CSCE 436/836: Embedded Systems
Lab 2: Basic Communication, Debugging, Thruster Control, and Sensor Reading
Instructor: Carrick Detweiler
carrick _at_ cse.unl.edu
University of Nebraska-Lincoln
Spring 2011
Started: Feb 8, 2011
Lab 2 Due: Feb 22, 2011
1 Overview
In this lab we will implement basic serial communication, thruster
control, and analog sensor reading. We will also learn about
techniques for debugging embedded systems. The serial communication
will enable simple keyboard-based control of the hovercraft. We will
control the thrusters by implementing PWM control of the power lines
of the motors. The output of the thrusters will be controlled by
sending serial port commands. Finally, we will read the analog
sensors on the hoverboard. This includes the gyro on the 3.3V Atmel
and the current and voltage feedback on the 5V Atmel.
Note that at the start of the lab we have only covered the serial
communication topics in the lecture portion of class. Over the next
two weeks we will cover most of the concepts needed to complete the
rest of this lab. However, the lab instructions are fairly detailed,
so you can work ahead of what we are covering in class or you can wait
if you would prefer. Regardless, use the readings listed on the
course website and the Atmel datasheet as a reference in lab.
Finally, as always, please read through the whole lab before starting.
You may want to complete some sections of the lab before others. In
particular, we will start using the 5V Atmel in this lab, but this is
not described until towards the end of the lab.
2 Materials
In this lab you will need the following supplies in addition to those
you already have:
- Five 2-56 by 3/4 inch socket head screws
- Five 2-56 plastic lock nuts
- Velcro
3 Safety
The safety instructions included here are a reminder from the previous
labs. Please read through them again to remind yourself of safe
operating practices when working with the Hoverboard and Hovercraft.
Please let the instructor know if you have any questions or concerns.
In this lab we will start controlling the thrusters from the
Hoverboard. Make sure that all fingers, hair, clothing, and other
items stay well clear of the thrusters at all times. This is
especially true when you first start working with them as they may
turn on unexpectedly as you develop your control code. Also remember
that you should turn off the thrusters when they are not in use as
they may be damaged if operated for an extended period of time.
Please be extremely careful with the hoverboard. Make sure that you
keep any metallic objects far away from the hoverboard. This includes
rings, necklaces, coins, pens, etc. Also, never place the hoverboard
on anything conductive as there are exposed elements on the back of
the hoverboard. You should also never have liquids near the
hoverboard. Be careful when connecting cables and wires to the
hoverboard to make sure that you connect it to the right port and in
the right orientation.
We will be using high-power, two cell Lithium-Polymer (LiPo) batteries
to power the hoverboard. Please be extremely careful with these
batteries and recall the safety information we discussed in class and
lab. In particular, only use the designated chargers in the lab to
charge the batteries. The hoverboards will not turn on if the voltage
on the battery is too low, however, you should also be aware of the
use of your battery and never leave it connected to the hoverboard
when not in use. A fully charged battery will have a voltage of 8.4V.
A battery that is about half charged will have a voltage of
approximately 7.4V, and a nearly discharged battery will have a
voltage of around 7.0V. A voltage below 6.0V can be dangerous,
especially if you try to recharge it. If this happens, please let the
instructor know immediately. With care, it is possible to revive an
over discharged battery if it is done quickly (but putting it on the
charger is not the way to do it and it is dangerous).
Unlike programming a normal computer, it is possible that a bug in
your code could physically damage the hoverboard or other devices.
Take care, especially when programming I/O pins as setting the wrong
state on the wrong pin could damage the processor or peripherals.
4 Getting Started [5 pts.]
To start this lab, we will mount the hoverboard on the hovercraft.
Also, you will need to download the code templates from the course
website.
4.1 Mounting Hoverboard on Hovercraft
You have been provided with 3/4 inch 2-56 socket-cap screws and lock
nuts. Screw these into the mounting holes, you may need to move, cut,
or adjust the electrical tape you used to cover the base of the
hoverboard (make sure you it is still covered). The screws will stick
out significantly. Then, choose a location for your hoverboard. It
is possible to adjust the position later, but you should try to
minimize the number of times you move it as it will result in
additional holes in your hovercraft base.
When choosing a location to mount your hoverboard, consider the
location of the battery, how thruster power wires will be routed to
connect to the hoverboard, and the weight distribution needed to
balance the hoverboard (with the battery weighing significantly more
than the hover board). Once you have decided on a position for the
hoverboard, simply press the screws down into the foam base. You can
remove the hoverboard at any time by just pulling it up. You should
be careful when removing and inserting the hoverboard so that you
don't enlarge the holes too much. While the fit may be loose, the
length of the screws should be sufficient to prevent unwanted motion
of the board.
You have also been giving some Velcro. You can use this to affix the
battery to the board. If your battery does not have Velcro on it
already, please make sure to attach the fuzzy side of the Velcro to
the battery.
Question: Where did you mount your hoverboard and battery? Why?
4.2 Base Code
Download the code templates from the course website. As always, you
should look over all of the code so that you have a sense of what
needs to be done. Look for STUDENT CODE comments, which
indicate where you need to fill in code.
I recommend that you integrate some of your code from lab1 in with the
code from lab2. In particular, it will be useful to be able to use
the LEDs and button, so copy that code to your lab2 directory and
modify the corresponding SConstruct file to include any additional
source files you have added. With this base configuration, you should
make sure that you can compile your code without errors or warnings.
You may also find the Morse code implementation useful, although we
will quickly move to using serial communication.
The code provided for the lab is for the 3.3V processor, which we will
start off programming. We will also program the 5V processor in this
lab. While the code will be very similar for the 5V Atmel, it will be
slightly different. The programming and build scripts, however, are
the same, so you can directly copy these for the 5V processor. For
this lab, you are only being provided with a very small amount of
code. The design of the majority of the code will be left up to you
for most of the lab.
5 Serial Communication [15 pts.]
In this part of the lab we will setup the two UART ports on the 3.3V
Atmel processors. We will communicate over the serial port by using
the USB to serial converter that you use to program the boards. This
USB adapter allows communication with a 3.3V UART. Later in this
section, you will learn how to communicate over the serial port from
the computer. You will want to read this section before you start
actually programming too much as you will need to connect from the
computer to the hoverboard to see if your UART code is working.
Look at the hoverboard datasheet and identify the pins that are used
for the serial ports. Make a note of this and also identify the two
serial port connectors on the actual hoverboard.
In the sample code there are a couple of functions that you need to
complete in uart.c. We will discuss each of these functions
briefly, look at the code template comments for additional details.
Section 5.2 contains activities to do once
the serial port is working, but it also contains some techniques that
you can use to debug your system if you are having problems.
void uartInit(uint8_t port, uint32_t baud)
This is the main initialization function. Depending on the value of
port, it will configure either UART0 or UART1. The USB to
serial converter we are using typically operates with 8-bits,
no-parity, and 1 stop bit. This function should setup these
parameters and enable the serial port operation of the ports.
Carefully read the section on the UARTs in the Atmel datasheet to help
set the correct parameters.
This function also sets the baud rate of the port. This should
probably just be a call to uartSetBaud(...), it is included as
a parameter here so that the default baud rate for the UART is clear
upon initialization.
void uartSetBaud(uint8_t port, uint32_t baud)
This function sets the baud rate of the specified port. Read
the datasheet to determine how to set the passed baud rate. Recall
that there is a normal and double speed mode. Depending on the
desired baud rate, one of these may have a lower error rate. It is
not required, but optionally, this function can choose which mode to
use. If you do not do this, you should set it such that a baud rate
of 19200 has a low error as this is the baud rate we will typically
use. Also remember that the 3.3V processor operates at 8MHz.
void uartSendByte(uint8_t port, char c)
Implement this function that sends the specified byte, c, over
the specified UART port.
uint8_t uartReceiveByte(uint8_t port, char *c)
Implement this function that checks to see if there is a byte to
receive over the specified port. If a byte has been received,
it should be stored in the location pointed to by c and return
1 to indicate a byte has been received. If there is no available
byte, it should return immediately and return 0.
This function is said to be non-blocking as it does not wait if there
is no data available. This is typically better as it allows other
code to execute if there is nothing available. Blocking code can
introduce bugs into embedded systems as it will get stuck in an
infinite loop if a hardware or other error prevents communication.
You should never use blocking code, unless you have a specified
timeout that will escape from the blocked code once a certain amount
of time has passed.
5.1 Communicating from the Netbooks
You can communicate with the Hoverboard using any serial terminal
program such as minicom, seyon, or something else (even hyperterminal
in windows). Here I will tell you how to use GNU screen. This
is one of my favorite programs. It allows you to have multiple
terminal windows open and is also able to communicate over serial
ports. First make sure you have it:
sudo apt-get install screen
Next, you can connect to a serial port by giving the command:
screen /dev/ttyUSB0 19200
which will connect to serial port /dev/ttyUSB0 at a baud rate
of 19200. Adjust the baud rate if you set your Hoverboard to
communicate at a different speed.
To exit screen you need to send the command ctrl-a k
(that is hold down the control key and press a, let go of the
control key and press k). If you forget to do this (and
instead "x" or kill the window), screen will remain running
and attached to the serial port in the background. You can reconnect
to it by typing screen -r.
Note that in order to program the board, you will have to
ctrl-a k the screen so that the program script
can use the serial port.
5.2 UART Communication
The first thing to do once you have written the UART functions is to
verify that it works. You can test the sending and receiving from the
serial ports separately. Modify the main loop code to start out by
initializing both serial ports to a baud rate of 19200. Next, test
the sending on both serial ports by sending a short (and different)
string (multiple characters) over both serial ports that repeats
continuously. You can send a string over UART0, for instance, by
executing the code:
uartSendByte(0,'t');
uartSendByte(0,'e');
uartSendByte(0,'s');
uartSendByte(0,'t');
You should make sure to add a pause after sending each string. This
is important because if you send continuously over a serial port it is
possible (especially if there is some error in the clk speed) that the
receiver will have trouble synchronizing with the start and stop bits.
Question: Did your serial port sending code work? If not, what
problems did you have?
If you receive garbage over the serial port and not the string you
expected, you probably have a problem somewhere setting the baud rate.
If you don't receive anything at all, try testing the receiver first
(below), which will tell you if you at least have the baud rate
correct. If the receiver works, it is likely you forgot to set the TX
pin to output.
To test the receiver, write some code that lets you turn on and off
the LEDs depending on which character you send. Define different keys
that turn on or off each of the LEDs. Try sending these characters
and see if it works.
Question: What keys did you use to turn on and off the LEDs? Did this
work the first time, if not what was the problem?
Now combine the sender and receiver. Keep the LED toggling code, but
add to it so that it echos back the same character that you sent if it
is one of the LED toggle commands and sends back a '?' if that key
does not correspond to a known keyboard command.
Question: Experiment with standard baud rates (those shown in the
table at the end of the UART section in the Atmel datasheet). Do all
of them work for communication with the computer? If some do not
work, why is that?
It is useful to be able to send strings without having to type each
individual character. Add a function that lets you send a string
over the UARTs. Refer to the Morse Code file from Lab1 if you are
unsure how this should be implemented.
Printing predefined strings is useful, but it is much better if you
can print formatted strings in printf-style. It is possible to
configure the system such that printf will work to print over
the serial port. However, we will start by learning how to print
formatted strings. It is relatively easy if we use the function
snprintf(...). We are using snprintf instead of
sprintf. This takes the additional argument "n", which
indicates the size of the buffer being written into. Never ever,
ever use sprintf as this one very common method of ensuring
that your code will have buffer overflow bugs, as sprintf
assumes that the buffer you pass it is of infinite size.
If you are unfamiliar with snprintf you can google it or look
at the man page (from the terminal, type man snprintf). Using
this function requires including stdio.h in the .c file. The
idea is that you can create a temporary buffer to write the string
into and then print it using the string sending code you just wrote
(perhaps using strlen(...) to find the length of the string).
Try printing out over the serial port some of the variables that are
changing in your code. You may want to create a convenience function
that enable easy printing of integers as decimal number or hex
numbers.
Question: To go to a new line in the output, you can send a \r\n.
\r is a carriage return and \n is a newline. Describe what
happens if you only send one or the other.
If you look at the program script that we use, you will see
that at the start the first thing it does is send the string
"reboot" over the serial port. Add a basic parser to your code
that will reboot the board if it receives that specific string. You
will notice in there is a reboot() command in main.c
that you can use to reboot the board. This means that you won't have
to press the reboot button, you can just start the program script
(unless your code is stuck somewhere and doesn't parse the reboot
command).
Question: There are a number of ways you can implement parsing the
serial input and looking for the reboot command. Describe how you
implemented the parser for the reboot command and discuss why you
choose to do it this way. In particular, address how you can also
receive and process other serial commands while looking for the reboot
string. Does your implementation allow you to use the characters in
the string reboot for other commands?
6 Debugging [15 pts.]
Debugging embedded systems is much more challenging than debugging
programs written on a PC. In this section we will learn some
techniques for debugging. We will apply these techniques to figure
out how some code that someone else wrote works, even though we do not
have the source code.
6.1 Outputting Information
The most basic method for debugging is to use an LED to show where the
program is in its execution. Simply turning on a green LED when the
code is executing is a helpful tool. LEDs can also be used to obtain
relatively precise information on the timing of events. By turning on
an LED only when the function you are interested in is executing, you
can see how long that function takes and obtain some information on
when it is executing. If it is a very short event you may need to use
an oscilloscope to see when it is actually occurring. Alternatively,
you can only toggle the LED on every Nth execution of the function you
are interested in by keeping track of a global counter. One thing to
keep in mind, however, is that turning on and off LEDs does incur a
slight overhead as it takes some clock cycles to turn them on and off.
LEDs are useful, but we only have two on each processor so it can take
a long time to debug complex sections of code. We can also use the
serial port to aid in debugging by printing out information. However,
printing strings over the serial port can take a significant amount of
time. So if timing is important, printing debug information over the
serial port can change the program timing.
One useful technique is to output the values of configuration
registers over the serial port. This allows you to verify that they
are set as you expected (and that some other part of the code did not
overwrite registers you thought you set properly). Try printing out
the values of the DDR registers and verify that they are set to the
expected values.
Question: How long does it take to send the string
Hello, World!\r\n over the serial port if the baud rate is
9600? What about 19200 or 115200?
6.2 Help from the Compiler
The compiler gives us a number of sources of information that can help
debug programs. First and foremost are the warnings and errors the
compiler produces. You have to fix errors in order to compile the
code. However, you also should pay attention to any compiler warnings
(the SConstruct file enables warnings from gcc by
passing the -Wall flag). If there are warnings, you should fix
them, even if your code works as you expect. Often the warnings will
indicate some potentially undefined behavior that could cause problems
in the future. Also, if you always ensure that there are no compiler
warnings, if you see one, you will know there might be a problem.
scons will only recompile files that have changed since the
last compile. Sometimes it is useful to recompile everything to make
sure there are no outstanding warnings or other problems. You can
tell scons to delete any compiled files by issuing the command
scons -c (this is like doing a make clean). You can
then recompile everything and look for warnings.
The SConstruct file is setup so that when you compile, the file
main.lss is also generated. This file contains the assembly
code that is generated by compiling the code. You can look at this
file to see the assembly that is generated from the C code you
write. At the top of this file is also the memory map. There are a
number of entries in this memory map. They are:
- .text goes into the processor's flash. This is the actual code that is executed.
- .bss goes into the processor's RAM. This is for any globally defined, but not initialized variables.
- .data goes into the processor's RAM. This is for any globally defined variables that are initialized.
- .stab and .stabstr which are just used for debugging information, they are not loaded into the processor when you program it
Note that some of these sections may not be present if they are not
used at all (in particular the .bss and .data sections).
By summing .bss and .data you can determine how much of
the processor's RAM has been used, excluding memory used by the stack
or heap (which you should not be using).
Question: In Section 5.2 you included
stdio.h in order to use snprintf. This triggered the
compiler to include some code from the stdio library. How much flash
and memory does including this require?
When you look at main.lss you can see that shows the C
code you wrote with the assembly. Sometimes the C code and
assembly do not quite line up (or the assembly does not make sense
when compared to the C code). This is because compiler
optimization has been turned on. This means the compilers will
rearrange and modify code in order to speed operation. If you look at
the SConstruct file you will see that -Os flag is passed
to the compiler. This enables size optimization. What this means is
that the compiler will optimize the code to try and make it smaller,
while also trying to make it execute faster. Making it smaller is
useful to ensure that we have sufficient space in the flash for the
code. Alternatively you can disable optimization by passing
-O0 to the compiler (that is a dash, capital "oh" and a
zero). There are also levels 1, 2, and 3 for optimization. Sometimes
it is useful to disable optimization by using -O0. This will
make the assembly code correspond much more directly to the C
code. This may allow you to spot bugs in the assembly that would be
difficult to find with optimization enabled1. Be aware that changing the optimization level can
drastically change the speed of program execution and the program
size. In general, it is good to keep -Os enabled.
Question: For the different optimization levels, what is the resultant
code size in flash for your code base?
6.3 Applying Debugging Techniques
In the 3.3V Atmel code templates there is a file keycode.h and
keycode.o. The file keycode.o is the precompiled object
file, you do not have access to the original keycode.c file
that generated the object file. The author of the original code was
testing some new code he wrote for the embedded system that controlled
his car's steering. Unfortunately, there was a bug in the code so he
drove off a cliff. The keycode file was used by an old project (an
encryption keypad) to generate some keys to encrypt a client's data.
The keypad broke and now we need to make a new one to
allow them to recover their keys. The .h and .o files were recovered
for the project, but not the .c files and unfortunately the author did
not comment the code well (at all) and functions are not named well.
Start by determining the amount of code and memory that the
keycode code uses. You can do this by modifying the
SConstruct file to include the object file or not and then
examine main.lss. Note that when you look at keycode
code in main.lss there are not any comments to help determine
what is going on. This is because the author either did not comment
the code at all or did not include debug information when compiling
(hopefully it is the later, but judging by the .h file it might not
be).
Question: How much code space and memory does keycode use?
Now look at keycode.h where the functions are defined. In
keycode.h you will notices that there are 4 functions
startOne(), startTwo(), startThree(), and
startFour(). These are presumably functions that need to be
run at the start to initialize the system. Unfortunately, it turns
out that starting them in this sequence does not work.
Some initial debugging has found that if you call one of these
functions out of order, the function will not return and the board
will reboot. If a function is called in the correct order, it will
return, but the board will still reboot shortly after. We believe
that the full start sequence must be called in order to prevent
rebooting.
Question: Determine the correct sequence to call these functions. How
did you go about doing this? Hint: using LEDs or the serial port may
be useful.
There are two additional functions delaySmall() and
delayBig(). We think that these may be used by some of the
other functions to perform delays.
Question: Approximately how many milliseconds do delaySmall()
and delayBig() delay for? How did you go about determining
this? You should only need your board and a stop watch to figure this
out.
Finally, there are two functions encodeChar(char c) and
encodeInt32(uint32_t i). Once the startup sequences have been
called these functions will operate. For the client we need to
recover the encoding of the character 'c' and 'd' and the 32 bit
integer 369288000. For some reason, these functions do not return
values, instead, it looks like the encodeChar(...) function
stores its result in the memory location
KEYCODE_CHAR_MEMORY_LOCATION and encodeInt32(...) stores
its result in KEYCODE_INT32_MEMORY_LOCATION. You can read the
memory location by casting the integer memory location into a memory
pointer of the correct size and then dereferencing that pointer.
Question: What are the encoded values for the characters 'c' and 'd'
and the 32 bit integer 369288000? Explain how you determined this.
Question: How long does do the two encode functions delay for? Answer
this in terms of milliseconds and in terms of the delay functions.
6.4 Additional Methods
There are numerous other methods to debug embedded systems. You will
discover many yourself as you program microcontrollers more. You
should also be aware of JTAG interfaces that are external devices that
connect to a microcontroller and allow line debugging, insertion of
break points, and the reading of registers. Not all processors
support JTAGs and they are often processor-specific and can be
expensive. They are, however, very useful for debugging systems. It
is difficult, however, to use them to debug real-time events.
Another technique is to use periodic interrupts (we will learn about
these later) to examine the state of different variables or registers
(you can even use it to sample the current size of the stack). This
can be useful, but again can interfere with real-time events.
7 PWM Motor Control [15 pts.]
In this section we will configure the pulse width modulator (PWM) pins
on the 3.3V Atmel to control and vary the power to the thrusters.
Start by looking at the hoverboard schematic. Notice that the
PWM pins are connected to a number of N-channel mosfets. These are
like switches connected to the ground lines of the motors. If we turn
them on, the motors will turn on at full speed (do not do this). We
cannot vary the voltage or current that goes to the motors (that would
be a complicated circuit) as we did with the power supply in lab to
control the motor speed. But we can use PWM to switch the mosfets
fast. Because motors act as inductors, if we switch fast enough,
the voltage the motor sees will be proportional to the duty cycle of
the switching. This means we can vary the duty cycle of the switching
to change the speed of the motor.
Question: Notice the resistors connected to the mosfets. What do you
think they do? Is the mosfet on when the control pin is high or low?
Create functions (perhaps in a new .c file) to initialize and control
the PWM pins. Note, you should limit the maximum speed the thrusters
can operate at to about 75% of maximum. Impose this limit in the
software you write. Higher power operation may reduce the lifetime of
the motors. Also remember that you should not leave the motors on
longer than you need to (and at most around 10 minutes).
You should configure the PWM duty cycle to be around 500Hz, remember
that the processor operates at 8MHz. This duty cycle is fast enough
to provide smooth operation for the motors. It is also slow enough so
that the mosfets are not switching too fast. Mosfets consume more
energy (heat up) the faster they switch, so it is better to drive them
at a lower switching frequency. If you set the frequency too high,
you may burn the mosfets. You will need to read the datasheet
carefully to figure out how to control each of the PWM pins (note that
each of the timers is slightly different).
Question: Describe some of the differences between the different
timer/PWM ports on the Atmel.
Now add to your code so that you can control each of the thrusters by
sending different key strokes. You may want to have a "faster" and
"slower" key for each thruster that increments/decrements the
thruster speed by 5 or 10% with each key press. Alternatively, you
could make it such that it only thrusts when a key is held down. This
will give you primitive control over the hovercraft.
Question: What PWM percentage do you need to use to have lift with the
lift thruster?
Question: Describe the key mapping you used and your experience in
operating the hovercraft with these key mappings.
8 Analog to Digital Converter [15 pts.]
We will now implement code to read analog signals on the analog to
digital (A2D or ADC) pins on the 3.3V Atmel processor. As always, you
should start by reading the section on the ADC in the Atmel datasheet
and reviewing the hoverboard schematic. Most of the ADC pins on the
3.3V Atmel are free pins that can be used to connect to external
analog sensors. The last two pins, however, are connected to the
analog outputs of the gyroscope. We will read these in this section.
The ADC has 10 bit resolution. Configure the ADC to use AVCC as the
voltage reference, with single-ended voltage input (as opposed to
using it in differential input mode), and in single conversion mode.
You will also need to make sure that the ADC clock rate is correct.
Do not worry about using the ADC noise canceler at this point. Note
that since the ADC gives 10 bit resolution and the registers are only
8 bits, you have to read two different registers (in a particular
order) and then shift the result to fit into a 16 bit integer.
Question: What is the actual ADC clock rate that you used?
Question: Given that the ADC has 10 bit resolution, what is the
resolution of the ADC in mV?
8.1 Reading the Gyroscope
The gyro gives a relatively stable mid-voltage reading when no
rotations are occurring. When you rotate in one direction the voltage
drops and it increases if you rotate in the other direction. In
addition to the ZOUT-1X output, there is also a 4X output. This is
the same signal as the 1X output, except the signal change due to rotation
is amplified 4 times. Try reading both of these channels (and
printing the value over the serial port) while keeping the board still
and also while rotating each direction.
Question: Find the datasheet for the gyro. What is maximum rotational
rate that can be measured? What is the mV per degree per second for
the two outputs? What about converted into bits per degree per second
as read by the ADC?
Note that you may need to limit the rotational thrust on your hovercraft
if the rotation rate exceeds that which can be measured by the gyro.
When the gyro is not rotating, it produces a stable voltage that is
nominally around 1.23V (as per the datasheet). This number will vary
slightly for each gyro. Determining this number accurately is
critical because we are interested in integrating the output from the
gyro (so that we get an absolute angle). If there is a small error in
the initial offset, the integrated error will grow quickly over time.
It is impossible to remove all offset error, but we will try to remove
most of it. Create a function that averages a number of readings from
the gyro channels to find the offset 2.
Question: In addition to reporting the offset determined by your
averaging function, plot the readings of the gyro channels over time
when the gyro is not moving. You can record data from your serial
port to a file by doing cat /dev/ttyUSB0 > datafile.txt. Is
there noise in the signal? Is it different on the two channels?
Note, you can also increase the serial port baud rate to increase the
number of samples you can capture.
Question: Now record a similar sequence of data with one of the
thrusters active (but not rotating the board). Is there more noise?
Question: Now record a sequence of data when you are using the
thrusters to rotate your hovercraft in both directions. You can
script a thrust sequence (starting, stopping, changing direction,
stopping, etc) that will start with the press of the button to make
collecting data easier. Analyze and describe this data.
Question: Create a function that averages a few successive readings
(enough to remove most of the noise you notices in the signal when not
moving). Print this on the serial port every few seconds and record
data for a number of minutes without moving the hovercraft. Is the
offset stable over time?
9 Setting up the 5V Atmel Processor [15 pts.]
Now that you are experts in configuring Atmel processors, you should
setup the 5V Atmel processor on the hoverboard. The sample code does
not contain any files for the 5V processor. You should create a new
directory for the 5V processor code and copy the build (you may need
to modify the list of sources) and programming scripts. Most of the
code should be very similar. Ideally, you could create a set of
common code that will be shared between the two processors and only
customize a few particular aspects that are different. In this
section there are brief instructions reminding you of the differences
between the processor configurations.
9.1 LEDs and Buttons
You should configure the LEDs and buttons on the 5V Atmel processor.
Note that they are on different pins than on the 3.3V processor.
9.2 Serial Ports
The 5V Atmel processor only has a connector for one UART (UART0). The
other is dedicated to a radio connector slot. We will use this later
in the semester to have radio control of the hoverboard so that you
can operate it remotely. So you only need to configure UART0,
although you are also free to configure UART1. The main difference
between the 5V and 3.3V Atmel is that the 5V processor runs at 20MHz.
Notice in the schematic that the TX pins on the serial ports have
diodes and pull-up resistors. This allows the connection of 3.3V UART
devices to the 5V processor by only allowing the output of 3.3V on the
TX line.
9.3 Analog to Digital Converter
The 5V Atmel ADC has current monitors and battery voltage monitors as
well as a number of free ADC pins. The configuration of the ADC will
be very similar to that of the 3.3V processor. The main difference is
the clock speed and the fact that AVCC is 5V instead of 3.3V.
9.4 Monitoring Current and Voltage
Look at the first page of the hoverboard schematic. In the lower left
there are two INA196 current monitor chips. These are amplifiers that
multiply the voltage drop over the small value resistors (R32 and R33)
that all of the current must pass through. U6 monitors the current
used by the thrusters and U7 monitors the current used by the rest of
the board.
Question: What is the maximum current that can be measured by both
current monitors? Hint: first find the gain of the INA196 and then
note that the maximum voltage the current monitors can output is 5V.
Question: Given that the Atmel has 10 bit ADC resolution, what is the
minimum change in current that it can detect?
Create functions that output the current measured by both current
monitors.
Question: Can you detect the change in current when you turn on or off
an LED? What if you hold the reset button on the other processor
down?
Overall, it is probably best not to use more than 10 or 15A
continuously for all the thrusters combined. Use the LEDs on the 5V
processor to provide visual feedback as to how much current is being
used. Perhaps the green light could be on if it is less than 10A and
the red light could turn on if you use more than 15A.
Question: Using the current monitor on the thrusters determine and
report how much current the thrusters use for different PWM levels.
The 5V Atmel also has a voltage monitor circuit using R31 and R34
which form a voltage divider. Write code to convert the raw ADC
reading into a voltage. This will give you the ability to monitor the
battery voltage and, perhaps, shut down thrusters if the battery
voltage drops too low.
Question: What is the maximum battery voltage the 5V Atmel could read?
Question: (Optional) Record and plot the battery voltage readings with
the thrusters off and with one or more on at different levels.
10 To Hand In
You should designate one person from your group as the point person
for this lab (each person needs to do this at least once over the
semester). This person is responsible for organizing and handing in
the report, but everyone must contribute to writing the text. You
should list all group members and indicate who was the point person on
this lab. Your lab should be submitted by email before the start of
class on the due date. A pdf formatted document is preferred.
Your lab report should have an introduction and conclusion and address
the various questions (highlighted as Question: ) throughout the lab in
detail. It should be well written and have a logical flow. Including
pictures, charts, and graphs may be useful in explaining the results.
There is no set page limit, but you should make sure to answer
questions in detail and explain how you arrived at your decisions.
You are also welcome to add additional insights and material to the
lab beyond answering the required questions. The clarity,
organization, grammar, and completeness of the report is worth 20
points of your lab report grade.
Question: Please include your code with the lab report. Note that you
will receive deductions if your code is not reasonably well commented.
You should comment the code as you write it, do not leave writing
comments until the end. I recommend structuring your code such that
for every section of this lab, a different function is called from the
main loop. Then you will be able to comment out just one function to
change between sections in the lab and you won't have to delete
working code from previous parts of the lab.
Question: For everyone in your group how many hours did each person spend on this part and the lab in total? Did you divide the work, if so how? Work on everything together?
Question: Please discuss and highlight any areas of this lab that you
found unclear or difficult.
Footnotes:
1In addition,
sometimes optimizations can cause unexpected behaviors. The most
common is that the compiler will remove code that it does not think
or needed. For instance, loops that do not seem to do anything.
This can cause a problem if you wanted to use that loop for a
delay.
2Note that the gyro
offset can also be impacted by operating temperature and other
factors that may change over time.
File translated from
TEX
by
TTH,
version 3.89.
On 11 Feb 2011, 09:38.