Other Parts Discussed in Thread: TUSB211, , USB-REDRIVER-EVM
I've got some boards made that implement the TUSB501-Q1 redrivers for USB 3.0 SS and a TUSB211 for High Speed. The design is based upon the USB-REDRIVER-EVM.
Changes which deviate from the USB-REDRIVER-EVM are that to ease routing I swapped polarity, which is allowed per the TUSB-501-Q1 datasheet Layout Guideline. I also added single channel ESD to each of the SS lines, added some additional decoupling caps on the power/gnd, and where possible I increased the 0201's to 0402 and the 0402's to 0603.
When I connect the device to the computer, USB 2.0 HS works great as it should, but I get no USB 3.0. I'm using USB Device Tree Viewer V3.1.6 and I can see that the device flags 1 for "Is SuperSpeed capable or higher" but no matter what flags 0 for "Is not operating at SuperSpeed or higher". I get both 0 and 0 flags if I configure things completely wrong, or remove a line altogether. This leads me to believe that the computer recognizes the USB 3.0 SS lines but something is out of whack so it won't use them. When I switch my design inline with the USB-REDRIVER-EVM instead it works fine in 3.0 SS mode. The highest speed oscilloscope I believe I have access to is 1GHz / 5GS-s (6GHz if I could find the probes for that...) so AFAIK I can't visually analyze the signal (if anybody knows of a sub-$5k way to view USB3 eye diagrams I'm listening).
To prevent any sort of chance of compounding issues through the troubleshooting process I've tried different configurations on 6 boards thus far--so in case one configuration burns something out I'm not re-using that same board with another configuration. I've also tried removing the ESD IC's in case there is some conflict with the ones I used.
In my design, when looking for problems, I've noticed that my design may be swapping controller/device, so to test this I reconfigured one of them so I can plug it in the other way around (the USB-B side plugging into the device instead of the computer, and respectively for the USB-A side). I've tried each jumper setting on each configuration.
The boards were manufactured by Sierra Circuits and I had them specified to a tolerance of +\-5% for differential pair impedance.
Please take a moment to look over my design and lend a hand if you are intimately familiar with high speed layout or the TUSB501-Q1 :) Much appreciated.
Schematic [Image]: https://drive.google.com/open?id=13IEpxX3Fg3p08x9jgqa_mD0TSwIKPEv5
Layout [Image]: https://drive.google.com/open?id=1poFoci4M8QHu7aNUZOdUdfavFijrezUU
Boards [Shown in different troubleshooting iterations; Image]: https://drive.google.com/open?id=1iqjGDdspsbnm3LnoZqCLzwgcJ8ye16DC
All relevant files (incl. layout file [Altium]) are located here: 
