deskthority

Does any of you know if there is a buckling spring switch that can be used in a diy keyboard design? I like the way a Cherry MX switch can easily be incorporated in your own designs, but I like the clickiness of a buckling spring switch. I have thought about cutting up a Model M keyboard, but there are too many issues with the plate and the membrane. Maybe it is possible to 3d-print a buckling spring switch?

AFAIK the buckling spring mechanism is still patented, so nobody but Unicomp can actually make them, and they don't make single switch modules.

If it wasn't for the difficulty to get custom membrane sheets, I would have already attempted to build one with 7bits symmetric stagger
From first sight taking apart my Model M, I thought re-using keycaps, springs and hammers would be quite possible, I just haven't gotten around to experimenting with the contact sheets yet. If you have a simple idea how to customize them share it!

wcass on geekhack seems to have found a way to make a custom membrane: http://geekhack.org/index.php?topic=40111
Now all that's left is making a custom barrel plate.

I'd really like to make a compact buckling spring keyboard, ideally with as much freedom in layout as discrete key switches provide.

Some way to adapt harvested buckling spring barrels into discrete switches would be revolutionary.

Some work is happening on discrete and modular capacitive buckling spring assemblies that one could easily make customs with, though it's on a backburner a bit with a priority for a generic controller first. But if you are or know of a materials engineer, shoot me a PM.

Also the patent has long expired.

Le's start a group buy and ask Unicomp to manufacture it for us.


I want one to fit the 122-key terminal keyboard shells.

Parak wrote:Some work is happening on discrete and modular capacitive buckling spring assemblies that one could easily make customs with, though it's on a backburner a bit with a priority for a generic controller first. But if you are or know of a materials engineer, shoot me a PM.

Also the patent has long expired.

Good to hear it.

What's different about a buckling spring controller, compared to the usual kind of thing done with Teensies?

I understand that capacitative (Model F style) switches are a whole different game, controller-wise, than membrane (Model M style) switches. And I'm yet to see a custom controller mod in a Topre (another capacitative switch technology); though if anyone's got a link…

7bit wrote:Let's start a group buy and ask Unicomp to manufacture it for us.

I get the impression that Unicomp doesn't feel up to designing or manufacturing anything different. They own the Model M machine and simply turn the crank. Otherwise, surely, they'd have tenkeyless at least!

Muirium wrote:
What's different about a buckling spring controller, compared to the usual kind of thing done with Teensies?

I understand that capacitative (Model F style) switches are a whole different game, controller-wise, than membrane (Model M style) switches. And I'm yet to see a custom controller mod in a Topre (another capacitative switch technology); though if anyone's got a link…

Yes, they are fundamentally different. Briefly put, a regular controller for typical switch like membrane, cherry, etc, strobes the switch matrix, detects which junctions are closed, and outputs relevant key presses based on that. A capacitive controller strobes the matrix, and reads the result via sense lines, where a substantial (usually predefined depending on a number of factors) change in capacitance represents a key press. One of the many advantages of the latter, for example, is that there's no need for diodes to get key rollover as changes are cumulative.

Also notably, such a capacitive matrix is a bit different from the way most capacitive sensitive devices work currently.

Hasu did a replacement controller for the HHKB here, but it leverages and interfaces with the existing capsense IC on the main pcb, so it's not a full on replacement. For the F type existing and future PCBs, the entire thing needs to be built from scratch, although it will still use a teensy or other micro as the core.

Instead of a PCB, I like the flexibility of the hand wired technique:

The custom project I'm working on with Matteo just now is another hand wired matrix. Just plain old Cherry MX switches, though.

I wonder what the chances are of building a discrete buckling spring switch mechanism based on simple electrical contact – the keyswitch equivalent of the Model M's membrane – because that seems to be the ideal modding recipe for really building your own. PCBs have their advantages of course, but ultimate customisation is where a hand wired matrix shines.

Muirium wrote: I wonder what the chances are of building a discrete buckling spring switch mechanism based on simple electrical contact – the keyswitch equivalent of the Model M's membrane – because that seems to be the ideal modding recipe for really building your own. PCBs have their advantages of course, but ultimate customisation is where a hand wired matrix shines.

Right, that's the goal, except I'd prefer to do it via capacitive sensing. One could certainly use a tiny membrane in a discrete module instead with pin terminations that hook into wire or pcb matrix.

I agree that capacitative is more appealing overall. Capacitative switches in a hand wired matrix sounds a great recipe for custom builds.

I may be wrong, but the increasing ubiquity of capacitative touch screens (like the one I write this on now) may mean the components of the technology are working their way into reach. Naturally, there's a difference between sensing fingers on a grid and the buckling of a spring on a switch plate, but tracking matrices of inductive values is more popular now than ever.

7bit wrote:Le's start a group buy and ask Unicomp to manufacture it for us.

I was just thinking the same thing. What can I say; GMTA.

But in case they don't want to, we could try to design and 3d print something ourselves, maybe even start a kickstarter...

The investment to make these parts is significant, as well as the R&D required. There's no way that Unicomp will do it - perhaps only if they ever get a lot more business than they do now.

There was a project to change the layout of the Model F by adding a new switch plate, circuit board and controller ... but it failed. I think it failed because they couldn't figure out how to do the capacitative sensing.

I once tried to make a IBM Model M2 into a tenkeyless by cutting the membrane and I tried to replace the missing leads with a PCB pen and conductive film, but my mods did not turn out to be very durable ... so I had to scrap the keyboard. I have donated parts from it to people who needed them.
However, I think that a PCB with surface-mounted diodes and a simple membrane with horizontal strips of metal could work.

What about Hall Effect sensing? I know that the striking end of the rocker in a BS key doesn't move all that far, but maybe it could be re-designed to actuate a magnet towards a HE sensor?

Having modular Buckling Spring switches with Cherry MX compatible stems would by my ideal.

It can be pretty much anything - contact based like an M, capacitive like an F, hall effect, optoelectric where it breaks a light beam, etc. Designing anything that hasn't been done before is an entirely different proposition of course. Wcass on geekhack did it via a membrane + pcb, and I (plus a few people that actually know this stuff unlike myself) have been focused on the replication of the capacitive design as it's a proven concept with at least the keycaps readily available from Unicomp. People are free to try and make a BS to MX converter of course, like they did for Alps.

you can make a custom BS keyboard using two steel plates - the top one with holes where the barrels should go. use an XT as a barrel donor (and maybe the case too). you can use model M hammers and springs with custom membranes.

i have wanted to order from this vendor in the UK to check it out, but shipping alone to the US is £11 so have been working other ideas first. you guys in the UK might try it out:
http://inkjetflex.com/site/rapid-prototyping/

i just picked up a rework station and some diodes and will be fitting diodes onto a PCB that i am using for a bottom membrane. the controller is a Teensy.

Now we're talking!

I'm in the UK. What are you after?

would you be interested in being a proxy? how much is postage from UK to US for an A4 size letter/envelope?

Glad to help out. What are thinking of sending, exactly? If it's light, it should be pretty cheap. The Royal Mail has post offices all over Britain and likes to quote everything face to face, letter on scale. But the vague, weasely website is here:

http://www.postoffice.co.uk/airmail

for about £30 i could order two top membrane sheets - 300mm x 150mm x 0.1mm delivered to a UK address. actually i should just suck it up and have them deliver to me directly.

The membrane switch is inexpensive and easy to assemble for large volume production -- but setup and tooling are expensive. Have you thought about just using discrete switches on a flexible (thin) circuit board? The switches could be a thin strip of beryllium-copper -- perhaps harvested from a scrapped switch ( e.g. a Cherry MY) -- soldered directly to the board. Alternatively, perhaps a very light-touch switch assembly:
http://www.panasonic.com/industrial/inc ... Switch.pdf
However, I don't know the dimensions under the hammer -- the clearance and throw -- or the actuating force available.

clearance under the hammer in the "up" possition is very small - maybe 1 mm. the micro switch that you reference might work, and only add about 2mm to switch height. there would even be enough room for a diode. this might turn an XT barrel into a discrete switch like a Cherry. but at £1 (or .6 in bulk) for each microswitch, even custom membranes are a better deal.

What about adding a metal strip along one edge of the barrels, and wiring up the bottom of the spring? That way when the spring buckles - it contacts the metal strip and closes the circuit / actuates the switch.

(I too have been wondering about a method to plate mount individual BS switches using the hard wired method.)

Interestingly, that was the original contact mechanism of buckling springs: http://deskthority.net/wiki/Buckling_Sp ... IBM_Patent

It might be hard to deal with bouncing though, as the foot can actually be more stable once the spring puts enough pressure on it. It might also get fiddly to wire up all springs.