 Resistor RES 100K OHM 1/4W 5% 0805 SMD 1 $0.14 $0.14 Sub Total $11.95

4.4.2 Octave Up

A completed circuit of the Octave Up pedal is presented below in Figure 4.4.2. The incoming signal from the guitar is amplified through the NE5532AI with an adjustable gain. The signal is then flipped up with the full-wave rectifier at the end of the circuit to give it its desirable sound.

Figure 4.4.2 Octave Up Finalized Schematic

98

Part Description Quantity Unit Cost Total Octave Up Op-Amp IC OPAMP GP 10MHZ DUAL 8DIP 1 $1.08 $1.08 Potentiometer 100K-Ohm Linear-Taper Potentiometer 2 $3.49 $6.98 Capacitor CAP ALUM 100UF 4V 20% SMD 2 $0.48 $0.96 Capacitor CAP ALUM 10UF 16V 20% SMD 2 $0.44 $0.88 Capacitor CAP FILM 10000PF 16VDC 0805 1 $0.67 $0.67 Capacitor CAP CER 0.47UF 16V 10% X5R 0402 1 $0.12 $0.12 Resistor RES 10K OHM 1/8W 5% 0805 SMD 4 $0.08 $0.31 Resistor RES 220K OHM .4W 5% 0805 SMD 1 $0.10 $0.10 Resistor RES 100K OHM .4W 5% 0805 SMD 1 $0.10 $0.10 Resistor RES 33K OHM 1/4W 5% 0805 SMD 1 $0.15 $0.15 Resistor RES 4.7K OHM 1/4W 5% 0805 SMD 1 $0.13 $0.13 Sub Total $11.48

4.4.3 Orange Squeezer

Figure 4.4.3 represents the Orange Squeezer compressor pedal. The guitar signal is driven by two J310 FET transistors. The signal is then amplified by a TL072ACD amplifier to increase its signal strength. The signal is finally controlled by a 10 k-ohm control knob.

Figure 4.4.3 Orange Squeezer Schematic

99

Unit Part Description Quantity Cost Total Orange Squeezer IC OPAMP JFET 3MHZ DUAL Op-Amp 8SOIC 1 $0.91 $0.91 IC SWITCH RF N-CH 25V 10MA Transistor SOT23 1 $0.44 $0.44 DIODE SWITCH 100V 0.15A Diode SOD123 1 $0.17 $0.17 10K-Ohm Linear-Taper Potentiometer Potentiometer 2 $3.49 $6.98 Resistor RES 1M OHM 1/4W 5% 0805 WIDE 1 $0.14 $0.14 Resistor RES 82K OHM 1/4W 5% 1206 SMD 1 $0.10 $0.10 RES 2.4K OHM 1/4W 1% 0805 Resistor SMD 1 $0.28 $0.28 RES 470K OHM 1/4W 5% 0805 Resistor WIDE 3 $0.14 $0.42 RES 390K OHM 1/4W 5% 0805 Resistor SMD 1 $0.23 $0.23 Resistor RES 10K OHM 1/4W 5% 0805 SMD 1 $0.23 $0.23 RES 220K OHM 1/4W 5% 0805 Resistor SMD 1 $0.23 $0.23 RES 100K OHM 1/4W 5% 0805 Resistor SMD 1 $0.23 $0.23 RES 1.5K OHM 1/4W 5% 0805 Resistor SMD 1 $0.14 $0.14 CAP CER 0.047UF 50V 10% X7R Capacitor 0805 2 $0.10 $0.20 CAP CER 4.7UF 25V 10% X5R Capacitor 1206 4 $0.25 $1.00 FILM CAP 2.2NF 20% 300VAC Capacitor MKPY2 1 $0.85 $0.85 Sub Total $12.55

100

4.4.4 Tremolo

The completed circuit schematic for the Tremolo effect is pictured in Figure 4.4.4. The LM555C creates a square wave that is fed into Q1. Q1 then inverts the -9V signal that is fed into the LM386 to produce the change in volume effect for the Tremolo pedal.

Figure 4.4.4 Tremolo Schematic

Part Description Quantity Unit Cost Total Tremolo Op-Amp IC AMP AUDIO PWR .325W AB 8SOIC 1 $0.91 $0.91 Timer IC OSC MONO TIMING 500KHZ 8-SOIC 1 $0.55 $0.55 Potentiometer 100K-Ohm Linear-Taper Potentiometer 1 $3.49 $3.49 Potentiometer 1M-Ohm Linear-Taper Potentiometer 1 $3.49 $3.49 Resistor RES 4.7K OHM 1/4W 5% 0805 SMD 1 $0.13 $0.13 Resistor RES 470 OHM 1/4W 5% 0805 SMD 1 $0.13 $0.13 Capacitor CAP ALUM 10UF 16V 20% SMD 2 $0.44 $0.88 Capacitor CAP ALUM 220UF 16V 20% SMD 1 $0.63 $0.63 Capacitor CAP CER 0.1UF 50V 10% SMD 1 $0.29 $0.29 Sub Total $10.50

101

4.4.5 Big Muff

Figure 4.4.5 is the revised circuit layout for the Big Muff. There are four transistor stages that produce the sound of the distortion. The first stage uses the BC109BP for the clean drive. The second and third stages use the BC549BP for the clipping drive. The final stage uses the 2N5088 for the tone recovery of the effect pedal.

Figure 4.4.5 Big Muff Schematic

Part Description Quantity Unit Cost Total Big Muff Transistor TRANS NPN PWR GP 4A 60V TO225AA 1 $0.76 $0.76 Transistor TRANSISTOR NPN 65V 100MA TO-92 2 $0.19 $0.38 Transistor NPN LL LN HI GAIN AMP TRANS TO92 1 $0.18 $0.18 Diode DIODE SWITCH 100V 0.15A SOD123 4 $0.17 $0.68 Resistor RES 36K OHM 1/4W 5% 0805 SMD 1 $0.13 $0.13 Resistor RES 100K OHM 1/4W 5% 0805 SMD 4 $0.15 $0.60 Resistor RES 120 OHM 1/4W 5% 0805 SMD 3 $0.14 $0.42 Resistor RES 470K OHM 1/4W 5% 0805 SMD 4 $0.14 $0.56 Resistor RES 39K OHM 1/4W 5% 0805 SMD 2 $0.13 $0.26 Resistor RES 1K OHM 1/4W 5% 0805 SMD 1 $0.15 $0.15 Resistor RES 10K OHM 1/4W 5% 0805 SMD 3 $0.15 $0.45 Resistor RES 27K OHM 1/4W 5% 0805 SMD 1 $0.12 $0.12

102

Resistor RES 2.2K OHM 1/4W 5% 0805 SMD 1 $0.13 $0.13 Potentiometer 100K-Ohm Linear-Taper Potentiometer 3 $3.49 $10.47 Capacitor CAP CER 1UF 16V 10% X5R 0603 4 $0.10 $0.40 Capacitor CAP CER 500PF 16V 10% X5R 0603 3 $0.10 $0.30 Capacitor CAP FILM 0.68UF 63VDC 2220 4 $2.32 $9.28 Capacitor CAP CER 0.47UF 100V 10% RADIAL 1 $0.75 $0.75 Capacitor CAP CER 0.1UF 50V 10% SMD 1 $0.29 $0.29 Sub Total $26.31

4.4.6 Distortion with Stutter Effect

The completed circuit layout for the Distortion pedal with stutter effect is presented below. The NE555 timer creates a square wave that is fed into the LM386 which creates the stutter sound. The LM386 then amplifies the signal to give it the distortion effect desired. The completed circuit of the Distortion pedal is shown below in Figure 4.4.6.

Figure 4.4.6 Distortion with Stutter Effect Schematic

Unit Part Description Quantity Cost Total Distortion with Stutter IC AMP AUDIO PWR .325W AB Op-Amp 8SOIC 1 $0.91 $0.91 IC OSC MONO TIMING 500KHZ Timer 8-SOIC 1 $0.55 $0.55 100K-Ohm Linear-Taper Potentiometer Potentiometer 2 $3.49 $6.98 10K-Ohm Linear-Taper Potentiometer Potentiometer 1 $3.49 $3.49 Resistor RES 470 OHM 1/4W 5% 0805 SMD 1 0.14 $0.14 Resistor RES 1K OHM 1/4W 5% 0805 SMD 1 0.13 $0.13

103

Capacitor CAP ALUM 10UF 16V 20% SMD 3 0.44 $1.32 Capacitor CAP ALUM 220UF 16V 20% SMD 1 0.56 $0.56 Capacitor CAP CER 0.1UF 50V 10% SMD 1 0.29 $0.29 Sub Total $14.37

4. 5 Digital Effects

4.5.1 Reverberation

The following Figure 4.5.1 is the code flow diagram of the reverberation effect. It shows the algorithm from the beginning where it clears the reverberation array to reading and writing on the CODEC then last the output. The scalar quantity is the “depth” value.

Figure 4.5.1 Reverberation Code Flow Chart

104

4.5.2 Fuzz

Figure 4.5.2 shows the fuzz effect. It takes the input and clips the wave form, which produces the distorted effect. The code flow diagram starts at the declaration of the variables used shows the reading off the CODEC, the processing, the writing on the CODEC and the output.

Figure 4.5.2 Fuzz Code Flow Chart

105

4.5.3 Echo

The figure below is for the echo effect. It starts with the clearing and demonstration of the initialization, reading off the CODEC, processing, writing on the CODEC, then outputting the information. The scalar quantity is the “depth” value. Figure 4.5.3 below is a block diagram of the echo effect.

Figure 4.5.3 Echo Code Flow Chart

106

4.5.4 Tin Can

The following Figure 4.5.4 notes the code flow of the tin can effect. It starts with the processing, omitting the initialization and reading and writing on the CODEC. The number of sample skipped is dependent of the amount that the sampling rate is down sampled.

Figure 4.5.4 Tin Can Code Flow Chart

107

4.5.5 Phase

The following code flow diagram in Figure 4.5.5 shows the phase effect algorithm. It is split into two pages. The phase effect uses down sampling with a variable delay size that is determined by the counters. After the down sampling and the current delay length the input up sampled for the output that is returned to the main function. The interval counter keeps track of the current size.

108

Figure 4.5.5 Phase Code Flow Char

109

4.5.6 Robot

The following Figure 4.5.6 is the Robot effect code flow diagram. It starts with the processing. The output is returned to the main function.

Figure 4.5.6 Robot Code Flow Chart

110

4.5.7 Fuzzy Tube

Figure 4.5.7 below is the fuzzy tube code flow diagram. It starts with the signal processing, the ends at the output which is returned to the main function.

Figure 4.5.7 Fuzzy Tube Code Flow Chart

111

Below is the completed parts list for the DSPC5515.

DSP PCB AT93C46DN-SH-T IC,SOIC,SERIAL EEPROM,3 WIRE 1 $0.28 $0.28 TMX320VC5515AZC H IC,BGA196,.65mm PITCH,DSP,VC5515 1 $0.28 $0.28 IC,TSOP48,FLASH,32M,3V,70nS,X8/X16 S29AL032D70TFI040 BOTTOM 1 $3.49 $3.49 FT2232HL IC,QFP64,DUAL USB UART 1 $6.71 $6.71 TLV320AIC3204IRH IC,QFN32,ULTRA LOW POWER STEREO BT AUDIO 1 $5.85 $5.85 C,SMT,PROTECTION ARRAY,2- TPD2E001DRLR CHANNEL 15-kV 1 $0.61 $0.61 TPS61041DBVR IC,SO5,BOOST CONVERTER,250mA 1 $1.58 $1.58 IC,DBV6,SINGLE D-TYPE FLIP-FLOP SN74LVC1G175DBV WITH 1 $0.35 $0.35 TPS76601DR IC,SO8,250mA TPS76601DVOLTAGE 1 $1.51 $1.51 IC,TSSOP14,QUAD BUS BUFFER SN74LV125APWR GATE,LOW 1 $0.46 $0.46 IC,SO5,LINEAR TPS76901DBVT REGULATOR,ADJUSTABLE, 1 $1.01 $1.01 CB3LV-3C-12M0000 OSC,SMT,12 MHZ 1 $1.05 $1.05 SSPT7F- 12.5PF20PPM CRYSTAL,SMT,32.768kHz 1 $0.83 $0.83 ZMM5234B-7-F DIODE,MELF,ZENER,6.2V,500mW 1 $0.34 $0.34 MBR0520LT1G DIODE,S0D-123,SCHOTTKY POWER,1A 1 $0.34 $0.34 VLF3010AT- 2R2M1R0 INDUCTOR,SMT,2.2uH,+/-20% 1 $0.90 $0.90 BLM18AG601SN1D FERRITE BEAD,SMT 0603,600 OHM 6 $0.10 $0.60 BLM21AG151SN1D FERRITE BEAD,SMT 0805,150 OHMS 2 $0.14 $0.28 BLM21PG221SN1D S FERRITE BEAD,SMT 0805,220 OHMS 3 $0.11 $0.33

CAP CAP,SMT 104 $0.47 $48.88 RES RES,SMT 92 $0.01 $.92 SWITCH,SMT,PUSHBUTTON,MOMENT KSC421G70SHLFS ARY,.25 1 $0.25 $0.25 CONN,PLUG,4 POS,RIGHT ANGLE,USB 48037-0001 TYPE A 1 $1.19 $1.19 10051922-1410ELF CONN,SMT,RECEPTACLE,14 POS.,.5mm 1 $0.70 $0.70 MUSB-05-S-B-SM- A-K-TR CONN,SMT,RECEPTACLE,5 POS,RIGHT 1 $0.94 $0.94 SJ1-3515-SMT CONN,SMT,STEREO JACK,3.5mm 2 $1.68 $3.36 $59.61 Sub Total

112

4.6 Bluetooth Module

The RN42 module was used to implement the wireless data transfer between the GUI and the DSP. The pins corresponding to GND, VDD, TX, RX, and CTS were the only pins of detail for this design project. The schematic for the RN42 (Figure 4.6.1), along with the schematic for the linear voltage regulator (Figure 4.6.2), is provided below with permission from MicrochipDirect.com.

Figure 4.6.1 RN-42 Module

113

Figure 4.6.2 Voltage Regulator

Unit Part Description Quantity Cost Total BT PCB RN-42 BT Module (with PCB trace antenna) 1 $14.56 $14.56 TC-1185 Voltage Regulator (5V to 3.3V) 1 $0.45 $0.45 D1 Green LED 1 $0.50 $0.50 D2 Red LED 1 $0.50 $0.50 2 (470 ohms) 1(3300 ohms) 2 (300k ohms) Resistors 2(100k ohms) 7 $2.00 $2.00 Capacitors 2 (1 microF) 2 $1.00 $1.00 Sub Total $19.01

4.7 Cabinet Housing

The housing for the amplifier was constructed from the various materials and parts shown in the parts list below. The frame of the amplifier was designed using wood. It had corresponding knobs which allowed the control of tone, volume, and power. Wood screws were used to form the housing and to prevent the speaker from movement. A mesh fabric was used to cover the speaker and prevent any extra hollow space.

The amplifier will look just a regular guitar amp, only larger for the added built in effects. The group will be designing and constructing the amplifier to fit the requirements for their designs. The amplifier will be assembled within a household garage. Painting, drilling, and general assembly will all be done by the members within the group.

Parts for the amplifier will be purchased from Home Depot, RadioShack, SkyCraft and Amazon.com. The speaker will be bought second-hand from an existing amplifier. This will drastically reduce the cost of the cabinet. The mesh fabric will be purchased from an arts and crafts store. The plan for the cabinet is to make it look as close to a name brand guitar amp as possible.

The appearance of the amplifier will be a retro look with the University of Central Florida Pegasus logo as a center piece. Colors of black and gold will flow throughout the outside of the amp.

In the final design we refurbished an vintage amplifier case, only small modifications were needed,

114

Unit Part Description Quantity Cost Total Cabinet Materials

Speaker 10" 50W Guitar Speaker 1 $50.00 $50.00 Housing Wood 1 $20.00 $20.00 Knobs Tonal, Volume, and Power 5 $8.00 $40.00 Screws Wood Screws 1 $6.00 $6.00 Fabric Mesh 1 $25.00 $25.00 Total $71.00

5 Testing

5.1 Pre Amplifier

Set up a DC bench power supply for output of +15V and -15V. Check to make sure the output voltages are correct by using Multimeter. When connecting the power supply to the pre amplifier board, make certainly that the polarities are correct. Using the sine wave of 100mV, 1k Hz from a signal generator as the input, check the output with an oscilloscope. Observe the output waveform. It should be the same as input waveform except the amplitude. If the output amplitude is too large for the power amplifier input, then adjust the feedback resistors for either Op Amps IC’s to either enlarge or lessen the gain of the pre amplifier.

5.2 Power Amplifier

Connect a Multimeter to measure the resistance between R14 and R16 (the effective resistance of the trim pot TR1). Turn trim pot TR1 all the way on one direction and the resistance should read about 2k ohm. Connect power amplifier circuit board to a bench dual DC power supply. Make sure the positive and negative power supply are connected correctly. Set the output to +30V and -30V. With these setup, do the following: 1. Turn on the power supply. Measure the voltage at the output terminal (the terminal will be connected to the speaker). Turn TR1 slowly towards the other way and stop when the output voltage is very closed to zero. 2. Connect a multimeter in series with the power supply. Verify the DC current is about 100mA. 3. Check the voltage across R4 is about 11.2V. This ensure the constant current source to the differential input is 2mA.

115

4. With a 1kHz input signal of 1V peak sine wave, use power meter to verify power output of at least 30W 5. With the above input signal, use oscilloscope to verify no visible harmonic distortion.

5.3 Power Supply

Before hooking up with any AC power source, make sure all fuses are placed and their values correct. Connect a voltage meter to the outputs of the power supply. Plug in the power supply board into the wall AC outlet. Watch the voltage meter to check the output voltages. They should be within 5% of the targeted output voltages. Listen for any excessive noise from the power transformer. Wait for five minutes and touch the outside of the transformer to look for any excessive heat. Use an oscilloscope to check AC ripples for every output voltage. The ripples should be within 1%.

5.4 Analog Effects

All boards will be laid out and test on PCB prototype boards before moving on to finalized PCB layouts on Altium. Using a power supply, boards will be powered on separately to test shortages, or inaccuracies within the design. Changes within values of resistors and capacitors will be tested for further changes to sound or output of effects. Once the boards have each been tested separately, all boards will be daisy-chained together to be tested simultaneously. Again, any corrections to be made will be noted and transferred to schematic and PCB on Altium. 5.5 Digital Effects

The effects were developed in TI’s code composer studio and implemented on the TMS320C5515 eZdsp. The eZdsp has eighth inch stereo jack input and outputs they were utilized in testing with a conversion jack to fourth inch guitar input cables. The operational amplifier in Figure 2.5.1 was constructed to amplify the input signal from the eZdsp. The gain was also controlled in software. Then the output was connected to an oscilloscope and measured. The wave forms are included in the design section . The PCB parts list has been complied and will be laid out during the second portion of the design course. Further testing will be continued.

5.6 User Interface

Interfacing between the user interface (using a Windows laptop) and the DSP was established first via USB. Various parameters were changed on the GUI to ensure that the data was properly transmitted to the DSP. The output signal was verified through a store-

116

bought amplifier to develop an understanding of the data transfer via USB. After the USB connection was established, the BT module could easily be managed and configured.

5.7 PCB Manufacturing and Assembly

For the manufacturing of the PCB boards, the group members will be using Advanced Circuits through their student program. There will be a total of three boards containing each subsystem of the guitar amplifier. All three boards are not expected to exceed 2- layers. The price for each board is $33 with the discounted student price. In agreement with Advanced Circuits, the boards will be delivered to the university. Once boards have been printed and delivered, the group members are going to assemble each one’s respective board. The university has provided a senior design lab for students to work on their PCB boards. The group members will use the lab to spin their boards for final revisions. Boards will then be tested and revised. If any further crucial revisions are necessary, corrections will be made and new PCB boards will be ordered.

6 Administrative Content 6.1 Milestone Chart Discussion

Below is the following schedule that was designed at the beginning of the semester. As a group a plan was laid out in order to make significant and relevant progress. The labor was divided into four sections. The sections included the power system, digital effects, analog effects, and user interface. The accomplishments were assigned each week, to each group member. The following figure specifies the accomplishments for each week and which member was responsible. The developed schedule allowed the group to work at their own pace while it also provided structure to efficiently plan out the duration of the summer semester design course.

117

The group had periodic weekly meeting. Weekly meetings ensured that each member could work in parallel. If any member’s information was needed for a subsystem outside their own, the information could be accounted for and scheduled so that each member could continue to make progress.

In addition to the weekly meets Google drive was the host of the working copy of the first semester’s draft. This allowed for real time editing of the paper and made sure that each member was working on the most current draft. Group member also used Gmail in order to relay scheduling information and design information. 6.2 Budget and Finance

The budget for the guitar amplifier project was determined by the group members. The total amount that was agreed upon was $800. No sponsorship or additional funding was provided, all funding was provided by group members exclusively.

After prototyping and projecting expected future costs the total budget was calculated. The total cost projected is $622.01. This is below the amount that the group agreed upon. There is a resalable amount of difference between what the group’s budget and the project cost so that the group can account for unexpected expenses.

The group’s decision was to divide the cost into equally amongst each other. For the purpose of designing and testing, each student will be purchasing their own test parts and may add it to the total cost that will be divided at the end of the project. Receipts will be kept for the purpose of refunding in case of sponsorship. Free samples will be deducted from the budget, although they may not be accounted for immediately. For the larger purchases, the group members will be buying the parts together because of the high cost.

Group members will be actively seeking websites that are student friendly for free parts and development boards. This is not excluding any discounted parts for students. In our budget we accounted for parts at retail cost, but the budget may decrease because of these factors. It is estimated that about %10 percent of the budget will either be free or discounted parts.

Although the group is not sponsored, pursuit of future sponsorship is going to be aggressively pursued. Through internships, some group members may get their employee to help with purchases and free shipping.

7 Conclusion

Overall the first design semester was considered successful by each group member. The first task that lead to our groups success that was the administrative decisions. The group took the time to schedule out the semester in a way that was realistic and effective. Periodic meeting we set so that to make sure that parallel progress could be made. In the

118

cases where the subsystems depended on each other, a pre meditated plain was set in order to make sure that the information that depended on each other could be relayed.

The budget was setup in the initial meeting so that we could move forward on design decisions. This allowed each member to understand how much they could spend on each respective system and would help to narrow down the initial search for proto type boards and parts.

The next step was the each of the four subsystems was thoroughly researched. The research led to informed decisions about the most effective way to design each group member’s corresponding subsystem. The research included the looking up all the circuit theory, circuit performance, prototyping boards, and back ground information necessary to move forward with the design.

With the knowledge of the budget and the research experience the initial prototyping boards were able to be selected. There were a variety of boards to choose from, but the budget constraints and knowing the specifications made picking out the prototyping boards an easy choice. The prototype boards were used for the digital effects and the Bluetooth. For the analog effects and the power system, pre amplifier, and power amplifier the circuits we simulated. The research and budget were in consideration when designing these system’s parts list was develop in order to fulfill the budgetary specifications. Once the subsystems were simulated or prototyped it gave a clear picture toward the final design. If there were changes needed then they would be addressed and if not they were further developed.

The results and testing gathered will be used in the final semester of the design course. The systems will be laid out on PCBs. All the subsystems will come together to create one working system. From the results gathered from this report, the group has confidence that the merger of all systems will be successful.

Appendices

A. Copyrights Permissions

119

D. Acknowledgements

Bob Cordell. Designing Audio Power Amplifiers. NY: McGraw-Hill, 2011

Bluetooth Advanced User Manual. http://www.microchipdirect.com

CT Circuits Today. http://www.circuitstoday.com.

Donald A. Neaman. Microelectronics Circuit Analysis and Design. NY: McGraw-Hill, 2007

Fastlanestoner.Amplifiers Types and Classifications. http://www.ultimate- guitar.com/column/the_guide_to/amplifiers_types_and_classifications.html. 2012

120

RN-42-EK User’s Guide. http://www.microchipdirect.com

Teemu Kyttala. Solid State Guitar Amplifiers. 2008

Build a FuzzBox http://www.hawestv.com/amp_projects/fuzzbox/fuzz1.htm

General gadgets http://www.generalguitargadgets.com/tech-pages/45-schematics/37-availableschematics

BeavisAudio http://www.beavisaudio.com/schematics/Dan-Armstrong-Orange- SqueezerSchematic.htm

BigMuff

Page http://www.bigmuffpage.com/Big_Muff_Pi_versions_schematics_part1.html

Instructables tremolo http://www.instructables.com/id/Make-Your-Own-Tremolo- Effects-Pedal/

Instructables stutter http://www.instructables.com/id/Distortion-Pedal-With-Stutter- Effect/

Mojo lama music http://www.mojolama.com/design-and-build-virtual-tremolo

Effects 101 http://www.bossus.com/go/effects_101/

How stuff works http://electronics.howstuffworks.com/gadgets/audio-music/guitar-pedal.htm

TMS320C5515 eZdsp Technical reference http://support.spectrumdigital.com/boards/usbstk5515/reva/files/usbstk5515_TechRef_Re vA.pdf

TMS320C5505 USB Stick Teaching Materials (CD) - C5000 teaching ROM

The Digital effect was biased off the teaching material provided and the code examples.

121