So our idea is to make a first operating version and, futurely, an "ultimate" version will all the big boi features. In this sense, I need to know if you guys agree with this or definitely want a first PCB with bluetooth or 3.1 USB or both.
Great project!
I would definitely prefer an "ultimate" version over the base initially. I can't imagine being able to justify the cost and waste of displacing multiple PCBs. I'd be much more willing to pay extra to know any PCB I invest is the best version possible. USB 3.1, Bluetooth - those features would definitely bring the HHKB to endgame territory and would make it superior to the new Hybrids. If cost is a concern, in the grand scheme of things, I honestly believe that anyone interested in buying an aftermarket HHKB PCB wouldn't bat an eye over a $24++ difference - people gut $200++ Novatouches just for the sliders to mount their $100++ GMK keycap sets.
But perfect is the enemy of good!
I know you're working on a billion things - and projected timelines on this?
Yeah, unfortunately all my projects were delayed and I wish I had more time to work on this.
Going for the ultimate version right away is not the best way to do it, at least in my view. That's because the most basic part -- the switch activation circuit -- is novel and, at this point, still experimental. Despite having one-switch protos that worked fine, I had to do a major rehaul of the activation circuit due to some current specifications. I am doing my best to make it the most robust and straightforward solution, but when it comes to analog sensing electronics, many variables come in place and they should all be considered. Placing components and guessing values is not an option as in most digital electronics. In short, I think that in this first moment the most important thing is to prove the concept of the switch activation circuitry, which is not simple on its own right, and then expand on that. This is called induction-based design and is a very secure and transparent way to manage a project.
Second, development of a USB 3.1 hub is not simple as 2.0 . I was re-thinking and this USB hub gave me an opportunity to do an "unified USB hub" daughterboard that could be open-sourced, much like ai03's Unified Daughterboard which latest version (C3) is of my authorship. This open up the possibility of other makers offering USB hub on their keyboards simply using that daugherboard. I also think that this opens up the possibility of expanding the options for the Faraday60. Initially what I was thinking on doing was offering a basic PCB with USB2.0 hub only, and then the more sophisticated PCB with USB3.1 . However, as I realized Faraday and the hub PCB are independent, one can easily make both daughterboards compatible with the motherboard. This means that it's better offering a USB2.0 hub DB, a USB3.1 hub DB, a basic motherboard and a more complete motherboard, so that the user can choose which to buy and upgrade accordingly. I think that this opens up a possibility, because as far as I have gathered, people are completely fine with a USB2.0 hub but would like bluetooth. So I think that there is no reason in tying the feature-rich version of the motherboard with the USB3.1 .
Third, I want to adapt Hasu's idea for USB-C and implement it on the basic PCB, and hopefully I can add both USBC and mini-B support for those that consider stock features better. I am however not sure on how to go about this, because the little castellated PCB needed is not simple to solder and I don't feel comfortable leaving that to the user to solder.
Finally, I'm waiting for QMK support on ARM Bluetooth; I received word from the developers in the QMK Discord server and apparently this is not a top priority. So I think that this initial version will use ARM, which is better, faster and cheaper, and the BL version will use AVR. Using ARM is a big deal for me because I think that it is the future of this hobby that people will eventually ditch AVR and start using ARMs. ARMs are also way cheaper than AVR's; for instance, STM32F072CBT6
costs US$1.1 at 100+ units while ATMEGA32U4
costs US$2.8 at 100+ units, nearly a third the price. This saves up some budget for me to use top-quality components in the sensing circuitry, where they matter most.
All in all, there are a lot of ideas going into this project and that is exactly why I chose HHKB Pro 2 to start, not FC660C or Realforces. It opens up a lot of fronts in custom keyboard design which I think will benefit the community as a whole, as I am going to open-source all of this. That way other makers can use the ideas and circuits to offer better products. But, then again, lots of things to consider and I'm going one thing at a time. So I won't have a precise timeline on when I am going to finish, but be certain I am in complete passion about this PCB and I intend fully to see it through.