The Silo Beam Switch - Beam Spring switches for the modern era



Oh, good, this is public now.

I’ve already commented on the board itself. It incorporates about 90% of the feedback from the XMIT Hall Effect boards. I wouldn’t do analog sense but HaaTa has wanted to do it for ages.

The switch as you see is made by Kaihua. The story here is that at CES in 2018 I brought along a Beam spring sampler and happened to show it to the CEO of Kaihua. The folks at I:C made sure he got some NOS Beam spring switches. That CEO is something of a switch junkie (surprise!) and was optimistic about being able to incorporate something like it into something modern.

With some effort these could be made into retrofits for existing XMIT boards. Though, I can’t promise the actuation would be at the right place.

I do wonder if Kaihua will make Beam spring contact-based switches.

(cross-post from


I’ve got a manufacturing concern.

What are you doing to make certain that the board stays together, and that the bottom of the contactless key switch mechanism is a constant distance from the sensors across the board? You’ll need to make sure the PCB is well fixtured for this.

Several Hall keyboard designs clamp or fixture the housing directly to the PCB so that this distance is constant.


Especially since you mentioned “I do wonder if Kaihua will make Beam spring contact-based switches”,

who has the rights to the switch design? If Input Club owns the rights to the switches, can Kaihua make a contact-based beamspring even if they wanted to?

I really wish a contact-based beamspring eventually comes out because I already have a IBM 5251 themed build built with clickies and these switches would be absolutely perfect for the build. Even the click sounds really similar to OG beamsprings from the video :smiley:


Why do we have to start with a full size board? What exactly is the reasoning? Has someone done sales research showing mainstream users want to pay for premium switches? So bummed about the form factor


I’m guessing the board for this is the same as for the other Hall Effect-revival switches, which are partly being touted for their analog sensor (which I can’t imagine appealing to most folks who aren’t gamers). Add to that the fact these switches are most likely to capture the existing, full-size-heavy, buckling spring enthusiasts and I think there’s a pretty fair rationale for starting from full-size and moving down.

Of course, I may also be biased since I’m into full-size boards and still haven’t seen any of the neat innovations from the community filter into the full-size form factor. I’ll believe the notion that it’s better to start from a small form factor and work up when I ever see it happen :wink:


Name one commonly played pc game that heavily relies on a numpad. I’m curious what a gamer needs a numpad for. For what it’s worth I’ve played PC games for 3 decades. I’ve never used a

I had a job that required employee ID lookup quite often and I relied on the keyboard numpad a lot so that I could enter ids while on the phone.

But I wouldn’t imagine a lot of corporations are not willing to buy anything other than a rubber dome for their employees. My company had over 2000 employees and I was the only person I know of to used my own keyboard from home. Everyone though I was crazy.


I only have used it in 3D software and Garry’s Mod, so the utility of it for me is… limited. I’d much rather have a good old TKL for the first run of these switches than I would a fullsize, but I guess IC has to start somewhere to fill in their lineup.


I’m not here to have an argument, I would just like to ask you to consider the fact that the feeling you are having is one that people who want full-size keyboards have had about almost every neat keyboard/innovation the community has made that has never made it to the full-size form factor.

You like the form factor that you like, and I like the form factor that I like. I happen to be getting catered-to first in this instance (for the first time I’ve ever seen) and I would love to see this technology make it to your form factor so you can enjoy it as much as I will.


I think Grand Theft Auto V needed a numpad to fly the helicopters. I only remember because I was so annoyed.


This is a concern of mine. I think that they’d take the HHKB approach and use a million screws in to standoffs, but you really want a stronger plate and you’re almost asking for integrated/thick plate like a Topre board.

I really hope there’s not that much in the PCBs as community acceptance and specialized layouts are a huge deal to me. I’m dissapointted in the Hakos, tbh, and if they were anything but BOX switches or if they required a different footprint, I’m sure they’d be extinct.


This is so interesting. Sure the form factor is not what I prefer but just the fact that this will be available in the future is already making me super excited.

The detailed article is also great.


I absolutely abhor having to use row 1 for inputting numbers. I am constantly inputting numbers to budget for projects, compare prices, etc. I don’t understand why a gamer would need the navigation cluster. We use wasd anyways so holding a modifier should make that cluster unnecessary. Someone make a Vibe with a 2u “0”! *desperate sobbing" …Anyways, to each their own, but I definitely can understand the need for a numpad.


If i were to input numbers all day long, a numpad is essential. But what I hate the most of a TKL is the place where the algebric symbols are, especially on non US or UK layouts.
Said that, a separate numpad is a good idea, and a clicky switch like jade or navy would work perfectly there.


The fact that these switches also seem to require at least 4.0mm travel is exciting to me, since I really prefer long travel as opposed to bottoming out. What Kailh’s been doing lately with 3.5mm travel is getting on my nerves.

I will certainly get the 108-key Keystone just so I can use these, but would really like to see them show up in other boards, like an ErgoDox Infinity Silo Edition, or similar. Really would prefer if the electronics necessary to work with these was open source, so that motivated people could make whatever design they want. A beamspring Dactyl / Dactyl-Manuform would be the ultimate typing experience.

Edit: @HaaTa Will Keystone be available with some kind of silver aluminum bezel? I’d love to have a full-size keyboard with that type of look. They are hard to come by.


We have been working on improved stabilizers though it’s been a long road…they are hard to prototype without spending money on tooling…and then it doesn’t work right. I wouldn’t hold your breath for these.

Fortunately, on Kira the stabilizers are quite good so I’m not as concerned as I was after the K-Type shipped.


All keyboards that use these switches must attach the plate to the pcb at multiple points. Input Club has already been doing this for keyboards with hotswap sockets going back to the K-Type.


It might be possible to have a contact-based beam spring switch. Though I’m not 100% sure it’s a good idea.

  1. I don’t currently see a way to fit the mechanism with a Cherry MX pcb footprint. Something more compact is needed.
  2. Matching the activation point with the click is near impossible. This is the danger of switches like Choc Whites and BOX Whites where a slight difference in tolerances is all you need to make things perceptibly bad.

Regardless, if this does happen, it’ll be after Keystone.


We’ll see what kind of case options will be available. For now we’re trying to keep things simple and focus on hall effect + beam spring.

Keystone will also have a steel plate, can’t have those switches moving around.


Do you already have design documentation somewhere to help people get started, or is there any existing resources to help people design hall effect boards? I imagine it’s probably not at that stage yet but figured I’d ask


Not yet. At this point the most I can say at this point is to look at analog hall effect sensors and MCU ADCs.

I’ll be releasing more details once I feel comfortable with the overall design (it’s designed, but not fully validated yet).