My current computer is also struggling in the USB area, I just have too much USB stuff.
If you're like me, you have "too many" USB 2 devices, and no USB-2-to-3 "converting" hubs. (Essentially, when using USB 2 devices, each root port is limited to 480 Mbit/s throughput, even though the root port is an USB 3 one. This is because USB 2 and USB 3 use physically different transceivers on each port.)
I've also found that for the USB 1.1 (Full speed, 12 Mbit/s max) stuff, using an USB isolator followed by an isolated wall wart (non-grounded, fully isolated USB supplies) powered USB hub, can make a lot of things more stable. Especially so for stuff I designed myself and am working on
.
I'm very eagerly awaiting cheap high speed isolators, now that both TI (ISOUSB211) and Analog Devices (ADUM4165) have released high-speed capable isolator chips. A simplified version of the
Olimex USB-ISO (for USB low or full speed only, no high speed support), with a downstream USB power input port (so that one can use a known good isolated USB power supply (no ground pin!), ubiquitous here in Europe, as long as one does not fall for the crappiest-of-the-bunch ones) to power the device.
In fact, I've already taken a look at the ADUM4165 and ISOUSB211 data sheets, with the intent of making my own heavy-duty isolators with Mean Well medical power supply (say, RPS-30-5 or RPS-45-5) for the isolated downstream. I'd love to have multiple separately isolated USB 2 ports (with a ganged cable to my computer), but I don't know enough about (isolated) DC-DC power conversion to work out how to do that with just one (possibly not so well isolated) power supply, without risking all sorts of EMI issues. All of
my designs thus far are based on datasheet application recommendations and TI Webench schematics. (With a microcontroller on the host side of one of the ports, for both monitoring the voltage and power use, but also for disconnecting both the power and the isolated port as needed. I'm especially interested in how "spiky" the current draw of a specific USB device is, at the sub-millisecond scale. It would be a very useful SBC test bench, especially if it had isolated, controlled and monitored 5V/9V/12V available for powering the target SBC.)