Interesting. Do those card revisions have the same PCI ID?
Which actually reminds me: You could easily create a patch to the pci.library so that it'd report any NEC USB card as having the same PCI ID as Spider 2. Assuming the chipset is the correct one it'd work. This way the spider driver would stay untouched and RDB destruction wouldn't happen either.
findpciclasscode $c0320 (uPD720100A USB 2.0 Host Controller)
(d0 = devicestructure)
compare vendorid $1033 = NEC, ok go on
compare subsystemvendorid $0e55 -> write $0ee4 to the structure
compare subsystemid $2928 -> write $089e to the structure
