Debugging USB on PIC32 (USB-host HID)

[gurksallad]

I have made a small PCB with a PIC32 on it. Using MCC I configured my project to use USB-host with a HID TPL. Then I extracted the keyboard example from the latest Microchip MLA, and I'm basically using app_host_hid_keyboard.* right into my project, just for testing. I added a lot of debug-messages in the functions as well just to get a grasp of where the code is running. Using MCC I also added a UART, so I can print the debug-messages on there and watch on my PC.

When running the project and attaching a keyboard, I see this on the debug terminal:

Code:

HOST: Initializing DETACHED state.
app_host: Attach keyboard
[>HOST: Starting settling delay. 

And then nothing happens. So obviously the code is running, the USB-stack is up, it recognizes a USB-device is inserted, but then goes dead. I have not changed anything in the USB-stack or in the example code from the MLA.

Any tips of where I should go from here?

[UKFlyer]

If you are outputting the debug messages using polling I would remove them as they will disrupt the timings of the USB. For debugging this sort of thing I use LEDs that can be turned on to indicate progress. This can be done without impacting any timing critical activities.

[towlerg]

Join the Axelson (author of USB complete etc.) forum http://janaxelson.com/forum/.

[gurksallad]

If you are outputting the debug messages using polling I would remove them as they will disrupt the timings of the USB. For debugging this sort of thing I use LEDs that can be turned on to indicate progress. This can be done without impacting any timing critical activities.

Hm, that's a good point. I do indeed use USB-polling, so maybe that's the problem here.

Unfortunately I cannot add LED's to the board as it's already fabricated, and hand-solder LEDs straight onto a TQFP-48 is a bit tricky. However, I did add a single "activity LED" on the board. Maybe I can bit-bang some binary numbers to that pin and measure with a logic analyzer to get a grasp of where the running code is at.

[gurksallad]

I poked around a bit and it seems the stack is stuck in this case switch after the keyboard is attached:

Code:

                        case SUBSUBSTATE_WAIT_FOR_SETTLING:
                            // Wait for the timer to finish in the background.
                            LED_Toggle();
                            break;

So it seems that the USB-connection is never settling. Now I just have to find out what that means.