Katy keyboard (or K80CS (Key80 Contoured Split))

CrimsonFlame · 06 Jul 2015, 23:17

@vvp:Looks like I was using oldstable by mistake. I set everything up to use the ubuntu ppa, but a typo made it prefer the oldstable, and I didn't catch it. I'm now running r5886, and everything is working great. I'm getting some weird lock up when loading libs, but if I close the non-rendering popup, it all works just fine. Who knows.

Thanks for the tips about Acetone. We've been meaning to pick up some acetone for the printer lab, but never managed. There's also the safety officer who complained last time we tried to order through official channels. Said it was enough of a hazard that we would have to put it in a chem safety cabinet, which we had no extras on site and would have to order ( a few hundred dollars of course). All for 16oz of acetone. No thanks

Of course if we just bought and stuck it in a drawer and claimed it was for personal use (nail care), it was all good.

vvp · 06 Jul 2015, 23:52

Great you solved the Kicad problem.

The safety officer is probably concerned with the fact that it evaporates quickly and burns very well. So bigger amounts must be handled with care (so that the acetone vapor does not reach concentrations when it could explode).

Just work with it in a good ventilated place or outside and there will not be any problem. If it would be dangerous then women would not be able to buy it in small amounts for a high price and with label "nail polish remover"

Also do not leave it in a car in summer (it boils at 57°C). If the container cannot handle higher pressures then this could be a problem.

CrimsonFlame · 09 Jul 2015, 17:53

Just an update:

I've tried printing the palm rest twice, and it keeps peeling up from the bed. I might just use them as it (it's a nice gentle roll off towards the back. Almost looks intentional !

If I feel determined, I'll try a full wipe-down of the bed (we used to clean it with Alcohol after each print, but our alcohol bottle walked off

), and maybe I'll use an ABS slurry on the bed.

vvp · 09 Jul 2015, 22:52

No big peeling is intentional.
My bed is covered with kapton tape and cleaned with acetone before printing. All parts hold well.

Cleaning with alcohol does not work that well. I tried it once and got pealing off too. Though it may have been in times when my bed was not leveled that well leveled.

If it peals off even after the bed was cleaned and If you think some of the adhesion pads need to be bigger or thicker then let me know which ones. I'll modify the model. It would be cool if you can post a picture of the problem.

I often get a slight peal off of corners (maybe by 0.5 mm at most). But only in the middle of the circular adhesion pads. It never propagates behind the circular border of the pads. I ignore these.

CrimsonFlame · 11 Jul 2015, 02:50

Hey vvp, where do you get your acetone? I picked up some 100% acetone nail polish remover from my local CVS today (as I happened to be there already), and went to town on the bottom plate. It however discolored the print. My bottle list denatonium beonzoate as an ingredient (bittering agent, I know), but perhaps it isn't added to the paint thinner type cans that you can pick up at a hardware store, and that's what's causing my problem. I might be able to hit my hw store this weekend.

Top Side:

Bottom :

Another thought, perhaps this is caused by the dye used in this batch of ABS? I'll try the acetone on another color of print to check.

vvp · 11 Jul 2015, 11:24

Edit: Edits in italic.

Oh shit! I forgot to mention this but also because this problem is not serious on red ABS, it is almost not there on the yellow ABS, and it is not there at all on white color ABS of course.
In other words the darker the color the bigger the problem and you selected the black color. I was kind of surprised with your color preference

This is one of the disadvantages of acetone painting as opposed to acetone vapor bath.
Natural ABS color color is white, just as the smuggles you see on your case. Acetone is melting the surface of the part when painting and the ABS itself sticks better to the brush than the color pigment added to the ABS. Or maybe acetone somehow can neutralize the pigment (I'm not a chemist). The more your acetone is contaminated with ABS the more white smudges. Notice how your acetone get whiter and whiter when you paint something with it.

To fix it you need to use clean acetone and "wash" the smudges with the brush. Make light moves over the surface. Do not press. It may be a good idea to scrub the most white places with a knife first (to remove most of the discolored ABS from the surface first). Other option is to paint the case. I do not have an idea what colors are good to use on ABS plastic but I saw same discussions on geekhack (and maybe even here on deskthority) about the suitable paints for ABS. If you cannot find it and have problem to "wash" the white smudges then post a question in a new thread about paints. Somebody will know for sure.

Next time select an ABS with lighter color if you can.

As for as acetone sources:

The best is simple clean acetone without any additives. It is available in shops which specialize in selling stuff for home builders or people who like to tinker or improve their homes (e.g. Baumax). It is sold as a solvent to aid removal of old (dry) paints or for cellulose based glues. In Slovakia, it sells under name "Aceton" or "P 6401". It sells in cans of 0.4 l and bigger. 0.4 l is plenty enough. You may be able to buy it also in shops which sell paints but these often do not carry it. They often carry only a "acetone" paint thinner discussed below. You may also find it in RC model shops, or if they do not carry it then they may be able to tell you where to buy it.

"Acetone" paint thinner. It is in quotation marks since it typically does not contain much acetone if any at all. In my country it is also known under name "C 6000". And this is what will be sold to you in Slovakia when you come to a paint shop and ask for acetone

It contains mostly (methyl, ethyl, buthyl) acetate, toluene and butanol. It will melt ABS a bit but it takes looooong time and it will not melt ABS completely anyway. Not good for making ABS-glue at all. But it is good for cleaning bed before printing with ABS.

I do not know what exactly is in the nail polish removers but they should contain significant percentage of acetone.
I read this somewhere in RepRap wiki.

How did go the printing of the palm part after the bed is cleaned?

Zekromtor · 13 Jul 2015, 08:39

Nice updates. I'm no longer a proponent for different alignment of horizontal rows, but I still love this board. It's the Kinesis killer.

vvp · 19 Jul 2015, 15:07

OK, I pushed the firmware out. It is not complete yet but the keyboard functionality itself, on the fly remap, and macros, and the LEDs (new on v0.7) should work fine. The photo sensor is not activated yet (i.e. the LEDs shine the same regardless of ambient light for now).

So you can get all 3 important sets of data to build this:

I'll add more information later on depending what questions will be there (if any).

Very rough estimate how much time it takes to build this. This assumes that you know how to use the tools needed. If you are new to something which needs to be done then it will take longer. I will measure the times the next time I'll be building it. So I'll give better estimates sometimes in the future.

the case
* setup print jobs: 7 * 0.5 h (plus the print time when you do not need to be there)
* clean up the parts and acetone painting: 2 * (3 h + 1.5 h + 1 h)
* glue the key part wit the palm part: 2 * 0.5 h
if you make one sided controller PCB at home then:
* toner transfer and chemical etching: 1.25 h
* drilling: 2 h
* putting in bridges (to simulate the front Cu side): 1.25 h
if you make one sided left side PCB at home then:
* toner transfer and chemical etching: 1.25 h
* drilling: 1 h
* putting in bridges (to simulate the front Cu side): 0.5 h
soldering parts on PCBs: 3 h + 0.5 h
prepare switches (put diodes on them): 3 h
glue switches to the plates: 2.5 h
vertical matrix wires (bare Ø 0.25 mm Cu wire): 3 h
horizontal matrix wires (enameled Ø 0.2 mm Cu wire): 4 h
connectors and wiring (pin housing / IDC / ethernet / USB): 4 h
compiling and putting the firmware on it: 0.5 h

So that is all together about 42 hours (what a nice number, I promise I did not tweak it

).

As for as the price. If you already have all the tools needed (including an FFF 3dPrinter) then you should be able to fit into 100 €. The things which can vary in price a lot are: the case (if you need to get it printed for you), the switches and the keycaps (you need (80+80) of these so there is quite a difference whether a piece costs 0.15€ or 0.8€).

Kko ordered ATxmega64a4u. Hopefully in about 3 weeks I'll know whether it comes with a bootloader preloaded and whether it is usable. Its only serious disadvantage is that it has only 4 kiB or RAM. That means that about 500 B needs to be removed from the RAM usage. Looks doable.

vvp · 19 Jul 2015, 15:10

@CrimsonFlame:
I added more remarks about acetone painting:
http://www.hck.sk/k80cs/index.html#acetone

kiltannen · 21 Jul 2015, 23:39

vvp wrote: Katy's controller and LCD can be put in the left hand side by just flipping the controller PCB and wiring the connectors a bit differently.

Just wanting to confirm here - it is really just the Palm print job that needs to be "flipped" for this, is that right?

vvp · 22 Jul 2015, 10:04

kiltannen wrote:
vvp wrote: Katy's controller and LCD can be put in the left hand side by just flipping the controller PCB and wiring the connectors a bit differently.
Just wanting to confirm here - it is really just the Palm print job that needs to be "flipped" for this, is that right?

No, if you want to use the controller PCB on the left side then:

print all the case files mirrored along ZY-plane (the plane which is in the middle between the keyboard halves)
- this means that both palm parts will be both mirrored and turned upside down (rotated by 180°)
- the bottom parts and keywel parts will be only mirrored along ZY-plane
if you would use my original controller PCB then the PCB would need to be turned by 180 degrees and (when compared to the controller on the right side):
- all the through hole components need to be soldered from the opposite side of the PCB
  - this is no problem for components having 2 leads only (you can swap them so that polarity is OK even when they are on the other side of the PCB)
  - components with more leads -> all the pin headers will have swapped pin assignments (1<->2, 3<->4, ...) - correct this when soldering wires to the receptacles
- all the surface mount components would be soldered as if the controller stay at the right side
when you are using your own controller PCB (because of the ATxmega breakout board) then just wire it so that it works (i.e. based on your schematic) and somehow fits to the case

vvp · 02 Aug 2015, 01:41

Oops! BIG TROUBLE!
I forgot to push the patch which swaps oLoad and iData to github. This was changed in past to be compatible with the standard Atmel boot loader (and kko's v0.7 electronics does not have it). If you already have the electronics done and tried to connect left and right sides with the firmware from github then this could fry PC0 port on ATxmega or QH output of SN74HC165 (or both). I'm sorry. I'll patch this on github by the end of this weekend.

vvp · 03 Aug 2015, 01:47

It is fixed. If you use my schematic then you want to be sure to have a version which has this patch:
849d6b9e12d23090ee93747893f76af22b1040d8
swap oLoad/iData to be compatible with the default Atmel bootloader

vvp · 03 Aug 2015, 13:12

And I forgot to mention. ATXmega64a4u arrived. We put it on a breakout board during the last weekend and we found out that it did not come with a bootloader preloaded.
At Farnell, ATXmega64a4u is always more expensive than ATXmega128a4u, so I do not understand why ATXmega64a4u is even being sold. Maybe it gets cheaper with extremely huge orders (tens of thousands?).

vvp · 09 Aug 2015, 16:46

I finished my v0.7. Kko's "v0.7" was slightly different (oLoad/iData swapped and small differences in the top row non-pinkie keys).

Anyway, this is the same as the published files. The only difference is that the controller PCB has also the components with reference numbers 10 and higher installed (and not necessarily the same as it is proposed in the schematic). So if you built it then this is what you get:

: full.jpg (340.59 KiB) Viewed 19756 times

: right-right.jpg (305.46 KiB) Viewed 19756 times

: right-back.jpg (300.23 KiB) Viewed 19756 times

: right-left.jpg (372.97 KiB) Viewed 19756 times

Here it is how does it look inside:

: inside-left.jpg (578.96 KiB) Viewed 19756 times

: inside-right.jpg (556.25 KiB) Viewed 19756 times

The left and right side PCBs are attached with a Ø 3 x 7 mm screw in one corner (the other corner is inserted into the corresponding slot). The controller PCB has one more optional screw. LCD PCB has metal mounting pads on one side. These were cut away and the PCB corners on that side are inserted into the little slots on the case. The other side is supposed to be mounted with Ø 2 x 7 mm screws (but I did not have screws so I used hex slot bolts). The top and the bottom of the case are connected with M3 x 6 mm hex slot bolts.

Columns of the switch matrix are connected with Ø 0.25 mm bare copper wire. Rows are connected with Ø 0.2 mm enameled copper wire. This way one does not need to bother with stripping insulation.

: inside-wiring.jpg (635.75 KiB) Viewed 19756 times

It works fine. This was written using the keyboard. I was asked to post some reachability pictures. I'll try to post them the next time. Hopefully the next weekend.

vvp · 29 Aug 2015, 23:14

Ok, the reach-ability pictures. If I would know it is so much booring work I would not agree to post them. Me, stupid

Here is the layout so that we know the positions of keys I'm talking about.

Spoiler:

The pictures are done using a hand which has 195 mm from the tip of the middle finger to the first wrinkle on the wrist. So it is probably bigger than average. The home position is at Space-J-K-L-; keys. Despite that most pictures have thumb positioned on the Alt key. It is done that way so that the wrist position is always about the same. I keep my wrist at the rests in such a position so that I can just reach the nearest bottom corner of the Alt key. Well I should keep the wrist just over the rests but I'm lazy.
As for as the comfort to reach a key, the rows are ordered like this (from the most easy to the most hard): home row, row just below it, row just above home row, the bottom row, the top row.
So here are the pictures of the top row reach-ability:

Spoiler:

The outer pinkie column. The top two keys are not a problem. No wrist movement needed. The bottom two keys (enter, keypad shift) require a slight wrist rotation.

Spoiler:

The inner pointing finger column. Pause key is easy to reach without wrist movement. The two bottom keys (application, insert) require a slight wrist rotation.

Spoiler:

The outer thumb cluster keys. All easy to reach without any wrist movement. Of course the alt key is pressed only in the corner but that is not bothering me.

Spoiler:

The inner thumb cluster keys. Space key is the home key for thumb so it is the most easy

. Shift key is the second most easy to reach thumb key. Keypad Shift key is easy (still not writ movement). Win key is the worst from the 4 inner thumb keys. I do also a slight wrist rotation there.

Spoiler:

I typically do not do chords on one thumb cluster if the chord needs only two modifiers. In such a case, I press one of the modifiers on one thumb cluster and the second modifier on the second thumb cluster. But there are exceptions of course:

More complicated chords on the thumb cluster which use only one thumb (no other fingers involved). From these I use only Win-Shift and Ctrl-Alt. Both of them are easy to do. It would be easier to do LayerShift-Space chord than Win-Shift but LayerShift-Space chord is not useful to me. The keys in these chords have different height. The idea is to press the higher key from a side and with the very tip of the thumb. This compensates for the keycap height difference. I do a tiny wrist rotation for Win-Shift chord.
Shift-Space and KeypadShift-Win chords would be much easier with 3dPrinted keycpas which close the space between them but I do not need these chords so I did not bother with custom keycaps. Both of them require significant wrist rotation and the second one also a small movement. Edit: With the special keycpas, there would not be any wrist movement for Shift-Space chord and there may be a slight wrist rotation or movement for the KeypadShift-Win chord.

Spoiler:

Chords on the thumb cluster I never do but that can be reasonably done. PgUp-PgDown chord is actually very easy to do without any wrist movement but it is not useful. Shift-PgUp and Win-PgDown require wrist rotation, the second one quite significant and maybe even a movement.

Spoiler:

Extreme chords - i.e. the ones which require more than 2 modifiers. I do all the modifiers with one hand by moving hand to the thumb keys. I use thumb for one key and fingers for the other thumb keys as necessary (one finger per key). The other hand presses the final key of the chord. This way complicated combinations like Ctr-Alt-Shift-F3 (or because of layering it is actually Ctrl-Alt-Shift-LayerShift-3) are easily doable. This is possible thanks to the fact that all modifiers are on thumb clusters. I never regretted moving Shift to the thumb cluster

Spoiler:

Edits in italic.

lydell · 30 Aug 2015, 11:07

Wow! Thank you so much for that extremely thorough showcase of you keyboard! It was way more than I expected, and is extremely helpful.

I really wish I had a Katy now …

Is that the 0.7 version? To me, it looks very finished. What’s up next? Will there be a 1.0?

vvp · 30 Aug 2015, 13:58

Yes, that was done with version 0.7.
I'll probably have one more tester locally who would like to built it himself. There is an interest. We will see whether he will actually do it. I'll provide only the raw printouts and help with the controller PCB. It would be interesting since he has a bit smaller hands. I'll will not do any changes to the keywell and the thumb cluster shape/placement (or any hardware change at all) at least till the end of this year but probably much longer (one more year?). I need to collect more feedback from myself and hopefully some more local users.

Anyway, there are some things as for as hardware which I'm considering even now:

Moving thumb cluster lower (i.e. nearer to the table). It can go lower about 4 mm without causing conflicts with the other parts inside. I do not know whether I'll actually like it lower but if not I can easily compensate with taller keycaps. So it is not a risk to do it.
Changing row spacing from the current effective average of 18 mm to about 16 mm. It is 18 mm on average at the tops of keycaps now. It is more at the base of the switches but the curved shape makes tops nearer than the bottoms. This could be useful for people with smaller hands but this also means that standard keycpas would not fit. It would need special keycaps just for Katy. The current row spacing is the smallest one which will work safely with common keycaps: DCS SP, DCS OEM, DSA SP. Because of the special keycap requirement there is a high chance this will not be actually done. Despite possible request from people with smaller hands.
Maybe add a trackpoint. Probably between N-key, ArrowDown-key and Space-key. But on both sides of course. That small void triangle there may provide enough space for it.
Maybe a built-in USB hub. There is enough space for it.
ESD procection .... nah ... just do not touch the exposed pins on the USB cable or the cable connecting the two halves just after you put on/off your favorite elecrifying pullover. (Some ESD is already built in the chips used. This would be just a stronger ESD on top of it.)

Changes I plan to be doing in the next half a year (or maybe much longer) are related to the firmware and the PC application.

The PC application (*) does not work on windows. It works well on linux. I'll make it work on windows and I'll make the strafe jump macro ... finally!
There are also some bug fixes and improvements in the firmware. It is not serious so I did not bother with them till now.
The password manager in the keyboard ???

(*) The layout picture in my previous message is just a part of a screenshot from the PC application - the tab which allows to view and redefine the layout. Of course on-the-fly remap/macros/layers work without any PC application.

As for as the hardware changes I mentioned: Since this is almost an open source project then there may not be any version 1.0 ever

There is about 50% chance there will be version 0.8. The problem is that what I have now works extremely well for me. So my motivation is not huge. Eeee, I'll almost for sure design some stand which will support small tilt to the sides but that does not warrant a new version number. It is just an additional stand. I think it may be useful but I do not miss it enough to be in a hurry.

As for as the software changes I mentioned: These will be done with something like 90% probability.

vivalarevolución · 30 Aug 2015, 14:02

Thanks for all the photos. It looks great.

Malkavian · 15 Oct 2015, 23:02

@vvp It's fucking awesome what you have done! Really, I love it! I've wanting a Maltron keyboard (and I don't like Kinesis one) since about 10 years, but then I haven't the money, and now... your keyboard is much better and probably much cheaper even if I don't move a muscle (I have neither a 3D printer, neither good electronics and soldering skills but have some skilled friends and you offered to print and sell pieces...)

I would like to have your great keyboard with Maltron L90, L89 (or maybe a mix) spanish layout. I'll be here impatiently waiting XDD

Thank you for such a great work

vvp · 16 Oct 2015, 20:46

Thanks.
Yes, I can provide the raw printouts of the case (about 65€) and the PCBs with parts soldered in and the firmware/bootloader loaded in the ATxmega controller. I can do a few finished cases but that would be rather expensive (about 160€). This is all done on an FFF 3dPrinter. The quality is bad compared to injection molded parts. Look at the pictures in the history to have an idea about quality. I'll be printing a case this or the next weekend. I can post some pictures.

I'm not sure about the price of the finished PCBs. The idea is: about 20€ more than the sum of the price of the components. That is for small amounts when it does not make sense to give it to a fab. That would put it very approximately somewhere around 40€. I'll do a BOM and then I can tell the price. The most expensive item there are the PCBs (about 8€). The PCBs are really expensive but that is the price from a local provider in small quantities (only 12 pieces were ordered). Price from DirtyPCBs should be around $3 per set but I did not order there yet.
The PCBs with ATxmega soldered (and firmware/bootloader pre-loaded) can ship only to European Union. Shipping outside would require to get an export permit and that looks like too much paperwork.

Even with these things done one needs to finish it: finish the case, put in switches with diodes, LCD, solder the matrix and the cables, connect the cables and screw in the screws/bolts

For now, it is primarily "do-it-yourself". And it can be done very cheaply if there is an access to an FFF 3dPrinter, soldering equipment, PDI programmer, and skills. And a lot of time

Btw: here are the expensive PCBs (as you can see I forgot to order the silk layers):

: pcb-front.jpg (57.67 KiB) Viewed 18581 times

: pcb-back.jpg (73.87 KiB) Viewed 18581 times

As for as the progress. I'm not sure I did anything from my last post. Well the PCBs were ordered and they arrived and I enabled photosensor and the LED brightness is now controlled based on it (this patch is not yet on github).

As for as the layout. One can define any layout one wants. All the keys (except program, and keypad shift/switch) are on-the-fly programmable as well as programmable from a PC application (working on linux only for now). Modifying physical location of program and keypad shift/switch can be done only by modifying the firmware source code. Chrisandreae posted a patch which allows to reassign these special keys from the PC application (i.e. no source code modification will be needed) but I did not merge it yet. As for as the keycaps with proper labels - I do not intend to solve this (definitely not in near time). People who do not care about keycap labels can get keycaps extremely cheaply from 7bit. It can be as low as 0.1€ per piece. Well, at least till his overflow stock sells out.

vvp · 17 Oct 2015, 15:18

For people who have an FFF/FDM 3dPrinter available: You can print the keycaps. I do not recommend it considering the price of unwanted/unpopular keycaps from 7bit. But if you would want to try it then here is a FreeCAD model of a keycap which is easily printable and works well enough. The license for the keycap model is public domain. Do whatever you want to do with it. I do not claim it is good for anything.

vvp · 04 Nov 2015, 01:20

The controller PCB populated:

: pcb-back-populated.jpg (157.87 KiB) Viewed 18482 times

: pcb-front-populated.jpg (110.05 KiB) Viewed 18482 times

It took about 3.75 h to solder parts to the right hand side by a person with significant soldering iron experience and only limited hot-air experience (about 20 h ?). All SMD parts took about 0.75 h using hot-air and the rest was the through-hole parts using an iron (lot of work with splitting pin headers to the required size and clipping the leads). Left hand side PCB took about 45 minutes.

Edit: The photo sensor support is pushed to github.

And the pictures of the raw printouts:

bottom part underside
Spoiler:

raw-bot-bot.jpg (376.19 KiB) Viewed 18482 times
bottom part upper side
Spoiler:

raw-bot-top.jpg (455.87 KiB) Viewed 18482 times
palm part underside
Spoiler:

raw-palm-bot.jpg (204.94 KiB) Viewed 18482 times
palm part upper side
Spoiler:

raw-palm-top.jpg (257.92 KiB) Viewed 18482 times
key part underside
Spoiler:

raw-key-bot.jpg (343.8 KiB) Viewed 18482 times
key part upper side
Spoiler:

raw-key-top.jpg (313.88 KiB) Viewed 18482 times

Zekromtor · 05 Nov 2015, 06:55

Man, you're still at it

Looks great!

CrimsonFlame · 05 Nov 2015, 15:40

Awesome! Glad to see you pushing ahead. I've had to shelve mine for a bit, until I've saved up enough to spring for the parts needed (pretty much everything but the printed parts

). Though I finally got my hands on a key tester and found the keys I like.

Ray · 05 Nov 2015, 18:49

Thanks for taking those detailed pictures. They really make me want to try something like this before I go on another take of a keyboard for my hands. They really show some ergonimic advantages very directly.
I will make another read through the thread to find a price estimate and see what the buzzer is used for

vvp · 06 Nov 2015, 00:27

Zekromtor, CrimsonFlame: Yeah, I'm still working on it, kind of. Last three weeks was a great weather here. Very beautiful autumn. Whenever I can, I ride my bicycle in the forest behind the town. The result is that I spend only few hours a week working on the keyboard ... if at all. But winter will come. That means a lot of mud and therefore no riding in the forests ... and maybe more time to spend on this project. I still plan to get the strafe jump macro working

Ray: Well, you can look in history or wait a bit more. I plan to post BOM. It will contain some ranges for prices of parts. The only problem is that I'm not really sure when I'll do it.
The buzzer is for:

Beeping when a key is pressed. You can switch this on or off. It is somewhat useful when you use linear switches and do not have any audible feedback when the actuation happened. Tactile switches do not need it much (there is a tactile feedback). Clicky switches are loud enough on their own. This beep is rather quiet now but it can be done more audible.
More audible beeps when switching between keypad and normal layer and especially when some of the LED state changes (e.g. when caps-lock gets activated/deactivated). This is quite useful.
It also indicated some error states. I may have removed this. Or maybe it is still there. I do not remember. The error states are better indicated on the LCD now.

Malkavian · 06 Nov 2015, 02:11

Ummm, since it have much less keys than a Maltron keyboard, could be not valid for me. In Linux "F" keys have more uses. And I want an spanish layout wich uses more symbols than english one... I will think about the possibilities...

Anyway your work is f***ing great!

vvp · 16 Nov 2015, 22:27

sva (the one with smaller hands) has built his Katy v0.7. I guess it is worth a post since it is the first version which was not built by me or kko. Me and kko do not count since we co-developed it

Spoiler:

Edit: And one more contoured keyboard you can build yourself.

vvp · 24 Nov 2015, 21:52

Ok, here is an example BOM for Katy.
The prices are approximately valid if you have free access to all the tools needed (notably it is an FFF RepRap 3d printer, soldering tools, and a PDI programmer). There will be significant price variability especially because of:

Parts. You should skip all parts with reference number 10 or more. That will save a tiny bit. You can skip all the pin headers and the receptacles and solder wires directly to the PCB. That will save a bit more. You can patch LCD hole in the case and skip the hole LCD (which is almost whooping 10 €). But this is all just peanuts compared to the items which follow.
PCBs. The price of one PCB set varies a lot based on the number of PCBs ordered. It was about 8 € per set for me. But if I would want only one PCB set then it would be about 40 €. Or you can home etch them too. They were designed so that they are easy to etch at home. In such a case you can have them for about 2 €.
Switches. You can get Gaterons even cheaper than 0.25 € per switch. Or you can get Cherry MX switches for much more than 0.25 €.
Keycaps. There is almost no upper limit here if you want to be fancy.
Case. If you do not have any access to a RepRap then the lowest price is probably about 65 € and not 24 €. Or you can try to get a crappy cheap Chinese filament and maybe you can print it below 20 €. Or you can order it from an SLS printer. It will look great but price will be around $ 270.