Well first of all here are (4 out of the 5) boards I received from JLCPCB
Nothing to add here: they just look and work fine, as expected.
Since I heavily copied the original design and placement the answer is yes, of course:
However, even though the board was OK, the 3D printed case was not at the beginning:
-
the cut for the USB slot was for a microUSB, but I intentionally went for a typeC
-
since I don't have a 3D printer, I had the case printed externally. It turned out that the board's holding parts were a bit too tall (so that the top part of the case was not closing properly)
Both of these problems were solved manually with the Dremel:
The yellowish layer you see on the sensor is its original cover. I did not removed it yet because I'm still waiting for the trackball to arrive so until then I'll keep the protection on.
Only a major fix was required compared to the original design:
- do not mount C13 and C14 since otherwise the oscillator does not start (verified with oscilloscope)
Then there is also another one, but this is due to the fact I was not able at all to find the ADNS-5050 component, so I replaced it with ADNS-5020 which is pin-to-pin compatible. In case you do the same then do not install also R3 and R4.
You can easily download QMK firmware from its repository, checkout at the master branch and build it with:
util/docker_cmd.sh make clean && util/docker_cmd.sh qmk compile -kb ploopyco/trackball_nano/rev1_001 -km default:production
The production keyword was added because we're dealing with a completely new MCU which does not have a bootloader yet (well, not the one QMK expects) so we build both the FW and the bootloader at once.
Yes, I used Docker. You can run the same commands without it, but perhaps you will need to install required packages.
Luckily there is no need for an external fancy programmer since the board can be programmed with avrdude using a RaspberryPi:
WARNING
In order not to uncorrectly power either the ploopy nano or the RPi, I strongly suggest to:
- disconnect the RPi from any power supply
- connect the RPi to the Ploopy
- connect power to the RPi and turn it on
Here is how I did the connection (but you can change it on the RPi side - check avrdude instructions for this):
| RPi pin # | wire color (RPi/Ploopy) | Ploopy TP |
|---|---|---|
| 1 | violet/red | 1 |
| 6 | blue/blue | 6 |
| 12 | gray/violet | 3 |
| 16 | green/orange | 2 |
| 18 | yellow/yellow | 4 |
| 32 | white/green | 5 |
(I added also the wirecolors in the table above so that you can also double check with the pictures).
Once you got the connection done and RPi turned on, you can SSH into it and the issue the following command to program your board:
sudo avrdude -p atmega32u4 -C ~/avrdude_gpio.conf -c pi_1 -v -U flash:w:ploopyco_trackball_nano_rev1_001_default_production.hex:i -U lfuse:w:0x5E:m -U hfuse:w:0x99:m -U efuse:w:0xC3:m -U lock:w:0x3F:m
WARNING
Of course care should be taken also when powering down:
- turn off the RPi first
- disconnect all power cables from it
- disconnect Ploopy
The ATMEGA32U4 has some internal fuses that are used to set some configuration in a non-volatile way (i.e. not lost by resetting the board). These configurations include the crystal oscillator and by default the microcontroller is set to use the internal RC one (check the datasheet for this):
- L-fuse = 0x52
- H-fuse = 0x99
- E-fuse = 0xFB
Internal RC oscillator is not OK for use with USB since it is not stable enough for that purpose and moreover the FW expects something at higher speed. That's why we set:
- L-fuse = 0x5E
- H-fuse = 0x99
- E-fuse = 0xC3
Of course these changes were also suggested in the Ploopy project's wiki: https://github.com/ploopyco/classic-trackball/wiki/Appendix-D%3A-Programming-QMK-on-Older-Ploopy-Devices#flash-the-bootloader
and elsewhere, like: https://www.reddit.com/r/olkb/comments/8sxgzb/replace_pro_micro_bootloader_with_qmk_dfu/)