The key cluster molds are being made

Hi there, and welcome to the monthly UHK status update!

TL;DR: Key cluster mold manufacturing is well underway. We’re getting close to the final design of the right side module molds, too.

Since our last monthly update, the design of the key cluster molds has been finalized, and now they’re under production. It’s always quite an experience to witness hunks of steel taking shape.

In the meantime, András and our contractor have been working hard to finalize the design of the right-side modules. As soon as this design gets finalized, our contractor will start to get these molds machined.

We’ve also ordered various parts, such as the optical sensor of the trackball, the mini trackball of the key cluster, and the FFC cables. This is quite an intense process for everyone involved, and we’re doing our best to hit our July ETA.

Feel free to change your shipping address any time.

Your tweets

You guys keep sending your awesome tweets, and we're always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2020-05-11.

Module ETA announcement

Hi there, and welcome to the monthly UHK status update!

TL;DR: We expect to start shipping the modules in July, and we’re making rapid progress.

As mentioned in our previous monthly update, there are a number of tasks to be done to mass produce the modules, but the most time consuming of all is the creation of their molds. The good news is that our mold making supplier just provided us an ETA: June 2020. This means that we should be able to begin shipping orders in July, likely completing all orders in August.

Our supplier is almost ready with the design of the key cluster module mold:

András has been working a lot on the 3D models of the modules recently, and all the modules should be finalized soon.

In the meantime, I’ve been busy optimizing the PCBs of the modules for manufacturing. It’s not rocket science, but there are a ton of details involved, and small tweaks can go a long way.

We’ve also started the procurement of the parts for the modules. Most of them are not very urgent, but there are exceptions like the optical sensor of the trackball module which can take up to 14 weeks to arrive. Luckily, we’re just in time to hit our ETA.

We’re knee deep into getting the modules manufactured, everything is going well, and we’ll be keeping you updated on a monthly basis as usual. You can change your shipping address any time.

UHK ninja belt

Special shootout goes to Hubert Łępicki who amused us with his UHK ninja belt (for the lack of a better term).

Hubert wouldn’t be a true ninja without demonstrating his belt in action. Check out the following video.

3rd party keycap sets

Some of you are on a never ending journey of pimping up your UHKs, and we’re excited to see your creations!

Further tweets

You guys keep sending your awesome tweets, and we're always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2020-04-10.

Every module prototype is functional

Hi there, and welcome to the monthly UHK status update!

TL;DR: The trackpoint and touchpad modules are functional, and we’ve got demo videos. The mouse mode of the UHK is usable to create digital art. Agent had a Linux security bug, so it’s worth upgrading.

Module demos

This update is quite late, but I think it’s the coolest one so far. I’ve been knee-deep in the firmware of the modules recently which bore some delicious fruits… I mean functional modules.

The trackpoint turned out great, and it’s very precise. Thanks to the dexterity of the thumb, I find the usability of the trackpoint module to be excellent despite the unusual trackpoint nib position compared to other trackpoint implementations.

The touchpad is also finally usable now. It’s excellent for swift, long distance movements, but it has to be optimized a bit more for precision movements. The touchpad IC has numerous parameters, so my hopes are high that precision control can be sufficiently improved purely on the firmware side.

For the sake of completeness and comparison, see the trackball module demo, too. This demo was featured in our previous update.

As you can see, all of the four module prototypes are functional. This is a huge milestone as far as the modules are concerned. Now that we have functional prototypes, it’s time to mass produce them.

Mass production will take a number of steps such as creating tooling, procuring parts, doing EMC tests, and minor manufacturing optimizations - just to name a few. Luckily, many of these steps can be done in parallel, but they’ll still take months. We’ll announce an ETA in our next monthly update.

Creating digital art with the UHK

Speaking of mouse control, although not as capable as the modules, let’s not forget about the venerable mouse layer of the UHK, as it’s very powerful for what it is. So much so that apparently it’s possible to create digital art with it. Give it up for Brandon Yu, who’s about to demonstrate the seemingly impossible.

I don’t know about you, but I’m officially blown away by Brandon’s skills and talent. Brandon also happens to be a game developer, so feel free to get in touch with him on Twitter.

New Agent features and security fix

Bill Gates used to say that 640K ought to be enough for everybody. I’m here to say that 32K is enough for every UHK user. Well, as far as the on-board memory of the UHK goes.

Now that Agent shows the allocation of the on-board EEPROM memory of the UHK, it’s easy to see that about 4K is used by the default configuration of the total 32K. Even if you have dozens of keymaps and macros, it should still be enough.

If you take another look at the above picture, you can notice the newly added configuration history section. Every time you save your configuration, a new entry gets created here, and you can restore any previous configuration with a click of a button.

Last, but not least, we’ve fixed a major security bug which affected the Linux versions of Agent. We suggest installing the latest Agent version which will fix the affected udev file upon starting it. Mad props go to Joel Eriksson of ClevCode for spotting this one!

Your tweets

You guys keep sending your awesome tweets, and we’re always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2020-03-10.

Controling the mouse with keyboard keys

Although we never advertised the UHK mouse layer as a perfect substitute for a mouse, it's surprisingly powerful in the right hands. So much so that apparently it’s possible to create digital art with it. Give it up for Brandon Yu, who’s about to demonstrate the seemingly impossible.

We don’t know about you, but we're officially blown away by Brandon’s skills and talent. Brandon also happens to be a game developer, so feel free to get in touch with him on Twitter.

Want to bring it up a notch? See the module demos.

The trackball module is fully functional

Hi there, and welcome to the monthly UHK status update!

TL;DR: The trackball module is now fully functional. Modules will be more portable than expected. Some great UHK reviews have been published.

Trackball module demo

I really wanted to show up the trackball module in the previous UHK newsletter, but couldn’t because the pointer moved in a very erratic fashion. I’m happy to report that I’ve managed to fix the problem, so here comes the obligatory demo.

We’re very pleased with the usability of both the trackball and key cluster modules. They allow for fine-grained mouse control without leaving the home row, and work just as envisioned. The mini trackball of the key cluster is very handy for scrolling, and the trackball module is a great tool for moving the pointer.

Although the schematic of the trackball prototype contained minor errors, the fix mostly involved changing a couple of lines of the trackball module firmware. This was followed by the refactoring of the trackball module firmware, specifically, and then a massive refactoring of the module firmware codebase in general. As a result, the source code of the modules shrunk to merely 100-200 lines per module, allowing me to develop the firmwares of the remaining modules with minimum effort.

I’ve already started to develop the firmware of the trackpoint module, and the new touchpad PCB has just arrived, ready to be assembled. I’m looking forward to making these modules work, too. We’ll be keeping you updated.

Module portability

The modules make the UHK less portable in the sense that one can’t just merge the halves and pick them up with one hand, but as it turns out, even with the modules, portability is exceptionally good for a modular keyboard.

It spontaneously occurred to us when playing with the modules that the left and right modules merge, just as the keyboard halves. This isn’t surprising given that the keyboard halves merge, and the shape of the modules is the inverse of the keyboard halves.

Then it also occurred to us that, given that the back of the modules feature magnets and magnet counterparts just as the UHK does, the modules and the UHK can be stuck together.

This allows for great portability despite the modular nature of the UHK, and as a bonus, it’s quite a conversation piece.

UHK reviews

LearnCode.academy has made an outstanding video review of the UHK, which we were quite flattered by. Will has praised many facets of the UHK, including its build quality, feature set, and ease of configurability.

Another spectacular review has been made by TechPowerUp. This one is extremely in-depth, contains tons of high quality pictures, and doesn’t only explore the outside of the UHK but its inside, too.

Your tweets

You guys keep sending your awesome tweets, and we’re always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2020-02-10.

Trackball and touchpad module progress

Hi there, and welcome to the monthly UHK status update!

TL;DR: We’ve made progress with the trackball and touchpad modules. Some outstanding UHK reviews have been published.

Trackball and touchpad module progress

I recently thought I’d finished the firmware for the trackball module and it would be ready to demonstrate for all of you. I can give you a demonstration, but it’s not what I was hoping for.

As you can see, the pointer moves in highly erratic fashion, and it heavily drifts toward the top right corner. I can affect the pointer by moving the trackball toward the bottom left corner, but the erratic movement usually overpowers my movements. I believe that this is a side-effect of insufficient filtering on the PCB, but it’s really just a wild guess on my part, and I’ve contacted PixArt, the manufacturer of the optical sensor who will help.

As for the touchpad, we’ve increased its resolution and sensing area.

The big footprint is for an optionally solderable 2x20 pin 0.1" connector that directly mates with the Azoteq CT210A-S dongle for configuration and diagnostics purposes.

The new design is a clear improvement over the old one, but we still have some work to do based Azoteq’s new feedback. The ground must be routed in a very specific way to maximize sensor performance. It’s not rocket science, just finicky. István is working on the next PCB version.

Next up, I’ll be working on the firmware for the trackpoint module, and will implement the suggestions of PixArt to make the trackball module work flawlessly. I hope that I’ll be able to showcase at least one, but hopefully both of these modules in our next monthly update. The key cluster module prototype is already fully functional, and the touchpad module will come last.

UHK reviews

Quite a few UHK reviews have been made recently. Let’s take a look at them!

ShopzadaPH has made an extremely thorough review featuring an unboxing, the keycaps, layout, switches, and configuration of the UHK. He really left no stone unturned, and the production quality is great. This one is highly recommended if you’re considering purchasing a UHK.

Thomas Ran is the king of the hill when it comes to mechanical keyboard reviews on YouTube, and he’s just reviewed the UHK. In his review, he emphasizes that he’s not an ergo guy, and yet, he’s done a great job of explaining the fundamentals of ergo boards and reviewing the UHK.

I met Christian Bäuerlein in 2018, as he's the organizer of Mechanicon, the meetup for mechanical keyboard enthusiasts in Germany. He recently offered to review the UHK, and given his formidable experience with mechanical keyboards, I gladly said yes. I think he’s done a great job of capturing the essence of the UHK in his review, including unboxing, assembly, layout, configuration, switches and keycaps.

Hardware.Info may well be the largest testing lab for computer hardware, peripherals and consumer electronics in the Benelux, which shows in their UHK review, as it’s so detailed, it’s almost like a mad science experiment. Thanks to the magic of Google Translate, English readers can read the translated English review, or read the original Dutch review.

Your tweets

You guys keep sending your awesome tweets, and we’re always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2020-01-10.

Key cluster and trackball module progress

Hi there, and welcome to the monthly UHK status update!

TL;DR: The key cluster module is now fully functional, and the trackball module is partially functional. Some awesome UHK reviews have been released.

Key cluster and trackball module progress

In our previous update, we shared a video about the latest key cluster module prototype. It worked reasonably well, but the responsiveness of its mini trackball left a lot to be desired. I’m happy to report that we’ve fixed this issue as you can see in the following video.

(UHKs are not backlit yet. We'll release a backlight upgrade kit at some point, and future UHK hardware versions will be backlit.)

As it turned out, the mini trackball only works well with bipolar hall-effect sensors, and now that we’re using the right kind of sensor, it picks up every tiny movement just as expected. I’m very pleased with the usability and responsivity of the mini trackball, and it feels like a great tool for scrolling in all directions.

As for the trackball module, I assembled the new PCB and gave it a try. As seen in the above video, the buttons work as expected. I haven’t yet implemented the most important part, which is the communication with the optical sensor IC - this will be my next task.

I’d like to make every module prototype work by the end of the year. The key cluster module already works well. The electronics of the trackball and trackpoint modules seem fine so far, and I have yet to write firmware for them, so they should be in close reach. The touchpad PCB is expected to be redesigned soon according to Azoteq’s instructions, so it may take a bit more time.

Following the prototyping phase of the modules, mass production will take additional time. We don’t have a solid ETA on the modules yet, and we’ll be keeping you updated.

UHK reviews

Minimalistik is a powerhouse when it comes to keyboard reviews, and he’s just made a UHK review. I believe he’s captured the essence of the UHK very well.

The Techne has also reviewed the UHK in his contagiously enthusiastic style which I particularly enjoy.

As an internationally known strength coach/personal trainer, Bernd Stößlein is not our typical customer, which is exactly what makes his UHK review so interesting.

Your tweets

You guys keep sending your awesome tweets, and we’re always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2019-12-10.

New Agent release and module progress

Hi there, and welcome to our monthly status update!

TL;DR: We’ve released a new Agent version after a long time without changes. We’ve made a functional key cluster module, and made progress with the trackpoint module.

New Agent release

It’s been a whopping ten months since we released the latest Agent version. We’ve actually been working on Agent since the latest release quite consistently, but weren’t able to publish a release due to the lack of a valid Windows Authenticode certificate. Long story short, we finally have a certificate, and recently released new Agent versions.

Feel free to check out the changelog on the GitHub Agent releases page. We’ve mostly fixed and polished a number of issues. A particularly useful feature is a dedicated Mac pointer speed preset which you should try out if the UHK mouse pointer movement feels slow on your Mac.

We’ve also implemented the fanciest UHK feature to this day: Agent shows whether the UHK is split or merged, and displays whether the left half is connected. Obligatory demo follows:

(UHKs are not backlit yet. We'll release a backlight upgrade kit at some point, and future UHK hardware versions will be backlit.)

Admittedly, this feature is pretty useless in itself, but it’ll actually be useful in the future. The same mechanism will be used to show the modules. Imagine connecting your modules, seeing them show up, and be able to configure them with a click of a button. And talking about modules...

Functional key cluster module

After a fair deal of prototyping, the key cluster module actually works. Again, obligatory demo follows:

You probably noticed the little thingie at the bottom of the key cluster module.

As you can see, it’s an FFC cable. Our current FFC cable manufacturer couldn’t make a cable of merely 13 mm length, so we used a much longer off-the-shelf cable for the time being. I actually doubt whether an FFC cable of such short length can be made, but an FPC (flexible printed circuit) can surely be made. But we’ll probably end up using a rigid-flex board as the best solution.

Apart from the above slight change, there’s another issue. I noticed that the responsiveness of the trackball is lacking compared to the previous prototype. The new, smaller hall-effect sensors are probably not sufficiently sensitive to pick up all the magnetic state changes of the mini trackball.

I wired the old mini trackball breakout board to the key cluster trackball board to be able to test it with the key cluster module, and the change in responsiveness was immediately apparent. The right board is super responsive, and the left one skips the beat very often, especially when moving it quickly.

I think we’ll revert to the previous hall-effect sensor, and try to pack them tighter to be able to fit them on the board.

Trackpoint module

We’ve made progress regarding the interconnection of the top and bottom trackpoint boards. There isn’t enough space for an FFC connector on the top trackpoint board which contains the actual trackball sensor, so the cable needs to be directly soldered to it. I designed an FPC for this purpose, and we plan to use hot bar soldering to affix it to the top board.

I used a soldering iron for prototyping purposes. So far, so good!

This module should work well, and I’m excited to write firmware for it, and for the rest of the modules.

UHK unboxing video

ShopzadaPH has made an awesome unboxing video of the UHK which you’re welcome to watch:

Your tweets

You guys keep sending your awesome tweets, and we’re always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2019-11-11.

Module PCBs assembled

Hi there, and welcome to our monthly status update!

First things first, we’ve finally delivered every UHK, including the ones with non-black cases! If you want to order a UHK, possibly with a non-black case, now’s the time. New orders ship in a week.

As for the modules, we keep making progress. Each of the module PCBs has been assembled since our last update.

Key cluster

The main key cluster PCB didn’t create any unexpected surprises.

His little brother, the key cluster trackball PCB, however, has proven to be more difficult.

The surface area of this board is merely 1 square centimeter or 0.15 square inch. This is not a whole lot of room for a dozen components, and we can’t make it larger due to the lack of space.

The PCB fab kept rejecting this board because the components were too close to the edges. The only solution was to use smaller magnetic sensor (which detect the rotation of the mini trackball). 

We ended up using a sensor with the HVSOF5 footprint, which is ridiculously small compared to the other components we use. I’m waiting for this part, and it’ll be quite an exercise to solder it. But I’m ready for the challenge, and so is my new microscope.

Trackball

The trackball board came out pretty nicely. It’s also completely useless, unfortunately.

While soldering the parts, I realized that the pinout of the ADNS-3530 optical sensors is backwards. The datasheet features the “top side” of the component, but as far as we’re concerned that’s the bottom side. Unlike trackballs, regular mice probably use this component in a reverse-mounted fashion, hence the misunderstanding.

Istvan has redesigned this board, and it’ll be fabbed soon.

Trackpoint

There’s not much to say about the trackpoint PCB. I haven’t found any issues yet. I think it’ll work well once I write firmware for it.

Touchpad

The touchpad PCB has turned out pretty well, too.

This board is hardly our final iteration though. Azoteq, the manufacturer of the IQS572 touch sensor IC that we use, reached out to us and offered to review our design.

They made suggestions on how to improve the ESD immunity of the design, but perhaps even more importantly, they suggested to use a finer pitched sensor matrix to improve the resolution and increase the sensing area of the touchpad.

We plan to implement their suggestions, and then we’ll send some samples to them for further review. The sensor IC is very sophisticated and has a myriad of configuration parameters, so having it tweaked by Azoteq will be extremely helpful. Azoteq provides the best support I’ve witnessed so far, and I’m very impressed by them.

As previously mentioned, we don’t have an ETA on the modules yet. We keep making progress, and we’ll announce the ETA when the time comes.

Your tweets

You guys keep sending your awesome tweets, and we’re always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2019-10-10.

Module PCBs are ready

Hi there, and welcome to our monthly status update!

István, our PCB designer, has been on steroids, and he finished the PCBs for every module! The boards are being fabricated right now, and are expected to arrive in a week - at which point I’ll assemble them.

We already showed an inside look of the key cluster module in an earlier post, so this time, I’d like to showcase the right-side modules. I’ll feature three images per module: the assembled version, the half-assembled version, and the latest PCB which is being fabricated.

Please note that the following modules are only prototypes. Their color is not representative, and neither is their surface quality, which will be way smoother once the modules get injection-molded. The color of the PCBs will also differ, as we’ll use black soldermask for the final boards.

Trackball

The trackball only has a single PCB. It utilizes the ADNS-3530 optical sensor, which happens to be the most compact optical sensor according to my knowledge. The retaining ring can be removed by rotating it counter-clockwise, so one can easily clean the ball.

Trackpoint

The trackpoint is composed of two boards. The top board is provided by our supplier and contains the actual trackpoint module. The bottom board is designed by us, and its purpose is to do protocol translation between the PS/2 protocol of the trackball PCB and the I2C module protocol of the UHK.

Touchpad

The touchpad module is composed of two boards. The bottom board is a trivial one which simply routes the pogo pin header to an FFC connector, supplying power and data to the top board. The top board does the actual sensing using the Azoteq IQS572 touchpad sensor IC. The top side of the touchpad will be covered by black film.

When I said that that the boards are ready, what I really meant is that these boards should be fully functional. Their design is not set in stone yet, but we expect only very minor changes going forward. Even our mechanical design is fairly advanced and should contain the mechanical features needed for injection-molding.

As I previously mentioned, we don’t have a solid ETA on the modules yet. As you can see, we’re making rapid progress, and we’ll get there, but we surely won’t rush them, as we want to get them right.

Your feedback

You guys keep sending your awesome tweets, and we’re always eager to read and feature them! If you got your UHK, please share your love!

We’ll be keeping you updated on all things UHK, and are looking forward to talking to you on 2019-09-10.