# How To Program Your Micro-Controller
This is about using `avr-dude`, `make` and getting a high-level understanding of what's happening under the hood so that you can tune your micro-controller yourself.
Notably, we'll cover the fuse programming for the ATTiny44 (echo ftdi board).
In practice, you can just read most of what is linked in the **Embedded Programming** class [there](http://academy.cba.mit.edu/classes/embedded_programming/index.html).
In fact, you should have done that already.
But for those who are overwhelmed, we will try to dissect some of the content
* [HTMAA's Embedding Programming](http://academy.cba.mit.edu/classes/embedded_programming/index.html)
* [Makefiles](https://makefiletutorial.com/) for executing pre-written commands
* [avrdude](https://www.ladyada.net/learn/avr/avrdude.html) (by Lady Ada) for programming through a programmer
* Existing tutorials for various programmers:
* [Sparkfun's avr programmer](https://learn.sparkfun.com/tutorials/pocket-avr-programmer-hookup-guide/al)
* [AdaFruit's USBTinyISP](https://learn.adafruit.com/usbtinyisp/avrdude)
* [List of AVR IC's and their packages](https://en.wikipedia.org/wiki/ATtiny_microcontroller_comparison_chart)
* [ATmega328P](http://www.digikey.com/product-detail/en/ATMEGA328P-AU/ATMEGA328P-AU-ND) (same as [Arduino Uno](https://en.wikipedia.org/wiki/Arduino_Uno)), ([datasheet](http://ww1.microchip.com/downloads/en/DeviceDoc/Atmel-7810-Automotive-Microcontrollers-ATmega328P_Datasheet.pdf))
* [ATmega16U2](http://www.digikey.com/product-detail/en/ATMEGA16U2-AU/ATMEGA16U2-AU-ND) (with 16**U** for USB support)
* [ATxmega16A4U](http://www.digikey.com/product-detail/en/ATXMEGA16A4U-AUR/ATXMEGA16A4U-AURCT-ND) (USB support)
## What are those files?
For most base projects of HTMAA, you are provided with a set of different files:
* `file.png` are typically images for tracing (trace) or cutting (outline)
* `file.c` is a C file that contains a C program, which runs on a micro-controller
* `file.make` (or anything ending in `.make`, so here typically `file.c.make` too) is to use the program `make` to call commands that allow you to do things
## What does `XXX` do?
First thing first, you should try reading the manual.
On Mac / Linux, you can access the manual of a specific command typically by typing
For example, we want to learn about `make`, so let's do that:
That should give you an interactive stream that you can go over (down/up arrows) and search through (`/` character, followed by search query).
To exit that manual, just press `q` (quit).
To get some help and figure out how to use the manual functions, press `h` (help).
Now, we're ready to start looking at those commands.
## What is [Make](https://www.gnu.org/software/make/manual/html_node/Introduction.html#Introduction)
* [Online make manual](https://www.gnu.org/software/make/manual/make.html)
* `man make`
Make allows you to automate some commands and simplify what you have to type to program things (though it can be used for programming anything, not just programs).
Typically, you'll be using the following pattern:
where `targetname` is a specific target to execute.
Targets are just names, behind which are defined sets of commands to do something.
Let's have a look at the basic FTDI echo files:
* [hello.ftdi.44.echo.c](http://academy.cba.mit.edu/classes/embedded_programming/hello.ftdi.44.echo.c) - the c program file (i.e. the code to compile and run onto the micro-controller)
* [hello.ftdi.44.echo.c.make](http://academy.cba.mit.edu/classes/embedded_programming/hello.ftdi.44.echo.c.make) - the makefile
The `.make` file contains the following (in the terminal: `cat hello.ftdi.44.echo.c.make`):
F_CPU = 20000000
CFLAGS=-mmcu=$(MMCU) -Wall -Os -DF_CPU=$(F_CPU)
avr-objcopy -O ihex $(PROJECT).out $(PROJECT).c.hex;\
avr-size --mcu=$(MMCU) --format=avr $(PROJECT).out
avr-gcc $(CFLAGS) -I./ -o $(PROJECT).out $(SOURCES)
avrdude -p t44 -c bsd -U flash:w:$(PROJECT).c.hex
avrdude -p t44 -P /dev/ttyUSB0 -c dasa -U flash:w:$(PROJECT).c.hex
avrdude -p t44 -P usb -c avrisp2 -U flash:w:$(PROJECT).c.hex
avrdude -p t44 -P usb -c avrisp2 -U lfuse:w:0x5E:m
avrdude -p t44 -P usb -c usbtiny -U flash:w:$(PROJECT).c.hex
avrdude -p t44 -P usb -c usbtiny -U lfuse:w:0x5E:m
avrdude -p t44 -P usb -c dragon_isp -U flash:w:$(PROJECT).c.hex
avrdude -p t44 -P usb -c atmelice_isp -U flash:w:$(PROJECT).c.hex
avrdude -p t44 -P usb -c atmelice_isp -U lfuse:w:0x5E:m
The **targets** we mentions are all the blocks starting like `targetname:`.
When calling for example
make -f hello.ftdi.44.echo.c.make program-usbtiny
Wait a second, what's that command?
Unfortunately, you cannot just use `make targetname` all the time.
It only works when the configuration for make (`.make` file) is actually named `Makefile` and available where you run `make`.
Most examples in the class are not named with that single same name because it would create lots of conflicts and there is nothing that associates that file.
By naming the file `myprogram.c.make`, we know that the makefile is associated with the program `myprogram.c`, but then we have to specify the `makefile` that makes uses.
Here is that argument in the manual:
Back to the command, `make -f hello.ftdi.44.echo.c.make program-usbtiny` ends up running the target `program-usbtiny` in the makefile.
avrdude -p t44 -P usb -c usbtiny -U flash:w:$(PROJECT).c.hex
* `program-usbtiny` is the target name (for the argument to `make`)
* `$(PROJECT).hex` (on the right of the target, after the `:` separator) is the list of dependencies that needs to be `made` before we can run the actual script below.
* `avrdude -p t44 -P usb -c usbtiny -U flash:w:$(PROJECT).c.hex` is the actual command being run
* Beware, those lines must start with a tabulation character `\t`, which is not visible but is different from a blank space
If we look at the dependency, it's name is actually using a MACRO definition that is available at the top of the make file:
The dependency name is either another *target* or a *file*, i.e. `hello.ftdi.44.echo.hex`.
If it's a file and the file is the last version, then it doesn't need to be computed again.
The rules are based on the existence of [the file dependencies and their relative timestamps](https://stackoverflow.com/questions/35588153/where-does-make-store-its-cache).
What about the target name? There is no explicit matching target name, but there is if we compute the MACROS (i.e. `$(PROJECT).hex:`)
And this has another dependency on the `.out` file, which is the object output from compiling the c file.
Fortunately, make outputs all the commands it runs, so you can just follow what's written on the terminal.
avr-gcc -mmcu=attiny44 -Wall -Os -DF_CPU=20000000 -I./ -o hello.ftdi.44.echo.out hello.ftdi.44.echo.c
avr-objcopy -O ihex hello.ftdi.44.echo.out hello.ftdi.44.echo.c.hex;\
avr-size --mcu=attiny44 --format=avr hello.ftdi.44.echo.out
The first command compiled the C file with `avr-gcc`.
* `-mmcu=attiny44` because assumes an attiny44 as mcu
* `-Wall` enables all **w**arnings
* `-Os` optimizes for **s**ize
* `-DF_CPU=20000000` defines the MACRO F_CPU and sets it to `20000000` (20MHz)
* `-I./` adds the current directory to search for included files (`#include "file"`)
* `-o hello.ftdi.44.echo.out` specifies the output name
* `hello.ftdi.44.echo.c` is the input file
Then the second command creates a `hex` file from the program output.
See [.hex vs .out](https://electronics.stackexchange.com/questions/417648/whats-the-difference-between-a-generated-hex-file-and-a-binary-file-in-embedded) for the difference.
TL;DR hex files are easy to read and used for not only programming but also checking that the programming went well, whereas out files are binary codes to be transferred to the memory for execution.
The last command computes information about the size that the program is going to take.
This is important if you are creating your own program, because you need to make sure it will fit in memory.
`avrdude` is a tool that allows us to send programs onto avr chips through a programmer interface such as the usbtiny (e.g., [FabTinyISP](http://fab.cba.mit.edu/classes/863.16/doc/projects/ftsmin/index.html)), or an [Atmel ICE](https://www.microchip.com/DevelopmentTools/ProductDetails/ATATMEL-ICE).
Here, the manual of `avrdude` is quite large, so instead, let's start by reading the help summary, which you can typically get by running the command with the argument `-h` or `--help`:
* `-p device` specify the type of device to program (in our case, it's an ATTiny44)
* `-P port` selects the port for programming (typically `usb`)
* `-c programmer` selects the programmer (e.g. `usbtiny`)
* `-U ...` requests a memory update (this is the actual action we're doing with avrdude)
avrdude -p t44 -P usb -c usbtiny -U flash:w:hello.ftdi.44.echo.c.hex
How did we know the device label was `t44`?
Well, the manual tells you that you can just query for the available names / devices with `avrdude -p ?`, which gives you a long list.
You can do the same to find the list of available programmers (and you'll find `usbtiny`).
Finally, the real deal - the memory programming through `-U`.
There are different actions: `-e` erases memory, likely not what you want, and `-U` updates the memory.
## `avrdude -U ...`
The format of the argument is (according to the help):
avrdude -U :r|w|v:[:format]
The manual is slightly more verbose and tells us what's available for the memory types, formats and operations:
In our case, `program-usbtiny` uses the following update command:
* Updating the `flash` memory
* **W**riting to it (not **r**eading nor **v**erifying)
* Using the hex file `hello.ftdi.44.echo.c.hex` that got generated
Why the `flash` memory?
You should read this [memory part](https://en.wikibooks.org/wiki/Embedded_Systems/Atmel_AVR#Memory) of the Embedded Systems from Wikibooks.
And you will definitely need to read the *datasheet* of your target micro-controller.
As a summary, there are different types of memory available. For the simple AVR systems, there are:
* **data** memory, I/O registers and SRAM - those are dynamic, changing during execution, and vanishing upon shutdown (aka *volatile* memory)
* **flash** program memory - this is where your program goes
* **[EEPFROM](https://en.wikipedia.org/wiki/EEPROM)** aka Electrically Eraseable Programmable Read-Only Memory, which allows you to store data that survives a restart / shutdown
* **fuses** are special types of memory that cannot be modified by the program and must be programmed separately
For interested readers:
* [flash vs eeprom](https://electronics.stackexchange.com/questions/69234/what-is-the-difference-between-flash-memory-and-eeprom/69275)
### avrdude and fuses
Fuses are specific bits of the memory that specify low-level configurations.
Thoses have to be programmed separately from the main program memory.
In our previous example with the ATTiny44, the fuse programming was done with the `program-usbtiny-fuses` target:
avrdude -p t44 -P usb -c usbtiny -U lfuse:w:0x5E:m
Here, the `-U` command targets the **low** bits of the fuse memory for the ATTiny44.
The actual value is specified as a hex number: `0x5E`.
### Hexadecimal Numbers
Hex numbers are numbers in hexadecimal base (16) and are typically prefixed with `0x`.
The symbols are:
| Decimal | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 |
| ------- | - | - | - | - | - | - | - | - | - | - | -- | -- | -- | -- | -- | -- |
| Hexa | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | A | B | C | D | E | F |
| Bin | 0000 | 0001 | 0010 | 0011 | 0100 | 0101 | 0110 | 0111 | 1000 | 1001 | 1010 | 1011 | 1100 | 1101 | 1110 | 1111 |
Looking at the individual bits (in groups of 4 since `2^4 = 16`) is typically the main reason for hexadecimal.
### Fuse meanings
Fuse bits describe low-level configurations.
There are tools available to figure out what fuses to write, notably Atmel Studio allows you to set those interactively during the programming step with help for each individual fuse of each device they support.
There are also [fuse calculators](http://www.engbedded.com/fusecalc) online.
But before you jump and set fuses, you should understand what they do, since they can make your micro-controller stay silent forever by disabling programming capabilities or changing the basic clock frequency while expecting a specific clock (that may not exist, and then you may need to change the hardware to make it work again).
Long story short - these are bits you should write carefully because some options are not reversible (or hard to reverse).
The first part is reading the datasheet about your micro-controller and its section related to fuses (e.g. page 192 of the att44 [datasheet](http://fab.cba.mit.edu/classes/863.09/people/ryan/week5/ATtiny44%20Data%20Sheet.pdf#page=159)).
In our ATTiny44 case, it mentions three different fuse bytes:
* **Fuse Extended Byte** (extra for self-programming)
* **Fuse High Byte**
* **Fuse Low Byte**
Let's have a look at the tables:
In our setup, we're just modifying the low byte and choosing value `0x5E`.
Let's deconstruct that:
`0x5E` is, in binary `01011110`, and the bits start (0-indexed) from the right (least significant bit).
Or in a table:
| `CKDIV8` | 0 |
| -------- | - |
| `CKOUT` | 1 |
| `SUT1` | 0 |
| `SUT0` | 1 |
| `CKSEL3` | 1 |
| `CKSEL2` | 1 |
| `CKSEL1` | 1 |
| `CKSEL0` | 0 |
The last four bits are about the clock selection, and their default value `0010` stands for the 8MHz internal RC oscillator as source of the clock.
In our program above, we're using `1110`, which we can understand by looking at the section on the clock (chapter 6, notably [page 24](http://fab.cba.mit.edu/classes/863.09/people/ryan/week5/ATtiny44%20Data%20Sheet.pdf#page=24), or just *search* for the fuse names):
We're in the last row case (`CKSEL3..0` is within `1000`-`1111` cases, for external crystal or resonator), which means we need to go look farther (at page 27):
And there we find that we're assuming an external crystal/resonator, with frequency from 8MHz and above (`CKSEL3..1` is `111`).
The last bit `CKSEL0` requires looking a bit further:
Since we have `CKSEL0=0`, it means we assume a ceramic resonator.
Given the additional `SUT1..0` as `01` (3rd and 4th bits from the left in `0101_1110`, we know that we have BOD (**B**rown-**o**ut **D**etector) enabled.
You can read about that [there](https://microchipdeveloper.com/8avr:bod).
The `CKSEL` fuses are the fuses you are the most likely to every change, whereas the other ones are typically very specific.
Finally, if we look at bits 7 (`CKDIV8` set as `0`) and 6 (`CKOUT` as `1`), we can see that it is such that:
* `CKDIV8=0` (programmed) so that the the initial prescaler is set to 1/8. Reading about the prescaler on [page 30](http://fab.cba.mit.edu/classes/863.09/people/ryan/week5/ATtiny44%20Data%20Sheet.pdf#page=30), we see that this just sets the initial prescaler value, but that is available in the program space, which can then be changed at runtime.
* `CKOUT=1` (unprogrammed) does not enable clock output. Had we set it to `0`, the same section near the prescaler on [page 29](http://fab.cba.mit.edu/classes/863.09/people/ryan/week5/ATtiny44%20Data%20Sheet.pdf#page=29) tells that pin `CKOUT` (`PB2`) would be outputting the clock signal.
This output can be useful if you need to debug your clock while using an external clock.
In this way, you can check whether your clock parameters (e.g. capacitors) are doing a good job and your frequency is what you expect.
1. Read the manuals
2. Read the datasheet (when you need it, no need to memorize)
3. Choose your fuses carefully, notably beware of
* `RSTDISBL` in the *Fuse High Byte*, since this prevent further programming (needed in FabTinyISP), or similar disable / enable options
* Choose the clock carefully and especially the meaning hidden behind each bit's value (e.g. ceramic vs crystal, frequency range ...)
* For debugging, consider outputting the clock signal