Katy keyboard (or K80CS (Key80 Contoured Split))

Kaibz

26 Jan 2015, 14:45

Thanks a lot for answering i really appreciate it.
Personally i do like the light sensor implementation.
Looking forward the future satus reports.
Cheers!

User avatar
vvp

12 Mar 2015, 14:03

Just to let you know that work is still being done on this.
Current status:
  • electronics is done (maybe I'll add one power filtering capacitor on the left side; it looks like it is not needed but at least a footprint can be added to the PCB)
  • v0.6 firmware ported to v0.7 electronics
  • keywell shape is fixed; there are these changes to keywell compared to v0.6:
    • smaller key spacing by 0.7 mm (it is 18.8 mm now)
    • steeper bottom and top rows
    • more stager on pinkie columns
    • the two top pinkie keys moved nearer
    • make it work better with DSA keycaps
As a result the case is about 2 mm taller, 5 mm narrower and the top two pinkie keys are about 5 mm nearer than they were in v0.6.
It still looks like it will be available by this midyear.
Some pictures of the test switch plate are below. Every time I change the plate shape I print mock-ups like these to check it out. And the renderings of the right hand side PCB. I do not have the real PCB at hand now.
top.jpg
top.jpg (78.77 KiB) Viewed 3016 times
front.jpg
front.jpg (53.35 KiB) Viewed 3016 times
left.jpg
left.jpg (51.38 KiB) Viewed 3016 times
right.jpg
right.jpg (56.24 KiB) Viewed 3016 times
dsa.jpg
dsa.jpg (60.95 KiB) Viewed 3016 times
controller-f.jpg
controller-f.jpg (66.74 KiB) Viewed 3016 times
controller-b.jpg
controller-b.jpg (78.01 KiB) Viewed 3016 times

lydell

14 Mar 2015, 20:05

I just wanted to say that this is the keyboard of my dreams. Truly impressive!

It would be very interesting to see some pictures with a hand on the keyboard, for example to see how well the pinkies can reach their keys or how easy the thumb keys are to press (without accidentally pressing others at the same time).

User avatar
vvp

15 Mar 2015, 22:06

I'll do the pictures of reach-ability when the first version of v0.7 is done.
The rest of the case is being worked on just now.

kiltannen

05 Apr 2015, 08:03

Hi - Massively impressive effort.

I am dead keen to make one of these - and probably end up with 1 at work as well.
The default definitions do not really matter since one can remap key positions easily using Prg-R <src> <dst> and the LCD shows small hints what needs to be done while remapping. Later, the layer definitions will be changeable/exportable/importable also using a special client application. That will be also used for advanced timed macros (e.g. a strafe jump sequence in a keyboard) and possibly also custom LED and LCD handling. But that is far future.

I am a little curious about some things:
  1. how far future?
  2. You show a QWERTY layout - I'm most interested in having the letter keys layed out with a Maltron style layout (or at least Dvorak) is that likely to be available right off or is that something I could only do "far futue"?
  3. you've gone with a keycap layout where the keys closest to your wrist are higher than the ones in the middle. How sure are you this is better than gently rising up?
  4. have you experimented with making some kind of lip in the keyswitch hole to latch the switch rather than gluing?
  5. Have you thought about including a USB port? Would be really usefull having one on the kbd for popping in thumbdrives... of course I don't have any idea how that would complicate the firmware &/or the circuit board...
Like I said quite interested, and the timing you are talking about probably works OK for me. Im not in a massive hurry.

I really like the idea of the printed case - it gives possibility for a better shape than the Massdrop acrylic approach. I do have access to getting print jobs done but dont yet have modelling skills...

User avatar
vvp

05 Apr 2015, 14:03

kiltannen wrote: Hi - Massively impressive effort.
Thanks. It turned out to be a bit bigger project than I originally intended. Especially because I did not expect the need to replace controller so soon. And I did not expect to replace CAD software too.
how far future?
  • Modifying the layer definition and macros from a PC application (store/restore layout) - this midyear.
  • kko wants to add a password manager to it soon - probably midyear. This will also use the AES in the controller. AES in the chip has also its bad consequences since USA puts export restrictions on the chip. So ATxmega chip availability is not world wide.
  • Me: strafe jump macro as a test case for advanced timed macros - somewhere between midyear and end of the year.
  • Probably by me: "unlimited" layer count (layer rename/copy/modify only from keyboard (no need for PC application)) - end of year.
  • Possibly me (maybe I'll rather add a trackpoint): custom LED/LCD handling - end of year or later.
All of this is probably not much work at all, but my free time is limited.
The complete firmware will be on github (hopefully within about a month), so any developers are welcome to hack it whatever way they want. For example, any LUFA based firmware for ATmega would be very easy to port.
You show a QWERTY layout - I'm most interested in having the letter keys layed out with a Maltron style layout (or at least Dvorak) is that likely to be available right off or is that something I could only do "far futue"?
You can use whatever keycaps you can source. Since it is already contoured, then I think DSA keycaps are acceptable. And that means just sourcing some labelled DSA set. One can put a keycap with a given label to any place he/she likes. All the keys use standard 1u keycaps so they are all interchangeable (if you use DSA). If there is enough interest we can organize a group buy. But so far from the responses I got it does not look like there will be enough people for a group buy.
you've gone with a keycap layout where the keys closest to your wrist are higher than the ones in the middle. How sure are you this is better than gently rising up?
I think it is better. Both Maltron and Kinesis have it that way too. Your finger will move in a arch around your knuckles. Keys should be in a similar arch. I cannot imagine how more typewriter style slope for the bottom rows (and possibly changing only switch angle) could work better but I did not even test this kind of shape. I tested a lot of variations but all of them had the general shape outline about the same.
have you experimented with making some kind of lip in the keyswitch hole to latch the switch rather than gluing?
A groove for CherryMX clip is in the model. But because of the precision of a reprap FFF method is not that good, it happens that the switch will not hold strong enough only using its clips. I believe that SLA or SLS 3dPrinters can print the case precisely enough so that gluing is not needed. But I targeted the the most common 3dPrinters (FFF/FDM method). One could make the switch hole a bit smaller so that it is harder to put the switch in. But then you run in a risk that the switch housing will deform and the switch stem will not move freely (may get struck at the bottom position after it is pressed.
Have you thought about including a USB port? Would be really usefull having one on the kbd for popping in thumbdrives... of course I don't have any idea how that would complicate the firmware &/or the circuit board...
Yes, but not for thumb-drives but for a mouse connection. The idea is that then a timed mouse/keyboard macro could be recorded directly on the keyboard instead of "typing" it in as a macro from a PC program. I decided not to complicate the first versions with USB host controller on the keyboard itself.
I really like the idea of the printed case - it gives possibility for a better shape than the Massdrop acrylic approach. I do have access to getting print jobs done but dont yet have modelling skills...
The STL files will be available. I wanted to provide also the STEP files but I was forced to switch FreeCAD for a cheap commercial CAD (*) and that does not allow exporting STEP files. If this would be a huge success so that it earns few thousands I would need to for CAD upgrade, then I could upgrade my CAD and provide STEP files. But based on responses I got so far there will be less than 10 users of Katy (not counting some people who are local to me).

(*) The problem with FreeCAD is that it often does not replay history correctly. So it is hard to use for more complicated shapes.

User avatar
vvp

05 Apr 2015, 14:20

And kko made some progress with his very first Katy v0.7.
Here are the pictures of an unfinished right hand side of the keybaord.
top-v0.7.jpg
top-v0.7.jpg (80.71 KiB) Viewed 2865 times
bot-v0.7.jpg
bot-v0.7.jpg (98.34 KiB) Viewed 2865 times

iN7h33nD

07 Apr 2015, 04:27

Just wanted to quickly post and express interest in this keyboard. It looks awesome and is exactly what I am looking for. Thanks for all the hard work your awesome!

kiltannen

08 Apr 2015, 08:12

Do you think it would be hard to put backlighting under the keys?

If I do go to the effort of building out a custom keyboard (And I'm FAST becoming unhappy :cry: with a standard kbd!) I think I would want the keys backlit - I use my PC a lot late at might - and would FAR rather have the light off - but don't yet touch type. Not likely to change either...

I tend to use 2 fingers from my left hand 2 fingers and a thinb from my right hand.

I would kind of like to learn to mostly use my left hand to type - leaving the right for pointing device work. BUT that would require a customisable keyboard WAIT that's the idea here! ;) another reason to take the plunge :D

I do have a friend with a 3D printer. He gets pretty damn awesome resolution so that should be covered. I'm not afraid of a soldering iron. My friend also has a CNC and can "etch" PCBs that way - or I could just go classic etching.

Either way - pretty keen to get started when you've actually got a working set of firmware...
Last edited by kiltannen on 16 May 2015, 13:17, edited 1 time in total.

pcaro

08 Apr 2015, 09:45


User avatar
vvp

08 Apr 2015, 11:34

kiltannen wrote: Do you think it would be hard to put backlighting under the keys?
The LEDs can be inserted into each switch housing. That is the easy part. Now there are two options:
  • All the LEDs lit the same way all the time the keyboard is powered - you can easily add this still using the PCB designed for v0.7.
  • Controlling brightness of each LED separately (e.g. so that you can display a key use heatmap). It is better to design a new PCB for this and replace the controller with something which has more GPIO lines or use a separate chip for LED driving. Also the connection between right and left side would need to change (probably to SPI). To pull this out one needs to have some knowledte how to design circuits. PCB mount points in both halves would need a modification too but one does not really need CAD skills fro this. The existing ones can be cut off a new ones glued in.
If I do go to the effort of building out a custom printer (And I'm FAST becoming unhappy :cry: with a standard kbd!) I think I would want the keys backlit - I use my PC a lot late at might - and would FAR rather have the light off - but don't yet touch type. Not likely to change either...
I tend to use 2 fingers from my left hand 2 fingers and a thumb from my right hand.
I would recommend to learn to touch type. Especially if you would like to use split keyboard so that you can adjust angles better. When touch typing you hardly ever look at the keyboard anyway so backlit is unimportant and some people find shining LEDs even distracting.
I would kind of like to learn to mostly use my left hand to type - leaving the right for pointing device work. BUT that would require a customisable keyboard WAIT that's the idea here! ;) another reason to take the plunge :D
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. The other side (which does not have controller/LCD) does not need to be connected. By defining two layer shift keys you can also fit in all the keys. Firmware would need to be modified for this. Simple configuration of the current firmware is not enough.

The idea of the current firmware goes like this. More layers (in the future, by midyear there will be only one layer). Each one layer has a direct mapping and keypad mapping (so one layer is actually two: normal and keypad). To access the keypad mapping one can use KeypadSwitch or KeypadShift. The keyboard has 80 keys. 1 used fixed for Program button, 1 used for KeypadSwitch, 2 used for KeypadShift. The rest 76 keys can have different functionality in normal layer and in the keypad layer.
I do have a friend with a 3D printer. He gets pretty damn awesome resolution so that should be covered. I'm not afraid of a soldering iron. My friend also has a CNC and can "etch" PCBs that way - or I could just go classic etching.
That is enough. The printer needs to have at least 19x18 cm build area to print this. It can be somewhat smaller if the printer has a heated chamber (of if it is LSA/SLS printer). If he/she will print the case for you for the material cost then you can built this very cheaply. Definitely below $200. Very likely below $100 if you wait for good group buys for switches and keycaps. And a lot of work. Your friend will probably agree to print the case for you for the material cost only. It is just starting the printer and letting it run. One does not need to babysit it. But if it is an FFF/FDM printer then the printed parts need some manual post-processing (cleaning up support and the two top parts need to be glued together). I doubt anybody will do the post-processing for free :). If he/she has SLA/SLS printer then there is little post-processing work.
Either way - pretty keen to get started when you've actually got a working set of firmware...
Firmware currently works as a keyboard. What needs to be done in the near term is:
  • possibly adjust messages to LCD,
  • lit the 4 LEDs added,
  • control brightness based on photo sensor,
  • maybe modify the control communication protocol with the PC application,
  • adjust the PC application itself,
  • make macros and macro recording work.

User avatar
vvp

08 Apr 2015, 11:58

pcaro wrote: If could be printed using this...: https://www.kickstarter.com/projects/ti ... 3d-printer
Based on this picture:Image
The print area is too small.
The minimum required print area is 19 cm x 18 cm without heat chamber.
If the printer has a heat chamber then it is 18.5 cm x 17 cm.
One could possibly do it by splitting the parts into more pieces and then glueing them together. But it is a lot of added manual work and some 3D CAD design work too.
Better find somebody who has a 3dPrinter around you locally. A lot of them has build area of about 20 cm x 20 cm. This is probably the most common bed size since original repraps started with it.

pcaro

08 Apr 2015, 15:17

Thanks @vvp for the info. I will look for a friend with 3dPrinter. Wishing see your project when finished. You are building a great keyboard.

kiltannen

09 Apr 2015, 01:25

kiltannen wrote: Do you think it would be hard to put backlighting under the keys?
vvp wrote: The LEDs can be inserted into each switch housing. That is the easy part. Now there are two options:
  • All the LEDs lit the same way all the time the keyboard is powered - you can easily add this still using the PCB designed for v0.7.
  • Controlling brightness of each LED separately <snip>. It is better to design a new PCB for this <snip>
I'm definitely thinking option 1 is the way to go. Maybe with a choice to turn all on or off via Keystroke of some kind...
Option 2 looks like being a never ending TCWB (Time Consuming Work Bomb) of a project

kiltannen wrote: don't yet touch type. Not likely to change either...
I tend to use 2 fingers from my left hand 2 fingers and a thumb from my right hand.
vvp wrote: I would recommend to learn to touch type.
BUT wouldn't touch typing if learnt using a custom layout make it REALLY hard to use a standard qwerty layout?
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. <snip> By defining two layer shift keys you can also fit in all the keys. Firmware would need to be modified for this. Simple configuration of the current firmware is not enough.
Sounds a little bit scary - like another TCWB. But maybe a worthwhile one AFTER having a working Katy up.
vvp wrote: The idea of the current firmware goes like this. More layers (in the future, by midyear there will be only one layer). Each one layer has a direct mapping and keypad mapping (so one layer is actually two: normal and keypad). To access the keypad mapping one can use KeypadSwitch or KeypadShift. The keyboard has 80 keys. 1 used fixed for Program button, 1 used for KeypadSwitch, 2 used for KeypadShift. The rest 76 keys can have different functionality in normal layer and in the keypad layer.
This sounds sensible - and quite useful. The one layer you expect available by midyear - will it be possible to design my own layout for the standard keyboard? EG have it as a DVORAK style layout? I'm guessing this would be a configuration of the firmware?
vvp wrote: The printer needs to have at least 19x18 cm build area to print this.

Build Area 240mm x 250mm x 170mm

vvp wrote: Firmware currently works as a keyboard.
Hmmm. Does this mean you have decided what processor you are going with for 0.7 and beyond? And are using that currently? And as it currently stands as a keyboard can I do the custom key layout I'm keen on? Or would I need to wait for the midyear firmware point?..

User avatar
vvp

09 Apr 2015, 11:42

kiltannen wrote:
vvp wrote: The LEDs can be inserted into each switch housing. That is the easy part. Now there are two options:
  • All the LEDs lit the same way all the time the keyboard is powered - you can easily add this still using the PCB designed for v0.7.
I'm definitely thinking option 1 is the way to go. Maybe with a choice to turn all on or off via Keystroke of some kind...
You can either add a small on/off switch to the case. This is the most easy option since it does not require any firmware modification.
The other option is to scratch the trace which leads from port PB0 to photo sensor power. Connect that port to a transistor input to amplify it. Connect the backlit LEDs to the transistor output. Connect the photo sensor sensor power directly to 3.3V. Modify the firmware so that you can control the backlit. Since you want backlit you hardly care about power consumption so this change is acceptable and most easy (if you want software control over backlit).
kiltannen wrote: BUT wouldn't touch typing if learnt using a custom layout make it REALLY hard to use a standard qwerty layout?
No, a lot of people (including me) touch type both on kinesis advantage and a standard row staggered keyboard. If you select layout as similar as possible between these two keyboards the skills are transferable almost in full. The only thing needed is about 5 evenings of typing exercises when doing the switch the first time. You can start learning even now on a standard keyboard and "transfer" the skill to Katy later.
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. <snip> By defining two layer shift keys you can also fit in all the keys. Firmware would need to be modified for this. Simple configuration of the current firmware is not enough.
Sounds a little bit scary - like another TCWB. But maybe a worthwhile one AFTER having a working Katy up.
This is actually easier than the hack to control backlit power from port PB0. When building the new Katy which would have controller on the left side, you just flip the PCB and connect the connectors a bit differently. You do not need to hack anything (scratch the PCB traces). You just connect connectors differently and the through hole parts will be soldered from the other side of the PCB. And you mirror the right hand side case before printing. That is it.
Edit: Of course, if you want more than one Keypad Shift (i.e. more than normal/keypad layers on that half of the keyboard) than you need to modify firmware. Otherwise the firmware can stay as it is.

kiltannen wrote: This sounds sensible - and quite useful. The one layer you expect available by midyear - will it be possible to design my own layout for the standard keyboard? EG have it as a DVORAK style layout? I'm guessing this would be a configuration of the firmware?
Yes, you can do this even without the PC application being installed. You can remap the keys within the keyboard itself. This works even now. The only keys which cannot be remapped are currently keypad switch, keypad shift and program keys. Maybe I'll make them re-mappable on-the-fly later too. You can change their position very easily in the firmware source code though.
kiltannen wrote:
vvp wrote: Firmware currently works as a keyboard.
Hmmm. Does this mean you have decided what processor you are going with for 0.7 and beyond? And are using that currently? And as it currently stands as a keyboard can I do the custom key layout I'm keen on? Or would I need to wait for the midyear firmware point?..
Yes, controller in v0.7 is ATxmega128a4u. The future versions (if there will be any done by me) may use a different controller. The minimum keyboard functionality and on-the-fly remap work now. One can do custom layout even now. All of it needs more testing so maybe some problem will appear but so far so good.
The PC application is being worked on just now.
All the data files and sources are not published yet. They need a bit of clean up and some how-to files etc. So there is still fair amount of work.

User avatar
vvp

23 Apr 2015, 10:49

Good news! Other people are making their own contoured keyboards. Hopefully we get much more projects like these.
https://geekhack.org/index.php?topic=71263.0
https://github.com/adereth/dactyl-cave/

zemm

04 May 2015, 18:59

I love your design! I'm waiting for Kinesis to upgrade Advantage to get my first contoured keyboard, but yours looks so good I'd love to get one for home use. Not having a 3d-printer near me, I have been pondering the option of trying to carve one out of wood (a long project), and use your layout and firmware (at least at the beginning, and maybe at some point roll my own, being a lazy software developer).

I've thought of a few designs, but since I have zero experience of contoured keyboards, and I trust your experience and prototyping on the matter, I would be very interested to get my hands on your design at some point for measures etc. (willing to pay something if that's what you are planning to do) :)

Btw. do you think DCS keycaps on already contoured keyboard are considerably better? Does Advantage use those?

User avatar
vvp

04 May 2015, 23:53

Yes, the new Kinesis Advantage should be better. The rubber keys will be replaced with the real ones, and firmware will be upgraded. Some other changes may be there too but I do not know about them.
I do not care that much about the rubber keys. I do not use function keys often. Even in Katy, they are just fine for me when they are in a second layer if both layer switch and layer shift are available. But the current Kinesis Advantage firmware does not support keypad shift and there is not enough keys on the thumb clusters for me.

One of the options is getting the current Kinesis Advantage. You probably can get a used one or a new one cheaply on massdrop: https://www.massdrop.com/buy/kinesis-advantage
And then you can replace controller and firmware to get more options: https://geekhack.org/index.php?topic=56 ... msg1620078
There was one more project announced. It will replace the whole Kinesis Advantage controller PCB and should also support 8 keys in each thumb cluster. But I cannot find a link to it now.

Katy's firmware is based on chrisandreae's firmware. The most notable difference is addition of layer (keypad) shift keys. And update for ATxmega128a4u. The update is needed because firmware does not fit to 31kiB FLASH when program interpreter is enabled(*). The program interpreter is needed for advanced macros (e.g. a strafe jump macro).

(*) The firmware contains a small virtual machine which interprets a byte code.

I use Kinesis Advantage for about 13 years now and I'm quite satisfied with it. The two main reasons I started Katy is because it is not easy to add 2 keys to each Kinesis' thumb cluster and I wanted a split keyboard too.

I do not have experience with carving. It looks like a lot of work. I probably would not be able to do it at all. I plan to release design files for everything for free except the 3D model of the case (which will be around 15€, probably on shareit). If you are willing to commit so much effort to it that you would carve it yourself then there is an option that I could print a case and send the raw prints cheaper. There is a lot of post processing work to get a usable part from a reprap raw print. You would need to do all the support clean up, string clean up, possibly sanding (if you care about smoother surface), acetone painting (or acetone vapour bath), and gluing yourself. Definitely less work than carving. Send me a PM after v0.7 is released if you are interested in this option. The release date still looks to be around this midyear.

Katy v0.6 is really good for my hands with DCS keycaps. I modified the keywell shape in v0.7 so that it is almost the same with DSA keycaps as v0.6 with DCS keycaps (plus more modifications to the pinkie columns). If your hands are like mine then v0.7 with DSA keycaps will be just fine. If your hands are smaller you may prefer DCS keycaps even on v0.7. My hand is 19.5 cm long form the middle-finger tip to the first wrinkle on my wrist. I'll post some pictures how keys are reachable when my version of v0.7 is done.

Kinesis Advantage uses DCS keycaps. It is good to have them there since the switches are built along a thin bent PCB. They cannot use a different height for some rows. Since Katy is manually wired I can not only change angles between rows but I can raise some rows too. The top row is raised a bit (so that DSA keycap in the first row feels like top row DCS keycap).

zemm

05 May 2015, 18:22

Thanks for your offer. Due to postages etc. stuff I guess I could find someone more local if I would go down the printing path at some point.

I was mostly considering on trying to print out 2d slices on paper out of your design, mainly from the curves and the spread of the rows, and then carve the rows out of separate pieces (possibly so that the spread could be fine tuned). Each row could then also be swapped later with a better design if needed. It would still be a hard work to prototype this way, so already fine-tuned starting point would be great, especially if it would work with DSA-caps :)

Anyway, this will be a long zen-ish project, so v0.7 and the plans are probably already available once I really get to working on it. I ordered some cherrys and do some testing at some point when they arrive, mainly about the attachment, durability, wiring and would it work at all.

And of course I did not mean to "roll my own" firmware but more to dive in to the alternatives and see the need of possibly customization like you did. For example I don't care about the VM programs, and possibly macros either, but layers are a must and LCD would seem handy, so I might also take a look at the ErgoDox fw if they add LCD.

I was also wondering by intuition (and zero experience), would the thumb keys be more natural when tilted more to the side, like with 4+ mouse buttons, so your thumb would move more like into the hand? I haven't seen it be done, so maybe it's a bad idea...

Ps. my hand is exactly the same 19,5cm, so looks good :D

Edit: ErgoDox: s/if they add LCD/since they added LCD/... oh gosh darn, I just found out about the ErgoDox Inifinity (too late for now). If there will be next run at reasonable time, that might be my choice over Advantage. One more reason to try follow Katy / try to build one to get a contoured kb too :D

User avatar
vvp

05 May 2015, 21:06

If you are willing to build more or less your custom case and you do not need VM programs then your firmware will probably easily fit into 28 kiB of FLASH and in such a case you can even use some common controller sold on a PCB (like Teensy 2.0 or ArduinoMicro). Here is a nice list of free firmwares to start from: https://github.com/tmk/tmk_keyboard/blo ... rojects.md
Adding an LCD support to them is easy. In the worst case you can copy-paste-modify from my branch of chrisandreae's firmware (I hope to push it to github in few weeks). Or from the future Ergodox Infinity firmware. Depends on what kind of LCD you use.

I do not know about carving. All the kinds of slices can be easily generated from an STL file if that would help you.
zemm wrote: I was also wondering by intuition (and zero experience), would the thumb keys be more natural when tilted more to the side, like with 4+ mouse buttons, so your thumb would move more like into the hand? I haven't seen it be done, so maybe it's a bad idea...
If you think you would like it that way then you can post questions to this thread: https://geekhack.org/index.php?topic=49721.0
I did not use that approach because it is hard to use when you want more than 4 keys in one thumb cluster. It also makes the keyboard taller.
zemm wrote: Edit: ErgoDox: s/if they add LCD/since they added LCD/... oh gosh darn, I just found out about the ErgoDox Inifinity (too late for now). If there will be next run at reasonable time, that might be my choice over Advantage. One more reason to try follow Katy / try to build one to get a contoured kb too :D
I considered getting an ergodox instead of building Katy. The main reason I did not do so is that I think the bottom row of keys is almost unusable without it being raised and tilted forward. Also it's thumb cluster is more far away from the keywell than it needs to be. The original erogdox had an option to have 8 keys in one thumb cluster.
There is a lot of talk about ergodox thumb cluster here: https://geekhack.org/index.php?topic=62848.0

kiltannen

17 May 2015, 11:08

Do you think this board here: https://www.mattairtech.com/index.php/d ... db-x4.html would be an OK starting point?

I'm just a little bit concerned about doing the soldering after having a look at the chip in your pics. While I'm reasonably OK with the idea of a simple solder job - With the chip package I'd certainly be a teensy bit concerned about creating un-intentional bridges.

User avatar
eldorange

17 May 2015, 11:29

Kinesis Advantage and Ergodox KILLER!!!

Do you have plans to release this killer keyboard in the market?

User avatar
vvp

18 May 2015, 01:54

kiltannen wrote: Do you think this board here: https://www.mattairtech.com/index.php/d ... db-x4.html would be an OK starting point?
It is one of the best breakout boards I have seen because it is very configurable. Almost everything can be disconnected using jumpers and that is what you would need to do if you would decide to use it. The only thing you would use from it is the power source circuit, the reset button, and the USB connector. Everything else would need to be disabled and pulled out either to Katy's PCB or your own. Probably your own ... since Katy's PCB does not have pulled out ATXmega pins outside and connecting wires to the ATXmega chip pads does not really solve your problem of avoiding soldering to such small pads. You could take Katy's kicad files and modify circuit/PCB for your needs.

The only other problem with the breakout board is that it uses two 10µF capacitors around LDO (power converter chip). Provided that the LDO is not some special kind (I doubt it would be) then it breaks USB specification which does not allow more than 10µF filtering capacitor in the power circuit at start up. It is not a big deal though. If it will ever cause a problem then likely it will only happen when connected to a hub (the hub or some other device connected to it will reset) and you would need to disconnect and reconnect again.

That being said, I think you just should try to solder the chip in. Really, it is easy to do when you have a factory produced PCB with solder resist, and an iron with temperature control. And depending how old you are you may need a magnifying glass :-) Just do not skimp with flux, set iron to around 280°C, stay on one pin/pad only a few seconds, and it will go fine. There is a lot of videos no youtube how to solder a TQFP package either a pin by pin or with a single drag over one side. The point is that you can easily see whether the chip pins are shorted or not. The solder will be repelled by the areas between pins because of solder resist on PCB and the flux. So the difference between shorted pins and separated pins is very visible. If you still would not be sure then the chip has ESD protection so you can solder it as the first component and use an ohm meter to check whether you do not have any shorts. I think it is almost impossible the ohm meter would put the chip in danger. I did it often there was never a problem with it.

I can build a few PCBs for people in EU. I do not know about people outside EU since ATXmega contains AES hardware and exporting it needs some additional paper work which I do not have a clue about. I bet people in all the countries which have good relations with USA can buy it locally. It does not seem that it would be enough interest to do lager scale production (which would amortize paperwork, delegating PCB production to some factory etc.).

Actually a few days ago we did some additional work on the LUFA update and we found that its footprint lowered so it fits on ATmega32u4 again now (even with the LCD and the interpreter). Still going back to ATmega is not our plan since ATXmega is easier to work with, cheaper, and has more GPIOs in a package with the same pin count.

User avatar
vvp

18 May 2015, 02:15

eldorange wrote: Do you have plans to release this killer keyboard in the market?
Like selling a finished keyboard in a nice box with 2 years of warranty and everything?
There does not seem to be enough interest for that. We can do it but so far it does not look plausible. The point is this does not make sense to do when the market is probably below 100 and almost for sure not above 1000 pieces. Look at ergodox, it was sold few hundred of times, probably below 1000 and it is still a DIY kit only. And it is easier to do compared to Katy. Well at least Ergodox Infinity has the SMD components soldered in when bought :-)
I'll know more after v0.7 is out.

kiltannen

18 May 2015, 07:55

I asked the support guys if they can supply with headers on:
support@mattair.net wrote:I don't normally install headers, so that the user can decide which side to mount them on. If you would like them soldered, I can do that (no charge). I also have angled headers in stock, if you need.
Seems like this would solve 90% of soldering? Keeping the 128 chip would give me the ability (technically) to add some code to support software control of the LEDs I plan to install on the keyswitches. I think that might be beyond my appetite for work though... :? I will probably just go with a physical switch mounted in the case.

I'm starting to get quietly excited.

I do have a fairly long lead time on the 3d printing (& I plan to switch left to right to play with the idea of left handed only...), so are you planning to make changes to those files or are they finalised? If so I'd like to start work on the case - can you let me know how I organise for the files with you? Is that maybe through PayPal?

User avatar
vvp

18 May 2015, 23:23

Ok, I'll try to get it out as soon as possible. Maybe in parts and not all at once. So that you can look at it.
Unfortunately, as for as the case, I want to print one more copy before the release. KKO's v0.7 is working as a keyboard with remap. Macros, LEDs, photo sensor are still not enabled there. And there is a small change in the case which was not in KKO's version. I need to 3d print it here and check it before I release it.

Anyway, you want to do significant changes to the PCB (better connection to a separate ATXmega breakout board *) and possibly the schematic too (if you want software controlled back-light). So you are probably interested in the kicad files first anyway. Hopefully the next weekend (i.e. in about 14 days) I'll try to get at least that out. I want a few small improvements and more descriptive silkscreen.

* Without that it does not make sense to use a separate ATXmega breakout board because you would not avoid soldering of some tiny stuff when connecting the breakout board to the PCB. It would be easier to solder an ATXmega chip directly than connecting tiny wires to the TQFP pads. My opinion about that is: Who cares, even if you solder it incorrectly and damage the chip. It's only about 3.5€. No bid deal. The only other reason why to use an ATXmega on a separate breakout board is that it already has some boot loader loaded. So there is no need for a PDI programmer.

kiltannen

19 May 2015, 02:19

vvp wrote: Ok, I'll try to get it out as soon as possible. <snip> Hopefully the next weekend (i.e. in about 14 days)
:mrgreen: (I do realise you are offering to release the PCB designs in that timeframe - and haven't yet put a timeframe on the 3d files for the case.)
vvp wrote: you want to do significant changes to the PCB (better connection to a separate ATXmega breakout board *) * Without that it does not make sense to use a separate ATXmega breakout board because you would not avoid soldering of some tiny stuff when connecting the breakout board to the PCB. <snip> The only other reason why to use an ATXmega on a separate breakout board is that it already has some boot loader loaded.
SO - the guy with the ATXMega board has offered to solder on header pins - even offering my choice of angled pins. And yes - definately getting a board with a bootloader is attractive to me... That gives me the chance to maybe not use/ modify your PCB at all - but just put some components together on some kind of breadboard, I am thinking like this: http://img.dxcdn.com/productimages/sku_140968_1.jpg or maybe this: http://img.dxcdn.com/productimages/sku_140784_1.jpg. I should be OK to solder headers onto those breadboards myself & the other few components you have on your board. That also saves me trying to figure out how to etch, and while etching is cool - it is also an extra manufacturing step I think I'd rather not put effort into.

SO having said that - what are the dimensions of your PCB? I think I can get the breadboard on it's way if I pick something fairly close to your PCB in size. That outfit (http://www.dx.com) tend to take around 5-8 weeks to land anything in NZ so ordering sooner rather than later would be a good idea. I am also thinking I should get the LCD on it's way from them, do you have a specific one in mind? Or maybe a form factor I should look for?

kiltannen

19 May 2015, 02:26

On a different note - I have been trying to figure out keyswitches - and it seems like if I ordered 100 cherrys from WASDKeyboards.com that would cost approx USD $60 plus shipping. Do you have an idea if that would be a good price - or could I do better with trying to jump on a Massdrop?

I haven't yet got any real ideas about keycaps... :(

User avatar
vvp

20 May 2015, 00:36

The case STL files will be about a week or two after the PCB files. But maybe I'll try to post some kicad files sooner (since you do not need PCB then it does not need to be polished (especially the stencil layers)). But at least you will see the schematic and you can create your BOM. And you can modify schematic and BOM for your back-lighting LEDs.

Right, you can use a universal board only and wire components with a magnet wire. I actually did this on my first PCB for the left side (katy v0.6). You can find a picture of it earlier in this thread.

Right hand side PCB dimensions are 50.292 x 66.040 mm. Left hand side PCB dimensions are 25.400 x 46.355 mm. You can look at the pictures of the internals of KKO's right hand side (v0.7) and my left hand side (v0.6) to get an idea how much space is around the PCBs. But you probably want to wait with ordering since you should be able to order it all from one shop and you can save on shipping charges. The LCD used is FDCC0802C-FLYYBW-51LK. But again you want to order together with other parts. I got all the parts for PCBs from one shop (farnell).

You can get switches for 0.65€ per switch from 7bit here on deskthority. You can get cheap Gateron switches from Asian sellers (probably in the range of 0.15 to 0.3€ per piece). You can get keycaps from 7bit too (if you do not care for legends you can get a keycap as cheap as 0.1€). Group buys are cheap too but they take a long time. And you seem to be in such a hurry that I have a hard time to get some breathing time :-D

infowolfe

20 May 2015, 05:05

vvp, I just wanted to say I've been lurking for a couple months, but I'm very excited to see how the k80cs turns out. Have you by chance looked into jacobolus's thumb cluster stuff and HaaTa's firmware/KLL (keyboard layout language)? The PCB layouts on the Infinity and ErgoDox Infinity keyboard kits marketed by MassDrop were designed by Parak and the firmware is github/kiibohd/controller (and HaaTa, jacobolus and Parak are all in #deskthority on freenode if you wanted to drop in).

If there's a group buy on cases or PCBs, I'll be in for 3, two for me and another for a guy I'm assembling an ErgoDox for. I've also got 3 other friends who'd probably go for one each as well. Alternatively, if/when the layout stabilizes a bit more
I'd be down to pay someone in Austin, Texas with a 3d printer to print me out at least 1 case set. If a serious group buy happens though, I'd be more than happy to forward this thread on to everybody at work (we're a python scientific computing shop) and I'm pretty sure there'd be a couple more people 'in' from that group.

Anyway, keep up the good work... you've designed one of, if not THE most compelling "original" (not a modified ape of another open source design) keyboards in years. I dig the Maltron-ish key pitch/layout, the split halves, and the LCD for layer tracking/output is icing on the cake.

Edit: forgot to request... it'd be really super awesome if by 1.0 it had a hub and at least 1 'child' usb header, for Yubikey/2FA hardware token.

Post Reply

Return to “Workshop”