In the nicest possible way I’m so happy to see others having the same issues as me, searches on the net and not finding anything similar made me think I was going mad!
The issue can be laid firmly at the door of MSI.
I purchased the X570 Content Creation when it came out and ignored the one tech site review that said strangely the EC-NVME Sabrent enclosure kept failing when plugged into this Motherboard – they put it down to immature drivers. In September 2019 I purchased my drive enclosure (the Sabrent) and sure enough it either wouldn’t show up when plugged in or it would register then fail or you’d see it but AS SSD type tests would either see it stall and freeze the PC or provide ridiculously low transfer rates (10MBps).
Anyhow, I gave up and resided myself to a waste of money…flash drives etc worked so I thought clearly not the Motherboard. Fast forward to this week. I now have three NVME enclosures with a range of NVME sticks inside. Plugged them in and all gave wildly different results in any disk testing software, the Sabrent (with a Sabrent NVME inside) still stalled and locked up the PC or ran at 10MBps, others provided some data transfer but not in line with USB 3.1 (Gen 2) speeds. Didn’t matter whether I used the front panel plugged into the motherboard or directly in the back. Sometimes one of the drives would give near max transfer rates then 30 mins later it wouldn’t be recognised. Pulling my hair our (or I would if I had any left!)
Then I suddenly remembered that I now had a 2080TI that I didn’t have when I first got the motherboard the year earlier and it has a USB-C port running outside of the motherboard control and using Nvidia drivers. All my drives whizz along at 3.1 (Gen 2) speed, all recognised instantly!! So effectively replicating Tesla’s result who used a separate PCIE card.
Back in 2018 there was concerns (on an X470 board) that the traces MSI use to go from controller to port were longer than other Mobo manufactures. There was a thought that this reduces either signal strength or power by a very small margin but enough (when dealing with single digit millivolts) to give the sort of flaky issues we have suffered when using NVME type enclosures and not using external power supplies. That would explain why flash disks/keyboards/mouses etc are all OK – they require very little signal/power to operate and different enclosures require different power/signal levels which is why mine produced different results but all far lower than expected. Whether this is true of all MSI boards or just some where there are minuscule differences in power of signal delivery we’ll never know (MSI will but of course they’ll never say).
I certain I can now purchase a PCIE card to plug into my front panel and get my transfer rates back.