If possible, could you run the same tests on an ODAC with and without the external PS and post the results? I think that many here including myself will be interested in your findings.
Well, if I had an ODAC - I would. But I don't, so I can't. Even though we've had disagreements in the past, I would defer to purrin's measurements/experience using an ODAC with a linear-regulated external power supply (it might've been the iFi, I can't remember). If my recollection is correct, he confirmed that there were differences.
There's also this:
running off a dedicated USB PCI on my desktop without other devices connected, into either a B22 or QRV-08, and AD2000 or modded T50RP headphones, the ODAC is etched, bright, rough and harsh. makes the BM DAC1 i recently sold after 6 years of enjoyment sound warm, smooth, and analog. with the Doodlebug inserted, less bright/etched and more fleshed out. more details and more natural sounding. still not my cup of coffee, but the Doodlebug is an improvement in my rig.
maybe my desktop USB power really sucks, or the implementation of the secret sauce Sabre is black magic. either way, isolation and clean power brings improvements.
huge props to A_A and tomb.
from a build perspective, the pads are generous and require a little extra heat. high quality board and a perfect first time SMD project.
Granted that's a subjective opinion, but one from a very-experienced, high-end DIY-er.
The other thing to consider from an engineering perspective, is that it's logical and a conservative application. It removes additional unknowns that have the potential to affect sound quality. It's in the direction of better - with zero risk - except for a small investment.