• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 63 subscribers
  • Views 510 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.

TLC5920 possible timing or current leakage problem

Josh Triska
Prodigy 20 points
Other Parts Discussed in Thread: TLC5920

Hello,

I have a design for a client which uses four cascaded TLC5920 chips driving one 16x8 LED matrix each. My issue is that on rows which have, say, one LED lit, the other LEDs in the same row (on the same COM pin) glow faintly. Order of operations to write to the four chips:

  1. Pull blank high (DSEL is not used and tied low)
  2. Clock out 64 bits of data while also setting CSEL pins (no particular attention is paid to timing between data clocking and CSEL transitions)
  3. Toggle latch to set outputs
  4. Pull blank low to enable common drivers

Blank and latch pins are tied together on the four chips. Refresh rate is 1kHz, data clock is 1MHz, and setup and hold times are at least 250 ns for everything. As far as I can see, this should work just fine according to the datasheet, but is there a  timing problem I don't see? Should I not allow transitions of the CSEL pins while data is clocking? I'm not sure why any current could be flowing through the other LEDs as blanking the output should completely shut down the low-side drivers.

Cheers,

Josh

  • 0
    TI__Guru 65180 points

    Hello Josh,

    Sounds as if you have some type of ghosting effect. This means you have some parasitic capacitance in the panel which needs to be charged otherwise such a effect can happen.

    The solution could be a resistor from VLED to each COM output and check if the issue is gone then. Maybe a 1 kOhm will work.

    Best regards,
    Brigitte

  • 0 in reply to Brigitte
    Prodigy 20 points

    Thank you for the suggestion. I will try a resistor from a COM line to VLED and see if it eliminates the ghosting on that row.

    ~Josh

  • 0 in reply to Brigitte
    Prodigy 10 points

    Hi Brigitte,

    I have a similar TLC5920 issue with my design. I'm using three TLC5920 parts cascaded. I see a similar ghosting issue with the third TLC5920 in the chain where one LED is faintly lit when the other LED (which shares a COM line) is lit. I think in my case I may have a timing issue. Using fly wire, I routed the control and data lines from the 1st driver to the 3rd driver. Everything else on the board remained the same. In this configuration, there was no ghosting. I can provide scope captures of the CSEL, Latch, Blank, etc. lines if necessary. Could a slight timing issue cause the ghosting? I can provide any additional information needed.

    Thanks,

    Mike

  • 0 in reply to MD
    TI__Guru 65180 points

    Hello Mike,

    If there is a timing issue, most probable the clock line is going high later than expected. I would expect that then the accidentally lit LED should always be the one before the one that is lit on purpose. If this is the case, please check the clock and data lines on the pins of the 3rd part. It is important that you measure on the pins and not on some connector on the board.

    Then I would recommend to add a driver to the clock line to make the signals clearer again.

    When you fly wire the clock and data lines to the 3rd chip are you also connecting GND with a wire? If yes, this could also be the root cause for the removing of the ghosting.

    What happens if you fly wire from the second to the third driver? Is the ghosting gone or is it still there?

    Best regards,
    Brigitte