Table of content

• Assignment
• My Work
• Group Assignment
• Idea
• Design
• Research
• Calculation
• 2D/3D Design
• Material (and re-design)
• Mill
• CAM Setting
• Tool path
• Preparation and Mill
• Assemble
• Problems and first-aid
• Rebuild
• Sound check
• Materials
• Files
• Lessons Learned
• References

As my final project is making sounds, I came up with idea that I will make a "something big original speaker".

Though it's my first time to make speaker, I challenged to make "backload horn" typed speaker that can emphasize low tone by stretched horn from back side of speaker unit. It's good chance to design and cut complex curve using large CNC and joint it tightly not for leaking noise.

According to above reference website and youtube videos, some says important design elements of BH speaker are:

1. So: Throat cross section on the back cavity of speaker - the cross section of the starting point of back horn
2. V: Volume of back cavity(chamber)
3. S(x): Neck(Sound path) crosssection area - more stretching neck(sound path), more larger cross section area.

1. So(Throat cross section) = "Diaphragm area of speaker(スピーカーの振動面積)" x 0.5(this is defined by Q0(振動先鋭度) of speaker(according to rule of thumb...see here(Japanese))

2. V(Volume of back cavity) = 10 * So / f (cross over frequency of speaker))

3. Sx(sound path cross section area)V(Volume of back cavity) = So * exp ^mx

I tried to check different portion, speaker width and size of back cavity using spreadsheet. Longer horn is better for sound but also I want make my speaker not too big.

Finally I calculated and decided So(throat cross section area), back cavity hight, back cavity depth and exponential neck width per 10 cm of neck.

1. draw external shape of speaker (from side aspect)

2. draw back cavity

3. draw neck shape by line. loner is better, but too narrow path cannot be made.

4. After drawing line, it goes thorough to the front panel of speaker.

5. Assign rectangle per 10cm of neck length. For each 10cm, the width of rectangle increases.

6. Make the rectangle line into curve. Utilized "3-point arc", "Center point arc" and "tangent arc"

Firstly, I tried to design square shape from side view like above.

However, square type looks like desktop PC and I think that shape is not so cool and takes unnecessary space. . That's why I changed the design like this - longer hight and shorter duration.

Then I re-design the shape of horn. Finally, I got speaker with:

• 440mm (height)
• 108mm (width)
• 160mm (duration)
• 1100mm (length of backload horn)

I designed the joint for front panel attaching speaker unit and the back panel for speaker cable terminal. It's good to make these panels by detachable joints as I want to change the size of speaker unit for experimental and maintenance reason.

As endmill is rounded shape and thicker than laser cutter or CNC for PCB milling machine, we need to open dog bone or similar shape when cutting inside of shapes.

As my speaker's surface is not so thick and some part only has 22.5mm, I designed the dog bone carefully so that I can keep thickness and strength of surface.

When making complex sketch with bunch of parameters on Fusion360, the sketch or entire design tends to be heavy. Sometime it's freezed or brings unexpected error. Though I've been in trouble for too heavy sketch problem, I found a solution to resolve this finally.

I tested it by cutting 2/5 size (my original plan was using 15mm of MDF by laser cutter. I made the design by 3mm of tilia plywood and check joint tightness by 3 * 2mm of that .

Tilia plywood - Ref: wiki, maker's site

The length of design looks to be consistent. However, the depth and width of slit in dog bone joint is not enough tight. This is because of laser cutter's kerf under condition of Universal laser cutter in Kannai and the material of 3mm of plywood. I definitely need to do test cut by actual material that I mill in larger CNC.

Originally, I planned to use 15mm of MDF. However, that material was sold out at CAINZ, home center near by Fablab Hamamatsu.

So, I needed to change the design of speaker. Some parts is redesigned by changing parameter of thickness while the other is not. It was because I used function like "mirror" or "circular pattern" and the pivot of the pattern was changed when I changed parameter.

• Spindle speed = 12000 rpm
• Feed rate = 700.0 mm/min
• Plunge rate = 300.0 mm/min

Making pocket, I used following settings.

• Start depth = 0.0mm
• Cut depth = 4mm
• Clear Pocket... = offset
• Cut Direction = Climb

CAM toolpath is made by "Cut2D". I ordered and distributed the image to physical area of plywood board.

Cut2D supports move the path image to a specific position, edit path, making tab and export the data to .tap file (G Code) etc.

For safety, wearing

• cover glass
• headset
• (mask)

Import G Code(.tap file) at the button of "Load G Code"

Start and reset buttons are both for CAM software and hardware button. Also, manipulating the XYZ axis are controlled by keyboard.

Step to setup the origin of tool is as follows

1. Move the mill to the left - under corner of the material
2. set origin (zero) of x axis and y axis
3. check z axis hight by gently pushing the Down keys. The appropriate height of z axis is checked by news paper between mill and material board, then set origin(zero) for the tool.

setting “zero” (origin) for x-axis ( and then, y-axis also), after moving the mill to the corner of the wooden board.

It took around 2.5 hours for cutting every parts of a speaker in total.

As we had time constraints for the remote location of Fablab Hamamatsu, I needed to change the plan from making stereo speaker into monoral speaker...

As I was the last user who had been using ZN1325 for FabAcademy for 4 days at Fablab Hamamatsu, I cleaned up the filter and cartridge covered by sawdust.

burring by belt sander at Take space

checked the quality of cutting - some drills are not penetrated through and the bottom thin board is pealed off in accident. I recut the material again as every part is not so big.

Partially using wood glue because speaker needs to be sealed tightly. I did NOT use wood glue for the part I made as a joint for speaker unit and back panel.

Joint for a panel of speaker cable terminal is tightly fixed.

cable terminal is assembled.

Using insertnut(鬼目ナット) for attaching tightly and detaching easily the cover for maintenance.

For fixing the position for attachment, I used a few dowel for each piece that I need to join.

This goes along with initial design of front panel slit.

This looks to be attached well, but I changed the design of this a little bit considering following problem and resolution.

This is the first assembled example of backload horn speaker. Measurement looks to be fine, but I want to correct some points as follows.

1. Failed to set the depth of pocket cut.

I wanted to do pocket cut by 5mm of depth firstly, but the mill goes to more than 5mm. That was due to incorrect value of setting on pocket cut in CAM.

Luckily it was not penetrated. So, I put wood pate to the part of too deep hole.

Afterward, I re-design the outline panel with more tight joint for front panel and hold for acrylic side window.

2. Unpenetrated profile and drill

As the material board it self and the sacrifice board seem to be slightly warped, the upper part of material was not fully penetrated on drill and profile cut.

Also, I setupu drill cut depth as 12.5mm against the 12mm thickness of plywood - 12.5mm looks not to be enough depth of drill cut.

For recovery,I cut off the last 1mm of unpenetrated part by cutter. For drill, I made a penetrated hole by electric drill.

3. Mismatched joint

The depth of pocket for back panel was 4mm depth but the joint length to be inserted to the pocket was 7mm.

I cut-off the joint by saw and shrink the length into 4mm.

4. Insufficient tags for narrow part

When cutting narrow part, the material was not enough fixed that the cutline became unstable. It seemed to be rare case, but as this picture, some part was not cut as designed.

Though the inside of speaker does not appeal their shape, I needed to smoothen by file as long as I can.

Actually, I forgot to pocket cut for attaching acrylic panel to the outline side panel. I discussed with my instructor, Tamiya-san, about how to recover this.

I decided to redesign the outline panel and make 2mm range of pocket offset to the window and make the joint parts a little thicker that can hold the acrylic panel with the offset.

This time, I successfully cut the pocket for slit that combines with front panel. I designed the depth of pocket as 7mm and the joint parts length as 6.58mm.

Sanding by 120 and 240 of sand paper, then oil finish to the lauan plywood - two coats to the "visible" area. The end surface of plywood absorbs large amount of oil than expected.

Redesigned pocket works as tight joint.

For the transparent panel on the side outline, I cut acrylic panel with 5mm thickness. The measurement of acrylic panel is just the same as the windows in side outline. The kerf of milling machine and laser cutter solves would solve the tight insertion.

For smoothing the surface of the speaker, I roughly did wet sanding by #2400 of sand paper.