Neo 2 Splitted Matrix Keyboard MB46-Neo

[burli]

My next keyboard will be a JD40 clone with QWERTZ layout and staggered rows, but I prefere splitted matrix keyboards with Neo 2 layout and I have a prototype of this. I tryed out to find the absolute minimum making use of the Neo 2 features. This is the result



It is absolutly possible to work with this layout and it's pretty handy to use Layer 4 for Enter, Backspace, Cursor … But sometimes it's also impractical. You always need two hands to delete something or just hit enter, if you are not in "touch typing mode". Also the most desktop shortcuts doesn't work like Alt+Tab

So I desided to expand the layout to it's final size



I extensively use layer 3 and 4 from Neo 2, but I add the movement keys to the Fn layer for two reasons. Some people prefer the right hand and desktop shortcuts like Ctrl+Alt+left/right doesn't work with layer 4

I think, this will be the final layout. Now it's time for the PCB design, but I still think about the controller. Should I use AVR or should I switch to ARM? I need tools for Linux

[joey]

Is each side a separate USB device? (you mentioned Alt+Tab not working)
If so look at how the Infinity Ergodox or standard Ergodox work (how they communicate between halves and appear as one device)

I strongly suggest ARM, we should move away from AVR

[burli]

Is each side a separate USB device? (you mentioned Alt+Tab not working)
If so look at how the Infinity Ergodox or standard Ergodox work (how they communicate between halves and appear as one device)

No, it's just one Teensy. It doesn't work only in the old version without native Tab key. In the extended version with Tab key everything is fine. The problem is, that the Neo 2 layer 4 doesn't send a tab scancode. If you press Mod4+ö (Neo ö) for Tab the scancode for AltR and v will be send. The Neo driver converts this to Tab. Same problem with the cursor keys

I strongly suggest ARM, we should move away from AVR

Well, true. But ARM much more complicated I think

[joey]

ARM is not much more complicated at all.
What kind of controller package would you use? QFP, QFN?

[burli]

What kind of controller package would you use? QFP, QFN?

Would prefer QFP, but I can solder QFN if I have to

[joey]

I used the MKL27Z256VLH4 from Freescale. Comes in a 64QFP 0.5 mm pitch package.
It's a great MCU in my opinion!

[burli]

I used the MKL27Z256VLH4 from Freescale. Comes in a 64QFP 0.5 mm pitch package.
It's a great MCU in my opinion!

Built in Bootloader. Hmm, does it work on Linux?

[joey]

Yes, works fine.

My current code is here: https://github.com/jgouly/keyboard-firmware
Hoping to find some time this weekend to upload the USB parts (which I have working locally).

[burli]

I'll give it a try

[joey]

Feel free to ask any questions. I've had two successful PCB designs with it.

[burli]

Ordered some Teensy LC to play. It's a Freescale KL26

[Brammm87]

Damn, this is impressive. I wonder, how does the US International layout do when typing German?

[joey]

Ordered some Teensy LC to play. It's a Freescale KL26

I have a FRDM-KL26Z, that's what I originally wrote my code for. There should only be minor changes to the USB and clock code, but I guess using the teensy code will be easier to start with

[burli]

Today I received sample keycaps for Cherry ML switches.



I decided to use Cherry ML swiches because this keyboard should be small and slim. The keycaps are a little bit smaller than MX and the distance between the switches is 18mm.

based on the availible keycaps I changed the design to this



I want to have a third thumb key. The first prototypes will be without a controller but with a socket for a Teensy or a ribbon cable connector