mirror of
https://github.com/Keychron/qmk_firmware.git
synced 2024-11-23 00:47:02 +06:00
136 lines
5.9 KiB
Markdown
136 lines
5.9 KiB
Markdown
|
## Hardware Information
|
||
|
|
||
|
The YANG HHKB BLE controller design is similiar to hasu's
|
||
|
controller. Most pins are compatiable.
|
||
|
|
||
|
**MCU**: ATmega32U4
|
||
|
**Bluetooth**: MDBT40 (nRF51822-based), with Adafruit Bluefruit LE UART Friend firmware.
|
||
|
**Power**: 3.3V
|
||
|
**CPU Frequency**: 8MHz
|
||
|
**Bootloader**: Lufa MassStorage
|
||
|
|
||
|
## Pin usage
|
||
|
|
||
|
| Description | HASU pin usage | YANG mod changed |
|
||
|
|:------------------------------------ | ---------------------- | -------------------------- |
|
||
|
| ~KEY: Lo(0) when key is pressed | PD7 input(with pullup) | |
|
||
|
| Hysteresis: Hi(1) if key was pressed | PB7 output | |
|
||
|
| Row selector bit0 | PB0 output | |
|
||
|
| Row selector bit1 | PB1 output | |
|
||
|
| Row selector bit2 | PB2 output | |
|
||
|
| Col selector bit0 | PB3 output | |
|
||
|
| Col selector bit1 | PB4 output | |
|
||
|
| Col selector bit2 | PB5 output | |
|
||
|
| Key unable | PB6 output | |
|
||
|
| Switch power | PD4 output | PD6 output (PMOS FET) |
|
||
|
| Bluetooth UART Rx | PC4 input | PD2 |
|
||
|
| Bluetooth UART Tx | PC5 output | PD3 |
|
||
|
| Bluetooth power | | PD5 output (low: power on) |
|
||
|
| LED 0 | | PF4 |
|
||
|
| LED 2 | | PF1 |
|
||
|
| LED 4 | | PF0 |
|
||
|
| Unused for PRO2 | PC6 | |
|
||
|
| Unused for PRO2 | PC7 | |
|
||
|
| Inner USB power | | PF7 |
|
||
|
|
||
|
## How to flash LUFA MassStorage bootloader on Linux
|
||
|
|
||
|
The FAT filesystem on Linux very often cannot flush the write cache,
|
||
|
leading to broken firmware in the flash.
|
||
|
|
||
|
We can use `dd` to write to the virtual block storage directly to
|
||
|
bypass the vfs layer.
|
||
|
|
||
|
```
|
||
|
dd if=FLASH.bin of=<path of virtual block device> seek=4
|
||
|
```
|
||
|
|
||
|
Skip 4 sectors because the default sector size of the virtual device
|
||
|
and dd is 512 bytes and the emulated flash file starts at 5th sector.
|
||
|
|
||
|
## How to find the path of the virtual block device
|
||
|
|
||
|
After the keyboard boots into flash mode, on Linux system you should
|
||
|
be able to find the block device in `dmesg` logs.
|
||
|
|
||
|
For exmaple if you type
|
||
|
|
||
|
```
|
||
|
sudo dmesg
|
||
|
```
|
||
|
|
||
|
You should find something like
|
||
|
|
||
|
```
|
||
|
[357885.143593] usb 1-1.4: USB disconnect, device number 24
|
||
|
[357885.627740] usb 1-1.4: new full-speed USB device number 25 using xhci_hcd
|
||
|
[357885.729486] usb 1-1.4: New USB device found, idVendor=03eb, idProduct=1962, bcdDevice= 0.01
|
||
|
[357885.729492] usb 1-1.4: New USB device strings: Mfr=0, Product=0, SerialNumber=0
|
||
|
[357885.745620] SCSI subsystem initialized
|
||
|
[357885.746712] usb-storage 1-1.4:1.0: USB Mass Storage device detected
|
||
|
[357885.746818] scsi host0: usb-storage 1-1.4:1.0
|
||
|
[357885.746919] usbcore: registered new interface driver usb-storage
|
||
|
[357885.747689] usbcore: registered new interface driver uas
|
||
|
[357886.766755] scsi 0:0:0:0: Direct-Access LUFA Bootloader 0.00 PQ: 0 ANSI: 0
|
||
|
[357886.773216] scsi 0:0:0:0: Attached scsi generic sg0 type 0
|
||
|
[357886.777474] sd 0:0:0:0: [sdx] 134 512-byte logical blocks: (68.6 kB/67.0 KiB)
|
||
|
[357886.780300] sd 0:0:0:0: [sdx] Write Protect is off
|
||
|
[357886.780302] sd 0:0:0:0: [sdx] Mode Sense: 00 00 00 00
|
||
|
[357886.783113] sd 0:0:0:0: [sdx] Asking for cache data failed
|
||
|
[357886.783114] sd 0:0:0:0: [sdx] Assuming drive cache: write through
|
||
|
[357886.842676] sdx:
|
||
|
[357886.859528] sd 0:0:0:0: [sdx] Attached SCSI removable disk
|
||
|
```
|
||
|
|
||
|
The `sdx` is the block device name and the full path is at `/dev/sdx`
|
||
|
The above flash command will become
|
||
|
|
||
|
```
|
||
|
dd if=FLASH.bin of=/dev/sdx seek=4
|
||
|
```
|
||
|
|
||
|
## Adafruit Bluefruit LE UART configuraton
|
||
|
|
||
|
The default baud rate used by the firmware is 76800 although adafruit
|
||
|
do not recommend using higher baudrates than 9600 because the nRF51
|
||
|
UART can drop characters.
|
||
|
|
||
|
Double speed mode to get more accurate async reading because the F_CPU
|
||
|
speed is 8MHz.
|
||
|
|
||
|
## Power saving mode design
|
||
|
|
||
|
Power saving is only enabled when USB is detached and using battery
|
||
|
power. Here we define several levels of power saving mode, each saves
|
||
|
more power but takes longer to resume operation.
|
||
|
|
||
|
1. Level 1: idle mode is activated after a short configurable time
|
||
|
(MATRIX_POWER_SAVE_TIMEOUT_MS) MCU is put into sleep mode and only
|
||
|
scan the matrix per 15ms. PORTB pins are set to input with pull-up
|
||
|
to save power. Sensing PCB is powered down between scans.
|
||
|
|
||
|
2. Level 2: after idling for longer (MATRIX_POWER_SAVE_TIMEOUT_L2_MS)
|
||
|
we entry this state. Matrix scan is skipped until the time lapses
|
||
|
900ms.
|
||
|
|
||
|
2. Level 3: sleep mode is activated after a longer timeout
|
||
|
(MATRIX_POWER_SAVE_TIMEOUT_L3_MS) Bluetooth module is powered down.
|
||
|
|
||
|
## Battery reading
|
||
|
|
||
|
VBAT is connected to AIN6 pin on the MDBT40 module and the AREF pin is
|
||
|
the reference voltage. Doing a ADC with AT+HWDAC=6 will return the
|
||
|
difference between VBAT and VREF.
|
||
|
|
||
|
It seems when fully charged the ADC read is 550. Likely VREF is 3311mV
|
||
|
and the fully charged VBAT is thus 3861mV.
|
||
|
|
||
|
Enable battery service with AT+BLEBATTEN=1 first then we can update the
|
||
|
battery level by using AT+BLEBATTVAL=%d
|
||
|
|
||
|
## References
|
||
|
|
||
|
* https://github.com/joric/qmk/wiki/hhkb_ble
|
||
|
* https://github.com/tomsmalley/custom-topre-guide
|
||
|
* https://github.com/abcminiuser/lufa/blob/master/Bootloaders/MassStorage/Lib/VirtualFAT.h
|