NUCLEO-F103RB USB Virtual COM Port Ceases to Function after 2-4 minutes

[jcrollman]

Posted on July 02, 2015 at 16:07

I have my Nucleo-F103RB virtual serial port over USB working for a short period of time. Usually around 2-4 minutes it will continue to push data across to my terminal. Then serial data will stop. USB is still functions, I can program the Nucleo just fine after serial data stalls. I see data being sent by the application micro (via serial) to the STLINK micro. No error is reported by windows device manager, the COM port is still functioning as far as Windows 7 (64-bit) is concerned. But serial data coming into the STLINK to the from the application micro never makes its way through to the USB. Reprogramming the micro or resetting the micro doesn't fix it. The only way to make it work again is by unplugging the nucleo from the computer.

I installed the STLINK009 is installed in windows. If I were a betting man, I'd say there is something wrong with the STLINK firmware, possibly sketchy support of USB 3.0.  I've tried a couple of older computers, and they all function correctly.

I've changed the baud rate, the data being sent, the com port number. I've tried multiple USB ports. I've tried every STLINK firmware available. Same result.

Anyone familiar with this issue?  I am working on a Dell Precision M3800.

[S C]

Posted on July 10, 2015 at 14:14

Thank you for all information.

I tried your application during more than 2 hours on an USB2 port and half an hour on an USB3 port, without any issue. Moreover I notice that the data rate is very slow compared to the stress tests we are doing, so I don't think no more this might be an overflow or rare boundary issue. I have a similar system (Windows 7 enterprise 64bits), older USB chip (Intel 7 Series), but I don't think this might explain the difference. I saw ''lot's'' of USB devices in your configuration, especially composite ones. Do you have some other virtual com port simultaneously connected ? Do you have any application that might interfere on a COM port with your terminal ? I don't know how to help you without reproducing the issue on our side; I'm afraid this is a kind of issue requiring an USB trace analyzer as starting point

[Tesla DeLorean]

Posted on July 10, 2015 at 16:28

I'm pretty sure it's not in the ST-LINK firmware or the driver side, which are common to all Nucleo's, here the F401RE is banging on for close to 60000 seconds.

...
This program runs since 59439 seconds.
This program runs since 59440 seconds.
This program runs since 59441 seconds.
This program runs since 59442 seconds.
This program runs since 59443 seconds.
This program runs since 59444 seconds.
This program runs since 59445 seconds.
This program runs since 59446 seconds.
This program runs since 59447 seconds.
This program runs since 59448 seconds.
This program runs since 59449 seconds.
This program runs since 59450 seconds.
This program runs since 59451 seconds.
This program runs since 59452 seconds.
This program runs since 59453 seconds.
This program runs since 59454 seconds.
This program runs since 59455 seconds.

Is the system in question going into low power, power saving or screen saving type modes?

[jcrollman]

Posted on July 10, 2015 at 21:31

No, I'm always using it. The computer is brand new, issued by IT a few weeks ago - I don't know what could be installed that might cause issues.  It does have one of the USB docks that I use. But I've unplugged it from the computer and it made no difference.

[Tesla DeLorean]

Posted on July 10, 2015 at 23:18

I don't know what could be installed that might cause issues.

Corporate IT a well known for loading all kinds of crap, and changing settings/policies average mortals would not. Aggressive AV, restrictive user rights and locking down USB access spring to mind. I guess you'd have to ask them? Attached is a quick blind port to the STM32F103RB-NUCLEO, you'd use the ST-LINK Utilities to push it on board. This was built with Keil, not mbed, and equivalent to the F401 code I've had running for 83K seconds now. If you've got multiple Nucleo's you could try another one to rule out some oddness related to the one currently disconnecting/failing. ________________ Attachments : STM32F103RB-NUCLEO-USART2-Rel1.hex : https://st--c.eu10.content.force.com/sfc/dist/version/download/?oid=00Db0000000YtG6&ids=0680X000006Hzxd&d=%2Fa%2F0X0000000bRS%2Fm0iZqQlVhSyFUstjYFdsc1sB9av5UeKqDQlkiYZ1Z0w&asPdf=false

[jcrollman]

Posted on July 10, 2015 at 23:50

The odd thing is that even after the failure, the nucleo still programs fine through the same USB. So its not like the USB isn't functioning, it is, but not for the VCP.  Reprogramming the board doesn't bring the VCP back up either. Only unplugging and plugging back in.