Mechanical Keyboard project

I fancied a new project as I had the itch to make something new. Seeing as I spend most of my time in front of a computer during the working hours of the day I wondered if I could make that experience better by having a nicer keyboard.

There’s a whole serious world of mechanical keyboard enthusiasts on Reddit and elsewhere with some great ideas and how-tos. A really great community, but also a really slippery slope. Once you get started I’m sure it’s really easy to keep spending some big money on custom kit.

I wanted to get something not too crazy to begin with – and something that I’m sure I would be able to use day to day, hence I was a little reticent to get an ortho or split keyboard as my first project. I also wanted to minimise the amount of soldering that I would be doing as my skills are not up to scratch (nor have they really ever been). Also I’m a big fan of the ISO layout and the correct size return key – so that layout was a must too!

Turns out it was a bit of a struggle finding a hot swappable ISO layout board – I almost gave up at a couple points in the process. Eventually I settled on a YMD75v3. It seemed to have decent reviews, some sweet LED support (both per key and under board) and, most importantly, was hot swappable and ISO. I probably didn’t need the extra row of function keys, so perhaps a 65% would have been okay, but getting hot swappable ISO in that size proved a little trickier.

The next thing I had to decide on was a case. While there are many places out there to get compatible and awesome looking cases for these 75% boards, I wanted to order from as few places as possible to save on shipping and customs (because face it, importing anything into Sweden is going to get a hefty customs charge, and the fewer fixed price ‘admin fees’ that I can pay the better…). So I was limited to the cases that the folks at YMDK were selling. Thankfully there were a few options, I could have chosen a metal case, but I wanted to go down the very LED route so I figured an acrylic case was a good way to show this off. Not to mention it was pretty much the cheapest option.

Next were switches. I’ve never really had the opportunity to try many different kinds of mechanical switches, and while I would have liked to order a testing board with a couple different options and try them all out, that would have taken far too long – and I don’t have that kind of patience. So I listened to a couple sound tests, fully knowing that I wanted something pretty quiet and linear, and settled on Gateron reds. YMDK sold Krytox GPL 205, so I figured what the hell why not lube these switches too for maximum smoothness?

Next up were the stabilisers. Having read lots of things on the MechanicalKeyboards subreddit about crappy clip in stabs I figured I should order nicer ones and not go for the standard package. So I picked up some gold plated screw in stabs that I fully planned on lubing for maximum smoothness.

LEDs were also a no brainer. At an extra $12 or so for 100 LEDs and 200 SIP sockets to make them hot swappable, it was all systems go. Not wanting to have a board that was too subtle I went with a colour called ‘slow rainbow’ which turned out to be a great choice. I bought them in the 2 x 3 x 4 mm size so they could fit nicely into the hollow on the front of every switch.

Keycaps were also a hard thing to choose. I wanted something cool looking and totally non standard, but most of them didn’t have an ISO set, were super expensive, or were not very translucent, so I’d miss out on the great work that I had done for the under key LEDs. I settled on some white PBT doubleshot pudding caps. They were cheap, and I figured that if they didn’t look very good then I could always swap them out and replace them for something nicer down the road.

Eventually, and I mean eventually in the sense of a week or so (which is pretty reasonable for FedEx) I had the parts in my hands and I could begin construction!

First step was to get those stabilisers lubed and attached to the board. Pretty straightforward and I could always add more lube later on if need be! Then I wanted to attach the plate next, but I realised that this might hamper my attempt at placing the SIP sockets for the LEDs. So I focussed on pushing all the sockets into place (this took a while, and required finding some blunt tool to make this a little easier) and clipping the legs from the rest of the sockets (also surprisingly time consuming with added danger from little bits of metal flying everywhere).

Once that was done, I could focus on prepping all of the switches. This involved opening all the switches (by hand, I didn’t get a tool, but after doing 50+ you begin to get the hang of it) and separating the top, bottom, stem and spring. Then using a very little paintbrush applying Krytox 205 to the stem and parts of the bottom casing. I put all the springs in a little bag added a little of the 205 and shook them for a while to distribute the lube everywhere. I’ve heard people using 105 or some other chemical for this, but I figured I’d use what I’d have around – and it was fine.

Then came the somewhat tedious process of assembling all the switches back together. Fine if you have a few, but I had 84 to go through. With a little background show on, this wasn’t so bad in the end.

Once reassembled I could place them on the board and click them into place. Remember no soldering here so this was a pretty straightforward process. I eventually could get the plate in place, and it was all starting to look great!

Then came the installing of the individual LEDs. I had to cut the legs a little shorter to make them fit flush with the switches, but once I’d done one I could figure out the ideal length for the rest.

Then came the case. This was a case (pun intended) of gently placing the board on the standoffs and finding a way of screwing the tiny screws in. I lost them in the board a couple times, but finally made better progress once I got a magnetic screwdriver head!

Then for the test! I connected the usb-c to usb-a cable to the computer and lights flickered on and then off. And a little warning appeared on the screen that a USB device was drawing too much power so it was cut off! Not the best start, but after trying a different cable or two, I realised the problem was with the cable and not the board. Phew.

A couple LEDs didn’t light, but I after a little jiggling around they lit up like the rest of their friends – maybe I had cut the legs a little too short in some cases. Once I’d confirmed everything was lit up, I started placing the keycaps. Maybe I hadn’t chosen the best quality ones ever, but once they were installed with the slowly glowing rainbow lights they actually looked pretty awesome.

I fired up VIA, installed that on the board (having to press a fun combination of different keys), and I was off to the races. Being a pretty big keyboard I didn’t need to think much about different layers, so the configuration was super straightforward. Plus, with VIA, reconfiguration at any point was mega simple.

And that was pretty much it! My first keyboard build: done! The switches were pretty light and fast (and over time would undoubtedly get a little faster as they broke in more) so it took a little getting used to, but nothing I couldn’t handle with a little practice.

Overall, a super fun project. Total cost about $200, so not super cheap, but I could have saved more with fewer LEDs, fewer keys, some soldering and no ISO. But hey, maybe something to try next time. Shout out to David from YMDK for answering my (stupid) questions.

I knew my first build wasn’t going to be my last build, so I thought I’d make a little list of things to try next time:

  • Split keyboards, why force yourself to have your hands next to one another when maybe that’s not the most natural position
  • Ortho, I’d entered the rabbit hole, so might as well go down a little further right? I wonder if it would make me any faster at typing?
  • Different switches. Gateron reds are apparently fast, which I really like, but should I go for something different next time?
  • Soldering! How hard could it be? As long as it wasn’t surface mounted components then it might be pretty doable!

Leave a Reply

Your email address will not be published. Required fields are marked *