Have you done double blind tests or measured the output?But I'm not here to preach something. It's just my humble opinion after actually trying it myself.
I believe you when you say it sounds different but I wonder if it is possible to identify the measurements.
As I keep pointing out, the analog output from the D90 and SA-10 can be shown to be audibly identical using PK metric calculations to the raw digital file but just hit the threshold of audibility when comparing the two recordings, suggesting that it is possible that they sound different. And the SA-10 is theoretically closer to the digital file.
I cannot reliably ABX it, but there are plenty of times I bought something new and cheap and it displaced the old expensive and times where I bought new and cheap and I returned the new one.