change F3 from CM1 to CM4
add debug flags for FFT
add bandpass filter
add different filtering apply function
add feature DYNAMIC_FILTER
replace mode GTUNE with DYNAMIC FILTER
move gyro frequency analysis into gyro loop instead of own task
CF/BF - Save 1280 bytes of flash and delete over 300 lines of code by
removing giant switch statement and static variables in the CLI.
fix
CF/BF - Remove some conditional code because there is a copy of every PG
so it will always be found*.
* unless a developer makes an error in the value table by specifying an
invalid PG.
CF/BF - Remove final usages of getCurrentAndDefaultConfigs to save more
flash space.
CF/BF - Change CLI command 'dump showdefaults' to simply 'dump defaults'.
* This will save a few bytes of flash space due to string re-use of the
'defaults' string.
CF/BF - Save a few more bytes by re-using the 'defaults' string again.
CF/BF - Fix CLI not being size optimized.
* The file had moved but the makefile was not updated to reflect that.
CF/BF - Add 'fc_init.c' and 'settings.c' to the size optimizations list.
CF/BF - Add some config source files to the size-optimized list.
CF/BF - De-duplicate CR/LF code by using appropriate methods. Promote
additional string re-use.
* This saves quite a lot of space when the file is size optimized, but grows it otherwise.
CF/BF - Further flash savings by removing more CR/LF on the
cliDump/cliDefaultPrintf format strings.
* It turns out that all the format strings ended in CR+LF so they could
all be removed.
CF/BF - Fix LF+CR being used instead of CR+LF for further flash savings.
CF/BF - Fix typo in max7456DmaInProgress. Always define an
implementation even when there is no TX-DMA.
CF/BF - Ensure that max7456 communication does not disrupt RTC6705
communication when using the same SPI port for both.
CF/BF - Port RTC6705 to use the VTX API.
* Support VTX channel/band/power on OSD for ALL vtx APIs - previously it only supported RTC6705.
* Remove all FEATURE_VTX usage.
* Remove all `#define VTX` usage.
Note that due to the async nature of tramp/smartaudio it's not possible
to know the current band/channel after requesting a change (as the
request hasn't been processed) so it makes no sense to try and save the
config - and on the tramp/smartaudio vtx the vtx maintains it's state.
So on an RTX6705 equipped FC the user should now MANUALLY save the
config using the save config stick position.
CF/BF - Move CMS for RTC6705 into vtx_rtc6705.c for consistency with
other VTX CMS menus.
CF/BF - Update RTC6705 via SOFT SPI to use VTX API.
* Achieved by simply aligning the API at the driver level and removing
all legacy conditional compilation checks.
CF/BF - Use new IO for RTC6705 CLK Hack
CF/BF - Port VTX button from CF v1.x.
Features:
* Channel Cycle
* Band Cycle
* Power Cycle
* Save Settings
* Works with any VTX using VTX API.
CF/BF - Allow support for Internal and External VTX - External takes
precedence when configured.
CF/BF - Fix display of VTX power in CMS.
* Issue occured when CMS was transitioned from integer to list of
strings.
CF/BF - Disable SMARTAUDIO and TRAMP on the SPRacingF3NEO due to flash
size.
CF/BF - Cleanup const usage of bass-by-value parameters. Cleanup
naming.
CF/BF - Cleanup naming. Removing noise-words. Improving consistency.
CF/BF - Improve readability by further naming cleanups.
CF/BF - Remove some magic numbers in the RTC6705 VTX code.
CF/BF - Use braces on some procedural macros.
CF/BF - Improve code readability by renaming rtc6705SetChannel to
rtc6705SetBandAndChannel.
CF/BF - Remove editor mishap.
CF/BF - Make the rtc6705Dev variable static.
CF/BF - remove duplicate state variablertc6705Dev_t - vtxDevice_t had
all we needed. Use brackets on some procedural style macros. Make all
the vtx vtable static const.
CF/BF - making some declarations const.
CF/BF - Fix vtx_power maximum value (out by one).
the flight controller code.
Likely the dependencies can be further improved. This is a
minimal-impact solution while there are a lot of other Betaflight PR's
still pending.
need implement fake eeprom & fake IO
need implement fake system function
can compile, stuck in isEEPROMContentValid()
EEPROM in memory work
EEPROM as file should work
fix some complie warning
MSP over TCP work (add dyad.c)
a little clean up
fix FLASH_CONFIG_Size in ld script & implement some pwmout
IO to simulator work!! need to check direction & scale!!
can fly but Gyro buggy
move dyad.c
fix busy-loop (limit to max 20kHz)
can simulatie in different speed now! (hard code)
add option for IMU calculation
add README.md
move dyad.c and fix F3 overrun the flash size
explanation SITL in README.md and reuse CFLAGS, ASFLAGS
- Specify source files in variables instead of manually writing
the make file rules, this removes a lot of duplication
- Use separate folders for all tests allowing for tests and common
source files to be built with different settings (defines)
```
+---------+
| FS-iA6B |
| |
| Ser RX |---|<---\ +------------+
| | | | FC |
| Sensor |---[R]--*-------| Uart TX |
+---------+ +------------+
```
R = 10Kohm, Diode 1N4148 or similar.
Both uart tx and rx channels are used so it's not possible to use the spare pin for rx of something else.
Configure the serial port like this to get both serial rx and ibus telemetry on the same port:
```
serial 1 4160 115200 57600 115200 115200
```
It is still possible to run the serial rx and ibus telemetry on two uarts like before, an example:
```
serial 1 64 115200 57600 0 115200
serial 2 4096 115200 57600 115200 115200
```
* Initial preparation for cleaning up multiple files per MCU
* Combined DSHOT for F3 and F4 (F7 todo)
* Build fixes (mixed some adjustments needed in transponder_ir.c)
* Fix F7 warnings/bugs
* Updated timer periods for transponder.
Fixed FURY including old pwm_output files.
* Added method to calculate prescaler and period on the fly for varying clock speeds.
This can be used for overclocking :)
* Consolidated led_strip
* Moved led strip to use system clock for timing (allowing overclocking)
* Fixed incorrect channel, and removed comparison warning.
* More cleanup
* Fix KIWIF4 target
* Cleaned up target.mk files
This prevents the git command log from command being executed during
every use of the variable REVISION and so speeds up the build by ~5sec
on my machine
* add travis directorie cache for arm sdk
* Travis. Test if make all in one job would work. And how slow/fast.
* Travis. Check that the old method with a couple specific targets still works.
* Rebased master and build all targets again.
* Make verbosity level 0 even less verbose, was still too much for travis in one mega-jumbo-all-target build.
* Even less output from make in verbosity level 0. Some Makefile symbol name cleanup too.
* arm_sdk_install did not check if allready downloaded and installed. Added checks and deps. Travis caching might work better now...
* Build all 56 targets to test Travis caching perf.
* Build all 56 targets to test Travis caching perf. Again. Now with correct yml arrsy syntax.
* Added installation of ccache and also to Travis cache. Caching the cache.
* Travis. Remove some unused APT libs, to see if this affects caching and build times.
* Travis. YAPT (Yet another perf trial). Remove all apt package install.
* Disaster. Put back installation if build-essential.
* Disaster still. Put back installation if zliblg-dev.
* Disaster, again. Put back installation of libc6 .
* Travis. Minimizing APT installation as it not be cached, issue 5876 on travis-ci github.
* Travis. Added comment on target specification.
Michael Jakob
jflyper
Anders Hoglund
Michael Keller
Petr Ledvina
rav
Sami Korhonen
Martin Budden
Dominic Clifton
cs8425
Magnus Ivarsson
J Blackman
blckmn
borisbstyle
Raphael Coeffic
Vlad Adamenko
blckmn
atomiclama
Bas Delfos
Steffen Windoffer
Konstantin Sharlaimov (DigitalEntity)
nathan
4712