Individual Buckling Spring Keyswitches? BS ErgoDox?

[krfkeith]

Despite trying a wide range of key switches (sadly, I've yet to try hall effect) nothing comes close, for me, to the feel of the good old buckling spring mechanism. I love the feel of the venerable Model M, but am not really a fan of the layout.  What I really would like is a buckling-spring based ErgoDox. Is there any technical reason why BS keyswitches could not be made individually like cherries?

Just a thought.

[Jersern]

I'd say it is possible since there isn't much too buckling switch keyboard and I think there are are knock-offs around so I'd say if someone really tried, that would be possible. A buckling spring keyboard is really just the spring and membrane. Now the question is will we get sued? (Yeah, probably)

[AKmalamute]

The patent expired in the late nineties so no suing anybody. It's just a cost / benefit analysis at this point...and BS is one of the costlier designs especially at low volumes.

if it happens though, I'm in.

[FrostyToast]

The problem is that you still need sophisticated tooling to get the barrel plate from what I heard.
However, doing a custom BS board would be cool.

[tbc]

So so so so in

[FrostyToast]

I guess you could take what would essentially be the cross section of a model m with the barrel, spring, and foam, and then maybe have it connected to an enabler type thing...

[Mandolin]

The problem is that you still need sophisticated tooling to get the barrel plate from what I heard.
However, doing a custom BS board would be cool.

Now that you mention it, there's no need to cnc milled metallic extrusion molds+ 50k equipment.

There are low technologic approaches to prototyping if you got a donor keyboard. To make a custom barrel plate you'd take appart single key sections cuts, order them like legos (copy, paste) according to the config of the final kbd, in respect of size, and angulation, patch the spaces between them and adapt the connection to the frame, or metallic plate of destiny with clay.

With the master prototype, a master mold could be made and then casted in plastic.

I don't know how to adapt the electronics of the switches to be able to register the key presses. A custom matrix and the soarer adapter could do.

I'm no legal expert, but as of now model M Patents are property of unicomp or IBM maybe, so if someone were to do this, it'd be for personal use only, at most.

[0100010]

You would need a donor Model F to harvest the barrels / hammers / springs and wcass to draw you up a PCB and barrel plate with integrated xwhatsit controllers.

[Puddsy]

You'd need to custom build a PCB, and you'd need to custom design a membrane.

It's "possible" but impractical.

[NeoNoeN]

Had a thought of this a couple of days ago and currently looking into model F donor boards for barrels, I'm thinking of finding the specifications for model M membranes to plate so I can figure out how to place the membrane on a PCB. Overlaying a membrane on a PCB would also allow for nkro (as well as be easier to design).

[FrostyToast]

Had a thought of this a couple of days ago and currently looking into model F donor boards for barrels, I'm thinking of finding the specifications for model M membranes to plate so I can figure out how to place the membrane on a PCB. Overlaying a membrane on a PCB would also allow for nkro (as well as be easier to design).

Again, the most concerning point is that we would have to effectively harvest model F barrels and find a way to put them together again.
This would not only be difficult, but expensive as heck.

[jacobolus]

There’s no reason to make custom boards with Model M membrane technology. Doing a custom Model F is a pretty similar level of difficulty from a manufacturing/design perspective, but a superior product.

What you need:
(1) XT keyboard as a donor for barrels/flippies and possibly keycaps
(2) Laser or waterjet cut steel plate, rolled on a slip roll
(3) Capacitive PCB (this is probably the hardest part, design-wise)
(4) Capsense controller (a couple different types of these have already been figured out by people in the community, and some can be bought already assembled)
(5) Some kind of case

[Snowdog993]

I wonder if you looked at these things....
http://www.shapeways.com/model/1180047/ibm-model-f-barrel-flat-quot-parak-quot-v2.html?materialId=75
I don't exactly know if that's what you're looking for, but it sure is expensive!

[NeoNoeN]

There�s no reason to make custom boards with Model M membrane technology. Doing a custom Model F is a pretty similar level of difficulty from a manufacturing/design perspective, but a superior product.

What you need:
(1) XT keyboard as a donor for barrels/flippies and possibly keycaps
(2) Laser or waterjet cut steel plate, rolled on a slip roll
(3) Capacitive PCB (this is probably the hardest part, design-wise)
(4) Capsense controller (a couple different types of these have already been figured out by people in the community, and some can be bought already assembled)
(5) Some kind of case

The reasoning behind using a custom membrane is because it should be easier than producing a capacitive PCB, using the membrane wouldn't have any effect on the functionality of the keyboard and make it much simpler to produce.

[jacobolus]

The reasoning behind using a custom membrane is because it should be easier than producing a capacitive PCB, using the membrane wouldn't have any effect on the functionality of the keyboard and make it much simpler to produce.

Do you have a source who will make you custom-layout membranes? The capacitive PCB takes some design work, but wcass has already done one custom Model F. As a PCB, there’s nothing too special about it: any PCB fab could make it. The tricky part is the PCB design, for which you want to have an EE on board for advice/help. I’m sure there are at least several people on this forum who could help with the PCB design (or could be commissioned to do all the design work) if you have a specific layout plan.

[NeoNoeN]

The reasoning behind using a custom membrane is because it should be easier than producing a capacitive PCB, using the membrane wouldn't have any effect on the functionality of the keyboard and make it much simpler to produce.

Do you have a source who will make you custom-layout membranes? The capacitive PCB takes some design work, but wcass has already done one custom Model F. As a PCB, there’s nothing too special about it: any PCB fab could make it. The tricky part is the PCB design, for which you want to have an EE on board for advice/help. I’m sure there are at least several people on this forum who could help with the PCB design (or could be commissioned to do all the design work) if you have a specific layout plan.

Ah okay, makes sense. I always assumed that making a custom PCB for a model F would be more difficult than producing a custom membrane, thanks for pointing me in the right direction I'll start working on something soon.

[berserkfan]

Ah okay, makes sense. I always assumed that making a custom PCB for a model F would be more difficult than producing a custom membrane, thanks for pointing me in the right direction I'll start working on something soon.

I enthusiastically support you!

It is my great hope to get a split layout Model F and I will definitely buy what you produce (ie help lower overall costs) as long as you don't have ergodox-style gaps for the thumbs which are too big for my hands.

[vivalarevolución]

i cringe when we talk about using Model F parts are donor boards.  Why not just figure out a way to manufacture the barrels, flippy plates and find some usable stock springs?  Go all the way, I say.

[NeoNoeN]

i cringe when we talk about using Model F parts are donor boards.  Why not just figure out a way to manufacture the barrels, flippy plates and find some usable stock springs?  Go all the way, I say.

That is very, very expensive. If we were to attempt that we would have to go to group buy and the MOQ would be extremely high. The most realistic thing we could do in terms of producing individual keyswitches is a standalone plate + barrel for a BS switch, could then use some of unicomp's spare hammer+spring and keys as the rest of the parts. I mean depending on how enthusiastic geekhack/mechanical keyboards/DT are we could do a massive inter-site group buy but that is the only way I could see the MOQ being achieved for producing custom parts such as these.

[krfkeith]

i cringe when we talk about using Model F parts are donor boards.  Why not just figure out a way to manufacture the barrels, flippy plates and find some usable stock springs?  Go all the way, I say.

That is very, very expensive. If we were to attempt that we would have to go to group buy and the MOQ would be extremely high. The most realistic thing we could do in terms of producing individual keyswitches is a standalone plate + barrel for a BS switch, could then use some of unicomp's spare hammer+spring and keys as the rest of the parts. I mean depending on how enthusiastic geekhack/mechanical keyboards/DT are we could do a massive inter-site group buy but that is the only way I could see the MOQ being achieved for producing custom parts such as these.

Hmmm, I would totally be down for this if we could make it happen. There's a 3D scanned barrel shapeways that another poster linked to. Do you think it would be possible to polish a 3D print of one of these to get something optimal?

[jacobolus]

Hmmm, I would totally be down for this if we could make it happen. There's a 3D scanned barrel shapeways that another poster linked to. Do you think it would be possible to polish a 3D print of one of these to get something optimal?

IBM was able to produce parts to very precise specifications. From what I’ve heard, getting buckling springs to work well and consistently is quite challenging. I don’t think 3d prints are going to be sufficient for this at reasonable cost in the near term, but you’re certainly welcome to try.

Making new tooling for barrels and flip plates would cost very large amounts of money, we’d need an up front commitment from hundreds of people to buy a very expensive product, with no certainty of success.

It’s a hell of a lot easier and cheaper to just use old XT boards while they’re still relatively easily obtainable.

[Mandolin]

Hmmm, I would totally be down for this if we could make it happen. There's a 3D scanned barrel shapeways that another poster linked to. Do you think it would be possible to polish a 3D print of one of these to get something optimal?

IBM was able to produce parts to very precise specifications. From what I’ve heard, getting buckling springs to work well and consistently is quite challenging. I don’t think 3d prints are going to be sufficient for this at reasonable cost in the near term, but you’re certainly welcome to try.

Making new tooling for barrels and flip plates would cost very large amounts of money, we’d need an up front commitment from hundreds of people to buy a very expensive product, with no certainty of success.

It’s a hell of a lot easier and cheaper to just use old XT boards while they’re still relatively easily obtainable.

I already posted a solution to do custom barrel plates that would cost much less than 1000 bucks with tolerances of 0.05 or 0.5% of dimensional variation. The precision of the barrel plate matter the most on the key cilinder section, specially inside it where it interacts with the keycap, spring, etc. The rest doesn't matter that much. A urethane rubber mold could yield 400 copies, but it could be made with silicone for long term use instead. An original donor "to trash" or a very high end 3d print would be necessary.

[jacobolus]

I already posted a solution

Okay. Well if you go get it working, I’m sure people will be interested. So get to it!

[NeoNoeN]

Hmmm, I would totally be down for this if we could make it happen. There's a 3D scanned barrel shapeways that another poster linked to. Do you think it would be possible to polish a 3D print of one of these to get something optimal?

IBM was able to produce parts to very precise specifications. From what I’ve heard, getting buckling springs to work well and consistently is quite challenging. I don’t think 3d prints are going to be sufficient for this at reasonable cost in the near term, but you’re certainly welcome to try.

Making new tooling for barrels and flip plates would cost very large amounts of money, we’d need an up front commitment from hundreds of people to buy a very expensive product, with no certainty of success.

It’s a hell of a lot easier and cheaper to just use old XT boards while they’re still relatively easily obtainable.

I already posted a solution to do custom barrel plates that would cost much less than 1000 bucks with tolerances of 0.05 or 0.5% of dimensional variation. The precision of the barrel plate matter the most on the key cilinder section, specially inside it where it interacts with the keycap, spring, etc. The rest doesn't matter that much. A urethane rubber mold could yield 400 copies, but it could be made with silicone for long term use instead. An original donor "to trash" or a very high end 3d print would be necessary.

I think creating individual barrels like that seen on the model F would be a better idea as it allowed people to create their own custom layouts. If you took it even further and integrated a mini capacitive PCB with pins you could easily sell a lot of these.

How would we make capacitive hammers without harvesting Model F keyboards? Could we custom mold metal hammers and just make them non-capacitive switches? We could technically make these cherry plate compatible.

[Mandolin]

I think creating individual barrels like that seen on the model F would be a better idea as it allowed people to create their own custom layouts. If you took it even further and integrated a mini capacitive PCB with pins you could easily sell a lot of these.

How would we make capacitive hammers without harvesting Model F keyboards? Could we custom mold metal hammers and just make them non-capacitive switches? We could technically make these cherry plate compatible.

Resin+metallic filler = model f hammer.
(It would need some tuning in respect of the amount used and be of a specific type. I'd need to test an model F hammer's capacitance and then use a filler to give me the same result. That's it.)

Instead of a barrel plate, I'd go for barrel modules.

[NeoNoeN]

Hasn't the patent for buckling springs expired?

[Mandolin]

Model F BS main patent is active and there are more involved.
Derivative works couldn't be made without IBM permission for each one:
http://www.google.com/patents/US4118611

I like my liberty, so, that's a strong no 

[krfkeith]

Model F BS main patent is active and there are more involved.
Derivative works couldn't be made without IBM permission for each one:
http://www.google.com/patents/US4118611

I like my liberty, so, that's a strong no 

How is that even possible? The Model F pre-dates the Model M, and I know for a fact the Model M patent(s) has expired.

[Mandolin]

How is that even possible? The Model F pre-dates the Model M, and I know for a fact the Model M patent(s) has expired.

Don't we have any lawyers in the forum? I'm not sure.

[Snowdog993]

I wonder if you looked at these things....
http://www.shapeways.com/model/1180047/ibm-model-f-barrel-flat-quot-parak-quot-v2.html?materialId=75
I don't exactly know if that's what you're looking for, but it sure is expensive!

I am quoting myself!  Okay look again....   Can be used with a normal Model F Pivot Plate (Flippy) and a Stem from a Model F or M. That means you can put type M buckling springs in that assembly! 

[NeoNoeN]

I was pretty sure that those only accept model F hammers which are more expensive to produce, plus 3D printing isn't the way to go if we mass produce, I looked into the model F 3D models on shapeways a while back and while it is good for prototyping it isn't viable for mass production.

[Snowdog993]

Look at that picture one more time...  let me hint.  Do a 180 on it.

[Mandolin]

If there isn't a legal constraint, I could do some custom fitted barrel plates or modules + hammers, but I've no means of mass producing them, and until then, it would be expensive, so just think in a switch tester.

To lower down the costs they could be spin or pressure casted.

[jacobolus]

There is absolutely no legal constraint. For patents from before 1995, the patent term is either 17 years from the issue date or 20 years from the filing date of the first related patent, whichever is longer. The Model F dates from 1981 and the model M dates from 1983. All of the relevant patents expired at least 10 years ago. US patent 4118611 which you linked before expired in 1997.

Unicomp might tell you they’re the only ones with rights to make buckling spring keyboards, trackpoints, &c., but they have no legal leg to stand on.

There may be trademark on the name “Model F” / “Model M”. I’d recommend any commercial product pick a different name.

[NeoNoeN]

If there isn't a legal constrain, I could do some custom fitted barrel plates or modules + hammers, but I've no means of mass producing them, and until then, it would be expensive, so just think in a switch tester.

To lower down the costs they could be spin or pressure casted.

I'm thinking I'm going to experiment with materials using shapeways first to check out which materials will be better. If we do go to mass production we would outsource it to China or something. However if you just feel like making one for yourself to try go ahead, but at our current conceptual stage it isn't really needed.

[Mandolin]

I thought the guy who filed the patent should've been dead for 50 years or something like that  ). US patent laws are not my thing.

I'm thinking I'm going to experiment with materials using shapeways first to check out which materials will be better. If we do go to mass production we would outsource it to China or something. However if you just feel like making one for yourself to try go ahead, but at our current conceptual stage it isn't really needed.

That's the best route!
If you get to do a revival model F, numpad or kit, I'm in for one.

Good luck 

[jacobolus]

I thought the guy who filed the patent should've been dead for 50 years or something like that  ). US patent laws are not my thing.

That’s copyright. (And it’s now death of the author + 70 years, or where that’s not applicable,  whichever is shorter 95 years from publication or 120 years from creation. Copyright term is utterly stupid. Patent term is IMO a bit too long and should vary from one field to another, but it’s not nearly so bad as copyright.)

[NeoNoeN]

Should we start up a making stuff together thread and git for this project? I don't want to be doing this by myself as I have my own educational obligations to attend to, but I can manage this project. If no one is interested in helping out I can start doing everything November next year ( This project is going to take a while regardless).

[jacobolus]

Should we start up a making stuff together thread and git for this project?

You can if you want, but I’d recommend finding someone who is actually willing to commit to do the work first. Otherwise I doubt anything will come of it.

[dorkvader]

making model F flip plates wouldn't be that hard in quantity.

If you put together a few thou in tooling, you'd only ave to do the prototyping to ensure an even mix of conductive material and other factors and then go to town on production. I think handcasting them will take too long.

THe other issue is the barrel: it must be relatively precise. Parts of the round barrels are not "used" and can clip into a plate like a cherry switch does. This would stabilize the switch quite well without the need for a mat

Though with bolt mods designed into the build and a flat plate it would be pretty easy to have a mat there.

Another option is to just cast them out of metal. Tin would probably work. Whatever's cheapest.

Another option is to make them out of nonconductive plastic and then tape or glue a conductive strip to the end. I'm not sure how this would work in production,  but if making a kit, letting the user put tape on 100 or so flip plates is pretty easy I think.

Springs: quite easy. You can get custom made springs in high quality very cheap in quantity. Even 10 keyboards worth, though I'd aim for 100 in the first run.

Then we need to call up unicomp or whoever and get them to sell us keycaps.

There are other design ideas (NKRO model M, different material than hard-to-source custom membranes, etc) I have but these pertain to the topic at hand.

Once I have a good solution and/or am done with the hall effect project, I plan to put time into this.

[vivalarevolución]

i cringe when we talk about using Model F parts are donor boards.  Why not just figure out a way to manufacture the barrels, flippy plates and find some usable stock springs?  Go all the way, I say.

That is very, very expensive. If we were to attempt that we would have to go to group buy and the MOQ would be extremely high. The most realistic thing we could do in terms of producing individual keyswitches is a standalone plate + barrel for a BS switch, could then use some of unicomp's spare hammer+spring and keys as the rest of the parts. I mean depending on how enthusiastic geekhack/mechanical keyboards/DT are we could do a massive inter-site group buy but that is the only way I could see the MOQ being achieved for producing custom parts such as these.

I never tried to use the Model M hammer and spring for a Model F.  I'm sure the Model M spring is fine, but I doubt the hammer will work smoothly.  I could just open up one of my Model F's, throw in some M springs, go to town, and report back.

[vivalarevolución]

There is absolutely no legal constraint. For patents from before 1995, the patent term is either 17 years from the issue date or 20 years from the filing date of the first related patent, whichever is longer. The Model F dates from 1981 and the model M dates from 1983. All of the relevant patents expired at least 10 years ago. US patent 4118611 which you linked before expired in 1997.

Unicomp might tell you they’re the only ones with rights to make buckling spring keyboards, trackpoints, &c., but they have no legal leg to stand on.

There may be trademark on the name “Model F” / “Model M”. I’d recommend any commercial product pick a different name.

I don't know squat about patent law, but would it matter if these patents were being violated outside the United States?

[berserkfan]

How is that even possible? The Model F pre-dates the Model M, and I know for a fact the Model M patent(s) has expired.

Don't we have any lawyers in the forum? I'm not sure.

The laws are clear on this; 20 years for patents granted in the USA before 1995. There has been little dispute or challenge on the time.

I remember reading that WCass' problem was, IBM's original patent didn't actually specify the formula for the capacitative content for the plastics. This is a trick which works very well in practice. EG Coca Cola has never been patented, but no one knows exactly how they make coke. So we don't know how to make the barrel and probably some one will have to do 1000 prototypes until he chances on the perfect formula for the plastic.

[Snowdog993]

What about fabricating (or fabrication) from aluminum?  I would think that might be cost effective?  This would be only to test the design and thickness of the barrels etc....  Then fabricate with ABS (most likely)  I'm no expert.

[NeoNoeN]

If we're going to do this should we go all the way and make it MX compatible? I know matt3o is working on his own stem but maybe with his permission we could incorporate something like it? If we're going with stems maybe we should also be looking at changing some stuff up so we can use costar and cherry stabs? I'm thinking that if we are going to make a modular bucking spring we might as well go all out and standardize/modernize it.

Another thought is that we could modify the housing so that they can clip into plates of the same dimensions of MX compatible switches, effectively making these switches interchangeable with ordinary cherry MX switches but with the feel of a bucking spring.

[jacobolus]

LOLOL. “Modernize” = make IBM’s keyswitch design from 1981 compatible with Cherry’s keycap mount from 1980 (I think, or maybe slightly before).

[NeoNoeN]

LOLOL. “Modernize” = make IBM’s keyswitch design from 1981 compatible with Cherry’s keycap mount from 1980 (I think, or maybe slightly before).

Standardize then, didn't really word it correctly.

[daerid]

A cherry mx compatible buckling spring ergodox?

[krfkeith]

Hi again!

So I've had an idea I've been thinking about for awhile. IIRC, making capacitive hammers is quite difficult, because, as someone else here noted, IBM never shared the exact formula: it was a trade secret (similar to Coke). I think getting capacitive to work is going to be very difficult. However, I do agree that capacitive sensing is superior to Model M style membrane sensing. My idea is this: why not use hall effect sensing? I know! Sounds crazy! But here me out:

First of all, tiny magnets can be very cheap. In fact, the page I linked to explicitly mentions that "these very small cubes are often used with Hall Effect sensors." Second of all, hall effect sensors can also be rather cheap, I've seen them as cheap as 25 cents a piece individually, or 13-14 cents when ordered in bulk. My idea is that we could use the Unicomp Model M hammers, and just drill little holes into them to embed the magnets.  There is a (rather crude) diagram of what I had in mind (the H represents the hall effect sensor)  included with this message.


If this would work, all we would need for an all-new parts buckling spring keyboard is the barrels. I do have to wonder whether it would be possible to use resin casting to make these. i don't know what kind of tolerances resin can produce, but I can try and do some research if there is interest.

[NeoNoeN]

I was thinking something like this would be easier to produce:

Sorry for napkin drawing.