• State
  • Replies 3 replies
  • Subscribers 38 subscribers
  • Views 39 views
  • Users 0 members are here
Support feedback
Options
Options
Related

MSPM0-SDK: MSPM0G3507 - XDS110 DAP Connection Error on One Board After Programming

Tulsi Das
Prodigy 120 points

Part Number: MSPM0-SDK
Other Parts Discussed in Thread: MSPM0G3507

Dear Texas Instruments Support Team,

I am using an MSPM0G3507 with CCS 20.4.0 and an XDS110 debugger. I am facing an intermittent debug connection issue on one of my custom boards and would appreciate your guidance.

Connection :
XDS110 debugger to external target(My custom board).

Wire used : GND, NRST, SWDIO, DWCLK

Target uses its own 3.3V from external power supply.
12V -> 5V regulator -> 3.3V regulator -> 3.3V perpect straight line by CRO no noise, NO ringing.

The same programmer with same cable other board are working fine.

Problem Description:

  • After initial 2-3 times flash programmed application operations , I am no longer able to connect using the XDS110.

  • The installed program working perfectly, The LCD operates correctly and the firmware executes as expected. But no further DUBUG connection possible.
  • CCS reports the following error:

"Texas Instruments XDS110 USB Debug Probe/CS_DAP_0 Error connecting to the target: DAP Connection Error..."

Observations:

  1. Double cross check and even replace NRST capacitor 0.01uF/50V, 0.47uF on Vcore, 0.1uF + 10uF on VCC.
  2. Triple connection check pin to connector NRST, SWCLK, SWDIO found correct.
  3. The same XDS110, CCS installation, PC, and 15 cm SWD cable work perfectly with another identical PCB.
  4. The faulty board continues running the application normally even though the debugger cannot connect.
  5. While probing SWCLK and SWDIO with an oscilloscope, the debugger unexpectedly connected successfully and the Factory Reset completed.
  6. Without the oscilloscope probes attached, the debugger again fails to connect.
  7. Resistance measurements with power removed are very similar:
  8. Resistance between  SWCLK to GND = 2.65 mega ohms.
  9. Lowering the SWD clock to 100 kHz does not resolve the issue.
  10. Adding an external 10 pF capacitor from SWCLK and SWDIO to GND does not reproduce the improvement observed when the oscilloscope probes are connected.
  11. I pull down 100K on PA18(BSL), but no effect.

Any suggestions or recommended debug procedures would be greatly appreciated.

Thank you for your support.

Best regards,

Tulsi Das

  • 0 in reply to Tulsi Das
    TI__Expert 7455 points

    Hi Tulsi,

    Because the debugger connects when oscilloscope probes are attached, the MCU is probably not permanently locked. This strongly suggests a marginal SWD connection, grounding issue, or board assembly defect.

    Please check the following:

    1. Connect the target’s 3.3V rail to the XDS110 VTREF/TVRef pin. VTREF does not power the board; it provides the probe with the target I/O voltage reference.
    2. Improve the SWD ground path. Keep SWCLK/SWDIO short and add a ground wire adjacent to SWCLK.
    3. Check continuity directly from the debug connector to the MCU pins. Inspect SWCLK, SWDIO, NRST, VTREF, and GND for poor solder joints, damaged vias, or flux contamination.
    4. Verify NRST has the recommended pull-up and approximately 10–100nF to ground.
    5. After correcting the hardware connection, perform the MSPM0 manual mailbox factory-reset procedure.

    The 100kΩ pull-down on PA18 is unrelated to the SWD connection problem. Also verify that the programmed NONMAIN configuration has not disabled SWD, but the oscilloscope-dependent behavior points primarily to a physical signal-integrity or grounding issue.

    -Brian

  • 0 in reply to Brian Dempsey
    Prodigy 120 points

    Hi Brian,

    Thank you for your reply.

    With my other running projects, I have programmed many thousands of MSPM0G3507 devices and have never encountered this type of SWD connection issue.

    In my current project, I have two boards (a Control Board and a Display Board) communicating via UART. Both boards exhibit the same SWD connection issue. I have checked the SWD connections multiple times and they appear to be correct. I also reduced the cable length and even soldered the SWD wires directly to the MCU pins, but the issue persists.

    The Control Board has an onboard SMPS that powers both boards. I suspect this SMPS may be the root cause because when I power the boards from an external battery instead of the onboard SMPS, SWD communication works reliably most of the time (although it still fails occasionally). I am currently investigating this further and will let you know the results.

    I also checked the MCU 3.3 V supply using an oscilloscope. The VDD rail appears clean and stable, with no noticeable noise or ripple.

    Regarding your suggestion to connect the target's 3.3 V rail to the XDS110 VTREF (TVRef) pin, I could not find this pin on my XDS110. I am using the XDS110 debugger integrated on the MSPM0 Launch Pad to debug an external target. Could you please let me know which pin corresponds to VTREF on the Launch Pad?

    Below are connections. I removed all jumper disconnect internal target and connect my custom board.

    Best Regards,

    Tulsi Das