Which will work great until you need to run for the bus or something and slam the lid down on yer thick-ass keyboard. Crunch!
1 Like
Being able to have a custom keyboard layout is what I’ve been dreaming off since seeing the first Framework announcement.
Has there been any news regarding fully custom keyboard input modules, ideally with low profile mecanical switches? What is missing for it to work? Is it just a vertical clearance issue?
If so the Kailh PG1316M could fit:
The PG1424 series could also be short enough:
Knowing that and the fact that the connection is a simple USB interface that’s simply laid flat, how hard is it to make a fully custom keyboard input module?
3 Likes
The math on those switch heights doesn’t quite work out. The total “hard” thickness available for an input module is only 3.1mm (aka the “key down” thickness), with a soft/pressable (“key up”) thickness limit of 3.7mm. The PG1316 switches are 5mm tall (including keycap and 0.8mm PCB, according to the write-up on the mikefive keyboard which uses them) so you’re ~1.3mm over the available soft space. In the keydown position, they are 3.1mm so technically compliant, but the LCD is going to be pressing all the keys down by about ~1.3mm when the lid is closed, which might leave a mark.
The PG1425 are similar, with a 6.4mm “key up” height and 4.40mm “key down” height (the measurements in the datasheet are a little misleading as they don’t include the 0.5mm switch legs).
That said, I think you could add a ~1.5mm shim to the laptop hinge without /too/ much difficulty, and I’ve got some parts I’m waiting to get back from being printed to validate that. If shimming is reasonable then a lot of options open up. But any keyboard build is going to have to be super height-conscious.
2 Likes
Thanks a lot for the detailed response!
As I can now see 3.7mm is really, really short indeed. My low profile keyboard wich is “short” by most standards is indeed way over 1cm (1.6mm hotswap sockets + 1.6mm PCB + Kailh v1s + keycaps).
And so, so far, there has not been a single fully custom Framework keyboard, right? Not even with more traditional silicone designs?
1 Like
38 Likes
Holy…
You guys are fabulous. I know you said you were trying to see if single key modules were possible. But I feared it would have to be shelved, just have too many other more important things to deal with. Great to see it’s progressing!
2 Likes
Yeah, this is amazing. I’m very excited to learn more about these modules and how they will work. Thanks for the update!
Here is the idea, of how you’d you them
You’d attach them to a plate, similar looking to this, which is used with custom mechanical keyboards. Except, it would need to provide the key matrix traces in addition to providing structure.
1 Like
Ah great, that looks achievable. Is the MCU going to be on the PCB lattice then too? Will that be something Framework can provide, or maybe at least provide some guidance for those that are entry-level hardware designers?
Since it’s meant for custom key arrangements, I don’t think Framework would be making the lattice plates. It’s possible they might do one. I think they would likely create a pcb example file at least. And guidelines for material choice to ensure it’s strong and rigid enough. Maybe polyimide with stainless steel stiffener.
I’m sure a lot of keyboard fans will jump on this & share their designs. Maybe someone will share an ordering step-by-step for those who aren’t experienced with custom pcb design and sending them to be made. Take these files, upload to jlcpcb, select these options, etc.
All assuming that Framework’s single-key module makes it to production. Everthing is in a really limited space, and something like this has never been done before.
1 Like
There’s already a microcontroller sample file in the InputModules git repo, under “MicrocontrollerInputModule”. It uses an RP2040. You’ll probably need an I/O expander, though. I’ve got some designs open sourced already for my X1 Carbon Gen 10 keyboard that can be generalized. Maybe make a blank PCB with the right shape for an input module and the microcontroller (and I/O expander?) in the corner, then folks would just need to “add their keys”.
5 Likes
Lordy, I hope so. It is with great regret that my work has provided me with a MacBook Pro. I have come to realize that the keyboard is really, really great. And typing on my Framework has become somewhat joyless.
I wish it weren’t so. I fervently do. But I am here scrounging around looking for relief from someone, anyone. Seeing this fills me with hope! I’ll have my eyes peeled on this.
I’m delighted this project is still going. But I cannot help feeling that instead of designing, engineering, testing to millions of cycles, putting into manufacture, distributing and then dealing with any after-sales issue (many of which will be customer caused) it would not have been much simpler to have just made the chassis a smidge deeper to accommodate off-the-peg Cherry ULP! 
Edit: added smiley because some apparently need it spelled out to them I’m purposefully being irreverent!
The chassis is already made. We’re years past that point. So you’re just hassling them over a mistake you perceive they made? Targeting a certain thinness. A little extra thickness can sometimes be enough for a big reviewer to decide to call something a bulky brick vs acceptable.
2 Likes
@David_Eastham i feel like this conversation has already been thoroughly played out, but just in the interest of trying to button it up one final time:
yes, they could have made the chassis a little deeper. they could have done a LOT of things, that SOMEONE would have found more desirable in some respect. but when you’re making a product, you have to make decisions. determine what features you need, versus what features you don’t need. analyze the tradeoffs between one option and its competing options. settle on the breakpoints of what best serves the design you are attempting to make into reality.
they made a different choice than you on this point. they made different choices than i would on several points. that does not make the choices they made incorrect, or ill reasoned, or ill intentioned, it just makes the choices they made different. belaboring it endlessly definitely isn’t going to make history change, and it’s probably not going to convince them to re-engineer it for what is very clearly an exceedingly niche case.
please, i beg of you, find a way to focus on a viable way forward for your needs.
5 Likes
Much as I like needling the zealots on this forum I think the announcement of these switches does deserve an expanded and more earnest response, so here goes.
No, I’m responding to recent developments on the forum, if anything a little late to the post back in July. I would love to think that they produced a photo of a prototype key-switch just to shut me up, but I have not really been active here for a little while. Looking up thread, my last post mentioning discrete key switches is back in July of last year.
Paint me skeptical, but I cannot get so excited about a picture of a black rectangle purporting to be this switch. Whilst it gives an idea of how it might look in practice many of us will still have questions. Hopefully those in the know can enlighten me:
You ask the first better than I could: will it work? Is the object in the photo functional?
@Tony_Grosinger asks the next question: is it going to be practical for the amateur to build their own keyboards - the whole point of the exercise after all?
As for deeper cases, the responses from @MJ1 and @eso have dug their heels in on the issue. These responses risk painted us into a corner where the discrete key-switches must to be made to work if custom keyboard formats can happen. Since the alternative apparently requires a loss of face from the firm and more strident members of the user community.
Good news here: I am quietly optimistic about the new MNT Reform model (the Next) which has just been announced. That boasts a proper mechanical keyboard by default, so loads of depth for customisation, from a firm whose concept of open source hardware is on a whole other level.
As for the weariness I detect in your posts, that is a reflection of how long it has taken to this point. I actually felt similar weariness - some might say strangely - at the picture of this supposed key-switch. Part of me would have been happier if the whole thing had been quietly put to bed and we could - as you suggest - move on rather than try to get our hopes up again.
1 Like
Hello hello. Pardon me, but does anyone know if there are KiCad or Eagle footprints for this? I’ve been trying to design a full-size split keyboard but I always get stuck on the PBC Design 
It seems to still be in the prototype stage.
That I can understand, but what key switch is it based on? Cause on can assume it’s PG1316, looking at the pins, but I rather ask to be sure.
Edit: Reference; https://www.reddit.com/r/ErgoMechKeyboards/comments/1cfg3vr/mikefive_a_kailh_pg1316_keyboard/
1 Like
Sorry, I must have missed notification of your reply. Very late response, but it wouldn’t likely be a Kailh PG1316 switch. The FWL16 has 3.7mm of height available
Kailh PG1316 is over that even without the keycap or PCB.
I don’t think there is any off-the-shelf keyswitch module that is thin enough. Cherry MX ULP is 3.5mm, but it needs to be connected to a PCB underneath it. And then room for a keycap. cherrymx.de/en/cherry-mx/mx-ultra-low-profile/mx-ulp-click.html
I would think Framework is working on a custom module, perhaps based off of components of keyswitch mechanism used in their current FWL16 keyboards.
2 Likes