geekhack
geekhack Projects => Making Stuff Together! => Topic started by: Oobly on Tue, 15 October 2013, 07:51:59
-
Inspired by a cut on my little finger, an attempt to learn touch typing, the subsequent yelling of "WHY!?!?!", discovery of keyboard history and horror at the answer to "Why" and a decision to do something about it all.
(Also Suka's many interesting minimal builds, Truly Ergonomic, ErgoDox, Nexus (ErgoGP) and the desire to have / build something just right for my needs)
The concept: Design a keyboard and character layout that allows optimal hand, thumb and finger positions and movements. I also like dedicated arrow key and navigation clusters for editing.
I arrived at this:
[attach=1]
Goodbye Ducky G2Pro, it was good while it lasted:
[attach=2]
[attach=3]
[attach=4]
-
Testing thumb cluster designs:
[attach=1]
[attach=2]
Making mounts:
[attach=3]
[attach=4]
-
Wiring:
[attach=1]
All coming together:
[attach=2]
-
Got the firmware working and, IT'S ALIVE!
[attach=1]
At this stage I started learning my layout and discovered the angle is wrong (back is too high)... so, new mount / casing design:
[attach=2]
-
Feels good :)
[attach=1]
I may still change the thumb cluster position and angle, depending on further testing for optimal ergo! I really like using the thumb clusters, all 4 keys are easy to press, including combinations.
Up to about 12wpm touch-typing with a modified AdNW / BU-Teck layout (I get about 40 floating my hands all over a "normal" qwerty board). I'm using this site to learn: http://programmer-dvorak.appspot.com/index.html#lessons
Lots of similarities between the layouts in terms of which characters are on which rows and alternation, but BU-Teck is more modern and designed with staggered vertical rows in mind (also moves U off the home row and puts L in a better spot).
Next steps: Blank keycaps (I have some blank POM (black) and PBT (white) caps on their way to me), improving the firmware, covering the holes / making it look better.
-
Reserved... will post layout here.
-
Another reserved...
-
This is awesome!! Nice work.
-
Wow. That looks badass. Looks like a better design than those ErgoDox's...
Amazing work!
-
Nice!
-
Really amazing! Love the thumb cluster. How did you fab those sweet cases?
-
Bent sheets.
We had a workshop for sheet working. This is no easy task.
-
haha, all of my experiments with bent sheets were total failures. Definitely not easy :)
-
This is wonderful!
-
That is awesome, I want one now.
-
Wow! Thanks for all the kind words, guys. This project has been a lot of work, but very rewarding :D
The cases are made from 1.5mm aluminium sheet, cut with a jigsaw and "tinsnips" and bent with clamps, steel straightedge and pliers. They suffered a bit when I lowered the front edge, the finish is far from perfect, but they are functional :)
I forgot to add a few details: The controller is a Teensy 2.0 with a really simple firmware based on the PJRC USB keyboard code. It works fine for typing, but I have plans to make it a bit better for gaming, adding NKRO and possibly some media controls.
The cable linking the two units is 2 pieces of USB cable with a 10-pin connector on each end. I would like to upgrade the connectors at some point, since these ones work loose too easily. I went with direct matrix connections (4 rows, 6 columns per unit) to simplify the firmware instead of using an I2C encoder like the ErgoDox.
I realised I can't seem to add attachments to an old post, so here are my layouts, hope they make sense...
Alpha keys layout (modified AdNW / BU-Teck layout):
[attach=1]
Layer button pressed = Symbols:
[attach=2]
Layer and Shift pressed = Functions and Numpad:
[attach=3]
I will be flashing a QWERTY layout to it so friends and colleagues can try it out without having to learn my layout too.
I may still change the layout a bit, for instance the parenthesis and curly braces may swap places since I seem to be hitting the wrong one a lot while practising.
I am working on a Finnish / Swedish layout also, optimised for typing Finnish, Swedish and English, but it'll be a while still before it's finalised.
I will add a switch to the left hand unit so you can switch between 2 layouts stored in the firmware.
-
That looks good man, but did you really murder the ducky to salvage the switches and other parts ? :(
-
That looks good man, but did you really murder the ducky to salvage the switches and other parts ? :(
And quite rightly so too. Overpriced antiquated staggered layout pile of crap. The plate's the only bit worth keeping, and that's only if you can misuse it as oobly's done. Kill them all, I say. KILL THEM ALL!
That's totally ace, that is. Absolutely fantastic work.
-
This is beyond cool. Good work! Looking forward to anything that comes from this experiment of yours.
-
That looks good man, but did you really murder the ducky to salvage the switches and other parts ? :(
Well... not quite. It's more like an unwilling organ donor. There are enough bits left to keep it alive for other cruel and unusual experiments.. Muahahaha! Muahaha!
[attach=1]
I sometimes tell it I may order some more switches and run it plateless in order to keep its hopes up, although I am more likely to use the rest of the switches for the next project. >:D
That looks good man, but did you really murder the ducky to salvage the switches and other parts ? :(
And quite rightly so too. Overpriced antiquated staggered layout pile of crap. The plate's the only bit worth keeping, and that's only if you can misuse it as oobly's done. Kill them all, I say. KILL THEM ALL!
That's totally ace, that is. Absolutely fantastic work.
Well said, sir! And thank you! :)
Actually, the PCB is nice, too. I think it's a good keyboard overall, but I wanted something less likely to ruin my wrists when learning to touch type. <- actually not sure how much it helps, but I was frustrated with the normal layout.
I've been rethinking the parenthesis, braces, brackets and greater than / less than symbols. I may separate them into a mirrored configuration, so you open them with the left hand and close them with the same finger of the right hand.
Keycaps may become a hassle, too, if people who try it want legends on them.
Been working on an an even more minimalist build in my head already (inspired by JDCarpe's Smallfry 40% board), basically the same as this one, but without the navigation clusters. It wouldn't fit all the possible characters in, but enough to be a daily driver. I have just enough key switches left over from the Ducky for it ;)
-
Using the plate sideways is a stroke of genius. I'm strongly considering doing this to an old chicony board with futabas I have lying about. Just checked, and it's pcb rather than plate mounted, but that shouldn't be too much of an issue. Tempted to add a pair of thumbsticks rather than / as well as the directional clusters, though.
-
Really nice Oobly, case looks good,from stainless?
i see your handwiring without diode it seems
what you use cable to connect 2 halves?
Thanks
-
Great work, nice board!
Thing looks really badass...congrats on a job well done! :thumb:
-
Damn.. you crazy... good work.
-
Great work, nice board!
Thing looks really badass...congrats on a job well done! :thumb:
Thank you!
Using the plate sideways is a stroke of genius. I'm strongly considering doing this to an old chicony board with futabas I have lying about. Just checked, and it's pcb rather than plate mounted, but that shouldn't be too much of an issue. Tempted to add a pair of thumbsticks rather than / as well as the directional clusters, though.
Thanks, it was one of my better ideas. I have since discovered that the pinkie areas don't quite have enough of a stagger compared to the other columns, so I have had to do some modifications:
Pinkie areas cut off, ready to be reattached lower:
[attach=1]
Thumb clusters mounted at a new angle:
[attach=2]
This is now approaching the perfect layout for my hands.
Damn.. you crazy... good work.
Thanks!
Really nice Oobly, case looks good,from stainless?
i see your handwiring without diode it seems
what you use cable to connect 2 halves?
Thanks
Thank you. Case is 1.5mm aluminium sheet. There are diodes (1N4148) on each switch, wired to the columns. I found some small 5x2 pin connectors and used 2 pieces of USB cable (5 wires) to make the connecting cable. 10 wires is enough because I have a 4x6 matrix (4 rows, 6 columns), although in future I want to add 4 more wires for RGB LED backlighting and to ground the casing, but that can wait. Not sure what connectors and wires I will use for that yet.
I may add an AVR in the right hand section and use I2C between them like the ErgoDox.
-
I realised I haven't posted a user perspective shot of the board yet, so here are a couple:
[attach=1]
[attach=2]
I have reinstalled the keycaps in "normal" position, with reversed row4 caps for the arrow and edit clusters. It's a bit more comfortable to use now and any standard keycap set can be used as long as you don't mind incorrect legends on the edit keys, arrows and thumb keys.
I have switched the () and {} characters in the symbols layer. Feels more natural to me. Will update the image soon.
-
Using the plate sideways is a stroke of genius. I'm strongly considering doing this to an old chicony board with futabas I have lying about. Just checked, and it's pcb rather than plate mounted, but that shouldn't be too much of an issue. Tempted to add a pair of thumbsticks rather than / as well as the directional clusters, though.
You'd have to check where the rows and columns connect and how they are routed on the PCB, but it may be fine. I have been thinking of adding one of these: https://www.sparkfun.com/products/10835
Not sure quite where to mount it yet. It will make the firmware a lot more complex, though. Will be nice not to have to move my hands to use the mouse :) Suka has used an IBM Trackpoint nub on some of his keyboard builds with great success. Check them out here: http://deskthority.net/workshop-f7/my-diy-keyboard-collection-or-how-i-became-a-kb-geek-t2534.html
-
How the board is laid out is less of an issue, I'll cut the traces where necessary and then hardwire anyway, using the PCB as a mounting substrate and nothing else.
I saw those joysticks modules, kinda neat, but I've got a bunch of gamepads which are broken in one way or another, they should be reusable. Trackpoint would be cool as an alternative, and I have a couple of PS/2 trackpads lying about with nothing to do. And a very old trackball. And a bunch of rotary encoders, might be fun to roll my own rollermouse.
Oh, choices, choices.
-
Thank you. Case is 1.5mm aluminium sheet. There are diodes (1N4148) on each switch, wired to the columns. I found some small 5x2 pin connectors and used 2 pieces of USB cable (5 wires) to make the connecting cable. 10 wires is enough because I have a 4x6 matrix (4 rows, 6 columns), although in future I want to add 4 more wires for RGB LED backlighting and to ground the casing, but that can wait. Not sure what connectors and wires I will use for that yet.
I may add an AVR in the right hand section and use I2C between them like the ErgoDox.
Oh yes i see diodes :)
What advantage use AVR and 12C?
-
Thank you. Case is 1.5mm aluminium sheet. There are diodes (1N4148) on each switch, wired to the columns. I found some small 5x2 pin connectors and used 2 pieces of USB cable (5 wires) to make the connecting cable. 10 wires is enough because I have a 4x6 matrix (4 rows, 6 columns), although in future I want to add 4 more wires for RGB LED backlighting and to ground the casing, but that can wait. Not sure what connectors and wires I will use for that yet.
I may add an AVR in the right hand section and use I2C between them like the ErgoDox.
Oh yes i see diodes :)
What advantage use AVR and 12C?
It only needs 3 wires (ground, signal and power to the AVR), so I can use a thinner cable to connect the 2 pieces.
-
Man, I step away from GH for a week and I missed all the goodies....This is really awesome. Can you explain a little more how the wire between the two halves works and how its different from the Ergodox one? I never really understood how the Ergodox cable worked in the first place.
-
As it stands, Oobly has 10 wires running between the two halves of his board. These are 4 columns and six rows of his matrix (or vice versa). For the 48 keys he has in total in the two parts, he needs a 6x8 matrix (6x8 uses the least number of pins, 14, for 48 keys), so one half needs a 6x4 matrix. So 4 of the wires are connected direct to pins on the microcontroller, and the other 6 are connected serially with the 6 rows on the other half of the board.
The i2c option requires a two microcontrollers, one for each half of the board. One simply scans a single 6x4 matrix and sends events over 12c. The other scans its own 6x4 matrix, listens on i2c for key events on the other half of the board, and sends the merged stream of events over USB. I believe that's how Ergodox works.
-
The i2c option requires a two microcontrollers, one for each half of the board.
...
I believe that's how Ergodox works.
No, the ErgoDox has a microcontroller in only the right half. In the left half is a MCP23018 "I/O expander", which is a much simpler, and cheaper, chip. It has two 8-bit ports, which the microcontroller reads and writes using simple commands over i2c.
However, because the PCB is symmetric, used for both the left and right, you could build them as two one-handed ErgoDox'es. Most operating systems that support USB do support multiple keyboards over USB.
But then you would have to put the components upside-down on the left keyboard, modify the firmware to not try to talk to the other half and program each keyboard separately.
-
Ah, I stand corrected. Thanks.
-
Man, I step away from GH for a week and I missed all the goodies....This is really awesome. Can you explain a little more how the wire between the two halves works and how its different from the Ergodox one? I never really understood how the Ergodox cable worked in the first place.
Thanks, Tufty and Findecanor have explained it very well I think (thanks, guys!).
I have been considering using a Teensy in both halves, each with two operating modes (acting as a full controller and listening to the other Teensy signals, or just sending the key events over I2C). That way each half can act as a standalone mini keyboard or game controller (think left handed gamer, for instance) and you can connect the USB cable to whichever side is more convenient. For the prototype I am happy with the simplicity of the current setup for now.
This post took a while to type, up to about 18wpm now :cool:
-
Oobly, this is really cool. I think your keyboard is the closest thing to what I'm looking for, but not quite. So I guess I gotta build my own vision too. :) But just wanted to say nice work, it's inspiring to see what you've done. I love the minimal layout too.
-
Oobly, this is really cool. I think your keyboard is the closest thing to what I'm looking for, but not quite. So I guess I gotta build my own vision too. :) But just wanted to say nice work, it's inspiring to see what you've done. I love the minimal layout too.
Thanks!
Have you had a look at AcidFire's Nexus project? It's less "minimal", but getting close in terms of layout. However I definitely think it's worth trying to build your own vision. It's very rewarding and you end up with exactly what you want, no compromise. Do you have some plans for what methods to use yet?
I've added blank white PBT caps from Banggood (I think they're Keycool caps) and trampoline modded all the switches. It feels great, now.
[attach=1]
I still want to move the thumb clusters a bit. They need to be even further back (towards the user) and a little less angled side-to-side (more parallel to the columns). And I need to make a nice palm rest of the correct height. I'm also finishing up the design of the second version. I will be using acrylic (curved) and mounting the base of the switches to it (so I can mod them easily and can use round mount holes instead of square). Still finalising some details, but I am hoping to be able to include a palm rest in the actual case design.
-
This is a really cool project, and an excellent idea to use a staggered backplane sideways.
I have a suggestion, though not easy to explain so please ask if I'm not clear. Staggering between regular keyboard row is not regular (from top to bottom for the 4 main rows it's 0.5, 0.25, 0.5). I see that you used the 0.5 offset between your index finger and middle finger, and the 0.25 offset between the middle finger and ring finger. This may seem natural since as most men you have a ring finger longer than your index.
However the index finger has more "controlling hardware" in the arm (muscles, tendons, etc.) than the ring finger, so it's more agile and may move longer distances more easily. Did you have a try reversing the staggering order so that you get a shorter offset between the index finger and middle finger, and a longer one between the middle finger and the ring finger? To test it you can either put each keypad upside down, or swap them. It might be more comfortable.
-
Very nice! Love it!
two questions:
How are you planning to curve the acrylic?
Have you tried any of the DSA profile caps? They seem more amenable to vertical mounting, etc.
-
I want this to be a product... :-*
-
This is a really cool project, and an excellent idea to use a staggered backplane sideways.
I have a suggestion, though not easy to explain so please ask if I'm not clear. Staggering between regular keyboard row is not regular (from top to bottom for the 4 main rows it's 0.5, 0.25, 0.5). I see that you used the 0.5 offset between your index finger and middle finger, and the 0.25 offset between the middle finger and ring finger. This may seem natural since as most men you have a ring finger longer than your index.
However the index finger has more "controlling hardware" in the arm (muscles, tendons, etc.) than the ring finger, so it's more agile and may move longer distances more easily. Did you have a try reversing the staggering order so that you get a shorter offset between the index finger and middle finger, and a longer one between the middle finger and the ring finger? To test it you can either put each keypad upside down, or swap them. It might be more comfortable.
I have tried both ways and I think I get what you mean. It's easier to "stretch" the index finger up, harder for the ring finger. However I find it more comfortable as it is. The fingers fall naturally on the home row keys (they find the keys naturally) and the movement of the ring finger up to the "top" row is fine, just as easy as the others.
In this case I have chosen resting position to be more important than range of movement and it works for me. I suppose the ideal would be to have smaller vertical spacing between keys (AND natural position for home keys), but that requires either non-standard keycaps or curving the mount plate (which is difficult with staggered columns). Filing the bottom angled part of standard keycaps could work, though. Also, a fully custom, curved keywell like Kinesis and Maltron use would be nice, but (kind of) hard to make.
Very nice! Love it!
two questions:
How are you planning to curve the acrylic?
Have you tried any of the DSA profile caps? They seem more amenable to vertical mounting, etc.
Thanks! I love your board design, too! I plan to use a heat gun, a metal form and some protective gloves to smooth it down at the "formable, but not yet sticky" stage. Yet to try it, but I think it's feasible. I'm still not sure about the integrated palm rests, though, since they would require some curved to straight edges / bends and wouldn't be removable, making the board less portable. I want to be able to carry it easily to work and back, from desk to desk, etc.
I haven't tried DSA, but the OEM profile is comfortable. It may actually feel worse with more uniform height sphericals since the heights of OEM almost create a curve in the vertical which matches nicely to the movement of the fingers.
I want this to be a product... :-*
I'll see how the acrylic version goes and who knows... Maybe I'll make a very limited run (this is just a thought at the moment, please don't hold me to this!). It all depends on how much work / time will go into each and if I can come up with a repeatable and affordable process. Maybe I could make some kits? Still far too early to tell. I need to finalise the thumb cluster position with third party testing locally before going further.
Also, the cable connection between the halves needs to be more robust. I have some thoughts on how to do this which I'll test with the acrylic build.
jeffgran, have you seen suka's builds here: http://deskthority.net/workshop-f7/my-diy-keyboard-collection-or-how-i-became-a-kb-geek-t2534-120.html
Very inspiring :)
-
Oobly, this is really cool. I think your keyboard is the closest thing to what I'm looking for, but not quite. So I guess I gotta build my own vision too. :) But just wanted to say nice work, it's inspiring to see what you've done. I love the minimal layout too.
Thanks!
Have you had a look at AcidFire's Nexus project? It's less "minimal", but getting close in terms of layout. However I definitely think it's worth trying to build your own vision. It's very rewarding and you end up with exactly what you want, no compromise. Do you have some plans for what methods to use yet?
Yeah, I've seen what AcidFire is working on and it looks pretty cool too. I think for me it's still a bit too derivative of the ergodox. I just put my ergodox together (from R3, finally!) last week and I have to say.... eh. It seems like I've heard at least 4 or 5 people comment on how the ergodox's thumb cluster leaves something to be desired. But to me, the thumb buttons are the main reason for such a crazy keyboard in the first place. I actually want all six thumb keys (maybe more!) but I want them to be easy and natural to reach with my thumb. I love what you've done because I think it just makes good common sense that your thumb cluster should be angled like that. If you look at your thumbs, you realize they move in the opposite direction from the fingers. In school we learn that the Apes are special because of "opposable" thumbs! Our thumbs don't move up and down, they move in and out.
@bearcat's three thumb keys are sweet too. But speaking for myself, I need more :). I want space, backspace, tab, enter, control, meta, layer1, layer2, and some of those (the control/layer keys) need to be accessible from either thumb. Here's the thread I started a while back with my preliminary design: http://geekhack.org/index.php?topic=46015.msg966802#msg966802
I laid out the keys in the pattern I propose there on a table top, and pretended to type, and it seemed like that was the best natural-feeling layout I could come up with. Although with that design I still don't have the opposed thumb cluster like you do. And doing the "foldable" design may be prohibitively difficult, at least for the prototyping stage of it. So I'm going to have to go through some design iterations I think.
-
jeffgran, have you seen suka's builds here: http://deskthority.net/workshop-f7/my-diy-keyboard-collection-or-how-i-became-a-kb-geek-t2534-120.html
Very inspiring :)
Yeah I did come across that a week or so ago. Interesting for sure. I have to say that I don't like his physical layouts all that much, but I love his persistence and creativity in trying all kinds of designs, and I was definitely interested in the "soft" layouts, i.e. the firmware (neo-layout, BU-teck, etc). None of them seem quite right for me (I'm experimenting with Workman right now), but I love the diversity that I'm seeing in layout options.
BTW, does anyone know of a way to properly distinguish what you mean when you say "layout" here? You could be referring to e.g. staggered vs matrix or tenkeyless vs. 60%, OR you could be referring to qwerty vs. dvorak vs. colemak vs. workman, etc. Should we invent a qualifier or is there already one? Like when I said "soft" layout vs. "physical" layout... but better?
-
First of all - this is my favorite now, this resembles the most to what i'm trying to make myself, but seems that your craftsmanship skill is more advanced.
Since you're not using those i2c's/ioexpanders, you should definitely take a look at tmk_keyboard firmware (https://github.com/tmk/tmk_keyboard), it will take you a while to learn by examples, but it works and has a lot of interesting features that may affect your logical layout, e.g. tap vs hold making different actions as in suka's bluecube keyboard (http://deskthority.net/workshop-f7/my-diy-keyboard-collection-or-how-i-became-a-kb-geek-t2534.html), where space and shift are the same key, as all the rest of main action buttons are paired with modifiers. That, by the way, may help putting Esc and Del to thumb cluster so there were easier to reach.
I don't insist, but you may also like workman layout (http://www.workmanlayout.com/), at least for alpha keys.
It probably makes sense for those switching from qwerty to keep correspondence of numerical keys to those special chars the same way it's on qwerty, e.g. layer+key=1, layer+shift+key=! et c.
Do you use navigation keys often? They look kind of out of reach.
You mentioned that you may want to drop some of the keys, making it total of 40 keys, what you were going to get rid of?
Do you use any kind of whist rest? or does it feel comfortable with hands flying over keyboard? I'm asking as a heavy typist, my shoulders feel terribly if i fly-type a lot.
-
Sorry for the wall of text, but here are some answers and motivations for some of the layout decisions, firmware choice, etc.
I use "physical layout" and "character layout", but have no idea what the proper terms are.
Tap and hold firmware can be tricky to get the timing right and it's sometimes too easy to press another key before releasing the previous, especially for space bar. It also introduces a delay in response for those keys which can be very bad when gaming, particularly games which require perfect timing. Jumping with space bar in a platform game, for example.
I chose not to use Workman, Norman, Colemak, Dvorak, Carpalx optimised, Programmer Dvorak, etc. for various reasons and settled on modified AdNW / BU-Teck.
1. Most of them are designed for standard horizontally staggered physical layout.
2. Some place characters in very weird spots.
3. I wanted a layout that has very little in common with QWERTY so I can keep my QWERTY skills alive while learning the new character layout on a new physical layout. The difference means I am learning a NEW set of skills for my new board instead of retraining the old (which requires UNLEARNING the old layout while at the same time learning the new, kind of hard after more than 20 years). It's working, too. I can still type just as fast on a "normal" board as I always have and I'm up to 20WPM on the new.
4. I prefer outward rolls to inward, but prefer alternation even more. Some layouts are designed specifically for inward rolls.
BU-Teck is designed for alternation, vertical stagger physical layout with thumb modifers, has well-placed characters for finger and hand use, is very unlike QWERTY and is well optimised and analysed for the most common trigrams in English. Unfortunately it also includes unnecessary characters and includes some weighting in its optimisation for German, which I don't need. I have thus modified it to suit me. Still may make some small changes, but in general I am happy with it. Happier than I would be with any of the other layouts I have come across so far.
I did also come up with my own character layout, but I haven't completely analysed it yet. So far it does very well in the online layout analyzer (http://patorjk.com/keyboard-layout-analyzer/#/main), but I suspect it won't be as good as my current one for prolonged typing.
The number keys and characters have been very easy to learn so far. I find it's not so much a matter of symbols being associated with the corresponding number, but rather the position, so learning new positions for each independently hasn't been an issue. In fact it's a lot nicer to open and close brackets, etc AND type numbers with the layout. Could be partly because of the "numpad" type layout for the numbers and "strong" positions of the brackets (in the index and ring finger home key positions).
I do use the arrows and nav keys, but not as often as I thought I would. I find them quite okay to use, since I change my hand position a little when editing. I use shift in combination with them for selecting areas to copy, paste, etc. For the 40 key version I would drop the navigation cluster and arrows.
The Esc and Del buttons are well within reach of my pinkies. About the same as Q and P on a "normal" board, but I would certainly consider putting them on thumb buttons if I had more. Sometimes it's nice to have ESC in the default "panic" location, top left, but Del has no particularly strong reason for being where it is.
The 40 key layout isn't finalised, but it would probably have something like the same alpha and symbol layouts as the current one. For the function layer it would have:
F1 on Esc, F2 to F11 on the top row, F12 on Del.
Numbers on right hand, 1 to 5 in middle row, 6 to 0 in lower.
PgUp, Home, Up, End on left hand home keys.
PgDn, Left, Down, Right on lower row.
That way I can do all the edit functions with the left hand while keeping my right on my mouse.
There really are a lot of options, perhaps the best (if I were to make these available as a kit for example) would be a board with the same physical layout as the current one, but with 2 more thumb buttons per side and completely reprogrammable (would allow more layers and possibly some macro keys), although I must say that I am very happy with the current design and "only" 4 thumb buttons per side as they really do everything I need and are very easily reached with as little movement as possible.
I will experiment with 6 thumb buttons on the acrylic version. Doesn't hurt to have extra buttons, I don't have to use them all :D
-
Thanks for comprehensive answer!
I think most of layout analyzers miss the physical layout thing, and it's a huge advantage that patorjk has ergodox included, so the guys making a split non-staggered keyboard with less buttons may analyze what's best for them. But as you mentioned, it's quite doubtful that they take into account series of pressed keys, and that inward/outward/alternate rolling (i personally feel that inwards is more natural).
You mentioned the kit, if when it comes to that point please put me to the list.
-
Hey awesome work. Let me know if you want help creating a shell for your device. I actually built a custom ergo controller shell for a friend for a decent price and came out pretty cool. We built out some custom hardware too.
Just throwing that out there. It's not cheap to do though, mainly because 3D printing isn't very cheep. But if your interested let me know and I can do that type of stuff. Cheers mate
-
Great work for sure.
Just thinking about how I could obtain something like this, looks like a cool keyboard. You are right about the thumbs moving in and out, not up and down like on the Ergodox (which I think is awesome, but there's still room for improvement).
I would think about 3d printing the base for the keys, I wonder how we could make a curved or angled plate more accessible for people (like me) who are not willing to hack up an expensive plate-mounted keyboard for a project. Also I'd rather a matrix layout.
I guess I could also get a plate lasercut with bending points and holes ready for standoffs. Just thinking out loud.
-
I also think it should be possible to machine a top case + plate as one piece, out of metal or wood. If you have access to a 5-axis CNC machine (or can hire someone who does), and you’re willing to do direct wiring instead of using a PCB, then you should be able to make a milled top case + plate that has a curved keywell/etc. If your CNC machine is only 3 axis, then I think you’d be stuck with flat pieces, but you could maybe join a few of them together afterwards at some angle.
-
I also think it should be possible to machine a top case + plate as one piece, out of metal or wood. If you have access to a 5-axis CNC machine (or can hire someone who does), and you’re willing to do direct wiring instead of using a PCB, then you should be able to make a milled top case + plate that has a curved keywell/etc. If your CNC machine is only 3 axis, then I think you’d be stuck with flat pieces, but you could maybe join a few of them together afterwards at some angle.
Sounds expensive...
How would you use wood as a plate? Would it fit close enough?
-
Such craftsmanship is beyond my ability right now, but I think I might follow this and build my own ergo keyboard someday...
-
Totally rockin!
-
I'm periodically polling this topic for news. How is it going with your keyboard?
-
I'm periodically polling this topic for news. How is it going with your keyboard?
Development is on hold for the moment, I'm afraid. Lots of "real life" keeping me busy, so not much project time. The keyboard is in daily use on my home PC and does the job very well. I do still plan to build a "Mark II" version (either from curved acrylic or with flat steel plates like the first one and wooden cases) with a controller in each side, but it has to wait a little until I get more time for keyboard projects.
The reason for having a controller in each side is that either can be used on its own as a game controller or one-handed keyboard. At startup the firmware checks if the USB bus has been configured (that there is a host present) and if not, goes into I2C (master) mode. If there is a USB host it goes into "combined" mode (I2C slave mode + scanning local key matrix + handling USB). So either side can be connected to the PC and then simply linked to the other via 4-core cable (RJ11).
I think steel is the best plate material, although aluminium could work for plates this size if it's the right alloy and acrylic / perspex works well enough, too, but the thicknes requires the switches to be glued in place.
-
I should totally have used RJ11 connectors. Glad to hear you're still alive!
-
I'm periodically polling this topic for news. How is it going with your keyboard?
Development is on hold for the moment, I'm afraid. Lots of "real life" keeping me busy, so not much project time. The keyboard is in daily use on my home PC and does the job very well. I do still plan to build a "Mark II" version (either from curved acrylic or with flat steel plates like the first one and wooden cases) with a controller in each side, but it has to wait a little until I get more time for keyboard projects.
The reason for having a controller in each side is that either can be used on its own as a game controller or one-handed keyboard. At startup the firmware checks if the USB bus has been configured (that there is a host present) and if not, goes into I2C (master) mode. If there is a USB host it goes into "combined" mode (I2C slave mode + scanning local key matrix + handling USB). So either side can be connected to the PC and then simply linked to the other via 4-core cable (RJ11).
I think steel is the best plate material, although aluminium could work for plates this size if it's the right alloy and acrylic / perspex works well enough, too, but the thicknes requires the switches to be glued in place.
I'm curious, why do you have the USB side going into slave mode?
Also, as bearcat said it's good to see you're still around :D
-
that is a sweet board, i had something like that in mind. i know nothing about putting that together or where to start! someday you have to add joystick to it and make a gaming version
-
I'm periodically polling this topic for news. How is it going with your keyboard?
Development is on hold for the moment, I'm afraid. Lots of "real life" keeping me busy, so not much project time. The keyboard is in daily use on my home PC and does the job very well. I do still plan to build a "Mark II" version (either from curved acrylic or with flat steel plates like the first one and wooden cases) with a controller in each side, but it has to wait a little until I get more time for keyboard projects.
The reason for having a controller in each side is that either can be used on its own as a game controller or one-handed keyboard. At startup the firmware checks if the USB bus has been configured (that there is a host present) and if not, goes into I2C (master) mode. If there is a USB host it goes into "combined" mode (I2C slave mode + scanning local key matrix + handling USB). So either side can be connected to the PC and then simply linked to the other via 4-core cable (RJ11).
I think steel is the best plate material, although aluminium could work for plates this size if it's the right alloy and acrylic / perspex works well enough, too, but the thicknes requires the switches to be glued in place.
I'm curious, why do you have the USB side going into slave mode?
Also, as bearcat said it's good to see you're still around :D
Just thought I'd post a heads-up. The design of prototype number 2 is progressing (slowly). It'll use acrylic for the main structure and plate. I'm experimenting with bending methods and a novel "minimalist" design and assembly method. I am also trying out built-in wrist rests and seeing how much it affects the portability. I'll be getting PCB's made for it, too.
The USB side goes into i2c slave mode to "listen" to the signals from the side in master mode sending the key down/up events. I still need to experiment with this to see if the processor can handle everything it needs to be doing and still have fast enough response. I'd like to keep the time from actuation of the switch to sending the USB packet to under 2ms, depending on how I implement debounce. Initially the next prototype controller will only use 6KRO, but I may extend it to NKRO if there is enough processing capability remaining to support it.
-
sweet
-
Unlurking just to commend you on your project. It's simply.. amazing.
How difficult would it be to make such a keyboard, for someone who has little to no clue as to the programming/electronics side of things?
Would definitely love instructions/kit to make something similar.
-
Hi, may i know what did u use to make the metal casing for the ergo?
-
I think I'm inspired enough by this that I won't rest easy until I've tailor made my own, lmao.
What do you guys think of mounting the keypads on some movable arms that you can lock into place? With the option to easily remove them of course.
-
I'm curious what angles/rotation you ultimately arrived at for the thumb cluster relative to the main board... how far down, how rotated on each axis, etc...
-
I'm curious what angles/rotation you ultimately arrived at for the thumb cluster relative to the main board... how far down, how rotated on each axis, etc...
dido, any details would be good to compare with
-
forgot to ask you, what did you use to cut the Ducky plate? That seems like pretty accurate cutting; something I haven't yet done with my own dremel and cutting wheels.
-
Would it be possible to 3d scan this so it could be more easily reproduced? I'm in love! Nice work dude!
-
absolutely epic build! oobly keep up the good work!
-
just awesome. Is it possible to buy one from you?
-
this is extremely cool Oobly!
-
This is interesting!
I'm a ModelM lover, not sure if mechanical switches would be for me :(
I'll open a thread of find a suitable one, but writing it to not forget it: I want to buy a big comfortable and robust armchair for computing, studying and electronics stuff (going to ask when I have the money). The main issue would be if I would adapt to a splitted keyboard like that IBM one, plus having it in a stable way and my CST L-TrackX there too.