Here’s a video to watch:

Though it wasn’t my goal with this project, it was certainly the original inspiration. I wanted to build a machine that would drop marbles on three different pipes to make different notes. It would drop the marbles in a regular pattern using gearing and sprockets, then bring them back up again via some crank powered system. Thankfully, this idea was shut down fairly early on. While a great idea, it was just a little beyond the scope of this project.

After that, I realized I couldn’t just scale back, but needed a complete redesign. Behold, the first drawing:

A pipe piano! Well, a key of a pipe piano! You’d press down the piano key and the hammer would hit the pipe. This built directly upon Mechanical Movement 420 and included a little octave switch to open or close one end of the pipe. However, even the octave switch was too complex because it would just dampen the tone. Also, this design assumed springs were magic and could just somehow hold the pipe and key up perfectly. I still needed more work. That brings us to the second design:

 

This here is essentially the final design. The only things that changed between this drawing and the final accurate dimensions, replacing the lever fulcrum with walls, and a slight hammer shape change. I drew up the design in Illustrator as a single file, then separated out component parts for individual work.

However, there was one piece I couldn’t sketch up in Illustrator: the pipe. I was lucky enough to find a copper pipe in the basement by the PVC. Copper is like, two-thirds of the was to bronze, so clearly it was the best choice for what’s basically a bell. With a quick google search, I found a formula for relating frequency to my pipe length, but none of it mattered after testing. Mounting the pipe with a bolt through the side threw any calculations out the window. It was best to tune by ear. And by ear, I mean a handy-dandy tuner app because my ear is terrible.

I realized cardboard was not nearly sturdy enough for prototyping the levers, so I went straight to a quick wooden model. It was rough, but it worked! A couple things learned from the Mk. 1:

• The lever arms need to be more than one board (0.205”) wide. There’s not enough surface area for them to hit each other.
• The springs need something holding them down to the bottom or they’ll bounce off and get stuck.
• The back needs to be two boards thick to support the pipe adequately.

I made the changes and cut another model out. This one was much more involved and, had I done postprocessing on it, could have been the final. I plasma cut the hammer head, increased thickness of the device by a board, and glued copies together to make thicker levers. This one really surprised me by how well it worked. I honestly thought I’d have a long uphill battle getting this thing to work, but the Mk. 2 proved successful.

Then came the final. I added some designs to the side and redesigned the hammer to be larger and held with pins. Other than that, it was the same as the Mk 2. I built the whole thing in a day, from start to linseed oil finish. I added the acrylic paint into the raster, then sanded it to remove my mistakes, making the notes com out really well! It sounds decently, but I wish there were more notes. I’m just glad my ambition didn’t kill me.