• State TI Thinks Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 1686 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.

RTOS/TM4C1294NCPDT: Issue with Ethernet on TM4C1294NCPDT Custom board

Hemant Kamat1
Intellectual 535 points
Part Number: TM4C1294NCPDT
Other Parts Discussed in Thread: LM3S8962

Tool/software: TI-RTOS

I have designed a custom board based on TM4C1294NCPDT. The board uses a integrated magnetics Ethernet connector interfaced with the MCU pins 53,54,56,57. Appropriate matching and filter circuit is connected per the MCU design guidelines. However, when I execute the tcpEcho sample code on my board, it does not get the dynamic IP address from my local DHCP server. However, the same code executes as required on the TM4C1294XL Launchpad.

I have double checked my schematic (attached herewith), the PCB layout and final manufactured PCB and have found no apparent issues.

I have used Wireshark to monitor DHCP request and acknowledge packets but see no activity on the LAN.

The rest of MCU functionality is working as required including debug messages on UART0.

I would appreciate if anybody could suggest solution or pointers to the issue

Regards,

Hemant6327.Gateway-0.9.1.sch.pdf

  • 0
    Guru 117855 points

    Would it not make "great sense" for you/others - in just such a frustrating position - to clearly specify the (exact) components you've employed (which likely were NOT those present upon the LPad) and which may have "caused/contributed" to your problem?

    Getting "close" may not prove sufficient - even slight component performance deviation (may) prove difficult to detect - yet still "corrupt capability."

    Desire for a "custom board" (never explained/justified - btw) so often causes such issues - the "degree of difficulty" of (proper) component selection, board layout, manufacture & assembly - may not have been "fully recognized nor appreciated."     Until now!      And now what?

  • 0
    TI__Guru**** 190746 points
    HI,
    Did you have 25MHz MOSC with less than +/- 50ppm tolerance? Please refer to the system design guideline on the pcb layout recommendation for Ethernet. www.ti.com/.../spma056.pdf
  • 0 in reply to Charles Tsai
    Intellectual 535 points
    Hi,

    I have used 25MHz MOSC with < +/-50ppm tolerance. I have referred to datasheet of TM4C1294NCPDT, Technical Brief and System Design Guidelines among other relevant documents from TI for the design. The LAN connector with magnetics is TRJ19111DNL from TRXCOM. The matching and filter components are within 1% tolerance and are placed within distance as specified in the design guideline documents.

    ETH0LED0 turns on (and turns off) when LAN cable is inserted (and removed) from the RJ45 socket. The ETH0LED1 blinks suggesting apparent packet movement. I have used Wireshark to monitor TCP/IP packets but no packets seem to leave the LAN connector.

    The tcpEcho application waits with the following message

    Starting the TCP Echo example
    System provider is set to SysMin. Halt the target to view any SysMin contents in ROV.
    Service Status: DHCPC : Enabled : : 000
    Service Status: DHCPC : Enabled : Running : 000

    and after some time gives the following message

    Service Status: DHCPC : Enabled : Fault : 002

    Let me know if you need more information. I have attached the schematic PDF with my post.

    Thanks for your time.

    Regards,

    Hemant
  • 0 in reply to Hemant Kamat1
    Guru 117855 points
    Charles - much good detail (now) provided - yet the, "Number of boards produced - which yield this failure - is absent."
    Far better case could be made if this issue "persists" across multiple boards - rather than a "single, rushed, cost-constrained build..."
  • 0 in reply to cb1_mobile
    Intellectual 535 points
    Hi,

    The issue persists in all boards that have been produced for testing.

    I have inspected all boards for any manufacturing defects. But none are visible. I will upload a few more diagrams/images that may give more details to help in narrowing down on the issue. I understand custom boards are difficult to debug remotely and I will provide as much info as required so that the e2e community can suggest failure modes and solutions.

    Thanks for your time.

    Regards,

    Hemant
  • 0 in reply to Hemant Kamat1
    Intellectual 535 points
    Hi All,

    Any inputs about how I can resolve this issue? Need to resolve the issue asap. Thanks for your time and apologies for pushing. Regards, Hemant
  • 0 in reply to Hemant Kamat1
    TI__Guru**** 190746 points
    Can you refer to the below similar post? The post might have been created on your behalf already.

    e2e.ti.com/.../2385618
  • 0 in reply to Hemant Kamat1
    Guru 117855 points

    Hemant Kamat1 said:
    The issue persists in all boards that have been produced for testing.

    My friend - you seek a (near) urgent response - yet provide an "response (above)" which provokes, "Question - not Answer!"

    Clearly we seek a sense of, "How many such boards have been built - AND - how many have been tested?       

    The diagnosis for "small numbers" is different than that for "larger numbers" - and "all boards produced" provides no such insight...     (such could be 3 or 5 or 250 (even - and OFTEN 1!) etc...)

  • 0 in reply to Charles Tsai
    Intellectual 535 points
    Hi Charles,

    I will check out the e2e.ti.com/.../2385618 post.

    I had used LM3S8962 for my earlier Ethernet board and that had worked. The design guidelines for LM3S8962 Ethernet suggested that the ETH0RXIP and ETH0RXIM pins be grounded through 49.9 ohms resistors and a 0.01uF capacitor. Do you think this change needs to be done on my current TM4C1294NCPDT design?

    Thanks and regards,

    Hemant
  • 0 in reply to Hemant Kamat1
    TI__Guru**** 190746 points

    Hi Hemant,

     The LM3 Ethernet is a complete different module compared to the TM4C129. As I noted earlier please reference the system guideline app note /TM4C129 datasheet/ TM4C1294XL LaunchPad schematic for TM4C129, definitely not the LM3.

     With that being said, I do also see 49.9 ohm resistor on the below LaunchPad schematic. However, it is 0.1uF on C16, not 0.01uF as you noted. Please also refer to the other post where Bob mentioned to check the chassis ground on your board.

  • 0 in reply to Hemant Kamat1
    Mastermind 8695 points

    Hemant Kamat1 said:
    Service Status: DHCPC : Enabled : Fault : 002

    The launchpad gives the same error if it can't find a DHCP server. Use XGCONF to set a static IP and try pinging it. This will also fail if the hardware is botched.

    Hardware seem fine at a glance. The caps Charles referenced are just for decoupling.The built in magnetic RJ45 has a DC isolating cap so the ground connection Bob referenced is OK too, but I'm not sure how any large DC voltage potential would dissipate without a resistor in parallel. Maybe Trxcom knows something I don't.

    You should set a timer that verifies system clock is running at 120MHz as a way of checking the 25MHz osc.

  • 0 in reply to Charles Tsai
    Intellectual 535 points
    Hi Charles,

    Thanks for the observations. I will work on them and checkout.

    Regards,

    Hemant
  • 0 in reply to Peter Borenstein
    Intellectual 535 points
    Peter,

    Thanks for your observations and suggestions.

    Regards,

    Hemant