LCD Sharp BoosterPack SPI examples no longer work on CC13x0.

[rei-vilo]

The LCD_Sharp128BoosterPack_SPI_main example no longer works on the Sharp 128 against the CC1350 LaunchPad.

Only the first Hello from line 45 is displayed.

Energia/libraries/LCD_SharpBoosterPack_SPI/examples/LCD_Sharp128BoosterPack_SPI_main/LCD_Sharp128BoosterPack_SPI_main.ino

Lines 45 to 46 in 3116b84

	myScreen.text(10, 10, "Hello!");
	myScreen.flush();

The program seems to freeze before line 52.

Energia/libraries/LCD_SharpBoosterPack_SPI/examples/LCD_Sharp128BoosterPack_SPI_main/LCD_Sharp128BoosterPack_SPI_main.ino

Line 52 in 3116b84

myScreen.reverseFlush();

Same for the LCD_SharpBoosterPack_SPI_main example on the Sharp 96 against the CC1350 LaunchPad.

[rei-vilo]

The two same examples LCD_Sharp128BoosterPack_SPI_main and LCD_SharpBoosterPack_SPI_main run smoothly on a CC1352 LaunchPad with the corresponding screens.

[rei-vilo]

Using a logic analyser shows the internal VCOM doesn't work as expected on the CC1310.

On the CC1350

SPI /CS goes high to select the screen, a first SCK signal goes on but nothing more. No second SCK signal and no /CS going low to deselect the screen.

On the CC1352

[rei-vilo]

Replacing

Energia/libraries/LCD_SharpBoosterPack_SPI/LCD_SharpBoosterPack_SPI.cpp

Lines 531 to 532 in 3116b84

	SPI.transfer((char)command);
	SPI.transfer((char)SHARP_LCD_TRAILER_BYTE);

by

        shiftOut(15, 7, MSBFIRST, (char)command);
        shiftOut(15, 7, MSBFIRST, (char)SHARP_LCD_TRAILER_BYTE);

seems to solve the issue.

But adding an extra SPI.begin(); before line 531 doesn't solve the issue.

• Could SPI be called from a timer task?
• Does the implementation of TI-RTOS on CC1350 differ from CC1352?

I don’t know the minutiae of the implementation of TI-RTOS in Energia MT.

[rei-vilo]

Here are the traces after adding the patch

        shiftOut(15, 7, MSBFIRST, (char)command);
        shiftOut(15, 7, MSBFIRST, (char)SHARP_LCD_TRAILER_BYTE);

The program displays a first screen, waits for 1 second and then displays a second screen.

Trace for the CC1352

Trace for the CC1350

[rei-vilo]

As the CC1352 replaces the CC1350, the workaround at #1061 is good-enough to patch the issue.