Mario,
One would be hard-pressed to understand how RMAN datafile backups or archivelog backups can cause "log file sync" on foreground transactions.
Except for updating the data dictionary views, RMAN does not perform transactions. While RMAN itself does not generate redo to affect the LGWR or affect redo generation in any way, it may be possible that RMAN is signaling LGWR in some manner, interrupting the flushing of redo from memory-based Log Buffer to storage-based online redo log files. If the latter is the case, such as waits on the RMAN sessions, then we should be able to find evidence of that, but first things first, and let's start at the beginning...
It would be helpful to drill down into the observed sessions experiencing "log file sync" waits, and display what SQL statements are experiencing the wait, display what user is executing the sessions, and begin to understand what is actually happening and what is actually affected. It is never enough simply to observe waits without understanding their context. When you find who is waiting and on what SQL, can you share that?
Also, can you share the MOS/Metalink note IDs which indicate that RMAN "disturbs" the database and causes "log file sync"?
Looking forward to getting to the bottom of this...
Thanks!
-Tim