ultimatehackingkeyboard / uhk60v1-electronics

The schematic and PCBs of the UHK 60 v1

Ultimate Hacking Keyboard electronics

Contained are the schematics and board layout KiCad files for the Ultimate Hacking Keyboard.

Please make sure to clone this repo with:

git clone --recursive https://github.com/UltimateHackingKeyboard/uhk60v1-electronics.git

Now, you should be able to open any of the 4 KiCad projects of the repo:

• left-main contains main PCB of left keyboard half.
• right-main contains the main PCB of the right keyboard half.
• display contains the PCB that interfaces our custom manufactured LED segment display with the left main PCB.
• connectors are very small PCBs that are soldered vertically to the left and right main circuit boards to interconnect the keyboard halves without needing a bridge cable.

uhk60v1-electronics's Issues

Add tags for all the past revisions

Move the parts farther from the pins of the MX switches which are too close

Move via farther from the pad of CR6 on the left main PCB

Travis script throwing error

I'm trying to use the travis diff script and it's throwing error:
root@12ba8a8b9dbe:/kicad# ./generate-png-diff.sh
Copying 187bd8e:drone-kicad-sample.kicad_pcb to ./plot/187bd8e/
Kicad files saved to: './plot/current' and './plot/187bd8e'
Converting ./plot/current/drone-kicad-sample.kicad_pcb to .pdf: Files will be saved to /tmp/pdf
Traceback (most recent call last):
File "/tmp/plot_board.py", line 172, in
popt.SetColor(YELLOW)
NameError: name 'YELLOW' is not defined
Converting ./plot/187bd8e/drone-kicad-sample.kicad_pcb to .pdf's Files will be saved to /tmp/pdf
Traceback (most recent call last):
File "/tmp/plot_board.py", line 172, in
popt.SetColor(YELLOW)
NameError: name 'YELLOW' is not defined

Add fixating pins to the Cherry MX module

Add fixating pins to the Cherry MX module so that PCB mounted switches can be used, too

There are two libraries that already have this

https://github.com/geekhack-org/kicad-library

https://github.com/c0z3n/cherrymx-eagle

The latter would need to be converted to kicad. (done)

• Ensure license is compatible with the UHK geekhack-org/kicad-library#1
• Get cherry mx library UltimateHackingKeyboard/ugl-kicad-lib#2

Replace 6P4C connectors with 4P4C

Replace 6P4C connectors with 4P4C. Make sure that the chosen connector and the wall of the case leaves enough space for the clip so that it can be pressed. Also, 4P4C probably almost always uses reverse wiring!

Add footprint for the optional CT05-3050-G1 reed switch

Want to add the opensource logo to the board?

Thoughts on adding the official open source logo to the silkscreen of pcb?

http://www.oshwa.org/definition/

https://www.youtube.com/watch?v=QrHoowPpdQ8

Create 4 layer versions of the left and right PCBs

Hopefully, #32 will resolve the EMC issues, and we'll pass the next EMC test with flying colors, but we can't be sure. At this point, we want to put the UHK to market as soon as possible, so as a safety measure, we will design 4 layer versions of the left and right PCBs. On the upcoming EMC test, we'll give a whirl to both versions and pick the 4 layer version if necessary.

After finalizing the left and right boards, the left-main directory should be copied as left-main-4-layers and right-main as right-main-4-layers, then the 4 layer versions should be redesigned accordingly. The middle layers should exclusively contain GND and VCC.

Where can I find the files, for example in stl format, of the case for this keyboard?

Where can I find the files, for example in stl format, of the case for this keyboard?

Should TODO be split into issues?

There is a list of remaining todos

https://github.com/UltimateHackingKeyboard/electronics/blob/master/TODO

Thoughts on opening a github issue for each?

• can assign tasks to people
• prevent duplicate work from two people working on same problem at same time
• encourage task specific discussions

I'm willing to do the tedious work of migrating the todo's to issues if you want to go down that route.

Where do you get the libraries from?

I know you are really busy, so I'll try and be as self sufficient as possible.

I'm updating the readme with instructions so others can easily contribute. There are 3 libraries required

• special
• opendous
• uhk libs

I've figured out how to install opendous and the uhk libs, but even after a couple hours of searching, I can't find the 'special' library

Do you know where that library is located? Once I find it I'll document it so others can find it.

Create automated system to compare kicad diffs

When a new merge request is made. Use travis-ci to generate gerbers/pdf/png files and then use linux compare to generate a visual diff.

http://www.evilmadscientist.com/2011/improving-open-source-hardware-visual-diffs/

https://github.com/blairbonnett-mirrors/kicad/blob/master/demos/python_scripts_examples/plot_board.py

• Create travis config file
• Create scripts to generate diffs based off of plot_board.py
• Create aws account to upload artifacts to s3

You can assign this to me as I'm currently working on it.

Version 2?

Would it be possible to the V2 versions of the schematics?

Increase minimum space between tracks to at least 12 mil wherever possible

Replace the spring connector with a vertical PCB that uses pogo pins

Redesign the pads of the vertical PCB to not interfere with the keycap stabilizer

Redesign the pads of the vertical PCB to not interfere with the keycap stabilizer. Mechanical solder pads should only be added at the edges and the middle pads should be reduced

USB line impedance seems to be really bad

In the PCB production version the USB signals are of very low quality. That probably explains some of the USB issues I have seen over the last months using my two UHKs.

In this capture, a reference signal (orange) is compared to the UHKs signal (yellow). An impedance mismatch is obvious. The signal corresponds to a UHK keyboard with stock USB cable connected to a USB port with an immediate transceiver. When connecting other devices to this setup, the signal looks almost like the reference.

I have a few suggestions for the next PCB revision (if there will be one, which I really hope):

• check USB design guidelines and trace impedance, esp. for any specifics for the microcontroller used. ( https://www.nxp.com/docs/en/supporting-information/Universal-Serial-Bus-Training.pdf )
• avoid layer hops whereever possible. at first glance, it seems feasible to move C13, R3 and D14 to the top layer and reroute the USB trace along their old position. that would reduce via-count by two per trace.
• avoid dangling traces. currently there are traces going from the USB port J1 north to the board edge. these are probably for electrical testing, QA and firmware flashing? maybe for these two a simple bed of nails and two in-line testbeds can be used? -- I just tested cutting off these traces, without any visible impact. so that probably is not not neccessary at 12MHz.

I know that these recommendations are coming very late and that you very likely already know about these issues and have thought about how to fix those, if ever. But I was missing any reference on the USB issues I have seen and thought that there should be at least a ticket on this.

Just for completeness, here are a few things I have seen over the last month:

• Linux: plenty of bursts of "hid-generic 0003:051D:0002.0006: control queue full" in dmesg. those come all day long. (thats probably a software/firmware bug, but I am not sure)
• when booting up, I sometimes had the problem that the system does not start properly at all. it just hangs in the bios. seems to have gotten better by using a USB hub inbetween.
• random USB disconnects and reconnects (only rarely)

Protect against reverse polarity bridge cables

Creating custom bridge cables is a possibility, and it's easy to overlook the wiring resulting in reverse-wired bridge cables that can damage the electronics. In order to avoid possible damages, we'll implement reverse voltage protection using the following schematic:

The above circuit should be added right between FB1 and C6 according to the following schematic:

Re-wrap and reposition the "Unscrew the 5 large below the keycaps..."

Re-wrap and reposition the "Unscrew the 5 large below the keycaps and the 1 screw on the PCB" text to be more readable.

Replace 5 pin TVS array footprints with 6 pin footprints

6 pin TVS arrays (SC-74, SOT-223-6, 6-TSSOP) are much more prevalent than 5 pin TVS arrays (SC-74A, SOT-223-5), and they're backwards compatible with 5 pin footprints, so this will enable us to choose a wider variety of parts.

Add pins for Matias switches if possible

Not impossible, but difficult

Change MKL03Z8VFK4 to MKL03Z32VFK4

We'll be using the MKL03Z32VFK4 MCU instead of the current MKL03Z8VFK4. It's the exact same MCU in every respect but it has 32K flash instead of 8K. A new schematic part should be cloned from the old MCU accordingly, so its name would reflect the name of the new MCU. The new schematic part will have to be added to the ugl-kicad-lib repo.

Reroute D+ and D- tracks according to USB design best practices

USB type C connector?

Look into the possibility of using a USB type C connector

Parts need unique names to allow for auto routing

If you attempt to run the auto router, you will get the following errors

The first is caused because the holes do not have unique values

The latter is because every via has the same value

Cherry MX LED lights

Allow modders to use cherry mx switches with LEDs

• Update library to use cherry mx pinout with led holes. example done: UltimateHackingKeyboard/ugl-kicad-lib#2
• Choose a led driver that handles PWM
• Determine if sufficient pins available on the AtMega for led driver
• Design charlieplexing matrix

Optional features

• Software defined brightness, [like code keyboard https://codekeyboards.com/])
• Individually toggle buttons (caps lock, num lock)

Related #13

Add more vias to the female connector board

Currently, there is only one via per pad on the female connector board. We should use more, especially for VCC and GND.

Reduce resistor values on the TX and RX lines

Reduce resistor values on the TX and RX lines of the board-to-board connectors because they didn't work on the 5.1 generation prototypes

Use larger pitch components?

Look into the possiblity of using larger pitch components so that we could could make our manufacturing tolerances less strict resulting in lower prices and increased yield

Reverse 4P4C connector pins

Currently, the wiring of the bridge cable is backwards, so we should reverse the wiring of either the left or right 4P4C connector.