• State TI Thinks Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 1245 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.

TLC5929: Ghosting phenomenon after SW modift the speed

chentsu chung
Mastermind 6180 points
Part Number: TLC5929
Other Parts Discussed in Thread: TLC59283

Hi,

all signals, including SCLK, SIN, LAT, BLANK are affected by shift_speed / gray_scale_speed change. Basically, when you change from 3 to 2, all signals become faster (on and off times for each signal are shorter) and when you change from 2 to 3, all signals become slower( on and off times for each signal are longer).

In any case, customer having problems even with speed set to 3.

 

One thing that I discovered last night when testing with revision B board is that it works ok, immediately after the board is powered on. I don't see any ghosting problems and all test patterns look ok, but as soon as I write "function control data", it starts showing ghosting. This happens even if I write the exact same value to "function control data" as it was after power-up, which is E67Fh. To make it clear here are steps that I do:

1. Power off LED module and power it on -----> Everything looks ok, no ghosting

2. Move horizontal line up and down -----> Everything looks ok, no ghosting

3. Send E6F7h to function control data ------> Ghosting starts

 

I will double-check my FPGA code over next few days to make sure there are no mistakes there, but it seems as soon as I write any function control data, LED driver outputs start to leak some current and that causes ghosting. Here are my questions:

 

1. Do I need to have all data latches to 0, before I make changes to function control data?

2. Do I need to set BLANK to hight, before I make changes to function control data?

3. Is it possible that default value after power-up for function control data latch is not E67Fh, but some other value?

4. Is there any other way to reset LED driver to the default value, other than power-up reset?


  • 0 in reply to chentsu chung
    TI__Genius 14295 points
    Hi, Chentsu,

    Please see my answers:
    1. No.
    2. No.
    3. No. The default value can't be changed.
    4. No. You have to power off Vcc to reset to default value.

    Please check your code for function control.

    Can you send different BC data to check whether the code is working?

    Thanks.
    Regards,
    Kenneth
  • 0 in reply to Kenneth Du
    Prodigy 40 points
    Hi Kenneth,
    yes, sending different BC values works correctly, as well as selecting between LOD and LSD LED detction modes. We also get correct status information, image data (data latch) works ok.
    In fact, code is over 4 years old and used quite a bit, until we got complaint from customer about "ghosting" or "leaking" LEDs, when they checked LED board inside a dark room.
    System is using 1/4 multiplexing and when we, for example, turn-on row 1, there are also some random, dimly lit LEDs in rows 2, 3 and 4 (visible in the dark room). Even if we adjust BC to very low value, like 1 or 2, these "ghosting" LEDs seem to be unafected, ie. they are still there, even at same brightness. If we only power-off and power-on whole LED module, these "ghosting" LEDs dissapear completely. However, as soon as we send anything to function control, including exactly same value as default value E67Fh, "ghosting" LEDs reappear in exactly same positions and there is no way to turn them off, until we power-off and power-on whole LED module. Ie. seems that an update to function control, enables all outputs with very low current, even when all data latches are set to 0 and everyhing should be off.
  • 0 in reply to Slavko Radman
    TI__Genius 14295 points
    Hi, Slavko,

    From your detailed description, it seems you are facing the common ghosting problem in multiplexing application.

    Since TLC5929 does't have pre-charge FETs, there's chance for ghosting happens.

    One way to test, you can get the output voltage waveform for the channel when "ghosting" LED happens.

    We also have TLC59283 and it integrates Pre-Charge FET for Ghosting Reduction and it's PIN compatibale with TLC5929.

    But some features are not the same. Can you help to check this device?

    Thanks.

    Regards,
    Kenneth
  • 0 in reply to Kenneth Du
    Prodigy 40 points

    Hi Kenneth,

    thank you for your answer, however, our system requires LED fail detection and global brightness controller, therefore TLC59283 is not an option in this case.

    Anyway, I am still somewhat confused by the fact that immediately after power-up TLC5929 does not show any ghosting and as soon as we send something to function code, ghosting starts. 

  • 0 in reply to Slavko Radman
    TI__Genius 14295 points
    Hi, Slavko,

    I understand.

    Can you help to check the output voltage waveform for the two situations to compare the difference bewteen no ghost and ghost?

    If you have two or more displays, can you check on another display?

    Thanks.
    Regards,
    Kenneth
  • 0 in reply to Slavko Radman
    TI__Genius 14295 points
    Hi, Slavko,

    I'll close this thread due to inactivity.

    If you have more questions, please reply this directly and it will open again.

    Please help to click "Resolve" if you think this is helpful.

    Thanks.
    Regards,
    Kenneth
  • 0 in reply to Kenneth Du
    Prodigy 40 points

    Hi Kenneth,

    my apologies, as I was busy testing various options, over the last couple of days. Changing different types of MOSFET (with ghosting elimination function) did not help much. Turing off all MOSFET's, before turning on the next row, helped a bit, but to my surprise, leaving TLC5929 in "output leakage detection" mode (bits 8 and 9 of function control set to 11b) permanently, improves the situation more than anything else.

  • 0 in reply to Slavko Radman
    TI__Genius 14295 points
    Hi, Slavko,

    MOSFET should not help since in your application the ghosting is under the row that's ON.

    As you can see from datasheet, "A very small current is sourced from the turned-off OUTn to detect leaking when the SIDLD bits are '11' and BLANK is low" , this should be helpful to eliminate the ghosting.

    Please let me know if you need other help.

    Thanks.
    Regards,
    Kenneth