• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 29 subscribers
  • Views 1551 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.

TMS570LS0432: Cannot connect with SEGGER J-Link EDU

Thomas Meier
Prodigy 50 points
Part Number: TMS570LS0432
Other Parts Discussed in Thread: SEGGER,

Hello,

I have a SEGGER J-Link EDU that I am attempting to connect to program/debug a TMS570LS0432 with. I am using this microcontroller on my own custom board, and have gotten the debugger to work with a development board for the microcontroller (TMS57004 Launchpad). When attempting to connect to the microcontroller, I receive the following error:

I'm sort of at a loss as to what the hold up could be. Some things of note that I've noticed/tried so far are:

1.) I've tried putting a 2.7k pullup from nRST to VCC on the JTAG header to ensure that the device is in reset when attempting to program, as per this thread, but it doesn't seem to fix the issue.
2.) The connections from the debugger to my board aren't using the 20 pin ribbon cable that came with the J-Link, as my board doesn't follow any standard pinout. I'm only connecting one of the ground pins on the debugger to the board, and although I'm sure it would be ideal if I connected more, I can't imagine this being my issue (I'm not trying to connect at a high speed).
3.) I've referenced the schematic for the TMS57004 and there's nothing different between my debugging circuit and the one on the Launchpad. I don't have any pullups on TMS, TDI, or the clock lines, but these are marked as DNP on the schematic (which seems off to me, but I'll trust it).
4.) The light on the J-Link turns orange after I attempt to connect and stays that way until I close J-Link Commander.

I'm running a little thin on ideas as to what could possibly be wrong with my setup. It's certainly possible that there's an issue with my board, but as far as I understand, the only thing that could cause a problem would be an issue with power supply, and that all seems to be in order, so I'm kind of at the end of what I can check without actually connecting the debug probe. Does anybody happen to know why I can't connect the J-Link EDU to a custom board? Are there any hardware/design considerations I may have missed? Is there any way to check if the microcontroller is even working if it's never been programmed and I can't connect to it with my debug probe?

Thank you for your time,
Thomas Meier

  • 0
    TI__Guru**** 186256 points
    Hi Thomas,

    I used JLINK on TI HDK before and it works. Did you try other emulators (for example XDS100V2) on your board? Did you probe the JTAG signals on the board if the TCK is clean?

    Is the nTRST pulled high or low on your board?
  • 0 in reply to QJ Wang
    Prodigy 50 points

    nTRST is pulled high by the debugger when attempting to program, and I've verified this with an oscilloscope. It is not held low or high by anything on the board, but I've also tried to just hold it high with a 10k pullup, with no change.

    I don't have a good scope on hand to accurately assess that the clock signal is clean or otherwise, but I've captured the signal with a logic analyzer and it seems to be consistent, if not a little beyond my ability to interpret. At 4000KHz programming speed, every eight pulses, there is one that is slightly longer than the others.

    I'm guessing that since it's so consistent, this is an inherent part of how JTAG works, although I'm including this note in case this is *not* intended.

    I don't happen to have an XDS debugger of any sort on hand, unless there is a way I could somehow route the one from a development board to my own hardware.

  • 0 in reply to Thomas Meier
    TI__Guru**** 186256 points
    Hi Thomas,

    The nTRST should always be pulsed low at power up to initialize the debug logic, so I recommend to pulldown nTRST on your board.
  • 0 in reply to QJ Wang
    Prodigy 50 points

    After screwing with this a bit more, I have a couple questions. I'm guessing its an issue with my hardware.

    1.) I have continuity across the Kelvin ground and the power supply ground. These are definitely not connected on my PCB; are these connected internally in the microcontroller?

    2.) I'm having an interesting issue concerning grounding because of a missed trace on my PCB. The schematic is attached; the -3.3V net (which is a misnomer, this is ground) is not connected to the GND net that the processor/1.2V regulator is on. This was fixed by adding a jumper between the GND and -3.3V nets. However, upon throwing a logic analyzer at it, I noticed that I get absolutely no activity on TMS, TDI, TDO, etc. I start getting things on these lines when disconnecting this ground. The interesting thing is that with these grounds left unconnected, and before attempting to program, I get 0V between the reference pin and output pin on the regulator, but after attempting to program, I get 1.2V across these. Note that the ground ("-3.3V') on the JTAG header is not connected to the same ground that the microprocessor is on; I have absolutely no idea how the debugger attempting to connect seems to ground the regulator.

    Is there anything I don't know about the microcontroller that could explain this? This is obviously a design issue, and is probably way outside the realm of what you're able to answer, but I though I'd ask in case you happen to know anything that could explain it.

    Thank you for your time,
    Thomas Meier

    8623.Schematic.pdf

  • 0 in reply to Thomas Meier
    TI__Guru**** 186256 points
    Hello Thomas,

    1. Kevlvin GND is connected to the other ground pins through metal on die. The only difference between Kelvin GND and the other GND pins is that: Kelvin GND is the "local" GND for the oscillator module. In other words, Kelvin GND is the GND pin next to he oscillator module physcially. Connecting Kevlvin GND to board GND should not cause abnormal operation of oscillator even if it is recommended. It may impact the Jitter of the oscillator.

    2. I read your schematics and noticed that the -3.3V, -5.5V and GND are connected together at power supply chips (U1, U2). I also saw another strange thing: VCCP, VCCIO1/2/3/4, nRST, and nPORRST pins are left unconnected.

    VCCP is power supply for PLL. VCCIO is power supply for I/Os. They have to connect to power supplies: 1.2V and 3.3V
    If you don't use nRST and nPORRST, please pull them up.

    3. nTRST (JTAG signal): please pull this signal down