[IC] The Cruciformer: a Tool for Fixing MX Stems

Render Gallery

Quick Info

  • This tool shapes new and old Box stems to 1.28mm in the right dimension (kind of works with Cherry Clears and other oversized stems as well)
  • Precision CNC’d from tool steel or Type III Anodized 7000 Series Aluminum
  • Tool and accompanying jig should work with loose and mounted switches
  • Anticipated $35-$25 MSRP (depending on MOQ of 100-250) the tool offers cost-recovery for 40 pretool switches or costs savings if you have a 60% or larger board and get switches at below their launch price What the tool does This tool and optional jig (Price TBA) that cuts Box/MX-type stems into appropriate dimensions whether soldered in a board or loose switches.
  • Interest Check is for buyers and for vendors (buy more than 10 units and you can get vendor pricing).
  • Link to google form here!

Why you need a tool
The difference between the bad box stems and Cherry MX cruciform is is ±0.025mm more on either side, or about one-third the thickness of a human hair from each side. Hand trimming stems is laborious, time intensive, and riskily imprecise – I know from experience. Both new and old stock Box switches have tolerance issues and even some of my Gateron stems and Cherry MX clear stems benefit from clean up. As someone in the Discord said: “You have been struck by the curse of the box switch, take 4 damage for every $100 value in keycaps in your inventory.” Read a little more about stemgate/boxgate here

This tool is the right tool
I have been working on this since about when stemgate/boxgate broke. Other approaches introduce as many problems as they solve with consistency in stems and raise durability concerns with the tools. Some home-brew trimmers get stuck or bind on either the cut or remaining walls. This tool’s internal geometry relieves friction on most faces and the jig also helps you grip the tool to pull it out. This tool cuts to 1.28mm on the critical dimension based on the averages of ≈2k measured cruciform sizes from common commercial switches (I also used the community spreadsheet of sizes and this thread for validation). The tool is somewhat self-centering, and the optional jig will take all the guessing out of trimming the stems whether you have them soldered on a board or loose. This tool will be good for hundreds to thousand(s) of cycles of plastic stems without sharpening. Based on my background in knife design/making, brass and non-treated stainless steel are poor choices for a cutting tool, even of softer plastics; this tool will be CNC’d from either tool steel or 7000 Series Aluminum with hard coat anodization (equal to mild steel raw) based on results of tolerance and durability testing – either material is ideal, we’re just working out which is the better one. The tool should leave stems looking pretty squared-off, but there should also be some distinct forensic marks so if you buy or sell switches, you’ll know they were fixed properly. The round-over on the cutting edge helps smoothly transition the cut edge to the existing stem and matches keycap grip on the stem without leaving a shelf or square corner that still stresses a keycap shape.When this tool trims a stem, it will cut all the over-dimensioned surfaces without making keycaps more likely to fall off - the nubs on a BOX stem reduce parallel surface area in a keycap stem for friction to grip, buy correcting the stem to parallel walls keycaps should hold more secure (more surface area for grip without the nub stretching).

Buying the tool is economic
The price point is pretty healthy if you want to safely use Box switches, even if you just want to fix the Box switches you already have. Take the price per switch, and add the price of the tool divided by number of switches to find out if you are saving money or rescuing sunk cost – e.g. I love my Box Royals and want to use them with my favorite keycaps, I got 100 of them at a price of about $0.30 each, bringing the per-switch price to if at the 250 unit cost of $25. Even adding in the cost of the tool, I am saving money over getting new stock. If I tried to buy an alternative and make 100 Cherry RGB Jailhouse Blues ($0.50/ea+ $20 JSpacers + lube), then I would spend more per switch.

Back it and I’ll release Free and Open Source parts
While I will not be open sourcing the cutting tool itself because it won’t do users any good, I will also be entirely open sourcing a 3D printable design for a switch grinder (jig for jeweler, needle, or ceramic file for grinding stem into shape or cleaning up the stem) and one that uses X-Acto blades if the project hits MOQ. Even if you don’t want to purchase a tool to fix the stems, I want to help provide options. If we hit the 100 MOQ, I’ll open source one of the versions, the other at 150 units, and the jig at 250. I have a number of future projects (cases, keyboards, tools, etc.) and I’m trying to support the keyboard open source community. A couple projects in, I’ll release the laser scans of various keycap profiles. If you want support open source and community projects, support this!

Timeline for release
The CNC shop is available to make the production run in early February.

1/30/2019 UPDATE:

CNC shop let me know that they’re going to machine the prototype run next week and I’ll get some video of it in action for the GB.

2/28/2019 UPDATE:

CNC shop said they were going to machine the prototype run the first week of February, but I haven’t gotten anything back from them since, I am reengaging for the third time in the next couple days and will share my results. I’ve been dealing with job changes and a death in the family still, but I’m trying to get this going. I will add this to the text of the post as well.

3/5/2019 UPDATE:

I have reached out to the CNC shop president and my independent rep because the wait is killing me. I’m going to start talking to other shops or in-community makers if I don’t get an answer back soon.

3/12/2019 UPDATE:

Cruciformer lives! I’ve met with the CNC shop a couple times since the last update. We’ve made some great progress and I’m super happy with having waited (although not 100% pleased with the wait, haha) because the tool has improved and we’ve tested it more. The first production run is being machined now and I’m getting ready to send some pre-production samples off to community members for feedback/reviews. I’m working on video of it in action and winding up my testing on accuracy, effectiveness, use, and longevity (all looking good so far!).

3/17/2019 UPDATE:
Cruciformer lives! Pre-order soon™. I’ve added an update below, including a picture!

3/28/2019 UPDATE:
Waiting on the final finishing fixtures, the shop got caught up in a much larger job and they’re doing the soft-jaw and fixtures for the tool and a couple hundred tools will go out to anodization.

4/4/2019 UPDATE:
Still waiting on completed fixtures to machine the little Hershey kiss shape off of the top so the tools can go out for anodizing. I will be going into the shop to complete the tools on Saturday. I’ll be 3D resin printing my Nexus Slider jig soon. I’m holding off on releasing testing data to make sure it is representative of all the tools in the batch – that’s probably when the pre-order will go live. (I’ve only got two keyboards of BOX switches left to test with). So far, the ABS DSA keycap hasn’t shown signs of stretching over the weeks it’s been on the fixed stem.

4/4/2019 UPDATE:
Pre-order is open! expected shipping by end of the month! It’s a soft launch, I’ll be dropping a video demo and the testing data soon. We’ve decided not to anodize the first batch because it lets users touch up or sharpen the tool, and that was more effective than the anodizing over a longer life time (anodization sharper for longer, but can’t sharpen; not anodized, sharpen forever and still works for dozens to around a hundred of cycles between sharpening). If there is additional demand, we’ll release an anodized version, but for now, just a nail file or emory board will bring it up to sharpness again. My SP SA ABS test caps haven’t shown any signs of stretching and the stems are trimmed to the right dimension even when the tool is dull (more numbers and final details on the jig to come).

4/25/2019 UPDATE:
I’ve added a longer update with explanations of the tweaks and the Proxies

5/1/2019 UPDATE:
We have the machine time to do the last corrective steps on Saturday and the jigs should be cut before then!


I’m in :slight_smile:

1 Like

in for 1

1 Like

Another IC. I sure hope this one goes somewhere!

Awesome to hear about the open source stuff too. Hope it works out!

1 Like

In for a couple, I have hundreds of box switches I’d like to recover.

1 Like

I’m definitely in for one if this make GB!

1 Like

I’m in!

1 Like

I’m in, as well! I’d love to try out BOX switches, but I’d hate to damage my valuable GMK and MaxKey sets.

1 Like


1 Like

So annoyed that despite all the delays Kira is coming with almost 100 switches that can break stems. Already put in an order for 100x Silent Cherry Red RGB’s. This would be useful to have though if I prefer the box ones that come with Kira. I’m in.


I hope this goes into actual GB soon! I wanna use some MT3 with some box clickies but I don’t wanna stretch the plastic like I did with my PBT sets.

1 Like

Appreciate the support!

Would you be interested in testing once I get final samples?

Nice! If you haven’t, head over tothe IC form and let me know you’re from Keebtalk (I’d think some of the best of the community is here).

Haha, me too! It helps that I’m working with a professional CNC shop that I can physically visit…

I’m glad to hear. I was quite surprised that nobody seemed to comment or care about it, there were even people saying things like “Interested in reverse engineering and making my own.”

This is my personal use case. I’m focused on making the tool as ergonomic as I can for higher switch counts. Don’t be afraid to get more than one so you can switchmod with your friends! :wink:

They really are some of the better switch architectures, once they get silent stems and silenced click bars I’ll be interested in swapping out more of my MX-style. The BOX linears are really good (2nd place behind Gat yellows or retool blacks), BOX clicky are worlds better than click jackets, and BOX tactiles are equal or more unique. My fear was stemgate would stifle the only innovation we’ve seen in switches.

Agreed. I emailed support about changing from BOX Royals to Cherry RGB Blues for Jailhouse mods, and if you got the metal frame there still might be time to change.

Same, haha. Sorry to hear about your old stems. How do you like MT3? What boards/cases do you put them on?

1 Like

Didn’t get in on either of the two preceding sets - but I’m in on MT3 Godspeed! I’m probably going to be putting it on my wooden 60%, but I may end up with it on my Novatouch when that comes in for the full Not Topre Hi-Pro Experience.

1 Like

I definitely would! I have over 200 switches that would love some mx stem fixing treatment hehe

1 Like

I’m in!
I hope you find a reasonably priced way to send them to Europe…

1 Like

I hope you can get in the GB!

David at Mechboards.co.uk said he’d proxy for UK/EU and that Brexit hasn’t changed anything so their shipping prices are the same as anyone else. I’ll line up whatever is the cheapest option for you all in the EU.

I’m still looking for Canadian and Asia proxies, if anyone has suggestions or favorite vendors, mention them or introduce me! :slight_smile:

1 Like

That’s great news!
I’m actually based in the UK and I really hope brexit doesn’t make delivery here even more complicated, because it is hard enough as it is…
I had good experiences with mechboards.co.uk in the past :+1:

1 Like

I’m in. This would solve my problem of bags and bags of unused boxgate switches.

1 Like

1/30/2019 UPDATE:

CNC shop let me know that they’re going to machine the prototype run next week and I’ll get some video of it in action for the GB.