Intermittent shorts causing problems

[tyguy2]

Hello all,

Currently I'm working with the TI chip TUSB9261 trying to connect an MSATA drive to a computer over USB 3. I've modeled the board off this reference design. I've attached schematics of my current implementation along with the reference design, however I'm having some very strange issues. Without the chip installed, the resistance between the 5v, 3v, 1.8v, and 1.1v rails to GND is all in 50k-1M ohm range, which makes sense considering all that should be connecting them is bulk capacitance. When I power the board without the chip mounted, the power supplies all behave normally, outputting the correct voltages.When I soldered the chip to my PCB, however, the resistance across the 1.1v  to GND rail dropped to a steady ~420 ohms. When I connected the PCB to my computer, Windows detected it, but gave me a "device malfunctioned" error, and it quickly disconnects. This looped for a while, with the board connecting, "malfunctioning," then disconnecting. After a while, the board ceased connecting altogether, and I measured the resistance across the 1.1v rail, which was now close to ~20 ohms. I left the board overnight and measured the resistance across the rail again, at which point it has returned to ~420 ohms. Removing the chip from the PCB gets rid of the short altogether.

I'm not really sure what to make of all of this, or how to fix it. Does anyone have any ideas or advice?

[tyguy2]

Quick update: Did some snooping, looks like the bulk capacitance on the 1.1v rail failed for some reason, causing the short. I cut the traces to that area of the board and ran a few wires to a perf board with new caps on it, and the short disappeared. Better yet, looks like things are starting to connect to the computer as well.

[tyguy2]

Another update: Looks like the short has returned. I removed the chip from the board, it looks like the short is internal to the chip. Any ideas?