Exactly. Redundancy == Even distribution of data
No matter what you start with, you would have to "create" drives of the same capacity. Whether you buy them that way, or partition them, the result is the same.
(Reminds me of another RAID / LUN story...)
The concept of mixing and matching drives is not new, and solutions were in place to work around mis-matched drives.
Again, I still consider it a work around. But if it's a non-critical machine, and you want to try something out...play a little.
Sometimes I should put a disclaimer in my signature:
Plenty of ideas are fun to try on your desktop, or "project" machine.
Risk of data loss is quite high when tinkering.
Don't try this on production servers.
That particular incident involved a High-Availability server (RDBMS) that needed more space, so I had to expand the array while the system was running, while also moving to 4GB disks -- gotta love Unix and SCA SCSI. When the office shut down for a holiday a few months later, I replaced the older disks with the new 4GB disks (backups, recreate arrays, restores, etc. )
I agree with the backup suggestion. I do that myself.
You are better off relegating the older, smaller drives to lighter duty. Synchronizing folders is really handy.
Another thing to consider:
Windows allows you to "mount" a volume to a folder, like Unix. So you can mount one of those older drives to the "C:\Data" folder, and you will have that entire drive for your data, while your operating system and programs are on your newer, faster drives.
It "feels" like one drive since all paths start with "C:" but physically they are separate and you can use more of your total disk space.
If you are just looking to expand capacity, I would give this a try.