According to the data sheet, writing to a Control A register's ENABLE
bit will trigger write synchronization and set SYNCBUSY's ENABLE bit
which will be automatically cleared by the hardware once the write
operation is complete.
It is necessary to wait until SYNCBUSY's ENABLE bit is clear.
Reading synchronized registers before that returns previous values.
Writing synchronized registers before that drops the write and generates
an error value in another register.
The data sheet recommends polling the ENABLE bit in this situation.
Reviewed-by: Matheus Afonso Martins Moreira <matheus@matheusmoreira.com>
GitHub-Pull-Request: https://github.com/joeycastillo/Sensor-Watch/pull/450
References: SAM L22 Family Data Sheet §§ 14.3.2, 14.3.3, 14.3.4, 14.3.5
This patch enhances the Sensor Watch simulator with a temperature input
allowing users to simulate and test the temperature sensor.
Reviewed-by: Matheus Afonso Martins Moreira <matheus@matheusmoreira.com>
GitHub-Pull-Request: https://github.com/joeycastillo/Sensor-Watch/pull/426
Improves handling of months in the Sensor Watch
by computing whether the given year is a leap year.
Reviewed-by: Matheus Afonso Martins Moreira <matheus@matheusmoreira.com>
GitHub-Pull-Request: https://github.com/joeycastillo/Sensor-Watch/pull/434
* Sensor Watch Pro board definition
* Disable hot-plugging on SWCLK
* Add rainbow test
* Move interrupt mapping to board config
* New color definition for Pro boards in makefile
This makes movement_play_signal synchronous when in LE mode, despite
using the underlying asynchronous API. It's a bit of a hack, but it
should work well enough for now.
This also moves the enabling/disabling of the buzzer into the
movement_play_signal function, so that watch faces no longer have to do
it.
* Introduce shell module for basic serial shell with argument parsing
* Introduce shell_cmd_list module for basic compile-time command
registration
* Harden USB handling to hang less and drop fewer inputs
- Service tud_task() with periodic TC0 timer interrupt
- Service cdc_task() with periodic TC1 timer interrupt
- Handle shell servicing in main app loop
- Add a circular buffering layer for reads/writes
* Change newline prints to also send carriage return
* Refactor filesystem commands for shell subsystem
* Introduce new shell commands:
- 'help' command
- 'flash' command to reset into bootloader
- 'stress' command to stress CDC writes
Testing:
* Shell validated on Sensor Watch Blue w/ Linux host
* Shell validated in emscripten emulator
* Tuned by spamming inputs during `stress` cmd until stack didn't crash
These keys are the shortcuts to "press" the alarm, light and mode
buttons. However, they prevent these letters from being input in the
debug console to send filesystem commands. Strangely, there was already
some code to allow typing these letters in the console output, but not
in the input.
When using the simulator, I encountered cases where the light would become stuck on, and the watch
would be unresponsive. In particular, this would occur when pressing the light button on the
sunrise sunset watch face.
I appears that this is caused by a divergence in out the callback mask is interpreted by the
hardware interface, and in the simulator in the following function.
void watch_rtc_disable_matching_periodic_callbacks(uint8_t mask)
In particular, a mask of 0xFE is intended to disable all except the 128hz callback at index 0,
but instead disables all except the 1hz callback at index 7 in the simulator.
* initial commit, added opt3001 light meter test app
* tested working light meter board, i2c communication still has issues
* fixed i2c; rudimentary lightmeter works!
* added aperture priority ui
* added aperture priority ui
* added README
* adjusted cal
* fixed bugs (HI shutter speed, lux mode toggle)
* made it possible to advance to the next face
* initialized lux variable
* lowered tolerance for HI and LO
* Changed EV display from always showing EV100 to showing EV[iso setting]
* dont display old ev when ISO changes
* changed mode and light behavior
* updated readme
* fixed indentation
* made lightmeter display logic more consistent
* made lightmeter display logic more consistent
* reverted rules.mk (for merge into upstream)
* reverted rules.mk (for merge into upstream)
* removed OPT3001 PCB model
* made lux mode default, corrected timeout behavior
---------
Co-authored-by: Christian Chapman <user@debian>
* buzzer sequences: first draft, does not work on hardware yet (but in simulator)
* buzzer sequences: add changes to movement.c
* buzzer sequences: add demo face to Makefile
* buzzer sequences: fix problem of interrupted sounds. Add logic for repeating sub sequences. Tidy up (move logic to watch_buzzer files, remove buzzer_demo_face)
* buzzer sequences: tidy up even more
* buzzer sequences: disable registering a 32 Hz tick callback for watch faces, so it will be used exclusively by the buzzer sequences functionality
* buzzer sequences: add callback slot functionality to watch_rtc and make watch_buzzer use it. Switch internal buzzer sequences tick frequency to 64 Hz. Revert changes to movement.c
* buzzer sequences: fix parameter sanity check in watch_rtc code
* buzzer sequences/watch_rtc: optimize calling tick callbacks in RTC_Handler
* buzzer sequences/watch_rtc: fix error in calling callback functions
* buzzer sequences: revert changes to watch_rtc logic. Instead, use TC3 as the source for timing the sound sequences.
* buzzer sequences: fix frequency of callback
* buzzer sequences: integrate changes from PR #162 (set both CCBUF and PERFBUF for correct buzzer tone)
* Intermediate changes
* Databank working
* Main commit for precision timing
First version where all functions are supposed to be working
* Fix math error in nanosec. File storage for location.
* Remove obsolete comments
* Missing page name on pages rotation - thanks to jeremy
* Delete file.diff
* Cleanup+tempchart
1) finetune must always reset last calibration time when doing non-0 time correction, even when you are not applying ppm correction.
2) Dithers over 31 periods not 10, more resolution with still no risk of overflow
3) Minute-boundery finetune fix. I also just got this 1-minute error after finetune...
4) Write frequency calibration value in 1 operation rather than 2. All RTC writes must be single operations to avoid partially correct data.
5) Some code cleanup
6) Tempchart face is added for temperature statistics
* Update set_time_hackwatch_face.c
* Math error in display code of finetune, allow to update correction time even without correction - by long alarm press
* Increase reliability of stopping & starting RTC timer
As it's quite dangerous operation
* hackwatch - days adjust down fix by long alarm
* unify style
* More comments & last style change
* Simulator support
RTC operations (watch_rtc_enable and watch_rtc_freqcorr_write) are in common libs.
* Unicode fix
* Crystal aging is now adjustable (AA page in nanosec - annual aging, ppm/year)
Aging is baked into fixed offset every time finetune is performed, as it relies on last adjustment time.
* Blink on non-0 page every minute in finetune to measure clock error
* Rolling back private changes
* Cleanup
* Cleanup
* Quality of life changes in nanosec
1. Does not calculate & apply ppm correction if less than 6 hours passed since previous adjustment (as it gives very high correction values which are unrealistic and unhelpful)
2. Idle timeout resets to face 0 only if no correction was made
* unify style
* Fix low-power errors in nanosec infrastructure, faster display in finetune
* Merge fix
* unify style
Co-authored-by: Jeremy O'Brien <neutral@fastmail.com>
Co-authored-by: joeycastillo <joeycastillo@utexas.edu>
