The two drives are exactly the same speed.

The ST340016A is a 2 head drive, it uses both sides of the platter.
While the ST340014A is a 1 head drive, only using one side.

If the physical layout of sectors per track were identical, then
the 14A could potentially be half the speed of the 16A. Although
optimization in the cache to take advantage of 2 heads is not
perfect.

I briefly looked at the specs on the Seagate web site. The drives
are the same rotational speed, 7200 rpm. And the 16A has a slightly
higher BPI (bits per inch) 595,000 vs 540,000 -- so you get 10%
improvement there.

So depending on how efficient the cache is I'd estimate the 14A
to be 60-70% of the speed of the 16A.

You're neglecting the fact that the data on the 80G/platter drive is packed twice as densely as the data on the 40G/platter drive. I'd guess the newer drive would be at least as fast as the old one, if not faster because of the increased areal density.

The TPI is twice (94,600 vs 58,000), but that just means the tracks are
closer together, which does help seek time. However, single head can
only read one track at a time, and the density of each track (BPI) only
increased 10% (595,000 vs 540,000).

Since the heads on a dual-sided platter don't move independently, only one track at a time can be read there as well. The heads have a lot less ground to oover on a single-surface drive. The increased track density should also lead to lower latency as well.

All in all, the two drives are about equivalently spec'd.

Independently moving heads? You might be thinking of the old Connor drives
that did that.

Temperature skew is the only reason multiple heads cannot be simultaneously
used. And that's only an issue for drives with many platters. Definitely not an
issue when using the 2 sides of a single platter. The email response I got
back from Western Digital confirms it:

"Multiple read/write head are used to increase performance
of read and write access as they (all heads) have the ability
to read/write on a specific data surface at the same time."

Haven't heard from Seagate, but I assume they do the same optimization.

Also in this case, having the tracks closer together doesn't change the
track to track seek time, because most of that seek time is resync'ng the
servos, not moving the heads. So the big win is reading large amounts of
data sequentially, like when defrag'ng, because with a 2 head drive you
can read twice as much data before moving the heads.

Thanks for the replies everyone!

Drive still seems extremely slow as compared to the original.  For example, when I go to the defrag window, click "defragment", it's a full 20 seconds before the drive even starts its analysis (green progress bar), and then the actual defrag is very slow.  This is much slower than before, and for a 1.8 P4/512  RAM system, seems pathetic.

I compared two computers from work. One is a P4 1.5/256 RAM, other was a P4 2.6/512 RAM and they were both much faster (WD caviar).  If it's not the HDD, what explains this unbelievably long lag to start disk defragmentation?

Thanks again for the help.

Jon