Can anyone explain these benchmark results?  

Yeah, there is NO WAY that your SATA interface is getting between 6GB and 9GB a second throughput.

Your NVMe drive is pretty much right where it should be. The SATA drive is going to be about 1/5 of that.

I mentioned the possibility of Rapid Mode affecting the benchmarl results for the SATA drive in my first post (second paragraph.) Since this drive get's used mostly for small read/write operations I want to keep Rapid Mode enabled.

I thought maybe Crystal ran enough data through the drive to overcome Rapid's caching ability, but I guess not. So the bottom line is - no real problem here after all.