DIY Paper Speaker Driver

Paper is not an unusual material to build loudspeakers ! Early loudspeaker were made with paper (I love this one from Gaumont) are still made with although they may have additional hardening lacquer wich also protect them from humidity.

I am always excited when my diy creation moves or make noise or sounds...In this instructable we'll have fun with a DIY paper speaker driver.

Common speaker driver (electrodynamic speaker driver) are quite simple. It is basically a diaphragm (like a paper cone) attached to a coil (which will be excited by an electric signal out from a amplifier).

A magnetic field is provided by permanent magnets allowing the coil to move in reaction to this electric signal.

The coil is linked to the magnet through some sort of spring plate (called spider) allowing it to vibrate, which vibrates the diaphragm and produces sound.

Here are a few resources on speaker drivers :

https://www.ht-audio.com/pages/SpeakerBasics.html

https://massless.info/images/Loudspeaker.pdf

In this instructable I propose a way to make a diaphragm, spider and coil tube out of paper.

Those three pieces are glued together, an enameled copper wire is wound around the paper tube to make the coil.

With a few magnets, some 3D printed part, we then have a speaker driver to play and experiment with !

Do not get me wrong... we are not going to build the best ever speaker driver...although they are simple, many factors affect their quality. For instance, for better sound we should separate / isolate acoustic from the front and the back, which I have not done in this simple design.

But it is a excellent basis for learning and exploring possible improvement (ie different spider or diaphragm geometry or material...possibilities are endless).

An ideal diaphragm would be rigid and light. A simple diaphragm could be a cone. However, the outer edge of a cone has too much flexibility and would not vibrate as well as the rest of the cone (commercial speaker driver have an additional suspension part to avoid this), to work around this I created a truncated hexagonal cone. It is much more rigid with its straight edges.

For the project we need paper...

1. cone : medium weight paper for the cone and tube (not sure but I you estimate it to be around 120 gsm)
2. tube : medium weight paper
3. spider : heavy weight paper (300 gsm)
4. paper glue
5. enameled copper wire 0.1 mm diameter (42 SWG)
6. kapton tape (single and double sided)
7. magnets (ring shaped)
8. steel plate and tube (optional)
9. 4 x 2mm screws
10. 3D printed parts
11. 4 x 2mm Heat-set threaded inserts
12. wire/connector to test the speaker driver

Tools :

1. bone folder tool (a must for clean and accurate work)
2. toothpick
3. cutter
4. soldering iron
5. laser cutter (or printer and cutter)
6. optional : something to help winding the coil ie drill, home made winding machine...

We first need to cut our paper to build this cone. Several options here, we could create a pyramid cone net and have only two edges glued together or cut individual pieces glued together.

My choice was to go with individual pieces, easier to design and more flexible in term of possible shapes.

I have access to a laser cutter (btw it is my 3D printer that I converted in a laser cutter. I made something close to this https://all3dp.com/2/convert-3d-printer-to-laser-cutter-laser-engraver-mod/).

I need some gcode file to feed it. Inkscape can create those gcode files for laser cutting with this plugin.

A good starting drawing file format for laser cutting is a Scalable Vector Graphics (SVG) file. The plugin would then easily create the gcode file we need.

Drawing shapes can be done "manually" with Inkscape, using the various tools available but for non simple shapes it would be difficult and more over if you change your mind about the desired dimensions or shape you may have to redraw the whole shape.

I wanted to programmatically draw in order to have accurate dimensions and be able to adjust shape or dimensions with a few variable.

Processing is a excellent tool for this purpose with many tutorials available, moreover it can export your drawing into an svg file (which Inkscape can open to create the gcode)

You do not have access to a laser cutter but want to play with paper speaker driver ? You can still print your svg drawing an cut along the lines with a good cutter !

I attach in this instructable a template processing file that I use as a starting sketch for new drawing.

Paper cone, coil tube and spider sketches are also attached.... I am not a good coder, there are probably useless variables in there, not enough comments and maybe errors...but they work

Attached are also svg file (either direct result from the code or reworked with Inkscape, like the cone one in order to place 6 pieces by copy paste)

It is a two stage version, for extra rigidity the outer edge folds on itself as a triangle and the middle part will be glue on itself

This is a example of parts and the code used to draw it

The spider is some sort of spring plate maintaining the cone in place but allowing it to vibrate.

We have many possibilities here playing with geometry, materials.

In the video example we can see an example of a loudspeaker made with a very flexible spider made with light paper. It is really cool to see it moving... now it is not a big success when it goes to the low end frequency (bass playing)...

This piece of paper will hold the coil, it is also the link to the diaphragm and spider

Using a bone folder will help getting accurate and sharp folds, here are some instructions

All lines are scored and then folded.

I start by gluing the middle part on itself, then the outer edger.

The six cone pieces are then glued together, one to another, progressing up to the full hexagon pyramid

In order to wind the enameled copper wire around the paper tube I made a "quick and dirty" wind machine (with an arduino and two stepper motors... but you do not need that although it makes it simpler to build some number of coils in case you want to experiment several speaker drivers...

I think that with patience you can get a decent coil winding by hand or with the help of a power drill.

I use a piece of wood with the proper diameter (15 mm, you can probably find already made birch dowels in stores) and wrapped the paper tube around it.

Tape the wire on it, a do a first wind around.

How many turns ? What is the proper coil size ? Well not sure about that....

Looking at how commercial coils are made, it appears that most coils have two layers, and based on the magnet size an 8mm coil looked good to me.

One thing to check is that the resulting coil will have a impedance that should match what amplifiers need. The most common nominal impedance for loudspeakers is 8 Ω. This is the result of wire diameter and wire length and this is why I choose a 0.1 mm diameter wire. For an 8mm coil it means about 80 rounds per layer.

If you build a coil with a larger diameter, it would increase the wire length and you will probably need to go with a larger diameter wire to stay close to 8 Ω.

One simple verification is to measure its DC resistance, It should be a bit below 8 Ω : https://soundcertified.com/what-will-2-4-8-ohm-impedance-speaker-measure-dc-resistance/

For my coils I use kapton tape which helps securing the wire on the paper tube.

1. I first wrap the paper around the dowel
2. Wrap double sided kapton tape around it
3. Wind the two layers (80 turns each)
4. Wrap a last single side kapton tape around the wire.
5. A tip of UV activated glue helps securing in place the two wire ends

We now have our three main pieces and glue them together.

The two wire ends go through the center spider hole.

Solder the two wire ends to a connector of your choice

What about polarity ? On commercial loudspeaker you have + and - connectors...

This one also has a polarity... but we do not really care.... Polarity is important if you have multiple loudspeakers to make sure they are not interfering which would be a problem for the sound quality.

With a single loudspeaker it is not an issue.

In order to hold the magnet (three rings stacked to make it 15mm height) in place and to allow the spider attachment I created a few 3D printed pieces.

the are a set of washers and an envelop.

One of the washer is a two part washer with threaded inserts, allowing an assembly in between the spider and the cone.

Four screws will hold everything in place.

This is quite different compared to commercial loudspeakers where the spider is usually taped on the magnet part.

My setup has a few advantages :

1. first is makes it easy to center the coil in the magnet part
2. second it allows assembly/de-assembly when you want to test various cone/spider/coil with the same magnet assembly

With or without a center pole. Loudspeakers have a center pole, a metallic part centered in the magnet, it helps having a nice magnetic field through the coil.

I turned one out of a piece of steel, It could be also made from a steel rod threaded into its center and attached to a steel back plate, ti could also be made of stacked washers assembled with a screw....

but for fun and simple test purpose this center pole is optional, you will get sounds out of your loud speaker without any center pole.

You may have noticed a little wood piece plugged into the center pole... this is only for a stylish finish :)

I created a wood stand for my speakers... nice and convenient to enjoy the tests...(and I like testing in style :-) )

Listening to some music with a DIY paper speaker driver is a good test (a bit of disappointment maybe but should be a lot of satisfaction :-) ...

I was really surprised by the sound quality I got out of it.... however it is far from good and for sure needs improvement to compete with good commercial ones.

Another possible test is to feed the speaker with a sound file that goes through a frequency range that human can hear, then records what goes out of the loudspeaker and perform a spectrum analysis (Audacity can do it)

If my loudspeaker was perfect, its frequency analysis (left chart) would perfectly match the frequency analysis of the input (right chart)... and it is not the case we can see significant weaknesses at some frequencies (ie 800 Hz)

Next step will be to explore different geometry (cone, spider) play with paper weight ...

Another thing is that I build this loudspeaker like a tweeter as there is no surround attaching the external part of the cone to any gasket which is ok for high frequency but not a good idea for lower frequency... my next tests may include one....