TM4C1233H6PGE: USB Pipe Reset

Hi,

Andrew Worwood said:

I can initiate the communication with the system, but at some point I see the host computer issue a pipe reset command down to my isolated system and then my communication is lost.

How long does it take for this happen? Seconds, minutes or hours?

Can you repeat the same problem running the same code on a LaunchPad?

Can you repeat the same problem on a different PC?

Can you repeat the same problem using a different USB cable?

Can you repeat the same problem on another of your custom board if you have one?

Do you have a USB analyzer to find out if any type of errors occurred before the host issued a pipe reset?

Please also refer to TM4C123 System Design Guideline for best practices on USB design. Refer to section 4.1 https://www.ti.com/lit/pdf/spma059

Andrew Worwood said:

Is there something I need to do in the embedded code to handle this pipe reset error condition to get my communications back?

What USB class device are you enumerating with the host? 

This usually occurs within a minute or 2.

I don't have a launch pad to try this on. I can get one if need be though.

Yes, I can repeat this on a different PC.

Changing the USB cable can slightly change the behavior, It might take a different amount of time to happen.

I do have another custom board that I cannot repeat the behavior on. It uses the same processor and design for USB, but it is powered differently. ie powered by the USB itself, and it has a different layout.

I don't have a hardware USB protocol analyzer, we just used a free USB sniffer software. I also have a differential high-speed probe, but I haven't seen a specific error with that yet.

we are enumerating a CDC USB device.

HI,

  I don't know what is the cause of your USB problem. I will suggest you compare your schematic against the LaunchPad and the best practices documented on the TM4C123 System Design Guideline to see if you find any subtle differences. 

 Can you run the stock USB CDC class examples on your board? Will you see the same problem? You can find the example at C:\ti\TivaWare_C_Series-2.2.0.295\examples\boards\ek-tm4c123gxl\usb_dev_serial.

Andrew Worwood said:

Yes, I can repeat this on a different PC.

Which Windows version do you use? Will a different Windows version make a difference?

 Please also check if the Windows is picking up the INF file for CDC class at C:\ti\TivaWare_C_Series-2.2.0.295\windows_drivers

I get the same issue when I run the stock USB CDC example. Windows version makes no difference.

HI,

  I think the best way to diagnose the problem is for you to acquire a LaunchPad ($16.99) and see if you can repeat the same problem. If you can get hold of a USB analyzer then it will certainly help but that is a much more expensive tool than a LaunchPad. 

https://www.ti.com/tool/EK-TM4C123GXL

Hi Andrew,

  I have not heard back from you. Do you have any update? If you have resolved the issue, can I close this ticket?

I was getting similar issues on some other devices. I did try using using one of the TI USB isolators and I seem to be having better luck with that. Although I am having a different issue now with USB. I am in a scenario where I need to stream back data as fast as possible. But during this blast of data it appears as though the usb buffer goes from max space available(1024 bytes) to 0 space available almost instantly and stays with 0 space available until I am done trying to send data back( up to 360S later). Should I continue this chat here, or make a new forum more related to this new issue?

Hi Andrew,

Andrew Worwood said:

I was getting similar issues on some other devices.

   When you said other devices, do you mean a different TM4C123 or another MCU from other vendors? Although I don't know really know the cause to your issue, it seems like it is not really related to the MCU if a USB isolator can make the problem go away. 

Andrew Worwood said:

But during this blast of data it appears as though the usb buffer goes from max space available(1024 bytes) to 0 space available almost instantly and stays with 0 space available until I am done trying to send data back( up to 360S later). Should I continue this chat here, or make a new forum more related to this new issue?

 Please open a new thread for a different and I can close this one. Have you tried to increase the buffer size to at least the size of your received payload? Does it make a difference?

We have a couple of products that use the TM4C123 MCU, one of them showed no issue at all, and another one did show the same issue. we are going to do more testing with the new USB isolator to see if we have any more issues.