[IC] RotaryPad - Now in final pre-GB phase

Interesting. I don’t know enough about the project or PCB design to understand, possibly… but what’s up with this break apart design? Is it just for users to have anywhere from 4-9 encoders? Actually, I’m completely lost as to how the snap off sections work. how da heck do they connect to the main board?

Yep, it allows you to go down to a 6 or 4 encoder config if you’re not wanting the original 9, but all while keeping a single design to get better volume pricing. For the routing, it looks like this going across the little tabs.

1 Like

Nifty.


Working prototype is working

6 Likes

3 Likes



New pics after the Durham meetup this weekend.

1 Like

So it’s finally about that time as I’ve gotten back into being able to work through stuff. More roadblocks on BitBang mean I’m back here again. I’m actually fairly close to having this just ready to go out the door, I just need to finalize some tidbits and work out kit contents.

Features and status:

  • Controller compatibility - Should be completely functional with any ProMicro footprint controller, with the Elite-C having a small exception of having a GPIO where the RAW pin sits on basically every other option.
  • Snap-away function - Takes slightly more force to break off than I expected, but works well with only a little cleanup needed.
  • Breakouts - SPI, I2C, and RGB breakout headers all available.
  • Case - Had a pretty big “well duh that works” moment in getting the bottom cover to work, so a few adjusted tolerances and the case it’s ready.
  • Encoders - Gotta see what PEC11R variants are actually in stock

Now to the awkward portion: QMK support still isn’t in main since the PR for matrix encoders has been belly up for almost as long as I’ve put this off. Slightly good news: I can probably just host a fork with the matrix code, probably. Good-er news: I’m actively working on KMK support for the board, which is why I’m planning to offer Adafruit KB2040 as the “preferred” controller in the kit. And also because chip shortage go brrr and it’s both well priced and available.

Gib feedback plz

1 Like

More updates time:

Testing case design changes: I was hoping for completely tool-less PCB fitment into the case, but getting that working well and also not being odd to work is taking some effort. I may just give up the tool-less dream and use screws after all.

Controller: I’m still making it ProMicro footprint, but the “official” controller of the board is going to be Adafruit’s KB2040. RP2040 supports KMK via CircuitPython and in soon™ territory for QMK support, USB C go brrr, and (most importantly) it’s actually available. I’m planning to offer these as an option to order alongside the rest of the kit. Also the pink is pretty sick.

Firmware: I’m now aiming for KMK as the default option because it’s easier/less involved than QMK to put it on a board, and I think most people will be plenty happy with being able to just change keycodes in a text file with no compile needed.

PCB: Here’s where I’ve been hung up since last update. As mentioned, QMK didn’t technically support the nonsense used to make the previous design work, and I’m too much of a Python scrub to translate it over to KMK myself. So the answer is I’m now migrating the hardware design over to a more “normal” encoder layout for the closest four and an IO expander for the rest. This also got interrupted when the first two options I wanted to use are just impossible to buy right now, so I’m on my third hardware change there. Good news is these are actually pretty cheap, and I’ll probably go with DIP so they’re easy to hand solder. The number of solder joints on the board actually drops a good bit with the diodes disappearing.

As always, feedback is appreciated and I hope to have more updates this coming week as I continue trying to get the PCB and case nailed down.

1 Like

New hardware revision being sent off to prototype. JLC had a fucking fit over the nice slotted setup I had going on before so I had to modify it to get protos for not 8.5x what they should have cost :confused: Electrically identical though, and I’m possibly looking away from JLC for the GB anyway thanks to the allure of pink soldermask.

Poll time: What color soldermask is anyone interested in?

3 Likes

Pink, eh? Sounds like it would look good with some rainbow sprinkles on it too.




Protos are here, including having to mess with the breakaway portions to make JLC happy. Next step is get one soldered up for testing and finalizing a basic firmware. Case needs a bit of design update but that should be pretty simple.

I’ve also been poking with pricing breakdown and it’s unfortunately a bit chunkier than I had expected and wanted it to be. Looks like at low MOQ a full 3x3 kit(PCB, KB2040, MCP23018, mounting hardware, 9x encoders, printed case) is going to trend at least $50, with the encoders and controller being almost 2/3 of the cost in that. And up to a 50 MOQ, which I highly doubt I should even think this could hit, the cost reduction is less than a 25% drop from the low MOQ costs. This is the cost of using decent hardware I guess since this is neither a $4 aliexpress ProMicro nor $1 ebay encoders, and also a nicer fab than JLC for final PCBs because even their smaller stuff is starting to get uglier.

1 Like