Close
0%
0%

TonnetzOne - Isomorphic Midi Keyboard

An isomorphic midi keyboard using the Tonnetz layout

Similar projects worth following
64 views
0 followers
An attempt to design and build an open source isomorphic keyboard.

Isomorphic keyboards have keys arranged so that each chord shape is the same regardless of the start note. A regular piano keyboard is not isomorphic - a perfect fifth chord of C is very different in shape to a perfect fifth chord of F#. With a regular keyboard a large number of chord shapes must be memorised, but an isomorphic keyboard has a much small pool of chord shapes.

The layout of the keyboard is a grid following the Tonnetz pattern of notes.

When the prototype keyboard is complete, all source files (code, schematic, pcb etc) will be made open source.

Requirements:

  • Compact chord shapes
    • Isomorphic chord shapes reduces the number of shapes to memorise
    • Compact shapes make it easier to hold more keys with fewer fingers
    • Cannot spread fingers well, closer is better
  • Light keys
    • Less tiring on hands
  • Velocity keys
    • Velocity produced by time between two switches
  • Min 7 octaves
  • Compact and lightweight
  • USB Midi

Status:

  1. Analysis and design of keyboard layout [Complete]
  2. Prototype A - Key Mechanism Prototype [Complete]
  3. Prototype B - Small section test [Awaiting PCBs]
  4. Full Keyboard - TODO

Current layout design:

  • 120 × Gateron White / Clear MX clone with low actuation force without clickyness
  • 120 × Panasonic EVQ-Q2B03W Tiny switch with low actuation force
  • 120 × 1N4148 Discrete Semiconductors / Diodes and Rectifiers
  • 120 × 2.5mm Rubber O-Ring Dampeners Thick dampeners for MX style switches
  • 70 × White DSA keycaps

View all 9 components

  • Prototype B - Small section test (in progress)

    Redgeneral11/15/2019 at 16:19 0 comments

    Status: In progress (PCBs arrived)


    Purpose: 

    • Small scale test of key mechanism (5x2 keys) 
    • Controller program test
    • Responsiveness test
    • USB midi functionality test

    PCBs:

    PCB 1 (Teensy, 74H595, 10 EVQQ2 switch, and male header):

    PCB 2 (10 MX type switches, and female dual entry header):

    PCB 3 (Plate):

    [PCB rendered by online pcb viewer]

    Music notes

    The test pcb should return the following notes:

    B0D#1F#1A#1C#2
    C1E1G1B1D2

  • Prototype A - single key test

    Redgeneral11/15/2019 at 16:18 0 comments

    Purpose:

    • Determine min and max distances between EVQQ2 pcb and the MX pcb - Achieved
    • Confirm key mechanism works - Achieved
    • Test possibility of reusing same PCB design for each level and plate - Feasible but not practical 

    Identical PCBs:

    The Mitosis Keyboard aimed to reduce costs by using PCBs that could be used for multiple roles, such as the MX plate and MX PCB mount. Sections could be broken off to turn the PCB into a plate.

    [Image source: Mitosis Keyboard - https://imgur.com/a/mwTFj ]

    When a PCB design is ordered, often it is produced in sets of 5 copies. Costs could be reduced if the same PCB design could be used for each layer (EVQQ2 layer, MX layer, and Plate).

    Unfortunately when I received the PCB, I was unable to remove the breakaway sections and so had to dremel them out.

    In the end, although feasible, it would be very difficult to design for.

    PCB render:

    Note: diodes are the wrong way round on the design (oops), but have been soldered correctly instead.

    Photos: 

    [Photos of assembled prototype to be added later]

  • Key Mechanism

    Redgeneral11/15/2019 at 14:32 0 comments

    Each key is made from two switches, with the velocity determined by the time difference between. The larger switch is a Cherry MX type, with the seeond being EVQQ2 type.

    This key mechanism is based on the one seen at http://wiki.lvl1.org/Isomorphic_Keyboard


    MX switch

    [Image source: cherrymx.de/en/dev.html]

    The MX internal switch mechanism contains a plunger that is stopped by a well that extends beyond the base. A hole is required in keyboard PCBs to accommodate this well.

    By cutting off the well, the plunger is able to extend beyond the base and strike a switch below it.

    The MX switch variant chosen is the Gateron White (a MX clone), due to the the low force required and silent operation.

    The MX switches are held in place using a plate that the MX switches clip onto.  Normally this plate would be laser cut, however a PCB without traces is used instead as it is cheaper. The holes for the switch body to enter are made during the PCB manufacturing.

    EVQQ2

    [Image source: EVQQ2 datasheet]

    The switch has a very low actuation force, so little resistance when pressed by the MX switch's plunger.

    One problem with this switch is that is a surface mount component and so not easy to solder by hand. 

    Headers

    • Male header on the EVQQ2 layer
    • Dual entry female header on the MX layer
    • Matrix columns shared between layers via headers
    • MX row output passed to EVQQ2 layer (which also has the Teensy LC controller)
    • Note: the space between PCBs means that the male header's plastic component must be smaller than usual
    • Due to height of female header, the Plate cannot cover this header

    Key caps

    • DSA profile
    • DSA keycaps are rated for all rows on a computer keyboard, so they all must have the same shape and no slope.
    • Tonnetz keyboard requires all the keys caps to be identical without slope.

  • Column circuit test

    Redgeneral10/20/2019 at 11:48 0 comments

    Three 74HC595 chips will be used to drive the 24 columns of the keyboard's button matrix. I built a prototype on breadboard to ensure the circuit design and code were correct, As this was just a test of the columns, LEDs are used instead of buttons.

    Code:

    Read more »

  • Tonnetz Layout

    Redgeneral10/18/2019 at 20:54 0 comments

    Layout:

    Specifications:

    • 7 Octaves + 2 Notes
    • 120 keys : 86 Notes -> 1.4 Keys : 1 Note

    The layout pattern loops - the bottom rows are repeated at the top of the next column along. For example, the F# and G at the bottom of column 2 are repeated at the top of column 3.

    Chords:

    With the chord keys being adjacent, it should be possible to two keys with the same finger.  Entire length of of chord might be playable with just four fingers.

    Octave map:

    Midi note values:

    Midi value for each key could either be calculated on the fly or stored in an array.

    After comparing all the isomorphic keyboards, I chose the Tonnetz layout because:

    • Compact
    • Low key to note ratio - reduces weight
    • Chord shapes seem reasonable - compact and requires fewer fingers
    • Full 7 octaves

  • Alternative layouts

    Redgeneral10/18/2019 at 12:07 0 comments

    Before starting work on this project, I considered several different isomorphic layouts 

    • Harmonic Table Layout
    • Wicki-Hayden Layout
    • All Fourths Layout
    • Janko Layout

    Below I show the layouts and major/minor chords of each and my reasons for not choosing them.


    Harmonic Table Layout:

    Specifications:

    • 7 Octaves
    • 252 keys : 84 notes -> 3 keys : 1 note

    Chords:

    [Blue coloured indicate alternative keys to change the type of chord - e.g. Diminished chord by using F# instead of G]

    • Hexagonal keys allows for chords to be played with fewer fingers
    • All key caps would have to be custom made
    • Vertical chord patters did not seem very comfortable to play

    This was my initial choice for a layout, but at 252 keys, I very quickly became apparent that the weight of the switches alone would be too much.

    There was a commercially available keyboard using this layout called the Axis-64 by C-thru, but it has now been discontinued.

    Wicki-Hayden Layout:

    [Note columns of sharps on left are repeated on the right]

    There are two general configurations:

    1. One tall keyboard - with the hands one above the other
    2. Two split keyboards - each hand has its own keyboard. 

    Chords:

    I find these chord shapes really uncomfortable (especially the blue modifiers) - this is the deal breaker for this layout.

    All Fourths Layout:

    Specifications:

    • 150 keys : 50 notes -> 3 keys : 1 note
    • 4 Octaves

    The All Fourths was a strong contender, but it's few octaves and large number of buttons were against it.

    Janko:

    [image from wikipedia]

    For me the Janko is not suitable for several reasons

    • Just as long and as heavy as a regular piano keyboard
    • The chord shapes are not very pleasant for my hands

    Janko keyboards are currently in production branded as Chromatone - https://chromatone.jp/

  • My interest in Isomorphic Instruments

    Redgeneral10/14/2019 at 21:31 0 comments

    My drive to make an isomorphic keyboard comes from my limitations. I have painful joints that can make playing regular musical difficult. 

    • I can't hold items of any weight for long so that eliminates wind instruments, guitars etc.
    • Playing percussion jolts the joints causing pain
    • Limits to how wide I can spread my fingers without increasing pain - this means piano keyboards keys are just too far apart
    • Stringed instruments often have very contorted hand shapes to hold down the notes - not pleasant on my hands

    Using a grid of keys in an isomorphic layout should be better for me to play

    • Keys closer together for less stretching
    • Chord patterns that require less twisted hand shapes than a regular piano keyboard
    • Potentially a smaller, lighter keyboard
    Hopefully with this project, I should end up with a musical instrument that I can play comfortably. 

  • It begins

    Redgeneral10/14/2019 at 20:30 0 comments

    Project work begins

View all 8 project logs

Enjoy this project?

Share

Discussions

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates