• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 27 subscribers
  • Views 641 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DLPC910: Flex cable matching

Andrew Holme
Prodigy 140 points
Part Number: DLPC910
Other Parts Discussed in Thread: TIDA-00570, , DLPC900

Reference design TIDA-00570_2514435A_HI-SPEED_DLPC910_MAIN_ESD has shunt-C, series-R matching network between DLPC910 and flex LGA on the following outputs: RESET_ADDR[3:0], RESET_IRQZ, RESET_MODE[1:0], RESET_OEZ, RESET_SEL[1:0], RESET_STROBE.

DLPC910 datasheet says these signals are LVDCI_33 outputs; but does not specify any impedance or matching requirements.

There is a note on reference design schematic page 11:  "RC filtering only needed due to impedance mismatch by TI flex cable."

We are planning to use the Amphenol 7840-001 flex cable to connect our custom board to a DLi6500 off-the-shelf dev kit.  Is "TI flex cable" same thing as Amphenol 7840-001?

Why is the matching network required?  Why is DCI not enough?  Should we replicate the RC matching network on our custom board?

I notice the DLPC900 reference design also has an RC matching network; but uses different R and C values, although the connector LGA footprint looks the same.  Why different values?

Thanks.

  • 0
    TI__Expert 4530 points
    Hi Andrew,

    Welcome to DLP section of TI E2E community. I will get back to you to end of today.

    -ykc
  • 0 in reply to ykc
    TI__Expert 4530 points

    Hi Andrew,

    Yes TI flex cable is Amphenol 7840-001.
    RC values are different for DLPC900 because circuitry used in DLPC910 (based on Xilinx FPGA) is not same as DLPC900 (ASIC) and they work at different speeds.
    If flex cable impedence Zline is different from DMD Zload then RC impedence matching is still needed even if DCI output is used.
    Regarding what is recommended impedence for these signals, I will get back to you after consulting the expert within a week.

    -ykc

  • 0 in reply to ykc
    TI__Expert 4530 points

    Hi Andrew,

    Source impedance is  50ohms as set by  resistors DLPC_VRN_BANK4 and DLPC_VRP_BANK4 = 50ohms.

    Line impedance for these signals in cable is unknown as this is very old design. You can contact the vendor for more details.

    The board designer used RC values to match the impedance. It is recommended to use these values as board has been validated for these values.

    -ykc

  • 0 in reply to ykc
    Prodigy 140 points

    When you say "load impedance on DMD side is also 50 ohm" do you mean the PCB trace impedances?  The inputs themselves are LVCMOS with 18,000 ohm internal pull-downs.  Source-termination would be the simplest way to drive these signals.  DCI is programmed to 50 ohm using VRN/VRP resistors.  An additional series resistor alone would increase this; however, could one not just change the VRN/VRP resistors if a non-50-ohm resistive termination was required?  Here we have (DCI // C) + R which is an unusual source-termination circuit, not purely resistive.  Please may I have more explanation?

  • 0 in reply to Andrew Holme
    TI__Expert 4530 points

    Hi Andrew,

    This is very old design. Board was validated for these values. Capacitors may have been added to reduce EMI. You can experiment with these values. From current series termination value it seems line impedance = 75 ohm.
    You can test by using DCI to set impedance to 75ohms and avoid series termination resistor.

    -ykc