I would keep 4 drives in raid 10 (user DBs, tempdb), use 2 drives in raid 1 (logs), maybe use the other drive to mirror the OS drive (assuming it is currently a single disk) or as a spare. That way the system can survive any one drive failure. An alternative might be to use raid 5 instead of raid 10 if there is very little write activity, but given the size of the databases I would expect the data to be largely cached anyway.