feat: RGB leds supports

[raphaelcoeffic]

Summary of Changes:

• Adds support for WS2812 LEDs (taranis targets only for now)
• New custom function allowing to select a LUA script that controls a LED strip (see Add basic RGB leds script support edgetx-sdcard#115)
• stm32_pulse_driver modified to initialise DMA only once during init
• Improved pulse driver support for complementary channels

[Eldenroot]

CZ:

#define TR_SF_MOD_BIND "Modul Bind"
#define TR_SF_RGBLEDS "RGB led"

[raphaelcoeffic]

@mha1 may I ask you to check if the changes in the timer / pulses drivers still work for all modules / protocols?

[mha1]

@raphaelcoeffic good one. batting my pitch to a home run.

I'll do some spot checks on what I think I can do . §xternal/internal MPM, external ELRS, PWM out, softserial in. That's about it I think .

[mha1]

Including quick functional tests:
external ELRS control, telemetry, ELRS LUA script-> ok
internal MPM control, telemetry, HOTT LUA script -> ok
external MPM control, telemetry, HOTT LUA script -> ok
module bay softserial-in on SPort pin -> ok

Just looking at signals, no functional test:
DSM2 out (all 3 sub protocols) at CPPM pin -> looks ok
XJT D16 out at CPPM pin -> idle low bitstream available at 11Hz, -> should be ok
PPM out at CPPM pin -> single pulse at 22Hz with correct idle level but no channel data -> not ok

Hope that helps

[raphaelcoeffic]

@mha1 should be fixed now.

[3djc]

Tested on X7

• Flash internal XJT : ok
• Flash receiver through SPORT : ok after fix: PCBREV detect not working #3922
• Use XJT : ok, both control and telemetry
• Flash external MPM (STM32): ok
• Multi telemetry: ok

[mha1]

Just looked at PPM for now. Channels are back but frame rate is fixed at 14ms regardless of the set PPM frame length.

This was done with PPM frame length set to 40ms

[raphaelcoeffic]

Channels are back but frame rate is fixed at 14ms regardless of the set PPM frame length.

Do you happen to have another module turned ON? (internal?)

[raphaelcoeffic]

For example, with a internal XJT module on QX7:

The internal module sends every 9ms, but as the PPM pulse is not finished, every other slot is skipped, so that you get 2x9ms = 18ms.

[mha1]

It looks a lot like MPM's 7ms. And indeed it was on. I was fooled because there is no more GUI indication (yellow) if there is a module set. But at least consistent. No indication for internal and external modules. Set internal module to off, all good now.

[mha1]

Was the highlighting Internal RF and External RF is set to other than off scraped on purpose?

[raphaelcoeffic]

Was the highlighting Internal RF and External RF is set to other than off scraped on purpose?

Probably a question for @JimB40 and @philmoz

[JimB40]

If on purpose you would see PR with rather long discussion. Most of UI behaviour that differs from 2.8 is due to some 'fixes'.

[mha1]

@raphaelcoeffic alll the serial out on CPPM/SPort and (soft and hardware) serial in on SPort seems to work too

[pfeerick]

Of concern: anything that uses soft serial, i.e. pulse drivers

[richardclli]

Hey, I need this one and see if it works in my PL18 branch.

[pfeerick]

Neither X9D+2019 nor Boxer are booting with the GitHub builds for this PR.

[raphaelcoeffic]

Neither X9D+2019 nor Boxer are booting with the GitHub builds for this PR.

Should be fixes now.

[pfeerick]

TX16S, X9D+2019, X9D, Zorro: All connected to pre-bound/bound as necessary RXs with internal MPM/ISRM/XJT, telemetry and control link good. Mix of Ghost, CRSF, Tracer, XJT/XJTLite and MPM external modules used across all the radios with no issues with github builds.