Introduction: Build a Programmable Mechanical Music Box

If you've ever seen those little music boxes you wind up, or crank, and they play a little tune over and over from a little metal drum of notes, but wish they did more than play the same 10-second tune over and over for eternity? If only you could change the song and write your own music for it... Now there's an idea.

After a year of design and work, I completed my Re-Programmable Music Box / Mechanical Synthesizer / Organ Grinder Thingy. It has many names, and is 100% non-electric. Just wood, metal, and good ol' people power.

I began this project as a sort of proof-of-concept, designing something from scratch without a lot to base it off of, and a whole lot of engineering problems to solve. Also, I didn't really know what I was doing. It was intended to be a learning and problem-solving experience. And it was a lot of fun.

If you like my work, please vote for it in the Woodworking Competition beginning October 3rd.

Demonstration Video

Step 1: Design and Planning

Since this project was starting entirely from scratch, I needed to make a flawless design that could be easily worked with, and a good design always reduces waste. I decided to use oak, and at $60 per 10ft plank I didn't want to waste any. Also, its pretty complex and precision is extremely important for everything to work correctly.

I started with the idea: a large wood cylinder will hold metal pegs. It will rotate and force the pegs to pluck metal tines, which are tuned to specific notes.

I selected 12 notes across, since it seemed like a very flexible number of notes and allowed me to fill it with a simple 8-note scale in the middle, with a few extra high and low notes. I could also tune in flats or sharps if I wanted some specifically, or get almost all of a chromatic scale in. I selected 32 notes "around" the cylinder, because that's what you need to play Pop Goes The Weasel, or any 8-bar song using quarter notes.

The cylinder is made of softwood, 8" in diameter, that I picked up from a nearby carpenter for free. I found a belt that fit snugly around it to use along with a few gears and a belt for the cranking mechanism.

The tine material was cut from the prongs of a garden rake. This works great because it is flexible but snaps right back into place. The tine holder design was sort of up in the air until the rest of the machine was done, so that I could do testing. Initial designs were too complex and tiny, but the final design (image 3 below) is about as simple as possible, I believe.

Originally I wanted to use almost all wood for this project, but it turned out wood wasn't going to offer the precision that I needed for the tines. I ended up doing it with a CNC at my college. In Step 7 I hypothesize how it could be accomplished without such expensive tools.

When designing the wooden frame, try to place the pieces to get as few exposed edges as possible. The edges you do have, try to align them so that they create borders and still look appealing to the eye.

Step 2: Materials and Tools

Materials:
  • Oak plank (10ft long by 8in wide by 3/4" thick is what I started with)
  • 8" long by 8" diameter wood cylinder
  • Aluminum plate
  • 1" wide 1/4" thick aluminum stock
  • 100 steel pegs
  • 384 1/8" diameter by 1/8" long magnets
  • Rake tines
  • 8" rubber belt
  • Smaller belt
  • 3x gears
  • Carriage bolt and nut
  • ~12" long 1/4" steel rod
  • 1/4" wooden dowel
  • Biscuits
  • Instrument mallet
Tools:
  • Table saw
  • Band saw
  • Drill press
  • Number drill set
  • Fractional drill set
  • Wood glue
  • Clamps
  • Drill bits
  • Biscuit cutter
  • Planer
  • 1" hole saw
  • Polyurethane wood sealant
  • Pliers
* Note: I'm going to assume you know how to use your planer, your table saw, your band saw, your drill press, etc. and I won't go into detail about the operation of the machines.  If you're using someone else's machines, have them teach you how to use them, and have someone around for safety.

Step 3: Woodwork

So its tough to find wood in the shapes you want at the lumber store, so I bought a 10 foot long piece that was only 8" wide.  My solution was to chop it into three 3.3 foot segments and glue them together to form a 2 foot by 3.3 foot oak board.

First, chop the plank into three even pieces.  Then, plane them down to 1/2" thickness, since 3/4" is too thick for such a small project.   Next, cut biscuit notches with the biscuit cutter, and glue them in place.  Be sure to mark the locations of the biscuits so that you can plan around them so they won't be visible after making cuts.

Clamp it together and let it dry for 24 hours.

Measure the circumference of the cylinder, and divide it by 32.  Mark 32 lines all the way around the cylinder.  Now, mark 12 holes along each line for each note.  There should be 384 points to drill now. 

To drill the cylinder, set a vice on the drill press bed and put a few rags inside the open jaws so that the wood won't get damaged.  Use a weight on a string to check that the line of holes you are drilling are straight vertical.  Find a number drill that is a few thousandths of an inch larger than 1/8" and drill the holes exactly 1" deep using the depth control on the press.

Next, glue a magnet into the bottom of each hole.  This is to prevent the pegs from falling out when the drum holds them upside down.  The best way to do this is to place a magnet on the end of a steel peg, put a drop of glue on it, and carefully push it down into the hole.  It should protrude from the hole by 1/8".  Then tap it lightly with a hammer to make sure it's in proper contact at the bottom.

Step 4: Woodwork, Pt II

Hop back on the computer and devise a means to get your shape dimensions onto paper.  You can export 2D images from your design software and print them to-scale, or just read off dimensions and mark them on the wood with a pencil and a square.  When placing things on the wood, remember to keep visible edges away from the biscuits!  It helps if your design is on separate cutouts so that you can play with the positioning to maximize wood yield.  Remember the table saw blade will consume about 1/8" of wood, so leave gaps between your pieces.

I'd like to make a safety note.  Table saws are horrible, ridiculously dangerous machines.  If you are not incredibly afraid of the saw, then DO NOT use it.  You need a very healthy amount of fear, so that you'll take every precaution when using it, or you'll loose your thumbs.  I'm serious.

Once your pieces are cut, theres a few holes to drill.  Drill the holes for the shaft that the drum will rotate on, and the sound-port holes on the front piece.  Don't drill the holes in the octagonal pieces that will go in the drum yet, since you'll need to properly find the center once they're in place. Now start clamping and gluing.  I'll admit, I was on a tight schedule at the time and had to get these parts done in a few hours, so I cheated and used a nail gun.  It was a terrible idea, nail guns cause splits in hardwood, so don't do it! Take the time to use glue and clamps properly. 

To make the crank handle, cut a 1.5" piece of dowel and a 2" piece of dowel.  Cut a piece of leftover oak into a 2" by 1/2" piece.  Drill a 1/4" deep hole near one end on one side, and another hole at the other end on the opposite side. Glue one dowel into each hole.  The 2" dowel will go inside the machine, and 2" may be longer than needed so it can be cut down accordingly.

The way I came up with to glue the end pieces into the wooden drum was like this; First, put wood glue around the outside of the first octagon and lay it flat on the table.  Then, put the barrel down on top of it so they fit in together.  Make sure they're lined up correctly and wipe off any excess or oozing glue.  Wait for it to dry.  Then take a center-finding tool* and locate the center of this side of the drum.  Drill a 1/4" to meet your shaft diameter.  Now, do the same thing with the second octagon, but this time, to make sure everything is pushed flat and the octagon isn't too far into the barrel, push it with a rod or dowel through the hole in the opposite end.  Once the glue is dry, find the center of the second side and drill a second hole.everything should be perfect!

Cut a 12" length of 1/4" solid steel rod for the shaft the drum will rotate on.  File the ends so they're smooth.

* A great Instructable for a center finding tool is https://www.instructables.com/id/Make-a-Center-Finder/

Step 5: Mechanics

Now we need to start working on the mechanical linkages and crank for the drum movement.  Since putting the crank right at drum-level would result in the crank hitting the table, I had to devise some transfer mechanisms to allow me to put it higher up.

I found some plastic gears which meshed correctly with the 8" rubber belt I had for the drum.  I glued two of the gears together and glued a long bolt into a hole I drilled through the side of the box.  On the end of the box, I drilled a hole and glued a nut into it.  This would let me insert a machine screw with a captive armature on it, which held a gear on the end.  This would let me tension a transfer belt to prevent slipping of the belts and gears but still make the crank easy to turn.

The result was what you see in the image below.  The system works pretty well, and I don't think there could be a much easier way to do it without fabricating custom gears.

An important thing to think about is the space you have.  There was only about an inch of clearance between the internal wall and the drum, so space was tight.

Step 6: Finishing the Wood

I decided to finish most of the wood early because I wasn't going to be able to keep working on the project for the next four months, and I didn't want it to get dirty or change shape from water in the air.  I just coated it with a few layers of Minwax Polyurethane to make it shiny and bring out the colour of the wood a little bit more.

After the polyurethane is dry, squeeze the 8" belt onto the edge of the drum.  I cut the belt in half because it was too wide and covered the first row of holes. If it doesn't hold in place by friction alone, you can use finishing nails to tack it in place.

Step 7: The Musical Mechanics

This was a major engineering challenge for quite a while.  I had a few designs that were a bust, they were just far too small to be practical.  I ended up distilling the ideas into what I think is as simple as could be possible. 

The tine, which is a piece of steel from a leaf rake, is sandwiched between two aluminum plates.  The tine is kinked at the tip and is roughly the length it needs to be, plus enough length to be held in the plates firmly.  This is secured with two small screw that pass through the top plate and into threaded holes in the bottom plate.  The bottom plate then sits atop a large aluminum plate with slots machined in it.  On the bottom are larger slots with captive nuts in them, held in place against the top of the music box.  A screw passes through a hole in the second plate and into one of the nuts so that its position can be controlled by sliding it up and down in the slot.

Using this method, the tine length can be tuned to the appropriate note, and then the tine holder can be moved towards the pins on the barrel, and lined up for just the right amount of plucking force.

To build the tine holders, I used 1" wide 1/4" thick aluminum stock.  Since I only had 15.3mm of space per tine, I cut the pieces down to 15mm wide, and then into 15mm and 30mm long segments.

The tines can be cut from the rake using a hack saw or angle grinder.  Cutting their lengths is a hit-or-miss guessing game, however.  Depending on their width, thickness, and composition their resonating characteristics will change greatly, so I can't give a list of predetermined lengths.  You'll have to just cut a few and experiment to find what works for your tines.  The tips can be bent with two pairs of pliers.

I used a CNC to cut the aluminum plate.  Now, I know that using a CNC is out of the realm of 99% of readers, but I had the resources and it was the best logical way to do the design I had created. 

If it had to be done using simple tools, I would say that sandwiching pieces of wood together to form the slots, or use a chisel to carve out the slots for the nuts, and a drill and needle files to do the rest.

Step 8: Final Assembly and Tuning

Tuning is a pain-staking process that takes an hour or two.  The tines can be tuned to lots of different scales, but I selected a simple set of notes from the piano. They are:

A3
B3
C4 (Middle C)
D4
E4
F4
G4
A4
B4
C5
D5
E5

The result is a pretty good range of notes that works for lots of little songs like Mary Had A Little Lamb, Hot Cross Buns, and Pop Goes The Weasel.

To tune, start by arranging all the tines by length, and then place them tightly in the tine holders.  Mount all the holders on the aluminum plate, but not near the drum. Tighten them down so they will resonate nicely.  Using a piano or a virtual piano (like http://www.virtualpiano.net/) play the note that you want to tune to.  Strike the corresponding tine with a plastic mallet. If it is too high, loosen the two screws holding the tine and wiggle the tine out further.  If it is too low, push it in further.  Once the notes sound right (and you'll know when) then you can tighten down the tine as tight as possible so it can't slip.  Then, slide the tine holder up to the drum and put a peg right near it.  You want the peg to just barely touch the tine as it comes up, and then pluck it.  You don't want the tine to bend upwards more than a few millimeters.  Too much force could break the tines over time or cause the gears and belts to slip.

Repeat this process until all the tines are tuned.  Set up a "test pattern" by placing pegs in diagonal lines across the drum, so it will play the scale when rotated.  Test it out!  If it sounds good, try out a song or two.  You should be able to use music written for recorder, that should work well on this machine.  Chords are possible, but elaborate ones may require more torque to pluck than the crank will allow.

Step 9: Improvements and Considerations

I'll admit, after all the time and effort I put into good design on this project, it didn't turn out as well as I had hoped.  Part of the problem is the insane precision required to do this properly, which could never be achieved with wood.  If I were to ever try and build something like this again, it would have to be made entirely from metal.

Other than that I don't think the design has any major flaws. I would love to see some comments on improvements and enhancements for the future.

I hope you enjoyed reading, thanks very much.