• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 100 subscribers
  • Views 318 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

OMAP35xx: LCD becomes white sometime

Jitendra Kumar
Intellectual 285 points

Hi,

I am using omap35xx EVM and u-boot  2008.10. When i am displaying the image over the LCD it displays it properly, but sometime the LCD becomes white.

And, when you again run execute the display routine, it displays the image over LCd  properly. This is occur's randomly.

Can any one suggest me some pointers.

Thanks in advance.

Regards

-Jitendra

  • 0
    TI__Mastermind 41665 points

    Are you displaying the image to the LCD from Linux or have you written some test routine into U-Boot? In either case I would do some verification on the software to ensure it is writing to the display buffer properly, it is hard to say what the real problem is without further detail.

  • 0 in reply to Bernie Thompson TI
    Intellectual 285 points

    Hi,

    thanks for reply.

    Yes, We have written a test routine in u-boot to display the data on LCD. i will added the verification code to check that what is being in lcd display buffer is as expected or not.

    I have one more query, I have camera sensor, from that am getting 10 bit bayer's raw data,  that is being dumped to memory after CCDC, I am converting this 10 bit bayer raw data  to RGB888 by right shifting 10 bit bayer's output by 2 bit and and then putting remaining 8 bit in a buffer 24 bit buffer as 3 byte of the 24 buferr represent RGB and all the three byte have same 8 bit data obatined after doing shifting operation, with this I am able to display a black and white image over LCD. But LCD has 18 bit data lines only,  and by default, the LCD is being configured with 16 Bit RGB565 format, It displays the image over full 640*480 dimension with is format of data,

    On other hand, with format changed to 24bit RGB (unpakced 32 bit RGB24 format). Then, the image getting displayed only in 3/4 part of LCD and 1/3 remains blank with some colourd dots. Does any thing more is to be reconfigure in the DSS to get full screen display over LCD?

    Thanks

    Jitendra

  • 0 in reply to Jitendra Kumar
    TI__Mastermind 41665 points

    Jitendra Kumar said:
    On other hand, with format changed to 24bit RGB (unpakced 32 bit RGB24 format). Then, the image getting displayed only in 3/4 part of LCD and 1/3 remains blank with some colourd dots. Does any thing more is to be reconfigure in the DSS to get full screen display over LCD?

    Do your camera sensor and LCD resolutions match? If they do match and the image should be filling the screen entirely than you may want to check how the DSS is expecting the data format to be, in particular that your 24 bit rgb formatting is matching one of the formats given in section 15.4.2.2.1 of the TRM for the graphics pipe or 15.4.2.3.1 for the video pipe. If the formats are not congruent you could end up with your image scaled like you describe. Note that the DSS will convert the incoming data from the video or graphics pipelines to whatever format the display requires, so the conversion from 24 bit for the image data to 16 bit for the LCD should happen automatically.

  • 0 in reply to Jitendra Kumar
    TI__Mastermind 43875 points

    Jitendra,

    As Bernie mentioned, the frame buffer and the output format can be completely different. The display sub-system will either expand or reduce the color depth accordingly.

    In your case you need to leave the DSS output format as 18 bit, but configure the DISPC_GFXATTRIBUTES register so that the GFXFORMAT is set to either 0x08 or 0x09 depending on if you are packing your 24 bit pixel or not.

    Another approach is to use the BITMAP8 CLUT mode, in this case you would not need to byte duplicate your monochrome pixel from 1 byte to 3 bytes, but use the CLUT to do the expansion on the fly for you. To do this you would need to configure the CLUT with a simple monochrome ramp.

    Note, whenever you change any display controller registers you must also then set the GOLDC bit in the register DISPC_CONTROL (bit 5). This will transfer the settings into the hardware.

    BR,

    Steve