Arc beats Python for a web-controlled stereo

I recently set up a web interface to my stereo, implementing it in both Python and Arc. I found that, much to my surprise, writing the Arc implementation was much easier than the Python implementation in this case.

In my stereo controller system, I go to a web page that has an image of the appropriate remote control, and click the button. The Javascript in the web page sends a hex code back to the server, which sends the code to an Arduino microcontroller board, which converts the code into an infrared signal. The stereo interprets this signal as a command from the remote control and performs the desired action. It sounds complicated, but it responds quickly, even if I use the browser on my cell phone. I wrote up more details on the system earlier.
Remote control on G1 phone
The web server for this project is fairly straightforward: it needs to serve some static HTML pages and images, and needs to receive POST messages and forward the data over the serial port to the Arduino. I originally wrote the code in Python:

import cgi
import serial

from BaseHTTPServer import HTTPServer
from SimpleHTTPServer import SimpleHTTPRequestHandler

class MyHandler(SimpleHTTPRequestHandler):
  def do_POST(self):
    if self.path == '/arduino':
      form = cgi.FieldStorage(fp=self.rfile, headers=self.headers,
      code = form['code'].value
      print 'Sent:', code
      self.send_header('Content-type', 'text/html')
    return self.do_GET()

arduino = serial.Serial('/dev/ttyUSB0', 9600, timeout=2)
server = HTTPServer(('', 8080), MyHandler).serve_forever()
The Python web server is based on SimpleHTTPServer. To handle POST requests to the /arduino URL, I made a simple handler subclass that pulls the data out of the response, sends it to the serial port, and returns a blank response. The last two lines open the serial port and start up the server. The static web page serving comes "for free."

After writing this, I decided to try writing a version in Arc:

(= srvops* (table))

(= arduino-serial (outfile "/dev/ttyUSB0" 'binary))

(defopr || req "/index.html")

(defop arduino req (w/stdout (stderr)
  (let code (arg req "code")
    (write code arduino-serial)
    (prn "Sent code:" code))))

(thread (serve 8080))
Since you may not be familiar with the Arc web server, I'll give a brief explanation. The first line clears out the server's default pages (login, whoami, prompt, etc) that make sense for news.yc, but not for me. The next line opens the serial port. Next, the mysterious || redirects the home page. The next four lines are the handler for "/arduino": the "code" parameter is pulled out of the request (req). The code is written to the serial port, and printed to the log. The last line starts a serving thread on port 8080.

I'm not going to count tokens since that's a silly metric, but the Arc code is a clear winner here. The first Python disadvantage is the hideous cgi.FieldStorage call to parse the POST data; apparently this is the way to do it, but it took me considerable time to get this right. The second Python disadvantage is the necessity to set up the return code and content-type for the response. I ended up with a browser-specific bug when I didn't have this right, since some Javascript interpreters expect a particular response. Arc automatically sends a valid response. The main Arc disadvantage is there's no real serial port support; the Arc code just opens the port and hopes for the best. On the other hand, Python's serial library lets you set the baud rate, timeouts, and other parameters.

Arc and Python have a few more minor differences. Python requires imports, which Arc doesn't. Arc has the silly clearing of srvops* to avoid unwanted pages. Starting the server thread is simpler in Arc but not as flexible.

To summarize, I found the Python code difficult and error-prone to write, and the Arc code was shorter and just worked. This was a surprise to me; usually I find Python straightforward and Arc gives me pain when I have to write a bunch of code for some trivial thing it doesn't support. (For example, generating a time string in my temperature monitoring.) In this case, however, Arc came out ahead. Perhaps this is because Arc's flagship application is a web server, so it handles web serving well.

Obviously this is a fairly trivial program - the complexity is actually in the Javascript code and the C++ code on the Arduino - so I'm not claiming this as an overall victory for Arc over Python. But it's still interesting to find Arc came out ahead in this case.

TV-B-Gone for the Arduino

I've ported the TV-B-Gone code to run on the Arduino board. If you haven't seen a TV-B-Gone, it's a cute gadget that you point at a TV that's bothering you, and it turns the TV off. Internally, it's an infrared remote that broadcasts more than 100 different off codes that work on almost any TV. I figured it would be interesting to get the TV-B-Gone running on the Arduino.
TV-B-Gone running on an Arduinio

Update (November 2010)

I have a new, improved version of the software. Details and download are here.

The software

I started with the TV-B-Gone firmware, which is available under Creative Commons license. The TV-B-Gone runs on an ATtiny85 microcontroller, not the ATmega328 used by the Arduino, so some porting was required. One tricky part of the port is the IR signals are generated using the low-level hardware timer for PWM (pulse-width modulation). Since the ATtiny85 runs at 8MHz, and the Arduino runs at 16MHz, the timing code needed to be re-written, both the 10us delay loop and the timer register operations. The quick summary is I'm using Timer 2, scaling the clock by 8, using OCR2A to control the frequency and OCR2B to control the duty cycle, and have output on pin 3 (OC2B). (You are not expected to understand this. See my article on Secrets of Arduino PWM for details of how these timers work.)
#define freq_to_timerval(x) (F_CPU / 8 / x - 1)
    pinMode(IRLED, OUTPUT);
    TCCR2A = _BV(COM2A0) | _BV(COM2B1) | _BV(WGM21) | _BV(WGM20);
    TCCR2B = _BV(WGM22) | _BV(CS21);
    OCR2A = freq; 
    OCR2B = freq / 3; // 33% duty cycle
Another tricky thing is the original firmware stores the IR codes in program space, not RAM. The microcontrollers use a Harvard architecture, which means they have separate memory and data paths for code and data, unlike normal processors that use a von Neumann architecture. Because RAM storage is so small on these microcontrollers, the IR codes are stored in program space, with the result you need to use special methods to access them, rather than normal C pointers. Program memory storage is indicated with the PROGMEM macro and pulled out of memory using special functions such as pgm_read_byte. This makes the code somewhat confusing, for example:
const struct IrCode *NApowerCodes[] PROGMEM = {

data_ptr = (PGM_P)pgm_read_word(NApowerCodes+i);  

The original firmware stores the codes as compressed indices into tables of durations. See the TV-B-Gone design for details. Unfortunately, the Arduino's compiler didn't like the zero-length array in the IrCode structure, so I needed to rewrite the long file of codes to include another level of indirection.

Apart from these factors, porting was straightforward, and I tried to keep the original code where possible. I ripped out all the power-up and watchdog code. I replaced the low-level serial code with the Arduino's Serial library. Finally, I packaged it up into an Arduino sketch, which you can download. (Update: get the improved version here.)

The hardware

The hardware is straightfoward. I use pin 13 for the status LED; some Arduinos conveniently have an LED already wired up. Pin 3 is the PWM output to the IR LED through a 100 ohm resistor. If you want more range, add a driver transistor and use multiple IR LEDs. Pin 0 (update: pin 5) selects North American or European codes; leave it unconnected for North America and ground it for European. Pin 1 (update: 2) is connected via a pushbutton to ground to trigger the code transmission. (Update: pins have changed in the new version.)

To use the Arduino TV-B-Gone, point the IR LED at your TV, push the button, and your TV should turn off when the circuit hits the right code, which could take up to a minute. The visible LED should flash for each code that is transmitted. If the circuit doesn't work, use a cellphone camera to verify that the IR LED is transmitting. Don't expect more than a few feet range unless you use a transistor to increase the power. The code will print debugging output to the serial port if you set DEBUG in main.h.


Is the Arduino TV-B-Gone practical? If you want a TV-B-Gone that you can carry around to amaze your friends, the kit or product is much more compact, has high-powered IR outputs for longer range, and is nicely packaged. However, the Arduino platform is convenient for experimenting, and many people already have it, so I expect people will find it interesting.

You might wonder why I'm not using my Arduino Infrared Library for this project. Actually, I plan to at some point, but I wanted to get the straightforward port working first. In addition, my library will need some extensions to support all the codes that the TV-B-Gone supports.

I hope you enjoy the Arduino TV-B-Gone, and remember to use it for good, not evil!

Controlling your stereo over the web with the Arduino infrared library

Here's how you can control your stereo over the web. Not only that, but any other electronics with a IR remote control can now be controlled through your browser. You can even use your smart phone's browser to control your stereo from across the room or around the world.
Remote control on G1 phone
This project has several pieces: a simple Python web server running on your computer, an Arduino with my infrared library, and an IR LED or other emitter. The Python web server provides a web page with a graphical remote control. Clicking on the web page sends a code to the web server, which sends it over the serial port to the Arduino, which sends it to the IR LED, which controls your device.

The web server

I used a rather trivial Python web server (based on SimpleHTTPServer) that performs two tasks. First, it provides the static HTML pages and images. Second, it receives the POST requests and sends them to the Arduino using the pyserial library.

The following code excerpt shows the handler that processes POSTs to /arduino by extracting the code value out of the POST data and sending it to the Arduino over the serial line. The Python server automatically provides static pages out of the current directory by default; this is how the HTML files and images are served. The code assumes the serial port is /dev/ttyUSB0, which is typically the case on Linux.

class MyHandler(SimpleHTTPRequestHandler):
  def do_POST(self):
    if self.path == '/arduino':
      form = cgi.FieldStorage(fp=self.rfile, headers=self.headers,
      code = form['code'].value
      self.send_header('Content-type', 'text/html')
    return self.do_GET()

arduino = serial.Serial('/dev/ttyUSB0', 9600, timeout=2)
server = HTTPServer(('', 8080), MyHandler).serve_forever()
The server can be accessed locally at http://localhost:8080. You may need to mess around with your firewall and router to access it externally; this is left as an exercise for the reader. I found that using my cell phone's browser via Wi-Fi worked surprisingly well, and over the cellular network there was just a slight lag (maybe 1/3 second).

The Arduino code

The code on the Arduino is pretty simple. It reads a command from the serial port and makes the appropriate IR library call. Commands consist of a character indicating the type of code, followed by 8 hex characters. For instance, "S0000004d1" sends the code 4d1 using Sony protocol; this is "play" on my Sony CD player. "N010e03fc" sends 010e03fc using NEC protocol; this turns my Harman Kardon stereo on. The full code is here, but some highlights:
void processSerialCode() {
  if (Serial.available() < 9) return;
  char type =;
  unsigned long code = 0;
  // Read 8 hex characters into code (omitted)
  if (type == 'N') {
    irsend.sendNEC(code, 32);
  else if (type == 'S') {
    // Send Sony code 3 times
    irsend.sendSony(code, 12);
    irsend.sendSony(code, 12);
    irsend.sendSony(code, 12);
  // More code for RC5 and RC6
In more detail, the Arduino waits for 9 characters to be available on the serial port. It then parses the hex value and calls the appropriate IR library send routine. The Arduino code does some special-case stuff for the different code types. Sony codes are transmitted three times as the protocol requires. The RC5 and RC6 protocol uses a toggle bit that is flipped on each transmission. (Disclaimer: I don't have RC5/RC6 devices, so this code is untested.) If your device uses a different protocol that the library doesn't support, you're out of luck unless you add the protocol to the library. That's probably not too hard; a couple people have already implemented new protocols.
Arduino controlling stereo via IR

The web page

Most of the smarts of the system are in the web page. In my setup, I have a HK-3370 stereo, and a Sony CDP-CE335 CD player. I took a picture of the remote for each and used an HTML image map to make each button clickable. Clicking on a button uses Ajax to POST the appropriate IR code to the server.

I use Ajax to send the code to the server to avoid reloading the web page on every click. The Javascript code is verbose but straightforward. The first part of the code creates the XML request object; unfortunately different browsers use different objects. The next part of the code creates and sends the POST request. The actual data sent to the server is, for instance, "code=N12345678" to send 0x12345678 using NEC protocol.

function button(value) {
  if (window.XMLHttpRequest) {
   request = new XMLHttpRequest();
  } else if (window.ActiveXObject) {
    try {
      request = new ActiveXObject("Msxml2.XMLHTTP");
    } catch (e) {
      try {
        request = new ActiveXObject("Microsoft.XMLHTTP");
      } catch (e) {}
  }'POST', '/arduino', true);
  request.setRequestHeader('Content-type', 'application/x-www-form-urlencoded');
  request.setRequestHeader('Content-length', value.length);
  request.setRequestHeader('Connection', 'close');
  request.send('code=' + value);
The clickable buttons are defined in the HTML through an image and an image map. Generating the image and the image map is the hard part of the whole project:
<img src="hk3370.png" width="268" height="800" border="0" usemap="#map" />
<map name="map">
<area shape="rect" coords="60,97,85,111" href="#" alt="on" onClick="button('N010e03fc')" />
<area shape="rect" coords="104,98,129,109" href="#" alt="off" onClick="button('N010ef906')" />
Each line in the map defines a region of the image as a button with a particular code. When clicked, this region will call the button function with the appropriate code, causing the code to be sent to the web server, the Arduino, and finally to the stereo. The alt text isn't actually used, but I recommend it to keep track of what button is associated with each line. The href causes the cursor to change when you go over the region, but isn't necessary.
Image of remote control in Internet Explorer

In this approach, the Arduino code and web server code are very simple, as they don't need to know the functions of the codes. A different way to implement this system would be to put the table of codes ("SONY_CD_ON" = 4d1, etc.) either in the web server code or the Arduino code.

To generate the web pages, I took a picture of the remote with my camera and cleaned up the picture with GIMP. I used the GIMP image map plugin to create the image map. I outlined each button, then filled in the URL with the appropriate IR code, and filled in the alt text with the name of the button. Finally, I copied the map file into the HTML file and edited it to use the Javascript function.

The easiest way to obtain the IR codes is to use the IRrecvDump example sketch included in my IR library. Simply press the button on your remote, see what code was sent, and put that code into the image map. Alternatively, you may be able to find codes in the LIRC database.

Once you get going, it's actually fairly quick to generate the image map. Select a button in the editor, click the physical button on the remote, copy the displayed value into the editor, and move on to the next button. As long as you don't get obsessive and start tweaking the regions to line up perfectly, it's pretty quick.

If you don't want to mess around with the image map, take a look at simple.html. This file shows how to use standard HTML buttons. Not as cool as the image, but much easier:

<input type="button" value="on" onClick="button('N010e03fc')" >
<input type="button" value="off" onClick="button('N010ef906')" >
Simple button arrangement

The hardware

The Arduino does the work of converting the hex code into an IR signal. The IR library uses digital PWM pin 3 as output, which must be connected to your IR emitter. I use a 100 ohm resistor to limit current. Wiring it up is trivial.
Schematic of IR connection to Arduino

I used a Tivo IR blaster that I had lying around, but a plain IR LED will work too. The following picture shows an IR blaster attached to my sterero. Note that the blaster needs to be positioned about the stereo's IR receiver; it may be easier to see the receiver with a flashlight.
IR blaster attached to stereo

Putting it all together

To summarize the steps:
  • Download the files and unpack them.
  • Get the infrared library and install it.
  • Aim the IR source at the device(s) to control, and connect it to the Arduino. (You may want to use a visible LED for testing.)
  • Connect the Arduino to the computer's USB port.
  • Compile and install IRwebremote.pde using the Arduino IDE.
  • Create HTML files for your remote controls. (This is the hard step.)
  • Run the Python server: python
  • Go to http://localhost:8080 and control the devices.
This project has multiple pieces and requires customization, so it's not a plug-and-play project. I only recommend trying this if you know what you're doing. Good luck!

Lorem Ipsue: when internationalization goes bad

I recently saw a Master cable lock for sale with the interesting text "Lore Ipsum!" and "Lorem Ipsue!" underneath the word "Pull". If you've done any graphic design or web mockups, you're probably familiar with the Lorem ipsum text that's traditionally used as a placeholder. This Latin-based text is used, for instance, when you want to show the style and layout of a website, but don't want people to get distracted by the words.
Lock with text 'Lore ipsum' and 'Lorem ipsue'
Apparently when they designed the lock packaging, they put in placeholders but forgot to replace them with the French and Spanish translations ("Tirez!" and "¡Tira!"). I find "Lorem Ipsue!" in place of "Lorem Ipsum" interesting; maybe it is supposed to be a more Spanish-sounding placeholder, but I would have been more impressed by ¡Lorem Ipsum!

The moral is: make sure you check your internationalization; just because it looks foreign doesn't mean it's right.

Notes on dual-booting Windows 7 and Linux

Dual-booting Windows 7 and Linux is mostly straightforward, but I ran into a few problems. The following notes describe my setup, mostly for my own benefit.

If you're looking for a comprehensive dual-boot tutorial, try apc or Loko.

My setup

I'm running Windows 7 and Fedora 11 Linux on a Gateway SX2800.


I use the Gnome Partition Editior (GParted). Windows 7 conveniently includes a partition software under Disk Management, but inconveniently it can't shrink a volume by a large amount.

My partitioning setup is a bit crazy. Because I bought my computer before Windows 7 was released, I used Windows 7 beta until I got the official upgrade DVD. One unpleasant surprise is you have to reinstall to get from the beta to Windows 7 Home Premium; you can't just upgrade. I had an empty E: partition, so I installed the offical release there, transferred my files with Windows Easy Transfer, and then spent an evening reinstalling all my software. I store most of my data on a FAT32 partition D:to avoid messing around with NTFS on Fedora.

You'll probably want something simpler than what I have, but my partition table is:

$fdisk /dev/sda
   Device Boot      Start         End      Blocks   Id  System
/dev/sda1               1        1698    13631488   27  Unknown
/dev/sda2   *        1699       14445   102390277+   7  HPFS/NTFS
/dev/sda3           14446       27193   102398310    7  HPFS/NTFS
/dev/sda4           27194       77825   406701540    5  Extended
/dev/sda5           27194       28289     8803588+  82  Linux swap / Solaris
/dev/sda6           28290       70176   336457296    b  W95 FAT32
/dev/sda7           70177       77825    61440561   83  Linux
sda1 is a mysterious Gateway recovery partition, sda2 is C:, sda3 is E:, sda6 is D: for most of my data, and sda7 is the Linux partition.


/dev/sda7               /                       ext3    defaults        1 1
tmpfs                   /dev/shm                tmpfs   defaults        0 0
devpts                  /dev/pts                devpts  gid=5,mode=620  0 0
sysfs                   /sys                    sysfs   defaults        0 0
proc                    /proc                   proc    defaults        0 0
/dev/sda5  swap                    swap    defaults        0 0
/dev/sda6  /d   vfat auto,rw,user,sync,exec,dev,suid,uid=500,gid=500,umask=000 0 2

My GRUB config

title Fedora (
 root (hd0,6)
 kernel /boot/vmlinuz- ro root=/dev/sda7 rhgb quiet
 initrd /boot/initrd-
title Windows 7
 rootnoverify (hd0,1)
 chainloader +1
Note the off-by-one: (hd0,6) boots off sda7. For Windows, you need to specify the partition with the Windows boot loader; if I try to boot E: directly with (hd0,2), I get BOOTMGR IS MISSING.

Editing the GRUB config from Linux rescue boot

Invariably I end up unable to boot Linux from my disk. To fix this, I boot from my Linux CD and then fix the GRUB config. The following assumes sda7 is the Linux partition:
mount /mnt /dev/sda7
cd /mnt/boot/grub
vi menu.lst
grub-install --root-directory=/mnt /dev/sda

Editing the Windows boot loader configuration

If you have multiple versions of Windows, you'll need to configure the Windows boot loader through bcdedit: details.

Other random errors

My Fedora Core boot failed with an infinite loop of "init: ttyx respawning too fast, stopped". The solution was to change selinux=enforcing to selinux=permissive in /etc/selinux/config.


The above notes are mostly for my own benefit the next time I need to configure GRUB, but maybe they'll help someone else. Disclaimer: it's very easy to lose data or mess up your system so it won't boot. Back everything up and make sure you have Linux and Windows rescue disks. Don't blame me if things go wrong.

IR Bubbles: Controlling a relay with an Arduino and an IR remote

An Arduino can be used to turn devices on and off via an infrared remote. As an entirely impractical example of how to use my Arduino infrared remote library to turn things on and off, I have set up my Arduino to turn a bubble maker on and off under remote control. The same techniques can be used to control an LED or anything else via IR remote.

Arduino blowing bubbles

The circuit is straightforward. An IR sensor is connected to the Arduino to detect the infrared signal. On the output side, I used a standard relay circuit. To control the bubble maker, I wired the relay in place of the power switch. A 5 volt source wasn't enough to close the relay I had handy, but 9 volts was. For this reason, I powered the Arduino off a 9V battery and connected the relay to Vin so the relay would get the full 9 volts. (A 9 volt AC adapter also works.)

Arduino IR relay schematic

The code

The following code controls the relay. Any detected IR code will toggle the output on or off. Press a button once on the remote and the output will turn on. Press a button again and the output will turn off. (You will, of course, need to download my Arduino infrared remote library to use this code.)
#include <IRremote.h>

int RECV_PIN = 11;
int OUTPUT_PIN = 4;

IRrecv irrecv(RECV_PIN);
decode_results results;

void setup()
  pinMode(13, OUTPUT);
  irrecv.enableIRIn(); // Start the receiver

int on = 0;
unsigned long last = millis();

void loop() {
  if (irrecv.decode(&results)) {
    // If it's been at least 1/4 second since the last
    // IR received, toggle the relay
    if (millis() - last > 250) {
      on = !on;
      digitalWrite(OUTPUT_PIN, on ? HIGH : LOW);
    last = millis();      
    irrecv.resume(); // Receive the next value
The above code only toggles the output if there is at least a 1/4 second (250 ms) gap since the last received signal. The motivation behind this is that remotes often repeat the signal as long as the button is held down, and we don't want the output to toggle continuously. One other thing to notice is the call to resume(); after a code is received, the library stops detection until resume() is called.

I should make it clear that although the above example uses a relay output, the relay is just for example. The above code will work with any arbitrary output: you can use it to control a LED or anything else that can be controlled by an Arduino output pin.

Using two buttons

Another way you might want to use an IR remote is have one button turn the output on, and another button turn it off. Here's the loop code to use a Sony DVD remote; the "play" button turns the output on, and the "stop" button turns the output off:
void loop() {
  if (irrecv.decode(&results)) {
    if (results.value == 0x4cb92) { // Sony DVD play
      digitalWrite(OUTPUT_PIN, HIGH);
    else if (results.value == 0x1cb92) { // Sony DVD stop
      digitalWrite(OUTPUT_PIN, LOW);
    irrecv.resume(); // Receive the next value
Just remember that you have to call resume after all decodes, not just the ones you're interested in.

Getting the codes

You may wonder how I got the Sony codes. One way is to look at the appropriate LIRC config file. (At some point I'll write up a full explanation of the LIRC config files.) An easier way is to just dump out to the serial port the codes that your Arduino receives. Press the buttons on your remote, and read out the values received:
void loop() {
  if (irrecv.decode(&results)) {
    Serial.println(results.value, HEX);
    irrecv.resume(); // Receive the next value

Using a specific button

A third way you might want to use an IR remote is to have a specific button toggle the output on and off. Here's code that uses the Tivo button on a Tivo Series 2 remote to toggle the output on and off. Again, you can get the button code from a LIRC config file, or by dumping the output to your serial port.
void loop() {
  if (irrecv.decode(&results)) {
    if (results.value == 0xa10ce00f) { // TIVO button
      // If it's been at least 1/4 second since the last
      // IR received, toggle the relay
      if (millis() - last > 250) {
        on = !on;
        digitalWrite(OUTPUT_PIN, on ? HIGH : LOW);
      last = millis();
    irrecv.resume(); // Receive the next value

The hardware setup

Arduino relay circuit

The hardware setup is pretty simple. The above photo shows the IR detector on the left, and the relay on the right. The drive transistor, resistor, and diode are behind the relay, and the blue terminal strip is in front. The black box to the right of the Arduino is a 9V battery holder. The box underneath is just to hold everything up.

You may wonder why I placed the input and output components so far apart, and why I didn't use the protoboard power and ground buses. It turns out that the motor on the bubble maker is a nasty source of RFI interference that would jam the IR detector's signal if the wires to the bubble maker came near the IR detector wiring. Capacitors didn't help, but distance did. Most likely this won't be a problem for you, but I figured I should mention it.

In conclusion, happy hacking with the infrared library. Let me know if you do anything interesting with it!

Simple genome analysis with Arc: In which I examine the XMRV genome and discover little of interest

I decided to examine the genome of the retrovirus XMRV to see what I could learn. This virus, with the absurd name Xenotropic Murine leukemia virus-Related Virus, was found a few years ago in many cases of prostate cancer and very recently found in many people suffering from chronic fatigue syndrome (more, more). If these studies turn out to be true, it would be quite remarkable that an obscure new retrovirus causes such differing diseases.

My goal was to take the XMRV genome (the sequence of c's, g's, a's, and t's that make up its RNA), and see if the combination "cg" is a lot more rare than you'd expect. This probably seems like a pretty random thing to do, but there's actually a biological reason.

One way the human immune system detects invaders is by looking for DNA containing a cytosine followed by a guanine (which is called a CpG dinucleotide). In mammals, most CpG dinucleotides are methylated (specially tagged), so any CpG that shows up is a sign of something invading. My hypothesis was that XMRV might have a very low number of CpG dinucleotides, helping it avoid the immune system and contributing to its ability to cause disease.

Conveniently, the genome of XMRV has been sequenced and can be downloaded, consisting of 8185 bases:

        1 gcgccagtca tccgatagac tgagtcgccc gggtacccgt gttcccaata aagccttttg
       61 ctgtttgcat ccgaagcgtg gcctcgctgt tccttgggag ggtctcctca gagtgattga
It would be a simple matter to use Python to count the number of CG pairs in the genome, but I figured using the Arc language would be more interesting.

The first step is to write a function to parse out the DNA sequence from the genome file; this consists of the lines between ORIGIN and //, with the spaces and numbers removed.

(def skip-to-origin (infile)
  (let line (readline infile)
    (if (is 'eof line) line ;quit on eof
        (litmatch "ORIGIN" line) line ;quit on ORIGIN
        (skip-to-origin infile)))) ; recurse

; Reads the data from one line
; Returns nil on EOF or // line
(def read-data-line (infile)
  (let line (readline infile)
    (if (is 'eof line) nil
        (litmatch "//" line) nil
 (keep [in _ #\c #\g #\a #\t] line))))

; Read the sequence data from ORIGIN to //
; Returns a single string
(def readseq (infile)
   (skip-to-origin infile)
   (string (drain (read-data-line infile))))
This code simply uses skip-to-origin to read up to the ORIGIN line, and then read-data-line to read the cgat data from each following line.

The next step is a histogram function to count the number of times each nucleotide occurs into a table. (Edit: I've shortened my original code with some advice from

((def hist (seq)
  (counts (coerce seq 'cons)))
After downloading the genome as "xmrv", I can load the sequence into seq and generate the histogram:
arc> (= seq (w/infile inf "xmrv" (readseq inf)))
"gcgccagtcatccgatagactgagtcgcccgggtacccgtgt ...etc"
arc> (len seq)
arc>(hist seq)
#hash((#\t . 1732) (#\g . 2057) (#\a . 2078) (#\c . 2318))
The sequence is 8185 nucleotides long as expected, with 1732 T's, 2057 G's, 2078 A's, and 2318 C's. To count the number of times each pair of nucleotides occurs, I made a more general function that will handle pairs or any other sequence of n nucleotides:
(def histn (seq n)
  (w/table h
    (for i 0 (- (len seq) n)
      (++ (h (cut seq i (+ i n)) 0)))
Since I got tired of looking at raw hash tables, I made a short function to format the output as well as generating percentages.
(def prettyhist (h)
  (let count (reduce + (vals h))
    (let sorted (sort (fn ((k1 v1) (k2 v2)) (> v1 v2))
                      (accum addit (each elt h (addit elt))))
      (each (k v) sorted
        (prn k ": " v " (" (num (* (/ v count) 100.) 2) "%)")))))
Running these functions:
arc> (prettyhist (histn seq 2))
cc: 862 (10.53%)
gg: 639 (7.81%)
ag: 616 (7.53%)
ct: 612 (7.48%)
aa: 588 (7.18%)
ga: 585 (7.15%)
ca: 537 (6.56%)
ac: 529 (6.46%)
tg: 494 (6.04%)
tc: 469 (5.73%)
gc: 458 (5.6%)
tt: 401 (4.9%)
gt: 375 (4.58%)
ta: 368 (4.5%)
at: 344 (4.2%)
cg: 307 (3.75%)
This shows that CG is the most uncommon combination, but not super-low. How does this compare to the frequence if the C/G/A/T frequency was the same, but they were ordered randomly? We can compute this by computing all the possibilities of taking two letters from the original sequence. Instead of actually summing all the combinations, we can just take the cross-product of the original histogram:
(def cross (h1 h2)
  (let h (table)
    (each (k1 v1) h1
      (each (k2 v2) h2
        (let combo (string k1 k2)
   (if (no (h combo))
     (= (h combo) 0))
   (= (h combo) (+ (h combo) (* v1 v2))))))
This gives us the result:
arc> (prettyhist (cross (hist seq) (hist seq)))
cc: 5373124 (8.02%)
ac: 4816804 (7.19%)
ca: 4816804 (7.19%)
cg: 4768126 (7.12%)
gc: 4768126 (7.12%)
aa: 4318084 (6.45%)
ag: 4274446 (6.38%)
ga: 4274446 (6.38%)
gg: 4231249 (6.32%)
tc: 4014776 (5.99%)
ct: 4014776 (5.99%)
at: 3599096 (5.37%)
ta: 3599096 (5.37%)
gt: 3562724 (5.32%)
tg: 3562724 (5.32%)
tt: 2999824 (4.48%)
This tells us that CG would appear 7.12% of the time if the sequence were randomly shuffled, but instead it appears only 3.75% of the time, so CG shows up about half as often as expected. This is in line with known results for the related MuLV virus, so it seems that there's nothing special about XMRV.

On the other hand, it's known that the HIV genome has a severely low amount of CpG:

arc> (prettyhist (histn (w/infile inf "hivbru" (readseq inf)) 2))
aa: 1096 (11.88%)
ag: 971 (10.52%)
ca: 766 (8.3%)
ga: 762 (8.26%)
at: 691 (7.49%)
ta: 665 (7.21%)
gg: 631 (6.84%)
tg: 548 (5.94%)
ac: 530 (5.74%)
tt: 524 (5.68%)
gc: 430 (4.66%)
ct: 428 (4.64%)
gt: 409 (4.43%)
cc: 381 (4.13%)
tc: 315 (3.41%)
cg: 81 (.88%)
I also took a look at the H1N1 flu sequences. The influenza genome is on 8 separate strands of RNA, so there are 8 separate sequences to process. (The HA segment is the "H" and the NA segment is the "N" in H1N1.) Many H1N1 sequences can be downloaded. I somewhat arbitrarily picked A/Beijing/02/2009(H1N1) since all 8 strands were sequenced. After saving the strands as flu1, ..., flu8, I ran the histogram on all the strands:
arc> (for i 1 8 (prn "---" i "---")
  (prettyhist (histn (w/infile inf (string "flu" i) (readseq inf)) 2)))
On all strands, "cg" was at the bottom of the frequency chart. Especially low-frequency strands are 1 (PB2) at 1.93% CpG, 2 (PB1) at 1.52%, 4 (HA) at 1.57%, and 6 (NA) at 1.77%. Strand 8 (NEP) was highest at 3.09%. So it looks like influenza is pretty low in CpG frequency, but not as low as HIV. (Edit: I've since found a paper that examines CpG in influenza in more detail.)

To conclude, Arc can be used for simple genome analysis. My original hypothesis that XMRV would have low levels of CpG dinucleotides holds, but not as dramatically as for HIV or H1N1 influenza. Apologies to biologists for my oversimplifications, and apologies to statisticians for my lack of p-values :-)

An Arduino universal remote: record and playback IR signals

I've implemented a simple IR universal remote that will record an IR code and retransmit it on demand as an example for my IR library. Handling IR codes is a bit more complex than it might seem, as many protocols require more than simply recording and playing back the signal.

To use the universal remote, simply point your remote control at the IR module and press a button on the remote control. Then press the Arduino button whenever you want to retransmit the code. My example only supports a single code at a time, but can be easily extended to support multiple codes.

The hardware

The IR recorder
The above picture shows the 9V battery pack, the Arduino board, and the proto board with (top-to-bottom) the IR LED, IR receiver, and pushbutton.

The circuitry is simple: an IR sensor module is connected to pin 11 to record the code, an IR LED is connected to pin 3 to transmit the code, and a control button is connected to pin 12. (My IR library article has details on the sensor and LED if you need them.)
Schematic of the IR recorder

The software

The code can be downloaded as part of my IRremote library download; it is the IRrecord example.

Handling different protocols

The code supports multiple IR protocols, each with its own slight complications:

Sony codes can be recorded and played back directly. The button must be held down long enough to transmit a couple times, as Sony devices typically require more than one transmission.

The common NEC protocol is complicated by its "repeat code". If you hold down a button, the remote transmits the code once followed by multiple transmissions of special repeat code. The universal remote records the code, not the repeat code. On playback, it transmits the code once, followed by the repeat code.

The RC5 and RC6 protocols handle repeated transmissions differently. They use two separate codes for each function, differing in a "toggle bit". The first time you hold down a button, the first code is transmitted repeatedly. The next time you hold down a button, the second code is transmitted repeatedly. Subsequent presses continue to alternate. The universal remote code flips the toggle bit each time it transmits.

The universal remote handles any other unknown protocol as a "raw" sequence of modulated IR on and off. The main complication is that IR sensor modules typically stretch out the length of the "on time" by ~100us, and shorten the "off time" correspondingly. The code compensates for this.

Most likely there are some codes that that can't handle, but it has worked with the remotes I've tried.

The code also prints debugging information to the serial console, which can be helpful for debugging any problems.

In conclusion, this is intended as a proof of concept rather than a useful product. The main limitation is supporting one code at a time, but it's straightforward to extend the code. Also note that the record and playback functions can be separated; if you know the IR codes you're dealing with, you can use just the necessary function.

A Multi-Protocol Infrared Remote Library for the Arduino

Code now on github

The most recent code is at If you have any issues, please report them there.

Do you want to control your Arduino with an IR remote? Do you want to use your Arduino to control your stereo or other devices? This IR remote library lets you both send and receive IR remote codes in multiple protocols. It supports NEC, Sony SIRC, Philips RC5, Philips RC6, and raw protocols. If you want additional protocols, they are straightforward to add. The library can even be used to record codes from your remote and re-transmit them, as a minimal universal remote.

Arduino IR remote

To use the library, download from github and follow the installation instructions in the readme.

How to send

This infrared remote library consists of two parts: IRsend transmits IR remote packets, while IRrecv receives and decodes an IR message. IRsend uses an infrared LED connected to output pin 3. To send a message, call the send method for the desired protocol with the data to send and the number of bits to send. The examples/IRsendDemo sketch provides a simple example of how to send codes:
#include <IRremote.h>
IRsend irsend;

void setup()

void loop() {
  if ( != -1) {
    for (int i = 0; i < 3; i++) {
      irsend.sendSony(0xa90, 12); // Sony TV power code
This sketch sends a Sony TV power on/off code whenever a character is sent to the serial port, allowing the Arduino to turn the TV on or off. (Note that Sony codes must be sent 3 times according to the protocol.)

How to receive

IRrecv uses an infrared detector connected to any digital input pin.

The examples/IRrecvDemo sketch provides a simple example of how to receive codes:

#include <IRremote.h>

int RECV_PIN = 11;
IRrecv irrecv(RECV_PIN);
decode_results results;

void setup()
  irrecv.enableIRIn(); // Start the receiver

void loop() {
  if (irrecv.decode(&results)) {
    Serial.println(results.value, HEX);
    irrecv.resume(); // Receive the next value
The IRrecv class performs the decoding, and is initialized with enableIRIn(). The decode() method is called to see if a code has been received; if so, it returns a nonzero value and puts the results into the decode_results structure. (For details of this structure, see the examples/IRrecvDump sketch.) Once a code has been decoded, the resume() method must be called to resume receiving codes. Note that decode() does not block; the sketch can perform other operations while waiting for a code because the codes are received by an interrupt routine.

Hardware setup

The library can use any of the digital input signals to receive the input from a 38KHz IR receiver module. It has been tested with the Radio Shack 276-640 IR receiver and the Panasonic PNA4602. Simply wire power to pin 1, ground to pin 2, and the pin 3 output to an Arduino digital input pin, e.g. 11. These receivers provide a filtered and demodulated inverted logic level output; you can't just use a photodiode or phototransistor. I have found these detectors have pretty good range and easily work across a room.

IR wiring

For output, connect an IR LED and appropriate resistor to PWM output pin 3. Make sure the polarity of the LED is correct, or it won't illuminate - the long lead is positive. I used a NTE 3027 LED (because that's what was handy) and 100 ohm resistor; the range is about 15 feet. For additional range, you can amplify the output with a transistor.

Some background on IR codes

An IR remote works by turning the LED on and off in a particular pattern. However, to prevent inteference from IR sources such as sunlight or lights, the LED is not turned on steadily, but is turned on and off at a modulation frequency (typically 36, 38, or 40KHz). The time when a modulated signal is being sent will be called a mark, and when the LED is off will be called a space.

Each key on the remote has a particular code (typically 12 to 32 bits) associated with it, and broadcasts this code when the key is pressed. If the key is held down, the remote usually repeatedly broadcasts the key code. For an NEC remote, a special repeat code is sent as the key is held down, rather than repeatedly sending the code. For Philips RC5 or RC6 remotes, a bit in the code is toggled each time a key is pressed; the receiver uses this toggle bit to determine when a key is pressed down a second time.

On the receiving end, the IR detector demodulates this signal, and outputs a logic-level signal indicating if it is receiving a signal or not. The IR detector will work best when its frequency matches the sender's frequency, but in practice it doesn't matter a whole lot.

The best source I've found for details on the various types of IR codes is SB IR knowledge base.

Handling raw codes

The library provides support for sending and receiving raw durations. This is intended mainly for debugging, but can also be used for protocols the library doesn't implement, or to provide universal remote functionality.

The raw data for received IR measures the duration of successive spaces and marks in 50us ticks. The first measurement is the gap, the space before the transmission starts. The last measurement is the final mark.

The raw data for sending IR holds the duration of successive marks and spaces in microseconds. The first value is the first mark, and the last value is the last mark.

There are two differences between the raw buffers for sending and for receiving. The send buffer values are in microseconds, while the receive buffer values are in 50 microsecond ticks. The send buffer starts with the duration of the first mark, while the receive buffer starts with the duration of the gap space before the first mark. The formats are different because I considered it useful for the library to measure gaps between transmissions, but not useful for the library to provide these gaps when transmitting. For receiving, 50us granularity is sufficient for decoding and avoids overflow of the gaps, while for transmitting, 50us granularity is more than 10% error so 1us granularity seemed better.

Obtaining codes for your remote

The easiest way to obtain codes to work with your device is to use this library to decode and print the codes from your existing remote.

Various libraries of codes are available online, often in proprietary formats. The Linux Infrared Remote Control project (LIRC), however, has an open format for describing codes for many remotes. Note that even if you can't find codes for your exact device model, a particular manufacturer will usually use the same codes for multiple products.

Beware that other sources may be inconsistent in how they handle these protocols, for instance reversing the order, flipping 1 and 0 bits, making start bits explicit, dropping leading or trailing bits, etc. In other words, if the IRremote library yields different codes than you find listed elsewhere, these inconsistencies are probably why.

Details of the receiving library

The IRrecv library consists of two parts. An interrupt routine is called every 50 microseconds, measures the length of the marks and spaces, and saves the durations in a buffer. The user calls a decoding routine to decode the buffered measurements into the code value that was sent (typically 11 to 32 bits).

The decode library tries decoding different protocols in succession, stopping if one succeeds. It returns a structure that contains the raw data, the decoded data, the number of bits in the decoded data, and the protocol used to decode the data.

For decoding, the MATCH macro determine if the measured mark or space time is approximately equal to the expected time.

The RC5/6 decoding is a bit different from the others because RC5/6 encode bits with mark + space or space + mark, rather than by durations of marks and spaces. The getRClevel helper method splits up the durations and gets the mark/space level of a single time interval.

For repeated transmissions (button held down), the decoding code will return the same decoded value over and over. The exception is NEC, which sends a special repeat code instead of repeating the transmission of the value. In this case, the decode routine returns a special REPEAT value.

In more detail, the receiver's interrupt code is called every time the TIMER1 overflows, which is set to happen after 50 microseconds. At each interrupt, the input status is checked and the timer counter is incremented. The interrupt routine times the durations of marks (receiving a modulated signal) and spaces (no signal received), and records the durations in a buffer. The first duration is the length of the gap before the transmission starts. This is followed by alternating mark and space measurements. All measurements are in "ticks" of 50 microseconds.

The interrupt routine is implemented as a state machine. It starts in STATE_IDLE, which waits for the gap to end. When a mark is received, it moves to STATE_MARK which times the duration of the mark. It then alternates between STATE_MARK and STATE_SPACE to time marks and spaces. When a space of sufficiently long duration is received, the state moves to STATE_STOP, indicating a full transmission is received. The interrupt routine continues to time the gap, but blocks in this state.

The STATE_STOP is used a a flag to indicate to the decode routine that a full transmission is available. When processing is done, the resume() method sets the state to STATE_IDLE so the interrupt routine can start recording the next transmission. There are a few things to note here. Gap timing continues during STATE_STOP and STATE_IDLE so an accurate measurement of the time between transmissions can be obtained. If resume() is not called before the next transmission starts, the partial transmission will be discarded. The motivation behind the stop/resume is to ensure the receive buffer is not overwritten while it is still being processed; debugging becomes very difficult if the buffer is constantly changing.

Details of the sending library

The transmission code is straightforward. To ensure accurate output frequencies and duty cycles, I use the PWM timer, rather than delay loops to modulate the output LED at the appropriate frequency. (See my Arduino PWM Secrets article for more details on the PWM timers.) At the low level, enableIROut sets up the timer for PWM output on pin 3 at the proper frequency. The mark() method sends a mark by enabling PWM output and delaying the specified time. The space() method sends a space by disabling PWM output and delaying the specified time.

The IRremote library treats the different protocols as follows:

NEC: 32 bits are transmitted, most-significant bit first. (protocol details)

Sony: 12 or more bits are transmitted, most-significant bit first. Typically 12 or 20 bits are used. Note that the official protocol is least-significant bit first. (protocol details) For more details, I've written an article that describes the Sony protocol in much more detail: Understanding Sony IR remote codes.

RC5: 12 or more bits are transmitted most-significant bit first. The message starts with the two start bits, which are not part of the code values. (protocol details)

RC6: 20 (typically) bits are transmitted, most-significant bit first. The message starts with a leader pulse, and a start bit, which is not part of the code values. The fourth bit is transmitted double-wide, since it is the trailer bit. (protocol details)

For Sony and RC5/6, each transmission must be repeated 3 times as specified in the protocol. The transmission code does not implement the RC5/6 toggle bit; that is up to the caller.

Adding new protocols

Manufacturers have implemented many more protocols than this library supports. Adding new protocols should be straightforward if you look at the existing library code. A few tips: It will be easier to work with a description of the protocol rather than trying to entirely reverse-engineer the protocol. The durations you receive are likely to be longer for marks and shorter for spaces than the protocol suggests. It's easy to be off-by-one with the last bit; the last space may be implicit.


To make it easier to debug problems with IR communication, I have optional debugging code in the library. Add #define DEBUG to the beginning of your code to enable debugging output on the serial console. You will need to delete the .o files and/or restart the IDE to force recompilation.

Problems with Transmission

If sending isn't working, first make sure your IR LED is actually transmitting. IR will usually show up on a video camera or cell phone camera, so this is a simple way to check. Try putting the LED right up to the receiver; don't expect a lot of range unless you amplify the output.

The next potential problem is if the receiver doesn't understand the transmitter, for instance if you are sending the wrong data or using the wrong protocol. If you have a remote, use this library to check what data it is sending and what protocol it is using.

An oscilloscope will provide a good view of what the Arduino or a remote is transmitting. You can use an IR photodiode to see what is getting transmitted; connect it directly to the oscilloscope and hold the transmitter right up to the photodiode. If you have an oscilloscope, just connect the oscilloscope to the photodiode. If you don't have an oscilloscope, you can use a sound card oscilloscope program such as xoscope.

The Sony and RC5/6 protocols specify that messages must be sent three times. I have found that receivers will ignore the message if only sent once, but will work if it is sent twice. For RC5/6, the toggle bit must be flipped by the calling code in successive transmissions, or else the receiver may only respond to a code once.

Finally, there may be bugs in this library. In particular, I don't have anything that receives RC5/RC6, so they are untested.

Problems with Receiving

If receiving isn't working, first make sure the Arduino is at least receiving raw codes. The LED on pin 13 of the Arduino will blink when IR is being received. If not, then there's probably a hardware issue.

If the codes are getting received but cannot be decoded, make sure the codes are in one of the supported protocols. If codes should be getting decoded, but are not, some of the measured times are probably not within the 20% tolerance of the expected times. You can print out the minimum and maximum expected values and compare with the raw measured values.

The examples/IRrecvDump sketch will dump out details of the received data. The dump method dumps out these durations but converts them to microseconds, and uses the convention of prefixing a space measurement with a minus sign. This makes it easier to keep the mark and space measurements straight.

IR sensors typically cause the mark to be measured as longer than expected and the space to be shorter than expected. The code extends marks by 100us to account for this (the value MARK_EXCESS). You may need to tweak the expected values or tolerances in this case.

The library does not support simultaneous sending and receiving of codes; transmitting will disable receiving.


I've used this library for several applications:

Other projects that use this library

Other Arduino IR projects

I was inspired by Building a Universal Remote with an Arduino; this doesn't live up to being a universal remote, but has a lot of information. The NECIRrcv library provided the interrupt handling code I use.

Arc + Arduino + ARM: Temperature monitoring

I'm using my Arc web server running on an ARM-based SheevaPlug to get analog data from my Arduino. The frame below and the graph of temperature and illumination are coming live are a screenshot from the SheevaPlug; they are dynamically generated by some simple Arc code which also collects the data from the Arduino. (Note: this is currently a screenshot, as I'm currently using the SheevaPlug for another project.)
The graph shows the daily cycle of illumination (green line), as well as temperature climbing during the day (red line). The temperature goes way, way up when the sun hits the temperature sensor directly. Perhaps I should find some shade for it. For one day, the data went crazy; this is where the wires from the sensor got knocked out of the Arduino by the vacuum cleaner.

The SheevaPlug and Arduino

The Arc web server is running on the ARM-based SheevaPlug, a small Linux server that plugs into the wall. It is connected by USB to an Arduino microcontoller, which does the analog to digital conversion. In the photo, the SheevaPlug also has an Ethernet cable attached. A 4-conductor wire connects the Arduino to the temperature and light sensors outside. This wire is old phone cable, but as described below, I probably should have used something more shielded.

The Arduino code

The Arduino sketch (download) simply reads the voltages and writes them to the serial port:
#define SUPPLY 5.14 // Supply voltage (measured)

void setup() {

void loop() {
  float v = analogRead(TEMP_PIN) * SUPPLY / 1024;
  float c = (v - .5) * 100;
  Serial.print(" ");
  Serial.println(analogRead(SUN_PIN), DEC);
Note that I've hardwired the supply voltage, as it's needed for the conversion. The Celsius temperature is obtained directly from the measured voltage (10mV per degree, and offset by .5V to allow negative temperatures).

The Arc code

The Arc server code (download) is straightforward. It leverages my earlier Arc example server code (download).

A background thread fetches the time / sun data lines from the serial port and writes the data to a file. It converts Celsius to Fahrenheit and limits the updates to one per minute:

(def logdata ()
  (w/appendfile outf "/tmp/data"
    (w/stdout outf
      (w/infile serial "/dev/ttyUSB0"
        (let oldtimestamp nil
          (while 1
            (with ((degc sun) (tokens (readline serial))
                  timestamp (gettimestamp))
              (when (isnt timestamp oldtimestamp)
                (let degf (+ 32 (* 9. (/ (coerce degc 'num) 5.)))
               (prn timestamp " " (num degf 2) " " sun)
                  (= oldtimestamp timestamp))))))))))

(new-bgthread 'logdata (fn () (logdata)) 0)
The simple web page is generated in Arc, but the graph itself is generated by gnuplot. Since this is currently a static page, it's a bit of overkill to use the Arc functions to generate the page.
(defop temperature req
  (system "gnuplot < gnuplotcmd")
    (tag h1 (prn "Temperature and light: Arc + Arduino + ARM"))
    (gentag img src "/graph.png")
    (tag br)
    (prn "This web page is being served by the") (link "Arc language" "") (pr " web server running on a ")
    (link "SheevaPlug" "")
    (pr " plug computer.") (pr "  For more details see ")
    (link "" "")
    (pr ".")
    (link "View source code for this page" "/source-t")))

Arc vs. Python

I implemented a similar Arduino graph server in Python a few weeks ago. Overall, both Python and Arc make it easy to set up a simple web server. Comparing the Python code with the Arc code reveals Arc's lack of libraries.

The first problem with Arc is it doesn't have a serial library, so I can't configure the baud rate and parameters on my serial port from Arc. I'm just assuming it's set up right, and that's working but not very robust.

The second problem I encountered was creating the timestamps on the data. The latest version of Arc has a timedate function to generate a timestamp but unfortunately that's only in GMT. I tried writing a routine to convert the time to the local timezone, but that got rather annoying. In addition, Arc doesn't have any printf-like formatting, so I had to make my own formatting routine to generate zero-padded strings of the form "12:05". I rapidly decided that writing timezone functions wasn't what I wanted to do, and ended up just using the Unix date command. (This confirm's "kens' law": Any sufficiently complicated Arc application requries the use of 'system to get things done.)

; return a timestamp of the form 2009-08-20 19:22
(= tz "America/Los_Angeles")
(def gettimestamp ()
  (trim (tostring (system (string "TZ=" tz " date +'%m-%d-%Y %H:%M'")))))
I'd have to say that Python is clearly the easier solution overall.

Hardware details

sensors The circuit is pretty trivial. It measures light with a photocell and temperature with a TMP36 temperature sensor (tutorial). (The TMP36 is the three-wire black part that looks like a transistor.) These parts generate voltages that are read by the Arduino's analog-to-digital converters. The temperature sensor's voltage directly gives the Celsius temperature, while the photocell's light measurement is not calibrated to anything.

I had several problems with temperature measurement. First, the Arduino's A/D converter converts relative to its power supply voltage, so the temperature is only as stable as the power supply, and must be manually calibrated. Second, the 10 feet of unshielded phone wire between the Arduino and the temperature sensor introduced a lot of noise. Notice the 10 degree fluctuations on the first day of measurement. I added bypass capacitors after the first day, and the measurements are much smoother. Finally, the temperature sensor heats up a lot when the direct sun hits it in the late afternoon, apparently reaching 140° F. I guess there's a reason why real meteorologists put their temperature sensors in sheltered boxes instead of directly in the sun.

If I were doing this again, I'd probably use a digital temperature sensor such as the One-wire DS18B20; this would avoid the analog calibration and noise issues.

It would be cool to replace the wire between the Arduino and the sensors with Xbee wireless networking. Since I don't have any Xbees, the wire will have to suffice for now.

The photocell voltage is generated from a simple resistor divider. After the second day I changed the resistor from 10K to 1K so the curve wouldn't saturate in the bright sun. (10K worked fine indoors, but outdoors is much brighter.) You can see the break in the green line where I changed resistors.


Arc and the SheevaPlug have been more reliable than I expected; my code has been running for a couple of weeks without problems. I think the SheevaPlug makes a good platform for this sort of project. Using Arc, however, is more of an "experimental curiosity"; I'd recommend a different language unless you really want to use Arc.

I have a few other related postings:

World's smallest Arc server

The frame below is being served was being served by the Arc language web server running on a SheevaPlug plug computer. (I've replaced the live connection with a screenshot, as I needed to use the SheevaPlug for something else.) Screenshot

The SheevaPlug

SheevaPlug in the wall The SheevaPlug is a compact, inexpensive ($99), low-power (5W), ARM-based Linux server built into a wall-plug. It's fairly powerful despite its size, with an Ethernet port, Flash storage, and a 1.2GHz processor. See for more information.

Marvell gave me a SheevaPlug and I thought it would be an interesting platform for experimenting with Arc as a web server. The SheevaPlug seems like a good platform for an always-running web server.

The picture shows the SheevaPlug plugged into the wall, with the Ethernet cable at the bottom. That's not a wall-wart power supply; that's the whole computer.

Since the plug runs Linux, I can just ssh into it and use it like any other Linux server. I installed mzscheme on the SheevaPlug (apt-get install mzscheme), downloaded the Arc language with wget, used iptables to map port 80 to port 8080, and so on. Dynamic DNS provides a DNS name for my SheevaPlug.


Arc is a new dialect of Lisp, designed by Paul Graham and Robert Morris. It's designed to be compact, flexible, and useful for exploratory programming.

The Arc language now (finally) runs on the current version of mzscheme (link). The ARM version of mzscheme 4.1.3 is easily installable on the SheevaPlug, making it straightforward to run Arc on the SheevaPlug.

The demo is a simple program using the Arc web server. It lets you vote on a set of choices and then graphs the result using a simple HTML/CSS graph. The Arc demo also serves its own source code here.

I'm using Arc on the SheevaPlug mostly to see if it works, since I've been using Arc for a while. For normal SheevaPlug use, you'd probably want to use Apache, PHP, Python, or another normal web server configuration; they all have ARM packages that you can simply install.

I'm just assuming this is the world's smallest Arc server; we'll see if anyone runs Arc on something smaller.

If the server is not working, either I shut it down to do something else or it crashed. Send me email and let me know. You can access the server without the iframe here.

I've also used a Python web server on the SheevaPlug, which is a more mainstream way to go than Arc. In that case, I connected an Arduino to the SheevaPlug, so my Arduino could be accessed over the web.

Admittedly, the demo is just a proof-of concept. The next step is to do something non-trivial with the SheevaPlug and Arc.

Secrets of Arduino PWM

Pulse-width modulation (PWM) can be implemented on the Arduino in several ways. This article explains simple PWM techniques, as well as how to use the PWM registers directly for more control over the duty cycle and frequency. This article focuses on the Arduino Diecimila and Duemilanove models, which use the ATmega168 or ATmega328.

If you're unfamiliar with Pulse Width Modulation, see the tutorial. Briefly, a PWM signal is a digital square wave, where the frequency is constant, but that fraction of the time the signal is on (the duty cycle) can be varied between 0 and 100%.
PWM examples
PWM has several uses:

  • Dimming an LED
  • Providing an analog output; if the digital output is filtered, it will provide an analog voltage between 0% and 100% .
  • Generating audio signals.
  • Providing variable speed control for motors.
  • Generating a modulated signal, for example to drive an infrared LED for a remote control.

Simple Pulse Width Modulation with analogWrite

The Arduino's programming language makes PWM easy to use; simply call analogWrite(pin, dutyCycle), where dutyCycle is a value from 0 to 255, and pin is one of the PWM pins (3, 5, 6, 9, 10, or 11). The analogWrite function provides a simple interface to the hardware PWM, but doesn't provide any control over frequency. (Note that despite the function name, the output is a digital signal.)

Probably 99% of the readers can stop here, and just use analogWrite, but there are other options that provide more flexibility.

Bit-banging Pulse Width Modulation

You can "manually" implement PWM on any pin by repeatedly turning the pin on and off for the desired times. e.g.
void setup()
  pinMode(13, OUTPUT);

void loop()
  digitalWrite(13, HIGH);
  delayMicroseconds(100); // Approximately 10% duty cycle @ 1KHz
  digitalWrite(13, LOW);
This technique has the advantage that it can use any digital output pin. In addition, you have full control the duty cycle and frequency. One major disadvantage is that any interrupts will affect the timing, which can cause considerable jitter unless you disable interrupts. A second disadvantage is you can't leave the output running while the processor does something else. Finally, it's difficult to determine the appropriate constants for a particular duty cycle and frequency unless you either carefully count cycles, or tweak the values while watching an oscilloscope.

Using the ATmega PWM registers directly

The ATmega168P/328P chip has three PWM timers, controlling 6 PWM outputs. By manipulating the chip's timer registers directly, you can obtain more control than the analogWrite function provides.

The AVR ATmega328P datasheet provides a detailed description of the PWM timers, but the datasheet can be difficult to understand, due to the many different control and output modes of the timers. The following attempts to clarify the use of the timers.

The ATmega328P has three timers known as Timer 0, Timer 1, and Timer 2. Each timer has two output compare registers that control the PWM width for the timer's two outputs: when the timer reaches the compare register value, the corresponding output is toggled. The two outputs for each timer will normally have the same frequency, but can have different duty cycles (depending on the respective output compare register).

Each of the timers has a prescaler that generates the timer clock by dividing the system clock by a prescale factor such as 1, 8, 64, 256, or 1024. The Arduino has a system clock of 16MHz and the timer clock frequency will be the system clock frequency divided by the prescale factor. Note that Timer 2 has a different set of prescale values from the other timers.

The timers are complicated by several different modes. The main PWM modes are "Fast PWM" and "Phase-correct PWM", which will be described below. The timer can either run from 0 to 255, or from 0 to a fixed value. (The 16-bit Timer 1 has additional modes to supports timer values up to 16 bits.) Each output can also be inverted.

The timers can also generate interrupts on overflow and/or match against either output compare register, but that's beyond the scope of this article.

Timer Registers

Several registers are used to control each timer. The Timer/Counter Control Registers TCCRnA and TCCRnB hold the main control bits for the timer. (Note that TCCRnA and TCCRnB do not correspond to the outputs A and B.) These registers hold several groups of bits:
  • Waveform Generation Mode bits (WGM): these control the overall mode of the timer. (These bits are split between TCCRnA and TCCRnB.)
  • Clock Select bits (CS): these control the clock prescaler
  • Compare Match Output A Mode bits (COMnA): these enable/disable/invert output A
  • Compare Match Output B Mode bits (COMnB): these enable/disable/invert output B
The Output Compare Registers OCRnA and OCRnB set the levels at which outputs A and B will be affected. When the timer value matches the register value, the corresponding output will be modified as specified by the mode.

The bits are slightly different for each timer, so consult the datasheet for details. Timer 1 is a 16-bit timer and has additional modes. Timer 2 has different prescaler values.

Fast PWM

In the simplest PWM mode, the timer repeatedly counts from 0 to 255. The output turns on when the timer is at 0, and turns off when the timer matches the output compare register. The higher the value in the output compare register, the higher the duty cycle. This mode is known as Fast PWM Mode.
The following diagram shows the outputs for two particular values of OCRnA and OCRnB. Note that both outputs have the same frequency, matching the frequency of a complete timer cycle.
Fast PWM Mode
The following code fragment sets up fast PWM on pins 3 and 11 (Timer 2). To summarize the register settings, setting the waveform generation mode bits WGM to 011 selects fast PWM. Setting the COM2A bits and COM2B bits to 10 provides non-inverted PWM for outputs A and B. Setting the CS bits to 100 sets the prescaler to divide the clock by 64. (Since the bits are different for the different timers, consult the datasheet for the right values.) The output compare registers are arbitrarily set to 180 and 50 to control the PWM duty cycle of outputs A and B. (Of course, you can modify the registers directly instead of using pinMode, but you do need to set the pins to output.)
  pinMode(3, OUTPUT);
  pinMode(11, OUTPUT);
  TCCR2A = _BV(COM2A1) | _BV(COM2B1) | _BV(WGM21) | _BV(WGM20);
  TCCR2B = _BV(CS22);
  OCR2A = 180;
  OCR2B = 50;
On the Arduino Duemilanove, these values yield:
  • Output A frequency: 16 MHz / 64 / 256 = 976.5625Hz
  • Output A duty cycle: (180+1) / 256 = 70.7%
  • Output B frequency: 16 MHz / 64 / 256 = 976.5625Hz
  • Output B duty cycle: (50+1) / 256 = 19.9%

The output frequency is the 16MHz system clock frequency, divided by the prescaler value (64), divided by the 256 cycles it takes for the timer to wrap around. Note that fast PWM holds the output high one cycle longer than the compare register value.

Phase-Correct PWM

The second PWM mode is called phase-correct PWM. In this mode, the timer counts from 0 to 255 and then back down to 0. The output turns off as the timer hits the output compare register value on the way up, and turns back on as the timer hits the output compare register value on the way down. The result is a more symmetrical output. The output frequency will be approximately half of the value for fast PWM mode, because the timer runs both up and down.
Phase-Correct PWM
The following code fragment sets up phase-correct PWM on pins 3 and 11 (Timer 2). The waveform generation mode bits WGM are set to to 001 for phase-correct PWM. The other bits are the same as for fast PWM.
  pinMode(3, OUTPUT);
  pinMode(11, OUTPUT);
  TCCR2A = _BV(COM2A1) | _BV(COM2B1) | _BV(WGM20);
  TCCR2B = _BV(CS22);
  OCR2A = 180;
  OCR2B = 50;
On the Arduino Duemilanove, these values yield:
  • Output A frequency: 16 MHz / 64 / 255 / 2 = 490.196Hz
  • Output A duty cycle: 180 / 255 = 70.6%
  • Output B frequency: 16 MHz / 64 / 255 / 2 = 490.196Hz
  • Output B duty cycle: 50 / 255 = 19.6%

Phase-correct PWM divides the frequency by two compared to fast PWM, because the timer goes both up and down. Somewhat surprisingly, the frequency is divided by 255 instead of 256, and the duty cycle calculations do not add one as for fast PWM. See the explanation below under "Off-by-one".

Varying the timer top limit: fast PWM

Both fast PWM and phase correct PWM have an additional mode that gives control over the output frequency. In this mode, the timer counts from 0 to OCRA (the value of output compare register A), rather than from 0 to 255. This gives much more control over the output frequency than the previous modes. (For even more frequency control, use the 16-bit Timer 1.)

Note that in this mode, only output B can be used for PWM; OCRA cannot be used both as the top value and the PWM compare value. However, there is a special-case mode "Toggle OCnA on Compare Match" that will toggle output A at the end of each cycle, generating a fixed 50% duty cycle and half frequency in this case. The examples will use this mode.

In the following diagram, the timer resets when it matches OCRnA, yielding a faster output frequency for OCnB than in the previous diagrams. Note how OCnA toggles once for each timer reset.
Fast PWM Mode with OCRA top
The following code fragment sets up fast PWM on pins 3 and 11 (Timer 2), using OCR2A as the top value for the timer. The waveform generation mode bits WGM are set to to 111 for fast PWM with OCRA controlling the top limit. The OCR2A top limit is arbitrarily set to 180, and the OCR2B compare register is arbitrarily set to 50. OCR2A's mode is set to "Toggle on Compare Match" by setting the COM2A bits to 01.

  pinMode(3, OUTPUT);
  pinMode(11, OUTPUT);
  TCCR2A = _BV(COM2A0) | _BV(COM2B1) | _BV(WGM21) | _BV(WGM20);
  TCCR2B = _BV(WGM22) | _BV(CS22);
  OCR2A = 180;
  OCR2B = 50;
On the Arduino Duemilanove, these values yield:
  • Output A frequency: 16 MHz / 64 / (180+1) / 2 = 690.6Hz
  • Output A duty cycle: 50%
  • Output B frequency: 16 MHz / 64 / (180+1) = 1381.2Hz
  • Output B duty cycle: (50+1) / (180+1) = 28.2%
Note that in this example, the timer goes from 0 to 180, which takes 181 clock cycles, so the output frequency is divided by 181. Output A has half the frequency of Output B because the Toggle on Compare Match mode toggles Output A once each complete timer cycle.

Varying the timer top limit: phase-correct PWM

Similarly, the timer can be configured in phase-correct PWM mode to reset when it reaches OCRnA.
Phase-Correct PWM with OCRA top
The following code fragment sets up phase-correct PWM on pins 3 and 11 (Timer 2), using OCR2A as the top value for the timer. The waveform generation mode bits WGM are set to to 101 for phase-correct PWM with OCRA controlling the top limit. The OCR2A top limit is arbitrarily set to 180, and the OCR2B compare register is arbitrarily set to 50. OCR2A's mode is set to "Toggle on Compare Match" by setting the COM2A bits to 01.
  pinMode(3, OUTPUT);
  pinMode(11, OUTPUT);
  TCCR2A = _BV(COM2A0) | _BV(COM2B1) | _BV(WGM20);
  TCCR2B = _BV(WGM22) | _BV(CS22);
  OCR2A = 180;
  OCR2B = 50;
On the Arduino Duemilanove, these values yield:
  • Output A frequency: 16 MHz / 64 / 180 / 2 / 2 = 347.2Hz
  • Output A duty cycle: 50%
  • Output B frequency: 16 MHz / 64 / 180 / 2 = 694.4Hz
  • Output B duty cycle: 50 / 180 = 27.8%
Note that in this example, the timer goes from 0 to 180 and back to 0, which takes 360 clock cycles. Thus, everything is divided by 180 or 360, unlike the fast PWM case, which divided everything by 181; see below for details.


You may have noticed that fast PWM and phase-correct PWM seem to be off-by-one with respect to each other, dividing by 256 versus 255 and adding one in various places. The documentation is a bit opaque here, so I'll explain in a bit of detail.

Suppose the timer is set to fast PWM mode and is set to count up to an OCRnA value of 3. The timer will take on the values 012301230123... Note that there are 4 clock cycles in each timer cycle. Thus, the frequency will be divided by 4, not 3. The duty cycle will be a multiple of 25%, since the output can be high for 0, 1, 2, 3, or 4 cycles out of the four. Likewise, if the timer counts up to 255, there will be 256 clock cycles in each timer cycle, and the duty cycle will be a multiple of 1/256. To summarize, fast PWM divides by N+1 where N is the maximum timer value (either OCRnA or 255).

Now consider phase-correct PWM mode with the timer counting up to an OCRnA value of 3. The timer values will be 012321012321... There are 6 clock cycles in each timer cycle (012321). Thus the frequency will be divided by 6. The duty cycle will be a multiple of 33%, since the output can be high for 0, 2, 4, or 6 of the 6 cycles. Likewise, if the timer counts up to 255 and back down, there will be 510 clock cycles in each timer cycle, and the duty cycle will be a multiple of 1/255. To summarize, phase-correct PWM divides by 2N, where N is the maximum timer value.

The second important timing difference is that fast PWM holds the output high for one cycle longer than the output compare register value. The motivation for this is that for fast PWM counting to 255, the duty cycle can be from 0 to 256 cycles, but the output compare register can only hold a value from 0 to 255. What happens to the missing value? The fast PWM mode keeps the output high for N+1 cycles when the output compare register is set to N so an output compare register value of 255 is 100% duty cycle, but an output compare register value of 0 is not 0% duty cycle but 1/256 duty cycle. This is unlike phase-correct PWM, where a register value of 255 is 100% duty cycle and a value of 0 is a 0% duty cycle.

Timers and the Arduino

The Arduino supports PWM on a subset of its output pins. It may not be immediately obvious which timer controls which output, but the following table will clarify the situation. It gives for each timer output the output pin on the Arduino (i.e. the silkscreened label on the board), the pin on the ATmega chip, and the name and bit of the output port. For instance Timer 0 output OC0A is connected to the Arduino output pin 6; it uses chip pin 12 which is also known as PD6.
Timer outputArduino outputChip pinPin name

The Arduino performs some initialization of the timers. The Arduino initializes the prescaler on all three timers to divide the clock by 64. Timer 0 is initialized to Fast PWM, while Timer 1 and Timer 2 is initialized to Phase Correct PWM. See the Arduino source file wiring.c for details.

The Arduino uses Timer 0 internally for the millis() and delay() functions, so be warned that changing the frequency of this timer will cause those functions to be erroneous. Using the PWM outputs is safe if you don't change the frequency, though.

The analogWrite(pin, duty_cycle) function sets the appropriate pin to PWM and sets the appropriate output compare register to duty_cycle (with the special case for duty cycle of 0 on Timer 0). The digitalWrite() function turns off PWM output if called on a timer pin. The relevant code is wiring_analog.c and wiring_digital.c.

If you use analogWrite(5, 0) you get a duty cycle of 0%, even though pin 5's timer (Timer 0) is using fast PWM. How can this be, when a fast PWM value of 0 yields a duty cycle of 1/256 as explained above? The answer is that analogWrite "cheats"; it has special-case code to explicitly turn off the pin when called on Timer 0 with a duty cycle of 0. As a consequency, the duty cycle of 1/256 is unavailable when you use analogWrite on Timer0, and there is a jump in the actual duty cycle between values of 0 and 1.

Some other Arduino models use dfferent AVR processors with similar timers. The Arduino Mega uses the ATmega1280 (datasheet), which has four 16-bit timers with 3 outputs each and two 8-bit timers with 2 outputs each. Only 14 of the PWM outputs are supported by the Arduino Wiring library, however. Some older Arduino models use the ATmega8 (datasheet), which has three timers but only 3 PWM outputs: Timer 0 has no PWM, Timer 1 is 16 bits and has two PWM outputs, and Timer 2 is 8 bits and has one PWM output.


It can be tricky to get the PWM outputs to work. Some tips:
  • You need to both enable the pin for output and enable the PWM mode on the pin in order to get any output. I.e. you need to do pinMode() and set the COM bits.
  • The different timers use the control bits and prescaler differently; check the documentation for the appropriate timer.
  • Some combinations of bits that you might expect to work are reserved, which means if you try to use them, they won't work. For example, toggle mode doesn't work with fast PWM to 255, or with output B.
  • Make sure the bits are set the way you think. Bit operations can be tricky, so print out the register values and make sure they are what you expect.
  • Make sure you're using the right output pins. See the table above.
  • You'll probably want a decoupling capacitor to avoid spikes on the output.
An oscilloscope is very handy for debugging PWM if you have access to one. If you don't have one, I recommend using your sound card and a program such as xoscope.


I hope this article helps explain the PWM modes of the Arduino. I found the documentation of the different modes somewhat opaque, and the off-by-one issues unexplained. Please let me know if you encounter any errors.

Arduino + SheevaPlug = Cool Hardware Platform

I recently started using an Arduino in combination with a SheevaPlug as a convenient platform for hardware hacking. The two work together well, with complementary strengths. This describes my experience.

The Arduino

The Arduino is a popular microcontroller platform consisting of a small, inexpensive board and an easy-to-use C-based development environment. The board has multiple digital I/O pins and analog inputs, and a USB serial port connection for programming and communication. The Arduino can be interfaced with a wide variety of hardware. (The Arduino is kind of like a Microchip PIC development board, but uses the ATmega328 AVR microcontroller.) The Arduino is about the size of a credit card (but thick). It has connectors on top which will receive a daughter board (inexplicably called a "shield"), or jumper wires.
The Arduino board
The Arduino is very easy to program and use. The Arduino is programmed using the Wiring language, which is an extension of C with many libraries to interface with various types of hardware. The Arduino IDE provides an editor and compiler for a "sketch" (which is the name for an Arduino program), and downloads the code to the Arduino board over USB. After seeing the Arduino hype at the Maker Faire, I got a Adafruit Arduino Starter Pack, and I've been happy with it. More information about the Arduino is at; the playground is a good place to start.

You can do a lot with an Arduino, but it has some limitations. There's only 30KB of memory for programming (yes, kilobytes). Many things you might want to connect (Ethernet, SD card, USB hard drive or flash) require additional hardware, and even then are pretty limited.

The SheevaPlug

The Marvell SheevaPlug is a very small Linux server with a 1.2GHz 88F6281 ARM processor and 512MB of RAM and 512MB of flash storage. The SheevaPlug has a Gigabit Ethernet connection, a SD card slot, and a serial USB console connection. This all fits into a box the size of three cassette tapes that sells for $99 and uses about 5 watts of power.
The SheevaPlug
The SheevaPlug is designed to be plugged directly into the wall as a wall-wart, but the plug snaps out and can be replaced by a power coard. I'm currently using the "desktop" configuration with a power cord, since it's easier to connect things. Marvell recently gave me a SheevaPlug, and I've been enjoying it a lot.

Because the SheevaPlug is a full Linux system, the standard Linux software packages are available and easily installed with apt-get (full list). You can run X-windows (of course you need to VPN in). You can run Apache and PHP on the SheevaPlug and use it as a web server. The GCC toolchain runs on it. You can program it in C++, Python, Scheme, or whatever.

The main information source on the SheevaPlug is Another useful site is, a wiki about the SheevaPlug which is actually served by a SheevaPlug.

There are a few limitations of the SheevaPlug. The SheevaPlug is headless, so you need to access it via ssh or VNC. You can't run x86 binaries. The SheevaPlug's performance is roughly at the Pentium III level (benchmark), so it won't replace a state-of the-art server. Also, floating point performance is poor since there's no hardware FPU. The Linux install has some annoying minor issues with apt-get, DNS, NTP, and passwd (fixes). The SheevaPlug isn't really suited for low-level electronics projects because it doesn't have any accessible general-purpose I/O pins. (The CPU has 50 digital GPIO pins, but the SheevaPlug doesn't expose them.)

Arduino + SheevaPlug

The SheevaPlug and Arduino make a great pair together for hardware hacking projects since the SheevaPlug can easily provide the web server, computing, and storage functionality that the Arduino lacks, while the Arduino makes it easy to interface to hardware circuits. Moreover, the two are easily connected with a USB cable, and then they can communicate through simple serial communication (modulo kernel details below).
SheevaPlug with Arduino
The above picture shows the SheevaPlug and Arduino connected together by a USB cable. The SheevaPlug also has an Ethernet connection (green) and power (on the right). The Arduino has some jumpers to a breadboard circuit (described below).

A sample project: illumination recording

For an initial project, I decided to collect data on the illumnination in my room. (Temperature would have been more useful, but I didn't have a temperature sensor handy.) I implemented a simple Arduino sketch (i.e. program) to read illumination from a photocell and write it to the serial port. The SheevaPlug collects the data, generates a graph using gnuplot, and provides the data as a web page on the Internet. (Sorry, no URL to access it since I'm not currently running this.)

The screenshot below shows the web page that the SheevaPlug serves showing illumination across a few days. There are a lot of spikes due to turning lights on and off, but you can see where the sun rises, light increases through the day, peaking shortly before sunset. This web page illustrates the ability of the Arduino to measure analog data combined with the SheevaPlug's ability to collect data, plot it, and serve web pages.
SheevaPlug screenshot of browser

The Arduino code

Hooking up the Arduino to measure illumination was trivial; I used a 10K resistor and a photocell connected to an analog input pin. (I made the following schematic with Eagle.)
schematic of photocell connection

The Arduino sketch below (download) simply reads the analog input once a second and writes the value to the serial port as a decimal number. (One annoyance is the Arduino IDE won't run on the ARM processor, so I have to connect the Arduino to a separate computer to download the sketch.)

int ainPin = 2;     // select the analog input pin

void setup() {

void loop() {
  Serial.println(analogRead(ainPin), DEC);    // read the value from the sensor and write to serial

The Python code

The Python code that runs on the SheevaPlug (download) is straightforward. I used a simple Python web server to provide the graph page and some static files. It uses gnuplot to generate the graph. Some highlights of the code:

The pyserial library (apt-get install python-serial) provides simple access to the serial port. One thread reads lines from the serial port and dumps them to the data file along with a timestamp. Only one sample per minute is kept, and the others discarded. The Arduino may appear as a different device, e.g. /dev/ttyUSB0. If no such device appears when you plug the Arduino into the SheevaPlug, you probably need the kernel update below.

class Arduino(threading.Thread):
  def run(self):
    f = open('/tmp/data', 'a')
    # Port may vary from /dev/ttyUSB1
    self.ser = serial.Serial('/dev/ttyUSB1', 9600, timeout=10)
    old_timestamp = None
    while 1:
      data = self.ser.readline().strip()
      if data:
        timestamp = time.strftime("%m/%d/%Y %H:%M", time.localtime())
        if timestamp != old_timestamp:
          # Only log once per minute
          print >>f, timestamp, data.strip()
          old_timestamp = timestamp

The other part of the Python code is an HTTP server using SimpleHTTPServer; a /graph URL runs gnuplot on the data file to generate a png image file. The returned web page loads the image, which is served by the Python server as a static file. (Since the SheevaPlug supports Apache, PHP, and mysql, I could have used them, but it seems like overkill for a simple demo.)

 def graph(self):
    g = os.popen('gnuplot', 'w')
    print >>g, GNUPLOT_CMD
    self.send_header('Content-type', 'text/html')

Kernel with FTDI Serial support

The one stumbling block I encountered when using the SheevaPlug and Arduino together is the SheevaPlug kernel doesn't include FTDI support, which is necessary to access the Arduino via USB. An upgraded kernel is available, and much to my surprise it installed with no problems.

Where from here?

So far, I've found the Arduino and SheevaPlug to work together well. I have various plans to use the Arduino + SheevaPlug for home monitoring and control, solar panel monitoring, IR remote control, and other random projects. If there's interest, maybe I'll write about them in the future.