Late to the game on this one, but Thomas “Chyrosran22” reviewed this board. I watched it a while ago, but IIRC, it boiled down to “pretty mediocre board, but fine, elevated by the keycaps.”

Hand wiring can be a lot of fun. I have done about a dozen boards, including the “pocket battleship” I use for work everyday. This looks perfect.

Did you go with KMK?

There are custom-made products for this. You don’t have to buy from them obviously, but search Amazon for baby proof washing machine. Should be under USD20.

For the very basics of KiCAD, their own intro is helpful. I had to wrap my head around the workflow: first schematic, then PCB.

Once you grasp that, some keyboard specific stuff is described by Joe Scotto in this video. He tends to rush his tutorials though, so best to go through the KiCad page first or revisit the video after.

For resources, I found Ai03’s library of footprints to be perfect, except that his vertical 2U were not oriented how I needed them, which I realized too late, LOL. Then, the kbplacer plugin by adamws was absolutely invaluable and saved me so much time. I used JLCPCB, because even when US-China tariffs were at their worst (and goodness knows they could be again if Trump gets a hangnail or something), JLC was still the cheapest option for me.

If you’ve never done a PCB before, I might humbly suggest keeping to a fixed layout for the first go-round. I did that on my first board and it worked perfectly for what I intended it to be. I got a little ambitious with this one, and ambition+inexperience+impatience led to a flawed project.

SKCM white. Both switches and keycaps (and the front feet, actually) came from a fried early-90s Focus keyboard that had a trackball where the arrow keys go, and the arrows were around it on mouse micro switches with little flappy “buttons” that are part of the case, also very mouse-like.

For the PCB, lots of YouTube and searching Geekhack, deskthority, and (yes) Reddit to see how to use KiCAD for mechanical keyboards. Mine uses the “cheat” of mounting a Raspberry Pi Pico clone to the underside so I don’t have to know as much about electronics (that part actually went perfectly).

The mistakes were two tiny bits of trace that got deleted but I didn’t see, and some placement issues for the Alps version that I had to work around. I have four of them left, so I’ll just use MX compatible switches and a normal sized spacebar for future builds and avoid the worst of it.

The very specific combination might not have been done, but full size PCBs are out there, especially as replacements for vintage boards. If you don’t mind ISO and only a row of LEDs rather than per-key, there’s THIS, which seems to have Bluetooth. They seem to have made some interesting choices with the numpad as well, which is for the most part NOT electrically distinct from the numrow.

There’s also this collection, which might send you in the right direction. If you can do some coding in QMK/ZMK, you might be able to make one of the BLE enabled Pro Micro clones work.

Same. But damn, she just wants to play with Legos and Project Sekai.

67g would be a good number to pop into various search engines. For the Outemus, the Oranges are sort of intended to be a slight refinement of their basic brown. I’m not surprised you found them similar feeling.

Zealio tactiles are supposed to have some pre-travel, as are “Moyu Blacks/Everglide Dark Jades”. I haven’t tried either, but it seems more common to push the tactile event to the top front of the curve than to leave it in the middle. NovelKeys Cream Tactiles may be have some weight and pre-travel as well. The force curve for Kinetic Labs Penguin also looks promising.

One outside the “box” (LOL) would be Kailh Box MUTE Jade. These are “silenced clicky” switches that are supposed to have a very distinct event in the middle but not have the sound profile that we clicky degenerates love but so many people don’t.

Even some Kalih Box clickies of his preferred weighting would be better, though I have to admit the bottom-tier Outemu “dustproof” Greens I have on some of my boards are surprisingly nice if you’re into clickies.

Nice choices! the QK seem to hit that sweet spot of Group-Buy custom features at 1/3 to 1/2 the price. I’ll get one someday, I reckon, but right now I’m having too much fun building junkboards from scratch.

Well, it came out great! I’ve always liked the concept of ortho with normal keycap kitting, though it’s kind of amusing that you went with a custom and unsculpted set for the one and only ortho that didn’t need them. Absolutely no judgment, though. I have spent a lot of time making “bespoke” boards in my garage that are barely any different from commercial offerings.

Is that texture on the case from your build plate or the fancy new topside fuzzy skin algorithms?

Hi. I have largely settled in on a pattern for making my boards, which I admit will always reveal their DIY nature when you look close, and sometimes even from afar, LOL!

1. Design layout at https://www.keyboard-layout-editor.com/
2. Import that information into the swillkb or ai03 plate generators
3. Edit the outside profile of the plate manually in 2D CAD software, and usually use that to also make a baseplate
4. Only done this once, as I usually hand-wire, but here is where I’d design the PCB and send it off for manufacturing, at least in the before-times when this wasn’t prohibitively expensive.
5. Import the plate into 3D CAD software and design a case around it. This is still a time-consuming undertaking for me, so a couple of times I’ve skipped it and just used standoffs to separate the switches and circuitry from the base plate.
6. 3D print the case and any other bits that need it, like feet or blockers or MCU shells.
7. Laser cut the plates from something that my cheap Diode engraver can get through, generally “Masonite” hardboard.
8. Install switches into the plate and solder it up; for handwires this takes an awhile.
9. Install and edit the firmware. So far, I’ve always used KMK, but at some point I’d like to move on to the more common QMK.
10. Assemble the rest of the keyboard.

I haven’t sold any DIY boards yet, but for the right customer, someone who understands the aesthetic limitations but still wants to pay too much for my time and needs something unique, I’d certainly consider it. I’m under no illusions that this is a large market, LOL.

The biggest issue is that the sockets were never really designed to be for enthusiasts changing switches all the time. They were designed for the factories to have multiple versions to sell with minimal retooling. If you are extremely diligent with removing the solder from the legs of the switches and keeping them straight, then yes, there’s nothing to prevent this from working. However, a little blob here and there will make it much more likely that you tear a pad when inserting the switch into the GMMK, and also more likely that the socket will be slightly deformed and never work quite right with any other switch.

WASD is dead. ZSA and Nuphy seem to be fine.

This one has Box Navy switches.

I said improving, not reducing! 🤣

LOL, it works for me, but undoubtedly part of it that I’m not a proper touch typist at all.

A major design element here is that no key is more than 1.75 “units”, meaning nothing needs stabilizer hardware. It’s a cheat to improving sound and definitely one for easing construction on my very cheap laser cutter (really more of an engraver, but it can get through some things). The open spaces are also meant to evoke the “HHKB” and its retro inspirations like the original Macintosh keyboard, and honestly it hasn’t been a problem. I have four “spacebars” of 1.25 u each, but two of them do something else when held down (Fn for one, Alt for the other).

The 3D-printed case could stand a little refinement, and if it ever actually cracks I’ll replace it, but so far it’s hanging in there, and I really like the typing feel.

Apart from the board itself, how do you like the G20 caps? I have a hard time imagining how laptop-flat with full travel would feel.

There’s always the Boston…

So, just in the past year or two they’ve finally released switches from new molds (“MX2A”) that are quite well regarded, but not special enough to make up for what must be very high operating costs. Unlike GMK, they squandered an early lead in the enthusiast space and never got it back. They just took much, much too long to relaize that the viable markets for mech boards are gamers and people with varying degrees of enthusiasm for the hardware itself, not just an aging cohort of office workers who refuse to put up with membranes.