Finally found it (and created Readme.md + GH Repo)

My long, arduous search for a 3.7V Li-po/Li-ion battery has finally come to an end.

Since my local area has quite strict import of dangerous Lipo battery rules, I had to go find a LiPo battery from local vendors. This actually took way longer than I expected, because I refused to believe that the Mouser Electronics and DigiKey websites were legit. Turns out, they literally have a headquarters here in Hong Kong. I can source my battery holders from them

The problem was always this - the battery was too big and bulky. A flat rectangular LiPo battery would fit into that box nicely, but it wasted a lot of space for the flat solar panel, and the space I had constrained the volume I could put. So I decided on 3.7V Li-ion batteries - the little cylindrical ones.

Here’s a cool thing I learnt. Many electronics projects use the 18650 battery, but I never knew what it meant until I found its AA(14400) and AAA(10440) counterparts. It is actually very systematic:

• 18(mm) is the diameter of the cell
• 65.0(mm) is the length of the cell

Isn’t that cool? Just wanted to share that.

So finally, I decided to use the 10440 batteries (3.7V, 350mAh), and I would also need keystone clips to handle the 2A+ current draw from the RPis. I can insert these clips into my 3D-printed case to save some space.

Things I need to do next:

• CAD case: hopefully this is less cortisol-demanding
• Firmware and software
• Finish Readme and submit this project for funding

I’ll see you soon, probably taking a break over the weekend due to some family visits.

PCB Part 2/2 (Finished)

As the saying goes: The definition is 50% of the work. These 10 hours (and some more unlogged hours) were some of the most perfectionist work I’ve done in a while, aligning everything nicely and drawing clean tracks. I really wanted this to be a quality piece of open source hardware instead of some lazy slop.

Steps

I’ll describe my process in these 10 hours, since I think it is not only good for me to reflect on the stuff I did, but also great (debatable) experiences I could share with others so you could learn from the inefficient work I’ve done:
1.Imported 3D models for 1-to-1 sizing and making sure it all fitted in the case, and transferring over to a CAD model with my tin.

• This actually took a while because of the niche modules I’m using, and I had to design many of them myself. I recommend you try to design it as simply as possible for layout. This will give you a rough idea of the size of each component. As you can see, I had many components with lots of details, and that lagged the hell out of my Shapr (and caused my KiCAD to crash as well, didn’t press ctrl + s again when checking DRC), so making simple polygons would be the way to go.
• A neat trick I learnt to quickly add 3D models to KiCAD was to edit the PCB in a text file using Notepad++ (and plugins) to find all and replace all.
• The 3D models also let you know how wrong you designed your footprints, so don’t skip this step.
• ALL of this should be done before routing, or else 1 reposition could waste so much time.

2.Create clean, manageable tracks.

• I spent a lot of time making sure the tracks were arranged in a logical way where I can draw the tracks through my 2-layer PCB.
• Vias were very useful when I topologically had no way to get to my endpoint.

3.Add mouse bites to finish it off.

• I used mouse bites to make the PCB into 1 piece instead of 2 separate ones so I can save a bit of money. Again, my goal for this cyberdeck is to make it relatively cheap. I used mousebite footprints from this amazing library: https://github.com/AchimPieters/mouse-bite.pretty

4.Switched protocols from USB to UART, because the USB pins were too difficult to solder on the other piece if I soldered it flush against the PCB - this is a trade-off between compactness and functionality.

So yeah, to the ones reading this devlog, hopefully you learnt something. I certainly wish I’d learnt all of this before I started the design so that I could’ve saved 5 hours.

Problems

• The battery and solar panel cannot fit together at the same time.
• The wires may not fit for crossing the channel (UART wires) when the lid is closed.
• It may not be rechargeable with a solar panel because the charging module would break, unless I make room for another switch.
• I cannot order LiPO batteries from overseas due to legal customs constraints (at least I don’t think I can unless I pay hefty fees)

What I need to do next

• Now that the PCB is done, I need to find a smaller solar panel and lipo battery to fit everything inside - at the cost of power, however.
• Once the space issue is solved, I need to create a CAD case for the keyboard and the screen.
• And finally, I can create the firmware and finish the project at last.

PCB layout 1/2 done

It felt like this was a very unproductive 6 hours, but the struggle genuinely paid off as the layout is essentially perfect right now.

The layout is CRUCIAL. Without a good layout for a compact and smart design, there is no difference between handwiring the components on a breadboard and using a PCB.

There are a few things I wanted to achieve in the PCB:

1. Understandable for anyone editing/fabricating this as it will be open source
2. Easy to edit - clean tracks
3. Compact (duh, for the altoids case)
4. “Mountable” on a 3D printed case/frame - so it looks cleaner

So far, goals 1,3,4 are done, and I have yet to draw the tracks because I’ve been so busy with making everything aligned in a grid (like the keyboard), and making sure everything was readable (literally had to handpick the silkscreen out of each component).

To make sure everything was fitting correctly, I drew some temporary silkscreen on top of everything to know exactly how big it would be.

Improvements I have made since the last one:

• Screen has been reconfigured to SPI instead of I2C by rewiring SDA to MOSI and SCK to CLK on the zero 2W.
• Added a switch for on-off; I’m currently struggling to find a good enough switch with enough current rating so it doesn’t melt and break.
• The diodes are placed under the buttons so that they can save space - this however, may lead to an uneven bottom, and the keyboard would need a good 3D modelled case. (At first, it was placed in the blank space above the keyboard as seen, but now, it can be made a lot more compact)

P.S. Tragedy struck during this session - 2 hours of work was lost because I didn’t save, and KiCAD crashed. Will never forget ctrl + s ever again.

Completed KiCAD schematic

I’m fairly confident with KiCAD by now so I did this pretty quickly, just under 6 hours.

Again, this is a nature themed mini cyberdeck that fits in an altoids tin, with an LCD screen, a PCB keyboard, and a rechargeable battery (possibly by solar panels).

What I have essentially done:

1. Found all the components I need from Aliexpress
2. Created and assigned footprints/symbols such as the battery charger and LCD
3. Changed up a lot of the initial plan - might add a solar panel to enhance the “Nature” theme and possibly make it a doomsday device as well XD, will confirm at the CAD design stage when I finally fit all of these components in that lil tiny box
4. Wired everything up! 0 ERC errors (we can ignore some of it because I just can’t figure out how to assing the pins functions right like wdym bidirectional is not allowed for power output)

This is great progress, I’m excited to wire the PCB up and become miserable once I find out these components dont fit on one board :( But that’s for tomorrow, see you then!

P.S. Also awaiting sanity check in the slack channel for the schematic, might make minor edits tomorrow

Idea and research

“When life gives you mints, you make a cyberdeck with the box.” -Tsun su or something

Overview of the project and the plan

I got a Ricola box that reminded me of the cool cyberdecks I’ve seen on Pinterest, and I really wanted to make one.

I wanted to make a cyberdeck that is nature-themed (really matches with the mountainous terrain on the box), and is a general-purpose computer.

I would also prefer it if this could be used as a portable gaming console XD, running Pokemon Red or Doom or Tetris, etc.

I measured the dimensions, and it turns out it was exactly like an Altoids tin, which most people usually use to make cyberdecks and other stuff.

This will likely be powered with a Raspberry Pi Zero W for its compactness (and the fact that I already have one), and use 1 PCB for the keyboard (probably QMK). This is subject to change, but still a rough idea of what I’m aiming for.

What I need to work on

• I need to figure out the different components I will use and source them from Alibaba, preferably for a lower price and faster shipping due to my geographical advantage.
• Need to design the PCB for keyboard matrix layout and figure out what microcontroller to use.