I can completely relate to the drive to optimize the USB path. Years ago, I was firmly in that same camp, deploying the iFi Micro and Nano iUSB3.0 power supplies, inline iPurifiers, and Gemini split-data/power cables across my main and secondary systems. When you are wrestling with standard computer audio outputs, adding these stages absolutely changes the sound, often bringing a welcome sense of calm and transparency by shifting the leakage current and power supply noise spectrum around.
But for the lurkers following this thread who are looking for true isolation, it’s worth stepping back and looking at the root cause of the problem.
Cascading five different USB isolators, reclockers, and hubs introduces a massive amount of hardware complexity, multiple power supplies, and up to ten physical impedance boundaries where high-speed digital signals can reflect and create deterministic jitter. It’s essentially treating the symptoms of USB's inherent liabilities—packet bursts and erratic CPU polling that generate EMI—with a long chain of hardware band-aids.
I've since discovered a much more elegant, modern way to achieve this goal at the protocol layer rather than the USB hardware layer.
Moving to an architecture like Diretta—specifically running Diretta Direct Stream (DDS) Mode 3 L2 frames over a simple, point-to-point CAT6 or CAT6a patch cable between a Host and Target—provides complete physical and galvanic isolation natively via the network interface. More importantly, it continuously paces the audio data to keep the endpoint’s CPU cycles perfectly flatlined and uniform. By eliminating the erratic processing spikes that plague USB transmission stacks, you stop the electrical noise from being generated at the source, leaving nothing to mitigate.
If you enjoy the science experiment and the puzzle of mixing and matching different isolation chips, the multi-box USB route is certainly a path you can take. But if your goal is an elegant, low-clutter system with maximum isolation, fixing the transmission architecture via a protocol like Diretta renders the entire "decrapifier" chain obsolete.
If you are at all curious, you can read my open source DIY "cookbook" for solving this problem architecturally here: Building a Dedicated Diretta Link with AudioLinux on Raspberry Pi

