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

[NathanA]

I will report that one person recently noted an issue with transposed keys and the cause of the issue ended up being their use of Combos in the Vial firmware. Removing the combos fixed the transposed issue.

QMK Combos is honestly a bit of a weird feature. It can cause perceived delay/latency in key output, and it would not surprise me to learn that this in turn could cause a sense of "transposed keys" for some people, depending on what keys they are using as part of a combo and how fast they type.

Run this experiment:

Configure a combo in Vial consisting of (say) keys Q, W, and E. Have them output R.

Now if you press (and hold) Q + W + E really fast together in any order, R will be output. As expected.

Now try typing Q, W, or E just by themselves. You will notice that it feels "slower" to type these characters. Try typing characters other than those, and those will all feel normal and "fast" to type.

This is actually because if you pay close attention, the keys that are part of a combo chord will only get typed after you release the key, instead of as you press it.

The reason for this is obvious, if you think about it for half a second: if you have keys that are part of a combo which normally output something (they aren't pure modifier keys), you don't want their normal function to trigger if you are trying to use them in a combo. So if (in this example) you press Q + W + E, you want to see R and only R. You don't want to see QWER. So when you press such a key that is part of a combo, QMK has to wait and see if you are in the middle of typing the combo or not before it decides to output the keystroke for that key, or not. So when the 'Q' key is part of a combo, it only knows for sure that you intended to type a Q if 1) you press and then release the Q key, or 2) you press and hold the key and the "time out period" configured in Vial under "QMK Settings" > "Combo" has passed.

This timeout period is set to 200ms (1/5th of a second) by default, which for some people may be too long. The trade-off is that if you set it to a shorter period of time, then you may have a harder time successfully executing a combo.

But if you have alpha keys configured as part of a combo, and you are a fast typer, the alpha keys that are part of a combo only get outputted after you release the key, but OTHER alpha keys get outputted the second you press them down. This, I suspect, is what leads to the feeling of "transposed" key input when using the feature.

(If you press a key down that's part of a combo, and then continue to hold it down and press another key that is not part of any combo, QMK will immediately output both the key that's part of the combo along with the next key you type, and do so in that order, which in theory should help with this. But I suspect this doesn't completely eliminate that sense of input latency in a lot of situations.)

[pilcher]

Please do share: how do you reinstall and reseat Model F springs when the flipper is installed in the keyboard?

I used tweezers to get my F104 working. I was ultimately successful, but I will say that I didn't find it to be an easy process. I suspect that there are a few reasons for this, possibly including lack of experience, the tweezers that I was using, and damage to the springs during shipping. (In your shoes, I'd probably consider including a couple of extra springs with the classic F104s, as it seems like they may be particularly susceptible to damage during shipping.)

Since then, I discovered the my wife has been saving the pump mechanisms from various hair care and cosmetic products, and she was willing to sacrifice one that had a tube of the perfect size for the tube + skewer method*. After testing this method with a flipper and spring from my first aid kit, I'm pretty sure that this is going to be by far the easiest method

* https://www.youtube.com/watch?v=SnBsDeKs54w

[idollar]

As a reply to some recent postings here, it is completely wrong to assume that double key presses and transposed keys are due to xwhatsit's controller and that another controller would fix the issue based on one person's experience. While nearly all these issues are fixed by the users themselves during setup, I have worked on several of these boards myself over the years and it has always been done with what is noted in the manual (working on the springs and ensuring there is no dust or debris inside the keyboard that could affect capacitive sensing).

I will report that one person recently noted an issue with transposed keys and the cause of the issue ended up being their use of Combos in the Vial firmware. Removing the combos fixed the transposed issue.

Please, do not put words in my mouth. I did not say that "double key presses and transposed keys are due to xwhatsit's controller and that another controller would fix the issue based on one person's experience."

What I wrote is the following:

"I used to have similar issues with the xwhatsit controller in my FSSK hand-made keyboards. I had to play with the flippers and the springs. Without opening the keyboard, it is not an easy task."

The sensing in capacitive keyboards is based on measuring the temporary change in voltage when a key is pressed or released. Two aspects affect this sensing:

• The reference voltage: To measure the voltage, a reference voltage is generated using a DAC. After that, the changes are measured using a comparator. Due to the fact that the voltage change will not be the same across all keys, the keyboard is divided into sets that are measured against the same voltage.
• The measuring speed: Since the voltage change does not last long, the sampling rate is crucial.

I would not argue that the xwhatsit controller cannot do the job, but sometimes the older technology it is based on does not perform accurate measurements. This is why one has to "play" with the springs and keys to get it functioning properly. It will work in the end.

The RP2040 is faster, and the use of PIO and different DACs reduces the impact of varying capacitance changes.

So, it is not "one person's experience," it is simply technology. There is no magic behind it.

xwhatsit's controllers are extremely solid and reliable and have been in use since 2014. I have mailed out over 5,000 keyboards with xwhatsit's controller since 2019, including hundreds of the new models in the past few months, and only a handful of folks have reported issues like double presses and transposed keys out of hundreds of keyboards.

The xwhatsit controller is great and has paved the way for all the work done in this forum that you have used in your keyboards. This does not contradict the fact that the Leyden Jar is simply better.

This is why I stated that "My recommended solution = Leyden Jar." Please focus on "my" and "recommended." I am not saying that this is the only solution. I am stating that this solution is simpler and more reliable than changing the springs.

Others likely had to adjust springs to correct various issues like keys not clicking but followed the manual to fix it, so they did not need to report the issue. As always, when I mail out hundreds of boards quickly after a container shipment arrives, everyone is more likely to see reports of issues posted on the forums just because the shipment volume is far higher than normal.

This could very likely be avoided by using a faster and better controller.

While I saw no issues with Rico's excellent Leyden Jar controller in my testing and hope to use it exclusively once xwhatsit stock is depleted, currently no one is using this in a New Model F since they are still being assembled and have not yet shipped, so no one can say that they are less likely to transpose keys in a statistically significant sample size.

I am using the Leyden Jar, and it is as good as it could be. I have built four keyboards with it. While this may not be many keyboards, I am convinced, based on my R&D, that many of these issues, and others, will simply disappear.


Note:

I would recommend using the QMK console: https://docs.qmk.fm/faq_debug.

Unfortunately, the xwhatsit is not large enough to host firmware with this option enabled.

Another reason to upgrade to be added to the list

[NathanA]

I am stating that this solution is simpler and more reliable than changing the springs.

First, I do agree that I think it would be a fascinating and informative test if somebody could take a keyboard that seems to have some pad sensing problems of some kind, and change ONLY the controller, from the xwhatsit to the Leyden Jar. Don't touch any springs, don't tweak or adjust anything about the keyboard physically. Simply replace one controller with another.

It sounds like you have run this test on your own keyboards that you have built. It is an interesting data point that you have gotten this the positive result that you have only by changing the controller and nothing else. I think it would now be of further interest to perhaps repeat the same experiment with an Ellipse keyboard that is having issues, and seeing if the same thing holds true. Increase the sample size of the experiment, and run similar tests beyond your own keyboards.

All of that said, "simpler" is definitely a matter of opinion here. I say that because not everybody has the talent / capability / practice to desolder and re-solder 30+ pins. I know I don't. I would need a good deal of practice with soldering under my belt first before I would feel comfortable attempting to replace the controller on my keyboard with an xwhatsit controller (a keyboard that works perfectly fine with the xwhatsit controller and has ever since I received it, for over 3 years!) If I were to just do it right now, with no practice or training, I will ruin something in the process, guaranteed.

So replacing the controller is not "simple" for me. This is why I am a software hobbyist and not a hardware one. If there is something keeping the xwhatsit controller or firmware from performing as well as it could and being as "forgiving" as possible of hardware tolerances, I would certainly like to see such things pursued, rather than simply shrug shoulders and give up. I'm not sure that I am the man for that job (pandrew's capacitive sense code is complex...at least to me...), but IF it is doable, then SOMEBODY should be able to step up to the plate.

One thing that I keep thinking about is how, okay, the ATmega has a fraction of the power of the RP2040. But the original processor that IBM used in the late 70s/early 80s for cap-sense keyboards had a fraction of the power of the ATmega! And the firmware IBM wrote for it seemed to be VERY forgiving! So it's not just a matter of, "better technology" or "newer technology". What you actually do with the technology matters. Clearly IBM could squeeze a lot out of very little. If IBM had the ATmega to work with back in the early 80s, they would have been amazed at how much power and memory it gave them!

[Rico]

Hello everyone, this is Rico!

With Ellipse acknowledgment please let me share the current status and future work of the Leyden Jar controller here.

First I'd like to apologize for my lack of presence in this forum; internet is a time killer and I try as much as possible to optimize my free time with my family and hobbies (like working on the Leyden Jar).
With no hard commitments I'll try to have a bigger presence on this forum in the future.

As you may know, future Ellipse's F122, B122 and B104 keyboards will use the Leyden Jar controller in place of our good old XWhatsIt controller.
I am thrilled to see my little project to be used in a large scale project like this (and a bit intimidated) and hope that you, the future users of those keyboards, will be happy of the result.
Most of my time is now focused of ironing out a few things here and there in the firmware code side, once this task will be done I may be able to work on new features (discussed below).

I'd like to mention user idollar that has been the third person after Ellipse and me to manufacture and use the Leyden Jar controller for it's own projects.
This means a lot to me as it shows that the project is mature and easy enough for other people to use.
He was also able to make the Leyden Jar Diagnostic Tool work on Linux with a bit of dirty hacking in the build scripts. This demonstrated that the tool code is cross-platform as I hope it would be, and that the only thing missing is to work on the build scripts to have a clean Linux support (more on that below).

Work done

Github projects architecture

The Leyden Jar project is distributed into three different Github repositories:

• The Leyden Jar controller here.
• The Leyden Jar Diagnostic Tool here.
• My fork and dedicated branch of Vial QMK here.

I recently reworked the Leyden Jar controller Github repository so that it is now the central hub to all other Github repositories used in the project.
That means you now have to only know the link to the Leyden Jar controller and have access to all other related projects and links from it's readme.
I also added a Ko-Fi button, in case you want to fund past and future developments of this project.

Keyboard firmwares

Currently supported boards are:

• F122 (tested by Ellipse)
• B122 (tested by Ellipse)
• B104 (tested by Ellipse)
• F104 (tested by Ellipse)
• F77 (tested by me on my own board).

The Leyden Jar controller Github project no longer contain the firmware source code, everything is now available in the dedicated branch of my Vial QMK Fork.
My Vial QMK fork also contains releases of firmwares binaries for supported keyboards.

The future

We will about to end this year 2024 so this is a list of what is planned for year 2025.

Port other keyboards to be used by the Leyden Jar controller

The Leyden Jar controller share the exact same principles as the XWhatsIt controller. Sure it is an improvement on several points but nothing that a keyboard user should notice by typing.
I am very often calling the Leyden Jar 'an XWhatsIt on steroids'

The XWhatsIt controller has been in use for more than 10 years now, it has proven to be very reliable and many people have a board with this controller and experience no issues at all.
With the fantastic work that NathanA did to use Vial it is now also very user friendly.

But at some more or less distant future there will be no way to produce XWhatsIt controllers because the ATMEGA controllers will be unavailable.

This is a long term project and there is no big urgency.

Split keyboard management

The connector used for driving the solenoid could also be used to implement the UART serial protocol commonly used on QMK for split keyboards.
There is a plan to validate the approach using some of my Leyden Jar controllers and by designing a PCB testing jig.
The goal of the testing jig would be:

• Mimick key presses that the controller would recognize.
• Have the controller slotted in it and not soldered.
• Not having to desolder my F77 again, the ribbon cable has seen enough abuse from having to change two different revisions of the Leyden Jar controller already.

Maybe in the future design a new controller resivion that would allow both UART and solenoid to be used at the same time?

PS2 protocol management

This one will be tough and I do not know if it will work in the end.
People requested this for quite a long time, be able to use Model F or beamspring boards with an old computer that only has legacy PS2 connectivity.
A daugherboard PCB has been designed and even manufactured by me but still not tested yet.
The testing jig described before will also come handy for this project.
The holy grail goals of this feature:

• PS2 connectivity while keeping QMK as core firmware.
• When connected with USB the keyboard works just like any VIA/Vial keyboard and you would be able to tweak your layers/macros/keys.
• When connected with PS2 the keyboard uses all VIA/Vial configuration (layers/macros/keys) but sends PS2 compatible key data.

This will request some deep dive in QMK source code to see if it can be done, and some subtle hacks if proven to be possible.

Leyden Jar Diagnostic tool documentation

The first time I provided the tool to Ellipse I did an experiment: I gave him no guidance on how to use it and asked if things were not clear to him.
He has been able to use the tool without the need of documentation and without asking me questions so it's use looks to be trivial enough.
Anyway I still think that some explanation should be nice, at least in the readme of the Github repository.

Leyden Jar Diagnostic tool Linux build

For the moment builds of this tool are for Windows systems only but idollar showed that having the application on Linux was possible.
This is where I need help as I am a Windows only developper.
That would be great if some of you that have knowledge of CMake build scripts (or are willing to learn) and is a Linux user could do some clean work on the CMake file for everyone to easily use the tool on Linux platform.
It may not be possible to provide one day binaries for Linux (it is such a fragmented platform with all it's distributions), but having an easy way to call CMake then build and having the tool binary would already be great.
If someone is interested to tackle this task please tell me, either here or DM me !

And why not also MAC if someone wants to address this?

That's all guys!
It was a loooooong post, sorry about that.

[Ellipse]

A big congratulations are in order for Rico's great success with the Leyden Jar controller!

Everyone please do let Rico and me know if you are interested in helping contribute to any of Rico's 2025 development areas for the open source Leyden Jar controller.

The highest priority is getting one controller to power and communicate with the other controller so that only one USB cable is needed as I would like to start the next round of the Split Ortho F keyboards which were all spoken for in the first run of those boards with two USB cables (as a note there was no issue sharing function layers and shift states between boards on Mac/OS/Unix thanks to NathanA's firmware).

Regarding the solenoid and UART functioning at the same time, here was a proposal sent to me:

Move SOLENOID_ENABLE to GPIO26, SOLENOID_PULSE to GPIO27, COL16 moves to GPIO28, and COL17 moves to GPIO29, and then have a firmware version that allows for the UART and Solenoid at the same time without sacrificing the solenoid.

The UART would run on GPIO28 and GPIO29 and it would take away columns 16 and 17 for firmware variations that require UART, while these columns would be normal keyboard matrix columns in all other firmware versions not requiring UART.

These columns 16 and 17 could both be connected to the through holes at the bottom and also to the expansion header (maybe the expansion header would need to be 4x2 or 5x2 instead of 3x2 or larger to accommodate the 4 wires needed for UART and power, or there would be a separate 2x2 header for UART).

This way you can sacrifice 2 columns to have both the solenoid and UART with split keyboards, and with the same controller you can also choose to have those 2 extra columns and the solenoid if you do not require the UART."

[Rico]

Thank you Ellipse!

To be honest the biggest help for a Linux noob like me would be someone to look at the Linux build of the Leyden Jar diagnostic Tool.

This proposal for solenoid and UART at the same time is perfectly legit.
I see that someone carefully studied the Leyden Jar schematics and read the RP2040 datasheet to do this proposal.
Now there is the problem of firmware fragmentation with two different PCB revisions to support, but I may have a solution for that :p

[idollar]

Hi Rico,

For the moment builds of this tool are for Windows systems only but idollar showed that having the application on Linux was possible.

As you know, I have compiled the diagnostic tool on Linux (Debian), and it works perfectly.
I had no problems with it.

For the moment builds of this tool are for Windows systems only but idollar showed that having the application on Linux was possible.
This is where I need help as I am a Windows only developper.
That would be great if some of you that have knowledge of CMake build scripts (or are willing to learn) and is a Linux user could do some clean work on the CMake file for everyone to easily use the tool on Linux platform.
It may not be possible to provide one day binaries for Linux (it is such a fragmented platform with all it's distributions), but having an easy way to call CMake then build and having the tool binary would already be great.
If someone is interested to tackle this task please tell me, either here or DM me !

As you mentioned, there is some work to be done to change a few CMake files.
In order to get the binaries, I had to manually modify some of the Makefiles generated by CMake.

I will take a look at the changes I had to make in the Makefiles and try to configure CMake to generate them.
I don’t have experience with CMake, but I am willing to learn.

I will keep you posted, probably in a private thread to avoid spam.
There’s no need to mention that I will team up with any other volunteer.

Cheers

i$

[Rico]

Great, thank you !!!

[idollar]

Hi Rico,

I have managed to compile The Leyden Jar tool in debian 12 (bookworm) by modifying the cmake confguration.

Following you can find the line I have modified and my addition:

Code: Select all

$ diff CMakeLists.txt.orig CMakeLists.txt
57a58
> 
67c68,72
< target_link_libraries(Leyden_Jar_Diagnostic_Tool PRIVATE hidapi jsoncpp_lib minlzlib)
---
> # i$ Original entry replaced
> # target_link_libraries(Leyden_Jar_Diagnostic_Tool PRIVATE hidapi jsoncpp_lib minlzlib)
> target_link_libraries(Leyden_Jar_Diagnostic_Tool PRIVATE hidapi-hidraw jsoncpp_lib minlzlib)
> # i$ Following line is added
> target_include_directories(Leyden_Jar_Diagnostic_Tool PRIVATE external/hidapi/hidapi)

I am not sure if the changes will affect the Windows build.

Cheers

[Ellipse]

Thanks idollar for your help with the Diagnostic Tool! I find linux/unix to be difficult, having used Windows for decades.

Over the years I have taken great interest in collecting and restoring consumer products and tools that are decades old, sometimes even about 100 years old, and still working just as well as the day they were made, including old electric Telechron / General Electric clocks, Western Electric rotary phones (candlesticks and 300-series phones), and of course the great IBM keyboards which led me to join these forums ten years ago. The main things they had in common were that they were built extremely well and these companies produced tons and tons of spare parts to keep their products running, what I aim for with the reproduction boards. For example the ancient metal framing square that I use for this project was in bad shape, completely rusted, but a vinegar bath brought it back to like new condition, and I am confident it can last for many more decades.

I have been having some interesting discussions over email with a number of folks about the battle between the competing ideologies of buy it for life, right (and responsibility) to maintain/repair etc. vs. the standard big company one-year unlimited warranty, planned obsolescence, and everything is done for you. While folks will disagree with one or the other stances, I offer my perspective in the writeup below, which has been added to the FAQ section on the product pages to help potential newcomers to the project. How can we get more non-disposable products into the world where appropriate?


"Is this type of keyboard right for me? More than 6,000 new Model F and Beam Spring keyboards have shipped to fans around the world in the past decade, but what am I really getting into here? Why would any keyboard require setup and maintenance? What should I expect? Am I willing to sacrifice some of my time to set up and maintain a keyboard that people have believed is worth maintaining for over 40 years, unlike nearly every other computer product that has come and gone?

Just as people have been doing for the past four decades with these keyboards on their own, for most things that come up you are expected to fix them. I am well aware this is not the standard attitude of manufacturers which sell you a short warranty, prevent you from maintaining your products afterwards, and want you to keep buying their products every few years after the previous one stops working. My goal with this product is that you buy it for life, tell your friends/family/coworkers about the project, and buy spare repair parts so if the project disappeared tomorrow you could still maintain and repair your keyboard 10 or 20 years from now. A key or spring not working? Take out your key puller! The good thing is that this is probably the most repairable mainstream, industrial-quality keyboard out there in widespread use and almost everything is easy to fix with nothing more than a screwdriver, key puller, tweezers, and pliers. Every single fix is spelled out with a step-by-step manual and video guide geared towards complete beginners.

A major part of the project philosophy is that you need to be self-sufficient and follow the manual which provides step-by step instructions, instead of trying to do things another way, and you agree to abide by the terms of this project in order to receive service in the event that something is not right (mainly you have to do the troubleshooting steps in the manual as the project doesn’t have the resources to hire folks to walk everyone through the manual on a one-on-one basis, and I am extremely busy coordinating the project and mailing out the order backlog (I’ve QCed and mailed out 6,000 keyboards and thousands of parts orders!).

This is not a hobbyist or kit project; the keyboard is fully assembled and every key position is tested, but there’s an exception: the user has to follow the step-by-step instructions to put on the keycaps and adjust any keys that are not working, and follow the other step by step instructions in the manual. The manual should not scare anyone; it includes every possible issue that could come up with the keyboard and the step by step way to fix it. Most folks only need a very small part of the manual, to install keycaps and adjust a few springs to get things fully working (it may take longer to go through the manual than to do the actual setup!). You won’t be left with a non-working keyboard if you follow the manual; the last step if the steps in the manual do not fix an issue is to contact me and I will help. To read more about the project philosophy, see the manual linked above.

To keep the cost of these keyboards as low as possible for people all around the world (so far to more than 70 countries!), and to avoid the need to charge everyone a lot more to pay for frequent mail-in repair service as with other products (think those $300 to $500 in-home or mail-in service warranties for prebuilt computer systems, etc.), this project has a different philosophy which empowers the end user to take responsibility for following the manual to set up and maintain their keyboard, instead of having others do the work for you; by ordering any item from this site you agree to abide by the project’s terms and conditions at the bottom of the home page and during checkout. One example is that this project’s concept of something being damaged or defective may be different from other standards, and you agree to accept these differences. For example, these products will come not completely working out of the box because you have to install the keycaps and set things up, which includes actions such as adjusting parts that are not perfect and exactly in place, including keys and springs. You don’t get to send everything back for someone else to do the work shown in the manual. Instead, you would go through the manual fully and contact me if the steps do not resolve the issue. Another example: all items sold here are made with cosmetic defects which are considered acceptable, such as surface variations for die cast and injection molded parts, minor paint chipping over time, and other things that you might just as likely see on the original IBM keyboards but not with other products. These defects have been vastly reduced with the later production rounds, however, so do not look to posts from years ago and judge the other 5,999 keyboards and later production rounds based on a few examples.

What attitude should I avoid? You can’t have the attitude “I won’t do what’s in the manual and I expect you to take it back and do the work for me.” It goes against everything explained here and it is why you have to agree to the terms and conditions before placing any order. This includes performing the troubleshooting steps that you are instructed to do, such as reseating and adjusting springs and other parts. If someone else is doing this work for you, you won’t become confident enough to be able to fix and maintain the board in the future and it will end up in a landfill, which is what I don’t want. This is not some product that is fully set up for you, just power it on and take it to the repair shop only for the duration of the one year warranty, and then get rid of it and buy the latest model when it breaks because of planned obsolescence and because it is not economical to repair. Do not try to come up with creative solutions to fix issues described in the manual; just follow each step in the manual which addresses every possibility including the last step that you need to email me, as a last resort, for support for something that is bad.

What if I am nervous? It is ok to not be confident about some parts of the manual especially if these types of keyboards are new to you, but once you learn it all becomes quite easy and it may only take you seconds or minutes to fix something in the future. No experience is needed to work on these keyboards. Over the years people have told me exactly this, so you should not worry, as long as you can follow along a step by step guide while being able to use a screwdriver, tweezers, pliers, and a key puller. If you’re unwilling to do what’s in the manual then you’ll have to ask a friend or family member for help as I can’t guide you. It’s not as difficult as programming the VCR, I promise!"

[Rico]

Thank you idollar !

Looks to be good
There is just one little modification to be done on top of yours to have both Windows and Linux compiling
Will do that soon!

[wobbled]

Thanks idollar for your help with the Diagnostic Tool! I find linux/unix to be difficult, having used Windows for decades.

Over the years I have taken great interest in collecting and restoring consumer products and tools that are decades old, sometimes even about 100 years old, and still working just as well as the day they were made, including old electric Telechron / General Electric clocks, Western Electric rotary phones (candlesticks and 300-series phones), and of course the great IBM keyboards which led me to join these forums ten years ago. The main things they had in common were that they were built extremely well and these companies produced tons and tons of spare parts to keep their products running, what I aim for with the reproduction boards. For example the ancient metal framing square that I use for this project was in bad shape, completely rusted, but a vinegar bath brought it back to like new condition, and I am confident it can last for many more decades.

I have been having some interesting discussions over email with a number of folks about the battle between the competing ideologies of buy it for life, right (and responsibility) to maintain/repair etc. vs. the standard big company one-year unlimited warranty, planned obsolescence, and everything is done for you. While folks will disagree with one or the other stances, I offer my perspective in the writeup below, which has been added to the FAQ section on the product pages to help potential newcomers to the project. How can we get more non-disposable products into the world where appropriate?


"Is this type of keyboard right for me? More than 6,000 new Model F and Beam Spring keyboards have shipped to fans around the world in the past decade, but what am I really getting into here? Why would any keyboard require setup and maintenance? What should I expect? Am I willing to sacrifice some of my time to set up and maintain a keyboard that people have believed is worth maintaining for over 40 years, unlike nearly every other computer product that has come and gone?

Just as people have been doing for the past four decades with these keyboards on their own, for most things that come up you are expected to fix them. I am well aware this is not the standard attitude of manufacturers which sell you a short warranty, prevent you from maintaining your products afterwards, and want you to keep buying their products every few years after the previous one stops working. My goal with this product is that you buy it for life, tell your friends/family/coworkers about the project, and buy spare repair parts so if the project disappeared tomorrow you could still maintain and repair your keyboard 10 or 20 years from now. A key or spring not working? Take out your key puller! The good thing is that this is probably the most repairable mainstream, industrial-quality keyboard out there in widespread use and almost everything is easy to fix with nothing more than a screwdriver, key puller, tweezers, and pliers. Every single fix is spelled out with a step-by-step manual and video guide geared towards complete beginners.

A major part of the project philosophy is that you need to be self-sufficient and follow the manual which provides step-by step instructions, instead of trying to do things another way, and you agree to abide by the terms of this project in order to receive service in the event that something is not right (mainly you have to do the troubleshooting steps in the manual as the project doesn’t have the resources to hire folks to walk everyone through the manual on a one-on-one basis, and I am extremely busy coordinating the project and mailing out the order backlog (I’ve QCed and mailed out 6,000 keyboards and thousands of parts orders!).

This is not a hobbyist or kit project; the keyboard is fully assembled and every key position is tested, but there’s an exception: the user has to follow the step-by-step instructions to put on the keycaps and adjust any keys that are not working, and follow the other step by step instructions in the manual. The manual should not scare anyone; it includes every possible issue that could come up with the keyboard and the step by step way to fix it. Most folks only need a very small part of the manual, to install keycaps and adjust a few springs to get things fully working (it may take longer to go through the manual than to do the actual setup!). You won’t be left with a non-working keyboard if you follow the manual; the last step if the steps in the manual do not fix an issue is to contact me and I will help. To read more about the project philosophy, see the manual linked above.

To keep the cost of these keyboards as low as possible for people all around the world (so far to more than 70 countries!), and to avoid the need to charge everyone a lot more to pay for frequent mail-in repair service as with other products (think those $300 to $500 in-home or mail-in service warranties for prebuilt computer systems, etc.), this project has a different philosophy which empowers the end user to take responsibility for following the manual to set up and maintain their keyboard, instead of having others do the work for you; by ordering any item from this site you agree to abide by the project’s terms and conditions at the bottom of the home page and during checkout. One example is that this project’s concept of something being damaged or defective may be different from other standards, and you agree to accept these differences. For example, these products will come not completely working out of the box because you have to install the keycaps and set things up, which includes actions such as adjusting parts that are not perfect and exactly in place, including keys and springs. You don’t get to send everything back for someone else to do the work shown in the manual. Instead, you would go through the manual fully and contact me if the steps do not resolve the issue. Another example: all items sold here are made with cosmetic defects which are considered acceptable, such as surface variations for die cast and injection molded parts, minor paint chipping over time, and other things that you might just as likely see on the original IBM keyboards but not with other products. These defects have been vastly reduced with the later production rounds, however, so do not look to posts from years ago and judge the other 5,999 keyboards and later production rounds based on a few examples.

What attitude should I avoid? You can’t have the attitude “I won’t do what’s in the manual and I expect you to take it back and do the work for me.” It goes against everything explained here and it is why you have to agree to the terms and conditions before placing any order. This includes performing the troubleshooting steps that you are instructed to do, such as reseating and adjusting springs and other parts. If someone else is doing this work for you, you won’t become confident enough to be able to fix and maintain the board in the future and it will end up in a landfill, which is what I don’t want. This is not some product that is fully set up for you, just power it on and take it to the repair shop only for the duration of the one year warranty, and then get rid of it and buy the latest model when it breaks because of planned obsolescence and because it is not economical to repair. Do not try to come up with creative solutions to fix issues described in the manual; just follow each step in the manual which addresses every possibility including the last step that you need to email me, as a last resort, for support for something that is bad.

What if I am nervous? It is ok to not be confident about some parts of the manual especially if these types of keyboards are new to you, but once you learn it all becomes quite easy and it may only take you seconds or minutes to fix something in the future. No experience is needed to work on these keyboards. Over the years people have told me exactly this, so you should not worry, as long as you can follow along a step by step guide while being able to use a screwdriver, tweezers, pliers, and a key puller. If you’re unwilling to do what’s in the manual then you’ll have to ask a friend or family member for help as I can’t guide you. It’s not as difficult as programming the VCR, I promise!"

What a whole load of bullshit
“ this project has a different philosophy which empowers the end user to take responsibility for following the manual to set up and maintain their keyboard”
Translates to
‘I dont want to pay the money to provide real support, so fix it yourself when your keyboard arrives with issues, or ask people for support on deskthority so I don’t have to deal with returns’.

You do realise that by selling to europe you have to honour their consumer rights? Your business practices don’t fly there. If i had an issue with one of your products it would be either a return of chargeback if its not something obvious.

And for fuck sake can you help others in the community like idollar who are asking you for your input on creating replacement flippers?
viewtopic.php?f=7&t=29392&p=521784#p521784
They dont want to pay your prices… you received support from this community at the start of your project detailing specs for model f components, yet you don’t give anything back to the community unless there’s money in it for you.

Please just be less of a scumbag and give back… Need I bring up you asking people to give you money to buy Deskthority even though you need this place to provide free support for your product that seems plagued with issues.

Give back.. and quit the greed.

[Ellipse]

The Chyrosran22 review has been posted for the new Model M case style F104 Model F Keyboard!

[Ordinary Witch]

it's nice that chyros got a mopar-style board to review since it has the same case color as the F122 I ordered, so this is a good comparison. I'm also getting the solenoid with it, and based on this video, I won't be disappointed when it arrives...

once the B122 project is done, it would be pretty cool if chyros got to review one of those, since he's a fan of >100%-sized keyboards. I asked him about this in his recent Q&A and he doesn't seem to be against the idea.

[DMA]

once the B122 project is done, it would be pretty cool if chyros got to review one of those, since he's a fan of >100%-sized keyboards. I asked him about this in his recent Q&A and he doesn't seem to be against the idea.

..to review B122, Chyros will need to find the missing keycaps first. Yes, I was made aware yesterday that B122 doesn't even come with a full set of keycaps. Boy I laughed.

[Ordinary Witch]

I suppose you could stick the relegendables on there that you can get from the store

[Ellipse]

Ordinary Witch yes I am hopeful that Chyros can review the new Round 2 beam spring boards next year!

DMA all of the beam spring keyboards do not include keycaps for those who prefer to use their own MX key sets. This allowed me to lower the base price of the keyboards since they were already quite costly to begin with. The beam spring boards are all MX key set compatible. For the PBT double shot key sets used in this project, the factory only made the mold for the regular full size keyboards, so you can either use your own set or the featured key set plus the 22 transparent relegendable MX compatible keys. I asked them about making additional double shot molds for additional keys and it would have added tens of thousands of dollars to the cost, which is more than all of the B122 keyboard project!

[wobbled]

Ellipse you gonna give back to the community and help share details in this thread?
viewtopic.php?f=7&t=29392&p=521792#p521792
I look forward to not hearing from you, and for your next wall of pure bollocks

[Ellipse]

Recently I was discussing the nature of this project some more on other forums. I am surprised that this has become a relatively lively discussion once more, about how scary the phrase "some assembly required" is to some folks these days. I've always been used to putting together furniture, building computers from parts, and restoring vintage stuff to working condition so I do not associate high end with "no assembly required, maintenance-free." For some things the whole point is in the joy of assembly (remember Lego bricks, K'NEX, Erector Set, Capsela, Lincoln Logs, etc.)

The new Model F project is here for the primary objective of creating lifelong Model F users who are confident enough to spend some time to set up and maintain these keyboards, not people who are used to a maintenance-free product that is discarded because no one knows how to fix it, and then it ends up in a landfill. Those folks may better be served with another product and there is no issue with discouraging folks who cannot adjust their outlook to match that of the project.

As was stated in the Project Management Triangle that I have brought up before, everyone would like products to be good, fast, and cheap but those are all opposing forces. To hire workers to assemble and mail orders faster would entail a higher cost. To set up keyboards for everyone (make it "good" for some perspectives) would require a higher cost. To keep the cost low, compromises have to be made and the biggest barrier to this project for most folks is the original $800 cost that IBM charged (adjusted for inflation). The first compromise was "fast" (speed) in that the project is essentially hand assembled with the help of some custom machines but with no expensive pick and place machines for keycaps, etc. Nothing matters if you are happy with the idea of a product that is high quality and fully ready to go, gets to you fast, has free tech support and no maintenance, if you can't afford it to begin with because it has IBM's price tag that was required to maintain all of IBM's fixed costs including automated installation machines, assembly workers, and support personnel. I will continue to remain steadfast in the project philosophy choosing high quality products at the lowest possible price for the most part, in order to maximize the number of folks who can experience a Model F, while trying to set expectations with folks who don't want to learn how to keep their keyboard working for life.

A suggestion about asking a US-based keyboard factory such as Unicomp to perform all of the setup labor for everyone highlights that some posters may not understand the differences between these two types of keyboards and about the nature of the opposing forces mentioned above. The Model M has zero parts that are shared with the Model F, with one exception: not even the springs and keycaps are the same, though the keys are compatible. The Model M I believe has a custom-made assembly jig that installs the keycaps, which would be cost prohibitive to remake and configure for the dozens of different keyboard layouts available on the Model F as opposed to the 4 layouts of the Model M (SSK and full size, ANSI and ISO for each). In other words these other factories would have to install keys by hand, so there are no savings with automation. This suggestion would only increase the cost of these keyboards because each person getting a keyboard would have to pay extra for someone to do the setup work for them. The only other alternative is to cheapen the quality of the keyboard to free up some of the cost to pay for assembly, which is not something I'd like to do. Unfortunately we just have to admit that there is not as large of a market for very high quality goods any more compared to what IBM was able to mass produce with their IBM PC's in the 1980s. Even IBM admitted this as they eventually had to discontinue / spin off their buckling spring keyboard operations when the market changed in the 1990s. Eventually some product's one-year support warranty will run out and then you'd be in the same place as you are when you get a Buy It For Life Model F keyboard, but with one difference: you wouldn't necessarily know how easy it is to repair the keyboard so you remain too timid to keep that key puller and tweezers handy for a quick maintenance fix.

It seems like those willing to install a few keycaps for a Model M would be perfectly at home with the Model F, which would require no different skill set from what would be expected of you as a Model F user, except you'd need more time to install the keys yourself and you may need to spend a few minutes moving some springs back into place. It sounds scary but it is not; the video in the manual shows that it is an easy process that can be learned.

How many folks are willing to pay $575 and wait several extra weeks for an already costly $400 keyboard just to have a US-based warehouse to install the keys and adjust some of the springs, if some springs and keys still need adjustment after shipping anyways? Either way outside of service professionals making house calls you would end up in the same place as you are now.

The project has had zero changes in this philosophy for the past 8 years and 5,000+ people who have supported the project - everyone has had to set up their keyboard, whether or not the keys were preinstalled. Every now and then a discussion comes up where some folks are surprised that they have to install keycaps or perform small maintenance to keep their keyboard running. To those who have been maintaining and collecting original Model F keyboards, the need for some small maintenance is definitely no surprise and it is in line with my own experience. I've had to completely disassemble and reassemble my original Model F keyboards, sometimes more than once.

Having to install keys is not buried, it is noted on the home page and on every product page. Very few folks think that it is the most important thing about a keyboard as many other high end keyboards have their keycaps separately bagged.

In summary, I feel the project's current stance on factory assembling as much as possible (everything but the keys and accessories), requiring everyone to learn how to set up and maintain their keyboards for life, and keeping the keyboards at less than half the cost of what IBM charged were the most important priorities that were achieved with this project.

So far I am unconvinced that the project could sustain even higher costs that would have to be passed on to each end user, or that volumes would increase so much from advertising these keyboards as having the keycaps pre-installed that the project would be able to recoup the additional costs by having a higher quantity of keyboards ordered. As some may recall, a couple hundred boards from the very first batch of some keyboard models had pre-installed blank keycaps with the factory adjusting key and spring seating where needed. In addition to the factory needing an extra month or two to perform this labor by hand, thus delaying the entire container shipment, these keyboards still needed some adjustment of springs and keys after they arrived off the container ship to me, and certainly some other keys fell out of place while in the final leg of transit from me to the end user. The time to adjust these keys even with a preinstalled board was not sustainable when trying to mail out a couple thousand boards and it resulted in everyone's boards taking longer, both for the factory to install them and for me to test and adjust each key and move through the backlog. For this reason I don't think installing keycaps is likely to reduce the need for any user setup and maintenance. I don't see a "maintenance-free" Model F as possible without additional compromises around cost and/or quality or a network of repair folks like IBM had when they rented the 4704 Model F keyboards to bank branches.

[wobbled]

Ellipse no one is wanting to read 6000 words of pure shite.

Be a good human being and replace someone’s case when it arrives with dents and paint damage. These are not factory issues, they are shipping damages that you are responsible for.

Everyone understands that some issues can arise and some time must be taken to set these boards up, no one denies that, but you actively go out of your way to deny peoples warranty by claiming a DENT / paint damage are ‘normal’ and ‘expected’ and your recommendation is a fucking paint pen?!?!?!?

Find me a car dealership that would sell you a new car with paint missing, and then claim its expected.

Ellipse you always post ridiculous amounts of shit to try and divert attention away from the biggest point here - you are being scummy. Give people refunds/ replacements when there are clear damages.

[Vla]

Ellipse no one is wanting to read 6000 words of pure shite.

Be a good human being and replace someone’s case when it arrives with dents and paint damage. These are not factory issues, they are shipping damages that you are responsible for.

Everyone understands that some issues can arise and some time must be taken to set these boards up, no one denies that, but you actively go out of your way to deny peoples warranty by claiming a DENT / paint damage are ‘normal’ and ‘expected’ and your recommendation is a fucking paint pen?!?!?!?

Find me a car dealership that would sell you a new car with paint missing, and then claim its expected.

Ellipse you always post ridiculous amounts of shit to try and divert attention away from the biggest point here - you are being scummy. Give people refunds/ replacements when there are clear damages.

Ellipse is running one of the worst stores in keyboard history as a borderline scammer. All the bad stuff he does is illegal in some countries...

[Ellipse]

Instead of focusing over and over again on the same few examples, I suggest that folks considering these keyboards read through all of the comments on the Chyros video for a good assorted sample and see whether the positive comments from new Model F users outnumber the negative ones 10 to 1, or vice versa. If I was refusing warranty service and all sorts of parts were arriving broken and unusable don't you think most comments would not be positive? The above generalized accusations are completely false and unhelpful.

The damage in the one example continuously being brought up was not a dented case; that is just false. Repeating the same false point that I did not replace this F104 user's dented case only goes to show that there are no real examples that support a claim of widespread issues. If there were, you wouldn't have to falsify or exaggerate to make a claim. Most issues were resolved by folks reading the manual to see how to adjust the springs and correctly install the keys.

The black painted case in question had a few mm of paint loss in the corner of the case in the back and some 1mm die casting surface defects in the back. That's it. I have repeated many times that this kind of paint loss can be fixed with a $5 sample can of paint or a paint pen, and all of the cases have casting marks. Again, I will note that from the beginning of the project, all cases have cosmetic damage. If you sign up for the project you agree that your case will not be perfect. This is clearly disclosed on every keyboard product page and in the terms and conditions that these keyboards are not perfect cosmetically. The project philosophy is not "I expect a perfect product with no cosmetic imperfections and I can return it for a free exchange otherwise" - instead, the philosophy is that if the user can fix something quickly and easily, then the user has full control over the fix, in exchange for bringing the cost of these keyboards down by 50% of what IBM charged, adjusted for inflation. If you can only have a perfect finish, you do not have to get a new board from this project. A company could certainly offer full service, fully set up, no-questions-asked repairs and in-home service, but the cost would approach the $800 to $1000 that IBM charged, so it is unfortunately impossible to achieve the project's objectives making the Model F as affordable as possible for everyone (see my earlier post above). No one can force the project to raise costs to cover "free" international shipping and free replacement cases for minor paint loss affecting a handful of the hundreds of keyboards that have shipped recently.

Chyros has fairly emphasized that you have to set up these keyboards and perform steps yourself for things like reinstalling keycaps and reseating springs. In the rare case that something arrives broken like a keycap then I have been replacing it for free, despite one YouTube user deciding to buy the caps anyways, right after I said I would replace a key for free if it was damaged.

From what most folks report, they spend about an hour or two to go through the manual and perform the setup steps, which primarily consist of installing keycaps, adjusting a few springs, and pressing each key to test it before everything is ready to go. The manual is set up for complete beginners to the Model F to get up and running, so most folks should not worry about overblown claims. At least 90% of the issues require just reinstalling a key or reseating a spring, though the manual is comprehensive and includes all possible issues that could come up with a Model F.

In summary the project fully disclosed a note on cosmetic defects on all keyboard product pages and that the end user has to follow the manual to set up the keyboard. No one is being left with a non-working keyboard. Broken keycaps and controllers have been replaced for free at my expense; no cases have suffered a dent in shipping and a warranty claim was denied, as was falsely claimed several times. Criticizing the project is all good and fine but let's try to keep the discussion cordial and try not to beat a dead horse as many folks have said about the same few people repeating the same points over the past few years. Nothing has changed regarding the core project philosophy.

[wobbled]

Once again with the wall of text…

You are seriously recommending that someone with paint damage on a brand new board use a paint pen? Literally no other retailer would get away with this, you are responsible for the safe arrival of these boards. Replace it and stop being cheap.

If you need to find more examples of people having a negative experience with you then take a minute to read through the keyboard institute discord (formerly DT) or even scroll back through this threads many pages and you’ll find complaints, they might be hard to find though as they are usually replied to with bible length bullshit from yourself

You will also find complaints in chyros’s video but you’ve already replied to them with ‘THIS IS FALSE’
Followed by BS.

[tron]

Would it be possible to offer a paid option on the site to pre-install keys and configure the keyboard before shipment?

For what it’s worth, I would rather do some minor installation and setup for a new model F vs restoring an older one. The casting dents and paint chips on the back side of the case are no big deal for me, however, if there are large visible chips/scuffs on the front side of the case, I would think the case should be sanded down and re-painted as most people would think this is unacceptable for a product in new condition. Probably best to handle on a case by case basis IMO.

[Ellipse]

tron fortunately we are not seeing major flakes or dents or things like that on the front. The newer cases are far better with the updated paint formulation but no uncoated paint finish will be free of cosmetic imperfections, whether on the F62/F77 or on the newer rounds. As noted in earlier posts I unfortunately do not have the time or personnel resources to set up everyone's keyboard, and it still doesn't create a setup-free experience as keycaps still fall out of place during shipping even when they are pre-installed. One of my earlier posts noted how some keyboards did have factory preinstalled keys (the unprinted keys) but setup was still required. There are hundreds of keyboard variations and dozens of key sets, so it would not be practical to guess at which variations folks will be ordering since each production batch is so small. Then I would have to spend time removing key sets if I guessed wrong, which would further delay the backlog.

I must be protective of my time because I am only one person moving through an ever-present backlog and people waiting for their orders. I have to make the project as much self-service as possible with all the various educational materials that have been put together, or I would not be able to provide these keyboards to as many folks as possible without having to hire additional staff and raise prices to cover these extra costs. That is the only way this project can survive at a fair cost to everyone. I would love to personally teach every person individually or in small groups like with an online seminar as that is also a great way to learn new skills, but I don't have the time sadly. It would be a cool idea if there were virtual "keyboard setup Zoom events" every week for folks to join in on to set up their keyboard after it arrives. What I did was put my efforts and experience into the manual and into a setup video so that everyone can see how to maintain these great old style keyboards.

As I keep mentioning, all Model F items are disclosed to have cosmetic imperfections and people know this before buying because it is written on the keyboard product pages. My position on the project philosophy is firm and everyone can make their own decision to accept this or consider another offering. It is easy to find faults with any product as nothing is perfect; there are always tradeoffs that some folks will disagree with. So yes, one tradeoff is that each board will not be replaced for cosmetic imperfections but parts will be replaced if they don't work. Another tradeoff is that the users are responsible for taking control of the setup and maintenance experience instead of returning the keyboard to a repair shop if one key does not click well, for example. Without these tradeoffs the project could not be priced at less than half of what IBM charged for these keyboards and the keyboards could not be offered around the world.

[abu nuwas]

Got my classic F104 + Arabic layout - loving it so far! Thank you, Elipse.

A stabilizer was missing, but I salvaged one from my Unicmp M. Installation was a breeze - not as intimidating as I thought.

One odd thing though - my num and scroll lock LEDs don't work. Caps lock works fine. Can this be fixed in software? I use linux.

Cheers!

[pelletik]

I got my F104 keyboard today, just in time for Christmas. Thank you, Joe @Ellipse!

I cracked open a beer and watched the hour-long setup video. The video is well done—thank you for creating it. An hour later, the keyboard was up and running. This is my third keyboard from you. My only regret is not purchasing the solenoid. I plan to purchase the F122 and a beamspring when they become available.

This is the best keyboard you’ve made. The limitations of the Model F AT and Model F XT are in their layouts, and the F104 is the solution. This reproduction is true to the original Model F. The switches are identical—I cannot tell the difference between my Model F XT and this keyboard. This keyboard will be my workhorse.

I’ve been using my F77 for several years and have really enjoyed it. However, it’s going into retirement for now.

Again, thank you for this project. Keep up the great work, and have a Merry Christmas!

[Ellipse]

Feedback request - for those outside the US and who received their new Model F keyboard through UPS, how was the delivery experience? Did the box and keyboard survive OK or did the box get roughed up in shipping?

abu nuwas glad your keyboard arrived and you set it up already! My guess is that maybe the two led leads might be making contact with each other, thus resulting in a short. I test all 3 LEDs on each keyboard as part of the testing process, so all LEDs were working when it starts its shipping journey. Of course please keep me posted on this and let me know if you are able to fix it. If not we need to determine which part is not working, whether it is the LED, the resistor, or just a loose connection of the ribbon cable. I have extra resistors and LEDs from the factory and can send replacement parts if needed of course, at no charge as I do for anyone who receives a non-functional part. I also suggest testing on another non-linux operating system to see if it is something specific to your install. I do not think I have seen any reports of new Model F LEDs not working. https://askubuntu.com/questions/812967/ ... mode-is-on

pelletik thank you for your positive note and glad you were able to get everything set up well! Please do let others know about the project, if you don't mind. And Merry Christmas to you too!

[genericusername57]

Feedback request - for those outside the US and who received their new Model F keyboard through UPS, how was the delivery experience? Did the box and keyboard survive OK or did the box get roughed up in shipping?

I got my F62 to Sweden about a year ago and it arrived in stellar condition with no dents to any packaging whatsoever and the keyboard itself has no imperfections that I can see. The only weird issue I've had is that my Tab key was brittle unlike the rest of them, so when adjusting the legs of the key it snapped clean off as soon as I applied pressure. I'm not sure if that's a manufacturing defect or something. The rest of the keys have been flexible and able to be adjusted slightly to reduce scratchiness. Since it was a clean break it was easily fixed with epoxy but I took the opportunity to order some keycaps from Unicomp to add some colour.



And yes I know the IBM badge wasn't on the originals but I like the look.