Drivers led implement driver for tlc59731 rgb controller

[javad123javad]

This patch series add driver support for TI TLC59731 RGB controller, along side the changes applied for STM32WB5MM_DK board and a demo samples based on this board.

[ajarmouni-st]

STM32WB5MM_DK doc should be updated to inform the user that RGB and IR LEDs share MCU resources, & so to use the RGB LED, JP5 must be ON and JP4 OFF (GPIO H1 becomes the chip select for the RGB device on spi1).
(from section 7.5.5 in https://www.st.com/resource/en/user_manual/um2825-discovery-kit-with-stm32wb5mmg-module-stmicroelectronics.pdf)

[simonguinot]

Hi @javad123javad,

Thanks for contributing support for the TLC59731 controller !

I started to review the driver but at some point (reading the datasheet) I realized that the TLC59731 is a LED strip controller (i.e. which can be cascaded). If it is correct, then you need to migrate this driver into the led_strip subsystem.

Additionally, while it is possible to use the CPU to send pulses through a GPIO (as you did), you may prefer to use an SPI controller instead. The idea is to program the SPI controller to send pulses through the MOSI pin. This method allows to offload the CPU. See the ws2812_spi driver as an example. I am not sure if it is possible here, but please check it.

[simonguinot]

In the TLC59731 controller specification I read that the brightness granularity of each channel is fixed to 256. So what is the purpose for the {min,max}_brightness properties ?

[javad123javad]

The overall Grayscale brightness of the RGB led is controlled by this properties. Furthermore, according to section 8.4 of the datasheet, it is possible to control the brightness of each channel:

The TLC59731 can adjust the brightness of each output channel using a pulse width modulation (PWM) control
scheme. The PWM data bit length for each output is 8 bits. The architecture of 8 bits per channel results in 256
brightness steps, from 0% to 99.98% ON-time duty cycle

[javad123javad]

Hi @javad123javad,

Thanks for contributing support for the TLC59731 controller !

I started to review the driver but at some point (reading the datasheet) I realized that the TLC59731 is a LED strip controller (i.e. which can be cascaded). If it is correct, then you need to migrate this driver into the led_strip subsystem.

Additionally, while it is possible to use the CPU to send pulses through a GPIO (as you did), you may prefer to use an SPI controller instead. The idea is to program the SPI controller to send pulses through the MOSI pin. This method allows to offload the CPU. See the ws2812_spi driver as an example. I am not sure if it is possible here, but please check it.

Hi @simonguinot
Thanks a lot for your comments. Before using GPIOs I have tried to use SPI to implement the driver, however it was not working. However, I will try to write the new version of the driver based on SPI.

[simonguinot]

Hi @javad123javad. Please close this PR and open a new one when it is ready.

[javad123javad]

Hi @javad123javad. Please close this PR and open a new one when it is ready.

Hi @simonguinot
It is almost finished. Is there any other changes that should I apply in a new PR?

[simonguinot]

Hi @javad123javad. Please close this PR and open a new one when it is ready.

Hi @simonguinot It is almost finished. Is there any other changes that should I apply in a new PR?

From what I can see it is definitely not ready. I can see a struct led_driver_api defined in the driver instead of a struct led_strip_driver_api. So there is still some work to do. Please close this PR. Reviewers don't need to be notified at each update while the development is still ongoing.