No question FIRST and foremost you want to be able to have good backups. You're working on that now.
You're getting somewhat bogged down in the details and nuances of trying to boot with an alternative boot CD, that's interesting as a learning experience, but try to keep the main thing the main thing, getting your stuff safely backed up.
It's possible that during the backup you may encounter a hiccup or two, if it hits upon some files that are affected by these unrecoverable errors. Ideally the backup technique you choose will allow you to LOG those errors and continue to backup all the rest and not merely "error out" and stop.
If you have such "error out" problems trying to copy or backup your data, check the tools options. For instance, gHost disk imaging software has options settings for whether to ignore read errors and continue trying to clone/backup image. XP will likely "error out" if you're merely trying to "copy and paste" from one drive to another. Look into Microsoft's ROBOCOPY command line tool and the command line options switches, it can log, it can even copy the modification times of folders. You may want to tweak the retries on error setting though, so it does not waste 2+ hrs retrying 2000 times on one/each bad file. Another alternative for future copying drive to drive is SynBack which has a nice GUI
You mention a boot attempt only "sees" 2 of 3 drives. Sometimes the bood CD is seen as drive "C" itself sometimes as drive "A", so it's going to depend. Clearly if the failing drive is not even seen that's going to be a problem. What you NEED to be CLEAR on is what your actual situation is. Do you have three (3) physical drives in your PC, or is it one physical hard disk patitioned into locgical drive letters C, D, and E? or some variation?
Some of what you need to do DEPENDS on what resources you have available. If you only have the ONE PC, then booting from a utility boot CD like a ultimate boot CD or Bart or what have you may be needed. On the other hand, if you have another PC that you can connect the "patient" drive onto as an auxillary / secondary drive, perhaps you can install and make use of backup and recovery tools on that "test bed" machine. Being careful of course, not to mix up which drive is the "operating room" machine's own drive(s), which drive is the patient, and which drive is the drive you're going to backup onto.
Obviously it's an important step to have a good BACKUP of all your stuff, all the drive letters that have irreplaceable stuff on them. You can backup all your stuff to external drive(s), which is probably a good idea anyway. If you have to, do you have what you need to reinstall Windows? all the programs you have, because if the drive cannot be saved sufficiently you may be doing that.
If you hard disk makes some click click noises, or ca-chunk ca-chunk, it's probably retrying multiple times on bad spots, it sends the heads to track zero park and then back trying to re-align the track position. BUT if your drive makes AWFUL terrible scraping or grinding noises, STOP. STOP using it and consult with experts. If the surface of the srive is scraping, an actual physical crash has ocurred, then the scrapings and filings inside are like boulders compared to the microscopic space between read/write heads and platter surface thanks to the Bernoulli effect, and will result in more scraping and shaving which would be obliterating the information on the drive surface. So then, you'll have to consider whether you risk further damage, or if you want to send it to a data recovery laboratory, such as gillware or ontrack data recovery.
But then as far as a hard disk failing, you want to know what and which one(s) you're dealing with. Hard disks are so cheap, you'll probably want to straight-up replace the drive that's failing. So you need to figure out what physical hard disks (vs logical drive letters) are in your PC, and what kind, are they PATA133 parallel ribbon or SATA, and what is the MAXIMUM hard disk size your PC BIOS supports (is it an older PC that cannot support hard disks bigger than the LBA limit of 127GB or will it or does it need a BIOS update)
My suggestion is you go purchase a suitable replacement hard disk (or if you have a "spare") and first "SCRUB" the destination drive, (that is thoroughly format it and check it for bad sectors and then erase it (don't do the wrong disk obviously)), and THEN use cloning software like Acronis True Image, Symantec Norton gHost, Paragon Drive Copy, or Runtime Drive Image, free Easeus Disk Copy, or free Clonezilla (I suggest Clonezilla) to try to duplicate your failing drive onto the new drive. Again, you may need to set options to ignore errors and keep on cloning as best as possible.
If the errors are significant and too numerous it's possible the resulting clone may have "holes", not boot and run Windows 100% properly. If that happens, you can then try SFC system file checker command and/or try a repair install of windows overtop (cross that bridge if you come to it)
Of the various ways your hard disk could be failing, bad sectors is a very likely, most likely culprit. I would NOT use madunix's suggestion of file recovery software, (an exhaustive and useful list) not YET, because that's for recovering files when the parition "table of contents" (file allocation table, master file table, inode table, etc) is corrupt or lost or files are accidentally deleted and have to be pieced together from sectors. What you need to do before that is figure out what the nature of the problem is. Are bad sectors are the cause of the problem, or if it's something more serious like physical crashing, the circuit board is fried, the spindle doesn't rotate to proper speed, the armature inside won't propely move/align, or what.
The hard drive diagnostics can go a long way toward telling you what is wrong. They are manufacturer dependent, you need to use the appropriate diagnostics for the manufacturer of the drive in question (and sometimes model, IBM deskstar/HItachi Ultrastar have their own) because if you use say Seagate's seatools on a Maxxstor drive the results will not be true or reliable or conclusive.
http://www.tacktech.com/display.cfm?ttid=287Be CAREFUL, some of these tools are command-line-ish and it can be difficult to tell the difference between NON-destructive scans and desctructive scans (a destructive scan would scrub / erase your drive's contents)
Another thing is the S.M.A.R.T. This is a feature of hard disks that keeps a sort of quality control logbook. Whenever the temperature of the hard disk is excessive, or anytime the drive fails to spin up to speed in a normal amount of delay or there are excessive bad sectors all-at-once, or the "spare" area for bad sectors is almost full, it raises a red flag. PC bioses can check the smart status druing the POST (power-on self test) when the computer if first turned on, and if the drive is bad, a good BIOS will stop and issue a warning. The disk diagnostics will check it too. Alternatively, enterprise-class OSes will monitor it. Alternatively, here's a little viewing tool
http://www.beyondlogic.org/solutions/smart/smart.htmBad sectors do happen, it's normal for drives to have bad sectors over time (think of little drop-outs on magnetic audio casette or VHS tape). But what to do about it? Drives by default always have "spare" sectors set aside for bad sectoring. Well, an expensive enterprise class disk has ECC error correcting redundancy and automatically deals with it. A RAID configuration also has redundancy so can deal with it. A consumer-level regular drive by itself doesn't. The problem is, if you used chkdsk with the /R option (or scandisk with the fix bad sectors box checked, same thing really) they do a LOWSY job of recovering the data off the bad sector(s), they "give up" after only a half dozen retries or so. Yes, they will set aside the bad sector forever, and they will reassign the bad spot of the affected file to a "spare" sector, and if it was able to read the data off the bad spot and write it to the spare, great, BUT if it did not, you now have a bunch of zeroes "holes" in the files and those files are essentially "corrupted". You would then be expected to recover from your last known good backup all files affected by chkdsk /r / scandisk bad sectoring. If you have good backups you can pick and choose what files to recover, great, if not, great big UH OH, those are corrupt. If they are files that are part of Windows or a program maybe you can reinstall those files. If they're you're data, you're really need to have had a backup.
If the bad sectors are on empty space, no problem. If the bad sectoring is done when you are formatting or reformatting, also no problem. If it happens to your data and you have backup, smallish problem of recover affected files from backup.
In the bad old days, bad sectors happened so often, that was actually normal procedure: Of course you did your backups often, verified them from time to time, and once a year (or more often in the really old bad days) you completely "scrubbed" your hard disk for bad sectors and then did a full recovery from backup. This also had the side-benefit of fully defragmenting.
If bad sectors affect your files and you DON'T have good backup, what about that BIG problem? Well, there is a possibility. There is a program called SpinRite 6 by Gibson Research
http://spinrite.com or a similar tool by a Russian developer called HDDRegenerator. What it does is it retries literally hundreds of times to read the data off the bad sectors, using signal strength analysis, data sampling and statistical comparison algorithms, it does it's best to recover as much of the data as possible off the affected sectors, often recovering most if not all 100% of the data off bad sectors. Interestingly, it also can be used as a PREVENTATIVE measure, setting asdie "iffy" sectors before they become bad, so that your data is not lost and prolonging the useful life of the drive.
chkdsk /F is still useful for fixing the integrity of the file allocation tables. Sometimes, when the FAT/MFT of the partition becomes corrupted (there are two copies for redundancy but it can still happen that the file allocation table is corrupt) it's because there were bad sectors in the all important track zero or on sectors where the FAT/MFT table of contents was stored. Hopefully SpinRite or HDDRegenerator can recover it, and you're ok. If unfortunately you used the /R option of chkdsk and it FAILED to recover the data off the bad sectors, well now you're a mess because the bad sectors have now been set-aside.
That's when some data recovery tools, like the list madunix posted, can come in handy. What those do, is look at the sequences of blocks and sectors that have file content in them, attempting to reassemble the files from those elemental file chains. It's the sort of thing the data recovery laboratories also use if the disk has crashed. However in those cases they open and disassemble the drive in a specal clean room, you know, the guys in the space bunny suits, carefully clean any scraping/shavings, transfer the platters to a specical expensive piece of equipment, program in the track/block/sector geometries, copy the data as best as possible to a temporary drive, and then try to piece together your files from the resulting pieces on the target.
Phshew
Okay, so, what actual drives and partitions do you have. Try to ascertain the nature and seriousness of the problem. Very important to get backups however possible. Then, proceed accordingly.
Hope this is helpful. Good luck.
