My DIY keyboard collection ( or how I became a KB-geek...)

[nisilhum]

Suka interisting what is the layout of your hypernano?
it looks fit 4 fingers and modifiers in addition pinkie

[suka]

Indeed, good to know it's not only possible but well tested. No issues with different host OS or rollover?

No issues with the trackpoint at least on Linux and several Windows versions since more than two years.
Rollover? Cannot say anything negative about, but I hardly press more than 2, very seldomly 3 keys at once and do not game with my custom boards so I would not notice issues during normal use. But with every key sporting a diode the maximum number of keys detected is only software limited by the default USB report size of 6 chars ( and that can be fixed for full NKRO as some other firmwares demonstrated, but I don't really see a point for it).

Suka interisting what is the layout of your hypernano?
it looks fit 4 fingers and modifiers in addition pinkie

It used to be something like this, index fingers on A & T. I think in the current version I shifted the thumb keys one column to the outside on each side

HyperNano normal alpha layer

HyperNano_L1.JPG (35.22 KiB) Viewed 20944 times

[nisilhum]

It used to be something like this, index fingers on A & T. I think in the current version I shifted the thumb keys one column to the outside on each side

HyperNano_L1.JPG

my bald yeah it's make sense now thankx

[Ladybug]

Thanks for posting! I am looking for a programmable keyboard which also allows me to try custom keyboard shapes and then choose my keyboard layout key by key in an exisiting program.
It would have 30 cm insulated wires between the keys and the PCB so I could arrange them on a wooden frame or one made out of gorilla plastic or some other material.

Or do you know where I can buy a board which is ready to use except the connection between the key switches and PCB? Is it feasible ( unbedenklich ) to have a lengthy wired connection beween those components (resistance issues)? I can solder so if it is technically possible I could solder the wired connection myself. Do you know of any group buys with those components?

I am a doctor and would like to contribute my medical knowlegde towards reducing impingement syndrome and rsi by experimenting with new shapes of keyboard that haven't been tried yet...
I would greatly appreciate your responce.

[suka]

Where's the working keyboard hidden?

More than just a teaser soon to follow...

[hypkx]

I read the whole thread, nice projects/keyboards.

[Ray]

Where's the working keyboard hidden?

Behind the Edelstoff?

Foldable case for transport? I can't wait for more than the teaser.

[suka]

Well, you got the case right at least And when did I mention something foldable is around the corner? And a case has already been made, so why not make _all_ columns flex^K - aargh, already getting ahead of myself...

The venerable RedTilt prototype has been my daily driver for over 3 1/2 years now, and I have nothing at all to complain about the general layout or functionality. And as you can see from the lack of updates to this thread there simply wasn't any good reason to continue the search for my personal holy grail of keyboards as the minimial tilted split board feels perfect in every regard to me.

But the prototyping character without a case and clear view (and access for dust and grime) of all its hacked up innards has been bothering me for a while even though the DIY appeal really shines here. Also, the improvised stand on screws and missing bottom shell made it hard to stow away or transport. And last but not least, the 2nd column on the pinkies has been of no use to me with my custom layouts, so finally I pulled myself together, started OpenSCAD once again, and was immediately up in Full Make Mode (tm), thinking up wildest options of ergo thumb clusters or re-thinking about the bowl-shaped experiments from my early phase.

In the end, though, I managed to distill my previous experiments into two evolved versions of my beloved RedTilt design, and one super secret next-gen combination of both: One minimalistic version, whereby the individual tilting of the columns has been given up in favour of an absolutely clean and transportable design, and a second one that takes the RedTilts ergonomic ideas to the next level by also rotating the rows, thus creating a bowl like shape.

So here they are, without much further ado:

Codename BlackFlat, to keep the senseless prototype names

IMG_20161120_145239899_HDR.jpg (874.01 KiB) Viewed 20285 times

Just as I am typing on in now: Trackpoint support is not yet connected, and I am commited to not keep the fixed cable for the next 3 1/2 years but order TRRS-Connectors or something similar with some port expanders next week. Matching supports are still missing and the connecting piece is way too high, but that will be easy to fix.

Tried to cleanly layout the wires this time to achieve minimal height

IMG_20161120_001543722.jpg (1.31 MiB) Viewed 20285 times

Inside of prototype and comparison to RedTilt

IMG_20161118_175152638.jpg (699.23 KiB) Viewed 20285 times

Upcoming bowl shaped variant

IMG_20161118_175102355.jpg (705.94 KiB) Viewed 20285 times

One more thing:
Are hinged boards my new 2017 project? Or is there something cooking already? Let's see...

Sneak peek 1 SCNR

IMG_20161115_214451216.jpg (787.17 KiB) Viewed 20285 times

Sneak peek 2

IMG_20161115_214816889.jpg (758.83 KiB) Viewed 20285 times

Apart from the trackpoint that still needs to be hooked up everything is working as expected, and the typing feel after only the few minutes this message took is not much different from the RedTilt I am used to. I guess with the column layouts I found a perfect match to my fingers, and the flexibility of a split board takes care of adjusting to different conditions in the workspace.

[suka]

Testing trackpoint with external reset circuitry instead of GPIO as previously used.



Everything wired and correct multimeter readings, but no cursor... Time for the logic analyzer I guess

[suka]

The trackpoint finally works, or more correctly put I finally figured I should probably betterconnect it to a working GPIO of the teensy I scrapped from my old BlueCube board (RIP). I thought I had cut off all fried pins but must have missed exactly the one I tried to put the PS/2 data signal on

Which combo do you prefer today?

IMG_20161122_214947554_HDR.jpg (634.13 KiB) Viewed 20224 times

Being technically more or less feature complete I decided to put the new design to the test and use it exclusively for a few days. And what can I say: Not too much of a difference even though I did not tent it for a longer period yet. Or maybe the switch from MX reds to browns again is overshadowing the subtle difference for now? Only my outer wrists got a little more pressure because I lazily put them on the table without noticing - since that was not at all possible with the much higher tilted variant I will have to monitor that bad habit again I guess since the board is so temptingly flat now.

Tilt options in different angles

IMG_20161122_215317377_HDR.jpg (784.26 KiB) Viewed 20224 times

Whatever the major difference will turn out to be, one thing changed notably already: While before people at work would rather give me pityful looks for typing on such a pile of junk parts and not even being able to afford proper printed caps now everybody passing by my desk is immediately drawn to the look of this creation. I guess because you now realize that it is not only the shape itself that is looking slightly irritating but also the minimal layout itself.

It took only some tiny software adjustments and I was able to use the new board for the entire last two days. I also snug a notifier LED into the frame this time, should not only be helpful for debugging but also in signalling the many different modes I use by now and tend to forget how to exit from them Macros and a passhash implementation are just such a great usability enhancement, but when you need to change them so seldomly it always was a minor nuisance to configure them without extra feedback from the firmware (yes, hid_listen and a good textual interface would be just fine, but everytime I work on a feature it feels I found a logical way to control it that I should not forget and then it all happens again ...

LED added

IMG_20161122_215334958.jpg (758.38 KiB) Viewed 20211 times

The transportation inter-frame works great, even though the new version came out a little too high - but this way the variable tenting modules could also be incorporated, and also a piece of cloth to prevent the caps from scratches, although that seems a little unlikely to happen.

So apart from giving it a few more days of typing time I'm waiting for the serial port expanders to arrive (yes, ordered them already as otherwise a new 3 1/2 year period of prototype status would have started for sure) so the interior can be greatly cleaned up and the cable made detachable. Hope to be able to fit them quickly as some other new component for a possible future project just arrived today
Although with the latest adjustments to the interior case elements it was not as hard anymore to get all cables and components aligned beneath the keyboard frame. The trackpoint stick need just a tiny bit more of clearance to the "C" cap, but I'd rather increase its initial reaction threshold before dremeling down these wonderful keycaps should a smaller extension not work!

Inside elements exactly fitting: But one too tight and one too brittle, duh!

IMG_20161122_215506559_HDR.jpg (619.62 KiB) Viewed 20211 times

Fun fact #73: 2*(1 MX blue + 3 Whites + 15 browns) take 965g in addition to the 130g of the upper half to fully compress when put upon each other

[suka]

I just could not resist: after a very interesting exchange with a fellow keyboard enthusiast last week I decided to finalize the bowl-shaped keyboard too, despite my satisfaction with the recent model.

But now all of the sudden it feels as if I have finally completed what I had in mind so many years, and this time without any compromises or prototype hacks: complete case, matching tenting supports, detachable connection cable and I2C on both sides for a minimal pin count.

OK, the Trackpoint and status LED are still missing, but that is the icing on the cake still to be done.




Inside still WIP since I was not sure the MCP23018 would fit and thus used long enough wires for alternatives,, but despite the many wires it is actually quite clean, especially at the teensy side

[tentator]

just wanted to say that I'm lurking your thread since years and love your project!
So far I never tried a splitted keybd but I also love the trackpoint as you do.. why do you have this white flat overlay on with a hole where the pointing stick then actually exits? Not clear if that is to "limit"/"contain" the max tilting of the pointstick or what..

tent:wq

[suka]

just wanted to say that I'm lurking your thread since years and love your project!

Thanks! Maybe I should split off another one with more detailed descriptions of individual features...

So far I never tried a splitted keybd but I also love the trackpoint as you do.. why do you have this white flat overlay on with a hole where the pointing stick then actually exits? Not clear if that is to "limit"/"contain" the max tilting of the pointstick or what..

You should definitely try a split board, so much better for a relaxed posture...

White flat overlay? Cannot follow you there. Do you mean the black flat holder of the trackpoint PCB?

[suka]

Ironed out the last couple of I2C bugs and put on some caps [*] to finally test drive that new baby - very nice albeit a little strange to type on, did not figure rotated rows would make such a difference. Seems a faster serial connection (or a reduction to only 16 keys per side ) could have been a better choice, though: not that there is a notable lag during normal use, but extremely fast consecutive hits seem to miss one out of maybe 15 presses. But hey, nothing optimized yet, so no worries, I'll give it a try this week...

IMG_20161212_022107502.jpg (41 KiB) Viewed 20221 times

[*] Too bad neither Retro nor Granite look nice in a black frame to me, so once again back to a layout proven to raise some eyebrows at work

[matt3o]

Skull Squadron goes well on anything

[suka]

Two complete days of typing on the latest "BlackBowl" creation and I must say I'm overly excited with it. I know you should never say never, but I feel I have nailed the ergonomic sweet-spot for me with it, and cannot see any more room for serious improvement. While I first found it a little strange to type on I now attribute that feeling to me being really tired at the time, and sitting at my working desk with a way to high table - it is an absolute blast!

The slight bowl or well shape is extremly satisfying to type on, decreases distance slightly and just seems to guide the fingers and hands back to centered positions with every little touch. So even when not tilted the board feels a lot more natural than anything else I've tried so far. I especially like the forward-tilted thumbkeys: makes them more natural to hit on the top rather than the edge, and makes up a little for the only little ergonomic nitpick I find - that I did not move the thumbs inwards further 1/2 key but rather kept the simpler rectangular layout.

Otherwise sizing and layout are true and tried, case and cable are finally presentable and since about 5 minutes the trackpoint is also mounted and functioning.

On the software-side of things, all the small features and major refactorings have finally also channeled into a pretty nice package, with proper PS/2 support in all modes thanks to TMK sources, on the fly macro recording, full configuration in EEPROM and changeable during runtime and also a passhash function compatible with Twik & HashIt! I2C has also improved enough that I notice no lags anymore at all, 125Hz scanning rate measured prove that impression ...

What's left to do? A RGB status LED for sure, a little carrying case or bag and plenty of nice-things-to-be-done-sometime are in my head, but unless something really urgent should surface I might just leave it as it is. If it wasn't for the new Kailh options I would probably see no reason to continue creating another board next year - but then there are always novelty ideas to be tested, so stay tuned

It did get crowded at the end: Never aim for too low a height or you will pay dearly during assembly :-)

IMG_20161213_220810358_HDR.jpg (403.22 KiB) Viewed 20140 times

[Tye]

Hey, suka,
Can you provide some input on how the TrackPoint stalk performs when it's between that set of keys? I am working on a custom board right now and I just can't get the courage to put a stalk there. I've moved it to the far left of the keys on the right hand board (so, one key to the left of where yours is). Does it interfere with regular typing or is it difficult to move because it's embedded in those caps? I really would like it closer like yours, though...

[suka]

Can you provide some input on how the TrackPoint stalk performs when it's between that set of keys? ... Does it interfere with regular typing or is it difficult to move because it's embedded in those caps?

In this particular configuration it does work marvelously if the extension is thin enough or you use a little larger spacing then the regular 19mm - I think I used 21mm minus the rotation. If you're willing to take away the corners of the surrounding caps - which I am absolutely not, at least with these keysets - you have no issues at all. I designed the extension to be 1.5mm diameter and cut it smaller around the bottom edge. Or tried an even thinner cross cross-section similar to MX stems, also fine.

Regarding the length, any increase in sensitivity through the longer leverage can easily be compensated through the trackpoint firmware. What I also did when I only had too long extensions at hand was to decrease sensitivity drastically when the trackpoint had not been used for a certain time and only ramp it up after the first intended stronger touch to prevent mouse mode from activating when unintentionally touching the nub. This can also easily and practically be combined with the Press-to-Select function as a selector, BTW.


So you see, plenty of options, and no reason not to mount it in a prime position if you have the room below the PCB. I guess the models where the stick is attached to its PCB by ribbon cable (not IBM?) would be even easier to place.

[Tye]

I didn't see anywhere which firmware you're using?
I played around with the acceleration curve in TMK on my Ergodox/Lenovo Trackpoint mod. Using a slow ramp up exponential type curve was the most interesting but in the end, even with a tall TrackPoint stalk, I stuck with the plain linear acceleration. I didn't even know the tap-to-select feature could be implemented third party. That's very interesting.
Grr, it's really tempting to reroute some traces on my new PCB to make room to experiment...

[mojes]

I'm in love with your last design.

[hbar]

Code: Select all

<delurking after several months of typing without thinking too much about the keyboard></>

Suka, that "BlackFlat" is almost identical to the "alpha" I'm typing on. Only mine has blank keycaps, a lasered stainless/acrylic case, and one more thumb key per side (disabled because they are useless). I'm glad to hear you dropped the outer pinky columns, I thought I wouldn't miss those so I omitted them right away.

The problem with labeled keycaps is that you don't get the right ones off-the-shelf, although uniform (non-sculpted) ones make things slightly easier.

As to the trackpoint: isn't the dome in the way of your fingers when you type? Mine is pure metal and never gets in the way, and I don't have to change those domes when they wear out (which is a pain on my laptop).

ħ

[wolfv]

The slight bowl or well shape is extremly satisfying to type on, decreases distance slightly and just seems to guide the fingers and hands back to centered positions with every little touch.

What a nice looking keyboard. I had not thought of centering by touch. Thanks for sharing.
(I was planning to make the columns in parallel planes, with staggered column heights. Will reconsider.)

Hope you don't mind answering some questions.

What is the layout of the thumb keys, modifier keys, and dual-role keys?

I am planning a very different bowl-shaped keyboard design.
Does text-based CAD work better than graphic CAD for bowl-shaped keyboard plates?
I have experience with Auto-desk Inventor and software programming, so I could go either way.
Would you mind sharing the OpenSCAD files for your bowl-shaped keyboard?

What is the purpose of the bare perfboard?

[geekabit]

Smile, you're on Hackaday.

https://hackaday.com/2021/12/02/keebin- ... -keyboard/

[Findecanor]

Smile, you're on Hackaday.

https://hackaday.com/2021/12/02/keebin- ... -keyboard/

This is a "geekhack thread" apparently