Hack Notes CVA 090408
(Today Just Skory's) Hacknotes 090408
 Everthing Coming Together
 A sad little story
Final soldering done, I plug in the motor array and battery. Everything worked! It was amazing! Oh my god! Excitedly, I stuffed all the bits into the armature and, well, I got nothing. Totally dead. Now, two days later, I know that just the act of putting the board in the armature knocked loose a cold solder joint which, up until this point of course, had been just conductive enough. That cold solder joint just happened to be the ground pin to which everything else is plugged. Trying over and over to get it to run with compass chip and shift register totally un-grounded was apparently enough to either scramble the bootloader of fry the ATMega chip entirely. I swapped chips with another board and problem solved, for now.
 Today's findings
- Inverted X and Y magnetic field strength in the code because the compass chip actually rests upside-down in the current armature.
- With new chipped swapped in, things were working damned fine (after resoldering and electrical taping some of the motor contacts...).
- But then, on battery power, weird things. It would work for about 30 seconds, and then either get stuck on motor 1 (always motor 1) or stop working at all. Reset did nothing, I had to litterally re-upload the code over USB.
- This behavior never came when running the thing off USB.
- Theory: Something is fuct with the on-board voltage regulator. Using two of these LiPo's was delivering 7.4V. The regulator seemed to be stepping that down OK, but why else this weird behavior?
- Solution: Despite thinking all along that we needed a full 5V (or more) to run everything properly, turns out it all runs just fine off of one LiPo at 3.7V.
- This doesn't really solve the problem of WHY the voltage regulator is screwing with the ATMega chip, but it makes it a non-issue, at least for now. It also, very nicely reduces the weight and complexity of the rig! Yey! Lesson, don't assume some things won't work without trying them first.
- If you refer to the description of the pockets on our armature, you'll recall we have a repeating pattern of three little pockets to hold pager motors and one big one to hold batteries or electronics.
- Turns out, 1.5" was way optimistic and too small to fit our final circuit board structure. **The batteries fit perfectly, as we'd measured them to do, but as the battery is rigid agaisnt the ankle, any pager motors next to a battery in one of those pockets are not held tightly against the skin.
- Solution: add external pockets for electronics and batteries. This also makes wiring easier, and allows the internal 1.5" pockets to be subdivided into more .5" pockets for even greater flexibility in pager motor spacing.
- Finally, now that I'm sitting here wearing this thing, I realize the crimps we used on the ribbon cable are terrible. They huge and very pointing against my supple and sensitive ankle. Worse, having a pointy plastic thing digging into your skin means that vibrations it transmits sometimes feel more distinctive than the pager motors themselves. This is very confusing.
 Next Steps
- Remake the pager motor array on a ribbon cable without plastic crimps (and add two more pager motors while I'm at it.
- Build a minially-sized, rigid (plastic?) casing for the electronics.
- Order all that stuff that we need to order already but haven't!