A Maker Wedding

Initially I wasn’t sure how much our wedding was truly going to represent my fiancée and I, after all, we wanted our family and friends to enjoy themselves and feel included — as with any large event there are a lot of expectations to manage. After deciding to craft my own Edison-style light fixtures for our reception I realized that the occasion was, in addition to a celebration of our life-long commitment to each other, an opportunity for us to showcase our creativity and perhaps introduce some of our family and friends to aspects of ourselves they may not have known existed.

In retrospect we probably took on too much, but it allowed us to feel the occasion was a true reflection of ourselves — for me this meant soldering, stripping, crimping, twisting, programming and no small amount of brow furrowing. None of these projects could’ve come together without the help of my wonderful wife Ester, who not only said yes, but also collaborated throughout and trusted me to deliver on some very important aspects of our big day. In addition, a big thanks to my dear old Dad who took time to help me with the lengthy task of wiring the Edison fixtures and to the friends and family who helped us setup and teardown these, and other installations.

Animated Arduino LED matrix lounge table top

Vinyl “flexi” record wedding invitations

XBee remote relay as photobooth RF camera trigger

Bachelor party wireless Arduino accelerometer Stab-O-Meter

JQuery Animated Wedding Website


Various puppet arms available at Obscura Antiques & Oddities, New York

Maker Wedding: Animated Arduino LED matrix lounge table top


Finally got around to making an LED table top, as it turns out — for my wedding reception. We decided to have a lounge area and an LED coffee table seemed like the perfect centerpiece for it. I decided instead of making a full table that I would make a table top that fit onto an existing ottoman. I affixed the LED strips to a plywood board which had a 2″ raised frame with aluminium duct tape, to help with brightness.

Arranging the LED’s in a proper matrix turned out to be quite a job as the strips I used came pre-wired and there isn’t all that much length between LED’s on the strip. I ended up having to cut and re-splice the connection leads for each row of the 7 x 7 matrix, after that the construction went quickly. You could get around this by using a more modular LED strip solution, I initially had ShiftBrites slated for this project, but I made something else with them and when I got around to this table top there were much less expensive options available.

I created an outer frame with a bevel to support a glass top. Initially I went with plexiglass but it would bow in the middle with anything of weight on the table, I didn’t want to add supports as this would disrupt the light diffusion, so I opted for a piece of tempered glass (actually intended for table tops to boot).

Adding adhesive obscuring film to the glass didn’t have the diffusion effect I’d hoped for so I sandwiched a sheet of white tracing paper between the tempered glass and a similarly sized piece of plexiglass and this gave the soft white diffused look I wanted.

The issue of programming the animations took a little longer. I wanted to use the disco(esc) animations available from the fine folks responsible for the 1E Disco Dance Floor — which I also used in my ohDisco! app for iPad. These animations are 32 x 16 and can have hundreds of frames — too much to load completely into the Arduino’s memory. But I didn’t want to have to deal with reducing the animations to 7 x 7, or reducing their total frame count, as this would affect the quality and overall impression. Instead I opted to add an SD card reader to the setup which stores the animations. The 7 x 7 section of each frame is loaded on-demand from each animation file and displayed on the table, with this setup the Arduino has no memory problems whatsoever and with a little more code it could index and play animations from the SD card without the need for code changes.

Worth noting is that the SdFat library used to interface with the Seeedstudio SD Card Shield wouldn’t run reliably (or at all sometimes) on an ATmega128 so be sure to use a more powerful Arduino running an ATmega328.

Parts

Arduino Sketch


const int chipSelect = 10;

#include <SdFat.h>
#include "SPI.h"
#include "Adafruit_WS2801.h"

uint8_t dataPin  = 2;
uint8_t clockPin = 3;   

SdFat sd;
SdFile myFile;

int rows = 32;
int cols = 16;

long framesize = rows*cols*3;
long rowsize = cols*3;

int ledrows = 6;
int ledcols = 6;

int rep = 0;
long reps = 5;

int brightness = 15;
int delaytime = 40;

char* files[]={
  "pulsar.ddf",
  "snake.ddf",
  "inter3.ddf",
  "inter4.ddf",
  "inter5.ddf",
  "rings.ddf",
  "rings2.ddf",
  "rings3.ddf",
  "matrix.ddf"
  };

int fileCount = 9;

char* file;
int frame = 0;
int frames;

Adafruit_WS2801 strip = Adafruit_WS2801(50, dataPin, clockPin);

// strip to matrix addressing array
byte addressMatrix[7][7] = {
  1,2,3,4,5,6,7,
  14,13,12,11,10,9,8,
  15,16,17,18,19,20,21,
  28,27,26,25,24,23,22,
  29,30,31,32,33,34,35,
  42,41,40,39,38,37,36,
  43,44,45,46,47,48,49
};

void setup() {
  Serial.begin(9600);
  randomSeed(analogRead(0));
  delay(400);  // catch Due reset problem
  if (!sd.begin(chipSelect, SPI_FULL_SPEED))
    sd.initErrorHalt();

  file = files[ random(fileCount) ];

  strip.begin();
  strip.setPixelColor(0, 0, 0, 0);
  strip.show();
}

void loop() {
  if (!myFile.open(file, O_READ)) {
    sd.errorHalt("failed");
    rep = reps + 1;
    return;
  }

  // seek to next frame
  if(myFile.fileSize() < ((frame*framesize)+1))
  {
    myFile.close(); 

    frame = 0;
    rep = rep + 1;

    if(rep > reps)
    {
      rep = 0;
      file = files[ random(fileCount) ];
      Serial.println(file);
    }

    return;
  }
  else
  {
    myFile.seekSet(frame*framesize);
  }

  // adjust reps for number of frames
  if(frame == 0)
  {
    frames = myFile.fileSize()/framesize;
    reps = 750/frames;
    Serial.println(frames,DEC);
    Serial.println(reps,DEC);
  }

  int data;

  int column = 0;
  int row = 0;

  while (row <= ledrows)
  {
    while (column <= ledcols)
    {
      data = myFile.read();
      // read red
      int r = map(data,0,255,0,255);
      // read green
      data = myFile.read();
      int g = map(data,0,255,0,255);
      // read blue
      data = myFile.read();
      int b = map(data,0,255,0,255);

      // set pixel address
      byte address = addressMatrix[row][column];

      // set pixel color
      strip.setPixelColor(address, map(r,0,255,0,brightness), map(g,0,255,0,brightness), map(b,0,255,0,brightness));

      // next column
      column = column + 1;
    }

    // reset column count
    column = 0;

    // increment row
    row = row + 1;

    // skip extra pixels
    myFile.seekSet((frame*framesize)+(row*rowsize));
  }

  // turn off first pixel (7x7 matrix, 1 unused pixel)
  strip.setPixelColor(0, 0, 0, 0);

  // send current frame to strip
  strip.show();

  // close the file
  myFile.close();

  // increment frame
  frame = frame + 1;

  // rest
  delay(delaytime);
}

Maker Wedding: Bachelor party wireless accelerometer Stab-O-Meter

Since I had disassembled the Wine-O-Meter I’d made for a friend’s bachelor party I needed to come up with something else for my own, I wanted to do an updated strongman competition. I decided to put together a wireless accelerometer to hopefully measure the speed and impact of various activities such as swinging a baseball bat, a sledgehammer, a hatchet, a tennis racket — get the idea? Sort of like the measurement tools used on shows like MythBusters or Deadliest Warrior. Along the lines of the Wine-O-Meter I dubbed the project the Stab-O-Meter as measuring arm movements reminded me of one of my favourite Futurama characters, Roberto.


My plan was to use an Arduino to read an accelerometer and use a pair of XBees to wireless relay the information to a laptop. The laptop would be running a Processing sketch to handle the high score display, reset and current readings. It took a little bit to find the right Arudino code to read the LIS331 Triple Axis Accelerometer I’d selected but it worked well once I found it. I don’t know a whole heck of a lot about accelerometers, but this one measures g-forces on three axis, x, y and z. After some trial and error I decided to add all positive g-force readings together and then add all negative g-force readings together. If the positive total was higher I used that as the current amalgamated reading otherwise I used the absolute sum of the negative values. Comment if you’re aware of a better way to translate x, y, z g-forces into a single number representing the speed of the motion (see Hank’s comment below).

Hank Cowdog

A neg X Acc means acc along the negative X axis. The magnitude of the acc is the important measurement, so a better approach would be to sum the squares of each X,Y,Z component and then take the square root (as per the Pythagorean Theorem). This computes the magnitude of the Acc regardless of the direction (or orientation of the accelerometer chips).

result = sqrt(xAcc*xAcc + yAcc*yAcc + zAcc*zAcc);


The Arduino sent the single number amalgamated reading in realtime (or as close as possible) via it’s serial connection to a XBee which in turn wirelessly relayed the serial data to a laptop running a processing sketch to read and deal with the data. The Processing sketch displayed a realtime reading bar on the right, the highest reading yet recorded in large numbers in the center and a RESET button to clear the current highest reading. With this system each contestant could reset the high score using the RESET button or the spacebar and the proceed to swing a bat or stab a tree or whatnot to find they’re personal best, which was then ranked against other’s scores on a white board.

This part worked great, however in impact scenarios (actually hitting something) it was too easy to max out the sensor, which has a max of 24g, so we restricted our games to non-impact swings. I had added hand wrap to the sensor case in order to secure it to the implement of choice, however I quickly realized that it also needed a non-slip surface for grip, I epoxied some rubber salvaged from a guitar effect pedal. Even with the hand wrap and the rubber footing the first full-force swing with a baseball bat sent the sensor soaring into a neighbouring house — duct tape provided the necessary upgrade in grip, but downgrade in polish.

The video below is, aside from my Roberto impression, an early test using a preliminary Processing sketch and no cases for the components. When I get a chance I’ll record a video of the finished setup, perhaps as I demolish my garage this weekend. Yes, it’s an odd video, but that’s what YouTube is for, right?

Parts

Arduino Sketch

// 3-axis Accelerometer
// Sparkfun Electronics Triple Axis Accelerometer Breakout - LIS331
// Arduino UNO

/* Wiring:
    UNO LIS331

    3.3V VCC
    GND GND
    10 CS
    11 SDA/SDI
    12 SA0/SDO
    13 SCL/SPC
    */

#include <SPI.h>
#include <stdlib.h>
#include <stdio.h>

#define SS 10 // Serial Select -> CS on LIS331
#define MOSI 11 // MasterOutSlaveIn -> SDI
#define MISO 12 // MasterInSlaveOut -> SDO
#define SCK 13 // Serial Clock -> SPC on LIS331

#define SCALE 0.0007324; // approximate scale factor for full range (+/-24g)
// scale factor: +/-24g = 48G range. 2^16 bits. 48/65536 = 0.0007324

// global acceleration values
double xAcc, yAcc, zAcc;

void setup()
{
  Serial.begin(9600);

  // Configure SPI
  SPI_SETUP();

  // Configure accelerometer
  Accelerometer_Setup();
}

void loop()
{
  readVal(); // get acc values and put into global variables

  int pos = 0;
  int neg = 0;

  if(xAcc > 0)
  {
    pos = pos + xAcc;
  }
  else
  {
    neg = neg + abs(xAcc);
  }

  if(yAcc > 0)
  {
    pos = pos + yAcc;
  }
  else
  {
    neg = neg + abs(yAcc);
  }

  if(zAcc > 0)
  {
    pos = pos + zAcc;
  }
  else
  {
    neg = neg + abs(zAcc);
  }

  int result = neg;

  if(pos > neg)
    result = pos;

  Serial.println(result,1);

   /*
    Serial.print(xAcc, 1);
    Serial.print(",");
    Serial.print(yAcc, 1);
    Serial.print(",");
    Serial.println(zAcc, 1);
  */

  delay(10);
}

// Read the accelerometer data and put values into global variables
void readVal()
{
  byte xAddressByteL = 0x28; // Low Byte of X value (the first data register)
  byte readBit = B10000000; // bit 0 (MSB) HIGH means read register
  byte incrementBit = B01000000; // bit 1 HIGH means keep incrementing registers
  // this allows us to keep reading the data registers by pushing an empty byte
  byte dataByte = xAddressByteL | readBit | incrementBit;
  byte b0 = 0x0; // an empty byte, to increment to subsequent registers

  digitalWrite(SS, LOW); // SS must be LOW to communicate
  delay(1);
  SPI.transfer(dataByte); // request a read, starting at X low byte
  byte xL = SPI.transfer(b0); // get the low byte of X data
  byte xH = SPI.transfer(b0); // get the high byte of X data
  byte yL = SPI.transfer(b0); // get the low byte of Y data
  byte yH = SPI.transfer(b0); // get the high byte of Y data
  byte zL = SPI.transfer(b0); // get the low byte of Z data
  byte zH = SPI.transfer(b0); // get the high byte of Z data
  delay(1);
  digitalWrite(SS, HIGH);

  // shift the high byte left 8 bits and merge the high and low
  int xVal = (xL | (xH <<8));
  int yVal = (yL | (yH <<8));
  int zVal = (zL | (zH <<8));

  // scale the values into G's
  xAcc = xVal * SCALE;
  yAcc = yVal * SCALE;
  zAcc = zVal * SCALE;
}

void SPI_SETUP()
{
  pinMode(SS, OUTPUT);

  // wake up the SPI bus
  SPI.begin();

  // This device reads MSB first:
  SPI.setBitOrder(MSBFIRST);

  /*
  SPI.setDataMode()
  Mode    Clock Polarity (CPOL) Clock Phase (CPHA)
  SPI_MODE0    0    0
  SPI_MODE1    0    1
  SPI_MODE2    1    0
  SPI_MODE3    1    1
  */
  SPI.setDataMode(SPI_MODE0);

  /*
  SPI.setClockDivider()
  sets SPI clock to a fraction of the system clock
  Arduino UNO system clock = 16 MHz
  Mode SPI Clock
  SPI_CLOCK_DIV2 8 MHz
  SPI_CLOCK_DIV4 4 MHz
  SPI_CLOCK_DIV8 2 MHz
  SPI_CLOCK_DIV16 1 MHz
  SPI_CLOCK_DIV32 500 Hz
  SPI_CLOCK_DIV64 250 Hz
  SPI_CLOCK_DIV128 125 Hz
  */

  SPI.setClockDivider(SPI_CLOCK_DIV16); // SPI clock 1000Hz
}

void Accelerometer_Setup()
{
  // Set up the accelerometer
  // write to Control register 1: address 20h
  byte addressByte = 0x20;
  /* Bits:
  PM2 PM1 PM0 DR1 DR0 Zen Yen Xen
  PM2PM1PM0: Power mode (001 = Normal Mode)
  DR1DR0: Data rate (00=50Hz, 01=100Hz, 10=400Hz, 11=1000Hz)
  Zen, Yen, Xen: Z enable, Y enable, X enable
  */
  byte ctrlRegByte = 0x37; // 00111111 : normal mode, 1000Hz, xyz enabled

  // Send the data for Control Register 1
  digitalWrite(SS, LOW);
  delay(1);
  SPI.transfer(addressByte);
  SPI.transfer(ctrlRegByte);
  delay(1);
  digitalWrite(SS, HIGH);

  delay(100);

  // write to Control Register 2: address 21h
  addressByte = 0x21;
  // This register configures high pass filter
  ctrlRegByte = 0x00; // High pass filter off

  // Send the data for Control Register 2
  digitalWrite(SS, LOW);
  delay(1);
  SPI.transfer(addressByte);
  SPI.transfer(ctrlRegByte);
  delay(1);
  digitalWrite(SS, HIGH);

  delay(100);

  // Control Register 3 configures Interrupts
  // Since I'm not using Interrupts, I'll leave it alone

  // write to Control Register 4: address 23h
  addressByte = 0x23;
  /* Bits:
  BDU BLE FS1 FS0 STsign 0 ST SIM
  BDU: Block data update (0=continuous update)
  BLE: Big/little endian data (0=accel data LSB at LOW address)
  FS1FS0: Full-scale selection (00 = +/-6G, 01 = +/-12G, 11 = +/-24G)
  STsign: selft-test sign (default 0=plus)
  ST: self-test enable (default 0=disabled)
  SIM: SPI mode selection(default 0=4 wire interface, 1=3 wire interface)
  */
  ctrlRegByte = 0x30; // 00110000 : 24G (full scale)

  // Send the data for Control Register 4
  digitalWrite(SS, LOW);
  delay(1);
  SPI.transfer(addressByte);
  SPI.transfer(ctrlRegByte);
  delay(1);
  digitalWrite(SS, HIGH);
}

Processing Sketch

 import pitaru.sonia_v2_9.*;
 import processing.serial.*;

 Sample beep;

 float high;
 int count;

 int inside = -1;
 int bx=850; // position in X of the up corner of the botton
 int by=460; // position in Y of the up corner of the botton
 int h=40;
 int w=100;

 float inByte=0;
 float drawByte=0;

 PFont f;

 Serial myPort;         // The serial port
 int xPos = 10;         // horizontal position of the graph

public void stop()
{
  Sonia.stop();
  super.stop();
}

 void setup () {
   // set the window size:
   size(1024, 550);

   high = 0;
   count = 0;

   f = createFont("Verdana",6,true);

   // List all the available serial ports
   println(Serial.list());
   // I know that the first port in the serial list on my mac
   // is always my  Arduino, so I open Serial.list()[0].
   // Open whatever port is the one you're using.
   myPort = new Serial(this, Serial.list()[0], 9600);
   // don't generate a serialEvent() unless you get a newline character:
   myPort.bufferUntil('\n');
   // set inital background:
   background(0);

  Sonia.start(this);
  beep = new Sample( "beep-02.wav" );
 }

void draw()
{
  if(keyPressed)
  {
    if(key == ' ')
    {
      high = inByte;
    }
  }

  background(0);

  //stroke(255,0,0);
  rect(xPos, 500 - inByte, xPos+20, inByte);

  textFont(f,25);
  fill(255);
  text(inByte, xPos - 10, 500 - inByte - 25);

  count = count + 1;

  if(count > 2)
  {
    count = 0;

    if(drawByte < high - 200)
    {
     beep.play();
     drawByte = drawByte + 100;
    }
    else if(drawByte < high - 10)
    {
       beep.play();
       drawByte = drawByte + 10;
    }
    else if(drawByte < high - 1)
    {
      beep.play();
      drawByte = drawByte + 1;
    }
    else if(drawByte < high - .1)
    {
      beep.play();
      drawByte = drawByte + .1;
    }
    else if(drawByte < high - .01)
    {
      beep.play();
      drawByte = drawByte + .01;
    }
    else if(drawByte < high - .001)
    {
      beep.play();
      drawByte = drawByte + .001;
    }
    else if(drawByte < high)
    {
      drawByte = high;
    }
  }

  if(drawByte > high)
  {
    drawByte = high;
  }

  textFont(f,140);
  fill(255);
  text(drawByte, 200, 325);

  rect(bx,by,w,h); // Button 

  textFont(f,25);
  fill(0);
  text("RESET", bx+10, by+30);
  fill(255);
}

void mousePressed(){
  if(!(((mouseX > (bx+w))
  ||(mouseY > (by+h)))
  ||((mouseX < bx)
  ||(mouseY < by))))
  {
      high = inByte;
  }
}

void serialEvent (Serial myPort) {
   String inString = myPort.readStringUntil('\n');

   if (inString != null)
   {
     // trim off any whitespace:
     inString = trim(inString);
     // convert to an int and map to the screen height:
     inByte = float(inString);
     inByte = map(inByte, 0, 1023, 0, 500);

     if(inByte > high)
     {
       high = inByte;
     }
   }
 }

Maker Wedding: Vinyl “flexi” record wedding invitations

wedding invites vinyl flexi records

My wife and I enjoy music immensely and are both enthusiastic musicians, with that in mind we decided to have vinyl “flexi” records cut for our wedding invites. You probably remember the flexible records or flexi-discs that sometimes came in magazines, or with children’s books — turns out the rumours of their demise are greatly exaggerated, they’re still around, perhaps even making a comeback.

After some preliminary investigation of the process we realized that dealing with a commercial recording just wasn’t going to work, copyright issues and permission issues would’ve bogged us down for months — so we decided to write and record our own collaborative instrumental song. A fun little tune we titled “Invited” was the result of a feverish week of writing, re-writing and recording (and re-recording).

The next hurdle was finding out the who and how of having them pressed, cut, printed, etc. The first place we found was Pirate Press but they were somewhat unresponsive and not at all impressed with our timeline or order size. We needed only about 150 and we were hoping to have them in 3 to 4 weeks, the regular turn-around is 5/6 weeks. Another place overseas (Northern Record Pressing, Norway) was quickly crossed off the list due to distance and our time constraints, in retrospect they may have worked out since we ended up dealing with international shipping anyway. With time running out I contacted David Read, The Vinyl Record Guru. David was friendly, helpful and while cautious regarding our timeline was willing to take on the project.

Once David was on board we mastered the recording and handed them off along with the printing template. We decided to print white ink on clear flexis to contrast the silver and black of various other items in our invite packages. Vinyl mastering was minimal as flexi audio quality leaves much to be desired we decided not to belabour the process, though some simple tips may have gone a long way, ie: center any low frequencies and assume anything below 40Hz or above 16Khz ain’t going to make it in. You can find more info here and here. You can also “simulate” what a recording is going to sound like on vinyl by either applying both a low and high pass filter or by running it through a VST such as iZotope’s Vinyl, which is what I’ve done for the simulated version below.
wedding invites vinyl flexi records
As it turned out, even though David is based in Canada, a lot of international shipping is required, the plates from Europe and the flexis themselves from the U.S. (perhaps from Pirate Press after all, I’m not sure), but in the end it took about 4 weeks to get the flexis after providing everything required — and they looked great!

A new needle for our portable record player from the good folks at Ring Audio in Toronto and we were off to the races. We paired the flexis with a sheet of translucent white vellum and stacked them with the rest of our invite package. The flexis are 7″ x 7″ and fit nicely into 7.5″ x 7.5″ envelopes — though these didn’t fit so nicely into some mailboxes. Another thing to keep in mind is that most paper places won’t stock this size of envelope, you’ll probably have to order them. We ordered ours from Paper-Papers and were quite pleased with the price and expedience, your local paper shop will most likely offer to order them for you but they’ll be ordering from the same supplier you can find online where you can have them shipped directly to you, without a markup.

All said and done, the flexis alone came in around $750, we had to settle for a minimum order of 250, by no means the most inexpensive invites but they came with the added benefit of having our original song on vinyl and, as musicians, that gives us the warm fuzzies.

Invited (MP3, Simulated Vinyl Version)
Invited (MP3)

Ester Pugliese (Lead Guitar, Electric Air Organ)
Phil Tucker (Rhythm Guitar, Lead Guitar Two/Three, Percussion)

Maker Wedding: XBee remote relay as photobooth RF camera trigger

XBee Button Relay Photo Trigger RF

If I didn’t have XBee’s kicking around from some other fanciful endeavour I probably would’ve purchased an inexpensive Aputure remote trigger for the photobooth at my wedding reception — but instead I decided to make an RF camera trigger system from parts I had around, mainly two XBees and a relay. XBee’s are great little wireless mesh transmitter/receivers, much like Tomax and Xamot what you do to one is instantly reflected by the other, so you need only connect a pin to high on the transmitting XBee and the same pin on the receiving XBee will go high — perfect for remotely triggering a relay.

crimson twins

The plan was to have one XBee with a simple pull-down button which would trigger a relay on the receiving XBee and whaddya know — it worked. The receiver relay was attached to a wired camera remote (Canon TC-80N3) I already had from a previous project which allowed the relay to trigger both a Canon Rebel XT and, after splicing in the original connector, a Canon 5D Mark II.
XBee Button Relay Photo Trigger RF
With this setup, two AA batteries for both the transmitter and receiver lasted for days. The only issue I found was that the Rebel XT needed about a 500 millisecond signal whereas the 5D Mark II would trigger instantly on a contact of any duration, I thought about a capacitor, timed-delay relay or 555 solution to keep the relay open longer, but never got around to implementing any solution — after all the 5D would capture the shot if not both cameras.

To be honest, this project survived to the night before, but I decided to scrap the photobooth entirely from the wedding reception to reduce complexity — the lighting turned out to be insufficient and, as the groom, I had no time left to deal with the camera setup. That being said, the theory is sound and it worked quite well for days in my living room, taking shot after shot of me and my fiancée lounging on our couch.

xbee-relay