Making the sensor

I fabricated my own sensor from a piezo-electric disk salvaged from a Radio Shack Tweeter. (Cost less than $10). Radio Shack also sells Piezo-electric elements like this one:

Which would probably work as well.

Also electronic watches with beepers in them will have very similar elements (usually cemented to the back cover). So that might be a place to scavange an element as well.



When you rip it apart, the piezo element you get looks something like this:





What is needed at this point then is to support the element from it edge and couple the center to the watch. I did this by modifying a cheap case holder like the one at the right.



I used a piece of 3/8" (9mm) baltic birch plywood to make the mount for the sensor. Any other reasonably knot free wood, plywood, or even a machinable plastic would probably work as well.

I cut a dado in one end to allow it to partially overlap the movable jaw on the watch holder above. In the end this step was pointless however. (See the last picture.)



Then used a forstner bit to create a well in the plywood to hold the peizo sensor. First a 1" bit to create a well large enough to hold the sensor, then a 3/4" bit to create a space behind the sensor to allow it room to vibrate freely. I used my Moto tool to add a couple of "ears" to the smaller cutout to provide clearance for the soldered connections on the sensor.



A small hole was added to the well for the cable. The sensor is then soldered onto the cable. I used a cable from a junker computer microphone, you could make your own as well. The "tip" of the mini-phone jack should connect center conductor and to one side of the piezo sensor, the sleeve should connect to shield of the cable and to the other side of the sensor. If your connector and/or cable has a third wire for the "ring" on the mini-phone connector leave it unconnected.

The sensor is then glued into place with hot glue.



To give the watch under test something to hold it I took a nylon washer, filed it's center hole to be somewhat more square, and the hot glued a stack of the nylon washer and a slightly smaller metal washer to the center of the piezo sensor.


Here's a watch under test. I immediately discovered a flaw in the design -- the jaws of the holder won't open wide enough for most men's watches. This little Bulova only barely fit.


To cure the problem I moved the sensor assembly to the other side of the jaw. The sensor is recessed just enough that it doesn't make contact, the plastic washer neatly just clears the top. I also added some self adhesive felt stickers to the top sides of both jaws to insure that the watch would not get scratched.

This picture show the sensor pressed against the side of the case. Since taking it I've discovered that you can get a much cleaner signal by pressing the crown against the sensor.


There is still more to do. At the moment I'm just plugging the element into the microphone input on my computer, but there would probably be some benefit to build a preamp for the sensor that incorporated a suitable high pass filter in the design.

Also I'm still thinking about adding some sort of outside framework so that it is easier to test in any orientation.