Your sandbox is just a plain folder. Its the hidden CVS folder in the root directory of your sandbox that contains information as to what was checked out, and where things have to be checked back in that CVS reads to figure out what to do when you commit.
Deleting this hidden CVS folder effectively detatches everything in your sandbox from CVS control. Doing another checkout on a branch in the same sandbox will recreate the CVS folder, and the files that are checked out will overwrite any existing files. But, this is what you would end up doing if you just created a second sandbox on your workstation for the branch copy. Using a another sandbox for the branch keeps things clean, and you can be sure about what the contents of the new sandbox are.
