Abacusynth: Housing Fabrication

May 04, 2022

My thesis, the Abacusynth, is a synthesizer inspired by an abacus.

First Prototypes

I began with some scrap wood connected by a steel rod, and I used small breadboards mounted with rubber bands to hold the sensors in place. I was using a time-of-flight sensor to measure distance and a breakbeam IR sensor to measure rotation.

Early Prototype

Prototyping the sensor with a rubber band

Designing the Enclosure

I sketched out what I wanted the final enclosure to look like. Initially I wanted knobs on both sides, one to control the waveform, and one to control the volume. This changed pretty quickly, since that added a lot of complexity. I opted for rotary encoders with push buttons so that I could combine multiple controls in a single knob.

Sketch of the final design

The first thing to do was to design the walls that hold the sensors and rods. I started with a cardboard rectangle with holes for the sensors: the distance sensor on top and the break-beam on the bottom.

Testing the sensors with laser cut cardboard

Once that was working a moved on to wood. The inner section would be removable, since it would not need any electronics — it would just hold the rods and spinners.

The first wooden inner section

I used 1/2” standoffs to connect the inner and outer sections. The PCB boards fit pretty snugly inside.

Building the wall structure with sensors in between two layers of wood

New Design

Designing multiple layers in Vectorworks

With the prototype working, I designed the final enclosure in Vectorworks. It consisted of multiple layers of wood that would stack to create the hollow walls.

Side view of the new design

The walls would be connected by 1/2” square dowels that I would cut by hand. They would fit within the square cutouts on the inner layer. The base would be 1/2” plywood, also cut by hand. I created a 3d render to make sure everything fit together.

3D render of the new design

I then arranged the panels for laser cutting. The inner wall sections would be made up of two 1/4” (really 0.2”) pieces and one 1/8” (really 0.12”) to add up to just over 1/2”, and the outer panels are all on on 1/8” wood.

Arranging the design for laser cutting

Time to cut! I started with cardboard to make sure the design worked.

Laser-cutting the new design in cardboard

All of the pieces

The assembled enclosure

The assembled enclosure — back view

Enclosure with the rods

Front panel

Ports on the back

The one mistake I noticed with the cardboard prototype is that I had reversed the PCBs in the design — the holes for the encoders were closer to the back than the front. Switching this allowed the PCBs to fit and also made the design better by allowing more room for the cables and the interactive pieces to be closer to the user.

Adding the Electronics

I could now add the electronics and see if everything worked. For more details on this, see the PCB development post.

Attaching the first PCB to the inner wall

Testing the electronics: Distance, rotation, and encoder

Testing with two rods

Connecting the wires and the speaker

More testing

Final Wooden Version

Now that I was confident in the design, it was time to cut the wood. I taped over it to minimize burn marks.

Laser-cutting the wood

The inner pieces on thicker wood

The plastic standoffs fit into hexagonal holes in the inner walls, so that the outer walls could both screw in to the same hole.

Plastic standoffs fit in the inner pieces

Inner section made up of three layers

The two assembled walls

The inner wall — the large hole is for the wires and the small square holes hold the support beams

The front panel is cut out of the thicker wood to allow for a more rounded sanded edge (after sanding)

Profile view

I wasn’t yet set on the front panel design so used the cardboard

Side view with the front panel

Assembling the removable inner section

Drilling it all Together

The inner walls are screwed to the base and the 1/2” dowel support rods. These screws are covered up by the outer wall.

Drilling the walls to the base

Using the laser cut pilot holes

Making sure the angle was square

The structure connected by the base and square dowel

Closer view

The outer walls would cover up the screws

Moving/Spinning is much more firm than with the cardboard prototype

One quirk of the design is that the shaft collars holding the rods in place make it a bit difficult to remove the inner section. I had to cut out a slot in the top and glue it to the inner piece so that you could slide out the entire section.

Cutting slots in the inner wall

Fitting rubber in the gaps

Glueing the cutout piece to the removable section

The notch

Removing the inner section

The old and the new…

Sanding

I sanded the visible edges that were burnt during the laser cutting process. This revealed the wood grain and made the surface smooth to the touch.

Sanding the edges

Looks and feels much better sanded

Sanding the outer-facing panels

Looking pretty good

Final Assembly

Ribbon cables from the sensors

Holes for the sensors

Adding the spinners to test if everything works

Testing the build

All four spinners attached to the inner section

All four rods working!

caption

Testing with the speaker mounted

The assembled, working enclosure

The front face, cut out of the thicker plywood

The back panel

Closeup of the front panel

Encoder Knobs

The final step was to build knobs for the four encoders on the left wall. I started by just sanding a piece of wood to get a feel for the size.

Sanded cylinder knob

I then used a drill bit to cut out cylindrical pieces. These were much cleaner and felt/looked nicer, however I needed a connection piece to hold firmly on to the potentiometer.

Wooden knob v2