This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

how to debug calibration of touchscreen on am335x

I'm using custom board based on am335x with resistive touchscreen  with tisdk 02.00.02.11

when system prompt to calibrate the touchscreen, the calibration tool respond to the touch most of the times. but after that the matrix-gui ignore my touch.

What I'v noted is that the tsc driver use the the right lower corner as 0,0 for (x,y) axes while the gui expect the top left corner as (0,0)

x value as reported  from the driver varies from 300 to 4000. while my screen is 800*480

evtest or ts_print seems to report touch event even when I'm not touching the screen with strange special range of value around 50, while real touch have more than 100 value. what could possibly cause this? do I have to fix this on the driver level? or it's normal.

is it possible to show a pointer when I touch the screen so that I can know where the system think where is the touch.

  • Hi,

    do I have to fix this on the driver level?

    I don't think the problem is in the driver level, as the TI ref boards work fine with ti_am335x_tsc.c driver..

    Try changing the tsc parameters in dts: x-plate-resistance, wire-config, charge-delay, adc-channels (use the channels, that are actually connected to your tsc).

    is it possible to show a pointer when I touch the screen so that I can know where the system think where is the touch.


    In drivers/input/touchscreen/ti_am335x_tsc.c add debug prints in titsc_read_coordinates(), that show the values calculated for x & y, you can also print the pressure value from titsc_irq().

    Best Regards,
    Yordan
  • Hi,
    Thanks, I've increased the charge-delay to 0x800, and the touch screen become responsive and it's working now.
    just one minor issue, one of ten tries of touching the screen, the touch is also duplicated in near point.
    increasing the charge-delay to even 0x20000 doesn't solve the issue. any idea?
    Thx,