4-Color Waveshare 2.13 inch e-Paper HAT(G) V2 does not work with ESP32 Dev Module

[michalin]

Can anybody help me? I tried to connect the display to the ESP32 dev modules, using following pins:

DIN (MOSI): 23
CLK (SCK): 18
CS: 15, or permanently connected to GND volt, as this is the only device.
DC: 27
RST: 26
BUSY: 25

I am using version 1.6.7 of the library, and the "Hello World" example.
in GxEPD2_display_selection_new_style.h, I uncommented:

#define GxEPD2_DISPLAY_CLASS GxEPD2_4C
#define GxEPD2_DRIVER_CLASS GxEPD2_213c_GDEY0213F51 // GDEY0213F51 122x250, JD79661 (FPC-A002 20.04.08)
GxEPD2_DISPLAY_CLASS<GxEPD2_DRIVER_CLASS, MAX_HEIGHT(GxEPD2_DRIVER_CLASS)> display(GxEPD2_DRIVER_CLASS(/*CS=*/ 15, /*DC=*/ 27, /*RST=*/ 26, /*BUSY=*/ 25)); // Waveshare ESP32 Driver Board

When I Use this configuration with the e-Paper Driver Hat from Waveshare, connected to the raw display, everything works fine:

But with the integrated driver board, the display does not work at all:

Other panels work with that driver circuit, so we can be sure that both wiring is correct and the board is not defective.

Any idea?

[ZinggJM]

@michalin, hi, welcome to Discussions in GxEPD2

The difference between the two Waveshare bords is the level converter on the 2.13 display bord, as far as I can see.

I guess that you have connected VCC of the display bord to 5V.
This is wrong when using a 3.3V processor, because the BUSY line then back-feeds 5V from the level converter to the processor pin.

Either connect 3.3V to the display bord, or use a series resistor ~10k in the BUSY line.

Good luck!

[michalin]

@ZinggJM ,
As far as I can see, both HAT modules have the same level converter: YE008. VCC is already connected to the 3.3V pin of the ESP32, and a 10k resistor in the busy line didn't help either. So I think there is no problem with wiring or hardware.
I recorded the signals with a logic analyzer, and here is what I came up with:

1: Raw display connected to e-Paper driver head: ~100ms after first RST (why are there two RST pulses, btw?), BUSY goes high, and the driver sends its initialization data. another ~100 ms later, BSY goes high again, the actual image data is sent and the display shows the "Hello World" as expected.

2. With the builtin driver circuit, BUSY also goes high and the same init bytes are sent. However after this, BSY remains low and the display data is never sent.

What I found is that the buillitn HAT module drags RST down while the ESP is resbooting, the other does not. Maybe this causes the problem?

FYI, it also works with the example code from Waveshare:

[ZinggJM]

I don't have this Waveshare board. So I can't test.

I assume this issue is caused by the "clever reset circuit" Waveshare uses on some boards.

void GxEPD2_213c_GDEY0213F51::_InitDisplay()
{
  if ((_rst >= 0) && (_hibernating || _initial_write))
  {
    digitalWrite(_rst, HIGH);
    delay(20); // At least 20ms delay
    digitalWrite(_rst, LOW); // Module reset
    delay(2);  // At least 40ms delay, 2ms for WS "clever" reset
    digitalWrite(_rst, HIGH);
    delay(2); // At least 50ms delay (32ms measured)
    _waitWhileBusy("_InitDisplay reset", power_on_time);
    _hibernating = false;
    _power_is_on = false;
  }

The reset pulse is already shortened for this panel, some other 4color need longer reset.
You can experiment changing this part.