I have two old disks both of which I've largely stopped using after Windows installs died on them. My plan was to recover what I could from the drives, and see if there was an area with no bad blocks that I could put a little partition for installing Linux onto (or, at the very least, a persistent disk image and settings save for knoppix). The problem is not only that bad blocks is reporting a lot of bad blocks, but that the number of bad blocks it reports changes every time I run it. I also get Input/Output errors when i run the scan.

From what I can tell, it seems that the drives just get to a certain point and stop working. Everything after that is seemingly reported as a bad block. For example, I ran badblocks on '/dev/hdb1' (the second of the two drives below), and got a similar result to the first. This time I ran it just after the other, and it reported every block as being bad. It must've done the test in a fraction of the time it spend on the other test, however, as the first test was running all day and was still going when I went to bed. Both were completed when I got up in the morning.

Here are the outputs from the scans:
root@Knoppix:/media/sda2# badblocks -svnf /dev/hdb2 -o "bb-b2.txt"
Checking for bad blocks in non-destructive read-write mode
From block 0 to 79184385
Checking for bad blocks (non-destructive read-write test)
Testing with random pattern: badblocks: Input/output error during ext2fs_sync_device
badblocks: Input/output error during ext2fs_sync_device
badblocks: Input/output error during ext2fs_sync_device
badblocks: Input/output error during ext2fs_sync_device
badblocks: Input/output error during ext2fs_sync_device
badblocks: Input/output error during ext2fs_sync_device
badblocks: Input/output error during ext2fs_sync_device
badblocks: Input/output error during ext2fs_sync_device
done
badblocks: Input/output error during ext2fs_sync_device
Pass completed, 15376236 bad blocks found.
root@Knoppix:/media/sda2# badblocks -svnf /dev/hdb1 -o "bb-b1.txt"
Checking for bad blocks in non-destructive read-write mode
From block 0 to 80895276
Checking for bad blocks (non-destructive read-write test)
Testing with random pattern: done
Pass completed, 80895276 bad blocks found.
To me this is indicative of a more fundamental problem. Perhaps some other hardware? The thing is that it reports seemingly randomly with both drives, so I'm wondering if it's the motherboard's controller that's a bit screwy. After that test, it's giving me an error when I try and mount the drives:
root@Knoppix:/media/sda2# mount /dev/hdb1
Error reading bootsector: Input/output error
Failed to startup volume: Input/output error
Failed to mount '/dev/hdb1': Input/output error
NTFS is inconsistent. Run chkdsk /f on Windows then reboot it TWICE!
The usage of the /f parameter is very IMPORTANT! No modification was
made to NTFS by this software.
See, the instructions to run something from windows are all nice, but windows isn't working for some strange reason.

I'm thus wondering: Do I get a new Hard Drive? If I do, do I bother going for SATA? Do I steer clear of the motherboard's HD controllers completely just in-case they're all screwy? Right now the 'sda' device that I'm using to store the check results on is a Flash drive, a VERY large flash drive at that, but it's very slow, sometimes writes blank files (under windows, not had that happen under Linux yet), and came with no warranty for about the amount I'd expect a drive 1/4 of the size to cost.