SSD1680 Epaper - buffer wraps when width is higher than 255 pixels

[saeugetier]

CircuitPython version

Adafruit CircuitPython 6.3.0 on 2021-06-01; Raspberry Pi Pico with rp2040

Code/REPL

DISPLAY_WIDTH = 296
DISPLAY_HEIGHT = 128

display = adafruit_ssd1680.SSD1680(display_bus, width=DISPLAY_WIDTH,
                                 height=DISPLAY_HEIGHT,
                                 rotation=90, busy_pin=epd_busy,
                                 highlight_color=0x000000)

Behavior

The display buffer is not displayed correctly. The width is higher than can be addressed with 1 byte. The right part of the display buffer is displayed on the left side overlapping the previous drawn buffer.

Description

Content is not displayed correctly.

Additional information

• Problem was found with a Waveshare 2,9 inch e-Paper Module using a SSD1680 chipset.

[saeugetier]

This is how the drawn display buffer looks:

[tannewt]

The bug may be in this file but I'm not sure where: https://github.com/adafruit/circuitpython/blob/main/shared-module/displayio/EPaperDisplay.c

[saeugetier]

I have recorded the communication between the Raspi Pico and the EPaper. I saw that the chip select is high between the first and second byte after the command. The chip select is not high after the command has been sent. Maybe the chip select ends the communication and the initialization is done with wrong parameters.

Recording via sigrok: EPaper.zip

[saeugetier]

I found the bug:
The command "set_row_window_command" has a wrong byte order: The content is command: 0x45, data: 0x00, 0x00, 0x01, 0x27. But correct is: command: 0x45, data: 0x00, 0x00, 0x27, 0x01.

The byte order has to be flipped here:
https://github.com/adafruit/circuitpython/blob/main/shared-module/displayio/display_core.c#L300-L303

[tannewt]

Ok, looks like we'll need a setting for MSB vs LSB. I assume it's working as-is for LCDs so this eink must do the opposite.

Want to make a PR for it? That display isn't used by Adafruit currently so it won't be a priority for Adafruit funded folks.

[lesamouraipourpre]

Maybe related, there is an issue on Blinka relating to missing MSB using SPI with MicroPython.
adafruit/Adafruit_Blinka#480

[saeugetier]

@tannewt I can prepare a pull request. But I need somebody for testing the Adafruit EPD, if they still work.

[tannewt]

We can ask some folks to test from the PR. We'll want to test some TFT as well since that code is used for all displays.