search

ebastler / UnMechanical

Capacitive touch keyboard running QMK firmware
CERN Open Hardware Licence Version 2 - Permissive
7 stars 2 forks source link

Request: add a "review" to the README #1

Open DeflateAwning opened 3 months ago

DeflateAwning commented 3 months ago

This project seems super cool! Wish I saw it advertised in a blog/reddit post, instead of re-inventing the idea myself while daydreaming about a keyboard so thin I could store it between my laptop's existing keyboard and its screen.

Would love to see a 2-3 sentence-long review of the keyboard in the README so I know where the project is at! Should I buy it as-is for $50 to play with, or does it need a new rev before ordering? Are the thumb keys usable at all (my thumbs have a bone that prevents the tips from touching an iPad keyboard very well, for example)?

ebastler commented 3 months ago

Hi! That's a good point, thanks. I must admit, I mainly pushed ahead with production of this board to be the first to publish such a keyboard design and due to technical interest, not so much due to actually wanting to use it (I am a terrible touch typist and struggle even with simple ergonomic boards like a Corne).

I made a proof of concept firmware back in the day that worked, but I failed in managing to get QMK split working with the capsense matrix scan code by @Cipulot . I know he expanded his code to support split operation by now, but afaik it has not been tested/adapted for this particular MCU yet. I mailed a pair of PCBs to him not too long ago, and hope he will be faster than me to get this working properly, at last.

I will give some more feedback once either Cip or I managed to create a working split-enabled firmware, as typing on half a keyboard is not particularly useful.

I also think some of the passives on the PCB need adjusting to get more sensitivity out of the capsense, but I have not dived into that yet. Basically something that can only be optimized through trial and error with a hot-air gun and patience.

In my testing, the board reliably recognized fingerprints, but it was not capable of doing so even through a very thin layer of plastic, for example 0.5-1mm thick 3D printed cover with grooves where the keys are to allow haptic homing. Swapping some passives in the feedback signal amplification part may, but is not guaranteed to improve this, and it may be that typing on the bare PCB remains the only option.

DeflateAwning commented 3 months ago

I guess the big question: does it need a new revision before I order it to play with it (and maybe hopefully help contribute some firmware debugging), or is it pretty well ready-to-order as-is?

ebastler commented 3 months ago

Cipulot said he'd like to experiment with the capacitors/resistors around the opamp which might improve reading strength by a lot - I'd wait for his results. I lacked a good sortiment of SMD parts myself so never got to do it.

I also noticed a bug in the USB-CC detection which explains some random behavior of not always being recognized immediately, I will fix that as well.

DeflateAwning commented 3 months ago

@Cipulot, any updates or insight? Advice on next steps, or what you've tried so far?