EE319K Laboratory Manual - The University of Texas at Austin

EE319K Laboratory Manual - The University of Texas at Austin
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 1
EE319K Laboratory Manual
Univ of Texas at Austin
Jon Valvano, Ramesh Yerraballi
Do not print the entire document; we will be making many changes.
Fall 2015 (8/9/15 version)
Table of Contents
LAB 1. DIGITAL LOGIC IMPLEMENTED ON A MICROCONTROLLER ................................................................... 6
LAB 2. FOR-LOOPS, WHILE-LOOPS, IF-THEN BRANCHING, SUBROUTINES, AND TIME DELAYS .............. 10
LAB 3. SWITCH AND LED INTERFACING...................................................................................................................... 13
LAB 4. MINIMALLY INTRUSIVE DEBUGGING METHODS ....................................................................................... 20
LAB 5. TRAFFIC LIGHT CONTROLLER ......................................................................................................................... 25
LAB 6. DIGITAL PIANO USING A DIGITAL TO ANALOG CONVERTER ................................................................ 31
SOLDERING GUIDE ............................................................................................................................................................. 41
LAB 7C. LCD DEVICE DRIVER FOR THE SITRONIX ST7735 (FALL 2015) ............................................................. 43
LAB 8. REAL-TIME POSITION MEASUREMENT SYSTEM ........................................................................................ 48
LAB 9. DISTRIBUTED DATA ACQUISITION SYSTEM ................................................................................................. 57
LAB 10. EMBEDDED SYSTEM DESIGN ........................................................................................................................... 66
This lab manual is updated each semester. If you find a mistake, or if you find a section vague or confusing, please email me
at ramesh@mail.utexas.edu . If you enjoy EE319K, you might also enjoy EE445L. Furthermore, if you really like EE319K,
you might consider specializing in embedded systems.
We will use the simulators and graders, but the automatic graders are provided only for debugging purposes. The TA will
assign grades for all components of the lab. There will be no automatic graders for Labs 7 – 10.
1) Download and install Keil 4.74 first.
https://www.keil.com/demo/eval/armv4.htm
2) Download and install the all examples, simulators and graders Labs 1-4,
http://edx-org-utaustinx.s3.amazonaws.com/UT601x/EE319K_InstallSpring2015.exe
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 2
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Laboratory grading policies: This is a programming class. Therefore, the quality of the software you write will significantly
affect your grade. In addition to writing software, each lab has specific activities that must be documented and turned in with
the lab. These deliverables will be screenshots, circuit diagrams, measurements or data analyses. When you get the program
finished, create one file of your source code, and all deliverables. Save this file as a pdf. Upload the pdf file to Canvas/SVN
as instructed, and bring the grading sheet to the demonstration. The TA will record the performance and demonstration
grades on the grading sheet. Labs are due during your scheduled lab period. A detailed schedule follows. Late assignments
incur a penalty. Exceptionally “good” programs may be given extra credit. If you do extra functions in your program, point
out and demonstrate the extra functions to the TA when you demonstrate your program. Partial credit, not to exceed a
maximum of 75% of full credit, may be awarded to a program that does not meet all specifications. Often it is better to
checkout late incurring penalties than to turn in an incomplete program. There are three components to the lab grade:
1. Deliverables 20% (completion grade, one pdf file, uploaded to Canvas/SVN, and showed to TA during demonstration):
2. Performance 35% (graded by the TA at the time of checkout):
How well does it work? Does it crash?
Does it handle correctly all situations as specified?
How clean is the user interface?
Possible 5% bonus, does it do more than specified?
3. Adhere to coding standard 5% (organization, style, comments etc.)
Comments explaining each subroutine (inputs, outputs, summary of behavior)
Modular programming (subroutines do only the tasks they were created for)
Code is clearly organized (not too many branches)
Good variable names (divisor is called divisor, result is called result etc.)
Consistent coding style throughout (bracket placement, proper indentation, variable naming scheme)
4. Demonstration 40% (graded by the TA at the time of checkout):
Can you explain to the TA how your software works? Be prepared to explain the data flow through your system.
Can you explain why you made certain software engineering tradeoffs?
Both partners if applicable must be present during the demonstration.
There is a lot of starter code in the book and on the internet. It is natural and appropriate for you to look at these examples.
However, during demonstration you are expected to understand both the algorithm used and the implementation details of
ALL software you are running at the time of checkout. In particular, the contents of these starter files will be examinable for
the lab checkouts as well as for in-class written exams. Even if your program works, you may get very poor score on the
entire lab (e.g., 0 to 50%) if you do not understand how it works. For Labs 3 through 9, if your system runs on the simulator,
but does not fully run on the real hardware, you may be assessed at least a 20% penalty on the entire lab.
Comments and policies about EE319K
Clearly describe the data and program structures in the software comments. Simple well-structured software with
descriptive labels is much easier to understand than complex software with voluminous comments. Use either top-down or
bottom up structure. Modular programming divides the problem into “natural divisions” with minimal coupling. The
interfaces should be simple and natural. The software will be graded on the correctness of the program function, the
comments, the interface to the user, the style, and the organization. The effort spent to write high quality software will be
rewarded both during debugging and when it is time to add features in the future.
1) EE319K is a “hands on” class and therefore, even though the labs are only 30% of the total grade, there is NO WAY to
make it through this class without taking all the labs seriously. It is during the lab experience that you will really learn.
2) The class changes every semester, so even if you are repeating the class, or have heard about the class from friends, please
read the lab manual and go to lecture.
3) The lab checkout times are arranged in advance, and may or may not take place during the lab slot to which you are
officially registered.
4) This lab manual describes the tasks for each lab and provides and examples of questions you can be asked. The questions
are meant to give you an idea of the topics, depth and scope of the questions. Your TA may ask similar but different
questions. You may also be asked ANY question related to the material in previous labs.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 3
5) In addition, the lab manual specifies a list of “deliverables”. There is a lab grade component for the deliverables and
missing some of components will affect the grade. You will create one pdf file with all the deliverables. You upload this file
to Canvas/SVN as instructed before the beginning of your check-out time. You also open the pdf file on the computer for the
TA to see during demonstration. You do not need to make hard copy prints in EE319K. It is OK to capture low resolution
screen shots, select the interesting parts in Paint, and then paste them into the document. (Ctrl-Alt-PrintScreen captures the
active window)
6) You must upload the deliverables before your time slot. Each lab will also have a Lab grading sheet. The grading sheets
are on the course web site, but you can find copies of the grading sheets in lab as well. If you begin to create the deliverables
file after the checkout started, you will be considered as not ready, and it will impact the grade.
7) Starting with Lab 3, each group of two students will be assigned a check-out time by the TA.
8) You will need to come to lab ahead of your assigned time, turn on your laptop, connect your laptop to your microcontroller
if needed and download your code. This way you are ready to demonstrate at the beginning of your assigned checkout time.
The checkout times are VERY short, and there will not be time for the TA to check you out if you need to spend half of the
checkout time setting up or creating documents.
9) Student(s) not ready with everything set-up on the time assigned, won’t be able to check-out during their slot, they will
have to do a “late-check-out” and loose at least 5 points if checked the same day. For more info, see table below.
10) For all EE319K labs, students work in teams of two. Labs 1 and 2 will be demonstrated and collected on a first come first
served basis. For Labs 3 through 9, you will be assigned a specific checkout time. If you are late or miss your slot, your grade
will likely be lower because you will not have time to present your lab to the TA. Partners must have the same lecture and lab
time (same unique number). Both partners should signup for the same demonstration slot. For Lab 1, you are free to choose
your own partner. For Lab 2, you are free to choose your own partner, which may be different or the same as the partner in
Lab 1. For Labs 3 through 9, you will choose your permanent partner, which may be the same or different as the partners in
Labs 1 and 2. You will need a Stellaris LaunchPad (either LM4F120 or TM4C123 version) starting with Lab 2. We cannot let
you to use the LM3S1968 board, because it will not match the lectures. Each partnership needs one board, but you are
encouraged to purchase a second board as a back up (boards get lost or broken). You will keep this permanent partner for
Labs 3 through 9. For Lab 10, students may again choose any student in the same lecture time, who may or may not have
been a previous partner. In cases where there are an odd number of students in a lecture, one student may have to work alone.
If you and your partner did not get along, our first, second and third option is to help you work out your differences. Both
partners are still both responsible for the lab even if you have difficulties working together. Only under extreme cases will we
split a partnership. It is common for the TA to assign a different grade to each partner based on individual performances
during your checkout demonstration. Both partners MUST be present for your assigned time. If your partner does not show
up for your assigned time, then one student may demonstrate, and the other student (the one not present) will receive a zero
and not be allowed to make it up.
11) Grading: you will receive your lab and homework grades through the Canvas. Please make a note of which TA checked
out each of your labs. Similarly, please note which TA collected your homework. The TA that checked you out should
upload your grade within a week after the check-out. There is no need to send emails and ask about it during that week.
However, if you don't see your grade after seven days, please send an email to the TA who checked you out and cc the email
to your professor. If it is not the TA you usually check-out with, please cc your regular TA as well. The deadline for
protesting your lab or homework grade is 14 days after the lab was checked or a HW assignment turned in.
12) Please DO NOT PRINT this whole lab manual since it will be updated multiple times throughout the semester. Changes
made to the manual after the first lab is due will be marked with %%% and colored in RED or BLUE. You can search for
the three percent signs and see exactly what has changed. Therefore if you decide to print, it will be best if you print only the
part of the manual for the lab on which you are currently working.
13) Office hours: The professors and TAs have office hours. You can find the office hours of the professors on their websites.
TA office hours will be posted on Canvas and a hard copy of the TA office hour schedule will be on the wall in the EE319K
lab. Lab and TA office hours will be in ECJ1.314. We share ECJ1.314 with EE302. You can attend any office hours, even if
it is not your professor/TA.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 4
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
14) Late check-outs: In case you didn’t make it to check-out during the slot you were assigned, you still can check a lab out
later during office hours of ANY TA. There is no need to ask/register for that, just show up. If you show up at the end of a
TA office hour, there may not be time to check you out. There is a penalty of 10 points for each late day, see the table below.
“0/-5 points” means that if you check out during your scheduled time, it is possible to get full credit, but checking out later
that same day will be considered late (-5 points). No check-outs will be allowed after Tuesday following the due dates.
Students in the Thursday lab have an advantage of having their labs due later, but students in the Tuesday and Wednesday
labs have more opportunities for late checkout. If you have special circumstances, please talk with your professor about it. If
you have something that will not allow you to check-out and you know about it ahead of time, consider asking to check-out
early during office hours of ANY TA. For early checkouts, please send an email to the TA you want to check-out early with,
explain why, and cc your original TA.
Date of your lab
Tuesday
Wednesday
Thursday
Friday (none)
Tuesday
0/-5 points
Wednesday
-10 points
0/-5 points
Thursday
-20 points
-10 points
0/-5 points
Friday
-30 points
-20 points
-10 points
Monday
-45 points
-35 points
-25 points
Tuesday
Not allowed
Not allowed
-40%
15) Lab lecture: We will post the times of lab lectures on Canvas. There will be multiple such lectures for each lab starting
with Lab 2. All the lectures for a particular lab will cover the same topics. So you will have the flexibility to attend the lecture
most convenient for your schedule. The lecture(s) will take place couple of days (up to a week) before your check-out time.
During the lecture a TA will answer questions you might have, present a possible solution (how your solution is suppose to
operate - not the code). You may not take photographs of the constructed hardware shown by the TA. The TA may give some
hints and will answer your questions. Therefore we strongly recommend you read the lab manual before coming to the lecture
so you’ll have some time to think of the lab and see if you have any questions.
16) Communication: Sometimes you need to ask a specific person a specific question. For example, you might ask your TA
about a missing grade, it would be nice to email Valvano about possible typos in the book, or you might want to ask your
professor about a special circumstance. In these cases, contact the specific person. Conversely, if you have a general question
about the LM4F120/TM4C123, the lab, or the lab equipment, the best approach is to post a question on the class forum. If
you wish to email all TAs and professors, use this email address: f14_ee319k@utlists.utexas.edu.
17) The schedule of the labs is available in the course syllabus. From time to time (marked in the schedule) we will use the
lab slot to provide a demonstration of tools. During weeks with demonstrations, you are required to attend the whole hour or
as long as that session will take. Other times, when we will do lab checkouts you will have to attend your checkout slot only.
When checking out a lab you must arrive ahead of your scheduled time to set everything up.
18) As mentioned in the syllabus, you will need a wire strippers, a solderless breadboard, and a multimeter for this class. In
case you plan to buy a multimeter, there is no need to spend a lot of money, a $10 multimeter will be good enough. The NI
box used in EE302 can be used as the multimeter for this lab.
19) The software tools you will need for this class are installed in all the ECE computers, so you can work in any of these.
More specifically, the development system will work on any Windows machine with a USB port. We strongly recommend
you perform labs on your personal laptop.
20) You can download the software tools we use for the class (Keil uVision and PCB Artist) from the web, install it on you
own machine. We know that these tools were successfully used on XP, Vista, Windows 7, Windows 8, Mac* and Linux* (*using virtual machines). However, we cannot assure you will be able to run it on any OS/hardware. Therefore, some students
might have to work in the lab. Moreover, if you work on your assignment on other machine, please bring your laptop to
checkout to reduce the number of problems that could arise. I.e., it doesn't help that you swear it worked yesterday at home, if
it doesn't work during the check-out.
21) We expect every group of two students to own a Nokia 5110. There is a hardware emulation mode that can be used if
your LCD gets lost or broken.
22) If you are purchasing a Launchpad board from another student, I suggest you have a TA test the board to see if it is
functional. We do have a board tester, so if you think you might have damaged your LaunchPad, you can ask a TA to test it.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 5
23) We do pass your lab software through a plagiarism checker. Any copy/pasting from current or previous students
constitutes scholastic dishonesty. Basically, you should not be looking at solutions to EE319K labs written by other students
(other than your partner of course). See the course syllabus for more information on plagiarism.
Useful web sites
Valvano’s web page
EE319K material
Data sheets
Book examples
C programming
Setup Keil uVision
Starter files
http://users.ece.utexas.edu/~valvano
http://users.ece.utexas.edu/~valvano/Volume1 (information specific to EE319K)
http://users.ece.utexas.edu/~valvano/Datasheets/
http://users.ece.utexas.edu/~valvano/arm/
(example code from the book)
http://users.ece.utexas.edu/~valvano/embed/toc1.htm
http://users.ece.utexas.edu/~valvano/Volume1/uvision/
http://edx-org-utaustinx.s3.amazonaws.com/UT601x/EE319K_InstallSpring2015.exe
Checkout procedure
1) Get a checkout sheet. They should be available in lab. If you can’t find one in lab, you can download and print one from
http://users.ece.utexas.edu/~valvano/Volume1.
2) For Labs 1 and 2, it is first come first served. You should plan to go to your lab section or your partner’s period. For Labs
3 to 9, there are specific checkout times, for which you will sign up for in advance.
3) Before you checkout you must upload the one pdf file with all the deliverables as listed for each lab.
4) You will need to come to lab ahead of your assigned time, turn on your laptop, connect your laptop to your microcontroller
if needed and download your code.
5) The one pdf file should be open on the computer for the TA to see.
6) At the start of your checkout time, you will show the deliverables to your TA and run your program for the TA. The TA
may wish to see certain conditions, so be ready to make small changes to your programs.
7) You will be asked questions to determine your level of understanding.
8) At the end of checkout, the TA will record all parts of your lab grade on the grading sheet, so you will know your score at
that time. These sheets will not be returned. Make yourself a note of which TA checked you out; this information will assist
you if you have to correct errors in lab grades.
To checkout early or late, go to a TA office hour. TA time during office hours is first come first served.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 6
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Lab 1. Digital Logic Implemented on a Microcontroller
Preparation
Read Chapters 1 and 2 of the book
Read Sections 3.1, 3.2, 3.3.1-3.3.5, 3.3.9, 3.3.10, 4.1.2 and 4.2 of the book
Install and run Keil uVision on your personal computer
Download, unzip, open, compile, and run these four projects (either in lab or on your personal computer)
http://edx-org-utaustinx.s3.amazonaws.com/UT601x/EE319K_InstallSpring2015.exe Labs 1-4
SimpleProject_xxxasm.zip
Random number generator with no I/O
GPIO_xxxasm.zip
Simple output to Port D
Switch_xxxasm.zip
Simple input from Port A, LED on Port F
InputOutput_xxxasm.zip
More complex with inputs and outputs to Port F
Lab1_EE319K.zip
Starter project with links to grading engine
Purpose
The general purpose of this laboratory is to familiarize you with the software development steps using the uVision
simulator. Starting with Lab 2, we will use uVision for both simulation and debugging on the real board, but for this lab, we
will use just the simulator. You will learn how to perform digital input/output on parallel ports of the LM4F120/TM4C123.
Software skills you will learn include port initialization, logic operations, and unconditional branching.
System Requirements
The specific device you will create is a digital lock with two binary switch inputs and one LED output. The LED
output represents the lock, and the operator will toggle the switches in order to unlock the door. Let PE2 (Port E bit 2) be the
Boolean variable representing the lock (0 means LED is off and door is locked, 1 means LED is on and door is unlocked).
Let PE3 (Port E bit 3 is one input switch) and PE4 (Port E bit 4 is a second switch) be Boolean variables representing the
state of the two switches (1 means the switch is not pressed, and 0 means the switch is pressed). The LED should come on if
both switches are pressed. The PE2=1 if and only if PE3=0 and PE4=0. The specific function you will implement is
PE2 = (not(PE3)) and (not(PE4))
Procedure
The basic approach to this lab will be to develop and debug your system using the simulator.
Part a) To create a Lab 1 Project, perform these tasks. Find a place on your harddrive to save all your LM4F120/TM4C123
software. In Figure 1.1 it is called LaunchPadware, but it will be EE319Kware when you install the .exe.
Figure 1.1. Directory structure with your Lab1 together with driverlib and inc directories.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 7
It is important for the directory structure to look like Figure 1.1. Download and unzip the starter configuration from
http://edx-org-utaustinx.s3.amazonaws.com/UT601x/EE319K_InstallSpring2015.exe into this location. We will place our
projects in folders like my Lab 1.
Find the four example projects listed above. Begin with the Lab1_EE319K project. Notice the folder for each project is in
EE319Kware. Either double click the uvproj file or open the project from within uVision. Make sure you can compile it and
run on the simulator. Please contact your TA if the starter project does not compile or run on the simulator. Startup.s
contains assembly source code to define the stack, reset vector, and interrupt vectors. All projects in this class will include
this file, and you should not need to make any changes to the Startup.s file. main.s will contain your assembly source
code for this lab. You will edit the main.s file.
Figure 1.2. Start Keil by opening the Lab1.uvproj file.
You should rename the Lab1 starter folder to include your EIDs. Add your names and the most recent date to the comments
at the top of main.s. This code shows the basic template for the first few labs. You will not need global variables in this lab.
THUMB
AREA
DATA, ALIGN=2
;global variables go here
ALIGN
AREA
|.text|, CODE, READONLY, ALIGN=2
EXPORT Start
Start
;initialization code goes here
loop
;the body of the code goes here
B
loop
ALIGN
END
Program 1.1. Assembly language template.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 8
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
To run the Lab 1 simulator, you must do two things. First, execute Project->Options and select the Debug tab. The debug
parameter field must include -dEE319KLab1. Second, the EE319KLab1.dll file must be added to your Keil\ARM\BIN
folder. The debugging dynamic link libraries should go into Keil\ARM\BIN when you install the .exe.
Figure 1.3. Two ways to debug the software using the simulator (DCM.DLL -pCM4 -dEE319KLab1).
Part b) Write a flow chart for this program. We expect 5 to 15 symbols in the flow chart.
Part c) Write pseudo code for this program. We expect 5 to 10 steps in the pseudo code. You may use any syntax you wish,
but the algorithm should be clear.
Part d) You will write assembly code that inputs from PE3 and PE4, and outputs to PE2. You can copy and paste the address
definitions for Port E from the lm4f120.s or tm4c123gh6pm.s file (listed below).
GPIO_PORTE_DATA_R
EQU
0x400243FC
GPIO_PORTE_DIR_R
EQU
0x40024400
GPIO_PORTE_AFSEL_R
EQU
0x40024420
GPIO_PORTE_DEN_R
EQU
0x4002451C
GPIO_PORTE_AMSEL_R
EQU
0x40024528
GPIO_PORTE_PCTL_R
EQU
0x4002452C
SYSCTL_RCGC2_R
EQU
0x400FE108
The opening comments include: file name, overall objectives, hardware connections, specific functions, author name, TA,
and date. The equ pseudo-op is used to define port addresses. Global variables are declared in RAM, and the main program
is placed in EEPROM. The 32-bit contents at ROM address 0x00000004 define where the computer will begin execution
after a power is turned on or after the reset button is pressed.
;****************** main.s ***************
; Program written by: Your Names
; Date Created: 8/25/2014
; Last Modified: 8/25/2014
; Section 1-2pm
TA: Wooseok Lee
; Lab number: 1
; Brief description of the program
; The overall objective of this system is a digital lock
; Hardware connections
; PE3 is switch input (1 means switch is not pressed, 0 means switch is pressed)
; PE4 is switch input (1 means switch is not pressed, 0 means switch is pressed)
; PE2 is LED output (0 means door is locked, 1 means door is unlocked)
; The specific operation of this system
;
unlock if both switches are pressed
Program 1.2. Required comments at the top of every file.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 9
One student should write the flowchart, and the other student should write the pseudo code. Students should work together
for the other parts.
To interact with the I/O during simulation, execute the Peripherals->TExaS Port E command. When running the simulator,
we check and uncheck bits in the I/O Port box to change input pins (1) and (2). We observe output pins in the window (3).
You can also see the other DIR AFSEL DEN and PUR registers. To have the simulator test your Lab 1 solution, click the
Grade button (4). Your score will be shown in the Score box (5). The Grading Controls are for the MOOC and are not used
for any part of you grade in EE319K this semester
Figure 1.4. In simulation mode, we interact with virtual hardware.
Demonstration
During the demonstration, you will be asked to run your program to verify proper operation. You should be able to
single step your program and explain what your program is doing and why. You need to know how to set and clear
breakpoints. You should know how to visualize Port E input/output in the simulator.
Deliverables (Items 1-4 are one pdf file uploaded to Canvas, have this file open during demo.)
0) Lab 1 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
1) Flowchart of the system
2) Pseudo code for the algorithm
3) Assembly source code of your final main.s program
4) One screen shot of the Port E window, showing PE2, PE3, and PE4 showing the case with the LED on.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 10
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Lab 2. For-loops, while-loops, if-then branching, subroutines, and time delays
Preparation
Read Sections 2.7, 3.3.8 and 4.6 of the book (plus the ones listed in Lab 1)
Download, unzip, open, compile, and run the starter project, which has debugging windows configured
Lab2_EE319K.zip
Starter project with links to grading engine
Purpose
The purpose of this lab is to learn simple programming structures in assembly. You will also learn how to estimate
how long it takes to run software, and use this estimation to create a time delay function. You learn how to use the
oscilloscope to measure time delay. Assembly software skills you will learn include masking, toggling, if-then, subroutines,
and looping.
System Requirements
The system has one input switch and one output LED. Figure 2.1 shows the system when simulated as the switch is
touched. A negative logic switch means the PF4 signal will be 1 (high, 3.3V) if the switch is not pressed, and the PF4 signal
will be 0 (low, +0V) if the switch is pressed. A positive logic blue LED interface means if the software outputs a 1 to PF2
(high, +3.3V) the LED will turn ON, and if the software outputs a 0 to PF2 (low, 0V) the blue LED will be OFF. In Lab 3,
you will attach a real switch and and real LED to your protoboard, and interface them to your microcontroller. In Lab 2 you
first debug in simulation and then run on the real board, but no external hardware will be required. Overall functionality of
this system is described in the following rules.
1) Make PF2 an output and make PF4 an input (enable PUR for PF4).
2) The system starts with the LED ON (make PF2 =1).
3) Delay for about 1 ms
4) If the switch is pressed (PF4 is 0), then toggle the LED once, else turn the LED ON.
5) Repeat steps 3 and 4 over and over
Figure 2.1. Example screenshots in simulation mode with the switch pressed then released. Delta=1.002ms.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 11
Time is very important in embedded systems. One of the simplest ways in which we manage time is determining
how long it takes to run our software. There are two ways to determine how long each instruction takes to execute. The first
method uses the ARM data sheet. For example, the following is a page from the Cortex-M4 Technical Reference Manual.
E.g., see pages 34-38 of http://users.ece.utexas.edu/~valvano/EE345L/Labs/Fall2011/CortexM4_TRM_r0p1.pdf
Figure 2.2. From the Technical Reference Manual page 36.
On the TM4C123 the default bus clock is 16 MHz ±1%. Starting in Lab 4 we will activate the phase lock loop (PLL) and the
bus clock will be exactly 80 MHz. For now, however, we will run at about 16 MHz. The following is a portion of a listing
file with a simple delay loop. The SUBS and BNE are executed 800 times. The SUBS takes 1 cycle and the BNE takes 1 to 3
(a branch takes 0 to 3 cycles to refill the pipeline). The minimum time to execute this code is 800*(1+1)/16 µs = 100 µs. The
maximum time to execute this code is 800*(1+4)/16 µs = 250 µs. Since it is impossible to get an accurate time value using
the cycle counting method, we will need another way to estimate execution speed. An accurate method to measure time uses
a logic analyzer or oscilloscope. In the simulator, we will use a simulated logic analyzer, and on the real board we will use an
oscilloscope. To measure execution time, we cause rising and falling edges on a digital output pin that occur at know places
within the software execution. We can use the logic analyzer or oscilloscope to measure the elapsed time between the rising
and falling edges. In this lab we will measure the time between edges on output PF2.
0x00000158 F44F7016
MOV R0,#800
0x0000015C 3801
wait SUBS R0,R0,#0x01
0x0000015E D1FD
BNE wait
(note: the BNE instruction executes in 3 cycles on the simulator, but in 2 cycles on the real board)
Procedure
The basic approach to Lab 2 will be to develop and debug your system using the simulator using a negative switch
(PF4) and a positive logic LED (PF2). In Lab 3, you will build and test an actual switch and LED. This lab will run on the
LaunchPad using SW1 on PF4 and the blue LED on PF2.
To run the Lab 2 simulator, you must do two things. First, execute Project->Options and select the Debug tab. The
debug parameter field must include -dEE319KLab2. Second, the EE319KLab2.dll file must be added to your
Keil\ARM\BIN folder.
Figure 2.3. Using TExaS to debug your software in simulation mode (DCM.DLL -pCM4 -dEE319KLab2).
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 12
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Part a) Design a subroutine that delays about 1 ms. First, draw a flowchart, and then write pseudocode. You will call the
subroutine with a BL instruction and return with a BX LR instruction. Any delay from 0.8 to 1.2 ms is ok. To implement a
delay, you could set a register to a large number, and then count it down to zero. With a bus clock of 16 MHz, there are
16,000 bus clock cycles in 1 ms. You need to know how long it takes to execute the loop once, then determine the number of
times you need to execute the loop to create the 1 ms delay. E.g., if the time to execute the loop once is 4 cycles, then
executing the loop 4000 times will be about 1 ms. In this simple estimation we are neglected the other instructions not inside
the loop, because they are 4000 times less important.
Part b) Write a main program in assembly that implements the input/output
system. Pseudo code and flowchart are shown, illustrating the basic steps
for the system. We recommend at this early stage of your design career you
access the entire I/O port using GPIO_PORTF_DATA_R. After you fully
understand how I/O works, then you can use bit-specific addressing to
access I/O ports.
main
Initialize
ports
LED on
Turn on the clock for Port F
Set the Port F direction register so
PF4 is an input and
Delay
PF2 is an output
Enable the PF4 and PF2 bits in the Port F DEN register
Set bit 4 in Port F PUR register so it will run on the real board
Not pressed
Pressed
Start
Set PF2 so the LED is ON
loop Delay about 1ms
LED on
LED toggle
Read the switch and test if the switch is pressed
If PF4=0 (the switch is pressed),
toggle PF2 (flip bit from 0 to 1, or from 1 to 0)
If PF4=1 (the switch is not pressed),
set PF2, so LED is ON
Go to loop
Part c) Test the program in simulation mode. Use the built-in logic analyzer to verify the LED is toggling at the rate at which
it was designed. In particular, capture two screenshots like Figure 2.1 showing when the switch is pressed, the LED is ON for
1 ms and OFF for 1 ms.
Part d) Simulators typically run slower than real hardware. Use the built in logic analyzer to measure how much Cortex-M
time is simulated in 10 seconds. Hit the reset twice to clear the time axis on the plot. Run the simulator for 10 human seconds
(real time with your watch), and then stop simulation. Obverse the logic analyzer time to determine the amount of Cortex-M
simulated. For example, my computer simulated 15 sec of Cortex-M time in 10 sec of human time, meaning the simulator
was running 150% faster than a real Cortex-M. There are many factors that affect this ratio, so do expect to see this ratio
vary a lot. The point of the exercise is to get a sense of how a simulator manages time.
Part e) Load your software onto the board and test it again. If the PF4 switch is not pressed, then the blue LED is on, which
you can see by noticing its bright color. If the PF4 switch is pressed, then the blue LED toggles 500 times a second, which
you can see by noticing its dim blue color.
main
Demonstration (both partners must be present, and demonstration grades for partners may be different)
You will show the TA your program operation on both the simulator and the real board. Be prepared to explain how
the delay function works. How would it be different if it were 100 ms instead of 1 ms? The TA will pick an instruction from
your code and ask you which addressing mode it uses. Execute the program step-by-step and run to cursor. What is a Reset
Vector and why do you need it? What does AREA do? What does EXPORT do?
Deliverables (Items 1-5 are one pdf file uploaded to Canvas, have this file open during demo.)
0) Lab 2 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
1) Two screenshots, like Figure 2.1, showing the system running on the simulator. One showing the touch and
release, and the other showing a close up of the toggling
2) Flowchart of the delay function
3) Pseudocode of the delay function
4) Assembly source code of your final program
5) Measurement of how much microcontroller time is simulated in 10 seconds of actual time.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 13
Lab 3. Switch and LED interfacing
Preparation
Read Sections 2.7, 3.3.8, and 4.2
Download
EE319K_TM4C123_artist.sch
PCB artist example drawing with all EE319K components
Lab3_EE319K.zip
Starter project with links to grading engine
Purpose
Lab 3 is the continuation of Lab 2; essentially, you will implement in hardware what you developed in simulation
back in Lab 2. The purpose of this lab is to learn how to interface a switch and an LED. You will perform explicit
measurements on the circuits in order to verify they are operational and to improve your understanding of how they work.
You will learn to use a free design tool called PCBArtist from Advanced Circuits for drawing electrical circuits.
This tool will be used in EE319K, EE445L and EE445M. To download this tool, go to http://www.4pcb.com/ When
downloading this free application from the internet, be careful NOT to accept any of the special offers (when asked yes or no,
say NO.) When installed, there will be a drawing application called PCB Artist, which we will use to draw electrical circuits.
Open the starter file EE319K_TM4C123_artist.sch using PCB Artist. We will also use PCB Artist to design PCB boards in
EE445L.
System Requirements
Figure 4.13 in the book shows a switch interface that implements positive logic. In this lab you will connect it to
PE0. The right side of Figure 4.14 shows an LED interface that implements positive logic. You will attach this switch and
LED to your protoboard (the white piece with all the holes), and interface them to your TM4C123. Overall functionality of
this system is the similar to Lab 2, with five changes: 1) the pin to which we connect the switch is moved to PE0, 2) you will
have to remove the PUR initialization because pull up is no longer needed. 3) the pin to which we connect the LED is moved
to PE1, 4) the switch is changed from negative to positive logic, and 5) you should increase the delay so it flashes about 8 Hz.
1) Make PE1 an output and make PE0 an input.
2) The system starts with the LED on (make PE1 =1).
3) Wait about 62 ms
4) If the switch is pressed (PE0 is 1), then toggle the LED once, else turn the LED on.
5) Steps 3 and 4 are repeated over and over
Procedure
Back in Lab 2, you developed and debugged your system using the simulator, and then used a built-in switch and
LED to run on the real board. During this lab you will build and test your hardware. The final demonstration will be run
stand-alone without connection to the PC. This stand-alone operation illustrates how embedded systems operate. Whenever
we call TExaS_Init, we activate the phase lock loop (PLL) and run at 80 MHz. %%%
To run the Lab 3 simulator, you must do two things. First, execute Project->Options and select the Debug tab. The debug
parameter field must include -dEE319KLab3. Second, the EE319KLab3.dll file must be added to your Keil\ARM\BIN
folder. The Grading Controls are for the MOOC and are not used as actual grades in EE319K this semester.
Figure 3.1. Using TExaS to debug your software in simulation mode (DCM.DLL -pCM4 -dEE319KLab3).
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 14
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Part a) Make a copy of the EE319K_TM4C123_artist.sch file and rename it with your EID and Lab3. Edit this file to show
the interface for the switch and the LED. When you are done, print the circuit as a pdf file, and email this pdf file to your TA
for approval. If you wish to build something different from Figure 3.3, please have a TA check your design. Your circuit
should look similar to book Figure 4.13 and the right side of Figure 4.14. The first step is to click and drag components you
will need so they are close to each other
Figure 3.2. PCB Artist drawing showing Port E, 3.3V power, a 10kΩ resistor, a switch, one 7406 gate, an LED, a 220Ω
resistor, a 0.1 µF capacitor, +5V power and ground.
The second step is to add traces (these will be actual wires when you build it). In general, we draw circuit diagrams so current
flows from top to bottom on the page.
Figure 3.3. PCB Artist drawing one possible Lab 3 drawing. When building the circuit, the +3.3V comes from the +3.3V
signal on the LaunchPad. The +5V comes from the VBUS signal on the LaunchPad. The LaunchPad ground must be
connected to the ground signals on your external circuit.
In the left side of Figure 3.4, wire1 and wire2 cross in the circuit diagram, but are not electrically connected, because there is
NO dot at the point of crossing. In the right side of Figure 3.4, wire1 and wire2 cross in the circuit diagram, and are
electrically connected, because there IS a dot at the point of crossing.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 15
Figure 3.4. When traces cross without a dot they are not connected. When traces cross with a dot they are connected.
Part b) Make the five changes listed in the requirements section to your Lab 2 system. The best approach to time delay will
be to use a hardware timer, which we will learn in Lab 4. In this lab, however, we do not expect you to use the hardware
timer. Again, use the logic analyzer on the simulator to verify the software operates properly. In this lab the delay was
increased so you will be able to see the LED flash with your eyes.
Part c) Engineers must be able to read datasheets during the design, implementation and debugging phases of their projects.
During the design phase, datasheets allow us to evaluate alternatives, selecting devices that balance cost, package size,
power, and performance. For example, we could have used other IC chips like the 7405, 74LS05, or 74HC05 to interface the
LED to the TM4C123. In particular, we chose the 7406 or 74LS06 because it has a large output current (IOL = 40 mA), 6
drivers, and is very inexpensive (59¢). During the implementation phase, the datasheet helps us identify which pins are
which. During the debugging phase, the datasheet specifies input/output parameters that we can test. Download the 7406 and
LED datasheets from the web
http://users.ece.utexas.edu/~valvano/Datasheets/7406.pdf
http://users.ece.utexas.edu/~valvano/Datasheets/LED_red.pdf
and find in the datasheet for the 7406 the two pictures as shown in Figure 3.5. Next, hold your actual 7406 chip and identify
the location of pin 1. Find in the datasheet the specification that says the output low voltage (VOL) will be 0.4V when the
output low current (IOL) is 16 mA (this is close to the operating point we will be using for the LED interface).
Figure 3.5. Connection diagram and physical package diagram for the 7406.
Using the data sheet, hold an LED and identify which pin is the anode and which is the cathode. Sometimes we are asked to
interface a device without a data sheet. Notice the switch has 4 pins in a rectangular shape, as shown in Figure 3.6. Each
button is a single-pole single-throw normally-open switch. All four pins are connected to the switch. Using your ohmmeter
determine which pairs of pins are internally connected, and across which pair of pins is the switch itself. In particular, draw
the internal connections of the switch, started in Figure 3.6, showing how the four pins are connected to the switch.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 16
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Figure 3.6. Connection diagram for the normally open switch.
To build circuits, we’ll use a solderless breadboard, also referred to as a protoboard. The holes in the protoboard are
internally connected in a systematic manner, as shown in Figure 3.7. The long rows of of 50 holes along the outer sides of the
protoboard are electrically connected. Some protoboards like the one in Figure 3.7 have four long rows (two on each side),
while others have just two long rows (one on each side). We refer to the long rows as power buses. If your protoboard has
only two long rows (one on each side, we will connect one row to +3.3V and another row to ground. If your protoboard has
two long rows on each side, then two rows will be ground, one row will be +3.3V and the last row will be +5V (from VBUS).
Use a black marker and label the voltage on each row. In the middle of the protoboard, you’ll find two groups of holes placed
in a 0.1 inch grid. Each adjacent row of five pins is electrically connected. We usually insert components into these holes. IC
chips are placed on the protoboard, such that the two rows of pins straddle the center valley. To make connections to the
TM4C123 we can run male-male solid wire from the bottom of the microcontroller board to the protoboard. For example,
assume we wish to connect TM4C123 PE1 output to the 7406 input pin 1. First, cut a 24 gauge solid wire long enough to
reach from PE1 and pin 1 of the 7406. Next, strip about 0.25 inch off each end. Place one end of the wire in the hole for the
PE1 and the other end in one of the four remaining of the 7406 pin 1.
If you are unsure about the wiring, please show it to your TA before plugging in the USB cable. I like to place my
voltmeter on the +3.3V power when I first power up a new circuit. If the +3.3V line doesn’t immediately jump to +3.3V, I
quickly disconnect the USB cable.
Figure 3.7. All the pins on each of the four long rows are connected. The 5 pins in each short column are connected. Use
male-male wires to connect signals on the LaunchPad to devices on the protoboard. Make sure ground wire connected
between the LaunchPad and your circuit. The +3.3V and VBUS (+5V) power can be wired from the LaunchPad to your
circuit. I like to connect the two dark blue rows to ground, one red row to +3.3V and the other red row to VBUS(+5V).
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 17
Part d) After the software has been debugged on the simulator, you will build the hardware on the real board. Please get your
design checked by the TA before you apply power (plug in the USB cable). Do not place or remove wires on the protoboard
while the power is on.
Before connecting the switch to the microcomputer, please take the measurements in Table 3.1 using your digital multimeter.
The input voltage (VPE0) is the signal that will eventually be connected to PE0. With a positive logic switch interface, the
resistor current will be VPE0/10kΩ. The voltages should be near +3.3 V or near 0 V and the currents will be less than 1 mA.
The goal is to verify the VPE0 voltage is low when the switch is not pressed and high when the switch is pressed.
The circuit in Figures 3.1 and 3.2 used R1 as the 10kΩ resistor. There are other 10k resistors in the PCB artist starter file; any
one of which could have been used.
Parameter
Resistance of the
10kΩ resistor, R1
Value
Units
ohms
Supply Voltage, V+3.3
volts
Input Voltage, VPE0
volts
Resistor current
mA
Input Voltage, VPE0
volts
Resistor current
mA
Conditions
with power off and
disconnected from circuit
(measured with ohmmeter)
Powered
(measured with voltmeter)
Powered, but
with switch not pressed
(measured with voltmeter)
Powered, but switch not pressed
I=VPE0/R1 (calculated and
measured with an ammeter)
Powered and
with switch pressed
(measured with voltmeter)
Powered and switch pressed
I=VPE0/R1 (calculated and
measured with an ammeter)
Table 3.1. Switch measurements.
Next, you can connect the input voltage to PE0 and use the debugger to observe the input pin to verify the proper operation
of the switch interface. You will have to single step through the code that initializes Port E, and PE0. You then execute the
Peripherals->TEXaS Port E command. As you single step you should see the actual input as controlled by the switch you
have interfaced, see Figure 3.1.
The next step is to build the LED output circuit. LEDs emit light when an electric current passes through them, as
shown in Figure 3.8. LEDs have polarity, meaning current must pass from anode to cathode to activate. The anode is labeled
a or + , and cathode is labeled k or -. The cathode is the short lead and there may be a slight flat spot on the body of round
LEDs. Thus, the anode is the longer lead. LEDs are not usually damaged by heat when soldering. Furthermore, LEDs will
not be damaged if you plug it in backwards. LEDs however won’t work plugged in backwards of course. Look up the pin
assignments in the 7406 data sheet. Be sure to connect +5V power to pin 14 and ground to pin 7. The 0.1 µF capacitor from
+5V to ground filters the power line. Every digital chip (e.g., 7406) should have a filter capacitor from its power line (i.e., pin
14 VCC) to ground. The capacitor in your kit is ceramic, which is not polarized, meaning it can be connected in either
direction.
Anode
Photo:
Cathode
Circuit symbol:
a +
k -
anode
cathode
Figure 3.8. LEDs.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 18
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
The circuit in Figures 3.2 and 3.3 used R10 as the 220Ω resistor. There are six 220Ω resistors R10 – R15 in the PCB artist
starter file, any one of which could have been used.
Take the measurements as described in Table 3.2. The R10 measurement occurs before R10 is inserted into the circuit.
Single step your software to make PE1 to output. Initially PE1 will be low. So take four measurements with PE1 low, rows
2,3,4,5 in Table 3.2. Then, single step some more until PE1 is high and measure the three voltages (rows 8,9,10 in Table 3.2).
When active, the LED voltage should be about 2 V, and the LED current should be about 10 mA. The remaining rows are
calculated values, based on these 8 measurements. The LED current (row 12) can be determined by calculation or by direct
measurement using the ammeter function. You should perform both ways to get LED current.
Part e) Debug your combined hardware/software system on the actual TM4C123 board.
Warning: NEVER INSERT/REMOVE WIRES/CHIPS WHEN THE POWER IS ON.
Row
1
Parameter
Resistance of the
220Ω resistor, R10
Value
Units
2
+5 V power supply
V+5
volts
3
TM4C123 Output, VPE1
input to 7406
volts
with PE1 = 0
(measured with voltmeter)
4
7406 Output, VkLED k-
volts
with PE1 = 0
(measured with voltmeter)
5
LED a+, Va+
Bottom side of R10
volts
with PE1 = 0
(measured with voltmeter)
6
LED voltage
volts
calculated as Va+ - Vk-
7
LED current
mA
calculated as (V+5 - Va+)/R10
and
measured with an ammeter
8
TM4C123 Output, VPE1`
input to 7406
volts
with PE1 = 1
(measured with voltmeter)
9
7406 Output, VkLED k-
volts
with PE1 = 1
(measured with voltmeter)
10
LED a+, Va+
Bottom side of R10
volts
with PE1 = 1
(measured with voltmeter)
11
LED voltage
volts
calculated as Va+ - Vk-
12
LED current
mA
calculated as (V+5 - Va+)/R10
and
measured with an ammeter
ohms
Table 3.2. LED measurements (assuming the 220 Ω resistor is labeled R10).
08/09/15
Conditions
with power off and
disconnected from circuit
(measured with ohmmeter)
(measured with voltmeter,
notice that the +5V power is
not exactly +5 volts)
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 19
Demonstration (both partners must be present, and demonstration grades for partners may be different)
You will show the TA your program operation on the actual TM4C123 board. The TA may look at your data and
expect you to understand how the data was collected and how the switch and LEDs work. Also be prepared to explain how
your software works and to discuss other ways the problem could have been solved. Why the 7406 was used to interface the
LED? I.e., why did we not connect the LED directly to the TM4C123. Why do you think you need the capacitor for 7406
chip? Why was the delay increased from 1 to 62 ms? How would you modify the software to change the rate at which LED
flickers? What operating point (voltage, current) exists when the LED is on? Sketch the approximate current versus voltage
curve of the LED. Explain how you use the resistor value to select the operating point. What is the difference between a
positive logic and negative logic interface for the switch or the LED? We may test to see if you can measure voltage, current
and/or resistance with your meter (so bring your meter to the demonstration).
Deliverables (Items 1, 2, 3, 4, 5 are one pdf file uploaded to Canvas, have this file open during demo.)
0) Lab 3 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
1) Circuit diagram, using PCBArtist, or hand drawn
2) Screenshot like Figure 3.9 showing your debugging in the simulator
3) Switch measurements (Table 3.1)
4) LED measurements (Table 3.2)
5) Assembly source code of your final program
Precautions to avoid damaging your system
0) Do not attach or remove wires on the protoboard when power is on. Always shut power off when making hardware
changes to the system.
1) Touch a grounded object before handling CMOS electronics. Try not to touch any exposed wires.
2) Do not plug or unplug the modules into the LaunchPad while the system is powered.
3) Do not use the TM4C123 with any external power sources, other than the USB cable. In particular, avoid connecting
signals to the TM4C123 that are not within the 0 to +5V range. In particular, voltages less than 0V or greater than +5V will
damage the microcontroller.
4) Do not use PC3,PC2,PC1,PC0. These are needed for the debugger.
5) You can’t use PA1 PA0 PD4 and PD5. These are connected to the serial port and USB.
6) See Figure 4.8 in the textbook. If you use both PD0 and PB6, remove R9 from your board. If you use both PD1 and PB7,
remove R10 from your board.
Figure 3.9. Simulation of Lab 3, showing PE0 high, and the PE1 output toggling at 8 Hz.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 20
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Lab 4. Minimally Intrusive Debugging Methods
Preparation
Read Sections 4.3, 4.4, 5.1, 5.2, 5.5, 5.7, 6.1, 6.2, and 6.9
Download, unzip, open, compile, and run the project
SysTick_xxxasm.zip
PLL and SysTick functions
Lab4_EE319K.zip
Starter project with links to grading engine
These videos https://www.youtube.com/playlist?list=PLyg2vmIzGxXHYEHQrxNxGcRg6vCTB20Ud relate to Lab 4.
Purpose
The purpose of this lab is to learn minimally intrusive debugging skills. When visualizing software running in realtime on an actual microcomputer, it is important use minimally intrusive debugging tools. The first objective of this lab is to
develop an instrument called a dump, which does not depend on the availability of a debugger. A dump allows you to
capture strategic information that will be viewed at a later time. Many programmers use the printf statement to display useful
information as the programming is being executed. On an embedded system we do not have the time or facilites to use printf.
Fortunately, we can use a dump in most debugging situations for which a printf is desired. The second useful debugging
technique you will learn is called a heartbeat. A heartbeat is a visual means to see that your software is still running. The
debugging techniques in this lab use both hardware and software and are appropriate for the real TM4C123. Software skills
you will learn include indexed addressing, array data structures, the PLL, the SysTick timer, and subroutines.
System Requirements
In this lab, you will design, implement, test and employ software debugging instruments to experimentally verify the correct
operation of your Lab 3 system. For the software approach, you will record the Port E value (includes both the input and
output signals) and the time during each execution through the outer loop of the main program of Lab 3 as your system runs
in real time. You will activate the PLL (by calling the TExaS_Init function, which in turn will execute Program 4.6 of the
book) to make the microcontroller run at 80 MHz. You will also activate the SysTick timer (call SysTick_Init shown in
Program 4.7 of the book), which will make the 24-bit counter decrement every 12.5 ns. We will use this counter
(NVIC_ST_CURRENT_R) to measure time differences up to 335.5 ms. To measure the current time, you simply read the 24bit NVIC_ST_CURRENT_R value. This software dump should store Port E and NVIC_ST_CURRENT_R data into arrays
while the system is running in real-time, and the information will be viewed at a later time. Software dumps are an effective
technique when debugging real-time software on an actual microcomputer. We will define the debugging instrument as
minimally intrusive if the time to collect and store the information is short compared to the time between when information
is collected. You will define an array capable of storing about 3 seconds worth of Port E measurements, and an array capable
of storing about 3 seconds worth of time measurements. For example, if the outer loop of your Lab 3 executes in about 62
ms, then the loop will be executed about 50 times in 3 seconds (3000/62), so the array sizes will need to be 50 elements each.
You may use either pointers or counters to store data in the arrays. You will collect performance data on the system with a
no-touch, touch, no-touch sequence during the 3-second measurement. Figure 4.1 shows how to configure the Lab 4
simulator. The Grading Controls are for the MOOC and are not used in EE319K this semester.
Figure 4.1. Using TExaS to debug your software in simulation mode (DCM.DLL -pCM4 -dEE319KLab4).
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 21
PF2 is already interfaced to an LED on the TM4C123 board itself. Write debugging software and add it to the Lab 3
system so that this LED always flashes while your program is running. In particular, initialize the direction register so PF2 is
an output, and add code that toggles the LED each time through the loop. A heartbeat of this type will be added to all
software written for Labs 4, 6, 7, 8, and 9 in this class. Lab 5 will use PF2 as part of the system output. It is a quick and
convenient way to see if your program is still running. When you get to EE445L you can implement multiple heartbeats at
various strategic places in the software and that toggle much faster than the eye can see. In these situations, you will use a
logic analyzer to visualize many high-speed digital signals all at once. However, in EE319K there will be one heartbeat on
PF2, and the heartbeat must occur slow enough to be seen with the eye.
Procedure
The basic approach to Lab 3 through Lab 9 will be to first develop and test the system using simulation mode. After
debugging instruments themselves are tested, you will collect measurements on the real TM4C123.
Part a) When your main program calls TExaS_Init, this subroutine will activate the PLL making bus clock 80 MHz.
Adjust the delay function so it delays 62 ms. Write two debugging subroutines that implement a dump instrument. If we
saved just the input/output data, then the dump would be called functional debugging, because we would capture input/output
data of the system, without information specifying when the input/output was collected. However, you will save both the
input/output data and the time, so the dump would be classified as performance debugging. The first subroutine
(Debug_Init) initializes your dump instrument. The initialization should activate the SysTick timer, place 0xFFFFFFFF
into the two arrays, and initialize pointers and/or counters as needed. The second subroutine (Debug_Capture) that saves
one data-point (PE0 input data, and PE1 output data) in the first array and the NVIC_ST_CURRENT_R in the second array.
Since there are only two bits to save in the first array, pack the information into one value for ease of visualization when
displayed in hexadecimal. Put the PE0 value into bit 4 and the PE1 value into bit 0. I.e.,
Input (PE0)
Output (PE1)
save data
0
0
0000,00002, or 0x00000000
0
1
0000,00012, or 0x00000001
1
0
0001,00002, or 0x00000010
1
1
0001,00012, or 0x00000011
In this way, you will be able to visualize the data in an effective manner. Place a call to Debug_Init at the beginning of the
system, and a call to Debug_Capture at the start of each execution of the outer loop. The basic steps involved in designing
the data structures for this debugging instrument are as follows
Allocate DataBuffer in RAM (to store 3 seconds of input/output data)
Allocate TimeBuffer in RAM (to store 3 seconds of timer data)
Allocate two pointers (DataPt, TimePt), one for each array, pointing to the place to save the next data
The basic steps involved in designing Debug_Init are as follows
Set all entries of the first buffer to 0xFFFFFFFF (meaning no data yet saved)
Set all entries of the second buffer to 0xFFFFFFFF (meaning no data yet saved)
Initialize the two pointers to the beginning of each buffer
Activate the SysTick timer (call SysTick_Init shown in Program 4.7 of the book)
The basic steps involved in designing Debug_Capture are as follows
Save any registers needed
Return immediately if the buffers are full (pointer past the end of the buffer)
Read Port E and the SysTick timer (NVIC_ST_CURRENT_R)
Mask capturing just bits 1,0 of the Port E data
Shift the Port E data bit 0 into bit 4 position, and bit 1 into bit 0 position
Dump this port information into DataBuffer using the pointer DataPt
Increment DataPt to next address
Dump time into TimeBuffer using the pointer TimePt
Increment TimePt to next address
Restore any registered saved and return
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 22
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
For regular functions we are free to use R0, R1, R2, R3, and R12 without preserving them. However, for debugging
instruments, we should preserve all registers, so that the original program is not affected by the execution of the debugging
instruments. The temporary variables may be implemented in registers. However, the buffers and the pointers should be
permanently allocated in global RAM. You can observe the debugging arrays using a Memory window. Look in the map file
to find the addresses of the buffers. After you have debugged your code in simulation mode, capture a screenshot showing the
results as the switch is pressed then not pressed. The dumped data should start with some 0x01 data, next it should oscillate
between 0x10, 0x11 as the switch is pressed, then return to 0x01 when the switch is released.
Part b) One simple way to estimate the execution speed is to assume each instruction requires about 2 cycles. By counting
instructions and multiplying by two, you can estimate the number of cycles required to execute your Debug_Capture
subroutine. Assuming the 12.5 ns bus cycle time, convert the number of cycles to time. Next, estimate the time between calls
to Debug_Capture. Calculate the percentage overhead required to run the debugging instrument (100 times execution
time divided by time between calls, in percent). This percentage will be a quantitative measure of the intrusiveness of your
debugging instrument. Type the estimations and calculations in as comments to your program.
Part c) Write the debugging instruments that implement the heartbeat (initialization and toggle). The heartbeat instruments
can be embedded as simple assembly language software without making them separate subroutines.
Part d) Debug your combined hardware/software system first with the simulator, then on the actual TM4C123 board.
Figure 4.2 is screenshot of Lab 4 in simulation mode. You can find the address of the buffers by looking in the map file. You
can dump the memory to a file using the command (the data is formatted in a little-endian hexadecimal format)
SAVE data.txt 0x20000000 , 0x20000190
Figure 4.2. Simulation output showing the input, output and heartbeat.
Let’s look at one measurement 0x00ABE5DA-0x0057CDB4 = 11265498-5754292=5511206. Since each cycle is 12.5ns,
5511206*12.5ns = 69ms. For this solution, the measurement agrees with the data collected by an oscilloscope and shown in
Figure 4.4.
Figure 4.3. Similar data is observed in the memory window in simulation and on the real board showing results of the dump.
Part e) The input is on PE0 and the output is on PE1. PF2 is a heartbeat. Run your debugging instrument capturing the
sequence of PE0 inputs and PE1 outputs as you touch, then release the switch. Use the collected data to measure the period
of the flashing LED. Your measurement should be accurate to 12.5 ns.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 23
Demonstration (both partners must be present, and demonstration grades for partners may be different)
You will show the TA your program operation on the actual TM4C123 board. You should be able to connect PF2 to an
oscilloscope to verify the main loop is running at about every 62 ms. The TA may look at your data and expect you to
understand how the data was collected and what the data means. Also be prepared to explain how your software works and to
discuss other ways the problem could have been solved. Why did you have to change the delay function after the PLL was
activated? How did you change it? Why? The TA will pick an instruction in your program and ask how much time does it
take that instruction to execute in µsec. Does it always take same amount of time to execute? You will be asked to create a
breakpoint, and add the port pin to the simulated logic analyzer. Is Debug_Capture minimally intrusive or non-intrusive?
What do you mean by intrusiveness? Is your code “friendly”? How do you define masking? How do you set/clear one bit in
without affecting other bits? What is the difference between the B, BL and BX instructions? How do you initialize the
SysTick? You should understand every step of the function SysTick_Init. How do you change the rate at which SysTick
counts? Describe three ways to measure the time for a software function to execute? How do you calculate the sizes of the
port data and the timestamp data? If you used 32-bit data for DataBuffer what would be the advantages of 8-bit data? If
you used 8-bit data for DataBuffer what would be the advantages of 32-bit? Could you have stored the time-stamp data in
8-bit, 16-bit, or 24-bit arrays? Why does the pointer to the time-stamp array need to be incremented by four, if you want to
point to the next element in the array? How do you allocate global variables?
Deliverables (Items 1, 2, 3, 4 are one pdf file uploaded to Canvas, have this file open during demo.)
0) Lab 4 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
1) A screenshot showing the system running in simulation mode. In the screenshot, please show the dumped data in
a memory window and the I/O window, as illustrated in Figures 4.1 and 4.2
2) Assembly listing of your final program with both the dump and heartbeat instruments
3) Estimation of the execution time of your debugging instrument Debug_Capture (part b)
4) Results of the debugging instrument (part e) and the calculation of the flashing LED period in msec.
How to transfer data from debugger to a computer file
1) Run your system so data is collected in memory, assume interesting stuff is from 0x20000000 to 0x20000190
2) Type SAVE data.txt 0x20000000 , 0x20000190 in the command window after the prompt (“>”), type enter
3) Open the data.txt file in NotePad, it looks like this
:020000042000DA
:1000000001000000010000000100000001000000EC
:1000100001000000010000000100000001000000DC
:1000200001000000010000000100000001000000CC
:1000300001000000010000000100000001000000BC
:10004000010000001000000011000000100000007E
:10005000110000001000000011000000100000005E
:10006000110000001000000011000000100000004E
:10007000110000001000000011000000100000003E
:10008000110000001000000011000000100000002E
:10009000110000000100000001000000010000004C
:1000A000010000000100000001000000010000004C
:1000B000010000000100000001000000010000003C
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 24
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
:1000C000010000000100000000FEFF00DAE5AB00C7
:1000D000B4CD57008EB50300689DAF0042855B002C
:1000E0001C6D0700F654B300D03C5F00AA240B003F
:1000F000840CB7005EF4620038DC0E0012C4BA0053
:10010000ECAB6600C6931200A07BBE0072636A006F
:10011000444B16001633C200E81A6E00BA021A00E9
:100120008CEAC5005ED2710030BA1D0002A2C9007F
:10013000D4897500A67121007859CD004A41790013
:100140001C292500EE10D100C0F87C0092E02800A8
:1001500064C8D40036B0800008982C00E27FD80034
:10016000BC678400964F30007037DC004A1F88005F
:1001700024073400FEEEDF00D8D68B00B2BE370075
:100180008CA6E300668E8F0040763B001A5EE70087
:01019000C8A6
:00000001FF
4) Strip off the first 9 characters of every line, and the last two characters of every line. Delete the first and last lines, leaving
the data shown above in bold. Each two characters is a byte in hex. 32-bit and 16-bit data are of course little endian.
Figure 4.4. The input and output signals are measured with a two channel oscilloscope. When the switch is pressed, PE0
goes high. When the software sees PE0 high it toggles PE1 every 62ms. In this solution with system actually toggled PE1
every 69 ms.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 25
Lab 5. Traffic Light Controller
Preparation
Read Sections 4.4, 6.1, 6.4, 6.5 in the textbook
See http://users.ece.utexas.edu/~valvano/Datasheets/LED_yellow.pdf
See http://users.ece.utexas.edu/~valvano/Datasheets/LED_green.pdf
Download, unzip, build and simulate the FSM example C10_TableTrafficLight.zip
PointerTrafficLight_xxx.zip edXLab10.dll and Lab5_EE319K.zip.
https://www.youtube.com/playlist?list=PLyg2vmIzGxXEle4_R2VA_J5uwTWktdWTu
Purpose
This lab has these major objectives: 1) the understanding and implementing linked data structures; 2) learning how
to create a segmented software system; and 3) the study of real-time synchronization by designing a finite state machine
controller. Software skills you will learn include advanced indexed addressing, linked data structures, creating fixed-time
delays using the SysTick timer, and debugging real-time systems.
System Requirements
All software in this lab must be developed in C. Keil uVision4 projects in C are a little different than the assembly
projects used for Labs 1 to 4. The project configuration and the startup.s file are different. The Lab 5 starter file has the
appropriate connections to the Lab 5 simulator/grader (edXLab10.dll). The call to TExaS_Init will activate 80 MHz PLL.
Consider a 4-corner intersection as shown in Figure 5.1. There are two one-way streets are labeled South (cars travel
South) and West (cars travel West). There are three inputs to your TM4C123, two are car sensors, and one is a walk button.
The South sensor will be true (1) if one or more cars are near the intersection on the South road. Similarly, the West sensor
will be true (1) if one or more cars are near the intersection on the West road. The Walk button will be pushed by a pedestrian
when he or she wishes to cross in any direction. To request a walk, a pedestrian must push and hold the walk button for at
least 2 seconds, after which person can release the button, and the walk request will be remembered. In contrast, when a car
sensor is released, it means no cars are waiting to enter the intersection. You will use 8 outputs from your microcomputer that
control the traffic lights, the “walk” light, and the “don’t walk” light. The “walk” light will be the green LED (PF3) on the
LaunchPad and the “don’t walk” light is the red LED (PF1) on the LaunchPad. When the “walk” condition is signified,
pedestrians are allowed to cross. When the “don’t walk” condition flashes (and the two traffic signals are red), pedestrians
should hurry up and finish crossing in any direction. When the “don’t walk” condition is on steady, pedestrians should not
enter the intersection. The red and green LED connected to PF1 and PF3 respectively must be used. All other inputs and
outputs must be built on the protoboard. The Lab 5 simulator/grader is implemented in edXLab10.dll.
Walk
South
R
Y
G
West
Walk
R Y G
GR
Don’t walk
PF3
PF1
Figure 5.1. Traffic Light Intersection (three inputs and eight outputs).
Traffic should not be allowed to crash. I.e., there should not be only a green or only a yellow LED on one road at the
same time there is only a green or only a yellow LED on the other road. You should exercise common sense when assigning
the length of time that the traffic light will spend in each state; so that the system changes at a speed convenient for the TA
(stuff changes fast enough so the TA doesn’t get bored, but not so fast that the TA can’t see what is happening). Cars should
not be allowed to hit the pedestrians. The walk sequence should be realistic, showing three separate conditions: 1) “walk”, 2)
“hurry up” using a flashing LED, and 3) “don’t walk”. You may assume the two car sensors remain active for as long as
service is required. However a request for walk is signified by a push and release, where the walk button is pushed for at least
2 seconds.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 26
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
There is no single, “best” way to implement your system. However, your scheme must use a linked data structure
stored in ROM. There should be a 1-1 mapping from the FSM states and the linked elements. A “good” solution will have
about 15 to 30 states in the finite state machine, and provides for input dependence. Try not to focus on the civil engineering
issues. I.e., the machine does not have to maximize traffic flow or minimize waiting. On the other hand if there are multiple
requests, the system should cycle through, servicing them all. Build a quality computer engineering solution that is easy to
understand and easy to change. We expect your solution to behave reasonably and have 15 to 30 states. It is unacceptable if
your machine is less than 15 states. If your machine has more than 30 states you have made it more complicated than we had
in mind. Because we have three inputs, there will be 8 next state arrows. One way to draw the FSM graph to make it easier to
read is to use X to signify don’t care. For example, compare the Figure 6.9 in the book to the FSM graph in Figure 5.2 below.
Drawing two arrows labeled 01 and 11 is the same as drawing one arrow with the label X1. When we implement the data
structure, we will expand the shorthand and explicitly list all possible next states.
Next if input is 01 or 11
X0
goN
100001
30
Wait time
X1
waitN
100010
5
0X
XX
goE
001100
30
1X
waitE
010100
5
XX
Output
Figure 5.2. FSM from the book Figure 6.20 redrawn with a short hand format.
This lab was written in a manner intended to give you a great deal of flexibility in how you draw the FSM graph,
while at the same time require very specific boundaries on how the FSM controller must be written. This flexibility causes
students to be question “when am I done?” or “is this enough for an A?” To clarify the distinction between computer
engineering, and civil engineering, I restate the computer engineering requirements. In particular do these 9 requirements
well and you can get a good grade on this lab.
0) The system provides for input dependence. This means each state has 8 arrows such that the next state
depends on the current state and the input. This means you can not solve the problem by simply cycling
through all the states regardless of the input. You should not implement a Mealy machine.
1) There must be a 1-1 mapping between state graph and data structure. For a Moore machine, this means
each state in the graph has a name, an output, a time to wait, and 8 next state arrows (one for each input).
The data structure has exactly these components: a name, an output, a time to wait, and 8 next state
pointers (one for each input). There is no more or no less information in the data structure then the
information in the state graph.
2) There can be no conditional branches (do-while, while-loop, if-then, or for-loops) in your system, other
than in SysTick_Wait and in SysTick_Wait10ms. See the Example 6.5 in the book. You will have
an unconditional while-loop in the main program that runs the FSM controller.
3) The state graph defines exactly what the system does in a clear and unambiguous fashion.
4) Each state has the same format of each state. This means every state has exact one name, one 8-bit output,
one time to wait, and 8 next pointers.
5) Please use good names and labels (easy to understand and easy to change). Examples of bad state names
are S0 and S1.
6) Do not allow cars to crash into each other. Do not allow pedestrians to walk in one direction while any
cars are allowed to go. Engineers do not want people to get hurt.
7) There should be 10 to 30 states with either a Moore finite state machine. What if I have less than 10
states? Usually students with less than 10 states did not flash the don’t walk light, or they flashed the
lights using a counter. Counters and variables violate the “no conditional branch” requirement. If you are
less than 10 to 15 states, you probably did not handle the 2 second walk button recognition.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 27
8) If the pedestrian pushes the walk button for 2 or more seconds, eventually a walk light must occur. If the
pedestrian pushed the walk button for less than 2 seconds, it is ok for the system to generate a walk signal
or to not generate a walk signal.
One way to handle the 2-second walk button requirement is to add a duplicate set of states. The first set of states
means the walk button is not pressed. The second set of states means the walk operation is requested. Go from the first set to
the second set whenever the walk is pushed. Go from the second back to the first whenever the walk condition is output. The
two sets of states allow you to remember that a walk has been requested; in this way the request is serviced when appropriate.
There are many civil engineering questions that students ask. How you choose to answer these questions will
determine how good a civil engineer you are, but will not affect your grade on this lab. For each question, there are many
possible answers, and you are free to choose how you want to answer it. It is reasonable however for the TA to ask how you
would have implemented other answers to these civil engineering questions using the same FSM structure.
0) How long should I wait in each state? Possible answer: 1 to 2 seconds of real TA time.
1) What happens if I push 2 or 3 buttons at a time? Possible answer: cycle through the requests servicing
them in a round robin fashion.
2) What if I push the walk button, but release it before 2 seconds are up? Possible answer: ignore the
request as if it never happened. Possible answer: service it or ignore it depending on exactly when it
occurred.
3) What if I push a car button, but release it before it is serviced? Possible answer: ignore the request as if it
never happened (e.g., car came to a red light, came to a full stop, and then made a legal turn). Possible
answer: service the request or ignore it depending on when it occurred.
4) Assume there are no cars and the light is green on the North, what if a car now comes on the East? Do I
have to recognize a new input right away or wait until the end of the wait time? Possible answer: no, just
wait until the end of the current wait, then service it. Possible answer: yes; break states with long waits
into multiple states with same output but shorter waits.
5) What if the walk button is pushed while the don’t walk light is flashing? Possible answer: ignore it, go to
a green light state and if the walk button is still pushed, then go to walk state again. Possible answer: if
no cars are waiting, go back to the walk state. Possible answer: remember that the button was pushed,
and go to a walk state after the next green light state.
6) Does the walk occur on just one street or both? Possible answer: stop all cars and let people walk across
either or both streets.
7) How do I signify a walk condition? Answer: You must use the on board green LED for walk, and on
board red LED as the don’t walk.
Figure 5.3. Massachusetts walk signified by red and yellow (phased out in 2011).
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 28
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
In real products that we market to consumers, we put the executable instructions and the finite state machine linked data
structure into the nonvolatile memory such as flash EEPROM. A good implementation will allow minor changes to the finite
machine (adding states, modifying times, removing states, moving transition arrows, changing the initial state) simply by
changing the linked data structure, without changing the executable instructions. Making changes to executable code requires
you to debug/verify the system again. If there is a 1-1 mapping from FSM to linked data structure, then if we just change the
state graph and follow the 1-1 mapping, we can be confident our new system operate the new FSM properly. Obviously, if
we add another input sensor or output light, it may be necessary to update the executable part of the software, re-assemble
and retest the system.
Procedure
The C compiler will create a segmented software system putting global variables into RAM, and local variables into
registers or on the stack. The compiler will place constants, and fixed data structures, and program object code into
EEPROM. Most microcontrollers have a rich set of timer functions. For this lab, you will the SysTick timer to wait a
prescribed amount of time. You may use the functions in Section 4.4, but you are responsible for understanding all the details
of Program 4.7. On the other hand you may use the Program 4.6 to activate the PLL. You are not responsible for the details
of how the PLL works, but are responsible for what the PLL does and why we use the PLL.
In general, using cycle-counting (for-loops) for delays has the problem of conditional branches and data-dependent
execution times. If an interrupt were to occur during a cycle-counting delay, then the delay would be inaccurate using the
cycle-counting method. In the above method, however, the timing will be very accurate, even if an interrupt were to occur
while the microcomputer was waiting. In more sophisticated systems, using timer interrupts provide a more flexible and even
more accurate mechanism for microcomputer synchronization. Interrupt synchronization will be used in Labs 6 through 10. A
linked solution may not run the fastest, or occupy the fewest memory bytes, but it is a structured technique that is easy to
understand, easy to implement, easy to debug, and easy to upgrade.
The basic approach to this lab will be to first develop and debug your system using the simulator. After the software is
debugged, you will interface actual lights and switches to the TM4C123, and run your software on the real TM4C123. As
you have experienced, the simulator requires different amount of actual time as compared to simulated time. On the other
hand, the correct simulation time is maintained in the SysTick timer, which is decremented every cycle of simulation time.
The simulator speed depends on the amount of information it needs to update into the windows. Because we do not want to
wait the minutes required for an actual intersection, the cars in this traffic intersection travel much faster than “real” cars. In
other words, you are encouraged to adjust the time delays so that the operation of your machine is convenient for you to
debug and for the TA to observe during demonstration.
If you are using PD0, PD1, PB7, PB6, PB1 or PB0, make sure R9, R10, R25, and R29 are removed from your
LaunchPad. R25 and R29 are not on the older LM4F120 LaunchPads, just the new TM4C123 LaunchPads. The TM4C123
LaunchPads I bought did not have R25 and R29 soldered on, so I just had to remove R9 and R10. The R9 R10 jumpers are
only needed for some MSP430 booster packs running on the TM4C123, so there is little chance you will ever need R9 and
R10.
Part a) Decide which port pins you will use for the inputs and outputs. Avoid the pins with hardware already connected.
Table 4.4 in the book lists the pins to avoid. Run the starter code in the simulator to see which ports are available for the
lights and switches; these choices are listed in Tables 5.1 and 5.2. The “don’t walk” and “walk” lights must be PF1 and PF3
respectively, but where to attach the others have some flexibility. In particular, Table 5.1 shows you three possibilities for
how you can connect the six LEDs that form the traffic lights. Table 5.2 shows you three possibilities for how you can
connect the three positive logic switches that constitute the input sensors. Obviously, you will not connect both inputs and
outputs to the same pin.
Red south
PA7
PB5
PE5
Yellow south
PA6
PB4
PE4
Green south
PA5
PB3
PE3
Red west
PA4
PB2
PE2
Yellow west
PA3
PB1
PE1
Green west
PA2
PB0
PE0
Table 5.1. Possible ports to interface the traffic lights (PF1=red don’t walk, PF3=green walk).
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Walk sensor
PA4
PB2
South sensor
PA3
PB1
West sensor
PA2
PB0
Table 5.2. Possible ports to interface the sensors.
Page 29
PE2
PE1
PE0
Part b) Design a finite state machine that implements a good traffic light system. Include a graphical picture of your finite
state machine showing the various states, inputs, outputs, wait times and transitions.
Part c) Write and debug the C code that implements the traffic light control system. During the debugging phase with the
simulator. In simulation mode, capture logic analyzer screen shots showing the operation when cars are present on both
roads, like Figure 5.3. Your performance grade on this lab will be determined by your TA during checkout. You may run the
automatic grader this semester, but this score will not be considered as part of your actual grade.
Part d) After you have debugged your system in simulation mode, you will implement it on the real board. Use the same
ports you used during simulation. EE319K_TM4C123_Artist.sch is a starter file you should use to draw your hardware
circuit diagram using the program PCB Artist. The first step is to interface three push button switches for the sensors. You
should implement positive logic switches. Do not place or remove wires on the protoboard while the power is on. Build the
switch circuits and test the voltages using a digital voltmeter. You can also use the debugger to observe the input pin to
verify the proper operation of the interface.
The next step is to build six LED output circuits. Build the system physically in a shape that matches a traffic
intersection, so the TA can better visualize the operation of the system. Look up the pin assignments in the 7406 data sheet.
Be sure to connect +5V power to pin 14 and ground to pin 7. Write a simple main program to test the LED interface. Set the
outputs high and low, and measure the three voltages (input to 7406, output from 7406 which is the LED cathode voltage,
and the LED anode voltage).
Part e) Debug your combined hardware/software system on the actual TM4C123 board.
Demonstration
During checkout, you will be asked to show both the simulated and actual TM4C123 systems to the TA. The TAs will
expect you to know how the SysTick_Wait function works, and know how to add more input signals and/or output
signals. An interesting question that may be asked during checkout is how you could experimentally prove your system
works. In other words, what data should be collected and how would you collect it? If there were an accident, could you
theoretically prove to the judge and jury that your software implements the FSM? What type of FSM do you have? What
other types are there? How many states does it have? In general, how many next-state arrows are there? Explain how the
linked data structure is used to implement the FSM. Explain the mathematical equation used to calculate the address of the
next state, depending on the current state and the input. Be prepared to write software that delays 1 second without using the
timer (you can use a calculator and manual). How do you prove the delay will be 1 second? What does it mean for the C
compiler to align objects in memory? Why does the compiler perform alignment? List some general qualities that would
characterize a good FSM.
Deliverables (Items 1, 2, 3, 4 are one pdf file uploaded to SVN, have this file open during demo.)
0) Lab 5 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
1) Logic analyzer screen shot showing the system running in simulation mode when cars are present on both roads.
2) Circuit diagram (with your name and date) using PCB Artist
3) Drawing of the finite state machine
All source files that you have changed or added (like main.s) should be committed to SVN. Please do not commit other file
types.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 30
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Your I/O window may look slightly different from these examples (e.g., you are free to assign signals to different port bits.)
Figure 5.4. Simulation showing cars on both South and West
___________________________________________________________________________________
Figure 5.5. Simulation showing walk pushed, and cars on both South and West
___________________________________________________________________________________
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 31
Lab 6. Digital Piano using a Digital to Analog Converter
Preparation
Read Sections 9.1, 9.2, 9.4, 9.6, 10.1, 10.2 and 10.3,
Starter files PeriodicSysTickInts_xxx Lab6_EE319K.zip (see dac.xls inside this project)
https://www.youtube.com/playlist?list=PLyg2vmIzGxXGxGIgbOrqYWALtXnDvxCCd
http://users.ece.utexas.edu/~valvano/Volume1/E-Book/C13_Interactives.htm
Purpose
There are three objectives for this lab: 1) to learn about DAC conversion; 2) to understand how digital data stored in
a computer could be used to represent sounds and music; 3) to study how the DAC can be used to create sounds.
Programming Requirements
All software for this lab must be written in C. You can debug your code in the simulator but your final code must
run on the board with a DAC circuit. Keil uVision4 projects in C are a little different than the assembly projects used for
Labs 1 to 4. The project configuration and the startup.s file are different. The Lab 6 starter file has the appropriate
connections to the Lab 6 simulator/grader. However, you should also look at PeriodicSysTickInts_xxx.zip.
System Requirements
Most digital music devices rely on high-speed DAC converters to create the analog waveforms required to produce
high-quality sound. In this lab you will create a very simple sound generation system that illustrates this application of the
DAC. Your goal is to create an embedded system that plays three notes, which will be a digital piano with three keys. The
first step is to design and test a 4-bit binary-weighted DAC, which converts 4 bits of digital output from the TM4C123 to an
analog signal. You are free to design your DAC with a precision of more than 4 bits. You will convert the four digital bits to
an analog output using a simple resistor network. During the static testing phase, you will connect the DAC analog output to
your voltmeter and measure resolution, range, precision and accuracy. During the dynamic testing phase you will connect the
DAC output to headphones, and listen to sounds created by your software. It doesn’t matter what range the DAC is, as long
as there is an approximately linear relationship between the digital data and the speaker current. The performance score of
this lab is not based on loudness, but sound quality. The quality of the music will depend on both hardware and software
factors. The precision of the DAC, external noise and the dynamic range of the speaker are some of the hardware factors.
Software factors include the DAC output rate and the number of data points stored in the sound data. You can create a 3k
resistor from two 1.5k resistors. You can create a 6k resistor from two 12k resistors.
Static
Dynamic
testing
testing
TM4C123
TM4C123
Bit3
Bit2
Bit3
Vout
Bit2
Voltmeter
bit1
Bit1
Bit0
Bit0
I out
Speaker
Figure 6.1. DAC allows the software to create music.
The second step is to design a low-level device driver for the DAC. Remember, the goal of a device driver is to
separate what the device does from how it works. “What it does” means the general descriptions and function prototypes of
DAC_Init and DAC_Out that are placed in DAC.h. “How it works” means the implementations of DAC_Init and
DAC_Out that will be placed in DAC.c. The third step is to design a low-level device driver for the three keys of the piano.
For example, if you could create public functions Piano_Init and Piano_In, where Piano_In returns a logical key
code for the pattern of keys that are pressed. You may design this driver however you wish, but the goal is to abstract the
details how it works (which port, which pin) from what it does (which key pressed). The fourth step is to organize the sound
generation software into a device driver. You will need a data structure to store the sound waveform. You are free to design
your own format, as long as it uses a formal data structure. Compressed data occupies less storage, but requires runtime
calculation. Although you will be playing only three notes, the design should allow additional notes to be added with minimal
effort. For example, if you could create public functions Sound_Init and Sound_Play(note), where the parameter
note specifies the frequency (pitch) of the sound. For example, calling Sound_Play(Off) makes it silent and calling
Sound_Play(C) plays the note C. A background thread within the sound driver implemented with SysTick interrupts will
fetch data out of your music structure and send them to the DAC. The last step is to write a main program that links the
modules together creating the digital piano.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 32
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Procedure
Part a) Decide which port pins you will use for the inputs and outputs. Avoid the pins with hardware already connected.
Table 4.4 in the book lists the pins to avoid. There is a Lab 6 simulator, but will we not use the grader in EE319K this
semester. To use the simulator/grader the choices are listed in Tables 6.1 and 6.2. In particular, Table 6.1 shows you three
possibilities for how you can connect the DAC output. Table 6.2 shows you three possibilities for how you can connect the
three positive logic switches that constitute the piano keys. Obviously, you will not connect both inputs and outputs to the
same pin.
DAC bit 5
PA7
PB5
PE5
DAC bit 4
PA6
PB4
PE4
DAC bit 3
PA5
PB3
PE3
DAC bit 2
PA4
PB2
PE2
DAC bit 1
PA3
PB1
PE1
DAC bit 0
PA2
PB0
PE0
Table 6.1. Possible ports to interface the DAC outputs (DAC bits 4 and 5 are optional).
Piano key 3
PA5
PD3
PE3
Piano key 2
PA4
PD2
PE2
Piano key 1
PA3
PD1
PE1
Piano key 0
PA2
PD0
PE0
Table 6.2. Possible ports to interface the piano key inputs (piano key 3 is optional).
Figure 6.1b. The screens needed to run Lab 6 in simulation.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 33
Draw the circuit required to interface the binary-weighted DAC to the TM4C123. Design the DAC converter using a simple
resistor-adding technique. A 4-bit binary-weighted DAC uses resistors in a 1/2/4/8 resistance ratio. Select values in the 1.5
kΩ to 240 kΩ range. For example, you could use 1.5 kΩ, 3 kΩ, 6 kΩ, and 12 kΩ. Notice that you could create double/half
resistance values by placing identical resistors in series/parallel. It is a good idea to email a pdf file of your design to your
TA and have him/her verify your design before you build it. You can solder 24 gauge solid wires to the audio jack to simplify
connecting your circuit to the headphones. Plug your headphones into your audio jack and use your ohmmeter to determine
which two wires to connect. You have the option of connecting just the left, just the right, or both channels of the
headphones. Draw interface circuits for three switches, which will be used for the piano keyboard.
For sound on one channel only,
connect middle pin to ground
Figure 6.2. A mono jack has three connections (for one channel connect pin 1 to ground and pin 3 to the DAC output).
Part b) You must write the entire code in C. In addition, if you perform the extra credit, it too should be implemented in C.
Write the C language device driver for the DAC interface. Include at least two functions that implement the DAC interface.
For example, you could implement the function DAC_Init() to initialize the DAC, and the function DAC_Out to send a
new data value to the DAC. Place all code that accesses the DAC in a separate DAC.c code file. Add a DAC.h header file
with the prototypes for public functions. Describe how to use a module in the comments of the header file. Describe how the
module works, how to test the module, and how to change module in the comments of the code file.
Part c) Begin with the static testing of the DAC. You will write a simple main program to test the DAC, similar in style as
Program 6.1. You are free to debug this system however you wish, but you must debug The DAC module separately. You
should initially debug your software in the simulator (Figure 6.3). You can single step this program, comparing digital Data
to the analog voltage at the Vout without the speaker attached (i.e., left side of Figure 6.1).
#include "DAC.h"
int main(void){
unsigned long Data; // 0 to 15 DAC output
PLL_Init();
// like Program 4.6 in the book, 80 MHz
DAC_Init();
for(;;) {
DAC_Out(Data);
Data = 0x0F&(Data+1); // 0,1,2...,14,15,0,1,2,...
}
}
Program 6.1. A simple program that outputs all DAC values in sequence.
Using Ohm’s Law and fact that the digital output voltages will be approximately 0 and 3.3 V, make a table of the theoretical
DAC voltage and as a function of digital value (without the speaker attached). Calculate resolution, range, precision and
accuracy. Complete Table 6.3 and attach it as a deliverable (alternatively you could enter this data into DACdata.xls).
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 34
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Bit3 bit2 bit1 bit0
0
Theoretical DAC voltage
Measured DAC voltage
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 6.3. Static performance evaluation of the DAC (if you implement a 6-bit DAC, then make a table with 16
measurements: 0,1,7,8,15,16,17,18,31,32,33,47,48,49,62,63).
Figure 6.3 shows the static testing using program 6.1 running in the simulator.
Figure 6.3. A screenshot in simulation mode showing the static testing of the DAC. (the current version has a bug requiring
you place a break point at the line that you output to the DAC so the scope window is updated)
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 35
Part d) Design and write the piano keyboard device driver software. These routines facilitate the use of the four piano keys.
Include at least two functions that implement the piano keyboard interface. For example, you could implement the function
Piano_Init() to initialize the switch inputs, and the function Piano_In that returns a logical key code for the pattern of
switches that are pressed. Place all code that directly accesses the three switches in a separate Piano.c code file. Add a
Piano.h header file with the prototypes for public functions. Add comments that
describe what it does in the Piano.h file and how it works in the Piano.c file.
Part e) Design and write the sound device driver software. The input to the sound
driver is the pitch of the note to play. SysTick interrupts will be used to set the
time in between outputs to the DAC. Add minimally intrusive debugging
instruments to allow you to visualize when interrupts are being processed. Include
at least two functions that implement the sound output. For example, you could
implement the function Sound_Init() to initialize the data structures, calls
DAC_Init, and initializes the SysTick interrupt. You could implement a function
Sound_Play(note) that starts sound output at the specified pitch. Place all
code that implements the waveform generation in a separate Sound.c code file.
Add a Sound.h header file with the prototypes for public functions. Add comments
that describe what it does in the Sound.h file and how it works in the Sound.c file.
When you wish to play a new note you should write to NVIC_ST_RELOAD_R,
changing the interrupt period, without completely initializing SysTick.
One approach to debugging is to attempt to create a sine wave with a
constant frequency as shown in Figures 6.4 and 6.5. Just run the SysTick periodic
interrupts and output one DAC value each interrupt. The toggling digital line
shows you the interrupts are happening, and the top sine wave shows your table
and DAC output are working.
Project include a main (Lab6) and four submodules(PLL, Sound, piano, and DAC)
Figure 6.4. The 4-bit DAC with a 32-element table is used to create a sine wave in simulation. The top trace is the DAC
output (without headphones) and the bottom trace is a debugging monitor that toggles at each SysTick interrupt. (the current
version has a bug requiring you place a break point at the line that you output to the DAC so the scope window is updated)
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 36
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Figure 6.5. The 4-bit DAC with a 32-element table is used to create a sine wave. In this system, the software was attempting
to create 2093 Hz, and the measured frequency was 2092 Hz. The top trace is the DAC output (without headphones) and the
bottom trace is a debugging monitor that toggles at each SysTick interrupt.
Part f) Write a main program to implement the three-key piano. Make a heartbeat connected to an LED so you can see your
program is running. Document clearly the operation of the routines. Figure 6.6 shows a possible data flow graph of the music
player. Debug the system first in the simulator then on the real TM4C123 with an oscilloscope. Take a photograph of the
scope traces (like Figure 6.10) to capture the waveform generated by your digital piano. When no buttons are pressed, the
output will be quiet. When Button 1 is pressed, output a sine wave at one frequency. When Button 2 is pressed, output a sine
wave at a second frequency. When Button 3 is pressed, output a sine wave at a third frequency. Only one button will be
pressed at a time. The sound lasts until the button is released.
Push
buttons
Piano
interface
main
Sound
interface
DAC
interface
DAC
hardware
Sin wave
Table
Figure 6.6. Data flows from the memory and the switches to the speaker.
Figure 6.7 shows a possible call graph of the Lab 6 system. Figure 6.8 is one possible way to implement the extra credit.
Dividing the system into modules allows for concurrent development and eases the reuse of code.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 37
main
Sound driver
Piano driver
Init In
Init Play ISR
Switch
hardware
SysTick
hardware
DAC driver
Init Out
Speaker
hardware
Figure 6.7. A call graph showing the three modules used by the regular part of Lab 6. Notice the SysTick hardware calls the
SysTick ISR.
main
Music driver
Song Stop ISR
Sound driver
Piano driver
Init In
Init Play ISR
Switch
hardware
SysTick
hardware
DAC driver
Init Out
Timer0A
hardware
Speaker
hardware
Figure 6.8. A call graph showing the three modules used by the extra credit part of Lab 6. Notice there are two interrupts.
Part g) Add a debugging monitor to your interrupt service routine. I.e., set a digital output pin high at the start of the ISR and
clear it low at the end of the ISR. The scopes in the lab should have probes attached. For this part you need to use the
oscilloscope debugging tool. Connect one channel to a digital output showing when the interrupts are occurring and a second
channel to the DAC output (without the headphones). Set the oscilloscope to trigger on the digital output occurring in the
ISR. Observe the relationship between software and hardware.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 38
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Figure 6.9. Photograph of the DAC output and debugging monitor. This system uses a 6-bit DAC and a 64-element table.
Notice the DAC is monotonic (each increment of the digital input causes the DAC output to go up. (You only need 4 bits). The
top trace is the 6-bit DAC output (without headphones) and the bottom trace is a debugging monitor that toggles at each
SysTick interrupt.
Demonstration (both partners must be present, and demonstration grades for partners may be different)
You should be able to demonstrate the three notes. Be prepared to explain how your software works. You should be
prepared to discuss alternative approaches and be able to justify your solution. The TA may look at your data and expect you
to understand how the data was collected and how DAC works. In particular, you should be able to design a DAC with 5 to
10 bits. What is the range, resolution and precision? You will tell the TA what frequency you are trying to generate, and they
may check the accuracy with a frequency meter or scope. TAs may ask you what frequency it is supposed to be, and then ask
you to prove it using calculations. Just having three different sounding waves is not enough, you must demonstrate the
frequency is proper and it is a sinewave (at least as good as you can get with a 4-bit DAC). You will be asked to attach your
DAC output to the scope (part g). Many students come for their checkout with systems that did not operate properly. You
may be asked SysTick interrupt and DAC questions. If the desired frequency is f, and there are n samples in the sine wave
table, what SysTick interrupt period would you use?
This lab mentions 32 samples per cycle. Increasing the DAC output rate and the number of points in the table is one
way of smoothing out the “steps” that in the DAC output waveform. If we double the number of samples from 32 to 64 to
128 and so on, keeping the DAC precision at 4-bit, will we keep getting a corresponding increase in quality of the DAC
output waveform?
As you increase the number of bits in the DAC you expect an increase in the quality of the output waveform. If we
increase the number of bits in the DAC from 4 to 6 to 8 and so on, keeping the number of points in the table fixed at 32, will
we keep getting a corresponding increase in quality of the DAC output waveform?
Deliverables (Items 1, 2, 3, 4, 5 are one pdf file uploaded to SVN, have this file open during demo.)
0) Lab 6 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
Create one pdf file with the following for electronic submission (UTEID1_UTEID2.pdf)
1) Circuit diagram showing the DAC and any other hardware used in this lab, PCB Artist
2) Software Design
Draw pictures of the data structures used to store the sound data
If you organized the system different than Figure 6.6 and 6.7, then draw its data flow and call graphs
3) A picture of the dual scope (part g) like Figures 6.5 or 6.11.
4) Measurement Data
Show the theoretical response of DAC voltage versus digital value (part c, Table 6.3)
Show the experimental response of DAC voltage versus digital value (part c, Table 6.3)
Calculate resolution, range, precision and accuracy
5) Brief, one sentence answers to the following questions
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 39
When does the interrupt trigger occur?
In which file is the interrupt vector?
List the steps that occur after the trigger occurs and before the processor executes the handler.
It looks like BX LR instruction simply moves LR into PC, how does this return from interrupt?
All source files that you have changed or added (like main.c DAC.c DAC.h) should be committed to SVN. Please do not
commit other file types.
Extra Credit. Extend the system so that is plays your favorite song (a sequence of notes, set at a specific tempo and includes
an envelope). Your goal is to play your favorite song. The song should contain at least 5 different pitches and at least 20
notes. But, what really matters is the organization is well-done using appropriate data structures that is easy to understand and
easy to adapt for other songs. This extra credit provides for up to 20 additional points, allowing for a score of 120 out of 100
for this lab. 10 possible points for 2 or more extra DAC
bits, and 10 possible points for a song worthy of
dancing. One possible approach is to use two interrupts.
A fast SysTick ISR outputs the sinewave to the DAC
(Figures 6.4, 6.11). The rate of this interrupt is set to specify the frequency (pitch) of the sound. A second slow Timer ISR
occurs at the tempo of the music. For example, if the song has just quarter notes at 120, then this interrupt occurs every 500
ms. If the song has eight notes, quarter notes and half notes, then this interrupt occurs at 250, 500, 1000 ms respectively.
During this second ISR, the frequency of the first ISR is modified according to the note that is to be played next.
Compressed data occupies less storage, but requires runtime calculations to decompress. On the other hand, a complete list of
points will be simpler to process, but requires more storage than is available on the TM4C123. The fourth step is to organize
the music software into a device driver. Although you will be playing only one song, the song data itself will be stored in the
main program, and the device driver will perform all the I/O and interrupts to make it happen. You will need public functions
Song and Stop, which perform operations like a cassette tape player. The Song function has an input parameter that
defines the song to play. If you complete the extra credit (with input switches that can be used to start and stop), then the
three-key piano functionality still must be completed. In other words, the extra credit part is in addition to the regular part.
You interface more switches (which you can get from checkout) or you could use the on-board switches to activate the Song
and Stop operations
Some web links about music
Simple Music Theory
http://library.thinkquest.org/15413/theory/theory.htm
Free sheet music
http://www.8notes.com/piano/
Figure 6.10. A song being played with a harmony and melody (see two peaks in the spectrum).
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 40
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Figure 6.11. This waveform sounds like a horn (6-bit DAC, 64-element table). The top trace is the 6-bit DAC output (without
headphones) and the bottom trace is a debugging monitor that toggles at each SysTick interrupt.
There should be no DAC_Out in the main loop. There should be zero or one DAC_Out each execution of the ISR. Some
students will discover they can make sounds without interrupts, by simply using a software delay like Labs 4, and 5. This
approach is not allowed. Other students will discover they can create sounds by performing the DAC_Out in the main loop.
This approach is also not allowed. 20 point Lab 6 point deduction if zero or one calls to DAC_Out do not occur in the ISR.
These other approaches are not allowed because we will be employing sounds in Lab 10, and we will need the sound to
completely run within the ISR. The Lab 10 main program will start and stop sounds, but the ISR will do all the DAC_Out
calls.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 41
Soldering Guide
Reference: © John Hewes 2006, The Electronics Club, www.kpsec.freeuk.com
Safety Precautions
• Do not solder if you are pregnant or think you might be pregnant
• Never touch the element or tip of the soldering iron.
They are very hot (about 400°C) and will give you a nasty burn.
• Take great care to avoid touching any metal parts of the iron.
The iron should have a heatproof flex for extra protection.
• Always return the soldering iron to its stand when not in use.
Never put it down on your workbench, even for a moment!
• Allow joints a minute or so to cool down before you touch them.
• Work in a well-ventilated area.
The smoke formed as you melt solder is mostly from the flux and quite irritating. Avoid breathing it by keeping you
head to the side of, not above, your work.
• Wash your hands after using solder.
Solder contains lead.
Treatment for minor burns
Most burns from soldering are likely to be minor and treatment is simple:
• Immediately cool the affected area under gently running cold water.
Keep the burn in the cold water for at least 5 minutes (15 minutes is recommended). You do not wish to freeze the
tissue, so running water is much better than ice.
• Do not apply any creams or ointments.
The burn will heal better without them. A dry dressing, such as a clean handkerchief, may be applied if you wish to
protect the area from dirt.
• Seek medical attention if the burn breaks the skin.
Preparing the soldering iron
• Place the soldering iron in its stand and plug in.
The iron will take a few minutes to reach its operating temperature of about 400°C.
• Dampen the sponge in the stand.
The best way to do this is to lift it out the stand and hold it under a cold tap for a moment, then squeeze to remove
excess water. It should be damp, not dripping wet. Don’t put the
electronics near the tap water
• Wait a few minutes for the soldering iron to warm up.
You can check if it is ready by trying to melt a little solder on the
tip.
• Wipe the tip of the iron on the damp sponge.
This will clean the tip.
• Melt a little solder on the tip of the iron.
This is called 'tinning' and it will help the heat to flow from the
iron’s tip to the joint. It only needs to be done when you plug in
the iron, and occasionally while soldering if you need to wipe the
tip clean on the sponge.
Preparation
• You have two hands.
One hand will hold the iron and the second hand will hold the solder. This means everything else must be
immobilized.
• Immobilize the two ends to be soldered such the two ends are in physical contact.
In EE319K we will be soldering a wire to a component. One simple way is to 1) use masking tape to fix the
component to a ceramic tile; 2) cut and strip solid wire; and 3) use masking tape to hold the wire so that the two
metallic parts are touching. Another approach is to 1) cut and strip solid wire; and 2) tightly wrap the stripped
portion of the wire around the component.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 42
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Making soldered joints
• Hold the soldering iron like a pen, near the base of the handle.
Imagine you are going to write your name! Remember to never touch the hot element or tip.
• Touch the soldering iron onto the joint to be made.
Make sure it touches both the component lead and the track. Hold the tip there for a few seconds and...
• Feed a little solder onto the joint.
It should flow smoothly onto the lead and track to form a
volcano shape as shown in the diagram below. Make sure
you apply the solder to the joint, not the iron.
• Remove the solder, then the iron, while keeping the joint still.
Allow the joint a few seconds to cool before you move
the circuit board.
• Inspect the joint closely.
It should look shiny and have a ‘volcano’ shape. If not,
you will need to reheat it and feed in a little more solder.
This time ensure that both the lead and track are heated
fully before applying solder.
Good
-volcano
-shiny
Bad
-ball, bumpy
-dull
Soldering advice for components
Some components require special care when soldering.
Many must be placed the correct way round and a few are easily damaged by the heat from soldering. Appropriate
warnings are given in the following, together with other advice which may be useful when soldering.
Components
Resistor
Pictures
Advice
No special precautions are required.
Connect either way round.
Diodes must be connected the correct way
round: a = anode, k = cathode. Use a heat
sink with germanium diodes.
Diodes
a
k
Chip holder
Ensure the notch is at the correct end. Do
not insert the IC at this stage to prevent it
being damaged by heat.
Ceramic, or nonpolarized
capacitor
No special precautions are required.
Connect either way round. Take care to
identify their value.
Electrolytic or polarized
capacitor
LED
-
+
a
k
+
Electrolytic capacitors must be connected
the correct way round, they are marked with
+ or - near one lead.
LEDs must be connected the correct way
round: a = anode (+), k = cathode (-).
flat
Transistors
Transistors have three leads and must be
connected the correct way round.
Integrated circuits
When all soldering is complete, carefully
insert ICs the correct way round in their
holders. Make sure all the pins are lined up
before pushing in firmly.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 43
Lab 7C. LCD Device Driver for the Sitronix ST7735 (Spring 2015)
Preparation
Read Sections 7.1, 7.2, 7.3, 7.4, 7.5, 7.6, and 7.7
Read the data sheet for the LCD display (in project folder)
1-8-tft-display.pdf, ST7735R_V0.2.pdf
Lab7C starter project in SVN
Purpose
This lab has these major objectives: 1) to interface an LCD interface that can be
used to display information on the embedded system; 2) to use indexed addressing
to access strings; 3) to learn how to design implement and test a device driver
using busy-wait synchronization; 4) to learn how to allocate and use local
variables on the stack; 5) to use fixed-point numbers to store non-integer values.
In Spring 2015, we switched from the monochrome Nokia to the color ST7735
because it will make the Lab 10 game more exciting. The LaunchPadDLL will
simulate the Sitronix ST7735, even though Lab 7C will not have an automatic
grader.
Figure 7.1. Sitronix ST7735
System Requirements
In this lab you will interface a Sitronix ST7735 LCD to the TM4C123 as shown in Program 7.1.
// pin 10 Backlight
+3.3 V
// pin 9 MISO
unconnected
// pin 8 SCK
PA2 (SSI0Clk)
// pin 7 MOSI
PA5 (SSI0Tx)
// pin 6 TFT_CS
PA3 (SSI0Fss)
// pin 5 CARD_CS
unconnected
// pin 4 D/C
PA6 (GPIO)
// pin 3 RESET
PA7 (GPIO)
// pin 2 VCC
+3.3 V
// pin 1 Gnd
ground
Program 7.1. Interface connections for the Sitronix ST7735.
D/C stands for data/command; you will make D/C high to send data and low to send a command. This lab will use “busywait” synchronization, which means before the software issues an output command to the LCD, it will wait until the display
is not busy. In particular, the software will wait for the previous LCD command to complete. For the 10ms wait function, we
suggest you use the cycle-counting approach to blind wait (like Lab 2) instead of SysTick (like Lab 5) because you will need
SysTick periodic interrupts for Labs 8, 9, and 10.
The objective of this lab is to develop a device driver for the Sitronix ST7735 LCD display. A device driver is a set of
functions that facilitate the usage of an I/O port. In particular, there are three components of a device driver. First component
is the description of the driver. Since this driver will be developed in assembly, your descriptions are placed in the comments
before each subroutine. In Labs 8, 9, 10, we will call these driver functions from C, so we also place the function prototypes
for the public functions in the header file LCD.h. It is during the design phase of a project that this information is specified.
The second component of a device driver is the implementation of the functions that perform the I/O. If the driver were being
developed in C, then the implementations would have been placed in the corresponding code file, e.g., LCD.C. When
developing a driver in assembly, the implementations are the instructions and comments placed inside the body of the
subroutines. In addition to public functions, a device driver can also have private functions. This interface will require a
private function that outputs to commands to the LCD (notice that private functions do not include LCD_ in their names).In
this lab, you are required to develop and test seven public functions (notice that public functions include LCD_ or IO_ in
their names). The third component is a main program that can used to test these functions. We have given you this main
program, which you can find in the Lab7Main.s file. Please use the starter project which connects all these files.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 44
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
In the IO.s file you will implement and test three functions to handle a user switch and heartbeat LED.
;------------IO_Init-----------; Initialize GPIO Port for a switch and an LED
; Input: none
; Output: none
; Invariables: This function must not permanently modify registers R4 to R11
;------------IO_HeartBeat-----------; Toggle the output state of the LED.
; Input: none
; Output: none
; Invariables: This function must not permanently modify registers R4 to R11
IO_HeartBeat
;------------IO_Touch-----------; First, wait for the key to be released, and then
;
wait for the key to be touched
; Input: none
; Output: none
; This is a public function
; Invariables: This function must not permanently modify registers R4 to R11
In the LCD.s file you will implement and test two functions to communicate directly with the Sitronix ST7735 LCD. You
will not write the initialization ritual, because it is given in the ST7735.c file. However you will write these two functions
that output to the LCD. Your writecommand function will be used to output 8-bit commands to the LCD, and your
writedata function will be used to output 8-bit data to the LCD.
; This is a helper function that sends an 8-bit command to the LCD.
; Input: R0 8-bit command to transmit
; Output: none
; Assumes: SSI0 and port A have already been initialized and enabled
writecommand
;1) Read SSI0_SR_R and check bit 4,
;2) If bit 4 is high, loop back to step 1 (wait for BUSY bit to be low)
;3) Clear D/C=PA6 to zero
;4) Write the command to SSI0_DR_R
;5) Read SSI0_SR_R and check bit 4,
;6) If bit 4 is high, loop back to step 5 (wait for BUSY bit to be low)
; This is a helper function that sends an 8-bit data to the LCD.
; Input: R0 8-bit data to transmit
; Output: none
; Assumes: SSI0 and port A have already been initialized and enabled
writedata
;1) Read SSI0_SR_R and check bit 1,
;2) If bit 1 is low loop back to step 1 (wait for TNF bit to be high)
;3) Set D/C=PA6 to one
;4) Write the 8-bit data to SSI0_DR_R
Built on top of your LCD.s file is the c file ST7735.c, which implements 6 public functions that can be used to display
characters and graphics on the display. You should not modify the ST7735.c file. One of the functions, LCD_OutChar,
outputs one ASCII character on the display. In your file print.s, you will implement and test two more display functions.
Your LCD_OutDec function will be used to output an unsigned 32-bit integer to the LCD, and your LCD_OutFix function
will be used to output an unsigned 32-bit fixed-point number to the LCD.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 45
;-----------------------LCD_OutDec----------------------; Output a 32-bit number in unsigned decimal format
; Input: R0 (call by value) 32-bit unsigned number
; Output: none
; Invariables: This function must not permanently modify registers R4 to R11
LCD_OutDec (R0 is a 32-bit n) You may implement this using either iteration or recursion. You must have at least one
local variable, allocated on the stack, and use symbolic binding for the local. If you use recursion, the base case (n<10)
requires one call to LCD_OutChar. You can use SP or stack frame register for accessing the local variable.
; -----------------------LCD_OutFix---------------------; Output characters to LCD display in fixed-point format
; unsigned decimal, resolution 0.001, range 0.000 to 9.999
; Inputs: R0 is an unsigned 32-bit number
; Outputs: none
; E.g., R0=0,
then output "0.000 "
;
R0=3,
then output "0.003 "
;
R0=89,
then output "0.089 "
;
R0=123, then output "0.123 "
;
R0=9999, then output "9.999 "
;
R0>9999, then output "*.*** "
; Invariables: This function must not permanently modify registers R4 to R11
LCD_OutFix (R0 is a 32-bit num) You must have at least one local variable, allocated on the stack, and use symbolic
binding for the local. You can use SP or stack frame register for accessing the local variable.
Parameter
LCD display
0
″0.000 ″
1
″0.001 ″
999
″0.999 ″
1000
″1.000 ″
9999
″9.999 ″
10000 or more
″*.*** ″
Table 7.1. Specification for the LCD_OutFix function (do not display the quotes”).
An important factor in device driver design is to separate the policies of the interface (how to use the programs, which are
defined in the comments placed at the top of each subroutine) from the mechanisms (how the programs are implemented,
which are described in the comments placed within the body of the subroutine.)
The third component of a device driver is a main program that calls the driver functions. This software has two purposes. For
the developer (you), it provides a means to test the driver functions. It should illustrate the full range of features available
with the system. The second purpose of the main program is to give your client or customer (e.g., the TA) examples of how
to use your driver.
Procedure
The basic approach to this lab will be to first develop and debug your system using the simulator. During this phase of the
project you will use the debugger to observe your software operation. After the software is debugged, you will run your
software on the real TM4C123. There are many functions to write in this lab, so it is important to develop the device driver in
small pieces. One technique you might find useful is desk checking. Basically, you hand-execute your functions with a
specific input parameter. For example, using just a pencil and paper think about the sequential steps that will occur when
LCD_OutDec or LCD_OutFix processes the input 187. Later, while you are debugging the actual functions on the
simulator, you can single step the program and compare the actual data with your expected data.
Part a) This lab is sufficiently complex that we require you to debug the three LCD.s programs on the simulator. If your
writedata and writecommand functions in LCD.s are correct, the main program will output the "Lab 7, welcome
to 319K!" message on LCD screen, as shown in Figure 7.2. When print.s is correct, the LCD will show both a decimal
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 46
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
number and a fixed point number, see Figure 7.3. The TestData array in the Lab7Main.s file lists the test values for
print.s.
Figure 7.2. The simulation output showing LCD.s is operational.
Figure 7.3. The simulation output showing print.s is operation; push PF4 to advance output.
Part b) To test the hardware display, you can download and run the example :
http://users.ece.utexas.edu/~valvano/arm/ST7735_4C123.zip
Do not use the ST7735_4C123 project to build your lab 7 solution, but it can be run to make sure your LCD and wiring are
correct.
Part c) After your functions are tested in simulation, you will then test them on the real board. See Figure 7.4.
Your LCD_OutDec and LCD_OutFix functions must use local variables with binding. To observe the stack when
assembly programs are running, set the memory view to observe 0x200003c0, and the type to be unsigned long. This will
show you to top 16 elements on the stack. Successive refinement is a development approach that can be used to solve
complex problems. If the problem is too complicated to envision a solution, you should redefine the problem and solve an
easier problem. If it is still too complicated, redefine it again, simplifying it even more. You could simplify LCD_OutFix
1) implement the variables in registers (rather than as local variables on the stack)
2) ignore special cases with illegal inputs
3) implement just one decimal digit
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 47
During the development phase, you implement and test the simpler problem then refine it, adding back the complexity
required to solve the original problem. You could simplify LCD_OutDec in a similar fashion.
Figure 7.4. The Sitronix ST7735 is a low-cost LCD with 160x128 pixels.
Demonstration (both partners must be present, and demonstration grades for partners may be different)
You will also be required to demonstrate the proper operation on the actual microcontroller. During demonstration
to the TA, you will run your system in the debugger and show the binding, allocation/initialization, access and deallocation of
the local variables. Each time a function is called, an activation record is created on the stack, which includes parameters
passed on the stack (none in this lab), registered saved, and the local variables. You will be asked to observe the stack in the
debugger and identify the activation records created during the execution of LCD_OutDec. TAs may ask you questions on
LCD interfacing, and programming. What is the difference between post and pre-increment modes of addressing? What does
the D/C signal line on the LCD signify? What does busy-wait synchronization mean in the context of communication with
the LCD? The TA will ask to see your implementation local variables and ask you to explain the four step process (binding,
allocation, access and deallocation). You should be able to draw stack pictures. How does AAPCS apply to this lab? Why is
AAPCS important?
Deliverables (Items 1, 2, 3 are one pdf file uploaded to SVN, have this file open during demo.)
0) Lab 7 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
All source files that you have changed or added (LCD.s, IO.s and print.s) should be committed to SVN. Please do not
commit other file types.
Hints
1) This will run in simulation mode, so you can test your low-level functions, which should output the "Lab 7,
welcome to 319K!" message on LCD screen.
2) It will be important to pass data into your functions in Register R0; this way you will be able to call your
functions from C code later in Labs 8, 9, and 10. It is important to save and restore Registers R4-R11 if you use them.
3) You must use the cycle-counting approach to implement the blind waits (Lab 2) instead of SysTick (Lab 5)
because you will need SysTick periodic interrupts for Labs 7, 8, and 9.
4) Because there is a blind cycle synchronization in ST7735.c, Delay1ms so please leave the compiler
optimization on Level 0 and do not click Optimize for Time
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 48
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Lab 8. Real-Time Position Measurement System
Preparation, starter project Lab8_EE319K
Read Sections 9.1, 9.2, 9.3, 9.4, 9.6, 9.8, 10.1, 10.4, and 10.6
http://users.ece.utexas.edu/~valvano/arm/ADCSWTrigger_4C123.zip
Purpose
This lab has these major objectives: 1) an introduction to
sampling analog signals using the ADC interface; 2) the development of
an ADC device driver; 3) learning data conversion and calibration
techniques; 4) the use of fixed-point numbers; 5) the development of an
interrupt-driven real-time sampling device driver; 6) the development of a
software system involving multiple files and a mixture of assembly and C; and 7) learn how to debug one module at a time.
System Requirements
You will design a position meter. Your software will use the 12-bit ADC built into the microcontroller. A linear
slide potentiometer (Bourns PTA2043-2015CPB103) converts position into resistance (0 ≤ R ≤ 10 kΩ). Depending on how
you solder the wires to the pot, the full scale range of position may be anywhere from 1.5 to 2 cm. You will use an electrical
circuit to convert resistance into voltage (Vin). Since the potentiometer has three leads, one possible solution is shown in
Figure 8.1. You may use any ADC channel. The TM4C123 ADC will convert voltage into a 12-bit digital number (0 to
4095). This ADC is a successive approximation device with a conversion time on the order of several µsec. Your software
will calculate position from the ADC sample as a decimal fixed-point number (resolution of 0.001 cm). The position
measurements will be displayed on the LCD using the LCD device driver developed in the Lab 7. A periodic interrupt will be
used to establish the real-time sampling. The main program and SysTick ISR must be written in C. You must use your LCD
device drivers that you developed as part of Lab 7. The ADC and PLL code can be written in either assembly or C, your
choice. The device drivers should be in separate files: adc.s (or adc.c), and pll.s ( or pll.c). Each driver file will have a
corresponding header file with the prototypes to public functions. The SysTick initialization, SysTick ISR, mailbox and the
main program will be in the main.c file.
Figure 8.1. Possible circuit to interface the sensor (use your ohmmeter on the sensor to find pin numbers 1,2,3).
The left of Figure 8.2 shows a possible data flow graph of this system. Dividing the system into modules allows for
concurrent development and eases the reuse of code. The right of Figure 8.2 shows a possible call graph.
Position Voltage
0 to 2 cm 0 to +3.3V
Position
Sensor
Sample
0 to 4095
ADC
hardware
Sample
0 to 4095
ADC
driver
SysTick
ISR
SysTick
hardware
LCD
display
LCD
driver
Mailbox
main
Fixed-point
0 to 2.000
Figure 8.2. Data flow graph and call graph of the position meter system. Notice the hardware calls the ISR.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 49
You should make the position resolution and accuracy as good as possible using the 12-bit ADC. The position
resolution is the smallest change in position that your system can reliably detect. In other words, if the resolution were 0.01
cm and the position were to change from 1.00 to 1.01 cm, then your device would be able to recognize the change.
Resolution will depend on the amount of electrical noise, the number of ADC bits, and the resolution of the output display
software. Considering just the errors due to the 12-bit ADC, we expect the resolution to be about 2cm/4096 or about 0.0005
cm. Accuracy is defined as the absolute difference between the true position and the value measured by your device.
Accuracy is dependent on the same parameters as resolution, but in addition it is also dependent on the reproducibility of the
transducer and the quality of the calibration procedure. Long-term drift, temperature dependence, and mechanical vibrations
can also affect accuracy.
Figure 8.3. Hardware setup for Lab 8, showing the LCD and slide pot. The slide pot is used to measure distance. the display
you developed in Lab 7. (The 1k resistor shown in this figure should not be used).
In this lab, you will be measuring the position of the armature (the movable part) on the slide potentiometer, see Figure
8.3. Due to the mass of the armature and the friction between the armature and the frame, the position signal has a very low
frequency response. One way to estimate the bandwidth of the position signal is to measure the maximum velocity at which
you can move the armature. For example if you can move the armature 2 cm in 0.1sec, its velocity will be 20cm/sec. If we
model the position as a signal sine wave x(t)=1cm*sin(2πft), we calculate the maximum velocity of this sine wave to be
1cm*2πf. Therefore we estimate the maximum frequency using 20cm/sec = 1cm*2πf, to be 3 Hz. A simpler way to estimate
maximum frequency is to attempt to oscillate it as fast as possible. For example, if we can oscillate it 10 times a second, we
estimate the maximum frequency to be 10 Hz. According to the Nyquist Theorem, we need a sampling rate greater than 20
Hz. Consequently, you will create a system with a sampling rate of 40 Hz. You will sample the ADC exactly every 0.025 sec
and calculate position using decimal fixed-point with ∆ of 0.001 cm. You should display the results on the LCD, including
units. In general, when we design a system we choose a display resolution to match the expected measurement resolution.
However in this case, the expected measurement resolution is 0.0005 cm, but the display resolution is 0.001 cm. This means
the display may be the limiting factor. However we expect electrical noise and uncertainty about exactly where the
measurement point is to determine accuracy and not the display or the ADC resolution. In most data acquisition systems the
noise and transducers are significant sources of error. A SysTick interrupt will be used to establish the real-time periodic
sampling.
Nyquist Theorem: If fmax is the largest frequency component of the analog signal, then you must sample more
than twice fmax in order to faithfully represent the signal in the digital samples. For example, if the analog signal
is A + B sin(2πft + φ) and the sampling rate is greater than 2f, you will be able to determine A, B, f, and φ from
the digital samples.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 50
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Valvano Postulate: If fmax is the largest frequency component of the analog signal, then you must sample more
than ten times fmax in order for the reconstructed digital samples to look like the original signal to the human eye
when plotted on a voltage versus time graph.
When a transducer is not linear, you could use a piece-wise linear interpolation to convert the ADC sample to
position (∆ of 0.001 cm.) One approach is to use piece-wise linear interpolation. In this approach, there are two small tables
Xtable and Ytable. The Xtable contains the ADC results and the Ytable contains the corresponding positions. The
ADC sample is passed into the lookup function. This function first searches the Xtable for two adjacent of points that
surround the current ADC sample. Next, the function uses linear interpolation to find the position that corresponds to the
ADC sample. You are free to implement the conversion in any acceptable manner, however your conversion function must
be written in C.
A second approach to the conversion is to implement Cubic Interpolation. One description of Cubic Interpolation
can be found in the following document online: http://paulbourke.net/miscellaneous/interpolation/. A third approach,
shown in Figure 8.7, is to fit a linear equation to convert the ADC sample to position (∆ of 0.001 cm.)
Procedure
The basic approach to this lab will be to debug each module separately. After each module is debugged, you will
combine them one at a time. For example: 1) just the ADC; 2) ADC and LCD; and 3) ADC, LCD and SysTick.
The analog signal connected to the microcomputer comes from a position sensor, such that the analog voltage ranges
from 0 to +3.3V as the position ranges from 0 to Pmax, where Pmax may be any value from 1.5 to 2 cm. The minimum position
is established by where the paper scale is glued to the frame. Glue the paper so that a position of 0 can be measured.
In the final system, you will use SysTick interrupts to establish 40 Hz sampling. In particular, the ADC should be
started exactly every 25 msec. The SysTick ISR will store the 12-bit ADC sample in a global variable (called a MailBox) and
set a flag. Read Section 9.3 in the book to see how a Mailbox can be used to pass data from the background into the
foreground. The main program will collect data from the Mailbox and convert the ADC sample (0 to 4095) into a 32-bit
unsigned decimal fixed-point number, with a ∆ of 0.001 cm. Lastly, your main program will use your LCD_OutFix
function from the previous lab to display the sampled signal on the LCD. Include units on your display.
Part a) There are many ways to build a transducer. Your design will require cutting, gluing, and/or soldering. The system in
Figure 8.3 was created by plugging the slide pot into the protoboard and wrapping a solid wire to the armature (to create a
cursor, and gluing a metric ruler on the potientiometer. The example in Figure 8.3 used a Xerox copy of a metric ruler, which
we define as the truth. Another method is to cut out a piece of wood, cardboard, or plastic a little larger than the
potentiometer. Next, you could glue the frame (the fixed part) of the potentiometer to this solid object. Then you could glue
the metric ruler on the frame but near the armature (the movable part) of the sensor. Lastly, you could solder three solid wires
to the slide potentiometer. If you do not know how to solder, ask your TA for a lesson. There is a soldering station in the lab
floor. Pmax will depend on where/how you solder the wires to the potentiometer.
Please review safety procedures with your instructor, or TA. Please wash your hands after handling the lead solder, please
solder in a well ventilated area, and if you drop the soldering iron let it fall to the ground (don’t try to catch it). If you are
pregnant or think you might be pregnant, please do not solder.
Label these three wires +3.3 (Pin3), Vin (Pin2), and ground (Pin1), as shown in Figure 8.1. The potentiometer may have
places where it is nonlinear. The full scale range may be any value from 1.5 to 2.0 cm, depending on your potentiometer.
Part b) You will notice the Lab 8 starter project a main program Lab8.c and three submodules (ST7735, ADC, and TExaS).
The main program must be in C, but ADC, and SysTick can be in C or assembly. One possible organization is illustrated in
Figure 8.4. Each module has a header file containing the prototypes for public functions (ST7735.h, ADC.h and TExaS.h).
Use the display you developed in Lab 7. The right side of Figure 8.4 shows the call graph. The Lab8 (main) calls ADC,
ST7735, and TExaS. The ADC module accesses the ADC hardware, the ST7735 module access the SSI and LCD hardware,
and the TExaS module accesses Timer5, ADC1, PD3, UART0 and the PLL. The TExaS module will activate the PLL at 80
MHz, sample PD3 using ADC1 at 10 kHz, and send the data output UART0 to the PC so it can be displayed by
TExaSdisplay.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 51
Lab8.c
Lab7
ADC
Write
these
ADC
Slide pot
ST7735
TExaS
SSI0
ADC1
LCD
PD3
Activates PLL at 80 MHz
and creates a scope on PD3
Figure 8.4. Screen shot showing one way the project could be organized. The main program must be C and the LCD driver
must be the combination of the three files LCD.s, ST7735.c and Print.s. The ADC could be C or assembly.
Part c) Write two functions: ADC_Init will initialize the ADC interface and ADC_In will sample the ADC. You are free
to pass parameters to these two functions however you wish. You are free to use any of the ADC channels. The example code
in the book uses channel 9 and sequencer 3. You can use whichever sequencer you wish, but you must use a different analog
pin than PE4=Ain9. You cannot use PD3 because PD3 is used by TExaS. You are required to understand every line of the
your ADC code. Write main program number 1, which tests these two ADC functions. In this system, there is no LCD, and
there are no interrupts. Debug this system on the real TM4C123 to show the sensor and ADC are operational. The first main
program will look something like the following:
#include "TExaS.h"
#include "ADC.h"
uint32_t Data;
int main(void){
TExaS_Init();
ADC_Init();
while(1){
Data = ADC_In();
}
}
// single step this program and look at Data
// Bus clock is 80 MHz
// turn on ADC, set channel to 1, PE2
// sample 12-bit channel 1, PE2
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 52
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Figure 8.5. Screen shot showing one way to debug the ADC software on the board. You adjust the slide potentiometer, and
then see the ADC result. You can observe the ADC result in four places. The ADC result is in R0 after the call to the
function; it is stored in the variable Data; and it is in the ADC register SSFIFO3.
Part d) Write main program number 2, which you can use to collect calibration data. In particular, this system should first
sample the ADC and then display the results as unsigned decimal numbers. In this system, there is no mailbox, and there are
no interrupts. You should use your LCD_OutDec developed in the previous lab. Connect PF2 to a real oscilloscope or real
logic analyzer and use it to measure the time it takes the ADC to perform one conversion, and the time it takes the LCD to
output one number. Figure 8.6 shows for this system the ADC requires 9 μsec to convert and the LCD requires (you measure
this) μsec to display. The second main program will look something like the following:
uint32_t Data;
int main(void){
TExaS_Init();
ADC_Init();
LCD_Init();
PortF_Init();
while(1){
PF2 = 0x04;
Data = ADC_In();
PF2 = 0x00;
LCD_Goto(0,0);
PF1 = 0x02;
LCD_OutDec(Data);
LCD_OutString("
PF1 = 0;
}
}
// Bus clock is 80 MHz
// turn on ADC, set channel to 1
//
//
//
//
use scope to measure execution time for ADC_In and LCD_OutDec
Profile ADC
sample 12-bit channel 1
end of ADC Profile
// Profile LCD
"); // these spaces are used to coverup characters from last output
// end of LCD Profile
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 53
Figure 8.6. Oscilloscope trace showing the execution time profile. There are a lot of parameters that effect speed, so please
expect your time measurements to be different. For this system it takes 9μs for the ADC to sample and 948 μs to output.
Collect five to ten calibration points and create a table showing the true position (as determined by reading the position of the
hair-line on the ruler), the analog input (measured with a digital voltmeter) and the ADC sample (measured with main
program 2). The full scale range of your slide pot will be different from the slide pot of the other students, which will affect
the gain (voltage versus position slope). Where you attach the paper ruler will affect the offset. Do not use the data in Table
8.1. Rather, collect your own data like the first three columns of Table 8.1.
Position
Analog input
ADC sample
Correct Fixed-point
Measured Fixed-point Output
0.10 cm
0.432
538
100
84
0.40 cm
1.043
1295
400
421
0.80 cm
1.722
2140
800
797
1.20 cm
2.455
3050
1200
1202
1.40 cm
2.796
3474
1400
1391
Table 8.1. Calibration results of the conversion from ADC sample to fixed-point (collect your own data).
Part e) Use this calibration data to write a function in C that converts a 12-bit binary ADC sample into a 32-bit unsigned
fixed-point number. The input parameter (12-bit ADC sample) to the function will be passed by value, and your function will
return the result (integer portion of the fixed-point number). Table 8.1 shows some example results; this data is plotted in
Figure 8.7. You are allowed to use a linear equation to convert the ADC sample into the fixed-point number. Please consider
overflow and dropout during this conversion. In this system, the calculation Position = 0.4455*Data-155.38 can be
approximated as
Position = (114*Data-39777)/256;
// DO NOT USE THESE CONSTANTS, MEASURE YOUR OWN
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 54
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Figure 8.7. Plot of the conversion transfer function calculating fixed-point distance as a function of ADC sample.
The third main program will look something like
uint32_t Data;
// 12-bit ADC
uint32_t Position; // 32-bit fixed-point 0.001 cm
int main(void){
TExaS_Init();
// Bus clock is 80 MHz
LCD_Init();
PortF_Init();
ADC_Init();
// turn on ADC, set channel to 1
while(1){
PF2 ^= 0x04;
// Heartbeat
Data = ADC_In(); // sample 12-bit channel 1
PF3 = 0x08;
// Profile Convert
Position = Convert(Data);
PF3 = 0;
// end of Convert Profile
PF1 = 0x02;
// Profile LCD
LCD_Goto(0,0);
LCD_OutDec(Data); LCD_OutString("
");
LCD_Goto(6,0);
LCD_OutFix(Position);
PF1 = 0;
// end of LCD Profile
}
}
Part f) Write a C function: SysTick_Init will initialize the SysTick system to interrupt at exactly 40 Hz (every 0.025
second). If you used SysTick to implement the blind wait for the LCD driver, you will have to go back to Lab 7 and remove
all accesses to SysTick from the LCD driver. If you did not use SysTick for the LCD waits, then there is no conflict, and you
can use SysTick code similar to Lab 6 to implement the 25-ms periodic interrupt needed for this lab.
Part g) Write a C SysTick interrupt handler that samples the ADC and enters the data in the mailbox. Using the interrupt
synchronization, the ADC will be sampled at equal time intervals. The toggle a heartbead LED each time the ADC is
sampled. The frequency of the toggle is a measure of the sampling rate. If you connect the scope to the LED pin, you can also
measure the execution time of the ISR. The interrupt service routine performs these tasks
1) toggle heartbead LED (change from 0 to 1, or from 1 to 0)
2) toggle heartbead LED (change from 0 to 1, or from 1 to 0)
3) sample the ADC
4) save the 12-bit ADC sample into the mailbox ADCMail
5) set the mailbox flag ADCStatus to signify new data is available
6) toggle heartbead LED (change from 0 to 1, or from 1 to 0)
7) return from interrupt
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 55
Part h) Write main program number 4, which initializes the PLL, timer, LCD, ADC and SysTick interrupts. After
initialization, this main program (foreground) performs these five tasks over and over.
1) wait for the mailbox flag ADCStatus to be true
2) read the 12-bit ADC sample from the mailbox ADCMail
3) clear the mailbox flag ADCStatus to signify the mailbox is now empty
4) convert the sample into a fixed-point number (variable integer is 0 to 2000)
5) output the fixed-point number on the LCD with units (LCD_GoTo is faster than LCD_Clear)
Debug this system on the real TM4C123. Use a logic analyzer or oscilloscope to observe the sampling rate. Take a
photograph or screen shot of the LED toggling that verifies the sampling rate is exactly 40 Hz, see Figure 8.8.
Figure 8.8. The oscilloscope connected to the LED pin verifies the sampling rate is 40 Hz. If you were to zoom in, you could
also measure the execution time of the ISR.
Part i) Use the system to collect another five to ten data points, creating a table showing the true position (xti as determined
by reading the position of the hair-line on the ruler), and measured position (xmi using your device). You can use the
interactive tool at http://users.ece.utexas.edu/~valvano/Volume1/Lab8Accuracy.html to calculate average accuracy by
calculating the average difference between truth and measurement,
1 n
Average accuracy (with units in cm) =
∑ x ti − x mi
n i =1
True position
xti
Measured Position
xmi
Error
xti - xmi
Table 8.2. Accuracy results of the position measurement system.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 56
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Demonstration (both partners must be present, and demonstration grades for partners may be different)
You will show the TA your program operation on the actual TM4C123 board. The TA may look at your data and
expect you to understand how the data was collected and how the ADC and interrupts work. You should be able to explain
how the potentiometer converts distance into resistance, and how the circuit converts resistance into voltage. Also be
prepared to explain how your software works and to discuss other ways the problem could have been solved. What will you
change in your program if the potentiometer were to be connected to a different ADC pin? How would this system be
different if the units of measurement were inches instead of cm? What’s your sampling rate? What do you mean by sampling
rate? What is the ADC range, resolution and precision? How do you initialize SysTick interrupt? How can you change your
sampling rate? Be prepared to prove what the sampling rate is using a calculator and the manual. Explain how, when an
interrupt occurs, control reaches the interrupt service routine. Why is it extremely poor style to output the converted data to
the LCD inside the SysTick ISR? Where is the interrupt vector located? What are the differences between an interrupt and a
subroutine? What will happen if you increase your sampling rate a lot? At what point do you think your program will crash?
What is the Nyquist Theorem? How does it apply to this lab?
Deliverables (Items 1, 2, 3, 4, 5 are one pdf file uploaded to SVN, have this file open during demo.)
0) Lab 8 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
Create one pdf file with the following for electronic submission (UTEID1_UTEID2.pdf)
1) Circuit diagram showing the position sensor, hand-drawn or PCB Artist, like Figure 8.1 (part a),
2) Time measurements and a photo showing the ADC/LCD execution time profile, like Figure 8.6 (part d)
3) Calibration data, like the first three columns of Table 8.1 (part d)
4) A photo or screenshot verifying the sampling rate is 40 Hz, like Figure 8.8 (part h)
5) Accuracy data and accuracy calculation, Table 8.2 (part i)
Commit to SVN the final version of distance meter with SysTick, ADC, convert, and main (parts c, e, f, g and h). All source
files that you have changed or added should be committed to SVN. Please do not commit other file types.
Hints
1) Debug this lab in parts: debugging each module separately. If you combine two working systems (e.g., ADC and
LCD), you should retest the system.
2) There are lots of details in this lab, please ask your instructor or your TA if you are confused.
3) Use your voltmeter to measure the voltage at the ADC input
4) A useful debugging monitor is to count the number of interrupts (to see if the interrupts are occurring)
5) When running on the real TM4C123, you cannot use breakpoints effectively in a real-time system like this,
because it takes too long to service a breakpoint (highly intrusive). However, you can dump strategic variables into memory
and view that data in a memory window.
6) It takes a long time for the LCD to initialize. It is necessary to allow the LCD to finish initialization before
starting the SysTick interrupt sampling. One way to do this is to enable interrupts only after all initializations are complete.
7) The fourth column of Table 8.1 is the desired software output, assuming a linear fit between ADC and position.
The fifth column of Table 8.1 is the actual software output using the linear equation implemented as integer operations.
Be careful when connecting the potentiometer to the computer, because if you mistakenly reverse two of the wires, you can
cause a short from +3.3V to ground.
Do not use your Lab 7 project because the startup.s file for an assembly main (Lab 7) is different from the startup.s file for a
C language main (Lab 8). Use the EE319K_Lab8 starter project.
We expect everyone to use their Lab 7 LCD driver for Lab 8. However, if you cannot get your Lab 7 LCD driver to work,
you could use the UART0 instead UART_xxx.zip. The UART0 generates output to a PuTTy terminal on your PC. You will
need to write a new OutFix function. However there will be a 10 points penalty if you do not use your Lab 7 software. To use
the UART0, you must remove TExaS and add the PLL. The ADC requires the PLL to be active.
If your Lab 7 solution works, but your LCD is broken, ask the TA to show you how to run with the Nokia emulator (runs on
the real board, but displays in a window on the PC).
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 57
Lab 9. Distributed Data Acquisition System
Preparation
Read Sections 8.2, and 11.4 on the UART
Read Sections 11.1, 11.2, and 11.3 on communication systems and the FIFO queues
Download these examples FIFO_xxx.zip, UART_xxx.zip, and UARTInts_xxx
(*****WARNING: YOU ARE RESPONSIBLE FOR EVERY LINE OF CODE IN YOUR SOLUTION except for the LCD
software in ST7735.c/Nokia5110.c and any submodules the software in ST7735.c/Nokia5110.c calls *****). You will copy
your Lab 8 solution into the Lab 9 folder.
Teams
To fully test your system you will need two boards. However, you can perform initial testing by connecting the
transmitter output to the receiver input on one board. We suggest you find another group with whom you will finish testing
your system. You will not be checking out the same team as a team with whom you debug. Both groups will have to design,
build and test separate systems. That is you WILL NOT split the lab into two parts with one team working on the Sender and
other working on the Receiver. Each team is responsible for their own system. At the time of checkout, the TA will select
another group with whom you will be checked out. It is also possible for there to be a TA board with which you will check
out. During checkout your system must operate as both a transmitter and a receiver. You cannot share code with other
EE319K teams past, present, or future.
Purpose
The objectives of this lab are: 1) to learn how the UART works using both busy-wait and interrupt synchronization;
2) to employ a FIFO queue to buffer data; and 3) to study synchronization issues in a distributed data acquisition system. This
lab is demonstrated on two TM4C123 boards. There is no simulation in Lab 9.
System Requirements
You will extend the system from Lab 8 to implement a distributed system, see Figure 9.1. In particular, each
TM4C123 will sample the data at 40 Hz, transmit the data to the other system, and the other system will display the results on
its LCD. Basically the hardware/software components from Lab 8 will be divided and distributed across two TM4C123
microcontrollers. Communication will be full duplex, so the slide pot position will be displayed to the other system.
Figure 9.1. The LM3S1968 on the right measures position and the one on the left displays the results. Both systems show the
optional debugging monitors on the OLEDs. (***needs a new photo, but the basic idea is the same***)
Figure 9.2 shows a possible data flow graph of the distributed data acquisition system.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 58
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Position
0 to 2 cm
Position
0 to 2 cm
Message
Position
Sensor
STX d1 .
d2 d3 d4 CR ETX
LCD
display
Voltage
0 to +3.3V
ADC
hardware
LCD
driver
Sample
0 to 4095
ADC
driver
Fixed-point
0 to 2.000
SysTick
hardware
Sample
0 to 4095
main
SysTick
ISR
UART1
Tx Rx
UART1
driver
Message
Message
Message
Message
UART1
ISR
FIFO
Computer 1
Message
Computer 2
Message
Message
UART1
ISR
FIFO
Message
Message
Rx
Tx
UART1
UART1
driver
SysTick
ISR
Message
Message
Sample
0 to 4095
SysTick
hardware
main
ADC
driver
Sample
0 to 4095
Fixed-point
0 to 2.000
LCD
driver
ADC
hardware
Voltage
0 to +3.3V
LCD
display
Position
Sensor
Position
0 to 2 cm
Position
0 to 2 cm
Figure 9.2. Data flows from the sensor on one computer to the LCD on the other computer. SysTick interrupts are used to
trigger the real-time ADC sampling. You will use a 3-wire serial cable to connect the two UART ports. The full scale range of
distance is exactly the same as Lab 8 and depends on the physical size of the slide potentiometer. Shown here as UART1, but
any available UART can be used.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 59
A sensor is attached to each computer, and the ADC (ADC_In function) generates a 12-bit digital value from 0 to 4095. The
SysTick periodic interrupt establishes the real-time ADC sampling at 40 Hz. Your convert function will create a calibrated
fixed-point representation of the distance. The calibration conversion occurs on the same computer as the sensor. You will
then encode this information into an 8-byte message and send the message from one computer to the other computer using
the universal asynchronous serial transmitter (UART). There are five choices for the UART hardware:
• PC4, PC5 UART1
• PC4, PC5 UART4
• PC5, PC6 UART3
• PE0, PE1 UART7
• PD6, PD7 UART2 (PD7 will need unlocking, similar to the way PF0 needed unlocking)
The UART transmitter must use busy-wait synchronization, and the UART receiver must use interrupt synchronization. In
order to allow your board to operate with any other EE319K student’s board, we will fix the communication protocol. Each
measurement will be sent as exactly eight serial frames at 100,000 bits/sec:
STX, ASCII1, '.', ASCII2, ASCII3, ASCII4, CR, ETX
For example, if the distance is 1.234 cm, then the following message should be sent.
0x02, 0x31, 0x2E, 0x32, 0x33, 0x34, 0x0D, 0x03
Students can borrow a 3-pin cable from the lab to use to connect the UART ports between two TM4C123 boards. One wire
connects the grounds together, and the other two wires connect Tx to Rx and Rx to Tx. Ask your TA to cut you a section of
cable.
If you were transmitting on a network with more than two nodes, the protocol would need to include the destination address.
Furthermore, if you wanted to transmit other types of data, the protocol would need include data type. If the channel were
noisy, then extra frames could be added to the message to detect and correct errors. Network addressing, message type
specification, data compression, error detection, and error correction are not to be implemented in this lab.
Figure 9.3 shows a possible call graph of the system. The main program will initialize SysTick, the ADC, the FIFO, the
LCD and the UART, but after initialization the main program only performs receiving functions. Notice the race condition
where main can be calling Get at the same time as the UART ISR calls Put. The call graph shows UART1, but you can
use any of the available UARTs. There is no race condition for the shared UART because the transmitting and receiving
functions are separate. Dividing the system into modules allows for concurrent development and eases the reuse of code.
Init
SysTick
ISR
SysTick
hardware
In
Init
ADC driver
ADC
hardware
Transmitting
Receiving
main
Init
Out UART driver
ISR
UART1
hardware
Init Get
FIFO
Put
LCD
driver
LCD
hardware
Figure 9.3. A call graph showing the modules used by the distributed data acquisition system.
There are many options for arming interrupts for the receiver. However, we require you to arm just RXRIS with ½ full FIFO.
This way you get one interrupt for each 8-frame message received. One method, as illustrated in the book Program 11.6, but
not allowed in Lab 9, arms both RXRIS (receive FIFO 1/8 full) and RTRIS (receiver idle). This way the bytes are passed one
at a time through the FIFO from ISR to main. The main program would get the bytes one at a time and process the message.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 60
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Another method, good approach but not allowed in Lab 9, is to arm just RTRIS. This way you get one interrupt a short time
after all 8 bytes have been received. The ISR would read all received bytes, decode the message, and then put the string into a
software FIFO. This main program would get the string from the software FIFO and output it to the LCD. If the average rate
at which the ISR puts data into the FIFO is slower than the rate at which data can be sent to the LCD, then the software FIFO
linking the ISR and main program will never become full. The main program in each computer will output the position of the
other sensor on its LCD.
Think about what the baud rate means and how it is used. Can you find a mathematical relationship between the ADC
sampling rate (assuming the 8-byte message is transmitted in each sample) and the slowest baud rate possible for the
transmission UART? For example, if I were to reduce the ADC is sampling rate, then I could also reduce the baud rate of the
transmission UART. Think about how the baud rate for the receiver is selected. Is there an advantage to setting the baud rate
of the receiver UART faster or slower than the baud rate of the transmitter? Assume that hardware and software FIFOs are
deep enough to not affect these decisions.
Procedure (shown below as UART1, but could be any of the available UARTs)
Transmitter software tasks
Part a) Write a C function: UART1_Init that will initialize your UART1 transmitter. This will be similar to book Program
8.1, but with UART1 and with 100,000 bps baud rate. Busy-wait synchronization will be used in the transmitter.
1) enable UART1 transmitter (no interrupts)
2) set the baud rate
Part b) Write a C function: UART1_OutChar for the transmitter that sends one byte using busy-wait synchronization. Place
all the UART1 routines in a separate UART1.s or UART1.c file.
1) Wait for TXFF in UART1_FR_R to be 0
2) Write a byte to UART1_DR_R
Part c) Create a simple array to store the 8 bytes needed for the message. Modify the SysTick interrupt handler from Lab 8
so that it samples the ADC at 40 Hz and sends the 8-frame message to the other computer using UART1. The SysTick
interrupt service routine performs these tasks
1) toggle a heartbeat (change from 0 to 1, or from 1 to 0),
2) sample the ADC
3) toggle a heartbeat (change from 0 to 1, or from 1 to 0),
4) convert to distance and create the 8-byte message
5) send the 8-byte message to the other computer (calls UART1_OutChar 8 times)
6) increment a TxCounter, used as debugging monitor of the number of ADC samples collected
7) toggle a heartbeat (change from 0 to 1, or from 1 to 0),
8) return from interrupt
Toggling the heartbeat three times allows you to see the LED toggle every 25ms, as well as measure the total execution time
of the ISR using an oscilloscope attached to the heartbeat. To understand the transmitter better answer these questions
Q1) The UART transmitter has a built-in hardware FIFO. How deep is the FIFO? (How many frames can it store?)
Q2) At a baud rate of 100,000 bits/sec (1 start, 8 data, 1 stop) how long does it take to send 8 frames?
Q3) The SysTick ISR runs every 25 ms, and each ISR calls UART1_OutChar 8 times. Does this hardware FIFO
ever fill? In other words will the busy-wait loop in UART1_OutChar ever loop back?
Part d) Write a temporary main program for testing the transmitter, which initializes the PLL, SysTick, ADC, UART1,
heartbeat, and enables interrupts. After initialization, this transmitter test main (foreground) performs a do-nothing loop. This
main program will only be used for testing, and will be replaced by a more complex main once the receiver is built. The ADC
sampling and UART1 transmissions occur in the SysTick interrupt service routine running in the background. It is possible,
but not required to debug the system in simulation mode. Figure 9.4 shows the transmitter being debugged in simulation
mode. Notice the various representations of the 1.424 cm position signal, the analog input voltage 1.5 V on the ADC, and the
02 31 2e 34 32 34 0d 03 message in the UART debugging window.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 61
Figure 9.4. Example screenshot showing the transmitter system in simulation mode. We are currently unable to simulate the
UART in interrupting mode .
To debug the transmitter on the real TM4C123, you can place the ADC sample and the interrupt counter in a
debugger Watch window. The interrupt TxCounter should increase by 40 every second, and the 12-bit ADC sample should
respond to changes in the slide pot. Attach a dual-channel oscilloscope to the heartbeat and UART outputs, you will see the 8
frames being transmitted at 100,000 bits/sec every 25 ms. Take two photographs of the scope output, one zoomed in to see
the first two frames, and one zoomed out to see the 25ms time between messages, as shown in Figures 9.5 and 9.6.
Figure 9.5. Example scope output measured on the transmitter. The top trace is heartbeat showing the SysTick interrupt has
occurred and the total execution time of the SysTick ISR is 42µs. The bottom trace is U1Tx showing the first frame (STX
=0x02, frame is start, bit0, bit1, bit2, bit3, bit4, bit5, bit6, bit7, and stop). The width of bit1 is 10µs, showing the baud rate is
100,000 bits/sec.
Questions to think about:
Q4) If the SysTick ISR sampled the ADC and performed 8 UART outputs, why did this ISR finish in only 42 µs?
Q5) Using the photo captured in Figure 9.5, approximately how long does it take to sample the ADC?
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 62
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Figure 9.6. Example scope output measured on the transmitter. The top trace is PG2 showing the SysTick interrupt is
occurring every 40 ms. The bottom trace is PD2 (U1Tx). The 8 frames of PD2 occur in 800 µs, so on this scale one cannot
see the individual frames. (This data is from Spring 2013. Your measurements should show SysTick occurring every 25ms)
Receiver software tasks
Part e) Design, implement and test a C language module implementing a statically allocated FIFO queue. You are allowed to
look at programs in the book and posted on the internet, but you are required to write your own FIFO. Examples of how to
make your FIFO different from the ones in the book include 1) store the data so the oldest is on top and one size counter
defines the FIFO status; 2) implement it so 5, 6, or 8 bytes are put and get at a time; 3) reverse the direction of the pointers so
both PutPt and GetPt decrement rather than increment; Make sure your FIFO is big enough to hold all data from one
message. Place the FIFO code in a separate fifo.c file. Add a fifo.h file that includes the prototypes for the functions.
The software design steps are
1) Define the names of the functions, input/output parameters, and calling sequence. Type these definitions in as comments
that exist at the top of the functions.
2) Write pseudo-code for the operations. Type the sequences of operations as comments that exist within the bodies of the
functions.
3) Write C code to handle the usual cases. I.e., at first, assume the FIFO is not full on a put, not empty on a get, and the
pointers do not need to be wrapped.
4) Write a main program in C to test the FIFO operations. An example main program is listed below. The example test
program assumes the FIFO can hold up to 6 elements. This main program has no UART, no LCD and no interrupts. Simply
make calls to the three FIFO routines and visualize the FIFO before, during and after the calls. Debug either in the simulator
or on the real board. In this example, Put and Get return a true if successful, Put takes a call by value input, and Get takes a
return by reference parameter. You are free to design the FIFO however you wish as long as it is different from the ones in
the book, and you full understand how it works.
5) Iterate steps 3 and 4 adding code to handle the special cases.
#include “fifo.h”
int Status[20];
// entries 0,7,12,19 should be false, others true
unsigned char GetData[10]; // entries 1 2 3 4 5 6 7 8 should be 1 2 3 4 5 6 7 8
int main(void){
Fifo_Init();
for(;;){
Status[0] = Fifo_Get(&GetData[0]); // should fail,
empty
Status[1] = Fifo_Put(1);
// should succeed, 1
Status[2] = Fifo_Put(2);
// should succeed, 1 2
Status[3] = Fifo_Put(3);
// should succeed, 1 2 3
08/09/15
Jon Valvano
Status[4]
Status[5]
Status[6]
Status[7]
Status[8]
Status[9]
Status[10]
Status[11]
Status[12]
Status[13]
Status[14]
Status[15]
Status[16]
Status[17]
Status[18]
Status[19]
EE319K Fall 2015 Lab Manual
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
Fifo_Put(4);
Fifo_Put(5);
Fifo_Put(6);
Fifo_Put(7);
Fifo_Get(&GetData[1]);
Fifo_Get(&GetData[2]);
Fifo_Put(7);
Fifo_Put(8);
Fifo_Put(9);
Fifo_Get(&GetData[3]);
Fifo_Get(&GetData[4]);
Fifo_Get(&GetData[5]);
Fifo_Get(&GetData[6]);
Fifo_Get(&GetData[7]);
Fifo_Get(&GetData[8]);
Fifo_Get(&GetData[9]);
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
should
should
should
should
should
should
should
should
should
should
should
should
should
should
should
should
succeed,
succeed,
succeed,
fail,
succeed,
succeed,
succeed,
succeed,
fail,
succeed,
succeed,
succeed,
succeed,
succeed,
succeed,
fail,
Page 63
1 2 3
1 2 3
1 2 3
1 2 3
2 3 4
3 4 5
3 4 5
3 4 5
3 4 5
4 5 6
5 6 7
6 7 8
7 8
8
empty
empty
4
4
4
4
5
6
6
6
6
7
8
5
5 6
5 6
6
7
7 8
7 8
8
}
}
Figure 9.7. Example simulator screen shot debugging the FIFO.
Part f) Write a C function: UART1_Init that will initialize the UART1 receiver for the receiver.
1) clear a global error count, initialize your FIFO (calls Fifo_Init)
2) enable UART1 receiver (arm interrupts for receiving)
3) set the baud rate to 100,000 bits/sec
In general, one must enable interrupts in order for UART1 interrupts to occur. It is good style however to finish all
initializations, then enable interrupts from the high-level main program. The receiver initialization with interrupts should be
added to the transmitter initialization using busy-wait.
Part g) Write the UART interrupt handler in C that receives data from the other computer and puts them into your FIFO
queue. Your software FIFO buffers data between the ISR receiving data and the main program displaying the data. If your
software FIFO is full, increment a global error count. If you get software FIFO full errors, then you have a bug that must be
removed. Do not worry about software FIFO empty situations. The software FIFO being empty is not an error. Define the
error count in the UART.c file and increment it in the UART ISR.
Arm RXRIS receive FIFO ½ full (interrupts when there are 8 frames): the interrupt service routine performs these tasks
1) toggle a heartbeat (change from 0 to 1, or from 1 to 0),
2) toggle a heartbeat (change from 0 to 1, or from 1 to 0),
3) as long as the RXFE bit in the UART1_FR_R is zero
Read bytes from UART1_DR_R
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 64
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Put all bytes into your software FIFO (should be exactly 8 bytes, but could be more possibly)
If your software FIFO is full (data lost) increment a global error count (but don’t loop back)
(the message will be interpreted in the main program)
4) Increment a RxCounter, used as debugging monitor of the number of UART messages received
5) acknowledge the interrupt by clearing the flag which requested the interrupt
UART1_ICR_R = 0x10; // this clears bit 4 (RXRIS) in the RIS register
6) toggle a heartbeat (change from 0 to 1, or from 1 to 0),
7) return from interrupt
Toggling heartbeat three times allows you to see the LED flash with your eyes, as well as measure the total execution time of
the ISR using an oscilloscope attached to heartbeat. Place all the UART1 routines in a separate UART1.s file (or UART1.h
and UART1.c). This driver is not the same as the UART driver in computer 1.
Part h) The body of the main program reads data from the FIFO and displays the measurement using the same LCD routines
developed in Lab 7 and used in Lab 8. The main program in this data acquisition system performs these tasks
1) initialize PLL
2) initialize transmitter functions (ADC, heartbeat, SysTick, UART transmission busy-wait)
3) initialize receiver functions (FIFO, LCD, heartbeat, UART receiver with interrupts)
4) enable interrupts
5) calls your FIFO get waiting until new data arrives. FIFO should return an Empty condition if the main program in the
receiver calls get and the FIFO is empty. For more information on how the ISR and main use the FIFO, read Section 11.3 in
the book and look at the left side of Figure 11.6. Also read Section 11.4 and look the the function UART_InChar in
Program 11.6. In this step you keep reading until you receive an STX,
6) output the fixed-point number (same format as Lab 8) with units.
The next five characters gotten from the FIFO should be the ASCII representation of the distance
7) repeat steps 5,6 over and over
To debug the system on one TM4C123, connect the transmitter output to the receiver input. In this mode, the system
should operate like Lab 8, where the sensor position is displayed on the LCD. Use the oscilloscope to visualize data like
Figures 9.5 and 9.6.
To debug the distributed system on the real TM4C123, use two PC computers close enough together so you can
connect the two TM4C123 LaunchPads with the special cable, yet run uVision4 on both computers. You can place strategic
variables (e.g., ADC sample, interrupt counter) in the Watch windows on both PCs. If you start both computers at the same
time all four interrupt counters should be approximately equal, and the data on both computers should respond to changes in
the slide pot. For the lab to work completely, the transmitter and receivers must receive synchronize properly regardless of
which computer is turned on first, so try turning the boards on/off in different orders and at different times
Demonstration (both partners must be present, and demonstration grades for partners may be different)
You will show the TA your program operation on the two TM4C123 boards. Your TA will connect a scope to either
transmitter heartbeat and TxD→RxD or receiver heartbeat and TxD→RxD and expect you to explain the relationship
between your executing software and the signals displayed on the scope. Also be prepared to explain how your software
works and to discuss other ways the problem could have been solved. How do you initialize UART? How do you input and
output using UART? What is the difference between busy-wait and interrupt synchronization? What synchronization method
does the transmitter UART use? What synchronization method does the receiver UART use? What sets RXFE, TXFF,
RXRIS, RTRIS and when are they set? What clears RXFE, TXFF, RXRIS, RTRIS and when are they cleared? What does the
PLL do? Why is the PLL used? There are both hardware and software FIFOs in this system. There are lots of FIFO code in
the book and on the web that you are encouraged to look at, but you are responsible for knowing how your FIFO works.
What does it mean if the FIFO is full? Empty? What should your system do if the FIFO is full when it calls PUT? Why?
What should your system do if the FIFO is empty when it calls GET? Why?
Deliverables (Items 1, 2, 3 are one pdf file uploaded to SVN, have this file open during demo.)
0) Lab 9 grading sheet. You can print it yourself or pick up a copy in lab. You fill out the information at the top.
Create one pdf file with the following for uploading to SVN
1) A circuit diagram showing position sensor, the UART, and the LCD.
2) Two photographs of a scope trace showing serial transmission data, like Figures 9.5 and 9.6.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 65
Commit to SVN the final version of the software that implements both transmitting and receiving on one computer. All
source files that you have changed or added should be committed to SVN. Please do not commit other file types.
Hints
1) Remember the port name UART1_DR_R links to two hardware FIFOs at the same address. A write to UART1_DR_R is
put into the transmit hardware FIFO and a read from UART1_DR_R gets from the receive hardware FIFO
2) You should be able to debug your system with any other EE319K group. You are not allowed to look at software written
by another group, and you certainly are not allowed to copy software from another group. You are allowed to discuss
strategies, flowcharts, pseudocode, and debugging techniques.
3) This system can be debugged on one board. Connect TxD to RxD.
4) This hint describes in more detail how to create a message. Assume the distance is 1.424 and your software calculates the
appropriate integer 1424. First you convert the digits to 1, 4, 2, 4 in a similar way as LCD_OutFix. Next, you make the digits
ASCII 0x31, 0x34, 0x32, 0x34. Add the STX(02), the decimal point (2E), the CR (0D) and ETX (03), and the message will
be 0x02, 0x31, 0x2E, 0x34, 0x32, 0x34, 0x0D, 0x03. To send this message you call UART_OutChar 8 times using busywait. These 8 bytes will go in the hardware FIFO of the UART transmitter. The 8 bytes will be sent one at a time, one byte
every 100usec.
5) In order to synchronize the computers regardless of which one is turned on first, the receiver must search a variable
number of input frames until it finds an STX. After receiving an STX, the next 7 frames are a message. Furthermore, the
receiver should discard messages that are not properly formatted (e.g., no decimal point, decimal digits, CR, or ETX)
Figure 9.8. Logic analyzer trace showing heartbeat in the transmitter ISR (PF1), the serial line (UART), heartbeat in the
receiver main (PF3) and the SSI signals in the receiver. Time to execute the transmitter ISR is 500us because hardware
averaging is selected for the ADC. The time to send the message is 800us, and the time to output to the LCD is 100us. We
see an overall delay of 1.4 ms from the start of the ADC to the completion of the output on the LCD.
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 66
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
Lab 10. Embedded System Design
Preparation
Use the starter project on SVN, paste in your LCD solutio (place BMP images on the LCD)
Restrictions on cost
To win the competition, you are restricted to only use parts that were given to you as part of one of the other labs,
plus additional resistors to make more bits in your DAC. If you wish to use extra parts not given as one of the labs, such as
joy sticks, speaker, and amplifiers, you may do so, but will be disqualified from winning the competition, unless specifically
approved by your professor.
Teams
Lab 10 will be performed in teams of two. However, you are free to select a new partner for Lab 10, or to remain
with your regular partner. If you are not working with your regular partner, the old partner and your TA must be notified by
the first Friday occurring after your Lab 9 due date. This advanced warning will give your old partner time to find an
alternative partner.
Purpose
The objectives of this lab are: 1) design, test, and debug a large C program; 2) to review I/O interfacing techniques
used in this class; and 3) to design a system that performs a useful task. There are four options for Lab 10. Option 1 is to
design a 80’s-style shoot-em up game like Space Invaders. Option 2 is to design a turn-based like Connect Four. Option 3
is to design a Pipe Dream game as specified below. The fourth option is to propose an alternative project similar in scope to
these other options. If you would like to propose a project prepare a one-page description by the first Monday after your Lab
9 due date. Good grades will be given to projects that have simple implementations, are functionally complete, and are
finished on time. Significant grade reductions will occur if the project is not completed on time.
Interrupts must be appropriately used control the input/output, and will make a profound impact on how the user
interacts with the game. You could use an edge-triggered interrupt to execute software whenever a button is pressed. You
could output sounds with the DAC using a fixed-frequency periodic interrupts. You could decide to move a sprite using a
periodic interrupt, although the actual LCD output should always be performed in the main program.
System Requirements for all games
• There must be at least one externally-interfaced button and one slide pot. Buttons and slide pot must affect game
play. The slide pot must be sampled by the ADC.
• There must be at least three images on the LCD display that move in relation to user input and/or time.
• There must be sounds appropriate for the game, generated by the DAC
developed in Lab 7. However, the interrupt can be fixed period.
• The score should be displayed on the screen (but it could be displayed
before or after the game action).
• At least two interrupt ISRs must used in appropriate manners.
• The game must be both simple to learn and fun to play.
System Requirements for Space Invaders, Asteroids, Missile Command,
Centipede, Snood, or Defender
You will design, implement and debug 80’s or 90’s-style video game. You
are free to simplify the rules but your game should be recognizable as one of these
six simple games. Buttons and the slide pot are inputs, and the LCD and sound (Lab
7) are the outputs. The slide pot is a simple yet effective means to move your ship.
Figure 10.1. Space Invaders
http://www.classicgaming.cc/classics/spaceinvaders/index.php
System Requirements for Connect Four
You will design, implement and debug a Connect Four game. Buttons and slide pots are inputs, and the LCD and
sound (Lab 7) are the outputs. The slide pot is a simple yet effective means to select which column to enter your chip. There
are two possible modes for the game. In the required mode, a human plays against artificial intelligence running on the
microcontroller. In addition to the above requirements, Connect Four must also satisfy
• The colored pieces must move on the LCD.
08/09/15
Jon Valvano
•
•
•
•
•
EE319K Fall 2015 Lab Manual
Page 67
You must implement at least a 6-row by 7-column board, but are allowed to create more rows or columns.
Aside from size of the board, you must implement the official Connect Four rules.
You must check for wins horizontal, vertical, and diagonals. You should also check for a draw.
You must not allow for illegal moves.
The game must have a man versus machine mode.
For more fun, you have the option of creating a two computer mode, where the AI in one microcontroller plays against a
second AI running on a second microcontroller. A similar fun option is having two microcontrollers that allow one human
looking at one display to play against another human looking at a second display. The boards are connected via the UART1
ports, like Lab 9, but with full duplex communication. When running in two computer mode, you need to specify baud rate,
communication protocol, and determine which side goes first. Both computers should recognize that a winner has occurred. If
you implement a two-computer option, it must be done in addition to the required man versus machine mode mode.
System Requirements for Pipe Dream
Pipe Dream is a game played on a grid. The player receives pipe pieces that go on the grid. Possible pieces are a 90
degree turn, a straight piece, and a cross (two straight pieces on top of each other.) Players are given these pieces randomly
and cannot rotate them, and they cannot replace any already placed pieces. Which piece is to be placed next is shown to the
player. The object of the game is to construct a path for water to flow using the pipe pieces given from a designated starting
point on the grid to another designated end point before time runs out. Players receive a score based on how long their final
path of piping was, but if they fail to connect the two points the game will be over. Inputs will select a grid space, and a
single button to place the next piece available in the selected space. There will be a sound for a pipe section being placed, as
well as game over and win sounds. This game must satisfy the System Requirements for all games listed above. Possible
additions include (but not required):
• Use two slide pots for moving
• Add a “queue” of pieces that shows not only the next piece to be placed, but the next 3-5 pieces as well
• Add obstacles on the grid where pipes may not be placed (or can be placed with a score penalty)
• Add already placed pipe sections that give bonus points if used in the solution
• Instead of a time limit, have a flow of water start running through the pipes. If the water
gets to an open pipe the player loses, and if the water gets to the end point the player wins.
Figure 10.2 A pipe dream screen.
System Requirements for Design Your Own Project
The first step is to make a proposal. In the Lab10Files folder you will find a proposal for Super Breakout
(SuperBreakout.doc). If selected your game will be added as an official choice that you and other EE319K students would be
free to choose. Good projects require students to integrate fundamental EE319K educational objects such as I/O interfacing,
data structures, interrupts, sound, and the effective implementation of real-time activities. This project must satisfy the
System Requirements for all games listed above.
Any TA is authorized to approve an alternate game. If the alternate game is approved the one-page description should be
printed and posted in the lab, and thus available for other students to implement.
Valvano has eight 2-axis joysticks. These joysticks can be checked out on a first come first served basis, and used for the
design your own game option. The interface is simply two slide pots, which you will need to connect to two ADC channels.
You will need a different sequencer from SS3 so that your software can trigger the ADC and collect conversions from two
channels. To checkout a joystick, first you must get your design your own game approved by your professor, and then come
to Valvano. These joysticks must be returned.
Procedure
Part a) In a system such as this each module must be individually tested. Your system will have four or more modules. Each
module has a separate header and code file. Possible examples for modules include slide pot input, button input, LCD, and
sound output. For each module design the I/O driver (header and code files) and a separate main program to test that
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 68
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
particular module. Develop and test sound outputs as needed for your game. There should be at least one external switch and
at least one slide pot.
Part b) In the game industry an entity that moves around the screen is called a sprite. You will find lots of sprites in the
Lab10Files directory of the starter project. You can create additional sprites as needed using a drawing program like Paint.
Most students will be able to complete Lab 10 using only the existing sprites in the starter package. Because of the way
pixels are packed onto the screen, we will limit the placing of sprites to even addresses along the x-axis. Sprites can be placed
at any position along the y-axis. Having a 2-pixel black border on the left and right of the image will simplify moving the
sprite 2 pixels to the left and right without needing to erase it. Similarly having a 1-pixel black border on the top and bottom
of the image will simplify moving the sprite 1 pixel up or down without needing to erase it. You can create your own sprites
using Paint by saving the images as 16-color BMP images. 24-bit BMP images can also be converted, but the files are 8 times
larger and will use up your allotment of 32k on the free version of the compiler. Figure 10.3 is an example BMP image.
Because of the black border, this image can be moved left/right 2 pixels, or up/down 1 pixel. Use the BmpConvert.exe
program to convert the BMP image into a two-dimensional array that can be displayed on the LCD using the function
LCD_DrawBMP(). For this part of the procedure you will need to write main programs for drawing and animating your
sprites.
Figure 10.3. Example BMP file. Each is 16-color, 16 pixels wide by 10 pixels high.
Program 10.1 shows an example BMP file in C program format. There are 0x76 bytes of header data. At locations 0x12-0x15
is the width in little endian format. In this case (shown in blue) the width for this sprite is 16 pixels. At locations 0x16-0x19
is the height also in little endian format.
const unsigned char Enemy10Point1[] = {
0x42,0x4D,0xC6,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x76,0x00,0x00,0x00,0x28,0x00,
0x00,0x00,
0x10,0x00,0x00,0x00, // width is 16 pixels
0x0A,0x00,0x00,0x00, // height is 16 pixels
0x01,0x00,0x04,0x00,0x00,0x00,
0x00,0x00,0x50,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x80,0x00,0x00,0x80,
0x00,0x00,0x00,0x80,0x80,0x00,0x80,0x00,0x00,0x00,0x80,0x00,0x80,0x00,0x80,0x80,
0x00,0x00,0x80,0x80,0x80,0x00,0xC0,0xC0,0xC0,0x00,0x00,0x00,0xFF,0x00,0x00,0xFF,
0x00,0x00,0x00,0xFF,0xFF,0x00,0xFF,0x00,0x00,0x00,0xFF,0x00,0xFF,0x00,0xFF,0xFF,
0x00,0x00,0xFF,0xFF,0xFF,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, // bottom row
0x00,0x0F,0x00,0x00,0x00,0x00,0xF0,0x00,
0x00,0x00,0xF0,0x00,0x00,0x0F,0x00,0x00,
0x00,0x0F,0xFF,0xFF,0xFF,0xFF,0xF0,0x00,
0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x00,
0x00,0xFF,0xF0,0xFF,0xFF,0x0F,0xFF,0x00,
0x00,0xF0,0xFF,0xFF,0xFF,0xFF,0x0F,0x00,
0x00,0xF0,0x0F,0x00,0x00,0xF0,0x0F,0x00,
0x00,0x00,0xF0,0x00,0x00,0x0F,0x00,0x00,
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, // top row
0xFF};
Program 10.1. Example BMP file written as a C constant allocated in ROM.
08/09/15
Jon Valvano
EE319K Fall 2015 Lab Manual
Page 69
In this case (shown in red) the height for this sprite is 10 pixels. The LCD LCD_DrawBMP() function restricts the width to
an even number. This function also assumes the entire image will fit onto the screen, and not stick off the side, the top, or the
bottom. There is other information in the header, but it is ignored. As mentioned earlier, you must save the BMP as a 16color image. These sixteen colors will map to the grey scale of the LCD. The 4-bit color 0xF is maximum brightness or
white, and the color 0x0 is off or black. Starting in position 0x76, the image data is stored as 4-bit color pixels, with two
pixels packed into each byte. Program 10.1 shows the purple data with 16 pixels per line in row-major. If the width of your
image is not a multiple of 16 pixels, the BMP format will pad extra bytes into each row so the number of bytes per row is
always divisible by 8. In this case, no padding is needed. The LCD_DrawBMP () function will automatically ignore the
padding.
Figure 10.4 shows the data portion of the BMP file as one digit hex with 0’s replaced with dots. The 2-D image is stored in
row-major format. Notice in Program 10.1 and Figure 10.4 the image is stored up-side down. When plotting it on the screen
the LCD LCD_DrawBMP() function will reverse it so it is seen right-side up.
................
...F........F...
....F......F....
...FFFFFFFFFF...
..FFFFFFFFFFFF..
..FFF.FFFF.FFF..
..F.FFFFFFFF.F..
..F..F....F..F..
....F......F....
................
Figure 10.4. The raw data from BMP file to illustrate how the image is stored (0s replaced with dots).
Part c) The starter project includes a random number generate. A C version of this function can be found in the book as
Programs 2.6, 2,7, and 3.12. To learn more about this simple method for creating random numbers, do a web search for
linear congruential multiplier. The book example will seed the number with a constant; this means you get exactly the
same random numbers each time you run the program. To make your game more random, you could seed the random number
sequence using the SysTick counter that exists at the time the user first pushes a button (copy the value from
NVIC_ST_CURRENT_R into the private variable M). You will need to extend this random number module to provide
random numbers as needed for your game. For example, if you wish to generate a random number between 1 and 5, you
could define this function
unsigned long Random5(void){
return ((Random32()>>24)%5)+1; // returns 1, 2, 3, 4, or 5
}
Seeding it with 1 will create the exact same sequence each execution. If you wish different results each time, seed it once
after a button has been pressed for the first time, assuming SysTick is running
Seed(NVIC_ST_CURRENT_R);
Part d) When designing a complex system, it is important to design implement and test low-level modules first (parts a,b,c).
In order to not be overwhelmed with the complexity, you must take two working subsystems and combine them. Do not
combinate all subsystems at the same time. Rather, we should add them one at a time. Start with a simple approach and add
features as you go along.
Commit to SVN the final version of the software that implements your game. All source files that you have changed or added
should be committed to SVN. Please do not commit other file types.
Step 1) Certification: do this before lab the last week of classes
You will show the TA your program operation on your TM4C123 board. The purpose of certification is to certify
the game meets specifications, and to guarantee the zip file is uploaded. Certification will occur during the lab periods and
office hours. There will not be time to certify during the class competition. If your system runs and meets the specifications,
it will be invited to the design competition. The TA will place your game into one or two categories. The wonderful category
will have a maximum of 100 points, and the supreme category has a possible grade of 110 points. We expect most projects
to be invited. The details of the competition will be explained in class. The dute date is 2 hours before the start of the last
valvano or ramesh@mail.utexas.edu 8/9/2015
Page 70
Embedded Systems: Introduction to ARM® Cortex™-M Microcontrollers
lecture class of the semester. Late checkouts are handled by the TA in the usual way. All late checkouts must be completed
by Friday 3pm, and all LCDs with the bag must be completed by 5pm Friday.
Step 2) Design competition: occurs during the last class period of the semester (details announced in class)
For each project you will print one page (size 12 font) of some of your Lab 10 code representing software that you
think is well-written and creative. You will pull power from a computer to run your game. One team member will be
stationed at your game to help explain and run your system. The other team member will be given a grading sheet and will
visit/evaluate 8 other projects. The graders should bring their own headphones to minimize the chance of spreading germs.
The graders will spend 4 minutes with each project, and then rotate to the next project. The grading is a rank order and will
be secret. The scores should be based on software quality, fun and function. Fun will be had by all.
Step 3) SuperFinals competition: occurs during the last week of classes (details announced in class)
The top scoring game from each class will be invited to the finals. In addition, the TAs will choose a second game from each
class. This will make a total of 6 games for the final. All students from EE319K are invited: cookies and carrots provided.
Each person entering the room will be given ONE voting card. If you are invited to compete, bring your game and a 30
second story of something special about your game. At noon we will write the names of the 6 groups on the white board
making 6 columns
Between at 12:05-12:45 we will demonstrate the games
A) Set up, 1 minute
Plug the audio jack into your device (we will have amplified speakers)
Position the LCD under the document camera
Plug the USB cable into power (we will have USB power)
B) Game play 3 minutes, it is best to talk narrate while you play, helpers allowed
C) Summarize major features so audience will remember you, 30 seconds
D) Disconnect, 30 seconds (at this point you should ceremoniously vote for yourself)
If you have not been invited to compete, and you are proud of your game, bring it and a power source. Set it up and invite
others to play. We expect and encourage talking and walking around.
After the demonstrations have completed, students who wish to vote will cast their vote by placing their ONE card
under/by the name of their favorite project
Hint
1) Say you want to convert a wav file (blah.wav) you want to use in your game. Here are the sequence of steps you can do in
matlab to make this happen:
1. Read the file and extract the sample as well as the frequency:
[Spls, fs] = wavread('blah.wav');
2. Downsample it, to get it to the frequency you want (11.025 kHz):
Spls = downsample(Spls, round(fs/11025));
3. Adjust the samples (originals are fractions between -1 and 1) to range between 0 and 15 (4-bit)
Spls = round((Spls+ 1)* 7.5);
4. write it to a file so you can cut and paste it into your C code:
file = fopen('blah.txt', 'w');
fprintf(file, 'unsigned char blah[] = {');
fprintf(file, '%d,', Spls);
fprintf(file, '};\n');
That’s it, you should have a file (called blah.txt) with a declaration you can cut and paste in your code. There may be an
extra commas in the output that you would have to remove. You could make a convert.m file with these statements inside a
subroutine and have a converter that can be invoked for conversion from matlab. (From Mitch Crooks)
08/09/15
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising