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

[NathanA]

I may have missed it but I didn't see anything in the manual that mentioned either way whether haptic was enabled by default on "first boot" (uninitialized EEPROM) or not, but plenty of talk about setting a toggle key to turn it on/off. I assumed it must be off by default, much like NKRO is ("when in doubt..."). But a quick glance at QMK sources I think shows my assumption to be wrong, as the very first line of haptic_reset() is this:

Code: Select all

void haptic_reset(void) {
    set_haptic_config_enable(true);

There's also a FORCE_NKRO constant that one can set to make NKRO the default, whereas there is no equivalent FORCE_HAPTIC.

I guess that the logic here must be that NKRO over standard USB keyboard HID doesn't always work universally in all contexts (BIOS, etc.) so the default should be disabled unless FORCE_NKRO is explicitly set, but if you have a haptic device wired up then why not have it enabled by default? Fair enough, I suppose...

[jandres]

Thanks everyone for the help, in particular NathanA, you have been immensely helpful. After trying just about everything suggested, I suspected that the issue was hardware related, so I opened up the keyboard to do some diagnostics. I noticed that when I toggled on the solenoid, a light would blink with each key stroke, leading me to believe that there may be something wrong with the solenoid itself.

In order to rule out the driver, I first connected a test light to the two pins that go to the solenoid. Sure enough, I confirmed that the driver is in fact sending an electrical current with each key stroke. To be doubly sure, I then connected my Fluke multimeter to verify voltage. To rule out a short in the wires, I checked for continuity to the solenoid. At this point there is no question that the issue is not firmware or driver related, but with the solenoid itself.

I have ordered a new solenoid. Hopefully that will resolve the issue. I will report back as soon as I have received and installed it.

[Ellipse]

jandres I just sent out the tracking number. I think you are correct that if the solenoid driver light blinks when you press a non-modifier key, everything but the solenoid is operating correctly.

When you receive the solenoid, please directly plug it in to the driver and test it before attaching the L bracket or making any other adjustments. I know of one or two folks who damaged the solenoid by screwing in the screw too far such that it broke through the protective blue tape and damaged the coil (they may have forgotten to attach the L bracket - with the L bracket the screw is not long enough to break through the tape) or maybe used another screw that was longer.

[Shihatsu]

Almost 5 years after my Order my Modelf was in today. I must admit I dont like it as much as expected. The hard and brutal truth is THAT I LOVE IT WAAAAAY MORE than expected. Mind the yelling, not only am I excited a. f. but I must also be louder than this MONSTROSITY of a Keyboard.
It is full ISO-DE, thats why I am late to the party. And i have to figure out why the hashkeys is not working, but I will stick to the troubleshooting guide next and come back later if I cannot solve it. A few questions from my side, as I didnt follow the news for the last two years or so:

How do I order "coolermaster masterkeys style numpad/modifier hybrid"?
Are there any keycaps in different colors available?
How do I achieve an angle of like 5degree?

Kudos to Ellipse for this dream becoming a reality, expect a second order soon! YOU are the man!

[DrunkUkrainian]

Hi, I tried to update the firmware to the VIA version. I followed the guide to the best of my ability but it's not outputting anything despite being picked up by VIA. What do?

[NathanA]

Hi, I tried to update the firmware to the VIA version. I followed the guide to the best of my ability but it's not outputting anything despite being picked up by VIA. What do?

Was the keyboard working before you flashed VIA?

...oh, no. Just noticed something. In Ellipse's firmware ZIP archive, only the QMK hex files have dates newer than April 2021. So the VIA hex files are still all the same ones that he originally published a year ago.

The keyboard controller PCB got a slight tweak/revision done to it late last year, which required an update to QMK in order to support the new revision. That's why Ellipse updated his QMK hex files in November. But if the VIA hex files never got updated, and if you just received your keyboard & if your keyboard was built using one of the revised controllers, then VIA might not work.

If the keyboard isn't registering any keypresses, there also might not be a way to get the keyboard to go back to bootloader mode in order to reflash it with working firmware without opening the keyboard up. That's because the pandrew utility is not supported by the VIA firmware.

[DrunkUkrainian]

Hi, I tried to update the firmware to the VIA version. I followed the guide to the best of my ability but it's not outputting anything despite being picked up by VIA. What do?

Was the keyboard working before you flashed VIA?

...oh, no. Just noticed something. In Ellipse's firmware ZIP archive, only the QMK hex files have dates newer than April 2021. So the VIA hex files are still all the same ones that he originally published a year ago.

The keyboard controller PCB got a slight tweak/revision done to it late last year, which required an update to QMK in order to support the new revision. That's why Ellipse updated his QMK hex files in November. But if the VIA hex files never got updated, and if you just received your keyboard & if your keyboard was built using one of the revised controllers, then VIA might not work.

If the keyboard isn't registering any keypresses, there also might not be a way to get the keyboard to go back to bootloader mode in order to reflash it with working firmware without opening the keyboard up. That's because the pandrew utility is not supported by the VIA firmware.

Alright I think the outdated VIA files are the issue. I'll try using the QMK layout generator here http://35.164.28.200:5000/#/xwhatsit/br ... LAYOUT_all to create my desired layout

[NathanA]

Alright I think the outdated VIA files are the issue. I'll try using the QMK layout generator here http://35.164.28.200:5000/#/xwhatsit/br ... LAYOUT_all to create my desired layout

Another option, if you're feeling adventurous, is to try Vial instead, which I've already built a copy of that incorporates the changes needed to support the new controller revision. (If you prefer the VIA utility to Vial, the Vial firmware is also compatible with the VIA host-side utility.)

To kick your keyboard back into bootloader mode without functional key sensing, you will need to open up the keyboard and gain access to the pads marked "PROG" on the underside of the PCB. Those two pads need to be bridged together while you plug the USB connection into the host. I just straighten out a paperclip, bend it in half, and touch the two ends to the two pads.

[DrunkUkrainian]

Alright I think the outdated VIA files are the issue. I'll try using the QMK layout generator here http://35.164.28.200:5000/#/xwhatsit/br ... LAYOUT_all to create my desired layout

Another option, if you're feeling adventurous, is to try Vial instead, which I've already built a copy of that incorporates the changes needed to support the new controller revision. (If you prefer the VIA utility to Vial, the Vial firmware is also compatible with the VIA host-side utility.)

To kick your keyboard back into bootloader mode without functional key sensing, you will need to open up the keyboard and gain access to the pads marked "PROG" on the underside of the PCB. Those two pads need to be bridged together while you plug the USB connection into the host. I just straighten out a paperclip, bend it in half, and touch the two ends to the two pads.

I was able to get it back into bootloader mode, and have it reprogrammed the way I liked it. I'll keep an eye out for an update but I think I'm happy with the way I have it now

[jandres]

jandres I just sent out the tracking number. I think you are correct that if the solenoid driver light blinks when you press a non-modifier key, everything but the solenoid is operating correctly.

When you receive the solenoid, please directly plug it in to the driver and test it before attaching the L bracket or making any other adjustments. I know of one or two folks who damaged the solenoid by screwing in the screw too far such that it broke through the protective blue tape and damaged the coil (they may have forgotten to attach the L bracket - with the L bracket the screw is not long enough to break through the tape) or maybe used another screw that was longer.

Awesome! I will definitely do as you suggest and will be careful with the screws. I may have very well damaged the solenoid when I first installed it. Although the screw didn't perforate the blue kapton tape, it was pretty tight against it.

Thanks again for all the help.

[maarteneh]

Regards. I am trying to USB-ify a 3278 keyboard and need advice. I just received a modelfkeyboards.com controller product item "xwhatsit Model F, Beam Spring USB controllers, solenoid drivers, solenoids - Beamspring USB-C (non-Displaywriter)". It has a white PCB labelled "Beamspring-USB-5251" with an atmega32u2 cpu. After reading the modelfkeyboard.com manual and physically installing the controller on my keyboard (along with the solenoid driver and upgraded solenoid), it did not function as an external keyboard when plugged into my Macbook.

I then went to the qmk configurator site and selected, compiled, and downloaded "xwhatsit/ibm/3101_3727_3278_87key/universal". I chose that one because it indicated "xwhatsit", "3278" and "87key" and "universal" in the name. I successfully flashed the downloaded .hex firmware file using qmk toolbox, which then validated the flash and reset the controller (it subsequently "disconnected").

After doing that, the controller is now not recognized by qmk toolbox (or by the Mac as a keyboard). Powering on with PROG or RESET grounded does not recover the controller into bootloader mode. The controller no longer shows up in qmk toolbox at all.

I suspect I flashed the wrong firmware. Perhaps "xwhatsit/brand_new_model_f/beamspring_full/universal" would have been a better choice given that I have a brand new xwhatsit Model F controller. Argh. Anyway, does anyone know how I can recover from this? Any suggestions appreciated.

[NathanA]

Powering on with PROG or RESET grounded does not recover the controller into bootloader mode. The controller no longer shows up in qmk toolbox at all.

QMK Toolbox can generally only interact with a keyboard in bootloader mode. The only exception I know if is if debug console was enabled in your build of QMK firmware, in which case the controller can output things to be printed on the toolbox console as well as accept some commands. So I'm sure that the reason that it "no longer shows up" in QMK Toolbox is simply because it's not in bootloader mode & you are having trouble getting it to kick over into bootloader mode.

Have you tried using the pandrew utility to kick the controller back into bootloader mode? I can't find a copy of it built for macOS on the official Model F Keyboards site, but it looks like a very helpful forum member made a macOS version, downloadable from here. Try launching this while the keyboard is plugged in and see if the "Enter Bootloader" button works (if you have QMK Toolbox running simultaneously, it should suddenly detect the keyboard in bootloader mode after you do this).

RESET I believe just pulls the reset line on the microcontroller (reboots the device). To kick the controller into bootloader mode physically, you need to be shorting both of the PROG pads together *while applying power to the controller* (that is to say: plugging the USB cable into the host). I have found that the timing can be a bit finicky, so if at first you don't succeed, try try again...but if you can get the pandrew utility method to work for you on your machine, it is by far the simpler and more reliable method.

(If you can't get the pandrew utility to work on your machine, and if it is not an Apple Silicon Mac that we are talking about here, you might try Boot Camping over to Windows and then running the Windows versions of QMK Toolbox and the pandrew utility. Or if you have a Windows machine handy to try that on, even better.)

[maarteneh]

Hi NathanA, thank you for your tips.

Regarding the timing of the PROG shorting work, I just left PROG shorted to ground and then plugged in the USB. From your description it sounds like the PROG shorting needs to be intermittent during power-up. That will be hard to do by myself with, for example, a paperclip and then plugging in the USB. I will get help and report back. Meanwhile, do you happen to know if there is a specific timing window for shorting prog?

BTW, I tried the pandrew utility but it does not detect the USB device either, so the 'enter bootloader' button is greyed out.

I am concerned I have overwritten the atmega32u2 USB bootloader code if indeed I flashed the wrong xwhatsit firmware. Does anyone know if the QMK configurator / QMK Toolbox flashing process exposes that risk (or is the bootloader code left untouched by firmware flashing)?

Any insights appreciated.

[sedevidi]

I am concerned I have overwritten the atmega32u2 USB bootloader code if indeed I flashed the wrong xwhatsit firmware. Does anyone know if the QMK configurator / QMK Toolbox flashing process exposes that risk (or is the bootloader code left untouched by firmware flashing)?

QMK uses dfu-programmer behind the scenes. NathanA dug into this a few posts ago. Reading the dfu-programmer man page, I'd say that overwriting the bootloader should be difficult, potentially impossible, or at least impossible to not notice. I have no experience at all with this; I just report what I read not that long ago.

[Shihatsu]

Heya, just a short question regarding programing the ModelF:
Is there a comprehensive Guide for people who just use Linux and ISO-DE? The documentation is still cluttered all over the place, different documents and such. Maybe there is something better hidden somewhere?

[alireza]

Hey all,

I received my device yesterday and as a software engineer knowing near nothing about hardware, I've spent hours so far trying to get this keyboard working on my Mac M1 and any help here is much appreciated.

Objective: I want to remap a few keys through VIA. Keys to be remapped include Caps Lock to be Left Command (= Left Win).

My steps are:

Downloaded pandrew. Clicked on erase EEPROM and then Clicked Enter Bootloader.

Downloaded 'dfu-programmer' from homebrew (QMK toolbox didn't work) and then:

Code: Select all

> sudo dfu-programmer atmega32u2 get manufacturer
Manufacturer Code: 0x58 (88)

Code: Select all

> sudo dfu-programmer atmega32u2 erase --force
Erasing flash...  Success
Checking memory from 0x0 to 0x6FFF...  Empty.

Code: Select all

> sudo dfu-programmer atmega32u2 flash Downloads/QMK-layout-files/eeprom_eraser.hex
Checking memory from 0x0 to 0x27F...  Empty.
0%                            100%  Programming 0x280 bytes...
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>]  Success
0%                            100%  Reading 0x7000 bytes...
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>]  Success
Validating...  Success
0x280 bytes written into 0x7000 bytes memory (2.23%).

But then:

Code: Select all

> sudo dfu-programmer atmega32u2 flash Downloads/QMK-layout-files/Via\ F77\ -\ ANSI-ISO\ -\ PrintScr.hex
Checking memory from 0x0 to 0x4E7F...  Not blank at 0x1.
The target memory for the program is not blank.
Use --force flag to override this error check.

So:

Code: Select all

> sudo dfu-programmer atmega32u2 erase --force
Erasing flash...  Success
Checking memory from 0x0 to 0x6FFF...  Empty.

Code: Select all

> sudo dfu-programmer atmega32u2 flash Downloads/QMK-layout-files/Via\ F77\ -\ ANSI-ISO\ -\ PrintScr.hex
Checking memory from 0x0 to 0x4E7F...  Empty.
0%                            100%  Programming 0x4E80 bytes...
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>]  Success
0%                            100%  Reading 0x7000 bytes...
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>]  Success
Validating...  Success
0x4E80 bytes written into 0x7000 bytes memory (70.09%).

Then downloaded VIA, loaded "Via program - load this if you have the F77" and made the changes.

Unplugged and plugged back the keyboard. It's not recognized. Now I see "Brand New Model F Keyboards" in " ioreg -p IOUSB" but no key presses. pandrew and dfu-programmer also no longer recognize the keyboard.

What should I do?


SOLVED: Put the keyboard into the bookloader by connecting the prog pins at the back.

[alireza]

I managed to flash my changes onto the board. Now the current question is that the wires for the solenoid look nothing like the instruction video.

The wires look like this:
https://www.modelfkeyboards.com/wp-cont ... -00-15.jpg

If I connect these wires such that the red wire is connected to the square pins on each board, then when toggle on the solenoid via fn+space+t, the solenoid board's light blinks but the solenoid doesn't move.

If I connect these wires any other way, a whistling sound is produced which makes me immediately unplug the keyboard.

Any guidance here is appreciated.


SOLVED: Red wires should be connected to the square pins. The power was set to the minimum though - saw slight movement if I pressed 4 keys at the same time. Increased it and now it's visible/audible.

[Ellipse]

Here is a photo showing the new tougher powdercoating option on the compact cases that was mentioned several months ago. Dark Blue and Carolina (Light) Blue are shown below.

For compact keyboards shipping this year, everyone gets the new improved option instead of the anodized option which was more easy to scratch. The new cases are also from the brand new die cast aluminum molds as noted before (makes the case two piece instead of three piece) The new cases are fully compatible with the older batch of internals in case anyone wanted to upgrade.

20220423_195914.jpg (252.97 KiB) Viewed 14937 times

[Irrationalforest]

Hi all!
New Model F Labs F77 owner here.
The keyboard feels great in most respects, and I really am enjoying typing on it.

One question: The space bar seems slightly unreliable. One in maybe twenty presses seems not to register when typing fast. I can more reliably (70%) reproduce this issue by pushing Backspace and then the space bar quickly afterwards.
I have reseated the spacebar numerous times and tried with NKRO one and off.

I don't think it's a hardware issue, since its very repeatable with specific key orders (see backspace+space above).

Currently running the Ellipse official profile with QMK from the Update page with working solenoid patch.
Any tips would be greatly appreciated!

Thanks

[wobbled]

Hi all!
New Model F Labs F77 owner here.
The keyboard feels great in most respects, and I really am enjoying typing on it.

One question: The space bar seems slightly unreliable. One in maybe twenty presses seems not to register when typing fast. I can more reliably (70%) reproduce this issue by pushing Backspace and then the space bar quickly afterwards.
I have reseated the spacebar numerous times and tried with NKRO one and off.

I don't think it's a hardware issue, since its very repeatable with specific key orders (see backspace+space above).

Currently running the Ellipse official profile with QMK from the Update page with working solenoid patch.
Any tips would be greatly appreciated!

Thanks

Any time I've had issues with unreliable key strokes on the New Model Fs or converted original Model Fs it has been the voltage threshold that needs adjusting. I don't know if this is possible with QMK.

[kekstee]

Guess it's my favourite F board, given that it's a usable layout and all. Very nice to finally use it after all those years.

The fact that it's flashed with a totally usable hhkb layout which I ordered is a really nice touch, didn't do anything but slap the caps on and type away. Love the keyfeel but it's going to stay a bit of a curiosity given the sound of it :D

[Ellipse]

Irrationalforest check out the manual on the project web site - initial setup and troubleshooting sections which include a section on space bar modding. It could help to adjust the metal space bar tabs, the spring, the wire stabilizer for the space bar, or the space bar itself ("wiggle method" on the key stem part).

If backspace has something to do with it maybe it could be the backspace spring or the seating of that key.

[ictoys]

Alright I think the outdated VIA files are the issue. I'll try using the QMK layout generator here http://35.164.28.200:5000/#/xwhatsit/br ... LAYOUT_all to create my desired layout

Hi, new owner of the F62 kishsaver. I tried using the VIA hex as well but no key input detected, so the file is outdated and there's no way to have it work currently? I decided to use the QMK layout generator as well for the time being.

[Ellipse]

Solenoid driver update:

Given the chip shortage the original xwhatsit solenoid driver cannot be produced at the moment, so I am looking to switch to pandrew's updated xwhatsit-based solenoid driver. It is great that pandrew updated the driver back in 2020 to use more widely available parts.

pandrew has noted that any solenoid that works on the original xwhatsit driver should theoretically work on the new driver but to be safe has anyone done any testing with the pandrew driver and my solenoids?

[alireza]

The "q" button on my new F77 is flaky, even with the default firmware: sometimes it works and sometimes it doesn't.
In pandrew. the key is being detected but the highest signal that it gets is 136 which is lower than all other keys.

How can I fix this? Is this a voltage adjustment problem? If so, any pointers to how to actually do this on Mac would be much appreciated.


P.S. And now the solenoid clicks if I press "tab" (didn't used to) but not with "q". What's going on?!

[Zcool31]

I have another silly question. How did the original vintage capacitive keyboards work reliably without the need to manually tweak voltages or thresholds?

[Muirium]

I have another silly question. How did the original vintage capacitive keyboards work reliably without the need to manually tweak voltages or thresholds?

IBM's controllers were better engineered than ours. I'm not technical enough to know the details, but I can vouch for their flawless reliability even after 30+ years with degrading foam and all the rest. My XTs, my AT and my Kishsaver's IBM original controllers all worked perfectly, even with significant differences between rusty old springs. IBM's engineers really excelled with these things!

But yes, I pulled out all those controllers (which sit in storage) because I need USB, and programmability is nice, too. When it comes to capsense, though, they are in a league of their own. No thresholds, no tweaks, it just always works.

(If you read Xwhatsit's manual, he mentions the matrix positions IBM put into all Model Fs and beamsprings which allow the controller to self calibrate. Some spots in the matrix are always high and some always low. Maybe my 3278 beamspring's stock controller still worked too, but I've no converter to try it on. I did try Hasu's 4704 converter with my Kishsaver back in the day and it did!)

[alireza]

The "q" button on my new F77 is flaky, even with the default firmware: sometimes it works and sometimes it doesn't.
In pandrew. the key is being detected but the highest signal that it gets is 136 which is lower than all other keys.

How can I fix this? Is this a voltage adjustment problem? If so, any pointers to how to actually do this on Mac would be much appreciated.


P.S. And now the solenoid clicks if I press "tab" (didn't used to) but not with "q". What's going on?!

If I remove the "w" keycap, the "q" seems to be working fine. Any ideas?

[Shihatsu]

Ellipse, I got a question - better: suggestion - for the F77 and F62 Ultra Compact Case: The cases a great, bot the bottom plate is absolutely lackluster. It would be kinda "meh" on an ordinary board, but on the ModelF Keyboards it is a bit ridiculous. It is like two levels below the rest of what you deliver. Feels like a bummer. Please don't missunderstand me - ALL of your work is highly appreciated, eternal love, man! But this specific part of the board just does not deliver the same feeling of luxuary and perfection.
The bottom plate should be improved via two things, one is imho mandatory:
Countersunk holes for the screws. Not sunken screws are just bad design, overall: They deliver no advantage, but look fugly and make moving the board without feet just strange, especially when you compare it directly to the OG zink case.
Second improvement would be a bigger one, but I'd love to see this: A second plate ground plate/addon with integrated lifting to give the board a nice 5% lifting - this would mimik the feet (albeit in the lower setting) of old XT/AT Model Fs AND would add even more heft to the board.
Would one or even both of these upgrades be possible?

[Ellipse]

How are folks programming the right side blocks 3, 4 and 5? If anyone has made JSON files that are the same as the ones I provided, but with adjustments for the right side blocks 3/4/5 please do share! I guess the secondary legends are set up as a function layer, or maybe a tap layer to mimic the functionality of the Num Pad key?

Shihatsu regarding your suggestions, as noted a while back there was an issue with the countersunk holes breaking through the original prototype. Overall they weakened the board as they could not be screwed in tight enough to avoid a gap between parts of the compact case. The new design with exposed screws fixed that issue. However I will note that with the more recent beam spring project those screws are countersunk and I have not seen the same issues appear. However the compact case design has been finalized with the current design, which uses custom-made torx screws from brand new molds that have already been produced. Regarding the tilt there are bumpers of varying heights that can adjust the angle of the board. I have no plans to recreate the molds for additional IBM style supports, though maybe some folks could design some 3d printed options? I recall one or two projects with 3d printed feet for the new Model F keyboards.

alireza as I noted over an email, if keys are not working it is a matter of following the manual on the project web site and especially the QC secrets video to adjust or possibly, as a last resort, replace the springs themselves if needed, not just reseating the keys.