In this approach, we are using our Tilt Sensors.
Each of these has a little metal ball inside that closes or opens a switch
between the two leads. In this way of using a falling ball within a cylinder, it
is very similar to our final Tilt Arm approach. The only difference is that the
standard tilt sensor was not reliable enough for our application. We were
getting very inconsistent readings from each sensor. Plus, the legs of those
tilt sensors are pretty thin wire, and so after a little fumbling around, they
would bend out of position.
The second approach that you can see in the above prototype is using hall
effect sensors, magnets and white plastic tubes. This method actually
proved to be very reliable and consistent. The only downside was that it
was much more expensive and would require sourcing tubing and/or getting
a custom molded cylinder part made.
Both of these approaches had one more down side: they required the user
to precisely bend the sensor to a 45 degree angle. We found that if this
angle was just slightly off, then the sensor would begin malfunctioning. So
we continued to brainstorm for a solution that would avoid this problem.
With another prototype we used out Tilt-a-Whirl Breakout - RPI-1031.
It worked pretty well, but we wanted to keep this design 100% PTH parts.
Also, it required two of them, which would add to the cost. Also, (just like
the other tilt solutions, they had to be positioned at a 45 degree angle,
which meant bending headers or tricky soldering.
During all of our prototyping, the accelerometer BOB was actually an option
in the back of our minds. Our Triple Axis Accelerometer Breakout -
MMA8452Q would work well for this application.
Page 17 of 20