OK, you have two fixed IP addresses 10.0.0.2, your server, and 10.0.0.56, your NAS. Can I assume that you use a managed switch to oversee the network and if so what do its logs say about the traffic between server and NAS?
Another thing to look at any updates you have made to the server. Is there anything that might interfere with CIFS/SMB? We don't use windows here (our servers run WSeB and workstations are a mixture of WSeB client and OS/2 Warp 4, although we do have a VM win 7 we can use for testing).
My service manager said that with those sized daily backups you should be using a high speed storage array rather than a simple NAS box. Our 24 workstations might, on the end of the month 'create all the paperwork', generate about 50 MB of incremental backup so I have to ask out of curiosity how are you managing your backups. That leads to another thought - is there something you are trying to backup that interferes with the CIFS/SMB implementation on your server? Either the server logs and/or the switch logs should show what point the data transfer fails. Also, if Cobian Backup has logging you should enable that as well.
Sorry I can't be more helpful.