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.

CCS: XDS110: Found 0 devices

Other Parts Discussed in Thread: CC1310, ENERGYTRACE

Tool/software: Code Composer Studio

I can see a TI XDS110 debug probe in Win10 device manager. If I plug in an actual XDS110 probe OR a CC1310 launchpad, the entry shows up without problem.

But Code Composer 8.2 doesnt' see it. And if I run 'xdsdfu -e' it reports "found 0 devices"

How can this be?

  • If I move this probe and cable back to laptop, xdsdfu sees it.

    On the desktop, I have uninstalled probe, specifying "remove software" over and over until the xds110 enumerates as an unknown device. And then re-installed driver. Still nothing.

    Laptop device manager reports "XDS110 Class Debug Probe" as version 4/5/2016 version 1.7.0.0. This is CCS 8.2 environment.

    Desktop reports the same.

    I cannot understand how Windows can enumerate this board and yet XDSDFU -e cannot. Also, RF Studio shows a hint of activity when the Launchpad is plugged in. It shows a spinning wheel graphic for a second which is interesting. But it never shows anything in the list of devices.

    I am at a loss where to go next on this.
  • Hi,

    I haven't seen this behaviour as well - especially if the device drivers are properly instantiated. 

    Just to confirm, is that what you see in your Control Panel?

    One additional detail: can you open an elevated command prompt and re-issue the xdsdfu -e command? 

    Also, can you re-try in other USB ports? In certain systems the USB version (2.0 or 3.0) may make a difference, as well as powered/unpowered HUBs. 

    Apart from this I am unsure what else may be at play in this scenario exclusively with the desktop.

    Hope this helps,

    Rafael

  • Hi Rafael, thanks for your quick response. Yes, moving to a new port fixed it. My PC has 2 USB (480Mbps max) speed ports for mouse/keyboard (black connectors), 4 USB 3 ports (blue) and 4 more 480Mbps (black) connectors. None of the USB 3.0 connectors work although they all enumerate and show as you have shown in HW manager. 

    I cannot remember ever encountering a product, even a USB mouse, that cared about ports

    I use several JTAG solutions in my daily work, and TI's is, by far and away, the most fragile. Once it's working it is fine for that session, but can sometimes take hours to get working again (as above).And the JTAG boxes are getting bigger and bigger, with more cables and adapters than ever before! No keys on any cable. The XDS110 cable has a red stripe on the cable, but that isn't pin 1! The pod requires a 180 degree rotation from what seems to be pin 1. The LEDs are hidden deep inside the box, and you have to peek inside a hole in the endcap to actually see what is going on.Not refined at all!

    On the flip side, Xilinx has really nailed JTAG on their Vivado platform. One thumb-sized pod for everything--from their $8 FPGA to their $1000 FPGAs, with a tiny universal connector that the keyed pod plugs into. The tools are bulletproof. Plug in or unplug at any time, the dev environment doesn't care. Ever. It diligently connects and starts logging FPGA core temperature on a dashboard panel. Accidentally short out a supply on the board, the JTAG doesn't care. It reconnects when it can and offers to re-download. Programming a 2Mbyte parts happens wicked fast. 

    If you are looking to study a direction on jtag for future improvements, please consider xilinx's vivado the high water mark. Their toolchain had gotten really bad at the end of ISE, but Vivado is impressive. 

  • Ugh, another day, another hour lost trying to sort this out. Now no USB ports work. But as before, the OS still sees the probe. And curiously, TI SmartRF studio can see the probe this time. But xdsfdu and CCS cannot. 

    I ask again: How can the OS (and SmartRF studio) see this but CCS cannot? This was not this bad with 8.0, but it's unbearable on 8.2. 

    What else can I try? I am at a loss.

  • Hi,

    I am sorry to hear about the issues you are having; a few comments:

    - The issues with USB ports and other details were never fully characterized internally and therefore have become quite an elusive issue to address - your case seems a rare extreme one, given that many developers use this product and the number of reports is proportionally small. I personally have tested the XDS110 in a variety of USB 2.0 and 3.0 ports/powered HUBs/unpowered HUBs in different brands of equipment with the rare glitch that tends to correct itself without a reproducible cause. What I can promise is that we will keep testing our products and investigating possible causes for this issue.  

    - The XDS110 is the most comprehensive pod of our product line - it comes with a number of pin adapters to address many legacy standards while implementing the new variants of JTAG debugging (1149.7, SWD/SWO) and having reasonable performance across our device families, ranging from the 16-bit DSPs, 32-bit single-core MCUs up to 10+ core SoC devices.

    - The enclosure size was certainly a factor debated internally, but the use of a standard and moderately sized enclosure helped with the certification requirements and to accommodate the EnergyTrace HDR accessory.

    - I didn't understand the remark about the red stripe not matching pin 1. Do you have something different than what is shown below?

    - The LEDs were a compromise solution so we could improve in other areas of the product such as the pin adapters or the breakout signal board which aggregate much more usefulness. I am not a big fan either. 

    Regarding your last report, the screenshot shows the SmartRF connected, which takes the driver hostage and prevents any concurrent connections to it (the xdsdfu wouldn't recognize the probe anyways). At any rate, the system you are using somehow seems to have a number of problems with this probe - especially considering that your laptop seems to be getting along with it. 

    What is the brand/model of your desktop? Who knows? Perhaps we have something similar around here to try it out. 

    Regards,

    Rafael 

  • Hi Rafael, thanks for your continued help.

    The "found 0 devices" happened whether or not the SmartRF was running. The problem was eventually fixed by a reboot of the desktop. This isn't exclusive to my desktop. It's not a great experience on my laptop either. Yes, ti works for many days on end, but when the problems start, it means a re-boot is coming

    Regarding the stripe...

    I use a 0.05" pitch 10 pin connector on my boards (same as launchpad). The breakout board that came withthe XDS110 probe is keyed. When the red stripe is plugged into the keyed breakout out, the red strip is located on the pin 9/10 side of the connector. Thus, it needs to be plugged in "upside down" in order to work. Also, there's no key on the provided 10 pin breakout board. This means every time you want to plug in the connector--even into a shrouded receptacle such as on the Launchpad--you must carefully align the pins visually. Plus, the combination of adapters doesn't sit correctly on the launchpads. the SWO pin on the 0.1" header interferes with the PCB on the XDS110 adapter board.

    Maybe I missed a connector combo?