Struggling with connecting multiple 3.5 RPi displays

[cvels]

Hi Bodmer,

I’m struggling with connecting multiple 3.5 displays to a NodeMCU ESP32 dev board.
I was only able to get it working when using 2 TFT SPI displays, based on the ILI9431 driver. (https://www.waveshare.com/wiki/2.4inch_LCD_Module )
The problem is that the ILI9431 (at least that is what I found on the net) is limited to a display size of something like 2.8inch.
My second attempt is to have 2 TFT SPI 3.5 inch displays also from Waveshare (https://www.waveshare.com/3.5inch-tft-touch-shield.htm ) working together on the same SPI bus. These displays are using driver chipset ILI9486 (RPI_ILI9486).
I’m using the CS trick to distinguish between the displays in the sketch, but for some reasons seems that the display is not picking up this.
At the moment, I’m only trying to have one display working …
When I define CS correctly in the setup11_RPi_touch_ILI9486.h, it’s all working. (I have tried different ways with user_setup.h where I defined RPI_ILI0486_Driver etc.
But when I define CS as -1 and do the CS handling in the sketch, the display stays white and not reacting at all.
In your code, I saw remarks about an RPi display is reacting slower/different on CS changes … maybe related.

IDE: Arduino
TFT_eSPI version: 2.4.72
interface SPI:

Any idea on what I’m doing wrong?
Many thx in advance,
Kind regards,
Carl
test TFT.txt
Setup11_RPi_touch_ILI9486.txt
l

[Bodmer]

I don't think this is an issue with the library is the display operates when CS is defined.

I would expect the display to operate if the CS is controlled by the sketch. Note that the chip select must be low during tft.init() or tft.begin().

[Frank-Bemelman]

If you have working setup with one display (CS defined to some pin) - I would expect the display would also work if you connect CS to ground, not to the pin. And then you could try again, connecting CS to a pin of your choice, and control it from the sketch. I would expect that all should work. Last step would then be to define CS as -1 and see if it keeps working. And add a second display.. That would be my approach to the problem you seem to have.

[cvels]

Hi Frank, I agree but that is the weird thing. I tried that before creating this topic, connecting the CS pin directly to the ground was also not helping. I'm missing something small and stupid ... but cannot find the light. That's why i have ask the question in the first place to find out if the library was doing somewhere a magic thing which makes the difference. And it's only with this display/driver (the other displays with ILI9431 are working as a charm)

[Bodmer]

I suspect the problem may be related to the serial to parallel interface on these RPi boards. The chip select manages a counter that generates the write pulse.

To check if the power demand and SPI bus loading is not the issue I would connect both displays to the exact same pins (including the same chip select pin) and check both displays show the same thing.

[cvels]

Hi Bodmer, I have connected both displays together using exact same pins (also CS). The result is the same.
When using your library, both screen shows the same content (= ok). When handling the CS inside the sketch (and put in the header files of the library "#define TFT_CS -1" ) both screens turns white. No content.

[Bodmer]

This shows that the RPi interface chip select cannot be controlled from the sketch. This is likely due to the way the serial to parallel interface in these designs generate the control signals to the display. It appears that there is no work-around for this behaviour.