I am currently trying to implement the TUSB8044A as a standalone hub. Due to an error in the connector design, in our current prototypes we can connect the usb 2.0 lines OR the usb 3.0 lines. We are now reiterating a new version, but i wanted to make sure to fix every issue before resubmitting for a new prototype run.
Currently, when connecting the usb 2.0 lines the hub is recognized without issues; the port gets enumerated and I can see the entire hub in tools such as USB treeView. Downstream ports can also successfully enumerate any 2.0 device that is connected. The same cannot be said unfortunately about the usb 3.0; if i connect the SS lines the hub does not get enumerated. Not sure if relevant, but I saw that when the usb 2.0 lines are connected the crystal starts oscillation as normal, but that is not the case for the USB 3.0; rails are similar in both cases (a smidge less noisy in the 3.0 case, but I assume that is related to the hub not doing much, since the crystal did not start oscillating).
Hence, I have a couple of questions:
- Should the device function when connecting only the usb 3.0 lines (as far as data lanes go; all rails, bias resistor and configuration are connected as normal) and without connecting the usb 2.0 lines?
- Is there any suggestion that you can provide in order to debug or pinpoint the issue here, so that I can fix it in the second prototype?
I have opened a ticket with TI support, which pointed me to this post. The checklist highlighted there should be ok, the only difference being that I used 220 nF series capacitors on the TX lines instead of 100 nF ones for the sake of BOM optimization; can this be a problem? Additionally, I don't have series caps on the RX lines, with the assumption that the lines are ac-coupled using capacitors mounted near the host (a laptop, in this case) transmitter.
Thank you.
Best regards
