Skip to content

20. Project development

This week I worked on developing my final project All Ways Sound.

Updates and changes (very important!)

During the 2nd half of the Fab Academy, things and points of view changed a lot.
One special checkpoint was the Mechanical/Machine assignment, in which I saw the whole course in just a unique work like that, all things and capabilities applied to one big effort.
And that, again, was a essay of the Final Project assignment.

The fact of working in group during the machine challenge let the good feeling of fronting the targets in a collaborative way. The group consolidated itself and new ideas for the final project appeared. In that way, the most clear and atractive conclusion was to make our final projects as modules of the same big project, they could interact between them to give (in this case) a performative result, like an art installation in which our autonomous and completed projects could work together to connect through the inputs and outputs.

So, my initial idea about the Pocket Sound Box got replaced with a type of sculpture in which sound and light could be driven by physical presence (movement), with my original idea of giving the public the important meaning of sound in our lives.

This new project is called All Ways Sound and is chosen as the idea to develop during the final project.


All Ways Sound (the -real- final project)

Concept

  • The presence of sound

    The project is a sculpture/installation focused on remind the people the relevance of sound in human life. Sound is in the air, in the water, in the stone, and in any elastic medium in which the sound waves could be propagated by pressure variations on it.
    Although this concept is more or less easy to explain and to understand with the help of the experience, the fact is that sound is constantly ommited as the energy that it is. It can make our lives better or worse than it is thought and, even with that energy rounding on our environment, it is treated as something on a second plane.
    With this project, I want to bring the presence of sound closer to understanding and awareness, so that the experience allows us to remember it later, in day to day.

  • The process flow

    The way that All Ways Sound device works as follows:

  • The device is placed somewhere where the public could perceive it though its sound and light and the physical device itself.

  • The device is emitting a kind of light and a kind of noise autonomously until something/someone gets near (over) it. (It needs a 12v power supply and a computer to work) This is thought to be the signal that calls the people to get nearer.
  • The device changes the sound and the light behaviour when finds a proximity interaction just in the air because of the presence of sonar input. Here is where the concept of sound must get its biggest meaning about its transmission through the elastics medium: the air.
  • The device returns to its default behaviour when leaving it again alone, so the conclussion points to an idea of the sound that has to be attended to avoid the noise that produces continously.

  • Materials recipe

    The materials used to produce this device are:

  • The interactive behaviour

    The device is designed to combine the ultrasonic inputs on the All Ways Sound board with the computer audio output previously processed on Pure Data software.

    • INPUT:
      The ATTiny44-SU microcontroller on the All Ways Sound board processes the HC-SR04 ultrasonic sensor values and prints them by serial.

    • INTERFACE:
      The computer receives this input from the board FTDI RX/TX through the USB port and puts it into Pure Data with the comport command. The OSC commmand generates sound waves which are processed according to the HC-SR04 values and the Pure Data modifiers, like frequency, signal level or reverb. Also, these sound waves become audible through the Pure Data DAC command and the computer sound card.

    • OUTPUT:
      The audio signal is sended through the audio wire from the computer audio output to the 8 OHM 10 W speaker, which needs an audio amplifier board to allow the speaker to work efficiently.
      The ATTiny44-SU microcontroller on the All Ways Sound board also sends the HC-SR04 values to the neopixel LED strip through a PWM pin to modulate the light.

    By this way, when any movement is made on the air space near the sensor, the sound and light change accordingly.
    By default, the systems is triggering noise and a light pattern.

Design and Technologies

All Ways Sound design takes in consideration a few ideas:

  • Amplify the sound physically
  • Human interaction size
  • Case for components visually unatractive
  • Rigid parts to stay up without supports
  • Simmetry, colour and bightness for esthetic


According to these criterias, the design can be unfolded on these parts and measures | fabrication technologies:

  • 2 almost similar big conichal (morph between and hexagon and a circle) PLA 3 mm thickness printed shapes for both amplifying system and for components case | Made with a big Delta printer
  • 1 central 180 mm ring-type MDF structure (connecting 6 30 mm thickness) to hold the inputs/outputs and to attach the PLA pieces | Made by cutting on CNC
  • 2 5 mm acrylic round pieces in where the inputs/outputs are attached (these pieces are part of the central structure) | Made by lasercutting
  • 1 round PCV 0,7 mm membrane at the center of the structure to work as a drum | Made by thermoforming
  • 1 hexagonal MDF 30 mm thickness base in which the components stay (power supply and boards) | Made by cutting on CNC
  • 1 case (square MDF 3 mm thickness, transparent 0.8 mm thickness glasspack parabolic shape) for boards and input/output connections | Made by lasercutting and thermoforming
  • 2 PCB dedicated to control the sensors and to amplify the sound signal | Made by precission milling and soldering

Production

In general, the process was inverted from the previously planned one. I knew that the electronics and programming jobs could be hard and difficult with any little trouble. That was the reason to begin with that.

So the first thing I made was to buy everything I would need, that is, the LED strip and the electronic components and many assembly parts.
As I began to have problems with the 12V power supply process, I decided to finish the whole design details and fabrication to have the way clear to just ‘fight’ with the electronics and the programming wth everything else finished and ‘working’.

The documentation process was made day by day on the train and the bus while going/going back to the lab, taking photos and videos, making the process/project structure and explaining as much as possible all the details of the idea and the process.

Here are shown some steps of the whole process:

  • 3D printing

    The 3D printing process was a little slow, because of the machine that was stucked because of the colorant pellets. I had to begin the process many times until it was clean and calibrated. It began at the beginning and finished almost at the end of the week!





  • CNC cutting

    The CNC cutting was one of the first jobs. The pieces were sucessful altought I had to repeat one of them because of an error on the GCODE. I regenerated the geometry and the machining job to solve it.



  • Laser cutting

    Laser cutting jobs were made on also at the beginning and on the middle of the week, with some tries searching for tolerances when changing materials.



  • Electronics

    This work was tested from the beginning and finished near to end. I had to buy another TDA2003 when the first board was connected inverted to the power supply.
    Finally, I decided to customize a little bit the board to indicate the components place and the GNDs too.







  • Assembly

    The assembly was made and remade many times, starting it from the beginning. Althought the design was made to make changes easily, the parts were big and the most work was done without the 3D printed parts.

















  • Programming

    The programming process was made at the beginning using the previous sonar code and following some tutorials, and the final details were made at the end. Luckyly, when the power supply stuff was solved, the code worked without problems.






Problems and solutions

Most of the problems I had were related with the electronics:

  • The 12V power supply issues
    Both the speaker and the LED strip are managed by more than the USB voltage. My way to front this was by using an ATX power supply to give them the additional power needed.

    Speaker

    It just needed an amplifier on a circuit connected to the ATX power supply, so I bought all components needed (including the TDA2003 transistor) and tested on a protoboard an example amplifier circuit to ‘move’ a 10W - 8 OHM speaker that I found on the Internet. At first the signal gain was so poor, until I noticed that one GND was not connected.
    Conclussion: connect all grounds even if they are from different voltages.

    I’ve read that, in the case of sound, this idea of connecting grounds can add some noise to the signal.

    Once the speaker was working correctly, I decided to draw and mill the board, soldering as many SMD components as I could find (some ceramic or big capacitors were very difficult to find on stock and it so late to buy them by the Internet). The problem, then, appeared when adding the audio terminals and the big components.

    BROKEN SOUND BOARDS

    I tried X times to put pins for input/output and laying down the big components until I finished a reliable board.
    The main issue was that traces used to separate from the board. Someone recommended me to use a less than 40W power solder iron to avoid this, but I think that, if I used the same tool for all my boards along the Fab Academy without that problem, the reason was not the solder iron, but the way to hold some components that should be attached better (with holes or glue) to be ready to front higher forces, like thicker wires.


    The solution was to do that and to put another kind of connections like male/female pins layed down on the board, as the FTDI example showed on week 05 (Electronics production). I was afraid that pin connections could not be as much powerful as audio and current could need.

    Another issue with the board was the accidentally inverse connection to VCC and GND when testing it, so a spark suddenly appeared somewhere on the board and it, as supposed, didn’t work after that. The TDA2003 transistor exploded.

    LED strip

    The idea was to empower the LED strip with the ATX power supply and to control the light modulation through an ATTiny44-SU PWM pin in which the ultrasonic device would be connected.
    For that, I researched on the Internet on how to do this and I found a tutorial (LINK) that used a TIP120 transistor with Arduino and a potentiometer.
    I adapted the circuit to my All Ways Sound board, putting 2 power supply inputs, 5V and 12V, and also an input pin connection to the 5th ATTiny44-SU PWM pin, with a resistor in between.
    Connecting the LED strip to the board, nothing happened when loading a code in which the HC-SR04 sensor inputs could modulate de 5th pin signal :(
    After making some test with the multimeter, I realised that 12V were reaching the TIP120 collector, but only 5V were at the emitter.
    Here began the real problems, because I knew the LED strip was working directly to the power supply, but wasn’t while following the tutorial schematics.
    I tried the circuit on an Arduino and the LED strip turned on but at low intensity, and, again, without been modulated.
    At this point I made my big mistake: I had to options: to research again on the Internet for another examples (LINK1, LINK2, …) and to look for details on the same and reviewed again and again tutorial. And I DIDN’T declined for the second. The bad thing is that the solution was there, and I lost a time I hadn’t frustrating myself surfing forums and tutorials. The good one is that all that effort helped me to understand better the LED strip and power supply world.
    In between of this “LED’s Quest” I searched everywhere any 5V LEDstrip without any success… until a friend of mine gave me a neopixel old piece. I made it work in Arduino and, when putting all the stuff together on my All Ways Sound board… problems returned. The code was again overpassing the 4Kb limits because of the Neopixel and sonar libraries. When trying to
    So… the solution was in another image at the first tutorial, where the connections were different from the schematics. Here is where the documentation becomes crucial between text and images. I was connecting almost fine the TIP120 transistor, but not as well as necessary. See the differences with the previous image:

    Finally, when hacking the board with a wire, the light came up, controlled by the HC-SR04 ultrasonic device. It is a ‘short’ explanation, but it took me a big effort to discover and solve it (also with the help of my partners).

Summary

  • What tasks have been completed, and what tasks remain?
    Except for the boards case and the neopixel extra, everything has been completed in general.

  • what has worked? what hasn’t?
    Everything worked, with big efforts!

  • what questions need to be resolved?
    No questions pending, everything is clear!

  • what will happen when?
    Next steps for the All Ways Sound device are the nearly Santiago de Compostela Maker Faire and October’s Madrid Maker Faire. Also, during the next months, the device will be slowly improved in fabrication details and software behaviours in order to explore the interaction between digital and physical fields.

  • what have you learned?
    I have learned a lot!
    Mostly about the human implication and capabilities on the present and the future. The milestones are:

    • The knowledge about the power of context
    • The comunication and organization
    • The day to day discipline
    • The digital fabrication variety and possibilities
    • The programming and automatization power

Results



Note: This week focused on develop the final project. If you are looking for files related to it, go to the final project page.