Two ILI431s - different colors

MatrixRat

Well-known member
Building Midi controller DevKit V2 using a T4.1, two ILI9431 displays and other goodies.

Have buffered all SPI signals and have both displays and XP2046 touch tested and working.

Picture says a thousand words:-

TwoILI9431.jpg

Display1 is left, Display2 is right. Swapped ILI9431s, same result.

The Display2 object seems to be missing some bits.

So, How do I find those missing bits? Not tried yet but Screens on unique SPI bus? Would like to avoid that as pins are at a premium.

Thanks for any suggestions.
 

Attachments

  • FillTwoScreens.ino
    11.9 KB · Views: 31
Display1 is left, Display2 is right. Swapped ILI9431s, same result
When you say "same result" can you clarify - do you mean that the left display is always fine no matter which physical device is wired there, or that the screen that has lower bitdepth on the right also has lower bitdepth when on the left?
 
do you mean that the left display is always fine no matter which physical device is wired there,

Yes. Both screens plug into the main board so are easy to swap. Will play with swapping DC and CS #defines but expect that lower bit depth will manifest on the left screen.

Editing to report that adjusting code saw no change. Will double check wiring and sprinkle some pullups.

SPI outputs are buffered via 100R with 74HCT245 and MISO has 74HCT125.
 
Last edited:
Ok, I've moved it with:-

Code:
ILI9341_t3 Display2(TFT1_CS, TFT1_DC);//Was ILI9341_t3 Display1(TFT1_CS, TFT1_DC);
ILI9341_t3 Display1(TFT2_CS, TFT2_DC);//Was ILI9341_t3 Display2(TFT2_CS, TFT2_DC);

That is to say that the left display has reduced bit depth. Interesting is that after a few hours of wheelbarrow and shovel work, neither display had reduced bit depth. Hmm, from a cold start..

Points to wiring and layout issue.
 
this is a strange bug. May I ask why you need to use 74HCT245/125 between the T4.1 and the ili9341 screens ?
 
After sleeping on it, realized that HCT powered from +3.3v was a silly old fart dumb choice. Surprised it worked at all.

Using protoboard with two 245s located underneath T4.1 as I need to buffer at least 10 other pins to the outside world so the opportunity to experiment with SPI buffering presented itself. Orientation is such that SCK has the shortest possible path to 245 input.
Next will swap to HCs.
 
Back
Top