Laptop fallen - Need help to diagnose HDD

My laptop fell while running around 2 feet. Here's what happened then.

1. Started my Laptop, didn't show GRUB, simply said HDD not found.
2. Restarted my laptop, showed grub, started Ubuntu, stuff kept crashing, I shut my computer down.
3. Restarted, back to no GRUB and HDD not found
4. Booted from Ubuntu live CD, My 640 gig drive is visible. I can mount it and read write stuff.

What do you think has happened? I've shut my computer down, I need some quick solid advice on how to diagnose my hard disk and find out if I can get it repaired or buy a new one.

This is a physical issue. You can't poke at it with software to discover the problem.

1. Back up everything important. [you should have already been doing this, btw]
2. Take your laptop apart and make sure nothing is broken or had come unplugged.
3. Reassemble your laptop.
4. See if it's back in working order.

You could also simply skip steps 2 through 4 and take it somewhere for service, but that just means you're going to be paying someone else to do those same 3 steps.

Does your hard disk have an accelerometer?

First, the (potentially) good news--depending how you look at it: some newer hard disks have built-in accelerometers which automatically spin down the drive and park the heads if they sense that the computer has been dropped. If you're lucky enough to have one of these, then it's possible that the data on the drive is safe, but something else is damaged--for example, maybe a connector was knocked loose or a solder point on the motherboard is broken.

SMART diagnostics

You can diagnose whether the disk is damaged by booting from a live Linux CD and running SMART diagnostics. Either boot from Ubuntu then install and run smartmontools, or boot from a CD that already has bundled SMART diags, such as PartEd Magic. Before you run an extended test, you should first check the SMART log for reallocated sectors and pending sector count. Realloc sector count indicates the number of sectors which are unwritable, and pending sector count indicates the number or sectors which have been unreadable and which will be marked as reallocated if subsequent writes to those sectors fail. Regardless of the results, try recovering as much important data as possible, as quickly as possible.

After you've gotten your data off, run an extended SMART test. For good measure, you could also try running badblocks -nvs, but it's probably not necessary. If the SMART diags report any reallocated sectors, throw the disk away and restore your data onto a new one (preferably an SSD; see below).

dd to the rescue

I'd highly recommend using dd (or ddrescue/dd_rescue) to copy as much data as possible before running any other, more disk-intensive data recovery tools.

DON'T run SpinRite except as a last-ditch effort

I'm sure someone here will recommend running SpinRite, but that is bad advice if you haven't already exhausted all other options. Only run SpinRite as your absolute last resort after you've already exhausted all other options within your budget. If your data is important enough to spend potentially thousands of dollars sending the drive in to a data recovery firm, don't run SpinRite at all. Why not? Because if your hard drive has not catastrophically failed, it could be in a state of continuing degradation. SpinRite's aggressive seeking and reattempts at reading the same sectors over and over again as it tries to divine the data that used to reside at those addresses could end up doing more harm than good.

Your new hard drive should be an SSD

When you replace the drive, consider an SSD. SSDs are extremely fast, are now priced within the reach of most computer-savvy consumers (though they're still more expennsive than magnetic media), and--most importantly--they are more or less shock-proof compared to conventional magnetic media.