(Model MF) Remodeling the Model M (aka.. the Mara)

[lot_lizard]

One thing, you mentioned that you intend to open everything up, which is just awesome. I don't know if there were any links in the later pages, but I have a suggestion... Could you open a github repository for all this stuff, and link to it in one of the opening posts? For all the mechanical and electric CAD files, production notes, firmware and layout files, etc. It would make collaboration and access much easier.

I have been lazy about this bit. I wanted to wait until nothing would be altered before I threw everything out there, and clean up the OpenSCAD code a bit, but I could (and should) finally get all of that up there. We do have a repo for it, but there is very little in there currently. Plan on it being the resting place for all of the MF stuff. At minimum I should throw all the STL's and PDF's up there (production pieces) right away. Thanks for reminding


https://github.com/lot-lizard/Model-MF

For a take on XT modernisation, check out XTant project at
workshop-f7/bringing-the-ibm-pc-xt-into ... 7-120.html

Exactly... the following is the repo WCass put out there for the XTant if it didn't jump out in the thread.


https://github.com/wcass-/XTant/tree/master/XTant

[Sigmoid]

Cool, can't wait to see the files. So you used OpenSCAD? How comfortable was it to work in for a large project such as this? I guess it's good for the task, when I was playing around with it around two years ago, it was still lacking a lot of functionality...

For a take on XT modernisation, check out XTant project at
workshop-f7/bringing-the-ibm-pc-xt-into ... 7-120.html

That thread is kinda cool, but I didn't mean modding existing XTs. I really don't want to start hacking up my XT keyboards, I'd feel bad messing them up, I can work around their outdated aspects, and besides, that gigantic whale of a spacebar is just pure lovely, kinda like the oversized hood of an old Jaguar...

(And besides, TKL? For me, one of the most lovable aspects of the XT keyboard is that it's not TKL... )

What I was thinking of is creating a ground-up new keyboard based on the XT layout as inspiration.

[Techno Trousers]

What I was thinking of is creating a ground-up new keyboard based on the XT layout as inspiration.

I think this is the first time I've ever heard from someone who actually prefers the XT layout. I'm glad that such a project will be possible for you to undertake in the future, though!

[lot_lizard]

So you used OpenSCAD? How comfortable was it to work in for a large project such as this? I guess it's good for the task, when I was playing around with it around two years ago, it was still lacking a lot of functionality...

This and the Cherry adapters were the first time I ever used it. To be honest, it was quite a horrid experience for the plates themselves. I know AutoCAD and Solidworks quite well, but this was intriguing to try, so gave it a whirl. The code itself was quite pleasant to use (I am... or was... a developer by trade), but the engine is just not ready for something like this because it is meshed. Generating simple renders of the top plates would literally take hours if I wanted any real detail because of having so many facets (curved plates, many holes, etc.). I typically worked in cylinder facets between 4-8 per hole (not horrid), and then would dial that up to 64-128 facets when generating renders. Eventually when things were hardened, I managed to export out to STEP, brought into something that was easier for me, and regenerated as a proper object model killing off all the individual paths (from scratch really using the import as a template).


Moral, I really like it for small parts and am glad to have learned it. Would certainly use again, and is great for sharing because it is literally all a C-based language (all the math is fun). It is definitely a toy product when compared proper ones though, just the nature of meshes unfortunately.

EDIT: corrected "Realworks" to "Solidworks". Ducking autocorrect

[Sigmoid]

I think this is the first time I've ever heard from someone who actually prefers the XT layout. I'm glad that such a project will be possible for you to undertake in the future, though!

Haha. Well I'm not saying the XT is perfect, just that I prefer it to any TKL any time.

Personally, I dislike TKL, as I consider the "nav cluster" to be the next most useless thing on keyboards after the Scroll Lock button (and hey, the nav cluster contains the Scroll Lock button, so yea...) Getting rid of the numpad (which does everything the nav cluster does, and better), and keeping the nav cluster just seems like bad engineering.

So the thing with the XT layout is that it's a great concept with poor execution. The upsides of the XT in my eyes are: a numpad coupled right onto the main keyboard area, no top Function row, ESC on the numeric row, Ctrl on the home row.

On a neutral side, the gigantic Spacebar is weird, but I find it charming - if we don't have Command keys, then I prefer it to the empty cutouts on AT keyboards. The function cluster is a nice retro touch, though I'm not sure it adds anything compared to having functions on fn+num combos. The huge numpad (+) button makes sense for cash register-style long additions (like when doing taxes or something), but it definitely feels like a cultural carryover from a bygone day and age - though I wouldn't really consider it a huge drawback either... Sharing Enter between the numpad and the main cluster is a strange idea, but I guess it works okay (for the numpad).

So the real drawbacks (or what I consider as such) are the thin Enter, far too far out to the right... Backspace feels too "far out" from the home position, too. No (*) and* (/) on the numpad, even though they would beautifully fit in the place that Scroll Lock (the most useless thing on a keyboard ever) is hogging. Also, no left alt.
* - EDIT: Yea just realized there IS a (*) key, I've reconfigured that as Fn so it kinda slipped my mind. Anyway...

So what I'd consider as an "XT inspired modern layout" would probably be ISO based (to maintain the general look and feel, while fixing the Enter key), with Alt and Cmd keys on both sides of the spacebar, a modern numpad, and maybe keeping the function keys on the side. It might make sense to consider mirroring the keyboard, with the numpad on the left hand - someone recently pointed out in another thread how the right hand side numpad and nav cluster is a throwback before the mouse became ubiquitous - probably a right hand mouse, left hand numpad arrangement would be pretty good for quick data entry...

Moral, I really like it for small parts and am glad to have learned it. Would certainly use again, and is great for sharing because it is literally all a C-based language (all the math is fun). It is definitely a toy product when compared proper ones though, just the nature of meshes unfortunately.

Yea, I kinda agree. I've never gotten to the point of complexity where rendering became a major issue, but it definitely gets sluggish as you add stuff... Te main reason I gave up on it was that something like edge rounding or adding a bevel increased the complexity of the whole script by several orders of magnitude... also no parametric surfaces iirc, and extrusion was kinda limited. But I do absolutely love the programming approach it takes.

These days I'm learning Fusion 360 for my 3d printing needs, it is a deep dive, but feels pretty awesome.

[drevyek]

Honestly, it sounds as if the perfect board for you is something like the Model F AT, or Apple's M0116. The AT can be easily modded to use ANSI or ISo layouts, and already has almost your entire wishlist of features. I've even seen a mod done with it to give it the standard 7u spacebar + Alt keys on the bottom row. AFAIK, the PCB has pads for those positions already, so no real modding is needed there, plus it accepts the xwhatsit, so you can have a USB cap-sense controller from the get go, as opposed to the Soarer's Converter workaround for the XT.

Here it is: https://geekhack.org/index.php?topic=48288.0

You can get an AT quickly for ~$100, or cheaper if you're very patient or know a guy (depending on location).

[Phenix]

Where? Im in Germany, and didnt found many AT in the last year (beside 'schlitzohr jessy' - way to dubious for my liking to buy from her, one sold in the FS and one on computerbase)

[drevyek]

Don't know about in Germany, but I've seen a few AT's up for sale, for relatively cheaply, in both Canada and the US. But, they were around $125, which may be closer to the going rate for them. I know that the AT was much less commonly distributed than was the XT, and only had a short production cycle before being replaced by the Model M.

[Sigmoid]

I got both my xt's in the US, and have done some ebay-scourings since then too... The AT Model Fs are an order of magnitude rarer than the xt, anywhere, and the ones I've seen available had either been fixed up and sold at a massive premium by a scalper, or were missing keys and looked like they had been harvested for barrels and keycaps by said scalpers...
In Europe I have had issues finding *any* model F's at all.

But sure, the AT Model F layout is pretty sweet. Their endangered status only shows how much this project is needed.

[Wingklip]

Lot_lizard I have some special ideas that might interest you in the way of cherry compatible buckling springs and beam springs in the same profile of the cherry switch.

One of our geekhack members directed me here, I wonder if you're interested?

[Techno Trousers]

Welcome to the party, Wingklip!

[andrewjoy]

I got both my xt's in the US, and have done some ebay-scourings since then too... The AT Model Fs are an order of magnitude rarer than the xt, anywhere, and the ones I've seen available had either been fixed up and sold at a massive premium by a scalper, or were missing keys and looked like they had been harvested for barrels and keycaps by said scalpers...
In Europe I have had issues finding *any* model F's at all.

But sure, the AT Model F layout is pretty sweet. Their endangered status only shows how much this project is needed.

I have an AT just sitting in my keyabord storage pile doing nothing thats in europe. And an XT , and 1 F122 ( other one is in use ) and a 107 , and a spare case PCB and plate for an XT .

So there is plenty of model F in europe, its just all in my house

[Wingklip]

Welcome to the party, Wingklip!

Tastes like paddle plop springs

[lot_lizard]

Lot_lizard I have some special ideas that might interest you in the way of cherry compatible buckling springs and beam springs in the same profile of the cherry switch.

One of our geekhack members directed me here, I wonder if you're interested?

We are always interested in alternative ideas that have merit. No need to ask... POST away!!!

We have no secrets here in regards to intellectual property. One of the reasons I haven't posted too much yet is that it is all still being confirmed (avoiding others improving bad designs versus finals). If you are more comfortable PMing me, feel free. As Techno said... welcome

[Wingklip]

I've been a little uncomfortable with sharing it with anyone due to the little debacle with the Binge kecap maker. I really need to find a way to quickly patent my ideas and make them too before sharing them online

[codemonkeymike]

If you are looking to make money from your idea @wingklip then you are going to have to do the leg work of prototyping, tooling, and manufacturing. Patents are for having exclusive rights on a idea so you can monitize it, thus is the rub. Unless you do have prototypes and tooling, but then you wouldn't need to ask for permission to help out you could just manufacturer a product.

[Wingklip]

Alright, I'll share one thing, and that is the proposal of using 2 springs for the buckling mechanism instead of one. The use of smaller diameter springs allows much lower profile at which it buckles and yet also the same weight as a conventional bs.

I'll leave it to you guys to prototype a cherry switch. please donate me some coin XD

[Sigmoid]

I think your direction and the purpose of this thread are only tangential. The goal here is to recreate the IBM Model F mechanism exactly or at least very closely - it has the advantage of having been meticulously documented (all the expired patents to go by), extremely well proven (years and years of use), and well loved by many.

I think you are definitely on to something here, with a buckling spring MX switch (if I understand you correctly, that's your goal). If you don't want to go the open route, maybe it's worth getting a team together, prototyping and then patenting it. I'm sure you'll have a good market for it if you can get to production - maybe crowd funding would work well for something like this.

[DarKou]

Really interesting thread !

[fohat]

The goal here is to recreate the IBM Model F mechanism exactly or at least very closely

I have maintained for years that the "switch" (not even really applicable to these mechanisms) is not the component that makes the Model F so spectacular, otherwise the Model M would be extremely similar in feel, which it isn't.

There is no doubt in my mind that it is the synergy of forces, tension and compression, between the various plates (mostly steel) that combine to make the entire assembly, as a unified whole, feel so "alive" and that this will never even remotely be replicated by attaching a bunch of tiny plastic boxes to a plastic circuit board.

[Wingklip]

The goal here is to recreate the IBM Model F mechanism exactly or at least very closely

I have maintained for years that the "switch" (not even really applicable to these mechanisms) is not the component that makes the Model F so spectacular, otherwise the Model M would be extremely similar in feel, which it isn't.

There is no doubt in my mind that it is the synergy of forces, tension and compression, between the various plates (mostly steel) that combine to make the entire assembly, as a unified whole, feel so "alive" and that this will never even remotely be replicated by attaching a bunch of tiny plastic boxes to a plastic circuit board.

But of course, the designer of the plastic circuit board may very well choose to make it with a metal backplate. I say you are quite wrong in the switch not comprising a major role - the model F has a slamming flipper whilst the model m has a tiny plop when the flipper hits. Of course, the spring vibration is key.

The only way we can bring this out of the dark ages of the eighties is to create a modular form in the clone shell of a popular switch type

[Sigmoid]

I have maintained for years that the "switch" (not even really applicable to these mechanisms) is not the component that makes the Model F so spectacular, otherwise the Model M would be extremely similar in feel, which it isn't.

There is no doubt in my mind that it is the synergy of forces, tension and compression, between the various plates (mostly steel) that combine to make the entire assembly, as a unified whole, feel so "alive" and that this will never even remotely be replicated by attaching a bunch of tiny plastic boxes to a plastic circuit board.

I'd say it probably has a whole lot to do with the spring tension (higher than in Model M springs), and the flipper (which is a big flappy thing as opposed to the teeny little plinker on the Model M).

I'm sure that the structure of the keyboards plays SOME role, let's not forget that the tension in a Model F is between the metal plates (front and back), the foam padding, the PCB and the barrels. The actual moving parts are actually free from any tension. I suspect the role all this stuff is playing is merely acoustic.

And even if it isn't, as far as I understand, @lot_lizard's project actually involves steel plates, much like the original F, so it should get pretty close in feel and sound.

[DMA]

I have maintained for years that the "switch" (not even really applicable to these mechanisms) is not the component that makes the Model F so spectacular, otherwise the Model M would be extremely similar in feel, which it isn't.

I'm sure that the structure of the keyboards plays SOME role

Fortunately that's relatively easy to check - there are several FSSKs in existence. Model M subassembly, model F flippers and springs.. M bolt mod effects should also be described somewhere around here.. No foam though - but with all those bolts you can tune it to be as close to F as possible and the difference would be what foam brings

One just needs a happy owner of all 3 with enough free time and a gopro (or equivalent)

[micrex22]

I have maintained for years that the "switch" (not even really applicable to these mechanisms) is not the component that makes the Model F so spectacular, otherwise the Model M would be extremely similar in feel, which it isn't.

There is no doubt in my mind that it is the synergy of forces, tension and compression, between the various plates (mostly steel) that combine to make the entire assembly, as a unified whole, feel so "alive" and that this will never even remotely be replicated by attaching a bunch of tiny plastic boxes to a plastic circuit board.

The Model M uses a stiffer spring, smaller flipper and also has a rubber mat--all of those components will dramatically change the feeling.

When Model F flippies are put in Model M assemblies such as the FSSK they feel (and sound) identical to an F XT *with enough tension*. Of course even with a bolt mod you have to be mindful of certain 'hotspot' areas and especially cracking on the smaller SSK.

[Techno Trousers]

For what it's worth, to me one of the best things about the Model F feel is the long key travel before activation. When I'm up to full speed touch typing, there's no other switch that comes close in terms of tactile feedback.

I don't know if that feel can be replicated in an enclosed, Cherry-sized package, but it sounds like an interesting experiment, if nothing else.

[Wingklip]

I will attempt a 2x scale print of a cherry switch and try a bspring in there

[Dynomite54]

Wow this is a great idea. Will you be adding rubber feet?

[lot_lizard]

Wow this is a great idea. Will you be adding rubber feet?

It's funny you say that. I have always thought that Wodan's HHKB feet were a clever idea (in this thread). Maybe he would be interested. It should be a follow on exercise though as a independent GB at this point to MF phase 1. Our scope is not small already, and I want to see people get everything as soon as we can. Make a post in Wodan's thread there and see if we can get some "traction" (bad pun)

[emdude]

Slip-on rubber feet would certainly be nice for the future MF cases.

[Wingklip]

http://www.rccoilspring.com/rectangular-springs.html
Found this that could work for low profile bucklesprings