STM32X40 40% Otholinear keyboard

[tszaboo]

Hi. I wanted to share some pictures of a board that I'm working on.
It isn't assembled yet, the diodes haven't arrived.
I call it STM32X40, I probably will come up with a more elegant name for it.
It is somewhat inspired by the Mystrium, but I'm not fan of through hole components where you need to cut the leads. It makes a mess of the room, and it takes a lot more time to solder than SMD.
SMD is easy.

Features:
- 40% Otholinear 47 key
- FR4 plate with silver-on-black color scheme, support for other plates with 19.05mm distance (BM40 tested)
- QMK support with STM32F411 microcontroller
- Hilbert curve on the bottom. https://en.wikipedia.org/wiki/Hilbert_curve
- One sided PCB routing without vias. The bottom side of the board has 0 tracks on it. Organic flow of the tracks.
- USB Type C on left side
- 0% LEDs
- PCB mounted stabilizer

[kajahtaa]

Looks great.

[Gorbon]

- 0% LEDs

Oh nooooooo! 

In all seriousness; very, very nice. It's a labor of love, lovingly made and I love it!

[suicidal_orange]

I am slightly disappointed that your 'organic' traces are very straight and angular compared to the literally hand-drawn single sided boards of olde but that doesn't make it any easier to do - must have taken ages to route that! (given the minor differences all over the place I don't believe it is the work of an autorouter)

I hope it works and am looking forward to seeing it cased up

[tszaboo]

Looks great.

Thanks

I am slightly disappointed that your 'organic' traces are very straight and angular compared to the literally hand-drawn single sided boards of olde but that doesn't make it any easier to do - must have taken ages to route that! (given the minor differences all over the place I don't believe it is the work of an autorouter)

I hope it works and am looking forward to seeing it cased up

It is difficult to keep isolation distances with arced/curved traces. If you have an arc with the radius of 2mm, another arc needs to be 1.65mm or similar radius. And yes, routing it took hours, several iterations.

- 0% LEDs

Oh nooooooo! 

The STM32 board has two LEDs, one is blue and can be controlled. So there is Blinkenlights on it after all.

So I just got the diodes and spacers yesterday.
QMK doesnt seem to be 100% stable on this microcontroller. It compiles, uploads.
Then the first kepress starts spitting qwertqwertqwert into notepad ad infinium, until I press another button.
And first row dowsnt work at all.
This is slightly disappointing, probably need to spend some time fixing the firmware.

[tszaboo]

I'm writing this post on the KB. It works, with a few bugs.
A9 and B2 pin of the controller should be left open, because it is used for other functions. There might be some workarounds, but it is time consuming to track them down.
For now, I've disconnected A9 and connected a row to pin A10. And desoldered a resistor from the microcontroller board, that might screw with the bootloader sometimes.
I'm going to clean the board on Monday with Isopropyl alcohol(I dont have that at home), and post images.
I also have to learn how to use this effectively, my WPM went down to almost nothing on first try.
The board sounds great with these Panda switches, the space is a bit thock.

[suicidal_orange]

I'm writing this post on the KB. It works, with a few bugs.

That's great to hear, a first edition board should have a couple of jumpers

[MisterWaffle]

Well done with keeping the PCB design clean. Always nice to see some more ARM boards out there. 

[Gorbon]

A9 and B2 pin of the controller should be left open, because it is used for other functions.

According to the pin-out from their GitHub page, B2 seems to be connected to the BOOT button, but I would've expected A9 to work. I also have the same board and building (or at least trying to build) a handwired prototype around it.

It works, with a few bugs …desoldered a resistor from the microcontroller board, that might screw with the bootloader sometimes.

If it's not too much trouble, could you mention what bugs you encountered and which resistor you had to desolder? Apparently there are a lot of counterfeits of this board (if it doesn't have the WeAct logo in the back, it certainly is), that could perhaps have all sorts of hardware issues.

I also have to learn how to use this effectively, my WPM went down to almost nothing on first try.

Personally, I find non-split, ortholinear boards that have no separation between the two halves quite unergonomic. They force you to twist your wrists outwards and rotate your palms to face the keyboard, in order to align your fingers with the keys. Maybe it's because I have large hands, others seem not to mind as much. Hopefully it'll work nicely for you.

[tszaboo]

A9 and B2 pin of the controller should be left open, because it is used for other functions.

According to the pin-out from their GitHub page, B2 seems to be connected to the BOOT button, but I would've expected A9 to work. I also have the same board and building (or at least trying to build) a handwired prototype around it.

(Attachment Link)

It works, with a few bugs …desoldered a resistor from the microcontroller board, that might screw with the bootloader sometimes.

If it's not too much trouble, could you mention what bugs you encountered and which resistor you had to desolder? Apparently there are a lot of counterfeits of this board (if it doesn't have the WeAct logo in the back, it certainly is), that could perhaps have all sorts of hardware issues.

I also have to learn how to use this effectively, my WPM went down to almost nothing on first try.

The button is BOOT0, B2 is BOOT1. It has a pulldown resistor, (that I've removed) that was messing with the circuit.
So that's R5 here:
https://stm32-base.org/assets/pdf/boards/original-schematic-STM32F411CEU6_WeAct_Black_Pill_V2.0.pdf
Not the best way to handle it, but it works. And it jumps to bootloader reliably if I press column IK.Down and the reset.
Next time I just skip this pin.
A9 doesn't work, because it is used as USB-OTG detect. There seems to be a workaround, that didn't work for me.
https://github.com/qmk/qmk_firmware/issues/7855

Personally, I find non-split, ortholinear boards that have no separation between the two halves quite unergonomic. They force you to twist your wrists outwards and rotate your palms to face the keyboard, in order to align your fingers with the keys. Maybe it's because I have large hands, others seem not to mind as much. Hopefully it'll work nicely for you.

I have to agree. The layout seems to be nice for FPS games, where one hand is on the KB other on the mouse. I'll try it for PCB design, where the concept is the same. Though I usually have to enter a lot of numbers when doing that, this is not the KB for that.
Programmers seems to like this layout. But I think the best use case for this layout is wireless for home cinema. Where you dont want to type too much and having a small KB is beneficial. Did anyone ever make a wireless 40%?

[tszaboo]

So assembled, tested, debugged, tried.
I have 3 extra sets of PCBs, plates, screws, stabs ,diodes and microcontrollers.