F104+SSK+122+62+77+50+Ergo orders now open! New Kishsaver+Industrial Model F Keyboards

[gipetto]

The details on the ergodox model can be found here: https://www.modelfkeyboards.com/product ... -keyboard/

Unfortunately, it operates like two separate keyboards without a direct connection between the two halves. I would be very interested if someone has a plan on how to fix this and have a shared matrix between the two halves. As it stands this (combined with the potentially hefty price) is a dealbreaker for me, I am very used to my layout that has multiple layers that stretch across both halves.

Other than that, I would be a dream come true and I would order one instantly if there is a way to retrofit some kind of solution.

hid-remapper can do this. you plug both split halves into a usb hub, and plug the hub into the rpi-pico. hid-remapper currently doesn't distinguish between devices, so both will appear as the same device, then setup a layer on a key. i think 4 layers is the limit.

there's software solutions on the pc that may work too. i've seen kmonad mentioned.

[Vizir]

it should certainly be possible as I have used split keebs from http://www.rgkbk.net and using split usb-c cables.

The link doesn't work, so just to confirm: do both cables go to the computer and no cable between the two halves of the keyboard? And then holding an Fn key (that is not Apple Fn, but rather a custom one defined) changes how keys behave on the other half where you are not holding the key.

For example, if you have mapped Fn + Tab to send Caps Lock key, does it work if you hold the Fn key on the right half and press Tab on the left half? It seems impossible without custom drivers (or some direct communication between the halves).

sorry for the delayed response. i have multiple splits from rgbkb and have used both a usb-c cable connecting the two keyboards together as well as a y cable which legonut from rgbkb sent. it appears he has removed the y cable from the shop. i'm not certain if the y cable had any custom wiring. using either the split or the separate usb-c cable between the 2 halves i can use one half to hold a layer/fn key and the other half recognizes that the layer shift has happened. i think this is built into qmk and the slave half communicates with the master half. come to think of it, even corne and other various splits do this. i hope my response was clear enough.

see this too
https://github.com/qmk/qmk_firmware/blo ... eyboard.md

[Arkku]

i think this is built into qmk and the slave half communicates with the master half. come to think of it, even corne and other various splits do this. i hope my response was clear enough.

see this too
https://github.com/qmk/qmk_firmware/blo ... eyboard.md

Yes, QMK indeed has support for master and slave halves of split keyboards... But as it says on the linked page, the two halves must be connected to each other. You also speak about a USB-C cable between the two halves: does one half have two USB-C connectors or how does that work? What happens if you connect each half to the computer with a normal USB-C cable, and don't connect them to each other?

(Meanwhile, as for the otholinear model F keyboard, it only has one USB-C connector per side, so the connection between the two halves would have to be done by some other means. The controller does have a pin header for LEDs and a solenoid, so that could be repurposed to link the two halves.)

[Vizir]

i think this is built into qmk and the slave half communicates with the master half. come to think of it, even corne and other various splits do this. i hope my response was clear enough.

see this too
https://github.com/qmk/qmk_firmware/blo ... eyboard.md

Yes, QMK indeed has support for master and slave halves of split keyboards... But as it says on the linked page, the two halves must be connected to each other. You also speak about a USB-C cable between the two halves: does one half have two USB-C connectors or how does that work? What happens if you connect each half to the computer with a normal USB-C cable, and don't connect them to each other?

(Meanwhile, as for the otholinear model F keyboard, it only has one USB-C connector per side, so the connection between the two halves would have to be done by some other means. The controller does have a pin header for LEDs and a solenoid, so that could be repurposed to link the two halves.)

let me see if i can find the usb-c y cable. i bought a couple and have given them away but might have 1 lying around somewhere. with only 1 usb-c connector per half, the master/slave option wouldn't work. might have to look at the controller to see if some exposed headers could be used... this might be tough.

edit: or add another controller (pro-micro or similar) which connects to both halves and acts as a passthrough/translator for the 2 halves and sends the translated keycode to the computer. ugh.

edit2: what gipetto wrote above!

[Ellipse]

Thanks gipetto; they both seem like good options to consider for the Model F Ortho. If anyone could recommend a comprehensive factory default programming for the ortholinear Model F (including function layers) that would be helpful. I am seeing various recommended configurations online for the Ergodox style boards and I don't know which one is most common. Also it would be great to know what would need to be programmed for the F Ortho with hid-remapper or kmonad in terms of a factory default.

Confirmed the F Ortho boards each use a regular USB-C controller, the same as what is used with the other boards. They do not connect to each other.

I'm guessing this kind of USB OTG hub would plug into a Raspberry Pi Pico? https://www.aliexpress.us/item/2255800246303987.html

genericusername57 the answers to your questions can be found in the manual on the project web site. There is a specific key combination shown there to adjust the solenoid extend/retract times. If you made your own firmware on the QMK beta site, you will need to update the firmware to allow for these key combinations (you can start with the json files linked to in the manual).

IBM capsense utility is only used for the original firmware, not for the current factory default QMK firmware, so it will not work.

[genericusername57]

genericusername57 the answers to your questions can be found in the manual on the project web site. There is a specific key combination shown there to adjust the solenoid extend/retract times. If you made your own firmware on the QMK beta site, you will need to update the firmware to allow for these key combinations (you can start with the json files linked to in the manual).

IBM capsense utility is only used for the original firmware, not for the current factory default QMK firmware, so it will not work.

I see, then maybe I'm just too fast for the solenoid Does adjusting the screws on the solenoid affect the speed of it or is it only volume? I have it set to maybe half because I feel that gives me the sound I want, but no matter how I adjust the dwell time back and forth I can't seem to get it to keep up with 100 WPM or thereabouts.

[gipetto]

Thanks gipetto; they both seem like good options to consider for the Model F Ortho. If anyone could recommend a comprehensive factory default programming for the ortholinear Model F (including function layers) that would be helpful. I am seeing various recommended configurations online for the Ergodox style boards and I don't know which one is most common. Also it would be great to know what would need to be programmed for the F Ortho with hid-remapper or kmonad in terms of a factory default.

Confirmed the F Ortho boards each use a regular USB-C controller, the same as what is used with the other boards. They do not connect to each other.

I'm guessing this kind of USB OTG hub would plug into a Raspberry Pi Pico? https://www.aliexpress.us/item/2255800246303987.html

That looks like it should work, though since I don't have any split keyboards or usb hubs on hand there's no way to know and it's not easy to try out. I've seen users mention they ran into trouble with unpowered hubs exceeding the current draw of devices, so there's a chance that it might not work, only way is to try it out.

you can see how to set the layers up here, On the upside, the pico has a lot of memory, so it might be possible to fork the code for more layers than 4. you can change middle button in this pic for any keyboard button.
https://twitter.com/jfedorynski/status/ ... 4635004930

There's not a lot of features in hid-remapper, since it's targeted at mousing devices but If the ergo split boards can run qmk, then hid-remapper may make up for the broken features due to lack of an inter split cable.

If anyone does try this out, you can export the settings to a .json file for backup and for others to use.

[Ellipse]

genericusername57 the solenoid can certainly keep up with 100wpm to my knowledge, if properly configured in the firmware. I recommend fully extending the bracket. The manual recommends a solenoid timing of 20ms. No firmware configuration is required if you are using the factory installed hex files or those provided in the manual. You would have to configure it if you hare making your own firmware on the beta QMK site. Also the solenoid uses a lot of power draw and I am not sure if certain USB ports may be underpowered.

[Ellipse]

Here are the first photos of the completed and tested M Style FSSK sample; it is configured with the HHKB Style Split Right Shift and Split Backspace. Since the ~ ` key is left of the 1 ! key, I was not sure which key to place in the rightmost key of the block, so I selected a blank key. Any ideas on how folks plan on configuring their FSSK layout with such a configuration?

For the M Style case keyboards, wcass has ingeniously updated the PCB design to allow for the same PCB to support both HHKB style split right shift and regular right shift, to cut down on the number of PCB variations.

As noted earlier, these are the rejected sample cases that have defects (casting defects, they did not drill all of the mounts and screws correctly so the inner assembly has sunk inside the keyboard, etc.). Fortunately the keyboard is fully functional in my testing with QMK today.

The case bottoms have 4 holes for optional screw-on bumpers, plus 2 holes to mount a solenoid. The case tops have two internal posts, one to mount a solenoid and one for the P clip (USB cable strain relief).

You can also notice the changes to the interconnections between top and bottom inner assembly
1. They now have additional tabs between the first two rows, just like the IBM F122 and F104 originals. I believe this was done to keep the plates as tight as possible, to allow for the snappiest of clicks upon actuation of each key. The Round 1 compact style F104 and FSSKs do not have these extra tabs. I can tell that the Round 2 boards have some extra clickiness/snappiness to them compared to the Round 1 boards, though both have sufficient clickiness/snappiness.
2. The controller tabs have been modified slightly on the M Style boards to make them simpler and a bit less flexible. Sometimes when unscrewing the controller and adjusting an old bottom inner assembly, the old tabs would move out of position which would require some pliers to move them back into position. They are also slightly taller, to allow for more room between the controller and BIA.

2023-04-26_16-27-52.jpg (1.47 MiB) Viewed 6937 times

2023-04-26_16-27-56.jpg (1.3 MiB) Viewed 6937 times

2023-04-26_16-28-06.jpg (1.56 MiB) Viewed 6937 times

2023-04-26_16-19-23.jpg (1.12 MiB) Viewed 6937 times

2023-04-26_16-19-48.jpg (1.24 MiB) Viewed 6937 times

2023-04-26_16-20-09.jpg (742.15 KiB) Viewed 6937 times

2023-04-26_16-20-28.jpg (1.35 MiB) Viewed 6937 times

2023-04-26_16-23-38.jpg (1.15 MiB) Viewed 6937 times

2023-04-26_16-23-44.jpg (1.51 MiB) Viewed 6937 times

2023-04-26_16-25-25.jpg (1.13 MiB) Viewed 6937 times

[cap]

Beautiful cases, Ellipse! I can't wait!

My preference for the HHKB configuration of these new keyboards is to configure the main block just like the HHKB. This leads to having redundant Esc keys on these new keyboards that have a function row. This preserves muscle memory with actual HHKBs or the F77 and F62, where there's no escaping putting the escape key left of '1'.

[sedevidi]

One of the keys on my F77 (H) was misbehaving more and more over time. We don't use it very much in french apparently. It had a lower sounding click. I had re-seated the key and spring many times, and finally removed the spring, which was a bit curved. It's able to roll regularly on a flat surface, though.
I replaced it with a new straight one, and the key is now 100% good.
So even a slightly curved spring is enough to make a very bad key.

[del20nd]

That case looks fantastic. It looks like the genuine article at a glance, which will make it a much more effective blunt force instrument than a Model M SSK. Your enemies won't expect the cast aluminum!

[Ellipse]

Here are the first photos of the completed and tested M Style F104 sample; it is configured with the US ANSI layout. As noted earlier, these are the rejected sample cases that have defects (casting defects, they did not drill all of the mounts and screws correctly so the inner assembly has sunk inside the keyboard, etc.). Fortunately the keyboard is fully functional in my testing with QMK. Also the texture is a little more visible in this photo than in the SSK photo though the textures are all the same. It is a nice texture reminiscent of the Model M style cases (originally plastic) as well as the IBM Model F cases (unpainted plastic, painted plastic, and painted metal).

Also shown below are some photos of the inside of the F104, indicating the locations of the P clip and where a solenoid driver and solenoid would go. The LEDs are designed to have the flexibility for movement so you can move them closer or further away from the LED overlay based on your LED lock lighting preferences.

2023-04-26_19-36-34.jpg (1.69 MiB) Viewed 6682 times

2023-04-26_19-36-43.jpg (1.6 MiB) Viewed 6682 times

2023-04-26_19-37-15.jpg (3.82 MiB) Viewed 6682 times

2023-04-21_14-25-20.jpg (1.1 MiB) Viewed 6682 times

2023-04-21_14-30-18.jpg (761.9 KiB) Viewed 6682 times

Someone asked about getting some touch up paint for their powdercoated cases.

I would just order a sample size of touch up paint for ~$8 from Home Depot.  Some locations have a paint matching machine where you could bring in your keyboard case to get the most accurate color, since colors vary slightly, but if you are okay with a close but not exact match you can refer the Home Depot staff member to the photo below.

2023-04-27_22-53-03.jpg (2.06 MiB) Viewed 6682 times

[depletedvespene]

Here are the first photos of the completed and tested M Style FSSK sample; it is configured with the HHKB Style Split Right Shift and Split Backspace. Since the ~ ` key is left of the 1 ! key, I was not sure which key to place in the rightmost key of the block, so I selected a blank key. Any ideas on how folks plan on configuring their FSSK layout with such a configuration?

Weeeeeeell... the larger amount of available keys on a TKL, as opposed to an F77 form factor, diminishes the needs for extra keys, but I'd still want my FSSK with an ISO Enter key (+ split left Shift) and unsplit Backspace, BUT... with the HHKB Style Split Right Shift and the split space bar as well, for consistency with my extant F77.

I took a quick look at the store page, but it's not clear to me how to add these customizations to the layout, beyond the "Standard (ANSI or ISO)", "HHKB-Style Split Right Shift and Split Backspace" and "HHKB-Style Split Right Shift and regular 2-unit wide backspace" options.


While we're at this... perhaps I missed this, but AFAIR, the specialty extra keys include keycaps with the "F62" and "F77" legends. Perhaps options for "F104", "FSSK", "F122" and "F50" keys should be considered?

[Muirium]

Surely you can’t be serious. If there’s F122, then you need 2 of them, promptly! What beats more keys? More more keys! Moar more!

[depletedvespene]

Surely you can’t be serious. If there’s F122, then you need 2 of them, promptly! What beats more keys? More more keys! Moar more!

Yes, but the addition of keys needs to make sense — it ain't more keys only for the sake of more keys. Otherwise, you end up with useless overgrown junk like the "Hyper 7".

If I could have my way, I'd convince Ellipse to do the "F30 powerpad" to complement the FSSK, and that'd be great.

[Ordinary Witch]

I personally plan to put as many keys on my F122 as possible when it ships. I'm always out of keys for modded minecraft modpacks... I am sort of relying on the leyden jar controller being used and modified to add two more matrix columns though.

[Ellipse]

depletedvespene there is a layout customization service I offer for folks who would prefer not to add/remove flippers themselves. The important thing is to decide on HHKB style or regular style right shift since the barrel holes are in physically different positions.

For now I would recommend the F50 as it is the closest option to an F30 that is available. I am open to additional models but low volume models would require a minimum order quantity (TBD) and would cost at least $500; also everyone getting one would need to split $1,000+ (TBD) in tooling costs in addition to the keyboard cost. One top request is an F20 dedicated 4x5 number pad for those who intend on putting the mouse between a smaller keyboard and the F20.

Ordinary Witch yes the Leyden Jar controller is expected to be the default on the F122s. The test version I received works well in my typing tests. The F122 can still be used without it, but some keys like backspace, num pad +/Enter cannot be split without the additional columns as detailed in a prior post.

[stangbat]

I received my Ultra Compact F104 Model F yesterday. After setup and testing, I noticed that the right Menu key (between Alt and Crtl) did not work. I've done all the troubleshooting, and I've set up and configured an F77 before this. So I hopefully know what I'm doing. On looking at F104 keyboard's page, I see Item 6: "All keyboards were factory set to the ANSI layout." Some research leads me to believe that a right menu key is not part of this ANSI layout, hence it not working?

This leads to another question. In trying to figure out why the key wasn't working, I downloaded the QMK Model F Utility from the manual page. The version for download doesn't recognize the F104. Then I found this post from Ellipse:

As an update pandrew has updated the QMK beta site to add the F104 and FSSK.

However, I can't find the updated QMK utility anywhere.

TLDR: Is the right menu key supposed to work on the configuration that was shipped? And if so, is an updated QMK utility available to help me troubleshoot?

Honestly, it isn't a big deal if this menu key doesn't work. I never use it. I'm just curious at this point.

[Ellipse]

That key is not recognized by default in any operating system since it is programmed to be a function key by default per my 4/19 post (I need to update the project site). You would need to reprogram it before it is recognized, but I recommend having one key as a Fn key. The latest signal level monitor utility will recognize it however.

This is a change from right Ctrl being the default Fn key for non-HHKB F62/F77 keyboards.

Regarding the additional key made available through the smaller spacebar, that key is set to menu or right click by default, though I don't know if all OSes recognize the key. I just tested it now in Windows 10 and it worked to send a right click signal in Chrome.

The latest pandrew utility needs to be used to recognize things as the program is updated with each new supported keyboard. I have just updated it on the project web site, along with updating the zip file of hex files. pandrew has also updated the beta QMK site if you wanted to customize the layout. In addition to the below link, you can also download this file directly from pandrew by following the link in pandrew's QMK documentation viewtopic.php?p=480196#p480196

https://www.modelfkeyboards.com/wp-cont ... loader.zip

https://www.modelfkeyboards.com/wp-cont ... -files.zip

[stangbat]

Great, thanks. I just used the updated utility and I see that the key is registering. Below is it showing up. If I correctly understand what you've described, what is currently mapped as the menu key is actually hidden under my space bar.

menu key.jpg (142.83 KiB) Viewed 6193 times

[shine]

this is awesome, Ellipse. I take my hat off to you

[Ellipse]

Leyden Jar controller update:
Good news from Rico. Rico just announced the completion and open sourcing of the Leyden Jar controller project with the 18x8 extended matrix! And wcass has designed the F122 PCB to accommodate these extra pads, so now we have everything that is needed for the full F122 usage.

The next step is to optimize the components / design for mass production at the lowest feasible cost that does not result in noticeable performance loss. At first glance the RP2040 based board seems to cost more than the ATMEGA-based controller, even with bulk quantity. It looks like the board uses several components that cost $0.50 or more each, compared to the ATMEGA-based design only using one such component, not counting the ATMEGA chip itself. Which components can be changed out for lower cost components, even if the design has to change a bit? The push buttons probably need to be eliminated for example, since they cost nearly $1 each.

https://www.keebtalk.com/t/the-leyden-j ... s/17489/23

https://github.com/mymakercorner/Leyden_Jar

PS/2 support - A welcome surprise is that Rico has designed an add-on board that adds PS/2 support, however it has not yet been programmed on the firmware side so some development will be needed. It plugs into the solenoid header, which would not be needed for LEDs nor the solenoid, given the separate LED through holes and the PS/2 connection not permitting the required power for a solenoid.

f122 pcb 18x8.png (45.46 KiB) Viewed 6127 times

Leyden_Jar_Rev2_Front.png (332.08 KiB) Viewed 6127 times

Some of the higher cost components:
Comment Footprint Matched Part Detail
MCP4716A0T-E/CH SOT-23-6 MCP4716A0T-E/CHC110057
SOT-23-6 Digital To Analog Converters (DACs) ROHS
PCA9557PW,118 TSSOP-16_4.4x5mm_P0.65mm PCA9557PW,118C141380
8 400kHz I²C, SMBus TSSOP-16 I/O Expanders ROHS
ADP122AUJZ-3.3-R7 TSOT-23-5 ADP122AUJZ-3.3-R7C143910
300mA 60dB@(10kHz,100kHz) Fixed 3.3V~3.3V Positive 5.5V TSOT-5L Linear Voltage Regulators (LDO) ROHS
LMV339 TSSOP-14_4.4x5mm_P0.65mm LMV339IPWRC273648
TSSOP-14 Comparators ROHS
W25Q16JVUXIQ USON8_3x2mm W25Q16JVUXIQC2843335
USON-8-EP(2x3) NOR FLASH ROHS
mSMD110-16V Fuse_1210_3225Metric SMD1210-150C70102
6V 1.5A 100A 3A 1210 Resettable Fuses ROHS
USBLC6-2SC6 SOT-23-6 USBLC6-2SC6C7519
6V 17V Unidirectional 5.25V SOT-23-6 ESD Protection Devices ROHS

[wolfman]

Update for getting getting the newest version of QMK working. There are changes to the API. Once I updated for the new API. The F77 the left side keypad does not register. There is also an issue with the right most key on the bottom row. I will provide an update once I get all the key working.

I have now have all the keys working correctly. I have started testing with the new API with the F77. The newest version of VIA does not like the USB VID of 0xfeed. I have changed the F77 firmware to use the pid.codes USB IDs that Ellipse requested, VID: 0x1209 PID: 0x4707. The source for the updated QMK firmware is in my QMK development branch.

[wolfman]

I forgot to add that I have updated my hack of the pandrew-xwhatsit-utility for the pid.codes USB IDs that Ellipse requested.

The repo for my xwhatsit QMK development branch:
https://github.com/matthew-wolf-n4mtt/q ... w_xwhatsit

The repo for my hack of the pandrew-xwhatsit-utility:
https://github.com/matthew-wolf-n4mtt/p ... it-utility

[cap]

The push buttons probably need to be eliminated for example, since they cost nearly $1 each.

On a $500 keyboard, I'm ready to spend $2, even $20, more for the convenience of push-button switches on the controller. I hope we can keep them.

[TerminalJunkie]

I'm trying to debug a key that stopped working on my F62. It sounds fine and reseating the keycap hasn't helped, so I downloaded pandrew's utility to look at signal levels. But I get the following message when I click "Signal Level Monitor": "Unknown keyboard (you may need to update the util version)!"

The utility is from qmk-pandrew-macos-util-20210321.dmg. Is there a newer version? Any other suggestions? Is arkku's firmware likely to work better?

As an aside, my XT keyboard started having the same issue last week. (Seasonal change? It has been rock stable for years.) It has a stock IBM controller and a Soarer's converter.

[Ellipse]

TerminalJunkie I think the mac version of the compiled utility needs to be updated, or you need to run the Windows or Linux version for diagnostics. It's odd since the F62 PCB is unchanged since at least 2021 when the utility was compiled, so it should not require an updated version. Maybe reflashing a hex file from the manual will help.

cap the pricing difference is $10,000 or more in upfront costs when you are looking at making 5000+ controllers with 2 buttons each. Rico noted to me that pushbuttons can be soldered onto the pads if anyone prefers using them instead of using a key combination to enter the bootloader.

That is good news wolfman. Hopefully the updates will be accepted by the QMK team.

[Miky2147]

Hi, I'm very interested in the new F104 Model F which will be released later this year. Will the keyboard arrive ready for use (except of course the keys that need to be entered)? Or will I have to assemble the LED Overlays as well, even though I don't need to choose custom LED Overlay? Will the keyboard already have its LED Overlay already assembled if I don't choose a different LED Overlay from the standard already included in the $420 price? Sorry but I did not understand this. Thank you

[Ellipse]

Miky2147 everyone will need to install the adhesive overlay sticker, the bumpers, and the keys as those are not installed for any of the models.