Unicode oracle issue - Wrong results

We have a strange situation. We have moved to E810 with all business data in non-unicode and system table are in unicode. Upgraded to latest version of oracle recommended by PSFT. The problem is when we do a SQL in any table within the system table on a string index, it just returns the first row. For example if we do the following SQL select * from F9860 where siobnm in ('F0911','F0411'), it just returns F0911. I think it has to do something with the unicode character set. Has anyone ran into this problem. We do not have any issues like this in business data which is not unicode. I would appreciate any help in this case.
Thanks

We have the same issue. We are currently on Oracle 9.2.0.4. PeopleSoft said to upgrade to 9.2.0.5. We are still in the process of doing this.