Behind the scenes #2023/38

Quick news, in the mailbox, games, typewriters, meetups, updates, etc.

Hey y'all,

This week I fought Vial, (is ` some kind of a forbidden character in macros?! – at least in the online version), made some noise with a newly acquired cute (and smelly) Triumph typewriter, and of course kept an eye on interesting keyboard projects to archive.

---

Welcome back for another weekly recap and behind-the-scenes write-up.

If you're joining us for the first time, you can read how this started out and what this is all about nowadays. If you like what you see, subscribe to the newsletter (free) and donate some bucks to keep this otherwise free and ad-free project alive.

---

Reading for the weekend

• Interview: Quentin Lebastard of Bastard Keyboards, designer of the Charybdis and Dilemma, answered my questions about his work and designs before the release of the Dilemma Max.
• I tested the M4CR0Pad by Ergohaven and thought it would be interesting to demonstrate the ease of setting it up, along with a typical use case: image processing.

• Nakamura Hideki announced the testamatta #011, a nice low-pro unibody angled split keyboard.
• New LEGO-compatible board announced by KBDcraft: Lilith is also 60% (similarly to Adam), but it comes in a new case with 5.5 degree typing angle and in both ANSI/ISO flavors!

• Already mentioned last week, Maximus is a unibody split with trackpad by Protieusz – inspired by Daraku-Neko's Ieneko42c. Now with case STLs and Gerbers.
• Cyberdeck: Giga40 update with HIGOLEPC – committed by mujimanic.
• Tools: Vial to Keymap Drawer by YellowAfterlife takes your Vial .vil layouts and converts them to YAML that you can pass to keymap-drawer. And keyfab, a promising keyboard layout design tool by Jaroslaw Weber.
• Mod: notshitashi made a Glove80 with a trackpoint – with details and lots of WIP photos.

• Keyboard art – So much attention to detail: EX-00 – a heavily customized AM AFA R1 by TheExStudio. A life-size human size keyswitch out of cardboard (by i_like_miniwheats). Keebio Cepstrum illustration/map by jadepug. And this Jabberwocky with the tasty wooden wrist rest belongs to this column too I guess.
• Weekly Diderot effect: colorful replacement feet for Halo keyboards. And a slightly over-engineered suspension system with gas cylinders. ('but why?!')

• Typing game in the making: Kanji Kitchen by MochiDev. Learn Japanese while clicking and clacking – with in-game keyboards and keyboard sounds. Coming later this year on Steam.

Donations

• I most humbly beg zzeneg's pardon because he's been a donor for four months now and I missed his subscription somehow. To be precise, I haven't missed it, the cause is even more embarrassing than that, which I'm not going to disclose. ;)
• And as always, many thanks to everyone who helped this project thus far, especially to donors with recurring donations.

In the mailbox

Vintage keyboards are so overpriced, aren't they? I don't feel like spending that much money sellers are asking for however cool, unique, strange, rare etc. they are. It seems overpricing turns me to typewriters.

You can find gorgeous typewriters for dirt cheap – even beautiful, working specimens in impeccable condition. A Triumph Gabriele has just arrived:

All right, I have to confess I also bought an overpriced keyboard. In my defense, I waited more than a month for the seller to reduce the initial price in several steps. Probably rubberdome but the model is somewhat historic – part of game-dev history. It wasn't even shipped yet so more on this next week.

Typing games

It is evident that I'm not a gamer. The last game I played was Touch Type Tale – in March this year.

However, when I read that the new ZSA split (Voyager) comes with Nanotale, a typing based game released in 2021, I took another look at Epistory (2016), its predecessor which I loved for its slow pace and melancholic atmosphere. I got it for free but I've never finished it. Until last weekend! It turned out I was really close to the end of the story when I abandoned it, and contacted the developer, Fishing Cactus, with some questions about Nanotale.

And Bruno was kind enough to send me a digital copy for free. Unfortunately, while Epistory ran perfectly on my PC, it seems its resources don't match Nanotale's system requirements (and I'm definitely not going to buy a new PC for gaming). It really lags which kills the experience. Nevertheless, Nanotale is something I can relate to: the heroine is an archivist, collecting various creatures of her world, cataloging its mysteries and wonders, and writing about them – slightly similar to what I do with keyboards on these pages. ;)

Another promising game is Kanji Kitchen, which really resonates with me too – combining typing with learning Japanese? I can't wait to try it. Cute graphics, customizable in-game keyboard with nice sound. MochiDev told me the estimated release date is around the end of this year.

Meetup database

Upcoming:

• キー部 6%, Tokyo (Sept 23.)

New updates:

• Hong Kong meetup photos – missed.
• Jakarta meetup photos (Sept 16, missed too…).

As always, this meetup database is both a calendar and an archive so feel free to send me upcoming events or even ones from the recent past to make this collection as comprehensive as possible.

Vendor database

New shops and updates to the database of keyboard vendors this week:

• ori kobo (Japan) added.

Developments

• A handful of shop logos added, some optimization here and there.

---

Still with me?

Time flies like an arrow. Fruit flies like a banana.

---

That's all for today. Thanks for checking by. As always: Keep learning and building!

Until next time,
Tamás

M4CR0Pad in action

I tested the M4CR0Pad by Ergohaven and thought it would be interesting to demonstrate the ease of setting it up, along with a typical use case.

Evgeny from Ergohaven was kind enough to send me one of his M4CR0Pads. (I received it for free, only had to pay VAT and customs.) Instead of merely presenting its specifications, I thought it would make sense to write about how I set it up to aid my work, especially image processing.

Specs, build quality, design

We are talking about a 12-key macropad (3x4) which comes with a rotary encoder and a display – all this in a 3D-printed white case. In the parcel there was also a decent quality USB-C to USB-C cable with a small USB-C to A adapter. Plus two small tenting legs of the High Stakes breed. (I got some more samples of these 3D-printed legs of the tilt system, but those are not part of the default bundle.)

The look and feel is similar to – surprise! – Ergohaven's Planeta I reviewed earlier and which I've been using at home ever since.

Seeing both the black (K:02) and white cases, I definitely prefer the latter. It's really classy and the 3D printing layer marks are less obvious. The white blank XDA caps perfectly match the case, and I love the accents too.

The display's resolution is 128x64px (slightly larger than the more common 128x32 ones) which works better in the vertical position imo.

The device has a decent mass, you won't toss it around your desk accidentally. (Anti-slip bumpons on the tilting legs only.)

Specs

• 12 keys, MX, hotswap
• rotary encoder
• 128x64px OLED display
• RP2040 MCU
• USB-C connector
• QMK firmware, Vial ready

Macropads in action

Well, macropads are cute and cool props, but to say the truth I'm not really a macropad person. Or to be precise, I wasn't. Because for the sake of this test I could identify one specific use case in my workflow which is responsible for some awkward finger gymnastics every day: image processing. (Using combos like Ctrl+Shift+V/E is not the most comfortable way to do things if you, like me, have your modifiers on a split's thumb cluster.)

Of course I could have done this exact same keymap by setting up a dedicated layer on my split or semi-split boards, but actually it seemed to make sense to outsource a bunch of functions which follow in a logical order – without the need of pressing and holding a lot of layer and modifier keys during the process.

Workflow

A do image processing almost entirely in GIMP. Nothing serious: cropping and resizing images, converting file formats, export. Pretty much that's it. (Plus an ancient image viewer which I use for batch resizing to make the weekly covers, but that's not a big deal.)

As a first step of setting up the macropad I tried to identify the major steps of my workflow, along with the associated hotkeys or keypress sequences:

• (Copy or Screenshot [Alt+PrtScr])
• Paste as a new view [Ctrl+Shift+V]
• Crop [Shift+C]
• Resize [Alt+I>S, enter size, Tab, Alt+S to confirm]
• Export [Ctrl+Shift+E]
• Navigate/Create subfolder [arrows, tab]
• enter filename
• confirm export settings modal [Alt+E, Alt+E]
• Resize (thumbnail – same as above)
• Export (thumbnail – same as above, with some navigation in the folder structure)
• Close view (Alt+W, confirm Alt+D)

The exact hotkeys may change depending on your language settings, but that's not the point. What's more important is to be clear about our goals and the means to achieve them, while possibly not introducing even more issues on the way. ;)

Issues and goals

My normal workflow requires some keyboard-mouse changes (which I can't really avoid entirely), a handful of awkward key combos, and inconvenient juggling with layer changes.

When offsetting some tasks to the macropad, we can address the problem of the awkward key combos and layer changes right off the bat, however, I had to keep an eye on the newly introduced macropad-keyboard changes. Just like keyboard-mouse changes, they are very ineffective which I'd like to avoid or minimize.

The problem in the above example is: you may have to enter (newly created) folder and file names, for which a macropad is not really ideal or an option at all. Instead of changing back to the keyboard for this step, I wanted a simple solution to do this on the macropad.

Accidentally, I use number-only folder and file names for most of my images, so all I had to do was to set up a secondary number layer on the macropad (the macropad comes as a numpad on the base layer but I needed it on a different one).

Even better, if I'd like to automate the process entirely, I may stick to a single default file name (entered by a macro) and rename it later, e.g. when transferring the image to the server (while using a keyboard anyway).

Programming

The programming went like a breeze. The M4CR0Pad is Vial compatible, so all you have to do e.g. after visiting vial.rocks is some pointing and clicking. "Programming" is really an overstatement, let's call the process configuring instead.

The most difficult part, at least if you haven't done this before, may be setting up macros. I had some issues with Vial in this regard, but you should be fine: it was a result of my language settings.

Keymap

With the above functions in mind, I set up two layers. A base layer and a number layer. (The macropad can handle 16 possible layers by default.) Here is my base layer:

As you can see, there was room even for Esc and Undo. (If you're wondering about the Ctrl+Y, I'm on a QWERTZish layout.)

And below is a screenshot of setting up the almighty macro doing most of the job – and which threatens with a real disaster if I ever trigger it outside of Gimp. :D

Result

After setting this all up, I leaned back and watched with satisfaction how photos are being resized and saved on a single keypress – with modal windows popping up and disappearing automatically. The only issue I'll have to work on is finetuning delays here and there. If Gimp lags or resizing/saving of a larger file takes more time, keypresses of the macro may come too quickly.

Conclusion

The Ergohaven M4CR0Pad is a nice little buddy, very easy to set up thanks to its Vial compatibility, which can be effectively used to address specific workflows as demonstrated above.

12 keys were more than enough for my use case – actually, in the end I could set up a single key with a long macro doing the whole job described above. :D

I haven't even exploited the display and rotary encoder, but even without those, my workflow is already much more convenient.

Resources

• Ergohaven
• M4CR0Pad – available from $50.

Maximus

Maximus is a unibody split with trackpad by Protieusz – inspired by the Ieneko42c.

As already mentioned last week, the Maximus by Protieusz, inspired by Daraku-Neko's Ieneko42c, is a 3D-printed monoblock split with a trackpad, plus 2 extra physical mouse keys (even though the trackpad can do gestures).

In the meantime Protieusz also published the STLs and Gerber files of this remix, which is powered by a Raspberry Pi Pico and is equipped with the 65 mm variant of the Azoteq trackpad.

The 65mm trackpad makes a big difference compared to the Ieneko42c's 43mm trackpad. Operation is much smoother – Protieusz.

Specs

• 38 keys (3x5+3 per half + mouse keys)
• MX, soldered
• SMD diodes
• Raspberry Pi Pico controller
• 65 mm Azoteq trackpad

Resources

• https://github.com/protieusz/Maximus

Testamatta #011

Nakamura Hideki announced the testamatta #011, a nice low-pro unibody angled split keyboard.

Nakamura Hideki (aka hide-key) from Yokohama shared his latest project: the monoblock split testamatta #011 with exposed diodes and a XIAO controller.

After smaller projects like the Christmas ornament keypad and Lion head macropad, as well as the ortholinear Testamatta keyboard #010 with the cool homemade keycaps, the #011 is a more ergonomic unibody – low-profile Choc switches with MX spacing.

Specs

• 46 keys (6x4-ish per half)
• low-pro Choc switches
• exposed SMD diodes
• Seeed XIAO controller

Resources

While not open-source, you can find the Gerbers in the repo:

• Gerbers and firmware.

Interview: Quentin Lebastard

Quentin Lebastard of Bastard Keyboards, designer of the Charybdis and Dilemma, talks about himself and his designs on the occasion of the upcoming release of the Dilemma Max.

Quentin Lebastard of Bastard Keyboards doesn’t really need an introduction. As designer of keyboards like the Dilemma and all the Charybdis variants, he not just sells his pre-built boards and kits but also regularly shares his source files. I asked him a few questions on the occasion of the upcoming release of the Dilemma Max.

Could you tell us a bit about yourself?

Hey! Great to be here. I'm Quentin, a French keyboard designer operating from The Netherlands.

Is Bastard Keyboards your full-time job now?

I've been working on this project for 4 years now, and I'm extremely thankful to be able to call this my full-time job. What started as a single-desk keyboard building service is now a full-blown office, constantly buzzing with prototypes and ideas.

Sure, a lot of what I do can be considered work – community management, inventory, emails – but on most days it still feels like having fun and coming up with ideas. It's never boring, and I'm going to the office every day with optimism and enthusiasm.

I'm also working with Bas now, who at the start was helping on the operation side of the company, but is now also working on some cool upcoming creative projects.

How and when did this all started out?

Five years ago I was working in IT and, like most people in that field do at some point, started developing RSI issues. Back then I had already started to get into mechanical keyboards – and had a Ducky One TKL with blue switches. I then realized that I needed to get a bit more serious about my health and started to look into ergonomic keyboards. The Dactyl seemed like the best option back then, and I had just gotten a 3d printer that I wanted to put to use… From there, I built a few of them which helped tremendously my RSI, but I was not happy with their design and build experience.

A lot of people get into the hobby but few of them sell their own products. How did you end up as a keyboard maker?

I would describe the process best as organic. In September 2019 I moved to The Netherlands on what was supposed to be a short holiday to see some friends after quitting my job. Back then I was very active on Discord and Reddit publishing pictures of my design experiments, and it was gathering a lot of interest.

The main issue with building Dactyls has always been their complexity – they require a good knowledge of 3d design and 3d printing, electronics design, wiring, QMK. It involves multiple topics and can become overwhelming very quick. Because of that, prebuilt Dactyls were appealing to a lot of people, and I saw that as an opportunity to help them.

From there on, I went through a lot of different ways of building them – first handwired, then using Amoebas, then flex PCBs. First with models generated through online Dactyl generators, then with my own designs. First with handwired Pro Micros, then with custom-designed RP2040 boards with shield PCBs.

Whenever I ended up with new ideas, new designs, the main goal was to help others. For this, open source made sense – but for those not willing to "get their hands dirty", selling prebuilts and kits is a way to help them while financing my full-time work on innovation.

I guess there were some bumps in the road.

Designing keyboards ("maker") is not related at all to selling them ("shop"). Selling products entails very different skills and topics, like communication, inventory, versioning, after sales support, packaging, build guide editing, you name it.

It's not for everyone and brings its own challenges – and it takes out a huge chunk of time that could otherwise be used for designing. Working on this as a team rather than just myself has been very helpful, and it enables my work on design so it's very much worth it.

Working in a field that's so new makes things even more complicated – standards have to be worked on, and any decisions taken upstream can have consequences on designs for years as I try to make everything as backwards compatible as possible. Because all of the work until now around commercially sold Dactyl-likes (Kinesis, Datahand…) had been kept closed source, I had to reinvent the wheel (and share the process). There has been countless projects around Dactyls, but nothing with a quality high enough that I was happy with – I wanted to take it from a "weekend project" that you have to tinker with and debug a lot, to a more streamlined experience that can also be reused in other designs.

Designing and building keyboards for others also brings the quality requirements a lot higher – while I can afford a few quirks on designs that I use myself, if I ship to others the quality standard needs to be there. For this I’ve always focused on starting things very simple, and from there slowly upgrading – the Dilemma is the most recent example: I started with a very barebones 3x5+2 that brought only one new thing (the trackpad), and from there slowly upgraded it (integrated mcu, rgb, rotary encoders, more keys…).

Could you draw a quick timeline with all the notable keyboard models you designed?

I don't have the exact dates on hand, but the main events that defined my work would be:

The original Scylla design. This was my first Dactyl design in fusion 360, that came out of wanting a better looking Dactyl that was easier to print, build, and maintain.

The Skeletyl: this started as kind of a joke, as I never expected a keyboard with so little keys to become so popular. I wanted to experiment with new design techniques and came up with the skeleton design method, which was also quickly adopted by other designers. This was the first design I open sourced, and the others followed quickly.

The Charybdis: this was… a challenge. Back then there were a few existing builds (e.g. Tractyl manuform), but they required obscure PCBs and a lot of tinkering. I had to design a sensor PCB from scratch, as well as a reusable 3d holder assembly for it. This was a bet that worked out – the parts are now reused by a lot of other designs. Not only this, but the code in QMK back then for trackballs was… fragmented. Drashna and Charly put in countless hours to make this a better experience, and we now have code directly in QMK's drivers with a ton of neat features

The Splinky: this is not exactly a keyboard, but it's the "brain" of all my 3d designs. I started using Pro Micros and Elite-Cs, but Corona hit and made those 32u4 extremely expensive to source. With the discord community and other vendors, we worked on a standard pinout for the new, cheaper RP2040 chip by Raspberry (RP2040_CE), which means all those new powerful RP2040 boards can be used interchangeably.

The Dilemma: I've wanted to work on flat keyboards for a while and already experimented with a Sweep-like and Atreus-like. I wanted something that could be used on the go (when bringing a Dactyl is complicated), and also brought something new to the table (trackpad), all while not being too alien that it would scare adopters. It worked out well, and the new versions are bringing more and more features.

What aspects of the hobby and manufacturing do you like or are you interested in the most?

What I like so much about my job is that it involves so many different aspects. When I started 4 years ago I was mostly invested into the 3d design part, and the last year it’s been mostly electronics.

I try to keep things fresh – at the moment I’m trying to invest a bit more into photography and video (for guides and tutorials). There’s nothing super concrete yet, but it’s important for me to keep going out of my comfort zone.

Recently I had to design the silkscreen for the Dilemma Max and decided to learn Illustrator for it – it was a real challenge as I was also on a tight timeline, but it turned out great and I’m now using this tool in other parts of my designs as well.

My most favorite part by far though is the community – there are so many crazy ideas out there, and working on open-source keyboards means I can be part of it and contribute to it. Seeing mods and forks of my keyboards is the best reward there can be.

Your recently announced product is the Max variant of the Dilemma. It seems to be a natural next step in the line-up. Was this a community request?

This was definitely a community request. In all my keyboards, I have 4x6, 3x6 and 3x5 variants. When I started developing the Dilemma I decided to go with an “extreme” package, 3x5, just like the Skeletyl. This was because it enabled me to do some rapid and cheap prototyping compared to something bigger and more complex. I had only one big objective – produce a modern split keyboard with a trackpad. While the trackpad part worked well, I had to sacrifice other parts (rgb, integrated components) just to keep things simple.

When looking at my 3d keyboards, the big format (4x6) is the most popular – and that’s understandable. For someone who has little to no experience in splits, going with a 4x6 means it’s a less-big of a jump. It’s less intimidating, and having a number row feels familiar compared to homerow mods and multiple layers. It definitely provides a less intimidating way of getting into ergo. This aligns with my goals – help people find a keyboard that works better with their body. Having a bigger one means they are more tempted to try it, and maybe later switch to the smaller version.

Because of this, the demand for a bigger Dilemma was there – and with the experience I got designing the Dilemma I felt confident I could go ahead with a bigger design.

Bringing a real product to the market is much more difficult than creating a prototype for yourself. What are some possible challenges which are not so obvious for the uninitiated?

That is definitely true, and something that’s hard to fully grasp unless you are also managing products. Like I mentioned previously, the quality requirements are much higher. Rather than sticking to very general concepts, I thought it’d maybe be interesting to give a few concrete examples.

Technical: When I first started working on the original Dilemma 3x5, trackpads in ergo keyboards were still a very new thing. After settling for the Cirque trackpad, I had to find a way to connect it – either soldering a few bodge wires and desoldering a resistor, or using an FPC connector. Having to desolder a resistor and bodge a few wires is fine when you’re designing a one-off for yourself, or even for enthusiasts, but I wanted something truly easily accessible. Because of that I decided to switch to an FPC connector, which meant ordering the boards assembled from the factory rather than bare. This brought a lot of issues and design complexity, but in the end made builds slightly easier.

Quality requirements, and where to stop. Going into the Dilemma Max, I felt confident because the Dilemma was already a polished product. Every time I make a new product I go through 3 or 4 revisions, and this time was no stranger. While things went overall well, I had a few issues with components placed too close, or zone patterns being different because I overlooked them. For some builders or vendors that’s in the acceptable range of defects, but not for me. You need to draw a line on where you decide a product is good enough to ship, and even if high that line needs to be there.

Firmware and all technicalities. However, hardware is not everything. For a product that works great out-of-the-box, you also need great firmware. Because I work on innovative products this has always been a problem – there was no standard code for the trackball when it came out, and very little features.

I’m extremely lucky to be in touch with amazing volunteers who write and maintain the code for all my keyboards. This means not only developing code for either new features or new keyboards, but also updating as the hardware changes (eg. with the move to RP2040 or to Via).

To ensure a smooth experience in both hardware and firmware, I have the habit of shipping early versions to enthusiasts and developers. While they get an “imperfect” version of the product, they always give some invaluable feedback and start working on the firmware. This enables me to find issues I’d otherwise either not see or much later, and get a firmware that’s much better than if I’d had to do it (I suck at QMK).

Diagonal stuff. Finally, because I’m creating keyboards that are going to be built by multiple people with different backgrounds in DIY, I need to make sure the process goes smoothly. This means that during the design process, the 3d and electronic parts have to be made in a way where it’s expected there will be mistakes. It can and will happen that the screw inserts are not installed properly, the components are installed in reverse, there are shorts and cut traces and burnt plastic. People will make mistakes, and I need to work on making those mistakes less likely to happen, but also make sure that the keyboard is easy to fix.

From the start, I needed to do some choices which don’t necessarily make a lot of sense for an expert (eg. big screws, using Splinkys, reversible PCBs, ribbon cables instead of FPC…), but these make the build process easier and less prone to mistakes.

On top of this, there’s of course the build guides – another aspect that comes on top of the simple design of the keyboard, involving photography and editing, video recording and script writing. In my experience, creating a build guide really puts all the little details in the spotlight – if there’s some part of the build that’s awkward, suddenly you have to explain it in details and it puts you in front of your design choices.

What’s your current setup? Favorite layout, switches, keycap profile?

When typing this I’m using a Charybdis Nano, nylon, metal plates with magsafe and tripod adapter, tangerine switches, and DES caps. For some time I was using a wireless CNano with trash switches, but it kept having issues with connecting to my Windows machine and the trackball dropping out, so I switched back to my old design.

I’m a fan of light linear switches, so I might find something even lighter than the tangerines for my next build. I don’t plan on moving away from DES though – shoutout to Asymplex, those are the most comfortable keycaps I've ever used, and his new thumb cluster for the Skeletyl is killer.

When traveling, I use a Dilemma with sunset switches and MBK keycaps.

What’s your favorite keyboard designed by someone else? ;)

I’m a huge fan of the designs by Geist, specifically the Totem and Klor. One of the most difficult things you can do as an artist is create something that can be appreciated even by the uninitiated, and in that aspect he absolutely nails it. There is so much attention to every detail, and an aesthetic that is carefully crafted, and it really shows in the final build – no detail is left to luck. I also got the chance of meeting him in Leipzig, super chill guy.

What’s next? Any upcoming projects or future plans you can tell us about?

In the future I want to start working again on metal plates for tripods. I’m also still working on upgrading the build guides, and working on transport cases. Very long-term, I’m still slowly working on adding bluetooth to our offer. Bas is also working on epoxy DES caps, to add it to our offer.

Final thoughts?

Come check the Discord! We’re a friendly bunch :)

Resources

• https://bastardkb.com/
• Discord
• Github