> Is there a way for HT Track to download the ACTUAL file 
> rather than generate a HTML one with a link?
No - because HTTrack names HTML files *before* downloading 
them, potentially creating numerous links to this file. And 
when the redirect is seen, it is too late to back-patch 
existing references.

I know this is probably the most annoying limit in httrack -
 not being able to handle on-the-fly redirects, or MIME 
typing changing (such as an .html producting a GIF file). 
But changing that would require some hard work, because the 
stack/parsing system is organized to be independent from 
the downloading stack (downloads are processed 
in "background", in separate socket pools)