Hack Notes CVA 090415

From Noisebridge
Revision as of 12:52, 16 April 2009 by Elgreengeeto (talk | contribs) (New page: =Skory’s Hacknotes 090415= ==Past Days’ Progress== ===ironing out bugs and smoothing out edges=== *I fixed my spacing issue. **My amateur soldering left me with a pager motor array t...)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Skory’s Hacknotes 090415

Past Days’ Progress

ironing out bugs and smoothing out edges

  • I fixed my spacing issue.
    • My amateur soldering left me with a pager motor array that just couldn’t be squished in the armature’s little pockets right. On one side they were too close, the other too far.
    • Of the two logical options that followed, I chose the lazy one of just deactivating one motor on the side that was too dense and telling the code that I had only seven motors.
    • The spacing is still not 100% perfect, but because of the non-circular shape of the ankle and the fact that the whole rig is likely to shift in use, 100% perfect vibrator spacing is a practical impossibility and I’m very satisfied with what I’ve got with seven motors at this point.

Some experimentation with timing.

  • It is clear that when one is immobile we don’t need to leave motors constantly on (and, as a matter of fact, having such can present a real blister danger.)
  • The current code turns a motor off after a given period of immobility (measured totally inexactly in processor cycles; let’s call this “time-to-off”).
  • If you move enough to switch to another motor within that period, the motors will not turn off.
  • If you remain immobile, there is still a regular pulse indicating the persistent orientation (call it “persistent pulse interval”).
  • I have experimented with time-to-offs of 10 seconds and 5 seconds, and persistent pulse intervals of 30 and 60 seconds. None of these are altogether satisfying.
  • My next plan is to experiment with having the time-to-off decrease while the persistent pulse interval increases as a function of time spent immobile.

A hike in Tilden

  • I went on a hike and it was simply brilliant.
    • I hiked a few trails which I have hiked or run probably well over a thousand times in Tilden Regional Park (on the edge of which I grew up).
    • Here’s the wild part: I know these trails really, really, really well. All of them. Turns out, many of them veer wildly from their orientations in my mental map. The most remarkable parts were in the middle stretches of well-sheltered trails, particularly bits in the middle surrounded by canyon walls and lots of trees. There were times I was walking nearly 90° off of the heading in which I thought I was walking. I was sometimes well off of my expected heading even in exposed bits of trail.
    • Intersections, however, were generally oriented within maybe 10-15° of what I expected. This suggests to me that my mental map is organized by how the trails connect to each other and where they begin and end, but does very little to account for how the trails actually get there.
    • I also noticed that much of the error I was perceiving was related to expecting sections of trail to be more or less headed in a cardinal direction. My brain seems to have done a lot of regularizing wiggly paths into a more manageable network of straight lines, enough so that I was really conscious of the fact that North was often really not where I expected it to be.

Riding BART

  • Riding BART if freaking amazing.
  • The field lines flip back and forth depending on acceleration of deceleration.
    • This is particularly noticeable during long stretches where the train is apply little bits of acceleration and deceleration to maintain a constant speed.
    • When stopped in a station the train keeps electromagnetic brakes engaged the whole time. I know this because I can feel it.
      • This. Is. “Why.”

A voltage bug!

  • With ten minutes left of my hike the thing just froze. This may be from the battery getting week.
  • What I do know: from whatever cause, somehow the code got scrambled such that it got stuck on one motor and the only way to reset it was to plug it into a computer and re-upload the code.
    • This is the same as what was happening with two batteries plugged in. It seems to be triggered by funny voltages.  ?

Yesterdays Adventures In Fabric

  • Picture this: Eric walking through the fabric store with a square inch of Velcro in his hand. He shoves it at every bit of fabric he sees, most of the time finding no resistance in removing it disappointingly. Here’s why.
  • We want to make a new armature. The armature is really what makes or breaks this project, after all. Here are the constraints:
    1. We have to hold the motors in place, yet have them easily repositioned.
    2. We need to have the fabric around the motors to be not-rigid enough that it does not bleed to much vibration, making it difficult to feel which pager motor is active.
    3. We need an enclosed channel of some sort in which to lay wiring. There needs to be someway to keep the pagermotors connected to this wire.
    4. There can’t be other, larger objects protruding further than the motors on the interior of the armature. This only pulls the fabric, and thence the motors, away from the skin.
    5. There needs to be a place on the exterior to hold the circuit board and battery.
    6. (Hella important) it needs to be easily manufactured.
    • Our first design addressed number 1 with little pockets, but these pockets really fail number 6.
    • Our first design addressed number 3 with a channel above the pockets. This however failed number 4 (which we hadn’t really thought of, oops!).
    • Number 5 really needs external pockets. Attempting to solve it with internal pockets fails number 4 at best, and just doesn’t really work at worst.
  • The new design:
    • We want to position the motors using Velcro. Glueing the motors to bits of velcro (or really, sandwiching them with pointy-side velcro on one side, and a thin, non-abrasive fabric on the other side) is not too hard.
    • Making the internal surface of the armature out of commercial fuzzy-side Velcro critically fails constraint number 2. Hence the trip to the fabric store to find what soft fabrics hold onto Velcro the best. The answer: Felt. (Pulling the velcro off slowly destroys the felt, but it seems it would take many tens of off-pulls to make it unusable, and I find this acceptable.)
    • I want to add a layer of thin foam under the felt to really isolate each motor’s vibration and to, hopefully, quieten the buzzing. (The current rig’s buzzing is hella audible in any quiet-ish room.)
      • Eric thinks this is overkill. Our designs will thus diverge. Great!
    • Now the channel holding the wiring will be behind the pager motor array. This should better address contraint number 4.
    • There will be one long, or many small, slits between this channel and the felt interior to pull the wires connecting to the pager motors through.
    • Again, on the outside, pockets for a battery and the circuit board.
    • This new rig ends up much less tall than the current one. This is desirable.