Initialization for ES3N28P

[uraich]

I bought a LCD Wiki ES3N28P board with a 320x240 LCD display controlled by an ILI9431 chip. I am trying to get a hello world UI running on this board.
I tried with this initialization file
basic.py

In display_driver_framework is saw def init(self, type=None): tries to import _ili9431_init.py which does not exist. There is however _ili9431_init_type1.py and _ili9431_init_type2.py, which are imported if type is 1 or 2 (where does type come from?)

In my case backlight is active high but
self._backlight_pin.value(not backlight_on_state) switches the backlight off.
I modified this, removing "not" to switch backlight on.
The program runs but I only see a small vertical line.
Can I use the driver to try and draw a line or a colored rectangle?
I wrote a small pin definition file, where I copied the pin allocations from the manual and modified the pin definitions correspondingly (hw_es3n28p,py)
I do not see _flush_cb being called.
Can you give me any hints how I can debug this problem? Can I draw some simple objects (e.g. filled rectangle) to test the driver?
I tried the ili9341 driver with standard micropython using my pin definition file and there things work (colors are not yet ok, but otherwise things work fine). So, I am confident that the pin definitions are correct.

[resnbl]

Some suggestions for your basic.py:

display = myili9341.ILI9341(
    data_bus=display_bus,
    display_width=240,
    display_height=320,
    reset_pin=None,
    backlight_pin=LCD_BL,
    color_space=lv.COLOR_FORMAT.RGB565,
    color_byte_order=myili9341.BYTE_ORDER_BGR,    # <--- or _RGB to correct color issues
    rgb565_byte_swap=True
)

display.set_power(True)
display.init(1)   # <--- or (2) if (1) doesn't help

The init(1) call causes a number of commands to be sent to the display controller for initialization. You need this.

[uraich]

I saw what the init command is doing. For the moment I only get a white screen with a small black vertical line. The color problem was on the driver with standard, non-lvgl micropython. init-type 2 gives the same result.
I expected the flush_cb to be called at a certain moment, but I do not see this.
display.setpower does nothing because I have no power pin.
Otherwise I did what you wrote. Do you have a complete startup sequence?
Is there any way to draw a simple graphics object on the display, without lvgl intervention to check the ili9341 driver?

[uraich]

I finally managed to fix things. The problem was that the examples for initialization were not too helpful and the Projects made with this binding certainly does not work (any more) as it is.

The main problem was a missing event loop in which lv_task_handler is called. Without this, the flush_cb is never called and nothing appears on the screen.
For the colors, ili9341.BYTE_ORDER_BGR and the byte swap was not enough. I also had to switch color inversion on.
So ... the screen works now and I can have a look at the touch controller. I must also re-install the normal ili9341 driver and the display_driver_framework.py
Once done, I might try to write some documentation on how to get an lvgl_micropython esp32s3 system going. Without this, it is really difficult to get off the ground.

[resnbl]

Glad you got it figured out.
I share your frustration with the lack of complete documentation. One of the challenges here is that the many manufacturers that use ILI9341 driver wire their TFTs and touch panels differently based on whatever is most convenient when laying out the PCB it is on.

Be glad you have an ESP32-S3 - many CYDs (like the ones I have) only have an ESP32-D0WD-V3 with no PSRAM at all. This will be interesting to see if I am actually able to fit my whole app on it...

[uraich]

Well, mine has 8 MB of PSRAM and 16 GB of flash! Then it has a microphone, a loadspeaker with driver and a few more gadgets. It should be nice to play AI on it as the ESP32S3 has dedicated vector instructions.
--- and all that for 20 Euros from China!
I still get this when compiling (actually trying to flash):
esptool.py v4.12.dev1
Wrote 0x2e52f0 bytes to file /opt/ucc/micros/esp32/lvgl_micropython/build/lvgl_micropy_ESP32_GENERIC_S3-SPIRAM_OCT-8.bin, ready to flash to offset 0x0
esptool args: ['-m', 'esptool', '--chip', 'esp32s3', '-p', '/dev/ttyACM0', '-b', '460800', '--before', 'default_reset', '--after', 'hard_reset', 'write_flash', '--flash_mode', 'dio', '--flash_size', '8MB', '--flash_freq', '80m', '--erase-all', '0x0', '/opt/ucc/micros/esp32/lvgl_micropython/build/lvgl_micropy_ESP32_GENERIC_S3-SPIRAM_OCT-8.bin']
usage: esptool [-h]
[--chip {auto,esp8266,esp32,esp32s2,esp32s3beta2,esp32s3,esp32c3,esp32c6beta,esp32h2beta1,esp32h2beta2,esp32c2,esp32c6,esp32c61,esp32c5,esp32c5beta3,esp32h2,esp32h21,esp32p4,esp32h4}]
[--port PORT] [--baud BAUD] [--port-filter PORT_FILTER] [--before {default_reset,usb_reset,no_reset,no_reset_no_sync}]
[--after {hard_reset,soft_reset,no_reset,no_reset_stub,watchdog_reset}] [--no-stub] [--trace] [--override-vddsdio [{1.8V,1.9V,OFF}]]
[--connect-attempts CONNECT_ATTEMPTS]
{load_ram,dump_mem,read_mem,write_mem,write_flash,run,image_info,make_image,elf2image,read_mac,chip_id,flash_id,read_flash_status,write_flash_status,read_flash,verify_flash,erase_flash,erase_region,read_flash_sfdp,merge_bin,get_security_info,version} ...
esptool: error: argument operation: invalid choice: 'esptool' (choose from load_ram, dump_mem, read_mem, write_mem, write_flash, run, image_info, make_image, elf2image, read_mac, chip_id, flash_id, read_flash_status, write_flash_status, read_flash, verify_flash, erase_flash, erase_region, read_flash_sfdp, merge_bin, get_security_info, version)

[resnbl]

The esptool args shown are for invoking esptool as a Python module. The full command line is:

python3 -m esptool --chip esp32s3 ...

In my case, I create a venv and "pip installed" esptool to get the latest version and to be able to run it as a command directly:

source venv/bin/activate
pip install esptool

This would make the command simply:

esptool --chip s32s2 ...

(You don't need the -m esptool arguments.)

[uraich]

I also modified

            if (
                backlight_on_state != STATE_PWM and
                self._backlight_pin is not None
            ):
                self._backlight_pin.value(backlight_on_state) # removed **not**

in display_driver_framework

[resnbl]

I think the intent in display_driver_framework is to turn the backlight off until the initialization occurs later. I think your code should include:

display.set_power(True)
display.init(1)
display.set_backlight(1)  # turn backlight on AFTER init() completes

In my code, I have the backlight configured for PWM. To do that, add

backlight_on_state=ili9341.STATE_PWM,

to your display initialization. Then change the last command above to:

display.set_backlight(100)

This allows you to control the brightness of the backlight, not just turn it on/off.

It may be too late for you now, but I found the display_configs .toml files useful to find out what settings are used by display drivers.

[uraich]

Many thanks Bob,
I went back on my decision and put the "not" back into the backlight setting. I thought it actually was intentional to have the backlight off by default and have the user switch it on.
True, I could try PWM for the backlight.
In my code I don't need to set the power, since the card does not have a separate power pin for the display.
In the meantime I had some success with the touch driver. Again it did not work first go: You must not switch the rotation on until the calibration is done. Otherwise you will only see the first red rectangle.
I see that the touch screen does something when trying it on a slider. Unfortunately there seems to be a mismatch between the touch coordinates and the screen coordinates. When I move right on the touch, the slider moves left.
But, I am pretty sure that I am not far away from being able to fully use the board.
Once lvgl works, I still have to try the SD card, the microphone and the speaker.
The esptool problem comes in when you use deploy within the compilation command. I know how to use esptool, but I think that there is a bug in builder/esp32.py
Thanks again
Uli

[resnbl]

Have you run indev.calibrate()? If you do that once, you should get the proper left/right settings. I think it needs to be run before any display.set_rotation(lv.DISPLAY_ROTATION._x) statement, but I'm not sure...

Let me know when you get the SD card working. That is something I haven 't tried yet.

[kdschlosser]

calibration only needs to be run on resistive based touch screens.