Buckling spring switch---plate and PCB mount, MX-compatible

[Arcanius]

The past few years have seen a lot of growth in our once-little community---something I didn't imagine would happen when I got my first mechanical keyboard (a Model M) five or six years ago. While the community has grown and our tastes have diversified, we still base all our new keyboard purchases on one or two switch types: Cherry et al or Matias's ALPS switches. I don't think I'm alone when I say I wish I could have a tenkeyless or 60% buckling spring keyboard. So a prototype or three were made.

I've made a modular buckling spring switch with a Cherry MX stem. The activation force is between 9-10 loonies, so around 65cN. Travel is 4.0mm. The switch is taller than a Cherry MX, as you can see, but this is a limitation due to the spring height necessary for buckling.

It will hold a 3mm LED. An RGB version could also be made.

The switch can be opened without taking it out of the plate, since the clips are mounted upside-down from a Cherry design. The sound is a bit muted, and a little higher than a Model M, but this is a bare switch without a plate or housing to resonate in.
Now that a working prototype is working, all that's left is to gauge demand and see how much it might cost to produce a few keyboard's worth of switches!
You can see an extra picture and gif here: https://imgur.com/a/KBHoi
Two close-miked recordings of the switch with a keycap on top. Headphone warning for recording 2 especially.
https://clyp.it/2sy5wnk4
https://clyp.it/0ejokept
Cheers!

[bocahgundul]

Nice project bro 

[CPTBadAss]

I'm extremely interested in this switch. Is it possible to make a typing video or make audio of the switch? How does it feel compared to BS?

[keshley]

This is pretty cool!

So this should solder up to any MX style contact pads?

[ika]

This is really neat. Have you considered mass-production possibilities? A video of the switch in action would be pretty cool.

[Arcanius]

So this should solder up to any MX style contact pads?

The dimensions and pin placement would require a new PCB design, unfortunately. I would have liked to make it match Cherry mount PCBs, but it wasn't feasible.
I am looking at mass-production possibilities, but will have to wait until Monday to get any cost estimates. The higher the volume, the lower the price.

[jaffers]

This is very cool 

Webm is totally awesome https://i.imgur.com/RwTOQmh.webm

Perhaps a metal pad for the spring to hit to make it louder?

[keshley]

So this should solder up to any MX style contact pads?

The dimensions and pin placement would require a new PCB design, unfortunately. I would have liked to make it match Cherry mount PCBs, but it wasn't feasible.
I am looking at mass-production possibilities, but will have to wait until Monday to get any cost estimates. The higher the volume, the lower the price.

Gotcha. I'll keep an eye out then

[LunarisDream]

History in the making

[Melvang]

You are a man after my own heart. 

The wife said she hears another bromance coming.

[nubbinator]

Really cool idea.  I have a couple of questions about it.  Correct me if I'm wrong, but it appears that your contact mechanism is triggered by the spring attached to a metal pole hitting a plate in the back rather than traditional BS method of a flipper compressing a membrane or a capacitive flipper completing the circuit between two pads.  A contact mechanism like that would pretty much require gold plated springs.

I wonder if, instead, it might be possible to make a channel in the channel for the MX style stem and custom make a flipper you could drop in that.  To make it capacitive/conductive, you could epoxy a thin shim or attach conductive foil in place or, if it's produced, use conductive plastic.  Then you could have two pads on the bottom inside that the flipper could contact with.  Something like that would be more expensive, but it seems to me that it would enable you to actually make it a drop in replacement for MX switches.

Here's a rough diagram of what I mean:

[Arcanius]

Then you could have two pads on the bottom inside that the flipper could contact with.  Something like that would be more expensive, but it seems to me that it would enable you to actually make it a drop in replacement for MX switches.

That was an option in the design, but the problem was with height; a custom flipper will increase the height by at least another 3-4mm, which would make the switch very tall to put in any normal enclosure.

Two quick close-miked recordings! They are just of the plain switch with a keycap on top. Headphone warning for recording 2 especially.
https://clyp.it/2sy5wnk4
https://clyp.it/0ejokept

[Melvang]

I concur with trying to stay away from the spring being part of the circuit.  IBM tried that with one of the very early versions but quickly got away from it due to excessive wear.

[nubbinator]

I concur with trying to stay away from the spring being part of the circuit.  IBM tried that with one of the very early versions but quickly got away from it due to excessive wear.

Yeah, that's my concern.  You also run into the problem of potential lack of consistency and corrosion. 

[Arcanius]

Yeah, that's my concern.  You also run into the problem of potential lack of consistency and corrosion. 

The consistency doesn't seem to be a problem so far for this and the two preceding prototypes. I was worried about bouncing times but 5ms debounce is enough, it seems. 10ms would make me feel very at ease.
The wear is something to definitely consider. The spring doesn't slide much up and down the contact plate, nor is it with much force, so the question is how many keypresses it might take to wear through the plate, considering it won't exactly be paper thin.

EDIT: debounce was off by a factor of 10!

[jacobolus]

You might be able to use the spring as the main switch without flippers if you use capsense or optical sensors or something along the wall of the switch housing, but it will make the engineering more difficult. As others have said, making the spring directly close an electrical circuit seems like a bad idea.

[Zekromtor]

What is the travel, and how far does it need to be depressed before it triggers?

My concern is corrosion over anything else. Gold plating the springs was mentioned earlier... I wonder if that is a feasible solution or not, both cost wise and with the flexing.

[Melvang]

What is the travel, and how far does it need to be depressed before it triggers?

My concern is corrosion over anything else. Gold plating the springs was mentioned earlier... I wonder if that is a feasible solution or not, both cost wise and with the flexing.

Gold plated springs is no problem, has been done before, actually pretty common in Korean Customs.  The issue is going to be any design that causes the spring to impact the side of the barrel.  Look into IBMs early patents.  My suggestion would be using a flipper but have a contact plate on the bottom of the flipper and have that close the connection.  There is a reason MX switches contact points are gold plated.  No corrosion, and very resistant to wear from arcing.  I realize that the current and voltages loads aren't high, but it is still a concern when you start talking about the number of times these will open and close circuits. 

I would recommend staying away from capacitive designs as the PCB fabrication and design requirements are quite difficult, precise, not to mention calibration, firmware, and controller hardware sourcing.  Staying with a momentary switch design will drastically simplify the rest of the design.

[Dernubenfrieken]

I heavily recommend these being made to the same size as MX switches, so that you don't have to custom design multiple layouts for plates and run group buys. Would be a lot better to be able to buy a standard 60% plate, ergodox, etc and drop these in, and simply have to get a new pcb/case

[E TwentyNine]

The dimensions and pin placement would require a new PCB design, unfortunately. I would have liked to make it match Cherry mount PCBs, but it wasn't feasible.

Can you get into specifics on this?   What are the hurdles to overcome?

[Zekromtor]

The dimensions and pin placement would require a new PCB design, unfortunately. I would have liked to make it match Cherry mount PCBs, but it wasn't feasible.

Can you get into specifics on this?   What are the hurdles to overcome?

Probably that the spring and spring pin need to be centered for the design to remain as simple as possible.

[Arcanius]

Probably that the spring and spring pin need to be centered for the design to remain as simple as possible.

That's exactly it. I have gotten enough requests to make it 14mmx14mm to convince me that it'd be a nice feature to have, so I will make it mountable on existing plates, with a requisite PCB change.
Ideally, I would sink the spring below the PCB, a la Cherry, but having the spring as part of the circuit complicates things; I'd have to run a lead upwards from the spring inside the switch housing. I'm looking into something like this that's also not too expensive to produce.

[Zekromtor]

Any info on the travel, and how far does it need to be depressed before it triggers? I'm looking for a switch that breaks as close to the bottom as possible.

[Arcanius]

Any info on the travel, and how far does it need to be depressed before it triggers? I'm looking for a switch that breaks as close to the bottom as possible.

It's going to be 4mm, with triggering around 2-2.2mm in, much like IBM buckling springs.

[ctm]

Wow! This is amazing!!!

[MOZ]

I would highly recommend having compatibility with either Cherry MX plate and PCB or Alps plate and PCB. both of these are common, specially new keyboards incorporating both.

[KingRama]

This is so damn cool man, can't wait to try one out one day

[Zekromtor]

How are you doing your prototyping? and who are you getting to make the batch?

[ideus]

A switch with the contacts at the same position of the MXs would be a hit, even if it is more expensive.

[merlin64]

As an intermediate step, would it be possible to just use the MX compatible portion and convert an IBM model M to use cherry keycaps?

[Techno Trousers]

^or better yet, a Model F

[merlin64]

Raising this back from the dead, cuz I'm still interested!! Any update to this project?

[Wingklip]

I am currently working on a beam spring type design as a cherry compatible.

I have a really good idea for the buckling spring cherry which solves it's height issue and makes it the same size as the cherry switch. But if you want to patent it, please put my name down as part for royalties, because I am piss poor 

[zoggynog]

I need these in my life so I can have a dark blue model M clone compatible with gmk laser.

Everyone get back to work!