Lab6_Haptics

Lab 6 HAPTICS

The goal of this lab is to introduce some simple one-dimensional active force-feedback using a motor and encoder. We start either with a small knob or with a stick that has a force sensor on the top; other devices could be attached to the motor shaft. Each haptic effect can be coupled to a sound by sending OSC to Pd. Different haptic effects suggest different sounds; different sounds suggest different effects. What are the most interesting mappings? Which are appropriate, controllable, even "expressive". Try each sound control with and without the motor turned on. Are there some that simply cannot be played without haptics?

PROFILES - First we compute forces as a function of just position: for every encoder position, there is a specific force. Think of it as a profile or landscape -- a resisting force feels like pushing something up a slope, as the force increases, the slope increases. Think of the spring-centering of motortest.c as a "valley", wall.c has no force until you reach the wall, bumps.c and detents.c are all variations of force as "slope". Textures can be rendered as series of bumps. Stability: A very stiff wall is hard to make "stable": first, the encoder and force values are discrete making for "stair steps" or "zippers" and the calculations in fixed-point reducing resolution, more importantly the update rate must be high. You can explore various resolutions and rates.

TIME - Forces that vary in time can be pre-programmed (tick.c). A short positive then negative force, if it is quick enough feels like a click. When pressing against a spring, a click simulates a contact being made.

DYNAMICS - We can increase the apparent "mass" - giving the knob a twist, gets it going. You can simulate simple mass-spring oscilators (think of it as a pendulum). Without measuring force (we only measure position) we cannot reduce the "mass" or "damping". The human hand acts as a "damper" stabilizing an oterwise oscillating system; the hand can also be a spring-mass coupled with the virtual spring-mass just like a drum stick bouncing on a snare.

FRICTION - can be simulated with a combination of texture, dynamics and even pre-programmed

0. Set up

    a. connect the ribbon cable from PORTX on the motor board to PORTD on the AVR. Check the orientation: make sure that both headers have pins 9-10 at the same side of the ribbon cable (e.g. BLACK)
    b. wire the motor's encoder (grey cable) to AVR/Protoboard: White to D2, Green to D3, Red to +5v and Black to GND. D2 and D3 are the two external interrupts and represent Phase A and Phase B from the encoder.
    c. connect the power for the motor directly to the 15v supply pins, Yellow on the left pin and Black on the center pin of the two that have been attached to U5. This is the same place that we picked up 15v for the OpAmps in Lab 4.
    d. plug in the Spyglass LCD to PORTC.
    e. for the stick, connect the FSR Red to +5v, Black to GND, White to A0 for a2d conversion.
    g. load Lab6 on to your 250a directory

   download: these program files files and unpack them in your lab 6 directory using tar -xvf

   h. add these files to your ~/avrlib/ccrma directory.

   cp -r /usr/ccrma/web/html/courses/250a/labs/lab6/avrlib_ccrma_lab6.tar ~/avrlib/ccrma
   cd ~/avrlib/ccrma
   tar -xvf avrlib_ccrma_lab6.tar

1. motorspy (spring-centering)

    a. run "motorspy.c"
    ~>cd ~/250a/lab6/programs/motorspy
    ~>make load    b. The LCD should display Encoder: Force: and Duty:
       i) Try moving the knob/stick left and right.
       ii) What is the maximum duty?
       iii) If you press on the FSR the forces will increase.
   b. Try 'motortest.c' It is the same program but with out the Spyglass display.
   c. Try 'motorosc.c' It is the same program but with out the Spyglass display and with OSC. After loading motorosc.c, start motorosc.pd.

2. wall

    a. run "wall.c"
    ~>cd ~/250a/lab6/programs/wall
    ~>make load
    b. Look at the code for wall.c.
        i) Make the wall "stiffer"; what if it's too stiff?
       ii) Make the walls further appart or closer together.

3. bell (one wall, sending to Pd)

    a. run "bell.c"
    ~>cd ~/250a/lab6/programs/bell
    ~/250a/lab6/programs/bell>make load
    b. run "wallbang.pd"
    ~/250a/lab6/programs/bell>pd bell.pd
    c. when you "hit" the wall, a sound should be triggered.
        i) Change the sound file in the Pd patch by loading "snare" or "kick".
       ii) By varying the wall stiffness (SW7 or SW7) can you get an intentional bounce?
    iii) What stiffness (k) best matches the sound chosen?
      iv) With the right stiffness, you should be able to get a "drum roll".

4. bumps and detents, plucks and friction.

    A "bump" is a slope up to a round top and a slope down on the other side.
    A "pluck" is a slope up like a wall but followed by an abrupt drop in force at the top.
    Stick-slip "friction" is many small plucks in a row.

    a. run "detent", "detents" and "pluck".
    b. can you make a "bump" instead of a "detent"?
    c. think of sending some OSC messages at the right time/positions.
    d. vary W, D, S to match the sounds from Pd.

detent

5. tick - timed forces

    a. run "tick.c"
    ~>cd ~/250a/lab6/programs/tick
    ~/250a/lab6/programs/tick>make load
    b. Look at the code for tick.c
        i) Make the ticks longer, larger, more frequent.

6. walltick - "tick" on wall "entry"

    a. run "walltick.c"
    ~>cd ~/250a/lab6/programs/walltick
    ~/250a/lab6/walltick>make load
    b. run "snare.pd"
    ~/250a/lab6/wallbang>pd Snare.pd
    c. what is the "right" place (distance into the wall), magnitude, etc.
    d. try matching the wall stiffness and the tick to different sounds.

7. pendulum

8. spring, mass and damping

references:
<~/avrlib/ccrma/osc.h>
<~/avrlib/ccrma/osc.c>
not included in <~/avrlib/docs/>

see schematics and manual for MotorBoard on Pascal's site: AVR boards
motor data sheet: <~/Lab6/RelianceMotor.jpg>