• State Resolved
  • Locked Locked
  • Replies 10 replies
  • Subscribers 98 subscribers
  • Views 460 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.

AM5718: SLEWCONTROL bit for VOUT1 port pads

David M. Alter
Guru 61235 points
Part Number: AM5718

An AM5718 customer is observing excessive ringing on the vout1 video port output signals.  During debug, they uncovered two mistakes.

First, the vout1 power domain was mistakenly connected to 3.3V instead of 1.8V.  The datasheet specifies that only the 1.8V connection is supported on the AM5718 (sprs957g, section 7.7, p.218:   "NOTE  VOUT1, VOUT2 and VOUT3 only qualified for use at 1.8 V.").  This will be fixed on the next revision of their board.

The second mistake they noticed is that they did not have the SLEWCONTROL bit set for slow slew rate for the port pads.  Again, per p.218 in the datasheet:

CAUTION, All pads/balls configured as vouti_* signals must be programmed to use slow  slew rate by setting the corresponding CTRL_CORE_PAD_*[SLEWCONTROL] register field to SLOW (0b1).

Now, they changed the SLEWCONTROL to slow slew, but observe no discernible change in the output signal waveform.  They expected the edges to slow down.

***Would having VOUT1 powered at 3.3V account for the no-effect of the SLEWCONTROL bit?

***What are the expected edge rise/fall times on the VOUT1 port?  (I do not see this in the datasheet).

***Is there any other setting or connection that would impact the slew rate of this ports signals?

Thank you and regards,

David

  • 0
    TI__Mastermind 46460 points

    David,

    We had to remove VOUT support for 3.3V and for faster slew rates to prevent a reliability impact due to the package design.  It should be functional on your prototype but compliance with the Data Manual is required for production boards.  Some ringing is allowed on these pins but it must not violate the VILmax and VIHmin. 

    Are they seeing ringing on all nets or just certain ones?  Do you see this on all boards built?  How many were built?  What is the trace length?  Are there any routing stubs or connectors in this path?  Can you post a scope capture of the clock and a data pin measured at both the source and load ends of these routes?

    Tom

  • 0 in reply to Tom Johnson 16214
    TI__Guru 61235 points

    Hi Tom,

    Thank you for your reply.

    Tom Johnson 16214 said:

    Are they seeing ringing on all nets or just certain ones?  Do you see this on all boards built?  How many were built?  What is the trace length?  Are there any routing stubs or connectors in this path?  Can you post a scope capture of the clock and a data pin measured at both the source and load ends of these routes?

    Customer has not checked all signals, but is seeing the ringing on clock and several data pins.

    They have checked 4 out of 10 boards built.  They see the problem on all 4 checked.

    Trace lengths are about 3.5 inches, with one board to board connector in the traces.

    Customer is working to get some scope shots for you.

    Regards,

    David

  • 0 in reply to David M. Alter
    TI__Mastermind 46460 points

    David,

    Are there series resistors on any of these nets?  What is the highest VOUT pixel clock rate planned for this board?

    Tom

  • 0 in reply to Tom Johnson 16214
    TI__Guru 61235 points

    Tom Johnson 16214 said:

    Are there series resistors on any of these nets?  What is the highest VOUT pixel clock rate planned for this board?

    Customer is planning to use 64 MHz pixel clock.  They do not have series R on the lines right now.  They are adding 22 Ohm series R on the clock line for next board rev, but are hoping to avoid adding R's to the whole bus due to board area restraints.

     

    - David

  • 0 in reply to David M. Alter
    TI__Mastermind 46460 points

    David,

    The inductance of the connector may make series resistors a requirement for all VOUT nets.  Of the 3.5", how much of this length is on the board containing the AM5718 and how much on the other board?  Can the resistors be added to the other board?  Of course, in the next board spin, they will use 1.8V which may also reduce the ringing considerably.

    Tom

  • 0 in reply to Tom Johnson 16214
    TI__Guru 61235 points

    Tom Johnson 16214 said:

    The inductance of the connector may make series resistors a requirement for all VOUT nets.  Of the 3.5", how much of this length is on the board containing the AM5718 and how much on the other board?  Can the resistors be added to the other board?  Of course, in the next board spin, they will use 1.8V which may also reduce the ringing considerably.

    Customer advises that most of the length of the trace is on the AM5718 board.  Only about 0.5 inches on the connecting board.

    - David

  • 0 in reply to David M. Alter
    TI__Mastermind 46460 points

    David,

    Are you still expecting to receive scope captures of the ringing, as requested?

    Tom

  • 0 in reply to Tom Johnson 16214
    TI__Guru 61235 points

    Tom,

    Thank you for checking.  I have queried the customer on status.  Will advise as soon as possible.

    - David

  • 0 in reply to David M. Alter
    TI__Guru 61235 points
    Tom,

    Customer advised:

    "For schedule reasons we decided to just make a best effort at improving the design and spin new revision at 1.8V with series termination on the clock line. We are expecting those boards in mid-March and will make a comparison with scope shots to confirm whether there is any issue remaining after the changes."

    Regards,
    David
  • 0 in reply to David M. Alter
    TI__Mastermind 46460 points

    David,

    Understood.  I will close this thread.  We can open a new one and reference this one if more discussion is needed.

    Tom