Let me try it another way.. with a little education on how TSM is different..
TSM does not have a concept of a "full backup" when you're doing file system backups. This is one of the fundamental key differences that seperates TSM from other products. Therefore in a way, you're trying to fit a square peg in a circular hole in your line of thinking. Lets talk briefly about how TSM works differently..
Most backup products deal with Full/Incremental/Differentials and your retention is based on when those were taken and how many of them you wan't to keep. So it makes perfect sense to do what you're asking with other products. You might want to keep your weekly fulls going back 6 months but throw away your daily incrementals (which would obviously mean any version of that file backed up to an incremental would potentially be lost).
In TSM, file system backups are *always* incremental. The first time you backup a server, it backs up all the files since they've never been backed up before. Every single subsequent backup it only backs up files that have changed. There is no need to backup files that haven't changed because the system already has them. It manages what needs to be backed up and what doesn't in it's database. Therefore it consumes far less tape in the long run.
Retention in TSM is also quite different. It's more granular. Instead of retaining based on F/I/D backups, you retain at a file level. You decide how many versions of a file you want to keep and for how long (any any combination of the two). This can be quite powerful. You can easily retain .MP3 files differently then .DOC files and so forth. You can also retain whole filesystems, paths, regular expressions, whatever differently with a high degree of control.
What does this do to the tapes where the data is ultimately stored? Well first of all it really opens up options on where TSM stores the data. Which tape? Where on the tape? Intermixed with other clients or not? This is quite configurable but by default it generally keeps each client's data together and ideally on it's own tapes since that's better come restore time.
The other thing it does to data on tapes is over time as certain files get expired, it creates "holes" in the tape. Initially when you first backup a server, it may fill an entire tape. Down the road, the system likely expires some of those files on that first tape but not all of them (no need to, they haven't changed). This is where concepts such as "reclamation" come in where the system reclaims a tape once the tape reaches a certain percentage of utilization. Intially it's 100% utilized. Over time that drops.. When it reachs whatever reclamation threshold you set, the system can take the "good" data still on that tape and copy it over to a new tape so you can reclaim the portion of the tape not in use anymore.
In your case, if you find your robot filling up but you have ample blank tapes, you have three solutions..
One, check out some of the FULL volumes from the library and put in scratch (blank) tapes. If a restore request comes along that requires a tape that is not in the library, an operator will need to mount the particular tape.
Two, adjust your retention settings to something more realistic that coincides with your need/desire to have ALL backed up data in the library and the library's capacity. This requires some evaluation on your part in determining what's sucking up all the space and making appropriate policy changes.
Three, buy a larger library.
All in all you'll find it very powerful and best to let TSM manage your media for you. You do need to learn the TSM concepts of retention and reclamation which is why i highly recommend you go educate yourself. There is terrific material available. Read my FAQ on here on how to learn TSM as a starter. It's my guess that you're not doing your homework on why you're consuming all that tape. With a little homework and understanding of how TSM retention works, you should easily be able to reign in the consumption to what you want. Most people are lazy (or unknowledgable), however, and get themselves into trouble by not adequately setting retention policies and/or architecting a proper system to handle their backup and restore needs.