It is insane to print your own switches when you can buy extremely high quality ones for 25 cents a piece. And that level of insanity makes this a great project.
The advent of the fluid amplifier has led to the development
of data processing and control systems wherein the
processing and control functions are carried out by elements
which operate entirely on fluid principles. That is, these
systems contain no moving parts other than the working fluid
which flows through the system.
Therefore, an object of the present invention is to provide
a pure fluid system for inserting information into a fluid
data processing device.
> Apart from the fact that the keyboard is quite unpleasant to type on, I'd say this project was mostly a success. [...] I think my mistake was optimising for reliability instead of feel.
Existing switches are extremely reliable (if their reliability isn’t satisfactory, find some used Alps switches from the late 1980s, or some Honeywell switches, or ...), so the only reasons I can think of to make your own are to fit in some unusual spatial condition where an off-the-shelf switch won’t work, or for a better feel.
Yeah I think this was something like the guy who built a toaster entitely from raw materials* such as iron ore. It's a terrible toaster, but there's something to be said for learning first person what goes into the modern things we take for granted.
This is a neat project. I printed some switches out when I saw his post on Twitter I believe. It dovetailed into work that I've been doing to figure out replacement switches, buttons, and sliders using DIY 3D techniques vs sourcing through NKK or what not. Primarily for cost since any custom switch design is like $25K for the first 100.
As others have pointed out, these keys feel "mushy". And I agree with that. It comes down to the spring constant of printed springs. It is really really hard to make it consistent from print to print. Fortunately springs are actually super cheap and you can make those yourself pretty easily too if you want. So if you back off the "everything has to be printed" mantra just a bit. You can make a more consistent feeling switch. (and less deep which has some benefit on front panel stuff)
> Spending some time typing on my homemade switches has given me a newfound appreciation for just how good Cherry MX switches are.
I know at least one mech keyboard enthusiast from reddit who went from cherrry mx brown -> clears -> ergo clears -> gateron brown -> zealios v1 -> zealios v2 -> custom switches -> cherrry mx brown making a full cycle.
I assume you're talking about the "feel" of Cherry switches. Which many people really like.
As for reliability, I don't know how readily available they are any more, but IMO the true "gold standard" is Hall Effect, because it's contactless.
Hall effect keyboards used to be quite common 45 years ago in heavy data entry applications (e.g. people who spent their days entering typewritten manuscripts submitted by authors, years before every writer used a word processor).
Contactless is great because there's no mechanical key bounce or contact wear. Some quick googling showed that Cherry (at least once upon a time) made hall effect keyswitches. But they're probably too expensive for most applications.
I went scissor switch -> reds -> blues -> low profile blues -> low profile browns -> butterfly -> scissor switch.
My favourite of them all is probably butterfly, but obviously there's no equivalent of that for desktop keyboards, so I am running a scissor switch keyboard now. However, I have very fond memories of the low profile blue switch.
I went Blue > Brown > Clear > Blue > Silent Red (on order, waiting for delivery)
I went full circle back to blues, and blue are still my favorite... but the rest of my family does not share my enjoyment of them. I will go back to blues after they go back to work/school. :)
> "I toyed with the idea of getting a PCB made up that I could directly solder my switches to. It would certainly have made it easier to put the keyboard together, but I thought outsourcing the manufacturing of a PCB kind of takes the edge off making your own switches..."
Totally! There are ways to make your own PCBs using presensitized boards and exposing it to light ([1] for example) to keep it DIY
Custom-built keyboards are very cool, but I will never understand putting all this love into a non-ergonomic design. Typing on something with no separation or tilt murders your wrists! Mechanical keyboard enthusiasts will spend ages refining & debating every possible aspect of the individual switches, then lay them out on a flat slab like it's a giveaway Dell keyboard from 2006. At least the design he chose avoids a number pad, which either makes you situate the keyboard more to one side of your body or pushes the mouse too far out to use comfortably.
This is pretty sweet. One thing I've always wanted to do is embed a hardware keylogger into the keyboard itself (optionally with bluetooth or some other radio - so I can retrieve contents of logs remotely).
Seems like with one of those you could even defeat 2FA (if you were a black hat). Swap out the victim's keyboard with the one with the embedded radio keylogger. Auto-detect when they are logging into something with 2FA. Put the characters they type into the web form and submit faster than they do.
Good job! However, according to the wiring seems like it wouldn't be able to register something like ctrl+shift+z which is used as redo hotkey almost everywhere.
The good thing about QMK is that you can remap anything you want to anything you want. For a hobby project, I don't know how much it would matter anyway.
A guy in the mechkeys discord posted a clip of his 3D printed mech switches yesterday as well. I believe he is using reed switches instead of a 3d printed leaf but its pretty cool.
Kind of surprisingly, doing a quick search for that term online shows up a 1960's patent for one:
https://patents.google.com/patent/US3034628A/en
If anyone wants a challenge... ;)> Apart from the fact that the keyboard is quite unpleasant to type on, I'd say this project was mostly a success. [...] I think my mistake was optimising for reliability instead of feel.
Existing switches are extremely reliable (if their reliability isn’t satisfactory, find some used Alps switches from the late 1980s, or some Honeywell switches, or ...), so the only reasons I can think of to make your own are to fit in some unusual spatial condition where an off-the-shelf switch won’t work, or for a better feel.
* https://www.ted.com/talks/thomas_thwaites_how_i_built_a_toas...
As others have pointed out, these keys feel "mushy". And I agree with that. It comes down to the spring constant of printed springs. It is really really hard to make it consistent from print to print. Fortunately springs are actually super cheap and you can make those yourself pretty easily too if you want. So if you back off the "everything has to be printed" mantra just a bit. You can make a more consistent feeling switch. (and less deep which has some benefit on front panel stuff)
I know at least one mech keyboard enthusiast from reddit who went from cherrry mx brown -> clears -> ergo clears -> gateron brown -> zealios v1 -> zealios v2 -> custom switches -> cherrry mx brown making a full cycle.
As for reliability, I don't know how readily available they are any more, but IMO the true "gold standard" is Hall Effect, because it's contactless.
Hall effect keyboards used to be quite common 45 years ago in heavy data entry applications (e.g. people who spent their days entering typewritten manuscripts submitted by authors, years before every writer used a word processor).
https://en.wikipedia.org/wiki/Keyboard_technology#Hall-effec...
Contactless is great because there's no mechanical key bounce or contact wear. Some quick googling showed that Cherry (at least once upon a time) made hall effect keyswitches. But they're probably too expensive for most applications.
not quite readily available, but that may change soon
My favourite of them all is probably butterfly, but obviously there's no equivalent of that for desktop keyboards, so I am running a scissor switch keyboard now. However, I have very fond memories of the low profile blue switch.
I went full circle back to blues, and blue are still my favorite... but the rest of my family does not share my enjoyment of them. I will go back to blues after they go back to work/school. :)
Totally! There are ways to make your own PCBs using presensitized boards and exposing it to light ([1] for example) to keep it DIY
[1] https://www.youtube.com/watch?v=7wAer7a3tU4
Seems like with one of those you could even defeat 2FA (if you were a black hat). Swap out the victim's keyboard with the one with the embedded radio keylogger. Auto-detect when they are logging into something with 2FA. Put the characters they type into the web form and submit faster than they do.
https://www.youtube.com/watch?v=MwRfQ3p_jFo