It is actually a problem that’s encountered even with 8434 serial-to-USB adapters occasionally but certainly the incidence of the issue is lower than with the CP210x chipset of 6510USB. So it’s more a question of degree than an absolute distinction. I believe that Davis might have been looking at alternatives to that USB chipset a year or two back when the noise problem was first recognised, but that every option they looked at still had some degree of vulnerability to the same problem. The USBXpress option does seem to be somewhat more resistant to the problem than the VCP (Virtual COM Port) approach and I guess it was felt that USBX was just as good an option as completely re-engineering the logger.
There’s no doubt that this is a complicated problem, since noise/RFI can take many forms - through-space (ie spurious wireless transmissions) as well as through the mains supply. Adding extra ferrite cores is probably the single simplest measure that’s likely to have any benefit but won’t resolve every situation. Using a good UPS on the mains supply may help. But often trying to identify and remove the source of the interference can be quite effective too.