submitted

finally, i'm about halfway done with the project. I'm submitting it, and then when all the parts arrive, I'm probably going to spend another 10-15 hours soldering it and placing components. Then, i'm going to have to spend at least 5-10 hours making firmware.

the image is of an older iteration :)

Finished the cost chart

Just made a chart of what things ill need to buy, and where to get them from. I'm still sad that im 6 dollars over, just because jlcpcb is charging a huge amount for shipping. Also got the pcb ready for manufacturing.

Changed the pcb

Added a mountinghole (which was a big mistake, i should've done that earlier), which made me redo some of my edge.cuts, making me decide to completely redo my case :(. I also took the chance to estimate how big my battery is going to be, and decide where to put my testpoints.

Finally made an approx list of how much it costs

I found cheap switches on ebay, after i checked them to be from a very reputable seller. I found an approx quote of pcb + case (haven't finished either though) from jlcpcb, even though import taxes would apply. I found out that the keycaps i can just 3d print. I also got rotary encoders for about $4. I got the XIAO for 20 dollars. I got hotswap sockets for 10 dollars.

Just realized that you needed to add a battery... :(

Well, I completely forgot that i needed to add a battery, so i had to restart my CAD process. I did some research, and found out that i had to use a lithium poly-something battery. I found one on adafruit, learnt about how other companies are putting the - and + in the wrong cable colors, and then made a cube in my CAD project to show the battery. I came up with the idea that i could put it underneath my keyboard to provide some tenting angle, which was something that i was going to do anyways.

Decided to make the case a 1-part case

My case is going to be a 1 part case, because im very short on time (probably because i spent so long selling with my pcb design), and i have to finish this quick, and i still have ZMK to deal with after the case. I'm just going to have a hollow square with the top open, and that's going to hold it, and i'll figure out some way to attach it later.

Made a visualization of full PCB to make a case.

I exported my pcb from kicad into a .STEP file, and then i stuck it in Shapr3D (which is the best 3d modelling application). I had to use my computer for modelling it, not my Ipad, since the switches had a lot of seperate parts (i think 7 per switch), and i had 60 switches (420 parts + pcb stuff). I then discovered GrabCAD, which has so many good, high-quality models of exactly what i needed for the visualization, so i put a Seed XIAO nRF52840 SMD on there (which i got from thingyverse), then went to GrabCAD and got models for my rotary encoders. I placed them where it's supposed to be, but i think the switches i got were a little off, and i'm hoping that that's just the model being incorrect, not my pcb :)

Found prices for stuff and fixed the edge.cuts

When i was trying to export the pcb as a 3d model to make my case after, kicad said that i needed to fix my edge.cuts, so i tried to do it by hand, but then i realized that you could just do it by using the heal shapes feature, so i did that, got the pcb model, then i went to keeb.io and found 70 kailh choc v1 switches for 40 dollars, and the keycaps for 9 dollars. I downloaded a model of the switch so i can start modeling the case.

Finally fixed the mousebites problem, added rotary encoders.

Yayy! another problem i had that i fixed! When i looked at my 3d render of the pcb, there was a problem with my pcb; the mousebites just wasn't there, and it was essentially just 2 separate pcbs. I saw this problem a long long time back and i decided that that's a problem for future me. I did them now, after like, 10 days of procrastinating 😅. Also, i realized that making my own keyboard and having it just be just like every other keyboard you can find isn't good enough for me, and i wanted something special. I remembered the hackpad tutorial from 1 year ago, and i added 2 rotary encoders, 1 on each side. I'm using NFC pins to do this, so when i get into ZMK and firmware, I'll need to do some stuff to remap the functionality.

encoder:

new mousebites:

Routed right side of the keyboard, finished the pcb (mostly)

I tried using the Layout Replicator plugin on kicad 9.0 to mirror the left and right side, but i thought that was a little too tedious, so i just went and did it by hand. Way faster than i thought it would be, since I was bascially just mirroring it by hand instead of thinking it all the way through. It's not the same as the left side, because i didn't want to do it exactly, since even though it would look better, I don't want to spend way too much time on something that's pretty simple. I'm going to start working on the mountingholes now, since i need to start making my 3d model, and the firmware. I also ran into a huge problem with my edge.cuts, and kicad is giving me a hard time with my mousebites :( I'm kind of stuck right now, going to try to solve it later.

Routed left side of the keyboard

Ok, i still have a GND problem that i had in the last journal, but im going to fix that later and hope it doesn't break anything:). I have a lot of vias, about 30, and im probably going to have the same amount on the right side (if i do the same thing). going well so far, going to work on the right side next.

finished left pcb

Finally figured out my problem!!!!!!!!

Finally, i figured out the problem that stopped me from working on this for days...
I didn't place the switches or the diodes in the right place, switch 01 would be connected to diode 02 instead of 01 and that caused me to basically ragequit the keyboard (im not very smart). I found this out, then i found a problem with git. I tried 3 times before, and i couldn't make a git repo bc of the error: LF will be replaced by CRLF next time git sees it. I didn't want my files to be corrupted, so i just decided to make a git repo later :). It turns out, that wasn't a good idea, because now, after solving my last problem with kicad, i needed to make a repo so i can actually prove i did the work. I tried doing everything i can to solve the problem, and now i realize that i needed to paste a command in that doesn't use the Unix file endings, and uses windows 11 file endings. Then, i found out that when i trid to push a repo with terminal, i got some error that i couldn't trace. i ended up looking for a long time in my filesystem (which is very unorganized) just to find that when i made my 3 repos earlier, i deleted the repos, and not the git files from my project. I looked at the .gitattributes files, and i found that there was a line of code that was stopping my project from going to github. i deleted them, and now it finally worked! Anyways, aside from solving the problems, i also did some actual work on kicad. I changed the placements of the switches, diodes, and the microcontrollers so that it won't try to route things to the other keyboard. this took way too long, and i could've done it in shorter time if i actually knew how to use a computer. (i still have more problems to solve for some reason, but im going to do that later :))

as you can see, kicad has nothing that wants to route cross-board (except the GND to GND, another problem)

Starting to route the pcb

I had a tough time with blueprint for the past few weeks, because i stopped working on it due to exams and other work, then when I started to get back to working on it, I had way too many problems, and I couldn't get help for a long time (not trying to be mean to the support channel). I finished making the layout of the pcb, and i'm almost done with making mousebites.

Finished the schematic (I think I did)

Think I finished the schematic, probably got things like the mountingholes, and the VBAT stuff wrong, but I don't know if I actually got it wrong, or if I did it right. Also found out that I might be getting something with the left/right sheets wrong, because when I was making the schematic for the switches, whenever I added a new switch, instead of the first switch being switch 1, and the next being switch 2, it was the first one being switch 1, and the next one being switch 3. I found out that this is because whatever I do is duplicated on the right side, but since it doesn't look like that in the tutorial, I'm a little scared.

Fixed my problem and started working on it again

Fixed my problem that I was facing. It turns out, I was dumb and i was trying to add a singular .kicad_mod file to a footprint library (I think this was the error, I might be wrong). When i fixed it, I went back to trying to complete the split keyboard!

Failed to make it from Youtube tutorials.

I first started off by looking at some Youtube videos to learn some terminology (blue switches, brown switches, red switches, mods, etc), then I looked at some Youtube tutorials, but they all failed me, and were really outdated. When they all failed, I decided to make this devlog before I start trying to make it using the guides.